Parental divorce and child educational attainmen

The implications of parental divorce on children's outcomes have been a topic of heated debate among policy-makers and the general public. Citing existing findings of a negative association between parental divorce and child outcomes, recent family policies have geared toward promoting marriage and restricting access to divorce. This line of argument relies on two crucial assumptions: (i) most children from divorced families exhibit worse outcomes; and (ii) the negative correlation between parental divorce and child outcomes is causal. However, the literature base has yet to provide consistent evidence in support of either assumption. The use of different study samples, measures, and methodology makes it difficult to reconcile these conflicting findings.;This dissertation examines the validity of the two assumptions in the framework of one study. Using the Panel Study of Income Dynamics, information on a random sample of children and their families are available annually since childbirth until the completion of their schooling. First, the potential differences in child educational outcomes by parental divorce timing and child gender are examined, by modeling child attainments as a dynamic household production process in a hazard framework. I find that there are significant gender differences in (i) outcomes by timing of divorce, (ii) channels in which these effects work through, and (iii) changes in these effects over time. Second, the assumption of causality of parental divorce on children's attainments is examined in a joint hazard model of child attainment and parental divorce. By modeling parental divorce as an endogenous process, I find no evidence that the observed negative child outcomes is attributable to parental divorce, but may be attributable to unobserved family and child factors.;Given that a large proportion of American children will experience parental divorce while growing up, any simple generalizations in policy design risk stigmatizing a large portion of the population. Based on the findings in this dissertation, policies that restrict access to divorce in general is not an appropriate policy tool, since (i) child outcomes are significantly different based on observed family characteristics, and (ii) what drive these outcomes may have been driving divorce as well.