Working Paper #759
Skill Transferability, Migration, and Development: Evidence from Population Resettlement in Indonesia
Sam Bazzi · Arya Gaduh · Alex Rothenberg · Maisy Wong
Between 1979 and 1988, the Transmigration Program in Indonesia relocated two million voluntary migrants from the Inner Islands of Java and Bali to the Outer Islands. We use this large-scale policy experiment to identify the importance of skill transferability for economic development. We proxy for migrants’ skill transferability using two novel measures of economic proximity that capture the agroclimatic and linguistic similarity between migrants’ origins and destinations. The plausibly exogenous assignment of migrants to destinations allows us to provide the first causal estimates of the impact of migrants’ economic proximity on local development outcomes in receiving areas. We show that this quasi-experimental variation provides a novel solution to identification problems in multi-sector Roy models. Transmigration villages exhibit significantly higher agricultural productivity if they were assigned migrants from regions of Java/Bali with more similar agroclimatic endowments and indigenous languages. Limited adaptation—in terms of occupational sorting, crop choices, and ex post migration—may explain why initial origin-by-destination match quality matters over the long-run. We also use a model to show that economic proximity proxies for migrants’ adjustment costs and hence is a measurable source of comparative advantage. Overall, our findings have implications for the design of resettlement programs and debates about adaptation to agroclimatic change. Using an exogenous budgetary shock and a place-based evaluation strategy, we find that poor matching of migrants to destinations may explain the absence of significant average treatment effects of this large resettlement program on local economic development.