June 2018
After decades of convergence, the gender gap in employment outcomes has recently plateaued in many rich countries, despite the fact that women have increased their investment in human capital over this period. We propose a hypothesis to reconcile these two trends: that when they are making key human capital decisions, women in modern cohorts underestimate the impact of motherhood on their future labor supply. Using an event-study framework, we show substantial and persistent employment effects of motherhood in U.K. and U.S. data. We then provide evidence that women do not anticipate these effects. Upon becoming parents, women (and especially more educated women) adopt more negative views toward female employment (e.g., they are more likely to say that women working hurts family life), suggesting that motherhood serves as an information shock to their beliefs. Women on average (and, again, more educated women in particular) report that parenthood is harder than they expected. We then look at longer horizons—are young women’s expectations about future labor supply correct when they make their key educational decisions? In fact, female high school seniors are increasingly and substantially overestimating the likelihood they will be in the labor market in their thirties, a sharp reversal from previous cohorts who substantially underestimated their future labor supply. Finally, we specify a model of women’s choice of educational investment in the face of uncertain employment costs of motherhood, which demonstrates that our results can be reconciled only if these costs increased unexpectedly across generations. We end by documenting a collage of empirical evidence consistent with such a trend.