Policy Research Working Paper 10878 High-Skilled Migration from Myanmar Responses to Signals of Political and Economic Stabilization Yashodhan Ghorpade Muhammad Saad Imtiaz Theingie Han Social Protection and Jobs Global Practice August 2024 Policy Research Working Paper 10878 Abstract In recent years, Myanmar has witnessed considerable income. Randomization within the survey indicates that economic and political instability, leading many young political stabilization would potentially reduce high-skilled people, particularly the higher-skilled, to consider migrat- workers’ desire to migrate by about 15 percent, especially ing abroad for improved prospects. This paper employs an among men, those living in high-conflict areas, and persons innovative method to quantify migration intentions among with lower absolute income but higher perceived relative high-skilled youth by analyzing the take-up of migration income. In contrast, prospects of economic stabilization at different wage premia. A randomized survey experiment do not have a significant effect on migration intentions. In then evaluates how hypothetical political and economic the absence of political stability and a reduction in conflict, stabilization scenarios impact these intentions. The findings economic stabilization is unlikely to reduce talent outflows show that 35 percent of the respondents would be willing among the young. to take a similar job abroad for pay equal to their current This paper is a product of the Social Protection and Jobs Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted atyghorpade@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team High-Skilled Migration from Myanmar: Responses to Signals of Political and Economic Stabilization Yashodhan Ghorpade, Muhammad Saad Imtiaz, and Theingie Han JEL Classification: J61, O15, D74, F2 Keywords: Migration, emigration, Myanmar, brain drain, high-skilled migration, conflict High-Skilled Migration from Myanmar: Responses to Signals of Political and Economic Stabilization Yashodhan Ghorpade, Muhammad Saad Imtiaz, and Theingie Han Abstract: In recent years, Myanmar has witnessed considerable economic and political instability, leading many young people, particularly the higher-skilled, to consider migrating abroad for improved prospects. This paper employs an innovative method to quantify migration intentions among high-skilled youth by analyzing the take-up of migration at different wage premia. A randomized survey experiment then evaluates how hypothetical political and economic stabilization scenarios impact these intentions. The findings show that 35 percent of the respondents would be willing to take a similar job abroad for pay equal to their current income. Randomization within the survey indicates that political stabilization would potentially reduce high-skilled workers’ desire to migrate by about 15 percent, especially among men, those living in high-conflict areas, and persons with lower absolute income but higher perceived relative income. In contrast, prospects of economic stabilization do not have a significant effect on migration intentions. In the absence of political stability and a reduction in conflict, economic stabilization is unlikely to reduce talent outflows among the young. 1. Introduction Emigration from Fragile and Conflict-Affected States (FCAS) is driven by political and security considerations in addition to the standard economic drivers of labor mobility. Economic theory postulates migration as an effective means for labor to realize higher returns arising from spatial differences in income earning opportunities. Classical models have described migration as an outcome of wage differences, which in turn is determined by wages, costs of migrating, and the probability of finding a job (Todaro, 1969; Harris and Todaro, 1970; Fields, 1975). The New Economics of Labour Migration posits three main motivations for economic migration: (i) relative deprivation, whereby poorer individuals migrate to increase household income; (ii) investment: as migration aims to enhance household investments through remittances directed towards agriculture or enterprises; and (iii) insurance, where migration serves as a means to safeguard against risks through the diversification of income sources (Stark, 1980; Stark and Bloom, 1985; Lucas and Stark, 1985). On the other hand, the literature on migration from FCAS has tended to focus on forced displacement aspects. Factors such as crime, conflict, violence, human rights violations, acute economic hardship, and climate and environmental factors have been identified as the main drivers of involuntary movements of people within and across borders (Ibáñez 2014). 2 Both approaches tend to miss out the complex nature of decisions underlying economic migration in settings marked by duress. Recent advances have acknowledged the distinct motivations and needs for support among ‘distressed migrants’ (World Bank 2023). Yet, relatively lesser is known about the combination of factors that determine their preferences and choices with respect to migration. The lack of adequate empirical evidence on economic migration under distress also limits effective policy responses to mobility dynamics in such settings. Policies designed purely for economic migrants or for the forcibly displaced may be insufficient or inappropriate for the needs of distressed migrants. This paper is motivated by the trade-off between stability and income for people facing multiple sources of uncertainty and vulnerability. In FCAS, migration decisions are not driven by wage differences alone, but also by concerns about workers’ own security, generalized security-related threats to families and communities, the perceived prospects for (economic and political) stability in the future and perceived and real effects of political instability on the economy (Ibáñez, 2014; Clemens, 2009). Some studies have used traditional economic models of migration to understand migration choices