Report No: AUS9112 The Labor Impact of Lao Export Growth February 2016 Currency Equivalent Exchange rate effective as of February 24, 2016 (from BCEL) : Currency Unit = LAK (Lao Kip) LAK 8,110 = US$ 1.00 Fiscal Year = October to September Acronyms and abbreviations ASEAN Association of Southeast Asian Nations EAP East Asia and Pacific EU European Union GDP Gross Domestic Product LDC Least Developed Country LECS Lao Expenditure and Consumption Survey SME Small and Medium Enterprise SOE State Owned Enterprise TDF-2 Second Trade Development Facility WTO World Trade Organization Acknowledgements This report was prepared by Elizabeth Ruppert Bulmer – Lead Economist, Jobs Cross-Cutting Solutions Area, and Claire H. Hollweg – Trade Economist, Trade and Competitiveness Global Practice. The authors are grateful for comments received from Julian Clark – Senior Economist, GTC02; Mombert Hoppe – Senior Trade Economist, GTC02; Thomas Farole – Lead Economist, Jobs Cross-Cutting Solutions Area; and Lars Sondergaard – Program Leader, EACTF. We would also like to thank the Government of Lao PDR, especially the Ministry of Industry and Commerce and the Lao Statistics Bureau for data access and collaboration. This report was prepared as a part of a trade-related analytical program of work carried out by the World Bank in Lao PDR, under the Second Trade Development Facility program, and financed by Australia, the European Union, Germany, Ireland, and the United States. Disclaimer: The findings, interpretations and conclusions expressed herein are those of the authors, and do not necessarily reflect the views of the World Bank Group, its Executive Directors, or the governments they represent. Regional Vice President : Axel van Trotsenburg Country Director : Ulrich Zachau Senior Practice Director : Cecile Fruman Country Manager : Sally Burningham Practice Manager : Mona E. Haddad Primary Authors : Elizabeth Ruppert Bulmer Claire H. Hollweg Table of Contents Executive Summary ......................................................................................................................... i Chapter 1: Introduction and Trade Context .................................................................................... 1 1.1. Introduction .......................................................................................................................... 1 1.2. Trade Context ....................................................................................................................... 2 Chapter 2. Lao PDR’s Labor Market and Recent Employment Trends ......................................... 6 2.1. Characteristics of the Labor Force ....................................................................................... 6 2.2. Sectoral Employment Trends ............................................................................................... 9 2.3. Productivity and Average Wages ....................................................................................... 12 2.4. Determinants of Earnings ................................................................................................... 16 Chapter 3. Labor Sophistication and Competitiveness in Export Markets ................................... 18 3.1. Snapshot of Lao PDR’s Export Basket .............................................................................. 18 3.2. Measuring the Labor Content of Exports ........................................................................... 19 Chapter 4. Export-Employment Elasticity .................................................................................... 29 4.1. Comparing Exporting Behavior and Labor Demand ......................................................... 29 4.2. Estimating the Employment Elasticity of Exporting.......................................................... 32 4.3. Impact of the Business Environment on the Export-Employment Elasticity..................... 33 Chapter 5: Labor Mobility and Adjustment Costs ........................................................................ 35 5.1 Defining Labor Mobility Costs and Labor Adjustment Costs ............................................ 35 5.2. Worker Transitions in Lao PDR......................................................................................... 36 5.3. Labor Mobility Costs ......................................................................................................... 39 5.4. Labor Adjustment Costs ..................................................................................................... 42 References ..................................................................................................................................... 47 Annexes......................................................................................................................................... 50 Executive Summary 1. Lao PDR has seen strong economic growth during the past decade with an average growth rate of 7.8 percent per annum. Much of this has been driven by the growing external appetite for Lao PDR’s natural resources, which have not generated broad-based job creation and income growth for a significant proportion of the Lao population. Manufacturing – and especially garment manufacturing – has the potential to generate jobs for the large unskilled segment of the Lao labor force. The garment sector did indeed provide substantial job creation in the 1990s, especially for women, but more recent trends suggest a relative contraction of the manufacturing sector since 2007/08, both in terms of employment and weaker wage growth. This contraction is mirrored in Lao PDR’s manufacturing trade outcomes o ver the past decade, pointing to a decline in competitiveness on the international stage. While exports can be an avenue to growth and job creation, the success of policies aiming to sustainably improve employment outcomes will depend on the realities of Lao export competitiveness and global market demand. 2. Based on a series of complementary methodological approaches and datasets implicitly linking trade to jobs, this report paints a mixed picture of Lao PDR’s recent export performance and how this has translated into job creation and improved job quality across the economy. The findings of these various analytical approaches provide insight into the potential drivers of Lao PDR’s labor outcomes including the role of exports, and what the Government and firms can do to better position Lao producers in external markets so that more and better jobs are created for Lao workers in the future. 3. Whereas the economy is undergoing some structural transformation away from agriculture, workers are primarily shifting into low-skill non-tradable service activities and there is no strong concurrent rural-to-urban shift. Agricultural employment continues to dominate the Lao labor market, but agricultural production accounts for only a small share of Lao exports. The most dynamic export sectors – mineral products, base metals, and wood and pulp – are not generating sufficient labor demand to absorb the large unskilled segment of the labor force. Mining wages are higher than other sectors, controlling for other factors, but mining activities are not labor-intensive. Even the garment sector, which accounts for a tenth of exports, employs less than 2 percent of the labor force, pays very low wages, and experienced slower real wage growth than the rest of the economy (except for hotels and restaurants). 4. Manufacturing in Lao PDR is more export-oriented than average compared to other countries, with export rates similar to Vietnam. Exporting firms tend to be larger than non- exporting firms, both in Lao PDR and elsewhere. However, comparing Lao PDR’s 2012 export basket to countries at a similar level of development, Lao exports tend to be less sophisticated than those from Senegal, Cambodia, Indonesia, Bangladesh and the Philippines, based on indirect (global trade-weighted average) measures of wage, value-added, and human and physical capital. In the case of garments, which have a high labor content, Lao garment exports may in fact be less competitive due to higher wages vis-à-vis Bangladesh and Cambodia. Lao PDR’s small scale of garment production also makes it less well placed to compete for large contracts in certain markets compared to Cambodia and Bangladesh, even if Lao producers have i found a competitive niche in European markets. In dynamic terms, the labor sophistication of Lao PDR’s export basket has remained relatively flat since 2006, and made little progress in closing the competitiveness gap vis-à-vis comparator countries. This suggests relatively low-tech production methods and/or weak education outcomes. These factors in turn limit exporters’ capacity or incentives to invest in more productive technologies. 5. A key obstacle to attracting investment in high-value manufacturing is the Lao labor force’s low schooling levels. With respect to wages, the Lao labor market rewards higher education levels with higher wages, although we find evidence of declining returns to each level of education between 2007/08 and 2012/13. For those with a tertiary degree, for example, the premium relative to less than primary fell from 58 percent in 2007/08 to 39 percent in 2012/13, suggesting that there is increasing supply of skilled labor to meet current demand (this may be partly driven by civil service wage distortions). This also translates into relatively lower incentives for workers to achieve higher education levels. This trend contrasts with other dynamic economies where the diversification into higher value products and the adoption of new technologies require more and more top skills and therefore generate increasing returns to the highest levels of education. 6. It is notable that in the garment sector in particular, wage growth for those with a vocational degree was much faster than the average for all garment workers (14 percent vs. 2.6 percent annual growth). Whereas this implies increasing returns to semi-skilled workers, consistent with reported shortages of skilled labor, the slower wage growth for unskilled or entry-level garment workers may explain why employers have difficulty retaining workers (factories report high turnover rates of new entrants, most of whom lack a secondary or vocational degree). This either means that employers are actually willing to incur the repeated recruitment and training costs of their rapidly changing unskilled labor force, or that profit margins are too small to accommodate higher wages for this majority segment of the garment sector workforce. 7. Data on worker movements between jobs and into and out of different sectors suggest that the Lao labor market is relatively fluid, but workers seeking better work opportunities in other sectors typically face significant costs to transition. High labor mobility costs are the norm for countries with a large share of primary employment, and Lao PDR is no exception. Skilled workers face relatively lower mobility costs than unskilled workers, while female workers face higher transition costs compared to men. The agriculture sector has the lowest entry costs, and as such acts as the “sector of last resort”. Policy Recommendations 8. Policies to facilitate better labor outcomes could focus on improving existing firms’ or producers’ competitiveness, helping firms to enter and/or grow, facilitating access to technology and higher value production, and enhancing the skills in the labor market to attract and induce higher-value activities. 9. Small profit margins in the low-value garments sector mean that Lao producers have little room for finding productivity gains vis-à-vis competitors such as Bangladesh and Cambodia. A more promising strategy would focus on creating more attractive niches or product lines for high ii income economies through e.g., upgrading labor standards and environmental standards, thus helping big brand retailers and buyers to meet their corporate social responsibility objectives. 10. High turnover and exit rates from garment factories back to farm work or migration results in productivity and production losses for the Lao economy. Policies to reshape the incentive structure of garment jobs (such as working conditions and/or career development) could entice workers to remain in relatively more productive work. 11. A large pool of skilled and competitive labor is an essential component both for attracting investment in higher-value activities, and for enabling firms to upgrade the quality of their output. This means policies to increase educational attainment, more support for basic education, curricula reform to meet market demand including soft skills and entrepreneurial skills, and aligning vocational training to private sector demand. 12. Exporting rather than selling to the domestic market is correlated with more skilled and unskilled employment, although the effect is quite small. Policies that help producers access new external markets could include joining (electronic) knowledge platforms, improving standards compliance and certification in line with international standards, reducing trading costs by improving trade logistics and trade facilitation systems, and providing enterprises with support for marketing assistance and training, for example. 13. Whereas labor regulations do not seem to be a binding constraint on hiring, there is anecdotal evidence of confusion surrounding the implementation of the new minimum wage regulation which could diminish labor demand, suggesting a need for clarification and outreach by the Ministry of Labor. In addition, other recent World Bank reports1 identify critical business climate challenges that present larger impediments for firms trying to export, and that should be addressed. 14. Lao’s relatively low-tech export basket limits its capacity to access higher income consumers seeking higher quality goods and services. Policies should therefore focus on facilitating access to technology and higher value production through for example more favorable (but still non-distortionary) investment incentives, or more generally by easing access to formal finance through banks or SME-targeted credit institutions. 15. The high labor mobility costs in Lao PDR prevent efficient labor allocation across sectors, and could generate significant welfare losses in the event of a sudden change in the economic environment by slowing and dampening the labor market response. Policies to facilitate labor mobility could include labor market information or matching services, or cost- share training that is demand-driven and targeted to specific private sector needs. 16. Diversifying production and employment away from the primary sector toward industry and services will also reduce labor mobility costs, enabling workers to transition more easily to better jobs. In addition to policies directly targeting lower mobility costs, there is scope for reforming distortionary public employment hiring, compensation and human resource management policies, and revisiting the role of SOEs. Reducing these distortions could curtail queuing for public sector jobs and free more educated workers to enter the private sector. 1 See World Bank 2011b, World Bank 2014a, World Bank 2014b, World Bank 2015 iii 17. The limitations of Lao PDR’s manufacturing sector – both in size and quality – mean that any positive policy impacts are likely to have only modest effects on aggregate welfare and inclusive growth. Policies will need to go beyond the manufacturing sector to address the limited employment and demand spillovers of mining production, and the limited market integration and value addition in agricultural production. Policies to enhance agricultural productivity, job quality and economic opportunities in rural areas will not only foster greater equity, but will also reduce urban-rural distortions and dampen the incentives to migrate. iv Chapter 1: Introduction and Trade Context 1.1. Introduction 1. As countries become increasingly integrated into the global economy, increased trade links with other countries translate into increased access to better or cheaper imports and increased demand for exports. Both can have an impact on consumers, producers and workers through household consumption, household production, and labor incentives. The channels through which increased trade integration can affect labor include: (i) the consumption channel, typically leading to an increase in purchasing power and therefore higher real wages, and (ii) the employment effect due to increased labor demand. The extent of these gains to trade will depend on the incidence of trade policies or trade shocks; in other words, the impact will depend on which products become less expensive, which sectors increase demand for skilled or unskilled labor, and which workers can access these new jobs. 2. Lao PDR has experienced very strong export growth in the past decade, and this report tests several key research questions linking export performance to employment. For example, does increased trade boost aggregate employment, or simply generate a reallocation of workers between sectors? Do increased trade opportunities and export growth generate more jobs only in specific trade-related sectors or across the economy? Does increased trade lead to skill upgrading and/or more demand for skilled workers, enhancing long-term competitiveness? Does export growth lead to better jobs in terms of skill level and higher wages? 3. This report utilizes a range of methodologies and datasets that implicitly link trade and jobs; by using these complementary analytical approaches, we generate multiple perspectives on Lao PDR’s recent labor market outcomes, and their implications for Lao PDR’s current and future trade competitiveness. The objective of this report is to provide evidence using the following analytical approaches:  Employment and wage trends in the traded and non-traded sectors. Using the Lao PDR Earnings and Consumption Survey (LECS waves 4 (2007/08) and 5 (2012/13)) and the 2010 Lao PDR Labor Force and Child Labor Survey, we examine patterns of labor supply and demand in the Lao labor market, the nature of employment and how it has changed over time, the patterns of structural transformation of employment in the Lao economy, and the ways in which Lao PDR’s increased trade integration has affected worker welfare. This report builds on similar analysis in the World Bank’s recent Lao Development Report (World Bank 2014b), exploring wage gaps across sectors and population groups, and estimating the returns to education, controlling for sector differences and other individual characteristics. Education and skills are critical aspects of labor supply and demand, as pointed out in World Bank (2014a); respondents to the 2012 Enterprise Survey named insufficient skills as the main impediment to private sector growth. Furthermore, education and skills can be a fundamental driver of mismatch that could explain unemployment or underemployment. In addition to broad labor market characteristics, the analysis pays particular attention to the manufacturing sector, which is highly concentrated in exported garments. The analysis uses data on sectoral employment, productivity and wage trends to draw inferences about the 1 quality of jobs across sectors, and estimates the returns to working in different sectors and/or provinces. The results highlight the existing incentives for skill acquisition and identify access and job quality gaps that will affect labor competitiveness in the future.  Labor sophistication of exports. With respect to the traded goods sector, the analysis considers the types of labor utilized in Lao’s manufacturing exports and how this compares to the labor content of exports from competitor countries, which can shed light on the relative competitive position of Lao exports. Looking to future competitiveness prospects, we analyze the labor sophistication of Lao PDR’s existing export basket, how this has changed in recent years, and how these shifts compare to neighboring Vietnam or Cambodia, or to other countries at a similar stage of development.  Export-employment elasticity. The analysis compares firms’ demand for labor across sectors, and the degree to which exporting firms exhibit employment practices that differ from non- exporters. Given ongoing and planned Government efforts to address the regulatory environment in Lao PDR including with respect to labor regulations, the analysis also explores whether more competitive environments give rise to higher labor demand, or demand for different types of labor.  Labor mobility and adjustment costs. This part of the analysis examines the circumstances under which a trade shock can benefit or harm workers through changing labor demand and the role that mobility costs play in impeding labor adjustment to the shock. The impact will depend on the nature and cost of mobility constraints facing Lao workers, as well as the sectoral distribution of employment, skill sets and wages, and the relative impact of a trade shock on particular sectors. Even for a favorable shock, not all workers will benefit equally, and some workers may lose. The analysis uses panel data to derive transition matrices (the movement of workers between jobs and sectors), from which we estimate labor mobility costs. This provides the basis for testing the labor response to a hypothetical trade-related price shock. 4. The findings of these separate analyses provide a multi-faceted picture of Lao’s labor market and the role of trade in creating competitive jobs in competitive sectors. The conclusions point to a set of policy implications for policymakers to bear in mind as they address Lao PDR’s long- term development and growth challenges. 1.2. Trade Context 5. Lao PDR has seen strong economic growth during the past decade, driven by the exploitation of natural resources and closer regional integration. The country is surrounded by some of the fastest growing economies in the world, and it has benefited significantly from external demand for tradables and increased foreign direct investment inflows. Real GDP grew at an average growth rate of 7.8 percent per annum over the past decade. In 2010, Lao PDR graduated from low income to (lower-) middle income country status, and by 2014, per capita Gross National Income had climbed to US$1,600. At this pace, Lao PDR is on track to achieve its long-term vision: to graduate from Least Developed Country (LDC) status by 2020. 2 6. Despite impressive poverty reduction and progress towards achieving the Millennium Development Goals over the past ten years, significant gaps remain. Between 2002/03 and 2012/13, inequality widened, as the gains for the bottom 40 percent were significantly less than for the rest of the population, a pattern typical of natural-resource led growth. While absolute poverty (based on the national poverty line) declined from 33.5 percent of the population in 2002/03 to 23.2 percent in 2012/13, the Gini coefficient rose from 32.5 to 36.2. Measured using the internationally comparable poverty line of US$1.25 a day in 2005 PPP terms, poverty declined from 41.4 percent in 2002/03 to 28.8 percent in 2012/13. Human development indicators have not improved commensurately with poverty reduction, however, and certain population groups in remote areas lag behind. 7. Against this background, Lao PDR’s economy is going through a structural shift from agriculture to natural-resource based industry. Natural resource-based sectors have been growing very rapidly during the past decade, averaging around 18 percent real annual growth between 2003 and 2015 (supported by mining and hydropower mega projects). As a result, the share of these sectors in GDP has tripled since 1998, increasing from 6 percent of GDP in 1998 to 18 percent in 2012 and beyond. The share of non-tradables (services) to GDP grew more moderately, driven by increases in tourism and trade in recent years as well as expansion in infrastructure construction. Similarly, manufacturing (garments, wood and wood products, construction materials, light manufacturing, handicrafts, and, increasingly, food, beverages, and related processing) as a share of GDP has expanded, but at a relatively slower rate. 8. Lao PDR continues to integrate more closely into the rapidly growing regional economy through the implementation of trade commitments both multilaterally and within the Association of Southeast Asian Nations (ASEAN), the maintenance of relatively low tariffs, and through improved physical infrastructure and connectivity with neighbors. This has resulted in strong growth in cross-border flows of goods, services and investment. However, much of this has been driven by the growing external appetite for Lao PDR’s natural resources. Although the resources sector has brought significant benefits to Lao PDR’s economy in the form of higher economic growth, an increase in fiscal revenues, and infrastructure improvements, an over-reliance on the sector as a primary source of economic growth poses serious risks to the sustainability of the country’s long-term development. There is also a risk of increased volatility as the economy becomes more susceptible to external shocks driven by commodity price movements. In addition, the sector may not generate the kind of broad-based jobs and incomes levels needed to benefit a significant proportion of the Lao population. 