Abstract
The article analyzes the influences of spatial, sector, educational and status variables on earnings changes in Brazil between 1992 and 2011. These factors are considered in its specificity and in its relations to social class. The study combines the use of measures of gross differences and adjusted differences by quantile regression. Models are estimated both in terms of absolute effects (in Brazilian Real) and in relative effects (percentages). The paper covers particularly the changes in median earnings, but also looks for changes in the upper and lower levels of the earnings distribution. The relative losses of earnings of the metropolitan areas have shown strongly intrinsic to the spatial dimension. The fall in gross (not adjusted) earnings discrepancies for the race and gender categories have involved very different combinations between direct and indirect effects. Reductions in the spatial, sector and educational inequalities have contributed to the decrease of the heterogeneity within the social classes. All scenarios designed by the use of statistical controls have revealed that diminished the earnings advantage of having more schooling. Earnings shifts were lower just between the social positions that have been less affected by changes in the dispersion or distribution of education and its direct and indirect effects on earnings.
Keywords:
Inequality; earnings; social class; economic segmentation; education; quantile regression