Wage effects of motherhood: a double selection approach

Wage differentials between mothers and childless women are estimated correcting for the selectivity bias resulting from two double selection processes: firstly, the motherhood decision and the employment decision, and secondly the motherhood decision and the decision to be employed in a less demanding job. We use Dutch data on women’s wages and construct an indicator for less demanding jobs. Our estimations indicate that the motherhood decision is strongly correlated with both employment and having a less demanding job. This suggests that ignoring these correlations will lead to inconsistent parameter estimations of wage equations. The selectivity corrected estimation of women’s wage differentials indicate that a large part of the wage differential is composed by discrimination compared to estimations without correction for selectivity.