9. Even under fairly modest projections, Lao trade performance is likely to become increasingly skewed towards natural resource (hydropower and minerals) exports in the years ahead (Figure 1). While this brings tremendous opportunities, it also presents risks associated with the lack of export diversification. Recent analytical work describes how lack of diversification can restrict economic growth through the early onset of diminishing marginal returns to a more limited range of export products, and through increased risk of terms-of-trade shocks. 2 Cross-country research on the “natural resources curse” actually finds that the core driver of the curse is not the presence of natural resources per se, but the degree of export 2 See World Bank (2011a). 3 concentration. 3 Slow job creation is a key risk if the growth of other sectors such as manufacturing and services is constrained. Figure 1: Mining and electricity are Figure 2: Exporters show lower productivity increasingly dominating annual exports compared to non-exporters Source: World Bank (2012a) Source: World Bank (2011b) 10. There is growing evidence that export competitiveness in non-natural resource sectors is being eroded as the natural resources boom gets underway. Lao export flows in non- natural resource-based sectors tend to be short-lived and rarely in sub-sectors offering scope for significant value addition (Figure 1). Similarly, exporting firms demonstrate substantially lower productivity compared to firms serving the domestic market4 (Figure 2). As in many other LDCs, a not insignificant share of non-natural resource-based exports is dependent on trade preferences associated with Lao PDR’s Least Developed Country status. There are also increasing signs that the kip, Lao PDR’s currency, is overvalued. Small and medium sized enterprises struggle to access finance and develop the skills necessary to became larger and more efficient. Finally, evidence from a recent Investment Climate Assessment suggests that the natural resource boom has also concealed the costs of inadequate workforce education and skills, which lag behind those of comparator countries, and have emerged as the main constraints to private sector expansion in Lao PDR. 11. Increased trade and integration have the potential to act as powerful drivers of pro- poor growth and shared prosperity. Whereas the relationship between trade and poverty is complex and the transmission channel between the two is not direct, consistent with cross- country evidence, increased trade and in particular trade-related job creation at the low end of the skills spectrum hold great potential for delivering welfare improvements to poorer households. Data from the Lao Expenditure and Consumption Survey indicates that a key determinant of poverty reduction in Lao PDR has been the extent to which communities are linked to markets and services.5 Given Lao PDR’s relatively high female labor force participation rate, and the 3 Lederman and Maloney (2009). 4 See World Bank (2011b) for more on the relatively poor productivity of exporting firms vis-à-vis firms exclusively serving the Lao domestic market. 5 A fuller discussion of poverty trends is included in background papers to the Lao PDR Development Report 2010. 4 predominance of women in tradable sectors such as garment manufacturing, tourism and agribusiness, trade can be a catalyst for increased women’s economic empowerment, with positive externalities for household welfare and human capital investment. 12. Surrounded by some of the fastest growing economies in the world, Lao PDR has deepened its integration with the regional market. Growing regional demand for Lao PDR’s natural resources, coupled with trade policy reforms both multilaterally and in the context of ASEAN, have boosted cross-border trade and investment between Lao PDR and its neighbors. Goods trade is concentrated by destination with Thailand, accounting for more than half of Lao PDR’s exports, followed by Vietnam, Australia and China. Thailand also supplies three-quarters of the country’s imports, followed by China, Vietnam and Japan. 13. To meet its goal of graduating from Least Developed Country status by 2020, Lao PDR will need to focus not only on continuing to expand exports both regionally and globally, but also to increasing the returns of expanded trade to workers through higher employment and incomes. The analysis that follows examines the extent to which workers have participated in and benefit from the recent trade growth. 5 Chapter 2. Lao PDR’s Labor Market and Recent Employment Trends 2.1. Characteristics of the Labor Force 14. The labor market in Lao PDR is characterized by very high and increasing labor force participation and extremely low and declining unemployment. Labor force participation rates increased from 85 percent in 2007/08 to nearly 94 percent in 2012/13, with increases observed in both urban and rural areas (see Figure 3 and Annex 1 Table 1.1). Unemployment fell over the same time period, from 0.4 percent to 0.2 percent. Rural residents have relatively higher participation rates and lower unemployment (Figure 4), explained in part by the high share of labor engaged in farming activities. Urban participation rates grew slightly faster than rural rates, and women’s LFP grew faster than men’s. By 2013, participation rates were nearly identical for men and women. The capital city of Vientiane, which had the lowest LFP rate in 2007/08, experienced the greatest increase, reaching 96 percent in 2012/13, while unemployment fell to 0.1 percent. Champasak also experienced rapid growth in labor force participation and a concurrent decline in unemployment to zero. These results jointly translate into strong job growth. Figure 3: Rising LFP, falling unemployment Figure 4: Signs of queuing for good jobs Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey. 15. Unemployment is slightly higher among more educated workers and those in urban areas, suggesting a degree of voluntary unemployment as some workers prefer to remain unemployed while waiting for high-paying job offers in line with their qualifications or expectations. In 2012/13 the unemployment rate was the lowest for the population with no formal schooling – 0.1 percent – and highest for the population with a tertiary education – 0.3 percent. Given that labor force participation rates are similarly high across all levels of schooling, the higher unemployment rate for tertiary graduates suggests that more skilled job seekers may choose voluntary unemployment while waiting – or “queuing” – for a good job in the formal sector. This pattern of queuing is very common, and can be explained by a range of possible factors, including the need to spend time and effort searching for a good job, or the 6 relatively higher reservation wages of more educated workers, who themselves may live in higher income households that can afford to support their unemployment spells. 16. Overall, schooling outcomes are low in Lao PDR. In 2012/13, the population had on average 4.7 years of schooling (5.4 for men compared to 4.1 for women) – the equivalent of less than a primary education. According to the coding in the LECS datasets, there are five grades in primary school, four grades in lower secondary, three grades of upper secondary, two grades of low vocational school, three years of middle technical school, and nine years of college / high technical school (because lower secondary school in Lao PDR is only three years, the fourth year was treated as the same as the first year of upper secondary / low vocational school for comparison purposes).6 Close to 20 percent of the population has not completed primary school, and another 37 percent has completed only primary schooling. About 15 percent of the workforce has completed upper secondary or vocational training, and another 6 percent has obtained tertiary degrees (see Annex 1, Table 1.2). Education quality is also very low, even at the foundational level; a 2012 early grade reading assessment found that 30 percent of 2nd graders could not read, and only 60 percent of 4th graders had achieved reading fluency (World Bank 2014c). As stressed in World Bank (2014b), low literacy rates have persistent negative effects on the productivity levels of Lao workers because basic literacy is the foundation for acquiring other skills. 17. Schooling rates are rising, especially for women, but women remain much more likely to have less than a primary degree compared to men. The percentage of females with an incomplete primary education fell from 34 percent in 2007/08 to 27 percent in 2012/13 (Annex 1, Table 1.2). Nevertheless, males are more likely to have completed primary school (19 percent of males had no schooling in 2012/13, compared to 27 percent of females). And among those who complete primary, women are less likely to continue their education. 18. The time trends in educational attainment are positive. Between 2007/08 and 2012/13, the share of tertiary degrees more than doubled from 2 percent to 5 percent (admittedly from a very low base) and vocational education also rose (Figure 5). We observe a significant improvement at the lower end, as the share of workers with less than a primary degree fell from 39 percent to 25 percent, with most upgrading to a completed primary education. At the same time, however, drop-out rates increased (World Bank 2013). 6 In Lao PDR, the schooling levels (and corresponding grades) are: primary (grades 1-5); lower secondary (grades 6- 8/9); and upper secondary (grades 9-11), which could be completed as either general education or vocational). Students then progress to either technical training – which could be either first level for 2 years (grades 12-13) or middle level for 3 years (grades 12-14) – or to tertiary (grades 12-20). The highest level attained is based on highest grade attained. Less than primary corresponds to no schooling or completing grades 1-4, primary corresponds to completing grades 5-7, lower secondary corresponds to completing grades 8-10, etc. In the LECS we cannot determine if individuals attained either first or middle level technical training; we therefore considered those who have attained two or three years there to have completed technical training. 7 Figure 5: Shifts in educational attainment Educational attainment 2007 2013 2% 4% 5% 8% 6% 25% 39% 10% 15% 17% 32% 37% Less than primary Primary Lower secondary Upper secondary Vocational Tertiary Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey. 19. We observe a strong urban-rural divide in educational attainment as well. About three- fourths of rural residents have a primary education or less, compared to about a third for urban residents (in 2012/13). This result may stem partly from more limited access to schools in remote rural areas, but may also be a function of the nature of available jobs in rural labor markets. Most farming, for example, is conducted on small family farms and therefore does not require formal education. As shown in Annex 1 Table 1.2, three-fourths of agriculture workers have only a primary education or less. Residents of the capital Vientiane have significantly higher rates of upper secondary, vocational and tertiary education compared to all other regions. It is interesting to note that vocational attainment actually fell by 2 percentage points in Vientiane between 2007/08 and 2012/13, while tertiary rates concurrently increased by 6 percentage points. Otherwise, we do not observe significant regional differences in educational attainment. 20. Garment sector workers are only moderately more skilled than farmers. On average, 60 percent of workers in the garment sector had a primary education or less in 2012/13, up from 44 percent in 2007/08. This shift to a less educated workforce is consistent with reported high turnover rates and increasing competition from migration, in response to which garment manufacturers are increasingly recruiting young women from remote areas. 21. According to the 2012/13 survey, public sector jobs, and among those, civil service jobs, are dominated by workers with post-secondary education (Figure 6). One-third of civil servants have vocational degrees, and another 41 percent have tertiary degrees. This high concentration of the most educated workers in government jobs helps explain the queuing behavior noted above. Employees of state-owned enterprises exhibit a more even distribution of education levels. The public sector skills bias stands in stark contrast to private sector workers, most of whom have only a primary education. In the analysis below, we explore the extent to which higher education levels are remunerated by higher wages. 8 Figure 6: Public employees have disproportionately higher education Educational attainment by population groups, 2013 100% Percent 80% 60% 40% 20% 0% Private Public SOE Civil servant Urban Rural Less than primary Primary Lower-secondary Upper secondary Vocational Tertiary Source: 2012/13 Lao PDR Earnings and Consumption Survey. 2.2. Sectoral Employment Trends 22. Employment in Lao PDR is dominated by agriculture, although we observe a significant structural transformation toward services. In 2012/13, 60 percent of workers were engaged in agriculture, down from 71 percent in 2007/08 (see Annex 1, Table 1.3). The services share of employment, meanwhile, increased from one fifth to one quarter during the same period, driven by an increase in private sector service jobs, particularly in wholesale and retail trade, accommodation and food services, construction, and financial and insurance activities. For example, the number of net jobs added to the wholesale and retail trade sector between 2007/08 and 2012/13 was 31 thousand, compared to 21 thousand net jobs added in accommodation and food services, 15 thousand in construction, and nearly 13 thousand in financial and insurance activities.7 The remaining sectors account for very small shares of total employment: less than 1 percent of workers are employed in mining, 1.5 percent in garment manufacturing, and about 5 percent in other manufacturing. Garment manufacturing employment is predominantly female, accounting for 3.1 percent of total female employment, compared to 0.4 percent of male employment. Gender differences are quite small for all other sectors. 23. The shift toward non-farm activities was not accompanied by a significant shift from rural to urban employment. About 29 percent of the workforce were located in urban areas in 2012/13, the same as in 2007/08, and only 12 percent resided in Vientiane capital (also unchanged between 2007/08 and 2012/13; Figure 7). This means that rural workers are finding non-farm employment without migrating to urban centers – or at least without permanently migrating to cities. Anecdotal evidence based on private sector consultations suggests that garment factories mostly hire young women from rural areas where there are few alternative job 7 Note that the weights used to derive employment levels are at the household level. 9 prospects beyond agriculture or migrating to Thailand. Moreover, rural areas closest to the Thai border are much more dependent on permanent and temporary outmigration because of the significantly higher wages in Thailand compared to local jobs. The wage gap is large even when comparing unskilled work in Thailand to relatively attractive jobs in Vientiane. In the largest garment firms, workers tend to live on site and work for a limited period (averaging less than one year) before returning to their villages. Figure 7: Shift out of agriculture toward services, but little shift from rural to urban Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey. 24. Net job creation has been driven by the services sector, more than offsetting the large net job losses in agriculture and manufacturing (Table 1). Between 2007/08 and 2012/13, over 90 thousand jobs were created in Lao PDR. This is in spite of the agriculture, forestry and fishing sector contracting by 97 thousand jobs, and the net loss of 11 thousand manufacturing jobs (despite strong job gains in beverages, wearing apparel, furniture, and repair and installation of machinery and equipment). Manufacturing job losses occurred primarily in textiles (-10,408), leather and related products (-3,082), wood and wood products (-12,294), and fabricated metal products (-4,274). Construction, trade, hotels and restaurants and public administration were all important for job creation, and other services were the most important job creator. 25. The small share of manufacturing jobs and their relative stagnation since 2007/08 suggests a lack of investment to increase the number of firms or to expand employment in existing manufacturing firms. Whereas labor regulations in Lao PDR do not seem to act as a binding constraint on hiring, there is anecdotal evidence of confusion surrounding the implementation of the new minimum wage regulation which could diminish labor demand. This suggests an important role for the Ministry of Labor for outreach and clarification for employers. There may be other business climate challenges that present larger impediments to firm creation and expansion, but these are not addressed here (see for example World Bank (2011b), World Bank (2014a), World Bank (2014b) and World Bank (2015)). 10 Table 1: Net job creation at the sectoral level, 2007/08-2012/13 Job creation Sector 2007/08-12/13 Agriculture, Forestry and Fishing -96,849 Mining and Quarrying 5,657 Manufacturing -11,238 o/w Garments 6,187 Construction 15,032 o/w Construction of building 24,662 o/w Civil engineering 6,836 Specialized construction activities -16,467 Wholesale and Retail Trade; Repair of Motor Vehicles and Motorcycles 31,102 Accommodation and Food Services Activities 21,390 Public Administration and Defense, Compulsory Social Security 25,092 Other Services 101,703 Total 91,889 Source: Authors’ calculations from the 2007/08 and 2012/13 LECS 26. Although the survey data do not allow us to measure the extent of informality using the traditional definition (i.e., employees with access to social security or employers or own account workers in a registered business), the predominance of farming and services jobs – both low productivity – suggests that a significant share of employment is informal. Disaggregating by employment status, over 40 percent of employed workers are in unpaid family work, and another 40 percent are own-account workers (Figure 8). About 9 percent of Lao PDR’s labor force are private wage employees (which could be formal employment in registered firms, or informal wage work), and 8 percent are employed by the public sector (1 percent in state owned enterprises, 7 percent as civil servants), with men and women equally represented. Figure 8: Employment is predominantly informal Source: 2010 Lao PDR Labor Force and Child Labor Survey; 2012/13 Lao PDR Earnings and Consumption Survey. 11 2.3. Productivity and Average Wages 27. The sectoral distribution of employment stands in stark contrast to the sectoral decomposition of GDP (and value added) and the composition of exports, suggesting that labor is not optimally allocated to the most productive or export-oriented sectors. As discussed in detail in Chapter 3 below, Lao’s exports are mainly concentrated in mineral products and base metals, neither of which is labor intensive, reflected in the high labor productivity of the mining sector. Mining value added per worker is 19 times higher than that of the economy at large (i.e., aggregate VA per worker), and increased in real terms by 3.4 percent on average between 2007/08 and 2012/13 (Table 2). Electricity, gas and water supply come a close second, with sectoral labor productivity 17 times higher than the average of the economy, and growing 8.2 percent per year between 2007/8 and 2012/13. Agriculture, forestry and fishing has the lowest labor productivity in the Lao economy while employing the largest share of workers who provide more than half of the economy’s total work hours (World Bank 2014 b). Education and other social services and hotels and restaurants also have very low productivity. Textiles and garments account for a tenth of exports, but employ an even smaller share of the labor force. According to the LECS 4 and 5 survey data, the number of garment jobs (textiles and wearing apparel) declined from 66,746 in 2007/08 to 62,526 in 2012/13, but this decline was due to negative job growth in textiles, contracting by 10,408 workers, while jobs in the wearing apparel sector increased by 6,187. 28. Manufacturing productivity increased significantly since 2007/08 as sectoral employment contracted. The national accounts value-added data do not allow us to disaggregate between garment and other manufacturing output, but the wage and employment data suggest that gains in value-added per worker were likely driven by other manufacturing, where employment contracted concurrent with a 7.1 percent increase in the average real wage. Garment wages grew relatively slower, averaging 2.6 percent per year in real terms. Whereas this translates into significant welfare gains for garment workers, the relative gains are weak compared to other sectors. Garment workers at the upper end of the education distribution benefited more than their unskilled counterparts, however, as wage growth for garment workers with a vocational degree averaged 14 percent per year. Although these wage trends are consistent with reported shortages of skilled labor in garments, they also raise questions about why garment employers struggle to retain new sector entrants, reflected by high turnover rates. New entrants generally lack a secondary or vocational degree, and require employer-provided training. If firms want to retain their trained workers, they could offer higher wages but do not. One potential explanation is that recruiting and training costs are low, and employers would rather bear these costs than offer higher wages. Alternatively, if profit margins are very small, it may not be feasible for firms to raise wages adequately to compete with alternatives (e.g., higher wages in Thailand) or to compensate for the unpleasant nature of factory work. Limited career mobility prospects and/or minimal knowledge spillovers to other occupations also act as impediments to attracting and retaining productive workers.8 8 See World Bank (2012b). 12 Table 2: Real output per worker and real wages experienced robust growth, but varied by sector Sector Real Annual Average Labor Productivity Growth (Value Added/Worker) 2007/08-2012/13 2007/08 2012/13 Real (2002) Real (2002 Kip) (2002 Kip) VA/ (2002) Worker Wages Agriculture, Forestry and Fishing 3,746,871 4,558,711 4.0 10.2 Mining and Quarrying 183,992,736 217,009,633 3.4 18.4 Manufacturinga 10,511,264 17,052,103 10.2 6.3 Electricity, Gas and Air-conditioning Supply; Water Supply; Sewage, Waste Management and Remediation Activities 124,264,782 184,290,557 8.2 4.8 Construction 10,581,651 22,867,669 16.7 7.9 Wholesale and Retail Trade; Repair of Motor Vehicles and Motorcycles; Other Services 15,852,896 19,431,128 4.2 7.4 Accommodation and Food Services Activities 14,336,576 8,060,957 -10.9 -0.7 Transportation and Storage; Information and Communications 24,177,380 32,816,684 6.3 3.2 Financial and Insurance Activities 158,851,651 80,276,992 -12.8 11.4 Real Estate Activities; Professional, Scientific and Technical Activities; Administrative and Support Service Activities 16,948,958 21,476,844 4.8 5.2 Public Administration and Defense, Compulsory Social Security 22,150,428 30,354,205 6.5 Education; Human Health and Social Work Activities; Arts, Entertainment and Recreation 7,280,585 6,005,364 -3.8 Activities of Households as Employers 21,690,873 12,723,531 -10.1 6.9 TOTAL 8,051,122 11,445,514 7.3 7.3 a Note that real garment wages grew by 2.6 percent between 2007/08 and 2012/13, while real other manufacturing wages grew by 5.9 percent. Real labor productivity growth is the annual average change in sector value added per worker denominated in real 2002 Kip; wage growth is the change in the sector average wage denominated in real 2002 Kip. Value-added for 2007/08 is averaged over 2007 and 2008, and value-added for 2012/13 is averaged over 2012 and 2013. Source: 2012/13 Lao PDR Earnings and Consumption Survey and United Nations National Accounts. 29. The low proportion of manufacturing jobs and their relative stagnation since 2007/08 suggests that Lao producers have not succeeded in attracting FDI to expand garment and other manufacturing operations, and existing Lao manufacturing firms have not expanded production or created jobs as fast as other segments of the economy. Attracting new investment could help shift garment production toward higher value products and markets, requiring more skilled labor in better and more remunerative jobs. The potential reasons why such investment has not materialized could be the subject of future research in order to understand which sectors are better performing and may offer more opportunities to compete in the future. In the next chapter, we explore the competitiveness of Lao manufacturing in the international context. 13 30. Real wages have largely experienced strong growth over time, averaging 7.3 percent annual growth between 2007/08 and 2012/13 (Table 2). 9 The data 10 show that real mining wages grew fastest, averaging 18 percent annual growth, although real agriculture wage growth was not far behind at 10 percent. Service sector and other manufacturing wages grew in the range of 5-7 percent annually in real terms, but garment sector wage growth was weaker, and hotels and restaurant wages experienced negative real growth. The cross-sector variations in wage and productivity growth provide a mixed picture. Real value added growth was highest in construction, averaging 20 percent per year between 2007/08 and 2012/13, with real wage growth only 7 percent per year. Real value added growth in the agriculture, forestry and fishing sector instead was only 3.1 percent per year, partly explained by the reduction in agricultural employment, but agriculture wages grew by 10 percent per year. Real rural wage growth exceeded urban wage growth, although urban workers still earn relatively more (64 percent on average). 