Returns to education and gender gap

Abstract

The Italian gender wage gap is related to a gap in returns to education which causes a sizable glass ceiling effect. The gap is detected by quantile regressions implemented in different subsets. Quantile regressions allow computation of both the average gap and the divergence in the tails of the wage distribution. Comparison of the equations estimated separately for men and women, reveals a divergence in wage determinants for the average and for all quantiles. The statistical relevance of this divergence is verified by a test of changing coefficients. By repeatedly implementing this test to compare subsets of different regions, cohorts, and education levels, it is possible to rank the factors affecting the gap and to pinpoint at which quantile their impact is greater. Gender turns out to be a relevant source of changes to the coefficients, particularly for the top quantiles, and the regional variable interacts with the returns to education gap, determining a sizeable glass ceiling on southern women’s careers.