We estimate the effects of the extension of compulsory schooling from 5 to 8 years in Turkey on marriage and birth outcomes of teenage girls, using a regression discontinuity design. This law increases girls' schooling by roughly 1 year on average, substantially reducing the probability of marriage by age 16 and first birth by age 17. These effects disappear after ages 17 for marriage and 18 for first birth. These results suggest that this law has strong compulsory-attendance effects during the newly mandated years and human capital effects afterward on marriage and birth outcomes. However, the human capital effects vanish after a couple of years, when students become free to leave school.