Returns to Regional Migration: Causal Effect or Selection on Wage Growth?

Abstract

At the interface of economy, geography and sociology the presentation deals with wage effects of geographic mobility. Investigating pecuniary returns to geographic mobility, researchers have to tackle the problem that migrants assess greater innate ability and motivation. Empirical studies show that migrants are favorably self-selected with respect to human capital characteristics. To get rid of potential ability bias former studies used different approaches. On the one hand regressions, matching approaches and Heckman procedures are employed. These methods can deal with selection based on observables only. On the other hand IV-approaches are used, but the instruments employed mostly disenchant their quality after deeper investigation. Therefore several authors suggest the use of fixed-effects estimates to tackle potential ability bias. Using conventional fixed-effects models previous studies control for time constant unobserved heterogeneity only. This approach is valid if selection works on wage-levels. If selection does not work on wage-levels but on wage growth conventional fixed-effects models are biased because the parallel trend assumption does not hold. To tackle this issue, the authors suggest a more general model that allows for individual specific wage profiles. The empirical test draws on the German Socio Economic Panel Data (1991-2010) and compares estimates between conventional fixed-effects models and the more general fixed-effects-individual-slope-model. In that vein, the presentation delivers new insights about the role of selection bias when investigating the pecuniary returns to geographic mobility.