A quantile decomposition approach for regional wage differentials in metropolitan Brazil

Abstract

This study intends a better understanding of regional wage gap in Brazil. To accomplish that, a methodology of quantile regression first presented by Koenker and Basset (1978) is used to estimates a regional gap curve for a better image of regional gap behavior along the yield curve. Then, a methodology of quantile detailed decomposition developed by Fortin, Lemieux and Firpo (2009) is used to decompose the determinants of this behavior. As a result, the regional gap seems to be more effective for those workers in the top of the distribution, and this effect is positively affected by regional differences in labor force’s composition.