Divorce and the cognitive achievement of children

Approximately four out of ten American children experience the divorce of their parents. This raises concern because studies in sociology, developmental psychology, and economics show that offspring of divorced parents fare worse than offspring of married parents. The belief that a two-parent family is the ideal environment for raising children is reflected in recent pro-marriage public policies. One difficulty with these policies is that there is substantial evidence that interparental conflict adversely affects children, and such policies may prolong conflict-ridden marriages. The first goal of my paper is to address the following counterfactual question: Would a child whose parents divorced have been better off had the child's parents not divorced? Another goal of this paper is to understand the impact on children's outcomes of implementable family policies, such as monetary incentives to stay married, better enforcement of child support guidelines, and modifications to existing child support regulations. To study the effects of these policies, one needs to know how family structure and children's outcomes are jointly determined. Thus, the third goal of my paper is to answer the question: Through what mechanism are children better or worse off when family structure changes? To address these questions, I develop and structurally estimate a sequential model of a couple's behavior (from marriage onwards). In the model, parents value the quality of their children, as measured by scores on cognitive tests. Parents can affect their children’s quality by investing time and goods, and by improving the quality of the marital relationship as captured by absence of conflict. Child inputs, labor supply, fertility, divorce, child support transfers, and children's quality are endogenously determined. Divorce enables parents to shield their children from conflict but precludes joint time spent by parents with their children. Conflict plays two roles in the model: it is an input in the child quality technology and may also be an impediment to the exploitation of the gains from trade within marriage. Conflict may trigger a divorce through either channel. When estimating the model, I allow for permanent unobserved heterogeneity in preferences and child quality technology. I use the estimated model to answer the first and last questions by simulating behavior excluding the divorce option at the point when partners first want to divorce. Specifically, I compare a child's cognitive achievement and the inputs invested in the child under the counterfactual and the baseline scenario. This comparison gives the counterfactual change in achievement for each child whose parents divorced. I address the question of how policies that change parents' incentives to stay married affect the wellbeing of children by simulating behavior when Arizona’s child support guidelines are fully enforced, or when West Virginia’s monthly $100 marriage bonus is paid to low income married couples. An interesting feature of Arizona’s guidelines is that the financial burden on the father depends on the amount of time he spends with his children. The main findings can be summarized as follows. First, I find that a child whose parents divorced would have been better off had divorce not occurred. Specifically, both the mean and the median within-child difference in test scores are positive, and for 97% of the children this difference is non negative. According to the model, the reason for the estimated gain in test scores is that, when the divorce option is no longer available, a child receives more maternal and paternal time, more goods, and parents engage in some but very little conflict. Second, I find that the West Virginia marriage bonus leaves the divorce rate unchanged; in fact, a bonus of $1200 per month would be required to reduce the divorce rate of low income couples by 5%. Finally, under perfect enforcement, Arizona’s guidelines do not increase the time that fathers spend with children as intended. They do increase mother-child time within marriage, reduce female labor force participation (especially after divorce), and reduce conflict within marriage.