Using panel data for workers who change jobs, changes in several labor outcomes after inter-city migration are estimated by comparing workers in similar circumstances who move to a new city –the treatment group—with those who stay in the same city –the control group. After matching the two groups using Mahalanobis distances over a wide range of covariates, the methodology of “difference-in-difference treatment effects on the treated” is used to estimate changes after migration. On average, migrants experience income gains but their dedication to formal employment becomes shorter. Income changes are very heterogenous, with low-wage workers and those formerly employed by small firms experiencing larger and more sustained gains. The propensity to migrate by groups of sex, age, wage level, initial dedication, initial firm size and size of city of origin is significantly and directly correlated with the expected cumulative income gains of migration, and inversely with the uncertainty of such gains.