31. Comparing wage levels across sectors and skill profiles, workers employed in mining earn the highest wages, in particular workers with a tertiary education. In 2012/13, the mining sector paid on average 3.1 million Kip in weekly nominal wages, while mining workers with a tertiary education were receiving on average 9.7 million Kip weekly. Agriculture wage workers (i.e., laborers employed by other farmers, not self-employed) and garment workers earned the least, with weekly wages averaging 1.3 million Kip in 2012/13, followed by service sector workers (1.5 million Kip) and other manufacturing workers (1.9 million Kip). 32. Women earn less than men across education levels. In 2007/08, for example, the average male wage was 32 percent higher than the average female wage, but by 2012/13 the disparity was 22 percent. Although this suggests a narrowing of the gender wage gap, the regression analysis below shows that, controlling for other factors, the gap actually widened. The gender gap is largest at the low end of the education distribution (e.g., 76 percent for workers with a primary degree) but declines for more educated workers (e.g., 27 percent for tertiary graduates). 33. The average earnings of vocational graduates have grown over time compared to upper secondary and tertiary graduates. In 2007/08 female workers who completed upper secondary school earned higher wages than those who completed vocational school, but the premium effectively disappeared by 2012/13. Similar trends were observed in mining (where the premium went from 185 percent to 2 percent) and services (where it went from +17 percent to -7 percent). This switch suggests that productive activities in these sectors may have become more skill- intensive as they shifted to higher value or quality of outputs, generating greater demand for vocational graduates. An alternative explanation – that the relative quality of the vocational and secondary education streams changed – is also possible although unlikely in the short period of time. A third possible explanation is that the returns to secondary education declined due to the increased supply of secondary graduates. The data do not allow us to identify the main explanatory factor. In the garment sector, the value of vocational skills has increased 9 Note that survey data on the incomes of farmers and own-account workers tends to be incomplete or unreliable, limiting the extent to which the analysis can address income challenges for a large segment of the labor market. The wage analysis herein necessarily focuses on wage earners in sectors dominated by wage jobs. 10 Nominal wage data from the LECS 2007/08 and 2012/13 are converted to real 2002 wages using the World Development Indicators’ CPI. Note that non-wage agriculture income is excluded, as is income earned by own- account workers. 14 significantly, as mentioned above, growing faster than other education categories. Annex 2 Tables 2.1-2.3 report real weekly wages by sector and region for the different survey periods. 34. Workers employed in state owned enterprises (SOEs)11 earn substantially higher wages than other public sector (i.e., civil servants) and private sector employees (Table 3). This may or may not reflect differences in productivity however, as public wages are not necessarily market driven.12 Whereas the average public sector wage is close to that of private sector wage employees (the latter was 1.6 million Kip per week in 2012/13), the wage premium for SOE workers is very high compared to civil servants (SOE employees earn 2.3 million Kip per week on average, compared to 1.3 million Kip for civil servants), and this trend holds across sectors (with the exception of agriculture). The premium is especially high for those in other manufacturing, and for those with a tertiary degree. Private sector wages increase linearly with educational attainment: private employees with a tertiary degree earn over three-quarters more on average than a worker with a primary education. Table 3: SOEs pay more on average and attract more educated workers (2012/13) Average Sectoral years of Average Average Average Share of distribution of schooling years of nominal wage nominal wage employees in SOE (non- schooling (non-SOE) (SOE) SOEs (%) employees (%) SOE) (SOE) Agriculturea 1,247,430 1,200,000 0.24 6.90 4.5 3.9 Garment 1,205,463 1,454,974 1.87 2.37 6.2 8.0 Mining 2,462,344 2,750,000 4.28 1.47 7.0 9.3 Other manufacturing 1,370,358 3,208,001 3.84 13.98 6.6 9.6 Services 1,433,563 1,794,501 3.40 75.28 9.0 10.1 a Agriculture wage employees, excludes farmers. Source: 2012/13 Lao PDR Earnings and Consumption Survey. 35. Given the lower wages of civil servants, but higher likelihood of having received a tertiary education, this may dampen the returns to a tertiary education in Lao PDR. In the analysis that follows, we estimate the size of wage premia for different sectors of work and different levels of education using regression analysis. 11 Examples of SOEs in Lao PDR: Lao Airlines, Electricite du Laos (Electricity company), Nam Pa Pa (Water supply company), ETL Public Company (Telecom company). 12 In fact, civil servant salaries experienced two consecutive years of 35-40 percent nominal increases in FY2013 and FY2014 (World Bank 2014b). 15 2.4. Determinants of Earnings 36. Our regression analysis on the determinants of earnings uses Mincer-type specifications to test for the factors driving wage differences for wages reported in the 2007/08 and 2012/13 Lao PDR Earnings and Consumption Surveys (LECS waves 4 and 5) and the 2010 Lao PDR Labor Force and Child Labor Survey (see Annex 3 for details). We estimate each year separately, and note that differences in the survey questionnaires and sampling will affect comparability. There are four categories of wage earners: civil servants, employee of public and private work 13 ; employee of a private business, and employee of a state enterprise. We exclude self-employed, unpaid family workers, and employees not paid a wage (i.e., we exclude earnings from household production and own-account workers and farmers).14 37. There are positive returns to years of schooling, but these seem to have declined since 2007/08 (see tables in Annex 3 for the regression results). We also note a large but declining urban bias in wages, about 72 percent higher in 2007/08, 23 percent in 2010, but falling to only 4 percent in 2012/13, and losing statistical significance (controlling for other factors such as sector, gender, years of schooling and enterprise type (i.e., public vs. private)). Wages in Vientiane Capital are statistically higher than wages elsewhere – 15 percent higher in 2012/13 – all else being equal. 38. When considering levels of education completed rather than years of education attended, higher levels of education are compensated with higher wages. In each survey year, returns to education are significant at all levels above primary and increase linearly, such that each level of education is compensated more than the one below it, and tertiary education has the highest returns. The time trends are difficult to ascertain because of changes in the questionnaires and the different sample frames used in the household and labor force surveys. The best comparison is regression 2 in Tables 3.4-3.6 in Annex 3, which controls for sector of work; the results indicate that the returns to each level of education declined between 2007/08 and 2012/13. 15 Whereas in 2007/08 the lower and upper secondary degrees were equally remunerated (equivalent to 34 percent higher than a worker with less than a primary degree), in 2012/13 there was a wider wage differential (lower secondary earned 16 percent more than less than primary, while upper secondary earned 21 percent more). The point estimate on the vocational degree did not change much (25 to 27 percent) but became significant in 2012/13. And for those with a tertiary degree, the premium relative to less than primary fell from 58 percent in 2007/08 to 39 percent in 2012/13, which is consistent with an increasing supply of skilled labor to meet current demand. The education data do not allow us to determine which types of tertiary degrees have the highest returns (e.g., an engineering degree versus a business degree). 13 Note that this category may reflect workers who hold at least 2 jobs, one in the public sector and one in the private sector. The survey questionnaire does not enable a disaggregation by job. 14 Excluding non-wage earners may bias the returns to education, although the direction of the bias is likely to be mixed (unskilled farmers may have very low earnings, while employers or self-employed workers may have very high earnings, despite low levels of education). The regressions also exclude individuals who work more than one job, because we cannot disaggregate wages by activity. Workers with missing observations for key variables such as years of education are also excluded from the sample. 15 The returns are higher across the board for the 2010 labor force survey data. 16 39. The gender wage gap actually increased between 2007/08 and 2012/13, controlling for other factors, in contrast to the nominal average wage trends reported above. In the regressions that control for sector but exclude urban, province or public-private controls (column 2 in Annex 3 Tables 3.1 and 3.4) we find that in 2007/08 women earned 14-15 percent less than men with identical education and sector of work, and this gap widened to 18-19 percent in 2012/13. When other controls are introduced (last column of Table 3.6), the gender wage gap is estimated to be 20 percent in 2012/13. 40. The marginal negative returns to working in the garment sector (relative to mining) have remained steady since 2007/08. In 2012/13, a garment sector worker with an otherwise identical profile in terms of education, gender and place of work earned 72 percent less compared to a miner. The returns to agricultural work are even lower, albeit much improved from 2007/08 (86 percent lower than mining in 2012/13). 41. Civil servants earn significantly less than their SOE counterparts. Whereas public employees earn about one-fourth less than private sector workers, controlling for sector and all other factors (last column of Table 3.6), the positive wage premium for SOE employees is over 20 percent, whereas civil servants earn a negative wage premium of 40 percent. 42. These wage regressions point to important earnings differences across sectors, education levels, firm type and location. The main conclusions point to mining as the best remunerated sector, the importance of education at all levels, particularly at the tertiary level, the higher wages in Vientiane Capital, and the disproportionate advantages to working in an SOE. 17 Chapter 3. Labor Sophistication and Competitiveness in Export Markets 3.1. Snapshot of Lao PDR’s Export Basket 43. According to mirror UN Comtrade data, Lao PDR’s goods exports (manufacturing plus non-manufacturing) are dominated by mineral products, base metals and wood and pulp, none of which is particularly labor intensive. Textile and garment exports, by contrast, rank fourth, accounting for only 10 percent of total exports in 2012. The value of textile and garment exports has increased steadily over the last decade, driven by exports to the EU. Disaggregating Lao exports by destination market shows that garments, vegetables and foodstuffs are the main exports to the EU, mineral products (mostly copper ores and concentrates) are destined to China, and base metals, mineral products and wood are the main exports to EAP countries (see Figure 9). Mirror data are used to describe the sectoral composition of Lao PDR’s export (using import data of formal imports reported by trading partners). Especially for wood exports, these figures will cover exports leaving Lao PDR informally due to the ban on logging and timber exports but which are declared as imports in neighboring EAP countries. But they do not cover goods that are secretly moved across borders. Figure 9: Sectoral composition of Lao PDR’s export basket over time and by destination market, in USD million 18 Notes: EU28 includes Austria, Belgium, Bulgaria, Cyprus, Croatia, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, and United Kingdom. EAP* includes Samoa, American Samoa, Cambodia, Fiji, Indonesia, Kiribati, Malaysia, Marshall Islands, Micronesia, Mongolia, Myanmar, Palau, Papua New Guinea, Philippines, Solomon Islands, Thailand, East Timor, Tonga, Tuvalu, Vanuatu, Vietnam, Hong Kong, Japan, South Korea, Macao, and Brunei. Base metal export growth in 2005, 2006, 2007 and 2008 is being driven by exports of refined copper (HS 740311) to Thailand and a lesser extent Vietnam. Mineral products export growth in 2009, 2010 and 2012 is being driven by exports of electrical energy (HS 271600) to Thailand and copper ores and concentrates (HS 260300) to China. Source: World Bank Staff calculations based on mirror data from UN Comtrade. 3.2. Measuring the Labor Content of Exports 44. Empirical studies show that what a country exports and where it exports affect the level and composition of labor demand in the exporting country and its economic growth performance. The structure of a country's exports – that is, what types of products it exports, in which sectors and with what levels of value-added – affects the types of labor needed to produce these exports. Moreover, the export structure tends to vary by destination market. Exporting to high-income countries, where consumers have a relatively high preference for quality, stimulates the production of high quality goods, and increases the demand for skilled labor (Brambilla et al. 2012). Hausmann, Hwang and Rodrik (2007) provide evidence that the types of products countries export matter for economic growth. And Verhoogen (2008) shows that more productive plants produce and export higher-quality goods than less productive plants, and they pay higher wages to maintain a higher-quality workforce. 45. The evolution of Lao PDR’s export basket in terms of product mix, product sophistication and destination markets is assessed to estimate its labor content and the implication for job creation and job quality. Understanding the position of Lao exports vis-à-vis competing exporters and how this has changed over time helps shed light on the challenges to diversifying Lao export markets and products with a view to increasing export-driven labor demand. Adapting the Hausmann, Hwang and Rodrick (2007) measure of export sophistication, and in line with Shirotori, Tumurchudur and Cadot (2010), we develop indirect measures of the labor- 19 related sophistication of Lao PDR’s export basket. We construct five indices or "EXPYs" 16 that measure the following aspects of labor sophistication: (i) median wage, (ii) median value added per worker, (iii) ratio of skilled to total workers, (iv) years of schooling, or “human capital”, and (v) capital stock per worker. These EXPYs reflect the global trade-weighted average labor content of the exports that appear in Lao PDR’s export basket. The logic behind these measures is that labor market and other characteristics of the economies that intensively export a given product provide information about the implications of specializing in certain products exported to certain markets, and will determine how export growth will contribute to job creation, wages and skill mix. Based on these observations, it is possible to draw inferences about how product and regional trade patterns can impact employment, wages and skill demand in Lao PDR within the regional competitive context. 46. For each of the five EXPY indices, we compare the time trends of their levels, both in the aggregate (i.e., for all exports) but also separated by destination market (i.e., calculating destination-specific EXPYs to reflect the export baskets to particular destination markets). We compare Lao exports destined to the European Union (EU2817), the US, China, India and EAP.18 The results for each of the five EXPYs for Lao PDR are shown in separate panels in Figure 10. By comparing the export-share-weighted EXPYs over time and to different destination markets, we gain insight into recent trends in the labor sophistication of Lao’s total merchandise exports, including manufactured exports. And comparing these time trends with exports from similar emerging economies such as Senegal, Cambodia, Indonesia, Bangladesh and the Philippines (in Figure 11) allows us to make bilateral comparisons of labor content and product sophistication. 47. The trends in Figure 10 indicate that the level of labor sophistication – as measured by the wage, value-added, share of skilled workers, and human capital – embedded in Lao exports to the world (solid red line) has changed little since 2006, although there was some improvement between 2003 and 2006. Lao exports to EAP reflect higher wage content, value- added and skill content compared to the basket of Lao exports to other markets, and Lao exports to India have relatively higher wage content and value added compared to exports to the EU and US. Lao exports to China, by contrast, tend to have a lower embedded skill content (measured by a lower ratio of skilled workers and by lower average years of schooling), as well as a low physical capital content, explained by the predominance of mineral and base metal exports. We observe the least improvement in labor sophistication – for all five measures – in exports destined to the EU, which are mostly garments (recall from Figure 9). 16 EXPYs can be interpreted as the export share-weighted average labor market outcome associated with Lao PDR’s export basket, and are calculated using two-steps. The first step uses Comtrade data on global exports and assigns to each product a PRODY value for the five indices, namely a wage PRODY reflecting the average international wage associated with that export, a value-added PRODY reflecting the average value added for that product, and so on for the average skill share, the average years of schooling and the average capital content. Each of these PRODYs is the global trade-weighted average, covering all countries in the world that export the product. The second step then weights the PRODYs appearing in Lao PDR’s export basket by the share of each product in Lao PDR’s total exports, denoted EXPY. 17 EU28 includes Austria, Belgium, Bulgaria, Cyprus, Croatia, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, and United Kingdom. 18 EAP* includes Samoa, American Samoa, Cambodia, Fiji, Indonesia, Kiribati, Malaysia, Marshall Islands, Micronesia, Mongolia, Myanmar, Palau, Papua New Guinea, Philippines, Solomon Islands, Thailand, East Timor, Tonga, Tuvalu, Vanuatu, Vietnam, Hong Kong, Japan, South Korea, Macao, and Brunei. 20 48. Comparing the labor content of Lao PDR’s export basket to the export baskets of comparator countries Senegal, Cambodia, Indonesia, Bangladesh and the Philippines (as measured by the EXPYs in Figure 11), we observe that Indonesian and Philippine exports are more sophisticated, reflected by higher embedded wage, value-added and skill content than Lao exports, and higher physical capital content. Senegalese exports also compare favorably to Lao exports, at least in terms of wages, value-added and physical capital content. Lao’s export basket is more sophisticated than those of Bangladesh and Cambodia, however, although this is not the case when we consider only textile and apparel products, which show very similar (low) levels of sophistication. In terms of skill content, Lao exports have the lowest average ratio of skilled workers relative to all the other comparator countries. These estimates of labor sophistication suggest that Lao PDR has maintained its relative ranking of competitiveness over the past decade, but is showing little progress in terms of closing the competitiveness gap . 21 Figure 10: Labor sophistication of Lao PDR’s export basket over time and by destination market 200 25 EXPY median value added (PPP thousands) 20 150 15 100 10 50 5 0 0 2000 2002 2004 2006 2008 2010 2012 2000 2002 2004 2006 2008 2010 2012 Year Year World EU28 China EAP* World EU28 China EAP* 80 10 EXPY human capital (years of school) 60 8 40 6 20 0 4 2000 2002 2004 2006 2008 2010 2012 2000 2002 2004 2006 2008 2010 2012 Year Year World EU28 China EAP* World EU28 China EAP* 120 80 40 0 2000 2002 2004 2006 2008 2010 2012 Year World EU28 China EAP* Note: The spike in 2004 for EAP is being driven by: Manufacture of veneer sheets; manufacture of plywood, laminate board, particle board and other panels and boards ISIC 2021 (jumping from 3% to 9%); Sawmilling and planing of wood ISIC 2010 (dropping from 85% to 77%); Preparation and spinning of textile fibers; weaving of textiles ISIC 1711 (jumping from 0.3% to 1.2%); and Manufacture of basic iron and steel ISIC 2710 (dropping from 2.3% to .1%). The spike in 2008 for EAP is being driven by: Manufacture of basic precious and non-ferrous metals ISIC 2720 (dropping from 74% to 69%). Source: World Bank Staff calculations based on mirror data from UN Comtrade, World Bank Business Environment and Enterprise Performance Surveys, and Shirotori, Tumurchudur, and Cadot (2010). 22 Figure 11: Labor sophistication of Lao PDR’s export basket to the world over time compared to other emerging markets 200 25 EXPY median value added (PPP thousands) 20 150 15 100 10 50 5 0 0 2000 2002 2004 2006 2008 2010 2012 2000 2002 2004 2006 2008 2010 2012 Year Year LAO KHM IDN BGD PHL LAO KHM IDN BGD PHL 80 10 EXPY human capital (years of school) 60 8 40 6 20 0 4 2000 2002 2004 2006 2008 2010 2012 2000 2002 2004 2006 2008 2010 2012 Year Year LAO KHM IDN BGD PHL LAO KHM IDN BGD PHL 120 80 40 0 2000 2002 2004 2006 2008 2010 2012 Year LAO KHM IDN BGD PHL Note: SEN indicates Senegal, KHM is Cambodia, IDN is Indonesia, BGD is Bangladesh and PHL is the Philippines. Source: World Bank Staff calculations based on mirror data from UN Comtrade, World Bank Business Environment and Enterprise Performance Surveys, and Shirotori, Tumurchudur, and Cadot (2010). 49. The above-described trends in the labor sophistication of Lao PDR’s export basket compared to other countries’ exports and by destination markets do not point to particular products or markets that have the potential to spur labor demand in Lao PDR. Whereas the preceding analysis compares the weighted average values of these export-labor indicators, the next step is to examine the distribution of these same indicators across the entire range of products that Lao PDR exports, in order to differentiate quality differences across product lines. We therefore shift 23 our attention to the labor content of Lao PDR’s export basket across the entire distribution of exports, measured for each EXPY characteristic (i.e., median wage, value added, skill share, human capital and physical capital), which better reflects heterogeneity in the export basket mix. To do this, we use cumulative distribution functions to plot the PRODY indicators across all products (rather than the weighted average of all PRODY indicators – i.e. the EXPYs – used above). This allows us to compare the evolution of the labor sophistication of the entire basket between two points in time, and we can differentiate differences at the lower or upper ends of the distribution. 50. Between 2000 and 2012, we observe that Lao exports became more sophisticated, reflected in relatively higher wages, value-added and human capital content across the entire distribution of exports (Figure 12). The entire distribution of PRODYs in 2012 lies to the right of the distribution in 2000, which implies that a greater share of Lao PDR’s exports are in products associated with higher wages, value-added and human capital content. For example, in 2000, about 80 percent of Lao PDR’s exports were associated with products that paid at most an annual wage equivalent to USD10,000. In 2012, this share had fallen to 60 percent, meaning that a greater share of Lao PDR’s exports were associated with countries that pay higher wages. The ratio of skilled workers decreased, however, except at the top end of the distribution. Lao PDR is not increasing its exports of goods that are produced with more skilled workforces. The physical capital content of the Lao export basket experienced some increase, but in the lower portion of the distribution only. Figure 12: Cumulative distribution of Lao PDR’s export basket to the world in 2000 and 2012 1 1 .8 .8 Cumulative distribution .6 .6 .4 .4 .2 .2 0 0 8 9 10 11 12 13 8 9 10 11 12 13 PRODY median wage (PPP thousands) PRODY median value added (PPP thousands) 2000 2012 2000 2012 24 1 1 .8 .8 Cumulative distribution .6 .6 .4 .4 .2 .2 0 0 0 .2 .4 .6 .8 1 2 4 6 8 10 12 PRODY skilled workers (share of workforce) PRODY human capital (years of school) 2000 2012 2000 2012 1 .8 .6 .4 .2 0 8 9 10 11 12 13 PRODY physical capital per worker (USD thousands) 2000 2012 Source: World Bank Staff calculations based on mirror data from UN Comtrade, World Bank Business Environment and Enterprise Performance Surveys, and Shirotori, Tumurchudur, and Cadot (2010). 