Data, sampling, measures: This study uses Wisconsin administrative data on a sample of 5,206 first-born children (focal children) whose mothers were not married when they gave birth between 1999 and 2002. The study documents child support and food stamps received by families (economic outcomes) for five years following the focal child's birth and considers how they relate to paternal imprisonment. Previous research suggests that the data used in the study are highly representative of nonmarital births in Wisconsin. The data are matched to accurate date-specific imprisonment records for both biological fathers and half-siblings' fathers, as well as records for child support, employment, and public assistance participation. The sample is restricted to children whose biological fathers and/or half-siblings' fathers were not imprisoned before they reached age three. Through this restriction and controlling for parents' demographic and pre-natal economic covariates, the analyses estimate the effects of paternal imprisonment in the fourth year of the focal child's life (a binary measure) on economic outcomes in the fifth year (in $). Control variables include characteristics of both parents (age, race, earnings prior to the child's birth).
Analytic/statistical approach: To control for the nonrandom formation of partnerships, the study uses difference-in-differences (DD) methods, as well as fixed effects methods as sensitivity tests. Specifically, in the DD analysis, the difference in economic outcomes between the second and fifth years is compared for children who experienced paternal imprisonment in the fourth year and those who did not.
Results: In the fourth year, about 3% of the children's fathers were imprisoned. The results from the DD (or fixed effects) models indicate that paternal imprisonment reduces child support by $701 ($696) per year and increases food stamp benefits by $208 ($253) per year. The results are robust to sensitivity analyses using alternative model specifications and alternative samples.
Conclusions and Implications: The study presents paternal imprisonment as an omitted variable in the existing literature on the determinants of welfare and child support receipt. This research suggests that high rates of incarceration may undermine the success of recent welfare policy reforms designed to improve child support collection and reduce welfare expenditures. In addition, these empirical results challenge the traditional cost-benefit analysis of incarceration, which ignores the costs of incarceration on the families left behind as well as on other public systems.