in FCAS. One approach has been to incorporate violent conflict as an additional “push” factor motivating migration. Others have distinctly incorporated factors such as ‘safety’ or ‘peace’ into the migrant’s utility function and argued that the individual seeks to maximize total utility, and not wages or the utility derived from them alone (Morrison and May 1994; Shrestha 2017). In such a framework, the prospective migrant weights stability/ peace offered by migration with income. Which of these prevail to determine migration choices will vary across individuals and locations, 1 and is ultimately an empirical question. A second motivation for this paper pertains to concerns over “brain drain” in impoverished and fragile settings. Mobility tends to be skill-biased: globally, tertiary-educated persons are over seven times as likely to emigrate as those with only primary education, and over three times as likely as those with only secondary education (World Bank, 2023). While the circulation of skills through migration can be a strong driver of economic development, the emigration of highly skilled persons merits some concern in fragile and low-income settings. In such settings, the private and social costs and benefits of high-skilled emigration may diverge widely, potentially resulting in sub-optimal outcomes and inefficiencies at the aggregate level. Pronounced wage differentials between the (fragile) origin country and destination countries may make it optimal for an individual worker to migrate abroad and secure higher earnings, which would undoubtedly improve private wellbeing outcomes. Moreover, if economic prospects in the sending country remain bleak, or worsen, the local demand for labor, including that of high- skilled workers, may reduce, as may potential earnings. Under such circumstances, it may be optimal for the skilled worker to emigrate to realize a higher return on their human capital through opportunities available abroad. However, if large numbers of highly skilled workers leave the country over a short period of time, the country may be deprived of critical human 1 Civilians living in ‘strategic’ locations are more likely to witness contestation (Azam and Hoeffler. 2002; Kalyvas, 2006; Balcells and Steele 2012), while individual characteristics such as wealth, political affiliation, and ethnic identity may exacerbate or limit exposure to violence (Ibáñez and Vélez, 2008; Balcells and Steele, 2012; Adhikari, 2013). 3 resources for specific services and industries, which in turn may affect the wellbeing of larger segments of the population. The net effect of large outward movements of highly skilled workers may therefore be negative, at least in the short- to medium-term, when the supply of highly skilled workers is unable to respond to gaps created by large-scale emigration. Under such circumstances, the sending country may face a difficult choice between two alternatives; (i) the loss of high-skilled workers to critical sectors due to emigration, resulting in a situation often referred to as ‘brain-drain’, and (ii) the diminishing returns to high skills if high-skilled workers remain, as a form of “brain-waste” whereby human capital does not realize its full potential (Garcia Pires, 2015). The terms on which high-skilled individuals in fragile settings may be willing to accept lower- skilled work are generally understudied. When under duress, skilled migrants may take up work abroad that they are overqualified for. This could be because like many economic migrants across different settings, their competencies and qualifications may not be recognized at destination. However, migrating under duress may also result in migrants having lesser bargaining power and time to secure jobs that are well matched to their skills and earnings potential. As a result, workers may end up in lower-productivity and less remunerative jobs, which in turn limits the full potential gains from migration. This could, at least in theory, and in the short- to medium- term, result in simultaneous “brain drain” from the sending country, and “brain waste” at destination. This paper seeks to understand better the motivations for migration among high-skilled youth in a complex setting simultaneously marked by economic and political instability. It also seeks to directly understand the extent to which migration choices respond to economic and political uncertainty, linking the discussion on ‘brain drain’ from fragile states to prospects of stabilization. We situate our study in Myanmar, a fragile and low-income country in East Asia that has recently witnessed successive shocks including the COVID-19 pandemic beginning 2020, the military takeover of the government in 2021, ensuing and intensifying violence and displacement, global food and fuel price spikes, and continued economic slowdown. The World Bank has documented a weak economic outlook for Myanmar characterized by sluggish economic growth, slowdown in the demand for labor, a steep depreciation of the kyat resulting in high levels of inflation, poverty and food insecurity (World Bank 2024). The GDP is still 10 percentage points below pre- pandemic levels, and close to a third of the population was estimated to be in poverty as of December 2023. These mutually reinforcing adverse indicators have led to a downward revision of the project growth rate of the Myanmar economy to 1 percent. Conflict intensity has risen sharply since the military takeover of the government in February 2021, and had even shown a slight upward trend in the first half of 2024. These factors have resulted in both voluntary and forced movements of people within and across borders. The International Food Policy Research Institute (IFPRI) estimates that between December 2021 and June 2023 alone, at least 18.5 percent 4 of the population of Myanmar (around 10 million people) had migrated, either internally or abroad (MAPSA 2024). As economic and security prospects remain bleak, there may be sustained