Autosseleção e impacto da migração de retorno sobre a distribuição salarial: análise para os migrantes da Região Nordeste

Abstract

The main objective of this paper is to analyze the self-selection and the effect of the return migration to the Northeast region on the workers' income. Using data from PNADs from 2004 and 2014, two methodologies are employed: one parametric and the other semi-parametric. The first one, for the analysis of the return migrant selectivity in terms of unobservable characteristics, is to estimate a mincerian regression, for the logarithm of the wage / hour against several controls and correcting the selection bias of the sample. For the selectivity of the return migrant in terms of observable characteristics, we examine the kernel density functions for the logarithm of the wage / hour. The other methodology is the semi-parametric method proposed by DiNardo, Fortin and Lemieux (1996), used to estimate counterfactual wage densities to know the impact of return migration on the wage distribution. The main evidences obtained show: (i) the negative selectivity of return migrants, both in terms of observable and unobservable traits, as compared to non-returnees; (ii) the effect of return migration is greater for the migrants with lower income levels, indicating that they have made the optimum choice when returning; (iii) the gender wage gap is visually high for non-returnee migrants; (iv) with the simulation the wage densities of men would shift slightly to the left, indicating a small reduction in the distribution of wages if returning male migrants had opted to remain in the Southeast region, especially for those with lower wage levels ; (v) in 2004, women tend to be tied movers, that is, if returning female migrants had opted not to return they would have a better distribution of wages than that obtained with the factual decision.