Joint versus Sole Custody after Divorce: Causal Effects on Child Outcomes and Family Environment

Tasks

Parental presence plays an important role in a child’s social and academic development. Hence, when a family is disrupted by divorce, questions about where a child will live and how they will have access to parents become of first-order importance. Custody arrangements play a crucial role in this respect: they determine how often and in what way a child spends time with each parent, and potentially impact the amount of interaction and conflict between parents. The aim of this project is to examine the causal effects of joint versus sole custody on children’s outcomes and their family environment. Establishing causality in this context is a very challenging task. There is every reason to conjecture that parents self-select into custody arrangements based on unobservable factors that also play a role for the development of a child. Thus, any observed correlation between child outcomes and custody likely reflects selection and cannot be interpreted as the causal effect of joint versus sole custody.

Methods

To overcome this identification challenge, the project will rely on an empirical strategy not yet used in the family economics literature. The strategy exploits one specific feature of the Swedish Court System: for almost ten years around the turn of the millennium, Swedish district courts randomly allocated custody cases to judges. This randomization of cases to judges provides plausibly exogenous variation in court-ordered custody arrangements provided judges systematically differ in their preference for awarding joint versus sole custody. That systematic heterogeneity in judges’ preferences indeed exists has been successfully established in the economics of crime literature that estimates the causal effects of incarceration. Crucially, when custody cases are randomly assigned to judges, the type of cases and the characteristics of parents and children involved will be uncorrelated with judges’ preferences, which will however affect final verdicts. In using a judge randomization design, the results of this project will be informative of the effect of joint versus sole custody in the population of “uncooperative” parents, where custody
arrangements are disputed and determined in court. Although it will not be possible to generalize the findings to the wider population of “cooperative” parents who divorce without a court-ordered custody arrangement, noncooperative parents are the relevant group from a policy perspective, as they are the ones most directly affected by legal reforms affecting joint versus sole custody.

Data and other sources

The project involves building up a unique data set. In a first step, data will be collected and digitized on the universe of all divorce cases handled by Swedish district courts between 1996 and 2020. In a second step, this case level data will be merged to several registries from Statistics Sweden to construct a large administrative dataset containing various child outcome measures as well as family environment characteristics.