pressure on the population to migrate outside Myanmar. This may be particularly concerning in the case of high- skilled workers that play a critical role in key professions and occupations in Myanmar, especially because of the low share of the population with higher education, relative to regional and neighboring peers (figure 1). High levels of emigration among the already limited numbers of high-skilled workers, such as medical professionals, educators, and engineers, can result in shortages in critical sectors of activity, which in turn can be detrimental to long-term development. Figure 1. Educational attainment, Bachelor’s or equivalent / at least Master's or equivalent, population 25+, total (%) – Myanmar and selected regional and neighboring comparator countries 35 30 25 20 15 0.26 10 5 9.09 0 Bachelors Master's Source: Authors’ calculations using World Development Indicators, for latest available year since 2010. *Data for population 25+ with bachelor’s degree or equivalent not available for Cambodia. Against this backdrop, we attempt to answer the following research questions: (i) How much do high-skilled youth value the prospects of migrating abroad for undertaking work that is similar to their current jobs? (ii) How would the emigration of high-skilled youth respond to prospects of political or economic stabilization? What form of stabilization – political, or economic – matters (more) for the emigration considerations of high-skilled youth? (iii) Which individual characteristics determine migration responses to signals of political or economic stabilization? 5 To answer these questions, we deploy a nationwide phone survey that includes a new approach to ascertain the valuation of migration options among respondents, and a randomized survey experiment that identifies the effects of political or economic stabilization on migration intentions. The remainder of this paper is organized as follows: Section 2 describes the survey conducted for this study including a new approach to measuring valuations of migration, and a survey experiment designed to estimate the effects of stabilization on migration intentions. Section 3 describes the survey data and summary statistics. Section 4 discusses the identification strategy. Section 5 presents the main results as well as key heterogeneous effects. Section 6 concludes with implications for policy and future research. 2. Study Design While international migration typically accompanies the development process, resulting in net gains for both sending and receiving countries especially in the medium- to long-term, a sudden and large-scale immigration of high-skilled workers can result in critical shortages of qualified workers in key sectors of economic activity in the sending country. When emigration is undertaken under duress, induced by economic and/ or political instability, such detrimental effects may be more pronounced. When migration is motivated by a need to escape duress at home, prospective emigrants may be willing to accept lower wages to move abroad, than under more normal circumstances. They may also be more willing to take up work that they are overqualified for. Both such choices represent a potential misallocation of labor, provoked by the difficult circumstances faced by potential emigrants at home. Our study seeks to understand the extent to which (i) political instability (resulting in civil conflict) and (ii) economic uncertainty drive the willingness of high-skilled youth in Myanmar to migrate abroad, the threshold wage differential at which they become indifferent between migrating and remaining at home, and the extent to which they may be amendable to undertake work for which they are overqualified. We examine these questions in the context of the military takeover of the Government in Myanmar since 2021, followed by high levels of violent conflict and economic collapse. As young people in Myanmar face diminished economic prospects, the appeal of emigrating abroad may increase, the reservation wage for work abroad may fall, and the hesitation to perform work they are overqualified for may reduce. Does the willingness of high-skilled youth to emigrate respond more to political or to economic prospects and expectations? We examine these questions using a randomized questionnaire module in which respondents will be randomly assigned one of three hypothetical scenarios representing (i) an improvement in the political situation including the cessation of violent conflict, (ii) an improvement in the economic situation including currency and price stabilization, and (iii) a base scenario representing things as they are at present. 6 Under each of these scenarios, we estimate the wage premium/ discount that would make high- skilled youth indifferent between migrating and staying in Myanmar. 2.1 Measuring the Willingness to Migrate We deploy an iterative question that elicits the respondents’ preference to migrate when offered different levels of wages in the migration destination of choice with respect to current earnings, holding other factors such as the cost of migrating and the type of work offered abroad constant. The main opening question is posed as follows: “Imagine a situation: You are offered a job in a country of your choice which involves the same type of work that you currently do, and which pays you what your currently earn. Your employer at the new job will also help you with relocation, including any travel costs and visas Will you take the job?” Respondents who reply to this opening question in the negative (i.e., are not willing to migrate to take up a similar job abroad for same nominal pay as current earnings), are asked a follow-up question to ascertain if they would be willing to migrate if the same job now came with a 10 percent increase in earnings. For respondents still unwilling to migrate, subsequent iterations propose further increases in the wage offered abroad in multiples of 10 percentage points up to a 100 percent earnings