51. Looking at the 2012 data for Lao exports to the EU28, China and EAP, Figure 13 illustrates that across most products in the distribution, Lao exports to the EU (dominated by garments) are similar to those produced by countries that pay lower wages and have lower value- added compared to Lao exports to China and EAP (dominated by mineral and base metal exports). For example, 80 percent of Lao exports to the EU are products associated with average wages around $9,000 per year or less. On the other hand, 90 percent of exports to EAP are products associated with average wages of $10,300 per year. Lao exports to the EU also embody a higher level of skill. Lao exports to EAP rank highest in labor sophistication, reflected by higher wages, value-added and human as well as physical capital for most exports. 25 Figure 13: Cumulative distribution of Lao PDR’s export basket to EU28, China and EAP in 2012 1 1 .8 .8 Cumulative distribution .6 .6 .4 .4 .2 .2 0 0 8 9 10 11 12 13 8 9 10 11 12 13 PRODY median wage (PPP thousands) PRODY median value added (PPP thousands) EU28 CHN EAP* EU28 CHN EAP* 1 1 .8 .8 Cumulative distribution .6 .6 .4 .4 .2 .2 0 0 0 .2 .4 .6 .8 1 2 4 6 8 10 12 PRODY skilled workers (share of workforce) PRODY human capital (years of school) EU28 CHN EAP* EU28 CHN EAP* 1 .8 .6 .4 .2 0 8 9 10 11 12 13 PRODY physical capital per worker (USD thousands) EU28 CHN EAP* Source: World Bank Staff calculations based on mirror data from UN Comtrade, World Bank Business Environment and Enterprise Performance Surveys, and Shirotori, Tumurchudur, and Cadot (2010). 52. Comparing the labor content of Lao PDR’s 2012 export basket to those of other emerging economies in Figure 14, Lao exports have a mixed performance in terms of wages, value-added, and human and physical capital. This suggests the prevalence of low- tech production methods and/or possible weaknesses of the education system as well. When 26 considering the subset of textile and apparel exports separately, Lao’s exports are even less well positioned vis-à-vis comparator countries than the basket as a whole. Figure 14: Cumulative distribution of Lao PDR’s export basket to the world compared to other emerging markets in 2012 1 1 .8 .8 Cumulative distribution .6 .6 .4 .4 .2 .2 0 0 8 9 10 11 12 13 8 9 10 11 12 13 PRODY median wage (PPP thousands) PRODY median value added (PPP thousands) LAO SEN KHM IDN BGD PHL LAO SEN KHM IDN BGD PHL 1 1 .8 .8 Cumulative distribution .6 .6 .4 .4 .2 .2 0 0 0 .2 .4 .6 .8 1 2 4 6 8 10 12 PRODY skilled workers (share of workforce) PRODY human capital (years of school) LAO SEN KHM IDN BGD PHL LAO SEN KHM IDN BGD PHL 1 .8 .6 .4 .2 0 8 9 10 11 12 13 PRODY physical capital per worker (USD thousands) LAO SEN KHM IDN BGD PHL Source: World Bank Staff calculations based on mirror data from UN Comtrade, World Bank Business Environment and Enterprise Performance Surveys, and Shirotori, Tumurchudur, and Cadot (2010). 27 53. Mediocre education outcomes – whether due to inadequate quality, misalignment with labor demand, or limited access to education services – can limit the potential sophistication of Lao exports if exporting firms cannot recruit labor with appropriate skills, which in turn dissuades investment in more productive technologies. Whereas Lao’s entry into the global marketplace led to some rapid gains in the early 2000s, more recent export performance has been lackluster and potentially reflects a weakening competitive position, at least regarding labor- intensive exports. 54. The small profit margins in the low-value garments subsector mean that Lao producers have little room for finding productivity gains vis-à-vis competitors such as Bangladesh and Cambodia. Vietnam’s ongoing shift to higher value production could create space in the market for Lao producers. A more promising strategy would focus on creating more attractive niches or garment product lines for high income economies through e.g., upgrading labor standards and environmental standards, helping big brand retailers and buyers to meet their corporate social responsibility objectives. The Government has recently undertaken a standards upgrading program with technical assistance from the World Bank and ILO to the Labor Inspectorate of the Ministry of Labor under the auspices of the 2nd Trade Development Facility. This technical assistance focuses on the garment manufacturing sector. 55. Beyond garment manufacturing, Lao’s relatively low-tech export basket limits its capacity to access higher income consumers seeking higher quality goods and services . Even within the regional context, Lao has new opportunities to provide duty-free goods and services to its ASEAN partners. Government policies should therefore focus on facilitating access to technology and higher value production through for example more favorable (but still non- distortionary) investment incentives, or more generally by easing access to formal finance such as through banks or SME-targeted credit institutions but in the context of a level playing field. 56. The fact that over two-thirds of Lao PDR’s manufacturing workers are employed in the garment subsector, where both wages and technology content are relatively low, implies the need to diversify production into higher value manufacturing as well as higher value services. Increasing productivity and job quality in both requires enhancing the quality of the labor force. This means policies to increase educational attainment, more support for basic education, curricula reform necessary to meet market demand including soft skills and entrepreneurial skills, and aligning vocational training to private sector demand, particularly export-oriented private sector demand. A large pool of skilled and competitive labor is an essential component both for attracting investment in higher-value activities, and for enabling firms to upgrade the quality of their goods and services. 28 Chapter 4. Export-Employment Elasticity 4.1. Comparing Exporting Behavior and Labor Demand 57. Exporting firms tend to employ more workers than non-exporters in the manufacturing sector, according to evidence from World Bank Enterprise Surveys; this holds true in Lao PDR as well as in a sample of comparator countries (see Annex 4 Table 4.1).19 In fact, Lao exporting firms are about five times larger than non-exporters, and employ both skilled workers and unskilled workers (based on number of full-time employees). The share of skilled employees in exporting firms has increased over time, accounting for more than half of total workers in 2012. But this higher skills ratio does not seem to reflect higher product quality (based on the preceding results). Non-exporters, by contrast, employ relatively more unskilled workers, a trend that holds in the sample of comparator countries as well. 58. Comparing Lao PDR to its larger neighbors Vietnam and Cambodia, both of which pursued development policies based on industrialization with particular focus on garment manufacturing exports, we see that firm sizes are much larger in both Vietnam and especially Cambodia, and that exporters rely disproportionately on skilled rather than unskilled labor (Annex 4 Table 4.2). 59. The findings of the 2012 and 2013 World Bank enterprise surveys indicate that Lao PDR has a higher-than-average share of manufacturing firms that export (43 percent of Lao PDR’s manufacturing firms export, compared to 35 percent of manufacturing firms in the full sample; see Annex 4 Table 4.3). The manufacturing sub-sectors that are most engaged in exporting are chemicals, electronics and garments. For example, all chemical and electronics manufacturers and 77 percent of garment manufacturers in Lao PDR export, and most of these firms’ production is for external markets. About half of garment manufacturers export, and among these, they export over three-fourths of their output. This reliance on exports contrasts with other countries in the sample, where about one-third of manufacturers export, and among those exporters, only half of their production is destined outside the country, the other half destined for domestic consumption. This implies a lower degree of market diversification for Lao manufacturing production. Vietnam’s garment sector is similarly export-intensive, while Cambodian producers demonstrate the highest degree of export orientation, particularly in the textile and garment subsectors, but in food and other manufacturing as well (see Annex 4 Table 4.4). 60. For the World Bank Enterprise Survey sample, more than two-thirds of Lao PDR’s manufacturing workers are employed in the garment sector, in contrast to the more diversified distribution of manufacturing employment across subsectors in the rest of the world 19 Comparator countries are China, Belarus, Nepal, West Bank and Gaza, Georgia, Azerbaijan, Kyrgyz Republic, Kazakhstan, Bangladesh, Armenia, Uganda, Ukraine, Uzbekistan, Mongolia, Moldova, Tanzania, Kenya, Bulgaria, Serbia, Zambia, Bosnia and Herzegovina, Albania, Kosovo, Croatia, FYR of Macedonia, Slovenia, Montenegro, Djibouti, Jordan, Romania, Latvia, Democratic Republic of the Congo, Tajikistan, Estonia, Hungary, Lithuania, Poland, Slovak Republic, Turkey, Ghana and Madagascar. 29 (Table 4). How good are these jobs? Based on the wage data in the World Bank Enterprise Survey, Lao garment workers earn more than those in the food, wood and textile subsectors, but significantly less than in more skilled/higher value-added sectors such as chemicals and electronics. The data also indicate that exporting firms pay more: Lao garment workers in exporting firms earn 70 percent more on average than garment workers in non-exporting firms.20 This wage disparity is not observed in the full sample, although exporting firms do tend to pay higher wages than non-exporting firms, and this holds for most sectors (columns 5 and 6 of Table 4). Table 4: Labor market characteristics of exporting vs. non-exporting manufacturing firms Lao PDR Full Sample 1 2 3 4 5 6 % o/w in % o/w in % o/w in % o/w in Total Non-Exporters Exporters Total Non-Exporters Exporters Employment share (%) Food 5 54 46 15 53 47 Wood products & furniture 5 44 56 3 44 56 Non-metallic and plastic materials 8 11 52 48 Metals and machinery 1 17 52 48 Leather 1 73 27 Textiles 2 57 43 14 24 76 Garments 69 2 98 18 12 88 Electronics 3 5 37 63 Chemicals and pharmaceuticals 0 9 40 60 Auto and auto components 7 4 35 65 Other manufacturing 4 33 67 Median wage (current USD) Total Non-Exporters Exporters Total Non-Exporters Exporters Food 3,199 2,724 3,548 4,863 4,630 5,664 Wood products & furniture 2,008 2,042 1,692 4,270 3,005 8,016 Non-metallic and plastic materials 4,737 4,737 7,716 6,539 10,087 Metals and machinery 5,559 5,559 8,310 6,557 10,957 Leather 5,641 5,545 6,203 Textiles 3,007 4,895 1,716 5,427 4,812 6,559 Garments 4,339 2,632 4,512 5,137 5,013 5,273 Electronics 35,483 35,483 9,190 8,514 10,958 Chemicals and pharmaceuticals 7,802 7,802 7,563 5,342 10,175 Auto and auto components 5,797 5,797 7,934 7,272 9,442 Other manufacturing 8,384 7,982 10,715 Source: World Bank Staff calculations based on data from World Bank enterprise surveys for 2012 and 2013. Full sample of countries includes China, Belarus, Nepal, West Bank and Gaza, Georgia, Azerbaijan, Kyrgyz Republic, Kazakhstan, Bangladesh, Armenia, Uganda, Ukraine, Uzbekistan, Mongolia, Moldova, Tanzania, Kenya, Bulgaria, Serbia, Zambia, Bosnia and Herzegovina, Albania, Kosovo, Croatia, FYR of Macedonia, Slovenia, Montenegro, Djibouti, Jordan, Romania, Latvia, Democratic Republic of the Congo, Tajikistan, Estonia, Hungary, Lithuania, Poland, Slovak Republic, Turkey, Ghana and Madagascar. Data for China and Lao PDR are for 2012. 20 This is likely driven in part by longer working hours; moreover, the calculation reflects no controls for worker characteristics such as education. 30 61. When we consider the sectoral distribution of skilled and unskilled workers in Lao PDR, we find that most skilled workers in Lao PDR’s manufacturing sector are employed in the garment sub-sector, but so are most unskilled manufacturing workers (see Figure 15). The sectoral ratios of skilled-to-unskilled workers is highest in metals and machinery (82 percent are skilled), followed by the chemicals sector (59 percent), and garments (56 percent). Cambodian manufacturing employment is even more concentrated in garment production, which accounts for 93 percent of all manufacturing employment and 77 percent of all skilled manufacturing employment (see Table 5). Despite Vietnam’s more diversified production, the largest share of its skilled manufacturing workers – 23 percent – are engaged in garment production. Figure 15: Distribution of Lao Manufacturing Workers by Skill Level, 2012 Source: World Bank Staff calculations based on data from World Bank enterprise surveys for 2012 for Lao PDR. Note: Skilled and unskilled shares of manufacturing labor are measured on the left axis; percent of sectoral labor that is skilled is measured on the right axis. 31 Table 5: Labor shares in skilled and unskilled manufacturing in Lao PDR, Vietnam and Cambodia Lao PDR (2012) Vietnam (2009) Cambodia (2007) % o/w % o/w % o/w % o/w % o/w % o/w Total Skilled Unskilled Total Skilled Unskilled Total Skilled Unskilled Employment share Food 5 37 63 11 41 59 1 45 55 Non-metallic and plastic materials 8 22 78 12 56 44 1 63 37 Metals and machinery 1 82 18 8 56 44 0 27 73 Textiles 2 53 47 8 62 38 4 59 41 Garments 69 56 44 36 64 36 93 83 17 Electronics 3 48 52 4 76 24 Chemicals and pharmaceuticals 0 59 41 3 39 61 Other manufacturing 12 47 53 18 45 55 1 55 45 Source: World Bank Staff calculations based on data from World Bank enterprise surveys for 2012 for Lao PDR, 2009 for Vietnam and 2007 for Cambodia. 4.2. Estimating the Employment Elasticity of Exporting 62. We test empirically for a correlation between exporting and labor outcomes, and whether there is evidence of causality. We use firm-level data to estimate the quasi-elasticity of employment and wages with respect to the share of exports in total sales for Lao PDR, as well as a global sample of countries (see Annex 5 for a description of the methodology and data). Recent literature has shown that exporters achieve better labor market outcomes than non- exporters by employing more labor and paying higher wages (Bambrilla, Lederman and Porto 2012). Cebeci, Lederman and Rojas (2013) attempt to identify a causal effect using an instrumental variable approach, and find that increases in Turkish firms’ export shares of total sales are associated with increases in employment levels within firms, but a higher export share generates no observed impact on wages. 63. Following this methodology, we use the panel component of the 2009 and 2012 World Bank Enterprise Surveys to test for a causal effect between firm-level changes in export share and firm employment and average wages (see Annex 5, Tables 5.1-5.3 for OLS (non-causal) and IV (causal) results). The regression analysis finds insignificant effects for the Lao PDR sample (likely due to its small size), and statistically significant but small positive correlations between export share and employment (for total, skilled and unskilled) for the global sample of firms. The point estimates are very small: a 1 percentage point increase in the share of a firm’s sales that are exported is associated with only a 0.002 percent increase in total employment. The effect is mildly stronger for unskilled employment. These results suggest that globally, firms that are becoming more export intensive tend to hire more workers, with a preference for unskilled over skilled workers, but the effect is very small. The correlation with female employment and wages, on the other hand, is insignificant. The regressions testing for a causal relationship also yield insignificant results, although the magnitudes of the point 32 estimates are notably higher. Moreover, there appears to be no significant “Lao-specific effect”.21 64. Given Lao PDR’s high export-orientation, and the small but positive correlation between exporting and firm size, policies that help producers access new external markets may help foster some increased labor demand. A range of pro-trade policies should be considered, including joining knowledge platforms such as international exporters’ associations and fora, particularly electronic networks that are easier to access and facilitate knowledge sharing, improving standards compliance and certification within Lao PDR in line with international standards, reducing trading costs through measures to improve trade facilitation systems and logistics, and providing enterprises with support for marketing assistance and training, for example. 4.3. Impact of the Business Environment on the Export-Employment Elasticity 65. How does the business climate affect the size of the export-employment elasticity? Does the business environment jointly affect firms’ exporting and employment decisions? Seker (2010) finds evidence that in countries where firms face more constraints to creating jobs, a smaller fraction of firms participate in export markets, and among firms that export, those that find it more costly to create new jobs export less intensively. If exporting rather than selling to the domestic market is linked to greater job creation, then inefficiencies in the business environment are expected to limit the positive impact of exporting on labor market outcomes. We test this hypothesis using the World Bank Business Environment and Enterprise Performance Survey’s 60-country dataset (surveys were conducted between 2002 and 2013) 22 and following the methodology in Lopez and Shankar (2011) to build a panel that includes countries where variation in competiveness indicators is observed (see Annex 5 for details of the methodology).23 Competitiveness indicators include institutions, macroeconomic environment, higher education and training, market efficiency, and business sophistication (taken from the World Economic Forum’s Global Competiveness Report). 66. When we consider the role of competitiveness indicators24 on firms’ employment and wage responses to exporting, we find evidence that firms that increase their export share tend to 21 The interaction of a dummy variable for firms from Lao PDR with the export share variable yields insignificant results (Annex 5 Table 5.3). 22 Sample of countries includes: Afghanistan, Albania, Angola, Argentina, Armenia, Azerbaijan, Bangladesh, Belarus, Bolivia, Bosnia and Herzegovina, Botswana, Bulgaria, Burkina Faso, Cape Verde, Cameroon, Chile, Colombia, Democratic Republic of the Congo, Croatia, Czech Republic, Ecuador, El Salvador, Estonia, Georgia, Guatemala, Honduras, Hungary, Kazakhstan, Kenya, Kyrgyz Republic, Lao PDR, Latvia, Lithuania, FYR of Macedonia, Malawi, Mali, Mexico, Moldova, Morocco, Nepal, Pakistan, Panama, Paraguay, Peru, Poland, Romania, Russian Federation, Rwanda, Senegal, Serbia, Slovak Republic, Tajikistan, Tanzania, Turkey, Uganda, Ukraine, Uruguay, Uzbekistan, Venezuela, and Zambia. 23 In other words, the impact of exports on employment will be allowed to depend on the level of the competiveness indicators in each country by also interacting exports with these indicators in the model’s specification. It will then be possible to assess the employment response to changes in exports conditional on different levels of competiveness. 24 Both aggregated into a composite competitiveness indicator and as separate competitiveness indicators. 33 have higher employment levels (as noted above), but the effect is more muted in economies with better overall competitiveness and macroeconomic environments, decreasing the elasticity by 0.25 percent and 0.37 percent, respectively (Annex 5 Table 5.4). This could be interpreted to mean that in countries with more competitive environments, exporting firms tend to shift to less labor-intensive exports. The regression coefficient magnitudes are not enough to make the overall relationship negative. The business sophistication indicator is significantly and positively correlated with skilled employment, but with a small coefficient value; for a given level of market efficiency, firms that increase the share of exports in total sales by 1 percentage point also have lower levels of skilled employment by 0.022 percent. But the presence of enhanced market efficiency partly offsets the effect (as reflected by the positive coefficient on the interaction term); increasing the market efficiency index by 1 unit is associated with an increased elasticity of -0.016. With respect to female employment and wages, the presence of different aspects of competitiveness in an economy appears to have no correlation. 67. The causal regression analysis suggests that the degree of competitiveness has an impact on skilled and female employment and wages (reported in Annex 5 Table 5.5). Increasing the share of exports in total sales by 1 percentage point causes firms in countries with a given level of macroeconomic competitiveness to employ fewer female workers by about 0.34 percent, while an improvement in the macroeconomic environment leads to net lower female employment by 0.26 percent (due to the positive coefficient on the interaction term). This is also the case for the higher education and training indicator, albeit with a smaller effect on the elasticity estimates and interaction term. Controlling for the level of institutions, improving the institutional environment reduces the positive relationship between exporting and skilled employment. But we observe no statistically significant effect on total and unskilled employment or wages. 68. Taken together, the elasticity analysis generates some evidence that exporting rather than selling to the domestic market is positively associated with employment levels (both skilled and unskilled). We also find evidence that competitiveness can affect the skilled employment levels of exporters, albeit negatively for countries with better institutions, and positively for countries with higher education and training. We find no evidence that these results do not extend to Lao PDR, suggesting that for firms in Lao PDR, increasing exports should lead to increased labor demand, but the skills mix of this increased demand will be affected by the competitiveness of institutions and the business environment. 34 Chapter 5: Labor Mobility and Adjustment Costs 69. Open economies are subject to a variety of external shocks. Changes in a country’s trade policies (such as a reduction in tariffs as a result of WTO accession or joining a regional trade agreement) or changes in prices in global markets can affect the relative prices faced by domestic firms that consume or produce traded goods, which in turn affects the relative demand for labor across industries. The resulting labor reallocation between sectors in response to these types of trade-related shocks has wage implications as well. Whereas the destruction and creation of jobs is fundamental to the efficient allocation of resources, the speed with which labor markets can adjust in response to trade-related shocks depends on labor mobility and the associated mobility costs. If labor reallocations across industries and employers are costly, the labor adjustment will take more time or will be more muted than expected. Idle labor or labor stuck in lower productivity firms and industries implies lost incomes and diminished output gains. Frictions to labor mobility, therefore, can translate into significant labor adjustment costs by reducing the potential gains to trade. Evidence suggests that the forgone gains to trade due to frictions in labor mobility can be substantial (Davidson and Matusz 2010). 70. This chapter analyzes the circumstances under which a trade shock can benefit or harm workers through changing labor demand and the resulting impact on employment and wages. It applies a recently developed model – the World Bank’s Trade and Labor Adjustment Costs Toolkit25 – to assess the potential distributional effects of future trade-related shocks. The impact will depend on the distribution of employment, skill sets and wages across sectors, and the relative impact of a trade shock on particular sectors. Even for a favorable shock, not all workers will benefit equally, and some workers may lose, in the case of liberalization that harms a sector’s competitiveness and leads to diminished labor demand and a fall in sectoral wages. 5.1 Defining Labor Mobility Costs and Labor Adjustment Costs 71. The analysis distinguishes between two key labor market concepts to analyze the costs of labor reallocations that follow shocks, the ways in which mobility frictions affect workers’ employment decisions, and the collective impact of worker responses in terms of net employment, wage and welfare gains. The analysis considers permanent shocks that result in economy-wide reallocations in the long run, rather than transitory and other short-run business cycle shocks.  Labor mobility costs are defined as the costs perceived by a worker to move to a different firm or industry, independent of the reason for the move. Labor mobility refers to the ability of workers to move between firms and industries in search of alternative employment opportunities, such as in response to wage differences. Because workers incur a cost to move, 25 To better understand the tension between trade openness and employment, the Trade Group in the World Bank developed a toolkit to analyze the transmission channels from trade policies to labor markets, and the magnitudes of their impact. Understanding the tension between trade openness and employment is increasingly important to development. After two decades of global trade liberalization and other structural reforms, most developing countries today are open and well integrated into the global economy. While openness helps diversify risk and is linked to growth and poverty reduction, it can also have destabilizing effects. See Hollweg et al. (2014). 