Child Support and Nonmarital Fertility with a New Partner

Background and Purpose: Child support enforcement is a central element of public polices to improve the economic wellbeing of single-parent families. Children in these families are at elevated risk of living in poverty. They are also at elevated risk of living in complex families in which one or both of their biological parents have had children with other partners. This confluence highlights the importance of considering the potential effects of child support receipt on subsequent fertility with new partners. However, theory is ambiguous: child support could increase new-partner fertility because it makes a mother more attractive as a potential partner, or it could decrease new-partner fertility because it increases her ability to live economically independently. Disentangling any causal effect is complicated by unobserved characteristics related to both child support receipt and subsequent fertility with new partners. In this study, we take advantage of a policy experiment that resulted in randomly assigned differences in child support income to investigate our research question: what is the effect of child support on subsequent nonmarital new-partner fertility? We examine nonmarital fertility because the administrative data we use include a limited subset of marital births. Methods: Our sample includes virtually all 15,300 unmarried mothers who participated in Wisconsin's TANF program in its first 9 months of operation (1997 - 1998), and were randomly assigned to the experimental group (those receiving all current child support paid) or control group (those receiving only partial child support paid). We follow these mothers over a six-year period, identifying any additional nonmarital births to the same and new fathers. A merged administrate dataset drawn from the public assistance programs and child support enforcement systems in Wisconsin is used to identify the sample mothers, their basic demographic and economic characteristics, and whether they have new nonmarital births. We utilize event history analyses to compare nonmarital new-partner fertility for the experimental and control groups, comparing the Kaplan-Meier survivor functions for the two groups and estimating a Cox regression model that accounts for any differences in the initial characteristics of the two groups. We test the experimental effects for several subgroups as well that are expected to have larger experimental effect or higher risks of subsequent fertility. Results: Simple descriptive analysis shows that mothers who receive more child support are less likely to have children with new partner. However, event history analyses, utilizing the randomized variation, do not show a causal effect of child support income on subsequent fertility with a new partner. Conclusions and Implications: Policy-makers may want to use the simple result to argue that more child support should be collected, since that might then cause a decrease in new-partner fertility. But our full analyses show the short-sightedness of building policy around simplistic answers: improvements in child support may have a variety of effects, but changes to fertility are unlikely to be among them. This research implies that policy-makers, researchers, and advocates need to be careful about drawing causal inferences from simple methods.