cut (i.e., 20, 30, 40 … 100 percent increases in pay), followed by discrete options of being paid two and three times as much as current earnings. For those who are still not willing to migrate, we ask a new question: “How many times would your salary have to be increased by for you to take up this job abroad?” with discrete possible options ranging from four to ten times current earnings (allowing only whole number responses), a discrete option of “more than ten times”, and a final option that is worded as “will never migrate no matter how much I am paid”. The earnings increment percentage at which the respondent indicates willingness to migrate is the premium required to induce the respondent to migrate. For example, for someone who will migrate if offered a 30 but not a 20 percent increase in earnings to take up a similar job abroad, the wage premium at which they are indifferent between migrating and staying in Myanmar would be assigned as 30 percent. For someone who will migrate for no less than a 5- times increase in pay, the migration-inducing premium will be 500 percent. For anyone who indicates that they would migrate for a job that pays them more than 10 times their current earnings or that they would not never migrate no matter how much they are paid, we assign a migration inducing premium of 1,100 percent. On the other hand, for respondents who reply to this opening question in the positive (i.e., are willing to migrate to take up a similar job abroad for same nominal pay as current earnings), follow-up question is posed to ask if they would still be willing to migrate if the same job now came with a 10 percent cut in earnings. For respondents who still indicate a willingness to migrate, subsequent iterations propose further reductions in the wage offered abroad in multiples of 10 percentage points up to a 90 percent earnings cut (i.e., 20, 30, 40 … 90 percent reductions in 7 pay). The earnings cut percentage at which the respondent is no longer willing to migrate is the (negative) premium required to induce the respondent to migrate. For example, for someone who is willing to take a 20 but not a 30 percent cut in earnings to take up a similar job abroad, the wage premium at which they are indifferent between migrating and staying in Myanmar would be assigned as -20 percent. The calculation of the migration inducing premium using this iterative question, as described above is also graphically depicted in the flowchart below. Figure 2. Iterative module to measure willingness to migrate (same type of work) Following this set-up, we measure the value attached to migration by respondents in terms of their current earnings using three different, but inter-related indicators: i. The likelihood of respondents accepting a similar job abroad for same level of earnings as current pay. ii. The migration-inducing wage premium, as described above. iii. The likelihood of respondents choosing never to migrate regardless of the earnings premium offered. In specifications that include control variables, we explicitly include the respondents’ estimation of cost-of-living differences between Myanmar and the migration destination of their choice2 to 2 Respondents are asked how much they estimate their living costs will be in the migration destination of their choice compared to current costs of living in Myanmar. For those who do not explicitly mention an intent to migrate to a specific country, the question is 8 account for perceived wage differences across migration destinations and in respondents’ assessments of cost-of-living differences. 2.2 Experimental Design: Measuring the Responsiveness of Migration Intentions to Alternate Scenarios of Stabilization We attempt to measure the degree to which migration intentions respond to signals of political or economic stabilization through a randomized survey experiment. We introduce hypothetical scenarios of political and economic stabilization in the opening question described in the preceding sub-section to measure the willingness to migrate. The sample is randomly assigned to two treatment arms representing political and economic stabilization, respectively, or a control group that provides no signal of stabilization. Political stabilization is signaled by prefacing the introductory question with a brief description of positive political developments such that the question would read as, “Imagine a situation: The political situation in Myanmar improves, violent conflict reduces substantially, and a political agreement is reached. You are offered a job in a country of your choice which involves the same type of work that you currently do, and which pays you what your currently earn. Your employer at the new job will also help you with relocation, including any travel costs and visas. Will you take the job?” subsequent iterative questions follow and the three outcomes of interest are measured in the same way as described earlier. Similarly, economic stabilization is signaled by prefacing the introductory question as, “Imagine a situation: The economic situation in Myanmar improves, growth picks up and prices stabilize. You are offered a job in a country of your choice which involves the same type of work that you currently do, and which pays you what your currently earn. Your employer at the new job will also help you with relocation, including any travel costs and visas. Will you take the job?” We then assess whether migration intentions vary with the prospect of political or economic stabilization by comparing migration outcome variables for these two scenarios with the control arm, i.e., the neutral framing of the question as described in Section 3. that does not make any mention of the economic or political situation and therefore implicitly elicits responses under the prevailing situation. Respondents are randomly assigned to the control group or to either of the two treatment arms (political and economic stabilization) with equal probability to be assigned to one of the three groups (control and two treatment arms). Random assignment ensures that observed differences between treatment arms and the control group can be interpreted as causal estimates of the effects of stabilization on migration intentions. modified to elicit their estimation of cost-of-living differences between Myanmar and a migration destination they would choose if they had to. 9 Figure 3. Randomization to measure the Responsiveness of Migration Intentions to alternate scenarios of stabilization. 