35 this “labor mobility cost” reflects what a worker perceives to be his or her welfare cost of finding alternative employment. Typical impediments to job switching are skills mismatches (wages forgone because of lower productivity), limitations to geographic mobility (administrative procedures for internal migration26 and direct relocation costs), and severance and hiring costs (including those imposed by law or convention). Other factors may be location preferences, job search costs, and even the psychological costs of changing jobs.  Labor adjustment costs are the total costs incurred by workers and the economy as a result of a shock when labor mobility is costly. These adjustment costs depend on the magnitude of labor mobility costs, as mobility frictions affect workers’ labor reallocation decisions. Large mobility costs mean sluggish reallocation of labor in response to a trade-related shock, reducing the potential benefits to workers and the economy as a whole. Labor mobility costs, by contrast, can exist without any adjustment-inducing shock. 72. As a first step, the toolkit estimates labor mobility costs using panel data on workers’ sector of employment, average sectoral wages, and individual worker characteristics to assess worker transitions between or into and out of jobs over time. The toolkit then simulates the economy- wide responses to a trade shock in the presence of costly labor mobility vis-à-vis the impact on employment and wages. These estimates are used to compute labor adjustment costs as the difference between optimal (without mobility costs) and actual worker welfare. Annex 6 provides details of the methodology and data. 5.2. Worker Transitions in Lao PDR 73. Transition matrices provide the basis for estimating labor mobility costs, and we estimate worker transitions between sectors using panel data from LECS-4 and LECS-5 on gross labor flows between sectors and different states of work (including unemployed and out of the labor market). 27 We distinguish the effects for different worker-types, disaggregating by skill level and gender, and we include the entire labor force including self-employed and own account workers. See Annex 6 Tables 6.1-6.3 for summary statistics on sectoral labor shares and average wages by type of worker. 74. The Lao labor market is relatively fluid. Each cell in Table 6 reports the share of workers transitioning from each origin sector (row) to all other destination sectors (column). (Annex 6 Table 6.4 contains the number of workers transitioning between each category.) The cells on the diagonal indicate the shares of workers remaining in their current work/sector status; these “stayers” represent the largest shares. The transition statistics give a sense of the fluidity of the Lao labor market, even during a relatively short timeframe, that is, between 2007/08 and 2012/13. In fact, the share of “stayers” in Lao PDR is noticeably smaller than in other countries, suggesting more labor market fluidity. Agriculture has the highest incidence of “stayers”, suggesting it is hardest for agriculture workers to find alternative employment in other sectors of the economy. 26 See World Bank (2014b) Box 3.2 for a description of the bureaucratic barriers facing Lao workers wanting to move to a different province. 27 Although the LECS survey was not designed as a panel (at either the household or individual level), it is possible to track households over the 2007/08 and 2012/13 waves, and subsequently individuals within those households based on demographic characteristics. 36 75. Agriculture also acts as the residual or “last resort” sector, absorbing workers unable to find alternative employment. The agriculture sector absorbed most unemployed or inactive workers (48 percent of those exiting unemployment and 33 percent of new labor force entrants flowed into agriculture). Only one quarter of those unemployed at the beginning of the period remained unemployed in the 2012/13, and nearly all inactive workers had entered into some type of employment by 2012/13 (Table 6). Table 6: Worker transitions across sectors and into and out of labor force status between 2007/08 and 2012/13 (%) Other manufacturing Out of labor force Unemployed Agriculture Garment Services Mining Unemployed 4.35 17.39 47.83 0.00 0.00 0.00 30.43 Out of labor force 1.09 38.04 33.45 0.12 2.66 2.78 21.86 Agriculture 0.19 11.33 73.00 0.16 1.13 2.94 11.25 Mining 0.00 0.00 70.00 0.00 0.00 0.00 30.00 Garment 0.00 11.67 37.50 0.83 30.83 2.50 16.67 Other manufacturing 0.33 9.27 34.44 0.33 1.32 28.81 25.50 Services 0.73 7.98 22.88 0.48 1.21 4.51 62.21 Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey. 76. Public sector service jobs have a low share of exits, and a low rate of entry, suggesting that these jobs are highly desirable despite lower wages, but are also difficult to obtain. We observe the largest number of transitions out of financial services work in Lao. 77. Wholesale and retail trade, and hotels and restaurants are important stepping stone sectors out of agriculture. In Lao PDR, workers exiting the agriculture, forestry and fishing sector are more likely to find jobs in commercial trade, hotels and restaurants than other sectors of the economy. And for workers exiting the wholesale and retail trade, hotels and restaurants sectors, they are most likely remain in services, the majority switching to community, social and individual services. 78. If leaving a sector (other than agriculture) in Lao PDR, workers were more likely to enter agriculture than unemployment. This is not generally the case in other countries, where workers were more likely to drop into unemployment or out of the labor force if leaving the manufacturing, trade, finance or public service sectors than become agriculture workers, for example. 79. Skilled workers have a higher propensity to find jobs in the services sector, whereas unskilled workers are more likely to enter the agriculture sector (Table 7). This is true for the unemployed and the newly active in the labor market, but it is also true for workers switching sectors of work. For example, 51 percent of skilled workers exiting agriculture find services jobs, 37 as do 73 percent of skilled workers exiting garments, and 48 percent of those exiting other manufacturing. For unskilled workers, these numbers are significantly lower. In addition, fewer skilled services workers exit services, whereas unskilled services workers were almost as likely to exit as to remain in the sector. Table 7: Worker transitions across sectors and into and out of labor force status between 2007/08 and 2012/13 (%), skilled vs. unskilled Skilled Unskilled Other manufacturing Other manufacturing Out of labor force Out of labor force Unemployed Unemployed Agriculture Agriculture Garment Garment Services Services Mining Mining Unemployed 0.00 0.00 0.00 0.00 0.00 0.00 100.00 0.00 22.22 50.00 0.00 0.00 0.00 27.78 Out of labor force 4.17 16.67 6.25 1.04 1.04 5.21 65.63 0.90 27.87 43.60 0.00 3.37 3.15 21.12 Agriculture 1.52 4.55 37.12 1.52 0.00 4.55 50.76 0.15 8.67 75.85 0.11 1.21 2.90 11.10 Mining 0.00 0.00 66.67 0.00 0.00 0.00 33.33 Garment 0.00 9.09 9.09 0.00 9.09 0.00 72.73 0.00 10.00 37.78 1.11 34.44 3.33 13.33 Other manufacturing 0.00 17.39 8.70 0.00 0.00 26.09 47.83 0.40 6.72 34.78 0.40 1.58 31.23 24.90 Services 1.85 4.31 8.00 0.00 0.31 3.38 82.15 0.37 8.60 28.13 0.74 1.60 5.16 55.41 Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey. 80. Female workers entering the garment sector are most likely to come from out of the labor force or agricultural work (Table 8). The high number of exits from garment work is even higher than those having ‘stayed’ in the garment sector, confirming very high job turnover. However, garments do not appear to be a stepping stone sector; workers exiting garments are most likely to return to agriculture (even more likely than to stay in garments). This points to the need for a change in the incentive structure of garment jobs (such as compensation policies, working conditions, or career development) to entice workers to remain in relatively productive work rather than returning to the farm or migrating. Policies could center on skill acquisition and upgrading to facilitate the shift to higher value-added activities, or support for career development, inter alia. 81. Despite the highly feminized work force in garments, more female workers enter other manufacturing or service sectors, which are significant employers of women. Female workers exiting agriculture are least likely to enter mining, and are in fact more likely to exit the labor force or enter services or other manufacturing rather than the garment sector, implying that garment jobs mostly attract workers who lack other options. 38 Table 8: Worker transitions across sectors and into and out of labor force status between 2007/08 and 2012/13 (%), male vs. female Male Female Other manufacturing Other manufacturing Out of labor force Out of labor force Unemployed Unemployed Agriculture Agriculture Garment Garment Services Services Mining Mining Unemployed 0.00 0.00 57.14 0.00 0.00 0.00 42.86 6.25 25.00 43.75 0.00 0.00 0.00 25.00 Out of labor force 1.96 36.93 35.62 0.00 0.33 3.59 21.57 0.57 38.70 32.18 0.19 4.02 2.30 22.03 Agriculture 0.26 8.58 74.66 0.26 0.15 3.37 12.72 0.11 14.39 71.16 0.06 2.23 2.46 9.59 Mining 0.00 0.00 40.00 0.00 0.00 0.00 60.00 0.00 0.00 100.00 0.00 0.00 0.00 0.00 Garment 0.00 0.00 0.00 12.50 37.50 0.00 50.00 0.00 12.50 40.18 0.00 30.36 2.68 14.29 Other manufacturing 0.53 5.88 37.97 0.53 1.07 28.34 25.67 0.00 14.78 28.70 0.00 1.74 29.57 25.22 Services 0.73 6.46 27.46 0.73 0.15 4.85 59.62 0.71 9.82 17.32 0.18 2.50 4.11 65.36 Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey. 5.3. Labor Mobility Costs 82. It is not possible to measure labor mobility costs directly, because they are not readily observable. Instead, we use an indirect method that combines the observed worker transitions between sectors with the inter-sectoral wage gaps to estimate the “labor mobility cost” to explain why workers do not transition into higher wage sectors to the extent that wage gaps are eliminated. 28 The methodology of this model differs from standard models of trade, which assume no frictions in the adjustment of labor. For example, workers are assumed to move without frictions from import-competing sectors to export-expanding sectors, after an economy opens to trade. But in reality, workers respond slowly to trade-related shocks. We estimate the welfare costs for a worker to switch industries/jobs using a dynamic rational-expectations model of costly labor adjustment. In each period, a worker can choose to move from her current industry to another one, but must pay a cost in doing so. The decision for a worker to move depends on her expected welfare gain, net of the cost of moving. When there are low transition rates across sectors of the economy despite high wage gaps, we interpret this to mean that it is costly for workers to move. So for a sector that is difficult to access, we would assign a high labor mobility cost for entering that type of job. Combining the transition data with the observed wage gaps between sectors leads to estimates of labor mobility costs for entering each sector, expressed as a ratio of the annual average wage.29 These transition costs are estimated for all workers, as well as for skilled and unskilled workers, and for men and women (see Table 9).30 28 Note that there may be multiple reasons why wage gaps persist, e.g., on the labor supply side (when it is physically or technically difficult due to skills mismatch to obtain a job in a certain sector), or on the labor demand side (such as a lack of new public sector job openings). Both are obstacles to matching labor supply with labor demand at a market clearing wage, and both therefore imply a high cost of transitioning to said job. 29 The cost has a common component, which does not vary across similar type workers in an industry, and an idiosyncratic component. By solving the dynamic rational-expectations model of costly labor adjustment, it is 39 83. Understanding which sectors are the least costly to enter, and for different types of workers, will shed light on which sectors may provide opportunities following a trade shock. The methodology used here is sensitive to the level of disaggregation: the higher degree of sectoral disaggregation, the fewer the observed transitions, the higher the estimated mobility costs. It is therefore crucial to exercise caution when interpreting the results. This type of exercise is most useful for comparing the relative mobility costs across different sectors of an economy – particularly those most closely affected by trade shocks – and different types of workers. Table 9. Labor mobility costs (as a share of variance of the idiosyncratic component, expressed as a ratio of the average annual wage) Total Skilled Unskilled Male Female Unemployed / out of labor force 0.879 0.865 0.695 1.213 0.549 Agriculture / mining 1.169 1.652 1.140 1.082 1.282 Garment 2.855 3.028 2.914 5.002 2.181 Other manufacturing 1.878 1.821 1.970 1.698 2.114 Services 1.586 0.503 1.341 1.378 1.837 Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey. Note: To estimate labor mobility costs, it was necessary to combine mining with agriculture due to too many zeros in the transition matrix and non-convergence in the estimates. For robustness, estimates were re-calculated excluding the mining sector altogether. Overall, the rankings across sectors remain the same, and the point estimates change only slightly in value. For example, the residual sector of unemployed / out of labor force is still the least costly sector for total, unskilled and females to enter, while agriculture is the least costly sector for male workers to enter, and services is the least costly sector for skilled workers to enter. 84. Labor mobility costs are significant in Lao PDR, averaging 1.28 times the average annual wage. Labor mobility costs in Lao PDR range from 1.17 times the average annual wage to enter agriculture or mining, to 2.8 times the average annual wage to enter garment manufacturing.31 These are however lower than in other ASEAN countries, where labor mobility costs range upwards of six times the average wage for some sectors. These results are consistent with more churning in the Lao labor market, relatively low unemployment rates, and a much lower incidence if “stayers”. Workers faced with these welfare costs will certainly weigh the decision to change jobs very carefully, even when the incentives – such as higher wages – are strong. 85. High labor mobility costs are typical for countries with a large share of primary employment. Arias et al. (2013) find evidence of a positive correlation between labor mobility costs and the employment share in the agriculture sector, suggesting that for countries in which employment opportunities are more diversified away from primary sectors, workers encounter lower mobility costs. They also find evidence that although educational attainment at the primary possible to derive an equilibrium condition that is a kind of Euler-equation, which lends itself to estimation. The structural model identifies workers’ transitions across sectors of an economy that depend on the wage gaps between those sectors and the mobility costs of entering a sector. From observed data on sectoral transitions and sectoral wage gaps, we are able to estimate the labor mobility cost using GMM-type estimations. 30 Note that we do not capture workers switching jobs within sectors, nor do we capture workers changing jobs multiple times between 2007/08 and 2012/13. 31 We include all individuals engaged in the agricultural sector in the transitions, regardless of whether they are wage workers, own-account farmers, or non-wage family farm workers. However, the average agricultural wage is calculated for wage earners only. This could bias the average agricultural wage either upward or downward, which may in turn bias the labor mobility cost estimate for entering agriculture. 40 and secondary levels does not seem to matter for mobility costs, countries with higher levels of tertiary education tend to have lower mobility costs. 86. Skilled workers face lower costs to transition across sectors compared to unskilled workers, although the results vary by sector. In Lao PDR, unskilled workers face higher labor mobility costs to enter other manufacturing and services. This suggests that employers may prefer workers with sector-specific knowledge, creating obstacles for agriculture workers trying to switch into more productive sectors, for example. 87. Female workers face greater costs to transition across sectors compared to male workers, although this is not true for garment factory work. In general, labor mobility costs are higher for female than male workers, consistent with the observed low number of worker transitions by women. But the very low share of men in garment work yields a very high mobility cost for men to enter this particular sector. 88. Leaving employment either by becoming unemployed or exiting the labor force involves the lowest transition cost for women, but not for men. For example, female workers face a labor mobility cost equivalent 0.55 times the average annual wage to exit employment – less than half the cost of entering any other sector. For male workers, however, it is less costly to enter agriculture and mining sector than to become unemployed or exit the labor force; this result arises from the fact that very few men choose unemployment or exit the labor force. 89. The (combined) agriculture and mining sector is one of the least costly to enter for all groups, and particularly for males and the unskilled. In Lao PDR, workers entering agriculture incur a cost equivalent to 1.17 times the annual average wage – driven up by mining and quarrying which are more costly for workers to enter. The labor mobility cost for unskilled workers to enter agriculture is 1.14 times the annual average wage, versus 1.65 for skilled workers; and the labor mobility cost for male workers is 1.08 times the annual average wage, compared to 1.28 for female workers. (Table 6.3 in Annex 6 reports the average sectoral wages for 2012/13 for different worker types.) 90. Workers in Lao PDR face the highest labor mobility costs to enter garment work; this counterintuitive result is explained by the very small share of total employment in garment manufacturing. Even though we observe high-frequency turnover rates, the number of transitions is very small compared to the worker flows into and out of other sectors or work status. Moreover, the panel data measures whether a worker has transitioned between two points in time, 2007/08 and 2012/13, and does not capture high-frequency transitions. In other countries, skilled labor usually incurs a lower labor mobility cost to entering manufacturing jobs, and this is also the case in Lao PDR for other manufacturing (excluding garments). 91. Some services sectors are more costly for workers to access than others. Social and public sector service jobs, for example, have high mobility costs, especially among unskilled workers, and this is consistent with the very low observed turnover in public sector jobs. Financial, real estate, insurance and business services are also costly to enter, but more so for unskilled workers than skilled workers. 41 5.4. Labor Adjustment Costs 92. The presence of high labor mobility costs dissuades labor supply to certain sectors, but these sectors may be the most productive and with the highest wages. The resulting sub-optimal labor supply decisions by workers can add up to significant welfare losses for the economy at large. Policies to reduce labor mobility costs – such as labor market information or matching services, or cost-share training that is demand-driven 32 – would reduce distortions affecting labor supply decisions, increase worker mobility, and in the case of a shock or other change in market signals, boost and accelerate the labor market response. 93. Using estimates of labor mobility costs, the analytical framework allows us to quantify the potential employment and wage effects of a hypothetical trade shock to a specific sector. This type of simulation of post-shock adjustment dynamics can be useful to policymakers and policy designs for mitigating or offsetting large anticipated labor adjustment costs of worker transitions between sectors over time when labor mobility is costly. 33 In this simulation analysis, we model an exogenous trade shock that translates into a 20 percent increase in the price of garments emanating from international markets.34 This could result from liberalizing the Lao garment sector through expanded access to global markets where global prices are 20 percent higher than Lao prices. This equilibrium analysis illustrates how such a shock could affect jobs and wages not only in the garment sector but in the rest of the economy due to shifts in relative labor demand and supply.35 The analysis also allows for movements in and out of unemployment or the labor force. Figure 16 illustrates the impact of the shock for the total economy over time; the graphs show the change in relative wages (the average sector’s wage relative to the average economy-wide wage), employment and welfare for each sector (services, garments, other manufacturing, agriculture/mining, and the residual sector which comprises unemployment and out of the labor force).36 94. Workers in the garment sector would see real wage increases immediately after the shock, while wages of workers in all other sectors remain relatively unchanged . The simulation of a 20 percent increase in the price of garments increases the profitability of the garment sector, which translates immediately into higher nominal wages in year 1 of the shock. The change in the economy’s price index is proportional to the garment share; because the share 32 Note that a garment skills development center was established by the Government under the World Bank-led Trade Development Facility precisely to provide skills training to garment factory employees covering both technical and management skills. The cost-sharing program was initially embraced by a significant number of employers, but demand subsequently fell, and the center has been unable to transition to financial independence following the initial subsidy period, despite diversifying its course offerings. 33 The analysis uses an equilibrium model in which the structure of the economy is specified using assumptions about the production function in each sector as well as demand functions and their parameters. Parameters are calculated using data on relative wages, labor and con sumption shares, and labor’s share of output across sectors of the economy. The production and demand functions are then used to calibrate the initial steady state of the economy. Refer to Annex 6 and to Hollweg et al. (2014) for additional information on the methodology. 34 Note that a negative price shock of the same magnitude would lead to the opposite signed effects on sectoral wages and employment. 35 Note that the model does not assume any change in the capital-labor elasticity, such that we are not modeling increased investment or re-investment of capital that might occur following a trade policy shock. 36 For robustness, separate simulations were performed excluding the mining sector. Only minor differences emerged, and therefore are not presented. The results are available from the authors upon request. 42 of garments in household consumption is low in Lao PDR, the impact of the garment price rise on real purchasing power is very small. And because of constant nominal wages elsewhere, real wages in other sectors do not experience significant changes (as reflected by flat real wages outside the garment sector in Figure 16). Figure 16: Post-shock wages, utility and labor shares, total economy Wages Utility 26.5 1 Relative average wage .8 26 Worker utility level .6 25.5 .4 25 .2 24.5 0 0 5 10 15 20 25 0 5 10 15 20 25 Time Time W1 W2 W3 V1 V2 V3 W4 W5 V4 V5 Labor .4 .3 .2 .1 0 0 5 10 15 20 25 Time L1 L2 L3 L4 L5 Note: Sector 1 is agriculture/mining; sector 2 is garments; sector 3 is other manufacturing; sector 4 is services; and sector 5 is unemployed or out of the labor force. The y-axis measures the average relative wage/level of utility/working age population share, and the x-axis shows time measured in years (the price shock occurs at time 0). Relative wages reflect embedded labor mobility costs and therefore do not average 1. Utility is the sectoral average of workers’ lifetime expected utility (see Annex 6 for more detail). Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey and Global Trade Analysis Project (GTAP). 95. The real wage rise in garments incentivizes workers to reallocate to garments, due to the change in intersectoral wage differentials. The relative rise in garment wages compared to other sectors creates incentives for workers to move into garments, both from other sectors of the economy as well as from unemployment or out of the labor force. Workers are most likely to enter from the agriculture/mining sector as well as the residual sector, although the predicted outflows are very small given that all other sectors of the economy have much larger labor shares than the garment sector (which is why Figure 16 shows little or no visible impact). Recall that labor mobility costs for entering garments are high, which may also be preventing workers from 43 entering the sector at optimal levels. The predicted outflow of labor from other sectors would also lead to modest wage increases in those sectors due to the diminished relative labor supply. 