Research

Works in Progress

Higher Gains for Lower Earners: Place Effects, Match Effects, and Dynamic Selection in Rural-Urban Migration

Across experimental and observational studies in the development literature, estimates of the returns to rural-urban migration vary widely, from 0% to more than 30%. I revisit individual-level panel data in Kenya to study how heterogeneity in returns by baseline earnings can clarify conflicting estimates across studies. I then propose a simple earnings decomposition to understand why returns are heterogeneous by baseline earnings. I empirically separate the role of heterogeneous place effects versus heterogeneous worker-location match effects. I find that high rural earners are more likely to use urban migration to insure against negative origin shocks (i.e., leverage temporarily large place effects to smooth income), but sacrifice a high rural earnings premium when they migrate (i.e., negative match effects) so that the nominal return is near zero. Low earners use urban migration as a "move to opportunity'" but have nominal gains only through place effects, not through positive match effects. Low earners select into migration less dynamically than high earners. Lastly, I apply the same decomposition to panel data in Indonesia to demonstrate how this framework generalizes to other settings.

Democracy Vouchers in the Medium Term (with Alan Griffith)