3. Data 3.1 Survey Data and Sampling Our analysis is based on primary phone survey data collected for this study between January and April 2024. Given the prevailing security and mobility restrictions across much of Myanmar we were compelled to use Computer-Assisted Telephone Interviews (CATI) instead of face-to-face interviews with respondents. The sample comprises 2,400 respondents drawn from all states and regions of Myanmar. These were chosen from an existing panel of around 300,000 households and individuals by a large local survey firm. A target respondent had to simultaneously meet three criteria: be between 20 and 45 years of age, have completed at least a university degree, and be currently employed and residing in Myanmar. Initial calls were made to respondents in the survey firm’s panel to ask if any household member met these criteria and was willing to participate in a survey. Survey interviews (of a typical duration of 20-25 minutes) were conducted soon after securing these confirmations. At least 25 participants were targeted in each State or Region to ensure minimum representation. Once this minimum target was met, interviews were conducted based on the availability of target respondents per the criteria regardless of location. 3 3 The final state/region-wise distribution of the sample closely mirrors the prevalence of employed graduates in the representative Myanmar Household Welfare Surveys conducted by the International Food Policy Research Institute (IFPRI) beginning in 2020. 10 3.2 Summary Statistics Table 1 shows the means and standard deviations of the main outcome and control variables, as well as key demographic, labor market and other characteristics of respondents and their households. Table 1: Summary statistics for key variables Mean SD N Panel A: Outcome variables Job offered abroad: Same as current job (= same skill level) Will accept job abroad for current nominal earnings 0.35 0.48 2400 Migration-inducing wage premium (%) 387.65 496.92 2400 Respondent will not migrate regardless of earnings premium offered (%) 0.23 0.42 2400 Panel B: Control variables Daily survey success rate 0.27 0.08 2398 After 11 February 2024 0.77 0.42 2400 Respondent is male 0.39 0.49 2400 Age of respondent (years) 30.88 5.83 2400 Respondent is married 0.38 0.49 2400 Respondent has children 0.29 0.45 2400 Perceived cost of living difference at destination (1 = same as Myanmar) 2.15 2.28 2281 Panel C: Other covariates Highest level of education is TVET diploma (GTI, GTC etc.) 0.03 0.16 2400 Highest level of education is undergraduate diploma 0.01 0.12 2400 Highest level of education is bachelor graduate 0.94 0.24 2400 Highest level of education is postgraduate diploma 0.01 0.08 2400 Highest level of education is master’s degree 0.01 0.11 2400 Highest level of education is PhD 0.00 0.02 2400 Employment type is employee 0.66 0.47 2400 Employment type is paid apprentice/intern 0.02 0.13 2400 Employment type is employer (that hires workers) 0.12 0.32 2400 Employment type is self-employed (no hired workers) 0.21 0.41 2400 Individual income (USD) 117.89 73.19 2391 Household income per capita (USD) 75.77 85.48 2394 Perceived ratio of individual income to average income for similar work 0.95 1.40 2337 Respondent has a high skilled occupation (ISCO codes 1-2) 0.29 0.46 2394 Willing to migrate 0.52 0.50 2400 Want to migrate alone 0.52 0.50 1245 Want to migrate with family 0.48 0.50 1245 No. of contacts abroad who can help find a job 1.42 3.02 2400 No. of contacts in Myanmar who can help find a job abroad 1.47 3.41 2400 No. of contacts abroad who may be able to host short term 1.03 2.39 2400 Has savings or liquid assets if needed for move abroad 0.63 0.48 2400 Has a valid passport 0.19 0.40 2400 11 Received remittances from abroad 0.16 0.37 2399 Has the ability to migrate (network, liquid assets, and passport) 0.10 0.30 2400 Experienced reduced wages or business 0.36 0.48 2400 Experienced any employment shock 0.53 0.50 2400 Believe political situation will improve economic situation (ordinal, 0-5) 4.42 0.90 2400 Believe economic situation will improve political situation (ordinal, 0-5) 3.17 1.55 2400 Overall risk appetite (ordinal, 1-10) 6.04 2.76 2400 No. of conflict events in township since Feb'21 (ACLED) 164.40 202.57 2400 No. of conflict events in township in the preceding year (ACLED) 33.25 52.12 2400 4. Identification Following the experimental set-up described in section 3.2, we attempt to estimate the following equation: = + 1 + 2 + Where: Yi is the migration intention outcome (as described in section 3.1) for individual respondent i S is the set of randomly assigned scenarios under which the respondent chooses whether to migrate (a baseline scenario presenting no signal of any changes in the future, and two scenarios representing political and economic stabilization, respectively) X is the matrix of individual- and household-level controls. 1 which represents the conditional effect of alternate scenarios on people’s migration intentions, is the main coefficient of interest. Given the randomized allocation of S to the respondents, 1 is treated as the estimate of the causal effect of the treatment arms on migration outcomes, with respect to the base category. Table 2 below shows the results from an ANOVA used to perform a joint test of significance with the null- hypothesis that the means of key demographic and other characteristics of the respondents are not statistically distinct from each-other across treatment arms. We see that the treatment arm assigned is on balance, not correlated with individual characteristics. 