96. The predicted labor reallocation has implications for wages in the garment sector, which decline slightly as workers move in over time. In the new steady state, however, real wages of garment workers are higher than in the initial steady state. The larger labor shares of the other sectors mean that any labor outflows in general would have minimal additional effects on those sectors’ equilibrium wages. 97. The welfare – or utility – of all workers is predicted to increase under the simulated scenario, most notably for those in the garment sector. This is because of higher real wages, which increases the purchasing power of all workers, albeit modestly. In addition, garment workers also receive higher nominal wages. 98. The adjustment period is predicted to be very short in the simulation, due to Lao PDR’s relatively low labor mobility costs. It takes only about two years to reach the new steady state. Simulations of trade-related price shocks in other countries suggests that the average adjustment period can be much longer – six years on average for developing countries. The shorter adjustment period reflects the fluidity of the Lao labor market. 99. The results of the above simulations are similar when disaggregated between skilled and unskilled workers and between men and women, with a few notable differences (see Annex 6 Figures 6.1-6.4). Unskilled workers in the garment sector experience a larger relative nominal wage increase immediately following the shock compared to skilled garment workers, and this translates into relatively larger utility gains for unskilled garment workers. And with respect to gender differences, the positive effects of the higher price of garments accrues primarily to female workers, since it is primarily women who move into the garment sector, consistent with actual employment trends in garments. This gender difference stems from the very high estimated labor mobility cost for males to enter the sector, itself a function of the small current share of men in garment work. Females are predicted to move into garments from all other sectors except services, including from unemployment and from out of the labor force. Whereas female labor shares in other sectors except services decline following the price shock, the sectoral labor shares for men are not affected. 100. Labor mobility costs reduce both the speed of transition to the new steady state and the gains from trade. The welfare losses resulting from this slower and smaller adjustment, referred to as labor adjustment costs, are calculated as the difference in aggregate worker welfare between optimal labor allocations after a trade shock (with instantaneous adjustment when mobility costs are zero) and actual labor allocations (when labor mobility costs are positive). 37 101. The magnitudes of the actual and potential welfare gains from the trade shock to garment prices (respectively reported in Tables 10 and 11) vary by sector and worker type. The actual welfare gains are greatest for female workers, averaging around 1 percent of initial welfare 37 Labor adjustment costs are calculated using the simulated adjustment dynamics after the trade-related shock to the garments sector. The adjustment dynamics of the transition from the pre-shock to post-shock steady states produce levels of worker welfare associated with predicted labor market outcomes. Labor adjustment costs are calculated by solving for this welfare when there are mobility costs and comparing it to welfare when mobility costs are zero. 44 (Table 10). Female garment workers show the largest welfare gains, equivalent to 1.3 percent of their initial welfare, due to the low initial real relative wages and therefore the greatest welfare benefits from real wage increases. Male workers, by contrast, face negative gains from trade, given by the higher prices of garments and lower employment opportunities in the garment sector. Unskilled workers benefit more than skilled workers, with unskilled garment workers benefiting more than other sectors. Overall, the net welfare effect of this type of trade liberalization in garments is positive, although there are predicted winners and losers. Table 10: Actual welfare gains from the trade shock (% of initial welfare) Total Skilled Unskilled Male Female Agriculture / mining 0.241 0.046 0.232 -0.187 0.893 Garment 0.624 0.329 0.603 -0.181 1.303 Other manufacturing 0.243 0.045 0.234 -0.187 0.895 Services 0.265 0.057 0.256 -0.187 0.971 Unemployed / out of labor force 0.265 0.058 0.255 -0.183 0.952 Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey and Global Trade Analysis Project (GTAP). 102. Workers in Lao PDR could gain even more from a positive price shock to garments if there were no labor mobility costs. The simulations confirm that potential gains from trade for workers are always positive with the exception of male workers, and again the magnitudes vary by sector and worker type (Table 11). In addition to benefiting the most in actual welfare gains, female workers also have the highest potential gains from the trade shock, particularly female garment workers. Table 11: Potential welfare gains from the trade shock (% of initial welfare) Total Skilled Unskilled Male Female Agriculture / mining 0.246 0.048 0.237 -0.187 0.904 Garment 0.584 0.308 0.565 -0.181 1.267 Other manufacturing 0.248 0.047 0.239 -0.187 0.906 Services 0.266 0.057 0.256 -0.186 0.964 Unemployed / out of labor force 0.266 0.058 0.255 -0.183 0.951 Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey and Global Trade Analysis Project (GTAP). 103. The predicted labor adjustment costs – the difference between potential and actual gains – are very low in Lao PDR, at less than 0.01 percent of initial welfare (Table 12). This estimate suggests that the Lao labor market is relatively agile in responding to exogenous price shocks in the garment sector. But when adjustment costs are expressed as a share of total potential gains from trade (consistent with the literature), the forgone gains to trade due to frictions in labor mobility appear to be more substantial. The forgone gains for workers in Lao PDR can be over 6 percent of the potential gains for workers in the garment sector (Table 13). 45 Table 12: Labor adjustment costs (% of initial welfare) Total Skilled Unskilled Male Female Agriculture / mining 0.005 0.002 0.005 0.000 0.012 Garment -0.040 -0.021 -0.038 0.000 -0.036 Other manufacturing 0.005 0.003 0.004 0.000 0.011 Services 0.001 0.000 0.001 0.000 -0.007 Unemployed / out of labor force 0.001 0.000 0.001 0.000 -0.001 Table 13: Forgone gains from trade (labor adjustment costs as a % of potential gains to trade) Total Skilled Unskilled Male Female Agriculture / mining 2.221 4.454 2.075 -0.090 1.283 Garment -6.791 -6.757 -6.777 -0.043 -2.843 Other manufacturing 2.021 6.234 1.820 0.047 1.204 Services 0.268 -0.419 0.223 -0.173 -0.693 Unemployed / out of labor force 0.253 0.263 0.208 -0.047 -0.084 104. Diversifying production and employment away from the primary sector toward industry and services could also reduce labor mobility costs by enabling workers to transition more easily to better jobs. This, in turn, would result in a more efficient allocation of labor across the Lao economy. Moreover, policies to facilitate labor mobility – such as labor information and job matching services – would also contribute to allocative efficiency. 105. These types of policies targeting labor mobility costs could be further complemented by reforms that address distortionary public employment policies such as hiring, compensation, and human resource management within the civil service and SOEs, and revisiting the role of SOEs. If public sector and private sector jobs become equally attractive to workers, this could potentially free up more educated workers to enter the private sector and boost private sector productivity. 106. The various limitations of Lao PDR’s manufacturing sector documented throughout this report – both in terms of size and quality – mean that any positive policy impacts are likely to have only modest effects on aggregate welfare and inclusive growth. To generate bigger bang for the buck, policies will need to move beyond the manufacturing sector to address the limited employment and demand spillovers of mining production, and the limited market integration and value addition in agricultural production. Policies to enhance agricultural productivity, job quality and economic opportunities in rural areas will not only foster greater equity, but will also reduce urban-rural distortions and the associated labor mobility costs, and dampen incentives to migrate. 46 References Arias, J., E. Artuç, D. Lederman, and D. Rojas (2013). “Trade, Informal Employment and Labor Adjustment Costs: Theory and Evidence from Brazil, Mexico, and Morocco.” World Bank, Washington, DC. Artuç, E., D. Lederman, and G. Porto (2013). “A Mapping of Labor Mobility Costs in the Developing World”, Policy Research Working Paper 6556, World Bank, Washington, DC. Brambilla, I., D. Lederman, and G. Porto (2012). “Exports, Export Destinations, and Skills.” American Economic Review, 102(7): 3406-38. Bernard, A., and B. Jensen (1995). “Exporters, Jobs and Wages in US Manufacturing, 1976 - 1987.” Brookings Papers on Economic Activity: Microeconomics, 67-112. Bernard, A., and B. Jensen (1999). “Exceptional Export Performance: Cause, Effect, or Both?” Journal of International Economics, 47(1): 1-25. Bernard, A., B. Jensen, S. Redding and P. Schott (2007). “Firms in International Trade.” Journal of Economic Perspectives, 21(3): 105-30. Calderón, C., and V. Poggio (2011). “Trade and Economic Growth: Evidence on the role of complementarities for CAFTA-DR countries.” In J. H. Lopez and R. Shankar (Eds.), Getting the Most out of Free Trade Agreements in Central America. Washington, DC: The International Bank for Reconstruction and Development / The World Bank, pp. 83- 122. Cebeci, T., D. Lederman, and D. Rojas (2013). “The Structure of Exports across Destinations and Labor-Market Outcomes: An Empirical Case Study of Turkey.” Unpublished. Clerides, S., S. Lach and J. Tybout (1998). “Is Learning by Exporting Important? Micro-dynamic evidence from Colombia, Mexico, and Morocco.” The Quarterly Journal of Economics, 113 (3): 903-47. Darvas (2012). “Real Effective Exchange Rates for 178 Countries: A New Database.” Bruegel Working Paper No. 06, Bruegel, Brussels. Davidson, C. and S. J. Matusz (2010). International Trade with Equilibrium Unemployment. Princeton, NJ: Princeton University Press. Hausmann, R., J. Hwang and D. Rodrik (2007). “What You Export Matters”, Journal of Economic Growth, Spring vol. 12(1), pp. 1-25. 47 Hollweg, C., D. Lederman, D. Rojas and E. Ruppert Bulmer (2014). Sticky Feet: How Labor Market Frictions Shape the Impact of International Trade on Jobs and Wages. Directions in Development: Trade, The World Bank, Washington, D.C. Lederman, D. and W. Maloney (2009) ‘Does Diversification Matter?’ in Newfarmer, Shaw and Walkenhorst, eds. Breaking Into New Market – Emerging Lessons for Export Diversification, Washington, DC: The World Bank. Lopez, J. and R. Shankar, eds. (2011). Getting the Most Out of Free Trade Agreements in Central America. Directions in Development: Trade, The World Bank, Washington, DC. Seker, M. (2010). “Difficulties in Job Creation and Exporting.” Enterprise Note Series No. 12, Enterprise Surveys, The World Bank, Washington, DC. Shirotori, M., B. Tumurchudur, and O. Cadot (2010). “Revealed Factor Intensity Indices at the Product Level”, UNCTAD Policy Issues in International Trade and Commodities Study Series No. 44, United Nations, Geneva. Verhoogen, E. (2008). “Trade, Quality Upgrading, and Wage Inequality in the Mexican Manufacturing Sector”, The Quarterly Journal of Economics, 123 (2):489-530. World Bank (2011a). Lao PDR Development Report 2010 – Natural Resource Management for Sustainable Development, World Bank: Washington, DC. World Bank (2011b). Lao PDR: Investment Climate Assessment – Polices to Promote Growth in the Non Resource Sectors, World Bank: Washington, DC. World Bank (2012a). “Lao Economic Monitor”, World Bank: Vientiane, Lao PDR. World Bank (2012b). “Lao PDR Labor Standards and Productivity in the Garments Export Sector: A Survey of Managers and Workers”, PREM Department, East Asia and Pacific Region, World Bank: Washington, DC. World Bank (2013). “Skills and Knowledge for Greater Growth and Competitiveness in Lao PDR: Skills for Quality Jobs and Development in Lao PDR”, Report no. ACS6593, Human Development Department, Social Protection and Labor Unit, East Asia and Pacific Region, World Bank: Washington, DC. World Bank (2014a). “Lao PDR: Investment Climate Assessment 2014 – Policy Uncertainty in the Midst of a Natural Resources Boom”, Report no. 90960-LA, Trade and Competitiveness Global Practice, East Asia and Pacific Region, World Bank Group: Washington, DC. 48 World Bank (2014b). “Lao Development Report 2014: Expanding Productive Employment for Broad-Based Growth”, Report no. ACS9577, World Bank Group: Washington, DC. World Bank (2014c). “2012 Lao PDR EGRA Survey Report” [mimeo].World Bank (2015). “A Comparative Overview of the Incidence of Non-Tariff Measures on Trade in Lao PDR”, Trade and Competitiveness Global Practice, World Bank: Washington, DC. World Economic Forum (2015). The Global Competiveness Report 2014-2015. http://www.weforum.org/reports/global-competitiveness-report-2014-2015 49 Annexes 50 Annex 1 Table 1.1: Lao PDR’s labor force participation and unemployment rates (%) 2007/08 (ECS) 2010 (LFS) 2012/13 (ECS) LFP rate Unemployment rate LFP rate Unemployment rate LFP rate Unemployment rate Total 85.1 0.4 84.3 0.6 93.5 0.2 Male 87.9 0.4 85.1 0.6 94.0 0.2 Female 82.5 0.5 83.5 0.5 93.0 0.1 Less than primary 88.9 0.2 94.5 0.4 97.0 0.2 Primary 86.7 0.4 95.3 0.5 97.3 0.3 Lower secondary 78.7 0.4 95.3 0.8 97.7 0.3 Upper secondary 80.7 0.6 94.4 0.9 96.7 0.1 Vocational 90.2 1.2 94.8 1.4 95.4 0.1 Tertiary 80.7 1.4 95.6 2.6 98.1 0.3 Urban 79.1 0.7 77.3 1.1 92.0 0.2 Rural 88.0 0.3 87.6 0.4 94.2 0.2 Vientiane Capital 76.4 0.6 77.9 1.8 96.2 0.1 Vientiane 83.5 0.5 85.5 0.4 89.8 0.3 Savannakhet 81.5 0.5 88.2 0.3 91.5 0.3 Champasak 89.3 0.5 86.0 1.1 96.8 0.0 Khammuane 84.3 0.1 84.7 0.3 97.2 0.1 Luangprabang 86.6 0.1 85.5 0.5 91.0 0.0 Xayabury 87.5 0.8 86.0 0.2 91.7 0.8 Other regions 88.5 0.3 83.6 0.3 93.6 0.2 Source: 2007/08 and 2012/13 data are from the Lao PDR Earnings and Consumption Survey, and 2010 data are from the Lao PDR Labor Force and Child Labor Survey. 51 Table 1.2: Distribution of educational attainment38 for different population groups 2007/08 Less than Primary Lower- Upper Vocational Tertiary primary secondary secondary Male 27.8 35.0 17.5 10.6 6.2 2.9 Female 33.7 36.2 15.2 8.4 4.6 1.9 Agriculture 38.2 39.7 14.6 5.2 1.8 0.5 Mining 18.2 59.9 10.1 3.7 8.1 Garment 16.3 31.2 28.2 17.0 6.0 1.2 Other manuf. 24.7 32.6 20.8 13.6 6.3 2.0 Services 12.5 25.1 19.6 19.9 15.1 7.9 Urban 16.1 29.2 20.1 17.7 11.0 6.0 Rural 37.5 38.5 14.8 5.7 2.8 0.7 Vientiane Capital 11.0 24.3 18.3 23.1 13.0 10.4 Vientiane 19.8 38.4 22.3 10.3 6.9 2.4 Savannakhet 31.4 32.0 16.5 10.3 6.4 3.3 Champasak 29.0 40.8 17.9 7.2 4.6 0.5 Khammuane 38.4 36.0 14.8 7.6 2.7 0.5 Luangprabang 35.4 36.1 13.3 9.2 5.2 0.8 Xayabury 17.1 50.3 19.9 8.7 3.2 0.9 Other regions 41.7 35.4 14.1 5.2 3.0 0.6 2012/13 Less than Primary Lower- Upper Vocational Tertiary primary secondary secondary Male 18.6 37.5 20.0 10.6 6.7 6.5 Female 26.7 38.6 15.8 8.8 5.1 5.1 Agriculture 30.1 45.8 16.4 4.9 1.8 0.9 Mining 17.5 33.9 28.7 8.3 2.3 9.2 Garment 16.0 44.9 18.2 15.9 4.3 0.7 Other manufacturing 13.6 35.8 27.1 13.6 6.4 3.4 Services 9.6 23.2 19.5 17.9 13.8 16.1 Public 3.2 4.7 10.3 14.5 28.8 38.5 SOE 6.5 17.4 19.9 19.2 16.8 20.1 Civil servant 2.7 2.9 8.9 13.9 30.5 41.1 Private 21.4 39.8 20.4 10.5 4.5 3.4 Urban 10.9 26.0 20.2 18.1 11.2 13.5 Rural 28.0 44.0 17.1 5.6 3.3 2.1 Vientiane Capital 7.1 23.2 16.1 26.1 10.9 16.5 Vientiane 14.8 34.9 25.4 11.6 8.4 5.0 Savannakhet 25.9 37.6 17.5 8.4 5.4 5.1 Champasak 22.3 43.4 19.2 6.4 5.1 3.5 Khammuane 24.0 43.3 19.2 6.5 4.1 2.8 Luangprabang 31.0 39.3 14.5 6.4 5.8 3.0 Xayabury 12.7 48.6 23.7 7.2 4.5 3.3 Other regions 30.5 40.2 15.7 5.3 4.0 4.2 Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Surveys. Note: The large shifts in educational attainment between the two survey periods suggests that the variable may have been coded differently in 2007/08 compared to 2012/13. The 2012/13 education results are very similar to the 2010 Labor Force Survey results. 38 Highest level completed. 52 Table 1.3: Descriptive statistics 2007/08 2010 2012/13 Total Male Female Total Male Female Total Male Female Weekly wage 790,969 865,763 648,830 898,053 957,683 783,632 1,522,206 1,643,325 1,319,593 Weekly hrs. worked 36.977 36.939 37.017 44.387 44.810 43.965 33.941 34.395 33.485 Part-time 0.473 0.470 0.476 0.261 0.246 0.275 0.524 0.510 0.539 Years of schooling 5.432 5.603 5.239 5.183 5.936 4.477 4.744 5.445 4.104 Less than primary 0.392 0.372 0.414 0.213 0.184 0.244 0.249 0.210 0.291 Primary 0.313 0.307 0.320 0.379 0.365 0.395 0.374 0.370 0.379 Lower secondary 0.149 0.158 0.140 0.203 0.214 0.190 0.171 0.190 0.150 Upper secondary 0.083 0.090 0.075 0.084 0.094 0.073 0.095 0.103 0.086 Vocational 0.042 0.049 0.035 0.067 0.076 0.057 0.057 0.065 0.048 Tertiary 0.021 0.024 0.017 0.054 0.067 0.041 0.054 0.063 0.045 Agriculture 0.710 0.700 0.719 0.714 0.693 0.735 0.606 0.606 0.605 Mining 0.002 0.003 0.001 0.005 0.008 0.003 0.003 0.005 0.002 Garment 0.021 0.004 0.038 0.014 0.004 0.025 0.017 0.004 0.031 Other manuf. 0.053 0.063 0.042 0.036 0.039 0.034 0.045 0.052 0.037 Services 0.215 0.230 0.200 0.230 0.257 0.203 0.246 0.257 0.234 Public - - - 0.126 0.146 0.099 0.138 0.137 0.139 SOE - - - 0.012 0.016 0.008 0.018 0.019 0.015 Civil servant - - - 0.114 0.130 0.091 0.120 0.118 0.123 Urban 0.288 0.286 0.291 0.321 0.316 0.326 0.288 0.283 0.294 Vientiane Capital 0.115 0.112 0.118 0.134 0.131 0.137 0.120 0.118 0.123 Vientiane 0.075 0.076 0.075 0.095 0.100 0.090 0.082 0.083 0.081 Savannakhet 0.140 0.142 0.138 0.141 0.139 0.143 0.143 0.145 0.141 Champasak 0.107 0.106 0.107 0.104 0.103 0.105 0.110 0.110 0.110 Khammuane 0.060 0.059 0.060 0.061 0.060 0.062 0.050 0.050 0.050 Luangprabang 0.073 0.070 0.075 0.067 0.066 0.069 0.071 0.071 0.071 Xayabury 0.060 0.061 0.058 0.061 0.062 0.059 0.059 0.058 0.059 Sources: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey and 2010 Lao PDR Labor Force and Child Labor Survey. 53 Annex 2 Table 2.1: Real weekly wage by education level, 2007/08 Less than Primary Lower- Upper Vocational Tertiary Total primary secondary secondary Male 414,020 645,940 542,668 533,889 540,903 649,514 562,053 Female 335,861 330,879 625,515 460,343 355,335 528,877 425,584 Agriculture 274,292 406,332 432,028 287,046 385,298 447,793 370,834 Mining 693,989 793,024 1,101,136 510,085 178,914 678,620 Garment 233,344 410,940 758,517 373,430 480,680 398,380 521,657 Other manufacturing 369,504 331,307 463,054 703,902 768,412 556,702 485,959 Services 549,875 691,862 639,460 548,483 468,153 628,528 579,557 Urban 547,292 615,754 633,502 568,699 499,917 646,194 583,847 Rural 257,424 436,285 437,579 344,027 385,918 378,658 381,081 Vientiane Capital 568,316 417,988 579,819 456,504 494,032 725,935 535,290 Vientiane 252,606 385,259 460,777 316,778 475,211 371,132 399,005 Savannakhet 403,585 947,365 673,812 754,130 513,848 436,246 645,266 Champasak 506,690 660,699 518,946 308,921 438,773 445,513 495,819 Khammuane 274,149 950,251 418,984 770,521 469,574 623,835 Luangprabang 272,240 242,314 358,982 372,446 410,153 224,610 325,407 Xayabury 124,042 241,203 633,033 343,238 394,444 663,847 347,951 Other regions 317,186 389,172 536,049 467,436 415,095 455,877 420,152 Total 385,554 531,540 569,004 512,503 470,916 609,389 515,403 Notes: Nominal wages deflated to 2010 LCU using the CPI from the World Bank World Development Indicators. Excludes income from agricultural production or household businesses. Source: 2007/08 Lao PDR Earnings and Consumption Survey. Table 2.2: Real weekly wage by education level, 2010 Less than Primary Lower- Upper Vocational Tertiary Total primary secondary secondary Male 426,844 381,088 549,024 626,822 494,336 721,431 551,565 Female 284,099 317,799 421,082 394,274 446,832 687,581 454,421 Agriculture 107,546 188,546 171,547 200,897 898,053 836,038 210,394 Mining 523,418 594,915 1,086,009 775,270 937,911 1,757,071 874,342 Garment 446,286 388,219 412,158 475,179 644,946 1,427,154 461,540 Other manufacturing 471,842 479,987 552,364 882,691 848,527 1,003,675 635,774 Services 456,473 373,092 505,539 510,850 456,981 681,358 520,974 Public 503,374 360,907 415,813 516,242 440,279 575,431 487,637 SOE 1,087,535 563,255 714,339 909,010 1,249,277 1,110,248 965,740 Civil servant 383,621 311,194 376,742 453,599 389,388 520,269 434,372 Private 347,697 358,195 563,880 598,825 656,767 1,171,720 555,679 Urban 479,419 377,565 482,941 548,879 497,902 742,301 555,118 Rural 291,623 340,699 565,579 552,144 443,758 559,229 456,548 Vientiane Capital 453,925 413,474 486,717 642,019 625,729 897,991 634,271 Vientiane 448,458 746,453 592,119 779,795 424,867 564,004 579,206 Savannakhet 238,437 243,459 434,695 336,904 458,550 565,499 400,565 Champasak 290,908 296,396 417,263 587,957 466,217 721,013 474,097 Khammuane 520,895 351,095 342,930 362,918 406,038 408,881 392,958 Luangprabang 306,136 280,192 533,539 392,287 527,471 768,053 478,684 Xayabury 61,779 178,955 611,272 452,448 361,107 308,793 336,223 Other regions 539,744 402,852 670,914 540,043 397,748 494,434 494,087 Total 372,597 358,749 514,199 550,047 476,762 709,930 518,872 Notes: Nominal wages deflated to 2010 LCU using the CPI from the World Bank World Development Indicators. Excludes income from agricultural production or household businesses. Source: 2010 Lao PDR Labor Force and Child Labor Survey. 54 Table 2.3: Real weekly wage by education level, 2012/13 Less than Primary Lower- Upper Vocational Tertiary Total primary secondary secondary Male 831,298 724,972 791,698 756,120 786,028 868,001 795,568 Feale 479,045 412,682 493,412 588,729 597,342 685,505 581,596 Agriculture 902,954 418,242 760,491 696,137 481,240 796,938 630,474 Mining 1,195,624 853,035 911,535 1,670,133 1,634,772 4,654,609 1,506,455 Garment 239,496 584,053 479,853 815,016 383,587 609,746 Other manufacturing 565,526 626,975 523,247 949,496 912,743 1,813,106 742,013 Services 628,613 703,903 738,642 651,463 704,293 734,866 706,490 Public 659,389 642,771 627,613 616,680 685,527 677,210 663,772 SOE 1,046,569 810,613 849,611 1,108,327 1,039,545 1,516,239 1,084,583 Civil servant 553,978 482,673 559,827 518,953 650,571 617,185 603,231 Private 736,561 623,567 734,795 766,385 782,416 1,126,539 776,308 Urban 953,872 659,570 715,019 714,185 708,501 797,230 744,137 Rural 517,012 600,177 675,737 659,773 703,200 765,834 661,017 Vientiane Capital 546,364 765,543 625,762 783,401 798,213 873,822 785,534 Vientiane 394,148 517,621 985,980 694,527 894,728 1,084,411 867,604 Savannakhet 2,123,647 448,688 664,779 659,291 606,477 743,476 711,487 Champasak 705,470 915,217 664,912 618,457 750,756 730,229 731,324 Khammuane 529,642 636,322 555,845 591,256 603,359 655,516 597,862 Luangprabang 632,913 538,346 815,695 687,216 575,049 950,439 701,149 Xayabury 482,870 563,231 833,614 514,052 475,228 377,382 552,334 Other regions 535,771 486,967 747,257 567,939 666,097 652,312 627,110 Total 708,991 627,450 699,339 699,499 706,832 791,297 716,638 Notes: Nominal wages deflated to 2010 LCU using the CPI from the World Bank World Development Indicators. Excludes income from agricultural production or household businesses. Source: 2012/13 Lao PDR Earnings and Consumption Survey. 