Returns to Migration

Summarizes and evaluates current studies attempting to measure the wages or earnings effects of migration. The principal general conclusion is that migration dummy variable coefficients estimated in wage level or wage growth equations are of doubtful relevance to actual gains or losses. Current estimates of migration effects do not address the relevant theoretical hypothesis of wealth-maximizing behavior. Three specific conclusions emerge. First, estimated parameters differ between studies — no strong conclusions follow from this body of research. Second, studies generally ignore the impact of selectivity bias on estimated wage or earnings effects; the value of forsaken alternatives is almost always wrongly inferred from the wages or earnings of workers making different choices. Finally, these studies fail to recognize that migration effects measured with cross-sections or short panels may not permit the inference of life-cycle profile shifts similar to the profile displacements inferred from education wage equation coefficients.