4 4 While a few variables show some statistically significant differences of a small magnitude, we later examine the robustness of our estimates to controlling for these variables. 12 Table 2: F-test of joint significance of treatment arm means (H0: ̅ S0 = ̅ S1 = ̅ S2) Variable Control Political Economic F-stat P-value Group Stabilization Stabilization (S0) (S1) (S2) Respondent is male 0.383 0.389 0.394 0.088 0.916 Age of respondent (years) 31.015 30.776 30.853 0.356 0.700 Respondent is married 0.400 0.366 0.386 1.030 0.357 Respondent has children 0.320 0.260 0.287 3.596 0.028 After 11 February 2024 0.771 0.779 0.762 0.300 0.741 Willing to migrate 0.521 0.511 0.526 0.187 0.829 Individual income (USD) 117 116 121 0.914 0.401 No. of conflict related events in 33 34 32 0.424 0.655 township during prior year (ACLED) No. of conflict related events in 162 169 163 0.262 0.769 township since Feb'21 (ACLED) Overall risk appetite (ordinal, 1-10) 6.002 6.022 6.091 0.220 0.803 No. of employment shocks 1.036 1.076 0.954 1.614 0.199 experienced No. of contacts abroad who can help 1.400 1.410 1.437 0.031 0.970 find a job No. of contacts in Myanmar who can 1.371 1.334 1.714 2.947 0.053 help find a job abroad No. of contacts abroad who may be 1.030 1.017 1.049 0.034 0.967 able to host short term Received remittances 0.152 0.161 0.165 0.281 0.755 Self-employed (no hired workers) 0.208 0.209 0.210 0.007 0.993 Employer (hires workers) 0.113 0.107 0.130 1.087 0.338 Paid apprentice/intern 0.016 0.022 0.014 0.782 0.457 Employee 0.664 0.662 0.646 0.349 0.705 PhD degree - 0.001 - 0.974 0.378 Master’s degree 0.013 0.007 0.018 1.878 0.153 Postgraduate diploma 0.006 0.009 0.007 0.208 0.812 Bachelor’s degree 0.936 0.937 0.941 0.098 0.907 Undergraduate diploma 0.015 0.020 0.010 1.125 0.325 Daily survey response rate 0.264 0.270 0.268 0.889 0.411 5. Results 5.1 Effects of Stabilization on Migration Intentions Table 3 shows the differences in migration intentions for the same type of job as the respondents’ current occupation under scenarios of political and economic stabilization, with and without controls. These estimates are relative to the reference category representing a status quo marked by considerable economic and political instability in the country. Across the three outcome measures of interest, we see that only political stabilization will reduce the value attached to 13 migrating abroad. Under the prevailing situation, around 34 percent of respondents would be willing to migrate to take up the same type of job abroad as their current occupation for same pay. The prospect of political stabilization would, however, reduce this share by 5.4 – 5.8 percentage points. Political stabilization would also increase the reservation wage for a job abroad by 59-67 percentage points (viz. the current expected wage premium of around 390 percent). Finally, when presented the prospect of political stabilization, the share of individuals who say they would never migrate increases by around 4 percentage points compared to under the status quo (23 percent). These estimates are similar with the inclusion of controls. However, across the three measures, we see that economic stabilization does not cause any significant changes in migration intentions. The coefficients for the economic stabilization treatment arm dummy are neither statistically significant, nor large in magnitude. This shows a sharp contrast in the likely effects of two potentially different types of stabilization scenarios: political stabilization reduces the wiliness to migrate and the value attached to migration, whereas economic stabilization does not. This suggests that any economic recovery without accompanying political stabilization is unlikely to affect talent outflows from Myanmar. Table 3. The effects of Stabilization scenarios on Migration Intention Outcomes for similar work Individual accepts similar Average Wage Premium to Individual never chooses to job for current job pay Induce Worker to Migrate migrate regardless of wage abroad for Similar Job (%) premium (up to 11x) Scenario: Political -0.054** -0.058** 58.6** 67.3*** 0.038* 0.043** Stabilization (PS) (0.024) (0.024) (24.7) (25.2) (0.021) (0.021) Scenario: Economic -0.034 -0.034 5.3 8.0 0.001 -0.001 Stabilization (ES) (0.024) (0.025) (24.7) (25.24) (0.021) (0.021) Mean 0.347 0.342 387.7 391.7 0.229 0.232 Controls No Yes No Yes No Yes |Coeff|/Mean (PS%) 15.6% 17% 15.1% 17.2% 16.6% 18.5% |Coeff|/Mean (ES%) 9.8% 9.9% 1.4% 2% 0.4% 0.4% N 2400 2279 2400 2279 2400 2279 Standard errors in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01 Reference Category: No signal for any changes in the future Controls include respondents’ gender, age, marital status, parenthood, self-assessed risk appetite, dummy for interviews conducted after 11 February 2024, perceived difference in costs of living between origin and intended destination, daily survey response rate 5.2 Heterogeneous Effects The annex presents a detailed breakup of aggregate effects by different sources of heterogeneity with respect to respondents’ characteristics. We see that men’s migration intentions are more responsive to prospects of stabilization compared to women’s. However, we do not see much difference between groups by age, occupation type, experience of labor market shocks, or the 14 ability to migrate. 5 We now zoom in on three important sources of heterogeneity to better understand the implications of potential stabilization on migration intentions. We see that the effects of political stabilization on reducing migration intentions are stronger in high-conflict areas. 