55 Annex 3: Estimating the Returns to Education Methodology Returns to investment in schooling in Lao PDR are estimated using the human capital earnings function method separately across different years (2007/08, 2010 and 2012/13). Two specifications are considered to measure first the returns for years of schooling and second the returns for levels of schooling. The earnings function with years of schooling is: 2 ln = + + 1 + 2 + + where ln is the natural log of earnings of individual , is the number of years of schooling, and 2 are years of experience and its square, and is a dummy variable equal to 1 for females (added as a compensatory factor). The coefficient of interest, , measures the average private rate of return to one additional year of schooling, regardless of the level of schooling. The earnings function method can also be used to estimate average private returns to different levels of schooling by converting the continuous years of schooling variable, , into a series of dummy variables representing different levels of schooling achieved. The earnings function with levels of schooling is: 2 ln = + 1 + 2 + 3 + 4 + 1 + 2 + + where , , , and are dummy variables indicating primary, lower secondary, upper secondary or vocational, and tertiary education (university or middle/high training) as the highest level achieved (and in another specification as level completed). Different stratifications are used by including controls to consider the returns to education by sector, enterprise type, region, and province. Dummy variables are defined by sector (agriculture, mining, garments, other manufacturing, and services), enterprise type (SOE and non-SOE), region (urban and rural), and province (Vientiane capital, Vientiane, Savannakhet, Champasak, Khammuane, Luangprabang, Xayabury, and other). Data and variable definitions The Impact of education on earnings is estimated for 2007/08 and 2012/13 using the Lao PDR Earnings and Consumption Survey (LECS waves 4 and 5) and for 2010 using the Lao PDR Labor Force and Child Labor Survey. Data definitions for the 2010 Lao PDR Labor Force and Child Labor Survey are as follows: Weekly wages are defined as the amount of money in KIP earned for the last seven days. Weekly hours worked are total hours worked in the last seven days, summed across individuals’ primary job and other jobs. Part-time work is defined as less than 40 hours a week (based on total weekly hours worked). Years of experience is calculated as age minus years of schooling minus 6. Less than primary is defined as 56 individuals who have not completed primary school. Years of schooling is defined as highest grade achieved.39 Level of schooling is a dummy variable of whether a student has completed that level, based on highest level achieved (less than primary, primary, lower secondary, upper secondary, vocational, and tertiary). Industry of employment relates to the main occupation in the past seven days, defined at the 2- digit ISIC level. Agriculture covers 01-03, mining 05-09, garments 13-14, manufacturing 10-33, and services are the remaining activities. Individuals with economic activity or employment status in the past seven days as a civil servant, employee of public and private work, employee of private business, employee of state enterprise, employer or self-employed are included in the sample. SOE or civil servant are those whose economic activity or employment status in the past seven days was an employee of a state enterprise or civil servant respectively, and public are those whose economic activity or employment status in the past seven days was SOE or civil servant. Urban are those whose village type is urban, while rural are those whose village type is rural with or without a road. Seven out of 17 provinces were selected based on having more than 5 percent of the sample population. Data definitions for the 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey (LECS) are as follows: Weekly wages are defined as wages in cash from the previous week (excluding income from agricultural production or household business). Because the questionnaire does not distinguish wages by activity, we only include individuals with one job. Years of experience is calculated as age minus years of schooling minus 6. Years of schooling is defined as highest grade achieved.40 Level of schooling is a dummy variable of whether a student has completed that level (less than primary, primary, lower secondary, upper secondary, vocational, and tertiary). Note that SOE, civil servant and public sector categories are not available in the 2007/08 LECS. Regression results Note that wages include only four categories of earners: civil servants, employee of public and private work; employee of private business, employee of state enterprise. As such, they exclude self-employed, unpaid family workers, and employees not paid a wage. They also exclude business and agricultural income. They also exclude individuals who work more than one job, because we cannot split wages per activity. 39 According to the coding of the question, there are five grades in primary school, four grades in lower secondary, three grades of upper secondary, two grades of low vocational school, three years of middle technical school, and nine years of college / high technical school. Because lower secondary school is Lao PDR is only three years, the fourth year was treated as the same as the first year of upper secondary / low vocational school. Individuals reporting non-standard curriculum are excluded from the analysis. 40 According to the coding of the question, there are five grades in primary school, three grades in lower secondary, three grades of upper secondary, three grades of vocational school, and five years of university / institute. 57 Table 3.1: Lao PDR earnings functions with year of schooling, 2007/08 No controls Sector Urban Province All controls Years of schooling 0.083*** 0.044*** 0.046*** 0.065*** 0.032*** (0.007) (0.007) (0.007) (0.008) (0.008) Years of experience 0.024*** 0.029*** 0.025*** 0.026*** 0.028*** (0.006) (0.006) (0.006) (0.006) (0.006) Experience-squared -0.000*** -0.001*** -0.001*** -0.000*** -0.001*** (0.000) (0.000) (0.000) (0.000) (0.000) Female -0.077 -0.146*** -0.184*** -0.114** -0.191*** (0.058) (0.055) (0.056) (0.058) (0.055) Agriculture -1.594*** -1.282*** (0.216) (0.217) Other manufacturing -0.826*** -0.686*** (0.226) (0.227) Garment -0.519** -0.404 (0.247) (0.249) Services -0.691*** -0.553** (0.216) (0.216) Urban 0.715*** 0.375*** (0.055) (0.061) Vientiane capital 0.327*** -0.062 (0.072) (0.073) Savannakhet 0.159* 0.059 (0.090) (0.084) Champasak 0.424*** 0.219* (0.119) (0.112) Luangprabang -0.318** -0.421*** (0.147) (0.139) Khammuane 0.212* 0.053 (0.119) (0.113) Vientiane 0.189 -0.053 (0.119) (0.113) Xayabury -0.446*** -0.346*** (0.105) (0.099) Constant 12.054*** 13.354*** 12.014*** 12.095*** 13.133*** (0.104) (0.222) (0.100) (0.106) (0.222) Observations 1,860 1,860 1,860 1,860 1,860 R-squared 0.08 0.20 0.16 0.12 0.24 Source: 2007/08 Lao PDR Earnings and Consumption Survey. Notes: Standard errors in parentheses. *10 percent; **5 percent; ***1 percent significance level. 58 Table 3.2: Lao PDR earnings functions with year of schooling, 2010 No controls Sector Public SOE/Civil Urban Province All controls servant Years of schooling 0.086*** 0.071*** 0.091*** 0.094*** 0.080*** 0.081*** 0.068*** (0.003) (0.004) (0.004) (0.004) (0.004) (0.003) (0.004) Years of experience 0.007** 0.010*** 0.008** 0.010*** 0.008** 0.010*** 0.011*** (0.004) (0.003) (0.004) (0.004) (0.004) (0.004) (0.003) Experience-squared -0.000 -0.000 -0.000 -0.000 -0.000 -0.000 -0.000 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Female -0.133*** -0.100*** -0.131*** -0.114*** -0.143*** -0.138*** -0.100*** (0.034) (0.032) (0.034) (0.033) (0.034) (0.033) (0.032) Agriculture -1.959*** -1.802*** (0.108) (0.107) Other manufacturing -0.508*** -0.493*** (0.105) (0.104) Garment -0.367*** -0.521*** (0.124) (0.123) Services -0.788*** -0.780*** (0.095) (0.095) Public -0.077** -0.005 (0.038) (0.039) SOE 0.611*** (0.073) Civil servant -0.173*** (0.038) Urban 0.227*** 0.049 (0.034) (0.033) Vientiane capital 0.235*** 0.139*** (0.043) (0.044) Savannakhet -0.420*** -0.388*** (0.055) (0.053) Champasak -0.131** -0.073 (0.061) (0.058) Luangprabang -0.093 -0.144** (0.067) (0.064) Khammuane -0.057 -0.170** (0.075) (0.072) Vientiane 0.247*** 0.138** (0.072) (0.070) Xayabury -0.651*** -0.488*** (0.072) (0.070) Constant 12.229*** 13.170*** 12.207*** 12.150*** 12.163*** 12.279*** 13.199*** (0.059) (0.104) (0.060) (0.060) (0.060) (0.065) (0.107) Observations 4,345 4,337 4,345 4,345 4,345 4,345 4,337 R-squared 0.14 0.24 0.14 0.16 0.15 0.19 0.27 Source: 2010 Lao PDR Labor Force and Child Labor Survey. Notes: Standard errors in parentheses. *10 percent; **5 percent; ***1 percent significance level. 59 Table 3.3: Lao PDR earnings functions with year of schooling, 2012/13 No controls Sector Public SOE/Civil Urban Province All controls servant Years of schooling 0.035*** 0.035*** 0.052*** 0.059*** 0.033*** 0.032*** 0.045*** (0.004) (0.004) (0.005) (0.005) (0.004) (0.004) (0.005) Years of experience 0.028*** 0.029*** 0.030*** 0.030*** 0.027*** 0.027*** 0.030*** (0.005) (0.004) (0.004) (0.004) (0.005) (0.004) (0.004) Experience-squared -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Female -0.196*** -0.178*** -0.187*** -0.182*** -0.202*** -0.224*** -0.182*** (0.037) (0.037) (0.036) (0.036) (0.037) (0.037) (0.037) Agriculture -0.894*** -0.924*** (0.125) (0.126) Other manufacturing -0.544*** -0.592*** (0.124) (0.124) Garment -0.620*** -0.740*** (0.161) (0.163) Services -0.678*** -0.656*** (0.115) (0.118) Public -0.285*** -0.235*** (0.039) (0.043) SOE 0.231*** (0.072) Civil servant -0.390*** (0.040) Urban 0.043 -0.041 (0.037) (0.037) Vientiane capital 0.255*** 0.146*** (0.044) (0.047) Savannakhet 0.123* 0.041 (0.071) (0.070) Champasak 0.225*** 0.171*** (0.066) (0.065) Luangprabang 0.168** 0.089 (0.082) (0.082) Khammuane 0.085 -0.043 (0.070) (0.070) Vientiane 0.271*** 0.130* (0.075) (0.076) Xayabury -0.133* -0.230*** (0.069) (0.068) Constant 13.301*** 13.948*** 13.220*** 13.163*** 13.298*** 13.235*** 13.906*** (0.071) (0.128) (0.071) (0.070) (0.071) (0.075) (0.134) Observations 1,912 1,912 1,912 1,912 1,912 1,912 1,912 R-squared 0.07 0.10 0.10 0.13 0.07 0.10 0.14 Source: 2012/13 Lao PDR Earnings and Consumption Survey. Notes: Standard errors in parentheses. *10 percent; **5 percent; ***1 percent significance level. 60 Table 3.4: Lao PDR earnings functions with levels of schooling completed, 2007/08 No controls Sector Urban Province All controls Primary 0.174** 0.066 0.074 0.182** 0.056 (0.083) (0.077) (0.080) (0.082) (0.077) Lower secondary 0.566*** 0.341*** 0.326*** 0.487*** 0.245*** (0.091) (0.086) (0.089) (0.090) (0.086) Upper secondary 0.727*** 0.347*** 0.388*** 0.593*** 0.245*** (0.094) (0.091) (0.094) (0.095) (0.092) Vocational 0.725*** 0.249 0.342* 0.655*** 0.185 (0.183) (0.174) (0.178) (0.182) (0.173) Tertiary 1.041*** 0.576*** 0.590*** 0.838*** 0.440*** (0.105) (0.105) (0.107) (0.111) (0.109) Years of experience 0.022*** 0.027*** 0.024*** 0.024*** 0.027*** (0.006) (0.006) (0.006) (0.006) (0.006) Experience-squared -0.000*** -0.000*** -0.000*** -0.000*** -0.001*** (0.000) (0.000) (0.000) (0.000) (0.000) Female -0.076 -0.142** -0.181*** -0.114** -0.189*** (0.058) (0.055) (0.056) (0.058) (0.055) Agriculture -1.631*** -1.310*** (0.216) (0.218) Other manufacturing -0.874*** -0.716*** (0.227) (0.228) Garment -0.587** -0.449* (0.249) (0.251) Services -0.727*** -0.577*** (0.217) (0.217) Urban 0.709*** 0.369*** (0.055) (0.061) Vientiane capital 0.331*** -0.059 (0.072) (0.074) Savannakhet 0.152* 0.053 (0.090) (0.085) Champasak 0.414*** 0.210* (0.119) (0.113) Luangprabang -0.314** -0.412*** (0.147) (0.139) Khammuane 0.197* 0.042 (0.119) (0.113) Vientiane 0.173 -0.064 (0.119) (0.113) Xayabury -0.441*** -0.341*** (0.105) (0.099) Constant 12.297*** 13.533*** 12.161*** 12.253*** 13.263*** (0.103) (0.225) (0.099) (0.106) (0.226) Observations 1,860 1,860 1,860 1,860 1,860 R-squared 0.09 0.21 0.16 0.12 0.24 Source: 2007/08 Lao PDR Earnings and Consumption Survey. Notes: Standard errors in parentheses. *10 percent; **5 percent; ***1 percent significance level. 61 Table 3.5: Lao PDR earnings functions with levels of schooling completed, 2010 No controls Sector Public SOE/Civil Urban Province All controls servant Primary 0.283*** 0.236*** 0.293*** 0.288*** 0.265*** 0.241*** 0.199*** (0.077) (0.074) (0.077) (0.076) (0.077) (0.076) (0.073) Lower secondary 0.814*** 0.581*** 0.841*** 0.851*** 0.764*** 0.720*** 0.524*** (0.077) (0.075) (0.078) (0.077) (0.078) (0.076) (0.075) Upper secondary 1.001*** 0.764*** 1.060*** 1.068*** 0.948*** 0.901*** 0.698*** (0.080) (0.079) (0.083) (0.082) (0.081) (0.079) (0.080) Vocational 1.034*** 0.849*** 1.115*** 1.167*** 0.991*** 0.996*** 0.827*** (0.076) (0.076) (0.081) (0.080) (0.076) (0.075) (0.079) Tertiary 1.346*** 1.162*** 1.426*** 1.455*** 1.259*** 1.226*** 1.071*** (0.078) (0.077) (0.082) (0.081) (0.079) (0.076) (0.081) Years of experience 0.014*** 0.016*** 0.015*** 0.016*** 0.014*** 0.014*** 0.016*** (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) Experience-squared -0.000** -0.000** -0.000* -0.000** -0.000** -0.000* -0.000** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Female -0.082** -0.076** -0.081** -0.067** -0.095*** -0.098*** -0.082** (0.034) (0.033) (0.034) (0.033) (0.034) (0.033) (0.032) Agriculture -1.961*** -1.810*** (0.113) (0.113) Other manufacturing -0.468*** -0.450*** (0.107) (0.106) Garment -0.377*** -0.510*** (0.125) (0.125) Services -0.798*** -0.776*** (0.097) (0.097) Public -0.109*** -0.015 (0.038) (0.039) SOE 0.562*** (0.072) Civil servant -0.204*** (0.038) Urban 0.194*** 0.051 (0.034) (0.033) Vientiane capital 0.264*** 0.178*** (0.043) (0.044) Savannakhet -0.286*** -0.281*** (0.056) (0.055) Champasak -0.066 0.002 (0.061) (0.058) Luangprabang -0.018 -0.061 (0.067) (0.064) Khammuane -0.109 -0.196*** (0.077) (0.074) Vientiane 0.309*** 0.198*** (0.072) (0.070) Xayabury -0.570*** -0.401*** (0.072) (0.070) Constant 12.254*** 13.211*** 12.247*** 12.216*** 12.191*** 12.301*** 13.214*** (0.082) (0.117) (0.082) (0.081) (0.082) (0.086) (0.121) Observations 4,178 4,170 4,178 4,178 4,178 4,178 4,170 R-squared 0.13 0.22 0.13 0.16 0.14 0.18 0.25 Source: 2010 Lao PDR Labor Force and Child Labor Survey. Notes: Standard errors in parentheses. *10 percent; **5 percent; ***1 percent significance level. 62 Table 3.6: Lao PDR earnings functions with levels of schooling completed, 2012/13 No controls Sector Public SOE/Civil Urban Province All controls servant Primary 0.011 -0.001 -0.004 -0.022 0.009 0.037 0.020 (0.078) (0.078) (0.077) (0.076) (0.078) (0.078) (0.076) Lower secondary 0.174** 0.156** 0.216*** 0.222*** 0.166** 0.169** 0.190** (0.078) (0.078) (0.077) (0.076) (0.079) (0.078) (0.077) Upper secondary 0.216*** 0.212*** 0.326*** 0.353*** 0.203** 0.191** 0.280*** (0.080) (0.081) (0.080) (0.079) (0.081) (0.080) (0.082) Vocational 0.248** 0.268** 0.442*** 0.479*** 0.239** 0.256** 0.417*** (0.118) (0.118) (0.119) (0.117) (0.118) (0.117) (0.118) Tertiary 0.375*** 0.386*** 0.572*** 0.642*** 0.355*** 0.361*** 0.524*** (0.077) (0.080) (0.081) (0.079) (0.079) (0.078) (0.084) Years of experience 0.024*** 0.026*** 0.026*** 0.026*** 0.024*** 0.024*** 0.027*** (0.005) (0.005) (0.005) (0.004) (0.005) (0.005) (0.005) Experience-squared -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Female -0.202*** -0.188*** -0.198*** -0.196*** -0.207*** -0.231*** -0.196*** (0.037) (0.037) (0.037) (0.036) (0.038) (0.037) (0.037) Agriculture -0.829*** -0.860*** (0.125) (0.127) Other manufacturing -0.559*** -0.606*** (0.124) (0.124) Garment -0.602*** -0.717*** (0.162) (0.163) Services -0.671*** -0.650*** (0.115) (0.117) Public -0.297*** -0.246*** (0.040) (0.043) SOE 0.212*** (0.072) Civil servant -0.403*** (0.041) Urban 0.041 -0.037 (0.037) (0.038) Vientiane capital 0.252*** 0.145*** (0.044) (0.047) Savannakhet 0.117 0.038 (0.072) (0.071) Champasak 0.217*** 0.158** (0.067) (0.067) Luangprabang 0.162* 0.090 (0.083) (0.082) Khammuane 0.084 -0.029 (0.071) (0.071) Vientiane 0.274*** 0.141* (0.076) (0.076) Xayabury -0.148** -0.245*** (0.069) (0.069) Constant 13.491*** 14.131*** 13.504*** 13.490*** 13.483*** 13.396*** 14.125*** (0.086) (0.137) (0.084) (0.083) (0.086) (0.090) (0.141) Observations 1,852 1,852 1,852 1,852 1,852 1,852 1,852 R-squared 0.06 0.08 0.09 0.12 0.06 0.09 0.13 Source: 2012/13 Lao PDR Earnings and Consumption Survey. Notes: Standard errors in parentheses. *10 percent; **5 percent; ***1 percent significance level. 63 Annex 4 Table 4.1: Size of exporting vs. non-exporting manufacturing firms (average number of full-time employees) Year Exporting Lao PDR Full sample status Employees Skilled Unskilled Employees Skilled Unskilled employees employees employees employees non-exporter 20 11 8 n.a. n.a. n.a. 2006 exporter 112 49 62 n.a. n.a. n.a. non-exporter 52 - - n.a. n.a. n.a. 2009 exporter 262 - - n.a. n.a. n.a. non-exporter 46 20 27 84 32 55 2012 exporter 202 109 93 249 125 126 Source: World Bank Staff calculations based on data from World Bank enterprise surveys. Full sample of countries includes China, Belarus, Nepal, West Bank and Gaza, Georgia, Azerbaijan, Kyrgyz Republic, Kazakhstan, Bangladesh, Armenia, Uganda, Ukraine, Uzbekistan, Mongolia, Moldova, Tanzania, Kenya, Bulgaria, Serbia, Zambia, Bosnia and Herzegovina, Albania, Kosovo, Croatia, FYR of Macedonia, Slovenia, Montenegro, Djibouti, Jordan, Romania, Latvia, Democratic Republic of the Congo, Tajikistan, Estonia, Hungary, Lithuania, Poland, Slovak Republic, Turkey, Ghana and Madagascar. Data for China and Lao PDR are for 2012, and data for the rest of the sample are for 2013. 64 Table 4.2: Size of exporting vs. non-exporting manufacturing firms: Comparing Lao PDR, Vietnam and Cambodia (number of full-time employees) Year Export Lao PDR Vietnam Cambodia ing status Employe Skilled Unskille Employe Skilled Unskille Employe Skilled Unskille es employe d es employe d es employe d es employe es employe es employe es es es non- export er 137 87 49 export 2005 er 584 416 164 non- export er 20 11 8 export 2006 er 112 49 62 non- export er 90 48 42 export 2007 er 1027 841 186 non- export er 52 102 50 45 export 2009 er 262 535 306 232 non- export er 46 20 27 export 2012 er 202 109 93 Source: World Bank Staff calculations based on data from World Bank enterprise surveys. 65 Table 4.3: Export intensity of manufacturing firms: Lao PDR and the rest of the World Lao PDR (2012) Full sample (2013) Industry Percent of Percent of exports in Percent of Percent of exports in exporting firms total sales exporting firms total sales Total 42.7 87.5 35.0 49.2 Food 33.3 72.5 28.7 46.5 Wood products & furniture 44.4 66.3 29.1 45.9 Non-metallic and plastic materials 31.4 40.3 Metals and machinery 35.0 42.6 Leather 16.1 34.1 Textiles 50.0 77.5 48.1 54.3 Garments 76.9 94.5 47.8 70.9 Electronics 100.0 100.0 39.4 44.1 Chemicals and pharmaceuticals 100.0 100.0 39.0 37.6 Auto and auto components 32.9 48.6 Other manufacturing 29.5 47.0 Source: World Bank Staff calculations based on data from World Bank enterprise surveys for 2012 and 2013. Full sample of countries includes China, Belarus, Nepal, West Bank and Gaza, Georgia, Azerbaijan, Kyrgyz Republic, Kazakhstan, Bangladesh, Armenia, Uganda, Ukraine, Uzbekistan, Mongolia, Moldova, Tanzania, Kenya, Bulgaria, Serbia, Zambia, Bosnia and Herzegovina, Albania, Kosovo, Croatia, FYR of Macedonia, Slovenia, Montenegro, Djibouti, Jordan, Romania, Latvia, Democratic Republic of the Congo, Tajikistan, Estonia, Hungary, Lithuania, Poland, Slovak Republic, Turkey, Ghana and Madagascar. Data for China and Lao PDR are for 2012. 66 Table 4.4: Lao PDR’s export intensity of manufacturing firms compared to Vietnam and Cambodia Lao PDR (2012) Vietnam (2009) Cambodia (2007) Industry Percent of Percent of Percent of Percent of Percent of Percent of exporting exports in exporting exports in exporting exports in firms total sales firms total sales firms total sales Total 42.7 87.5 47.2 63.0 34.1 97.2 Food 33.3 72.5 46.3 63.6 18.2 75.0 Non-metallic and plastic materials 39.7 52.5 33.3 55.0 Metals and machinery 31.3 47.2 Textiles 50.0 77.5 48.5 64.3 100.0 95.0 Garments 76.9 94.5 74.6 88.2 97.8 99.9 Electronics 100.0 100.0 63.6 29.4 Chemicals and pharmaceuticals 100.0 100.0 44.0 15.4 Other manufacturing 40.0 66.3 46.5 64.0 25.0 70.0 Source: World Bank Staff calculations based on data from World Bank enterprise surveys for 2012 for Lao PDR, 2009 for Vietnam and 2007 for Cambodia. 67 Annex 5: Testing the Link between Exporting and Labor-Market Outcomes in Lao PDR and a Global Sample of Countries Methodology To identify the effect of firms’ sales structure on labor outcomes, we estimate the effect of within -firm changes in the share of sales that are exported on the level of employment (total, skilled, unskilled, female) and average wages (total, female) of the firm, the latter being a proxy for the quality of workers employed in each firm. The analysis follows a methodology similar to Bambrilla, Lederman and Porto (2012) and Cebeci, Lederman and Rojas (2013). The relationship between exporting and labor-market outcomes is specified as: ,, = + + + 1 ,, + 2 ,, + ,, where ,, stands for the log of firm ’s employment level or average wage in industry in year; ,, is the share of firm-level exports in total sales; ,,, is the log of total firm-level sales; are firm-, industry-, and year-specific effects to control for other factors that can affect firm size and sales; 1 is a first parameter to be estimated, which is an export-sales elasticity of employment and wages; and the error term ,, is assumed to be white noise. To control for firm-specific effects, the model is estimated in first-differences, which wipes out the influence of time-invariant firm-specific characteristics. Industry and year dummy variables are included to control for industry- and year-specific effects. Controlling for changes in total sales avoids potential biases in the estimation of 1 due to exogenous changes in export conditions. For example, a positive external shock that increases exports could possibly also increase domestic sales if firms invest part of the money from increased exports to stimulate domestic sales as well. Such an effect would over-estimate the quasi-elasticity of labor market outcomes vis-à-vis the share of exports in total sales. Alternatively, the positive external shock could decrease domestic sales if the supply response is inelastic and the firm decides to serve only the export market at the expense of the domestic market, in which case the bias would go in the opposite direction. In practice, the econometric model that is estimated is: ∆,, = + + 1 ∆,, + 2 ∆,, + ∆,, . OLS estimates of the elasticities 1 may be biased due to reverse causality, as changes in firm size and wages may affect the structure of firms’ sales (given that exporters are larger and more productive than non-exporters). Following Brambilla, Lederman and Porto (2012) and Cebeci, Lederman and Rojas (2013), our methodology will use exogenous fluctuations in exchange rates interacted with firms’ initial exposure to external markets as instrumental variables to identify causal effects of export growth within a firm (as a share of total sales) on a firm’s labor market outcomes. The first-stage regression specification is: ∆,, = + + 1 ,,=0 ∗ ∆ + 2 ∆,, + ,, where ,,=0 ∗ ∆ is the initial share of exports in total sales of each firm41 times the real effective (trade weighted) exchange rate of the country. A valid instrument is one that is correlated with the 41 The first year a firm appears in the sample is when it first exported, such that the initial share of exports is not zero for any firm. 68 endogenous regressor and uncorrelated with the error term. The choice of instruments is valid given that (i) a change in a countries’ effective exchange rate affects export competitiveness of a firm, (ii) the initial level of firms’ exposure to exports is pre-determined, and (iii) the assumption that a country’s trade- weighted exchange rate is not influenced by any single firm. Because the specification in levels has firm- level fixed effects, which control for time-invariant firm-specific factors, first differencing the model is not necessary to remove the effect of such factors. However, estimating the specification in changes arguably allows for a stronger instrument, given that changes in the effective exchange rate as opposed to the level are a better exogenous correlate to export competitiveness. Moreover, while the effective exchange rate is common to all firms within a country at a point in time, the interaction with a firms’ initial export or destination-market share allows the instrument to vary over time and across firms (so it can explain variation in changes in export shares over time and across firms).42 We test the model using firm-level survey data. After the preceding relationship has been estimated for Lao PDR, we build a panel dataset of firms across a large number of countries worldwide. This dataset enables us to estimate the average elasticity in the world as a baseline for comparison. The first and second stages of the estimation therefore become: ∆,,, = + + + 1 ,,,=0 ∗ ∆, + 2 ∆,,, + ,,, ∆,,, = + + + 1 ∆,,, + 2 ∆,,, + ∆,,, where is the fixed effect for country . Finally, the analysis considers whether a country’s level of competitiveness affects the magnitude of the elasticity; that is, whether the level of competitiveness affects the employment and wage responses to within-firm changes in sales structure. We use a panel dataset of firms across countries worldwide where variation in competitiveness indicators can be observed. The impact of exports on employment is allowed to depend on the level of the competitiveness indicator in each country by also interacting the firm’s change in export shares with these indicators in the model’s specification. The methodology is based on that of Calderón and Poggio (2011) and the second-stage estimation model is: ∆,,, = + + + 1 ∆,,, + 2 ∆,,, ∗ , + 3 ∆,,, + ∆,,, where , represents a competiveness indicator for country in year . The competitiveness indicators that we test relate to institutions, macroeconomy, higher education and training, market efficiency, business sophistication, and an index combining all of these. As there are now two endogenous variables, we use two instruments so the coefficients are exactly identified, where the second instrument is equal to the first instrument interacted with the competitiveness indicator. The first-stage estimation model is: ∆,,, = + + + 1 ,,,=0 ∗ ∆, + 2 ,,,=0 ∗ , ∗ ∆, + 3 ∆,,, + ,,, ∆,,, ∗ , = + + + 1 ,,,=0 ∗ ∆, + 2 ,,,=0 ∗ , ∗ ∆, + 3 ∆,,, + ,,, . 42 Changes in firms’ sales may also be endogenous to the specification, for example if firm size is correlated with unobservable firm characteristics that change over time. Notwithstanding, because we are less interested in estimating this coefficient, and because not including it potentially biases the coefficient we are interested in estimating, we maintain the specification. 