Internal migration and human capital accumulation among youth in developing countries

This thesis consists of a short introduction, an overview of the context and data used, three self-contained analytical chapters on migration and human capital in developing countries, and overall conclusions. The first paper documents detailed patterns of internal mobility and estimates the predictors of migration by gender and reasons for moving using a Linear Probability Model and a Multinomial Logit Model, respectively. My main finding is that young migrants are a very heterogeneous segment of the population and that “favourable selection” only holds for those who move for studies. The second paper explores the impact of internal migration on cognitive and psychosocial skills by estimating a 2-Stage Least Squares (2SLS) model, using weather shocks as instruments for migration. My key finding is that migration affects both cognitive and psychosocial skills, but these effects differ across migrants depending on their reason for moving: those that move for studies have higher cognitive and psychosocial skills than non migrants do, whereas those that moved for family formation have lower cognitive skills than non-migrants do. Lastly, the third paper focuses on whether age at migration has an impact on cognitive and psychosocial skills. I use sibling pairs to estimate a household fixed effects-2SLS model, using weather shocks as instruments for migration. My main finding is that younger migrants perform better than older migrants, although this effect can be offset or even dominated by input responses as a result of migration. In conclusion, this thesis provides evidence of the heterogeneous character of young migrants, as well as supports the claims that migration matters for skills formation and that the earlier it takes place, the better.