Parental divorce is a prevalent childhood event. A long literature attempts to estimate the impact of family dissolution on children's human capital formation. Previous studies applying sibling fixed effects estimators find that the timing of divorce has no direct effects on children's outcomes and conclude that the observed raw associations between child age at parental divorce and adult outcomes are driven by selection of parents into divorce. We apply the same methods on new data sources consisting of the universe of all children that experienced parental divorces in Denmark from 1982 onwards. We find small but precisely estimated negative average effects of early family dissolution on children's human capital formation measured from adolescence to the mid-twenties. By studying additional outcomes, we find significant evidence that parental divorce in early childhood leads to higher risk of mental health problems of children in adulthood. Furthermore, we find suggestive evidence that the timing of divorce plays an especially pertinent role for boys and for children of highly educated parents.