The Intergenerational Consequences of Incarceration

In 2007, over 50% of US prisoners were parents with minor children, accounting for 2.3\% of the total children in the United States (Glaze and Maruschak, 2008). Theoretical work suggests that parental incarceration could negatively affect the long-term educational, labor force and criminal outcomes of these children, but despite the social and economic importance of this phenomenon, researchers have yet to generate good estimates of these effects. Existing studies suffer from at least three shortcomings. First, previous research has typically been based on non-causal comparisons of children with incarcerated parents to those with non-incarcerated parents. Such comparisons may be frustrated by the possibility that incarcerated and non-incarcerated parents may differ in ways other than their incarceration (endogeneity bias). Second, to circumvent such possibilities, researchers have used long lists of control variables to make causal comparisons, but these comparisons often over-control for potential outcomes, and tend therefore to find muted effects (Murray et al., 2012). Third, the literature thus far has focused on short-term outcomes for children, or long-term outcomes for incarcerated parents, but has not addressed long-term outcomes for children.
This project will provide new evidence on the effect of parental incarceration, addressing issues of statistical bias (endogeneity) with an quasi-experimental design that compares children whose parents, upon arrest, were randomly assigned to judges with differing propensities to incarcerate the defendant. This approach allows identification of the causal effect of parental imprisonment (and of incarceration length) for individuals with parents on the margin of incarceration (Loeffler, 2013).
I will execute this research design in two samples. For the first, I have obtained historical prison records covering all inmates in Iowa between 1885 and 1970, which I can link to the publicly available full-count Censuses of 1880 to 1940. I can then estimate how changes to parental prison sentence length affect long-run income and arrests rates for children. Second, I will link modern court records from Cook County to Illinois Vital Records data. With the parent-child match in hand, I will match the children to a number of rich administrative data sources, allowing estimation of the effect of parental incarceration on long term outcomes like earnings, family structure, education, and criminality.
Both sets of estimates will contribute greatly to the current policy debate surrounding the long- term costs of both increased incarceration and longer sentence lengths, and to the academic understanding of intergenerational transmission of parental traits to their children.ca/ncf