Flexible estimation of wage distributions in the presence of covariates

An estimator of conditional wage distributions based on a piecewise-linear specification of the conditional hazard function is proposed. Under a minimal set of assumptions, the estimator is flexible enough to capture almost any underlying relationship, and is not affected by the curse of dimensionality. It also allows us to derive estimates of the conditional Lorenz curves and Gini indices. The methodology is used to investigate the wage trends in Spain in 1994-1999. The estimation results provide evidence that there has been strong decreases in both the returns to schooling and the inequality indices for workers with low levels of experience; these decreases may partly be explained by the "overeducation"phenomenon, which intensified in this period.