69 Variable definitions and data sources We utilize firm-level data to estimate the effects of within-firm changes in sales structure on firm-level employment and average wages. The datasets include firm-level employment (total, skilled, unskilled, and female workers), firm-level average wage (total), the share of exports in firm-level total sales, and total sales. For the country-specific analysis for Lao PDR as well as the cross-country analysis we use World Bank Business Environment and Enterprise Performance Surveys for 60 countries43 conducted intermittently during the years 2002-2013. Data on effective exchange rates is collected from Darvas (2012). Data on indicators of competiveness is collected from the World Economic Forum’s Global Competiveness Report. Export share: Percentage of establishment’s sales that are exported. Firm-level export shares are collected from World Bank enterprise surveys, which includes direct or indirect exports (sold domestically to a third party that exports products). Note that because a panel dimension is necessary to conduct the analysis of changes in export shares and changes in labor market outcomes, the selection of firms was limited to those observed in the data across multiple years. A firm was included in the sample if it was observed in any two consecutive sample years, had exported in one of those years, and had non- zero exports in the first year or a previous year. Real sales: Firms’ total turnover. Sales are collected from the World Bank enterprise surveys, then adjusted to real 2005 international currency units using the consumer price index and the PPP conversion factor reported by the World Bank World Development Indicators. Employment: Number of persons employed. Firm-level employment is collected from World Bank enterprise surveys, and measures the number of permanent, full-time employees at the end of last fiscal year plus number of full-time seasonal/temporary workers employed at the end of last fiscal year. 44 Permanent, full-time employees are defined in the World Bank enterprise survey questionnaire as all paid employees that are contracted for a term of one or more fiscal years and/or have a guaranteed renewal of their employment contract and that work eight or more hours per day. Full-time, temporary workers are defined as all paid short-term (i.e., for less than a fiscal year) employees with no guarantee of renewal of their employment contract and work eight or more hours per day.45 Female employment: Number of female persons employed. Firm-level female employment is collected from World Bank enterprise surveys. 43 Sample of countries includes: Afghanistan, Albania, Angola, Argentina, Armenia, Azerbaijan, Bangladesh, Belarus, Bolivia, Bosnia and Herzegovina, Botswana, Bulgaria, Burkina Faso, Cape Verde, Cameroon, Chile, Colombia, Democratic Republic of the Congo, Croatia, Czech Republic, Ecuador, El Salvador, Estonia, Georgia, Guatemala, Honduras, Hungary, Kazakhstan, Kenya, Kyrgyz Republic, Lao PDR, Latvia, Lithuania, FYR of Macedonia, Malawi, Mali, Mexico, Moldova, Morocco, Nepal, Pakistan, Panama, Paraguay, Peru, Poland, Romania, Russian Federation, Rwanda, Senegal, Serbia, Slovak Republic, Tajikistan, Tanzania, Turkey, Uganda, Ukraine, Uruguay, Uzbekistan, Venezuela, and Zambia. 44 It is possible that our ability to only measure full-time workers (permanent or temporary) may be driving our results. 45 Whereas the World Bank enterprise survey questionnaire asks for the average length of employment of all full- time temporary employees in the fiscal year, data on this variable is missing for most firms. 70 Skilled employment: Number of full-time employees who were skilled production workers at the end of the fiscal year. Firm-level skilled employment is collected from World Bank enterprise surveys. Unskilled employment is total employment minus skilled employment.46 Real wages: Total annual cost of labor or remuneration (including wage, salary, premium, bonus, social payments, etc. of both permanent and temporary employees) which was accrued or paid in kind to employees (including income tax) during the year divided by employment. Firm-level average wages are calculated from the World Bank enterprise surveys, then adjusted to real 2005 international currency units using the consumer price index and the PPP conversion factor reported by the World Bank World Development Indicators. Real effective exchange rate: The nominal effective exchange rate (a measure of the value of a currency against a weighted average of several foreign currencies) divided by a price deflator or index of costs , where 2007=100. The real effective exchange rate was collected from the IMF’s International Financi al Statistics, country-specific sources and Darvas (2012). Indicators of competiveness: National indicators of competitiveness capture factors that affect productivity and competiveness, and are grouped into different pillars. The six indicators used in the analysis relate to institutions, macro economy, higher education and training, market efficiency, business sophistication, and overall competitiveness (which combines these five factors). Data is collected from the World Economic Forum’s Global Competiveness Report. 46 The World Bank enterprise survey questionnaire asks what percent of the establishment’s labor force employed at the end of the fiscal year had a university degree, but this data is not available in the panel surveys. 71 Regression results Table 5.1: Lao PDR’s elasticity of labor market outcomes with respect to exporting OLS IV Employment Female employment Wages Employment Female employment Wages Export share -0.0008 -0.0021 -0.0041 0.0125933 0.0076 -0.0365 (0.004) (0.004) (0.006) (0.015) (0.014) (0.030) Observations 38 33 35 38 33 35 Note: * Significant at the 15 % level, ** at the 10 % level, *** at the 5% level. Standard errors are in parentheses. Source: World Bank Staff calculations based on data from World Bank enterprise surveys. Table 5.2: Global elasticity of labor market outcomes with respect to exporting OLS IV Employm Skilled Unskilled Female Wages Employment Skilled Unskilled Female Wages ent employment employment employment employment employment employment Export share 0.0021*** 0.0014* 0.0022*** 0.0011 -0.0003 -0.0015 0.0503 -0.0100 -0.0167 -0.0263 (0.000) (0.001) (0.001) (0.002) (0.001) (0.025) (0.046) (0.021) (0.022) (0.035) Observations 2,299 1,661 1,516 241 1,628 2,299 1,661 1,516 241 1,628 Note: * Significant at the 20 % level, ** at the 10 % level, *** at the 5% level. Standard errors are in parentheses. Source: World Bank Staff calculations based on data from World Bank enterprise surveys. Table 5.3: Global elasticity of labor market outcomes with respect to exporting, Lao PDR-specific effect OLS IV Employment Female employment Wages Employment Female employment Wages Export share 0.0020*** 0.0013 -0.0002 -0.0014 -0.0076 -0.0273 (0.0005148) (0.0019254) (0.0009826) (0.0258677) (0.0195538) (0.0371537) Export 0.0013586 -0.000925 -0.0009287 -0.0007436 0.0336622 -0.0388899 share*LaoPDR (0.002) (0.004) (0.003) (0.027) (0.047) (0.061) Observations 2,299 241 1,628 2,299 241 1,628 Note: * Significant at the 15 % level, ** at the 10 % level, *** at the 5% level. Standard errors are in parentheses. Source: World Bank Staff calculations based on data from World Bank enterprise surveys. 72 Table 5.4: Impact of export share on employment and wages, conditional on competitive indicators, OLS estimates OLS Overall Institution Macroeconomi Higher Market Business competitivenes s c environment educatio efficienc sophisticatio s n and y n training Employmen Export share 0.0116** 0.0030 0.0187*** 0.0050** 0.0028 -0.0005 t (0.007) (0.004) (0.006) (0.003) (0.005) (0.007) Export share*Competivenes -0.0025* -0.0003 -0.0037*** -0.0009 -0.0002 0.0006 s indicator (0.002) (0.001) (0.001) (0.001) (0.001) (0.002) Skilled Export share -0.0110 -0.0028 0.0039 -0.0029 -0.0103 -0.0217** employmen (0.013) (0.009) (0.012) (0.005) (0.011) (0.013) t Export share*Competivenes 0.0031 0.0011 -0.0006 0.0012 0.0030 0.0057** s indicator (0.003) (0.002) (0.003) (0.001) (0.003) (0.003) Unskilled Export share 0.0102 0.0109* 0.0097 0.0079* 0.0021 0.0050 employmen (0.012) (0.008) (0.010) (0.005) (0.010) (0.012) t Export share*Competivenes -0.0021 -0.0025 -0.0017 -0.0017 -0.0001 -0.0008 s indicator (0.003) (0.002) (0.002) (0.001) (0.003) (0.003) Female Export share 0.0324 0.0016 0.0217 0.0223 0.0037 -0.0002 employmen (0.042) (0.029) (0.057) (0.018) (0.041) (0.049) t Export share*Competivenes -0.0079 -0.0002 -0.0047 -0.0063 -0.0007 0.0003 s indicator (0.011) (0.007) (0.013) (0.005) (0.011) (0.012) Wages Export share -0.0086 0.0030 -0.0096 -0.0031 -0.0053 -0.0001 (0.013) (0.008) (0.012) (0.005) (0.011) (0.015) Export share*Competivenes 0.0021 -0.0009 0.0020 0.0008 0.0013 -0.0001 s indicator (0.003) (0.002) (0.002) (0.001) (0.003) (0.004) Note: * Significant at the 15 % level, ** at the 10 % level, *** at the 5% level. Standard errors are in parentheses. Industry fixed effects are defined at the 2-digit level of the ISIC classification. Source: World Bank Staff calculations based on data from World Bank enterprise surveys and World Economic Forum’s Global Competiveness Report. 73 Table 5.5: Impact of export share on employment and wages, conditional on competitive indicators, IV estimates IV Overall Institution Macroeconomi Higher Market Business competitivenes s c environment educatio efficienc sophisticatio s n and y n training Employmen Export share 0.0333 -0.0040 0.2106 0.0334 0.0647 0.0532 t (0.084) (0.077) (0.469) (0.046) (0.122) (0.125) Export share*Competivenes -0.0049 0.0031 -0.0388 -0.0042 -0.0122 -0.0095 s indicator (0.021) (0.017) (0.096) (0.012) (0.032) (0.032) Skilled Export share -4.0524 0.1753** -0.0439 -0.2506* 1.1256 -3.4719 employmen (21.111) (0.093) (0.188) (0.194) (1.965) (12.078) t Export share*Competivenes 0.9373 -0.0388* 0.0185 0.0590* -0.2645 0.8203 s indicator (4.881) (0.026) (0.041) (0.045) (0.464) (2.848) Unskilled Export share 0.2408 0.0032 -0.1357 0.0584 0.8418 0.3708 employmen (0.349) (0.061) (0.113) (0.074) (3.014) (0.594) t Export share*Competivenes -0.0540 -0.0010 0.0291 -0.0118 -0.2006 -0.0859 s indicator (0.082) (0.016) (0.024) (0.017) (0.727) (0.141) Female Export share 0.0143 -0.6811 -0.3392* -0.0893 -0.0130 0.1355 employmen (0.194) (1.179) (0.214) (0.153) (0.196) (0.369) t Export share*Competivenes -0.0084 0.1668 0.0775* 0.0233 -0.0010 -0.0394 s indicator (0.048) (0.301) (0.052) (0.043) (0.051) (0.091) Wages Export share 0.0027 0.0236 0.0229 -0.0015 -0.0224 -0.0691 (0.117) (0.115) (0.229) (0.071) (0.167) (0.284) Export share*Competivenes -0.0056 -0.0140 -0.0089 -0.0043 -0.0003 0.0104 s indicator (0.031) (0.036) (0.048) (0.018) (0.045) (0.071) Note: * Significant at the 15 % level, ** at the 10 % level, *** at the 5% level. Standard errors are in parentheses. Industry fixed effects are defined at the 2-digit level of the ISIC classification. Source: World Bank Staff calculations based on data from World Bank enterprise surveys and World Economic Forum’s Global Competiveness Report. 74 Annex 6: Labor mobility and labor adjustment costs in Lao PDR Methodology The Trade and Labor Adjustment Costs toolkit was designed to analyze how labor markets in developing countries adjust to permanent trade-related shocks. Examples of such shocks are changes in trade policy, whether at home or abroad, and enduring changes in international trade patterns that affect prices in global markets. These are distinguished from transitory shocks and other short-run business cycle fluctuations in that the shocks that result from economy-wide reallocations of labor are not temporary but permanent. Economic Model underpinning the Trade and Labor Adjustment Costs Toolkit Labor Mobility Costs are estimated based on observed worker transitions between sectors in response to differences in sectoral wages. Using a structural model of workers’ choice of employment sector, a worker employed in sector chooses to remain employed in sector or move to sector but by incurring a cost (for simplicity we assume the economy has only two sectors). This cost has a fixed component (average mobility cost caused by labor market frictions), and a worker-specific component , (the idiosyncratic cost of moving from sector to sector ) that captures personal circumstances such as family constraints or other preferences. The worker’s expected welfare in sector , , is the present discounted value of her real wage, a sector-specific fixed non-pecuniary benefit, and an option value reflecting the possibility of moving to a different sector with a higher wage. If the wage in sector rises, a worker in sector will experience an increase in welfare due to the higher option value even if she never actually moves. None of the components (wage, sector-specific non-pecuniary benefit, and option value) is specific to the worker, only the sector, whereas the idiosyncratic moving cost is particular to the worker. In each period, the worker decides whether or not to move, based on which sector offers a higher expected welfare net of moving costs. The expected welfare benefit of moving from sector i to sector j, ( − ), depends on the wage differential between sectors. The worker will move from sector to sector if the expected welfare benefit of moving ( − ) exceeds the cost of doing so ( + , ), namely if: − ≥ + , . Labor mobility costs can be estimated from the model using data on observed employment flows and wage differentials between sectors. The model of sectoral employment choices generates flows of workers across sectors of the economy, where the solution to the model is the employment allocation. The flows of workers across sectors depend on the model’s parameters, inclusive of the mobility costs . It is then possible to estimate these parameters by matching the predicted flows of workers simulated by the model with the observed flows of workers in the data. Different estimation methodologies are used, depending on the data available. The resulting mobility cost estimates represent a key input variable for simulating the dynamic adjustment paths to the new equilibrium employment-wage outcomes in the affected sector and the remaining sectors of the economy following an exogenous trade-related shock. The market-clearing employment and wage path solutions reflect workers’ optimization of their utility dependent on expected wages and costs to change sectors. 75 Labor Adjustment Costs are estimated for each country facing a hypothetical trade-related sectoral shock, and are calculated as the difference in workers’ welfare between the potential post-shock equilibrium outcome with zero labor mobility costs and the actual post-shock equilibrium in the presence of labor mobility costs. The change in relative prices and real wages following the shock will induce some workers to reallocate their labor across sectors. The magnitude of this reallocation depends on the size of the labor mobility costs. The new resulting equilibrium welfare of a worker, , is compared to her initial pre- shock welfare, , and her potential maximum welfare, , if mobility costs were zero. The maximum potential gains to trade (PG) are therefore − , and actual gains to trade (G) are − . Labor adjustment costs (LAC) representing the forgone welfare gains to trade due to labor mobility costs are therefore: LAC = PG – G = Vmax – V. Source: Artuç, Lederman, and Porto (2013) Data Estimation of labor mobility costs requires panel data on workers’ sector of employment, average sectoral wages, and individual worker characteristics over at least two time periods. Such data was obtained from the Lao PDR Expenditure and Consumption Survey (LECS). The LECS is a household survey conducted over five rounds, the latest two in 2007/08 and 2012/13. Although the survey was not designed as a panel survey (at either the household or individual level), the household-level identifier was kept constant between the last two rounds, making it possible to track households over the 2007/08 and 2012/13 waves, and subsequently individuals within those households based on demographic characteristics (gender and year of birth). Within the households that could be matched over time (32 percent of the sample), about 30 percent of the individuals could be tracked. While this creates possible attrition bias within the sample, this is a caveat to the analysis, as this data is the only source available in Lao PDR that meets the requirements. In addition, labor mobility costs were estimated first for the whole population, and subsequently by skilled versus unskilled workers, and male versus female workers. Labor adjustment costs are subsequently backed out from simulations of trade-related shocks with costly labor mobility, which requires additional information on sectoral consumption shares and the sectoral wage bill as a share of value added. Ideally, such data would be obtained from country-level national accounts (which was not available for Lao PDR). Instead, the needed data was obtained from the data underlying the GTAP database for 2011. Given the reduced sample size, sectors of employment were chosen to be as disaggregated as possible, while ensuring enough degrees of freedom for the estimations. The sectors of employment were based on the ISIC Rev. 3 industrial classification. They include: agriculture, forestry, fishing (1-3); mining and quarrying (5-9); garment (13 and 14); other manufacturing (10-12, 15-33); and services (35-99)47. The sample includes individuals aged 15 and older, to be consistent with the standard international definition 47 Services include: electricity, gas and water supply; construction; trade, restaurants and accommodation services; transport, storage and communications; financial institutions, real estate, leasing and services companies; and community, social and individual services. 76 of the labor force. For individuals reporting multiple sectors of employment, the sector in which the most hours were worked was used. In addition, the methodology allows for individuals to flow in and out of unemployment and the labor force, by including a residual sector. Unemployed individuals were identified as those that were inactive (had not worked during the past seven days, and did not have a permanent job if had not worked) but had searched for work in the past seven days. Out of labor force were identified as those that were inactive and had not searched for work in the past seven days. Average sectoral wages were determined by using the individual-level income data recorded in the LECS and calculating the average income at the sectoral level of the sample. It considers wages in cash during the last month, but excludes household business income and income from agricultural production. Because the survey does not ask for wages by source of income (just total wages last month), only individuals who report one activity are included. Unemployed and out of the labor force are assumed to earn no income. The skill level of a worker is based on educational attainment. An individual was considered skilled if she had obtained/completed a level of education higher than upper-secondary (vocational or university/institute). Table 6.1. Sectoral employment share (%), 2007/08 Total Skilled Unskilled Male Female Agriculture 73.5 1.9 66.5 36.4 37.2 Mining 0.3 0.0 0.3 0.2 0.1 Garment 1.8 0.2 1.9 0.1 1.7 Other manufacturing 4.9 0.5 5.2 3.1 1.8 Services 19.4 6.0 17.6 10.6 8.8 Source: 2007/08 Lao PDR Earnings and Consumption Survey. Table 6.2. Sectoral employment share (%), 2012/13 Total Skilled Unskilled Male Female Agriculture 68.6 1.7 61.6 34.3 34.3 Mining 0.4 0.1 0.4 0.3 0.1 Garment 1.7 0.1 1.7 0.2 1.5 Other manufacturing 4.6 0.4 4.7 2.7 1.9 Services 24.7 9.2 20.1 13.3 11.4 Source: 2007/08 Lao PDR Earnings and Consumption Survey. Table 6.3. Average weekly sectoral wages (Kip), 2012/13 Total Skilled Unskilled Male Female Agriculture 827,890 983,987 857,810 980,399 562,034 Mining 3,181,510 8,478,571 2,310,369 3,692,105 1,417,636 Garment 1,016,258 1,162,500 1,029,611 1,075,385 1,006,145 Other manufacturing 1,904,433 5,077,188 1,512,014 2,214,654 1,249,865 Services 1,658,153 1,536,658 1,847,048 1,672,816 1,633,123 Source: 2007/08 Lao PDR Earnings and Consumption Survey. 77 Table 6.4. Worker transitions across sectors and into and out of labor force status between 2007/08 and 2012/13, total (number of workers) Out of labor force manufacturing Unemployed Agriculture Garment Services Mining Other Unemployed 1 4 11 0 0 0 7 Out of labor force 9 315 277 1 22 23 181 Agriculture 7 421 2706 6 41 106 422 Mining 0 0 7 0 0 0 3 Garment 0 14 45 1 37 3 20 Other manufacturing 1 28 102 1 4 89 76 Services 9 98 284 6 16 56 772 Source: 2007/08 Lao PDR Earnings and Consumption Survey. Table 6.5. Worker transitions across sectors and into and out of labor force status between 2007/08 and 2012/13, skilled vs. unskilled (number of workers) Skilled Unskilled Other manufacturing Other manufacturing Out of labor force Out of labor force Unemployed Unemployed Agriculture Agriculture Garment Garment Services Services Mining Mining Unemployed 0 0 0 0 0 0 2 0 4 9 0 0 0 5 Out of labor force 4 16 6 1 1 5 63 4 124 194 0 15 14 94 Agriculture 2 6 48 2 0 6 68 4 237 2065 3 32 76 306 Mining 0 0 0 0 0 0 0 0 0 6 0 0 0 3 Garment 0 1 1 0 1 0 8 0 9 34 1 31 3 12 Other manufacturing 0 4 2 0 0 6 11 1 17 86 1 4 81 62 Services 6 14 26 0 1 11 267 3 69 228 6 14 42 451 Source: 2007/08 Lao PDR Earnings and Consumption Survey. 78 Table 6.6. Worker transitions across sectors and into and out of labor force status between 2007/08 and 2012/13, male vs. female (number of workers) Male Female Other manufacturing Other manufacturing Out of labor force Out of labor force Unemployed Unemployed Agriculture Agriculture Garment Garment Services Services Mining Mining Unemployed 0 0 4 0 0 0 3 1 4 7 0 0 0 4 Out of labor force 6 113 109 0 1 11 66 3 202 168 1 21 12 115 Agriculture 5 168 1462 5 3 64 251 2 253 1244 1 38 42 171 Mining 0 0 2 0 0 0 3 0 0 5 0 0 0 0 Garment 0 0 0 1 3 0 4 0 14 45 0 34 3 16 Other manufacturing 1 11 71 1 2 54 47 0 17 31 0 2 35 29 Services 5 44 185 5 1 33 407 4 54 99 1 15 23 365 Source: 2007/08 Lao PDR Earnings and Consumption Survey. Simulations of a price shock The simulations are based on a model of trade shocks and labor market adjustment developed by the World Bank’s Trade Department (see Hollweg et al. 2014). In each period, a worker can choose to move from his/her current sector to another sector or work status (unemployed or out of the labor force), but incurs a cost to move. The cost has a common component, which does not vary across time or similar workers in a sector, and a time-varying idiosyncratic component (the differences in the idiosyncratic benefits of being employed in a particular sector). By solving this model, we derive an equilibrium condition, which is a kind of Euler equation. Each worker makes the choice in each period that maximizes his/her lifetime expected utility. The utility function consists of: the average sectoral wage, a non-wage benefit of being employed in a sector, and the option value of moving to a different sector in the future. Figures 6.1-6.4 below plot the sector’s average wage relative to the economy-wide wage (incorporating sectoral labor mobility costs), average utility, and the working age population share for each sector and each time period. 79 Figure 6.1: Post-shock wages, utility and labor shares, skilled workers Wages Utility Labor 1.5 .5 34 Share of working age population Relative average wage .4 Worker utility level 33.5 1 .3 .2 33 .5 .1 32.5 0 0 0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25 Time Time Time W1 W2 W3 V1 V2 V3 L1 L2 L3 W4 W5 V4 V5 L4 L5 Figure 6.2: Post-shock wages, utility and labor shares, unskilled workers Wages Utility Labor 28 .4 1 Share of working age population Relative average wage .8 27.5 .3 Worker utility level .6 27 .2 .4 26.5 .1 .2 26 0 0 0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25 Time Time Time W1 W2 W3 V1 V2 V3 L1 L2 L3 W4 W5 V4 V5 L4 L5 Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey and Global Trade Analysis Project (GTAP). Note: Sector 1 is agriculture/mining; sector 2 is garments; sector 3 is other manufacturing; sector 4 is services; and sector 5 is unemployed or out of the labor force. The y-axis measures the relative average wage/level of utility/working age population share, and the x-axis shows time measured in years (the price shock occurs at time 0). Relative wages reflect embedded labor mobility costs and therefore do not average 1. 80 Figure 6.3: Post-shock wages, utility and labor shares, male workers Wages Utility Labor 28.5 .4 1 Share of working age population Relative average wage .8 28 .3 Worker utility level .6 27.5 .2 .4 27 .1 .2 26.5 0 0 0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25 Time Time Time W1 W2 W3 V1 V2 V3 L1 L2 L3 W4 W5 V4 V5 L4 L5 Figure 6.4: Post-shock wages, utility and labor shares, female workers Wages 24 Utility Labor .4 Share of working age population .8 Relative average wage 23.5 Worker utility level .6 .3 23 .4 .2 22.5 .2 22 .1 0 0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25 Time Time Time W1 W2 W3 V1 V2 V3 L1 L2 L3 W4 W5 V4 V5 L4 L5 Source: 2007/08 and 2012/13 Lao PDR Earnings and Consumption Survey and Global Trade Analysis Project (GTAP). Note: Sector 1 is agriculture/mining; sector 2 is garments; sector 3 is other manufacturing; sector 4 is services; and sector 5 is unemployed or out of the labor force. The y-axis measures the relative average wage/level of utility/working age population share, and the x-axis shows time measured in years (the price shock occurs at time 0). Relative wages reflect embedded labor mobility costs and therefore do not average 1. 81 The World Bank Lao PDR Patouxay Nehru Road P.O Box: 345 Tel: +856 (21) 266 200 Fax: +856 (21) 266 299 Website: www.worldbank.org/lao The World Bank 1818 H Street, NW Washington, D.C. 20433, USA Tel: +1 (202) 4731000 Fax: +1 (202) 4776391 Website: www.worldbank.org 82