6 This could be because political stability is especially valued in conflict-affected areas where a return to political stabilization would represent a bigger change from the status quo. Interestingly, even in less affected areas, economic stabilization does not appear to affect migration intentions. Conflict Exposure Table 4. Heterogeneity by exposure to conflict events Individual accepts similar Average Wage Premium to Individual never chooses to job for current job pay Induce Worker to Migrate migrate regardless of wage abroad (%) premium (up to 11x) Low High Low High Low High Conflict Conflict Conflict Conflict Conflict Conflict Scenario: Political -0.056* -0.064* 59.924 76.237** 0.017 0.067** Stabilization (PS) (0.034) (0.034) (36.560) (34.650) (0.032) (0.029) Scenario: Economic -0.040 -0.030 -30.008 47.856 -0.048 0.043 Stabilization (ES) (0.034) (0.035) (36.146) (35.330) (0.031) (0.029) Mean 0.310 0.374 415.142 368.941 0.244 0.219 N 1121 1158 1121 1158 1121 1158 Standard errors in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01 Reference Category: No signal for any changes in the future Controls include respondents’ gender, age, marital status, parenthood, self-assessed risk appetite, dummy for interviews conducted after 11 February 2024, perceived difference in costs of living between origin and intended destination, daily survey response rate. We now examine variations in effects by income levels of the respondent. We measure differences by absolute income and by a subjective assessment of the respondent’s income compared to others with similar experience and qualifications in the same occupation in Myanmar. We see that the effects of political stabilization on migration intentions are driven by lower-income individuals in the sample. Equally, however, we see that those who perceive their earnings to be on par with or higher than comparable peers respond more to prospects of political stabilization than those who believe their income is lower than peers’. 5 Measured through a combination of having a passport, as well as social networks and financial resources that can finance migration. 6 Defined as townships with above-median number of conflict events since 2021 in the sample. Conflict events data is taken from ACLED. 15 Absolute Income Table 5. Heterogeneity by Absolute Income Individual accepts similar Average Wage Premium to Individual never chooses to job for current job pay Induce Worker to Migrate migrate regardless of wage abroad (%) premium (up to 11x) Absolute income Low High Low High Low High Scenario: Political -0.080*** -0.036 89.263*** 45.724 0.049* 0.045 Stabilization (PS) (0.031) (0.038) (34.026) (36.823) (0.029) (0.032) Scenario: Economic -0.074** 0.008 46.089 -23.490 0.028 -0.026 Stabilization (ES) (0.032) (0.038) (35.367) (35.197) (0.030) (0.029) Mean 0.310 0.384 449.467 316.811 0.268 0.184 N 1286 993 1286 993 1286 993 Standard errors in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01 Reference Category: No signal for any changes in the future Controls include respondents’ gender, age, marital status, parenthood, self-assessed risk appetite, dummy for interviews conducted after 11 February 2024, daily survey response rate. Perceived Relative Income Table 6. Heterogeneity by Perceived Relative Income Individual accepts similar Average Wage Premium to Individual never chooses to job for current job pay Induce Worker to Migrate migrate regardless of wage abroad for Similar Job (%) premium (up to 11x) Perceived relative Underpaid On par/ Underpaid On par/ Underpaid On par/ income above above above Scenario: Political -0.033 -0.072** 53.244 76.929** 0.019 0.061** Stabilization (PS) (0.040) (0.030) (42.534) (31.477) (0.035) (0.027) Scenario: Economic -0.010 -0.048 29.231 -3.476 0.013 -0.007 Stabilization (ES) (0.041) (0.031) (42.688) (31.560) (0.035) (0.027) Mean 0.333 0.348 386.729 394.479 0.212 0.243 N 827 1452 827 1452 827 1452 Standard errors in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01 Reference Category: No signal for any changes in the future Controls include respondents’ gender, age, marital status, parenthood, self-assessed risk appetite, dummy for interviews conducted after 11 February 2024, perceived difference in costs of living between origin and intended destination, daily survey response rate. Taken together, these findings suggest that while low-income individuals’ interest in migration would reduce with political stabilization, this is likely only when they believe they are paid on par or above comparable individuals. Political stabilization would not diminish migration intentions of lower-income individuals who believe they are underpaid, nor of higher paid individuals. In other words, political stabilization plays a greater role in determining migration 16 decisions once individuals perceive themselves as being well paid in (local) relative terms. While they believe they are underpaid, political stabilization may do little to affect migration intentions. The migration intentions of richer individuals, regardless of their relative income status, do not generally respond to prospects of stabilization. Table 7. Heterogeneity by Absolute and Perceived Relative Income levels Absolute Income: Low Absolute Income High Absolute Income Relative Income: Underpaid Paid on-par/ above Underpaid Paid on-par/ above Individual accepts similar job for current job pay abroad Scenario: Political Stabilization (PS) -0.017 -0.137*** -0.104 -0.013 (0.046) (0.041) (0.085) (0.044) Scenario: Economic Stabilization (ES) -0.036 -0.099** 0.007 -0.002 (0.046) (0.044) (0.087) (0.043) Average Wage Premium to Induce Worker to Migrate for Similar Job (%) Scenario: Political Stabilization (PS) 30.1 141.6*** 150.7** 18.1 (50.6) (46.5) (75.9) (42.4) Scenario: Economic Stabilization (ES) 46.4 43.6 44.7 -33.4 (52.4) (48.5) (70.4) (41) Individual never chooses to migrate regardless of wage premium (up to 11x) Scenario: Political Stabilization (PS) -0.004 0.093** 0.105* 0.034 (0.042) (0.041) (0.062) (0.037) Scenario: Economic Stabilization (ES) 0.029 0.029 0.014 -0.031 (0.045) (0.042) (0.055) (0.035) N 587 699 240 753 Standard errors in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01 Reference Category: No signal for any changes in the future Controls include