Association for Public Policy Analysis & Management

A large literature shows that men earn higher wages than women, even after accounting for measurable differences in human capital. At the same time, a growing literature shows that mothers earn less than childless women and that fathers earn more than childless men. These findings suggest that a gender gap in the wage returns to parenthood – positive returns for men versus negative returns for women – contributes to the gender wage gap. Studies of gender and parenthood effects have rarely been integrated, however, so we do not know the magnitude of this contribution: how much of the gender gap is a family gap?

This study asks what proportion of the wage gap between men and women in the 1990s and 2000s was due to the combination of wage penalties experienced by mothers and wage bonuses experienced by fathers. In addition, because previous research has found that the wage effects of parenthood differ across demographic groups, we ask whether the role of parenthood in the gender wage gap differs by race or educational attainment.

To estimate wage gaps between parents and childless adults net of differences in human capital that may affect productivity, it is essential to measure actual labor-market experience. In addition, to avoid misclassifying parents who do not live with their children as childless, it is desirable to measure parenthood using birth histories in addition to or instead of household composition. For these reasons, previous studies of parenthood and wages have relied primarily on longitudinal data from the National Longitudinal Surveys or the Panel Study of Income Dynamics. Unfortunately, because the samples from these sources are smaller than the samples from large-scale cross-sectional and short panel surveys, and because childless men and women are a minority of adults, researchers have had limited power to estimate parenthood effects for demographic subgroups.

We overcome the limitations of the longitudinal data used in previous studies using a Census data product, the SIPP Synthetic Beta, which links survey responses from participants in the Survey of Income and Program Participation (SIPP) with the earnings histories of the participants as recorded by the Social Security Administration (SSA). The SIPP survey responses include a count of biological children. We use the SSA earnings histories to construct a measure of labor-market experience. A second contribution of this study is thus methodological: we use the linked SIPP-SSA data to demonstrate how administrative records can be used to augment data from large, nationally representative surveys with a crucial measure of labor-market experience.