The total fertility rate is well below its replacement level of 2.1 children in high- income countries. Why do women choose such low fertility levels? We study how labor market frictions affect the fertility of college-educated women. We focus on two frictions: uncertainty created by dual labor markets (the coexistence of jobs with temporary and open-ended contracts) and inflexibility of work schedules. Using rich administrative data from the Spanish Social Security records, we show that women are less likely to be promoted to permanent jobs than men. Temporary contracts are also associated with a lower probability of first birth. With Time Use data, we also show that women with children are less likely to work in jobs with split-shift schedules, which come with a fixed time cost. We then build a life-cycle model in which married women decide whether to work or not, how many children to have, and when to have them. In the model, women face a trade-off between having children early and waiting and building their careers. We show that reforms that reduce the labor market duality and eliminate split-shift schedules increase the completed fertility of college-educated from 1.52 to 1.88. These reforms enable women to have more children and have them early in their life-cycle. They also increase the labor force participation of women and eliminate the employment gap between mothers and non-mothers.