respondents’ gender, age, marital status, parenthood, self-assessed risk appetite, dummy for interviews conducted after 11 February 2024, perceived difference in costs of living between origin and intended destination, daily survey response rate. 6. Conclusion This paper has examined emigration intentions and valuations among high-skilled youth in Myanmar. Our iterative survey questions allow a novel way to measure the willingness to migrate, which we find varies by the type of job potentially available at destination. When offered the option to migrate abroad to take up a similar job as their current occupation and no explicit signal of any stabilization in Myanmar, around 34 percent of respondents would be ready to take up such work for no additional earnings premium, i.e., at the same nominal earnings level as their current job in Myanmar. On average, respondents would like to receive a little less than a 400 percent increase in nominal earnings for accepting a similar job abroad. Nonetheless, around 23 percent of respondents would prefer not to migrate even when offered a wage that is as high as 17 11 times their current income. Using these three measures, we also see that respondents would be less inclined to migrate for jobs that they are overqualified for, unless they receive a substantial earnings premium. In a context marked by economic and political upheaval, economic and political stability can drastically alter migration intentions. Our survey experiment shows that while political stabilization, understood as a cessation of violent conflict and reaching a political settlement, would reduce migration intentions, economic stabilization alone may not have any such effect. This finding has critical policy implications as it shows the limits of relying on plausible economic improvements to stem talent outflows without addressing political instability and conflict in the country. If political instability continues or worsens, one can expect a gradual exodus of high- skilled individuals out of the country. This in turn could have pernicious implications for the functioning and productivity of key sectors and occupations that rely on high-skilled workers. A disproportionate exit of skilled workers could also harm economy-wide productivity in the long run. Any efforts to stymie such movements by authorities can only aggravate migrants’ vulnerabilities; as long as the underlying intent to migrate remains strong, individuals may be compelled to rely on irregular and risky channels. While this paper has used cross-sectional data to study migration intentions and their interaction with stabilization prospects, there is little robust causal evidence on actual migration flows in response to stabilization in complex settings. Future research could potentially build on some of the methods developed in this paper and use panel data to study high-skilled migration flows from settings characterized by multiple sources of instability. A more detailed analysis of individual characteristics that drive migration responses to stabilization efforts would also deepen some of the initial insights offered in this paper. 18 References Adhikari, p. (2013). ‘Conflict-Induced Displacement: Understanding the Causes of Flight’, American Journal of Political Science, 57(1): 82–9. Azam, J. P and A. Hoeffler (2002). ‘Violence Against Civilians in Civil Wars: Looting or Terror?’, Journal of Peace Research, 39(4): 461–85. Balcells, L. and A. Steele (2012). ‘Warfare, Political Identities, and Displacement in Spain And Colombia’, HICN Working papers 124. Clemens, Michael A. 2009. “Skill Flow: A Fundamental Reconsideration of Skilled-Worker Mobility and Development.” CGD Working Paper 180 (August), Center for Global Development, Washington, DC World Bank (2024) Myanmar Economic Monitor: Livelihoods Under Threat (English). Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/099061124195517221/P5006631cca59607d182041fae 76ab566cc Harris, J.R. And M.P. Todaro (1970). ‘Migration, Unemployment and Development: A Two- Sector Analysis’, The American Economic Review, 60(1): 126–42 Ibáñez, A. M. 2014. "Growth in forced displacement: cross-country, sub-national and household evidence on potential determinants," Chapters, in: Robert E.B. Lucas (ed.), International Handbook on Migration and Economic Development, chapter 13, pages 350-387, Edward Elgar Publishing. Ibáñez, A. M. and C. E. Vélez (2008). ‘Civil Conflict and Forced Migration: The Micro- Determinants and the Welfare Losses of Displacement in Colombia’, World Development, 36(4): 659–76 Kalyvas, S.N. (2006). The Logic of Violence in Civil War. New York: Cambridge University Press. Lucas, Robert EB, and Oded Stark. "Motivations to remit: Evidence from Botswana." Journal of Political Economy 93.5 (1985): 901-918. Morrison, A. R., and R.A. May (1994). ‘Escape From Terror: Violence and Migration In Post- Revolutionary Guatemala’, Latin American Research Review, 29(2): 111–32 19 Myanmar Agriculture Policy Support Activity (MAPSA). 2024. Migration in Myanmar. Myanmar SSP Research Note 106. Washington, DC: International Food Policy Research Institute (IFPRI). https://hdl.handle.net/10568/139828 Shrestha, Maheshwor. (2017) Push and pull: A study of international migration from Nepal (English). Policy Research working paper,no. WPS 7965 Washington, D.C. : World Bank Group. http://documents.worldbank.org/curated/en/318581486560991532/Push-and-pull-a-study-of- international-migration-from-Nepal Stark, O. (1991). The Migration of Labor. Cambridge, MA: Basil Blackwell. Stark, O. and D.E. Bloom (1985). ‘The New Economics of Labor Migration’, The American Economic Review, 75(2): 173–8 Todaro, M. P (1989). Economic Development in the Third World. New York: Longman World Bank. 2023. World Development Report 2023: Migrants, Refugees, and Societies. Washington, DC: World Bank. doi:10.1596/978-1-4648-1941-4 20 Annex Basic Heterogeneities a. By Labor Market Characteristics 21 b. by Demographic Characteristics 22