“A Large-Scale Survey of International Migrants from Rural Bangladesh: Longitudinal Evidence on Migration Costs, Earnings and Health”
Abstract: Popular attention has focused on the harsh conditions facing overseas guest workers from countries such as Bangladesh to the states of the Persian Gulf, with the assumption of negative health consequences. In contrast, the global empirical literature on migrant health finds generally positive health outcomes for migrants relative to those left-behind, due in large part to self-selection. Yet most such studies match separate datasets from sending and destination country rather than using a binational panel survey. Few focus on guest workers or migrants to non-OECD destinations. The Matlab Health and Socioeconomic Survey (MHSS) is a binational panel survey that follows a representative sample of a rural area of Bangladesh from 1996-7 to 2012-4. Between survey rounds, one quarter of young adult males moved outside the country (most to the Persian Gulf), with one-third migrating internally. All out-migrants were followed, including festival and phone interviews with overseas migrants. MHSS2 achieved a 92% reinterview rate, including 87% of overseas migrants. This paper provides preliminary assessment of data quality for phone and festival interviews, and measures the effects of migrant status (international, internal), return migrant and country of destination on livelihoods, physical health and mental health. Regression estimates account for the potential confounding effects of current and past socioeconomic characteristics of the migrant, his parents, family and community using data back to 1974. Propensity score estimates account for the effects of self-selection into migration, while bounding exercises address the potential effects of differential inter-survey mortality.