Is There a Foster Care-To-Prison Pipeline? Evidence from Quasi-Randomly Assigned Investigators
Introduction
Involvement with child protective services (CPS) and the foster care system is remarkably common in the U.S. Children are placed into foster care when government authorities determine that they have been abused or neglected and it is not safe for them to remain in their homes. By age 18, more than one third of children undergo a formal CPS investigation for alleged maltreatment, and 5 percent enter foster care. Contact with CPS is even more prevalent among Black children, with over half subject to an investigation and up to 9 percent entering foster care during their childhood.
Children placed into foster care are particularly likely to be involved in the criminal justice system as adults. Nearly one-fifth of the prison population in the U.S. consists of former foster children, and about 70% of youth who exit foster care as legal adults are arrested at least once by age 26. Decades of research show a positive association between foster care placement and criminality, and policymakers often cite a “foster care-to-prison pipeline.”
Despite these grim statistics, there is little evidence for a causal relationship between foster care and later-in-life crime. Yet without looking at data, it is unclear whether foster care would reduce or increase adult crime. On one hand, keeping children in harmful home environments could lead to worse adult outcomes. On the other hand, separating children from their families could result in trauma and instability.
Baron and Gross (2023) provide new empirical evidence on the causal relationship between foster care and adult crime. To do so, they assembled a comprehensive administrative dataset linking CPS and adult criminal justice records in Michigan. The paper studies over 100,000 CPS investigations in the state involving children ages 6 to 16 between 2008 and 2016. The empirical strategy exploits plausibly exogenous variation in foster care placements created by the rotational assignment of CPS investigators: investigators are assigned to cases based on a rotational list, not for reasons specific to the child or family, and they have substantial discretion over whether to recommend foster care placement. These decisions are partly subjective, with some investigators being stricter than others. Using investigator stringency as an instrument, we compare the outcomes of children who, by chance, are assigned a strict investigator and placed into foster care to those who are not placed only because they are assigned a more lenient investigator.
Baron and Gross (2023) find that foster care placement substantially reduces the likelihood of later-life criminal involvement for children at the margin of placement. Specifically, children placed in foster care are 25 percentage points less likely to be arrested by age 19 compared to a control mean of 37%, which represents a 68% decrease. The reductions are even more pronounced for convictions (28 percentage points or 81%) and incarcerations (21 percentage points or 80%). These significant declines persist through age 21 for the subset of youth we can observe at older ages.
Exploring potential mechanisms behind the observed decline in adult crime, the paper provides evidence that children’s birth parents made positive changes following their children’s placement into foster care. Most children in our study reunified with their parents after spending one to two years in foster care, and parents themselves were less likely to have criminal justice contact after the removal of their children.
MVPF = 2.4
Net Cost
The net cost of a foster care placement to the government includes both the direct costs of each placement as well as the cost savings from less criminal activity (for example, savings from fewer incarcerations). The direct cost of each placement in Michigan is about $50,000 (ChildTrends, 2021). To estimate the cost savings associated with reduced crime, we pair our detailed criminal justice records, which indicate the types of crimes (if any) children committed, with estimates from Heckman et al. (2010) for the police and court costs associated with each arrest and the incarceration costs for a given incarceration spell. We then construct a variable equal to zero for children who were never arrested by age 19, and equal to the sum of the cost of each arrest and incarceration for each child with subsequent criminal justice contact. Because foster care placement occurs years before an adult arrest, we discount the social cost savings using a 3 to 5% rate (Anders, Barr and Smith, 2022). Using this variable as the outcome in our main specification, we find that these cost savings range from $12,000 to $14,000, depending on the discount rate. Combining the direct cost of each placement with these cost savings, the net cost of foster care to the government is between $36,000 and $38,000.
Willingness to Pay
To calculate the benefits to policy recipients, we measure the social benefit of a foster care placement as the reduction in social costs from the placement’s effects on children’s later-in-life crime. We again pair our detailed criminal justice records which show which types of crimes (if any) children committed with social cost estimates for each type in Chalfin (2015). We construct a variable equal to each child’s social cost of adult crime, equal to zero for children who were never convicted by age 19 and equal to the sum of the cost of each conviction for children who were convicted. For example, if a child was convicted twice, first for homicide and then for carjacking, then we define the child’s social cost as the sum of the social costs of homicide and carjacking. Because foster care placement occurs years before an adult conviction, we again discount the social cost savings using a 3 to 5% rate. Using this variable as the outcome in our main specification, we show that the reduction in social costs ranges from about $84,000 to $95,00 depending on the discount rate.
MVPF Summary
We calculate the MVPF as the reduction in the social cost of crime divided by the net cost of a foster care placement. The MVPF ranges from 2.22 to 2.63 depending on the discount rate, which means that society receives more than $2 in benefits for every $1 in government costs.
It is important to acknowledge that our estimates are limited in scope. For society’s willingness to pay, birth parents presumably have a high willingness to pay to avoid child removal, and this is not taken into account in the estimates above. At the same time, there are several reasons why our estimates could be understated. First, we only measure reductions in criminal justice contact through age 19. Second, our measure of “adult criminality” is almost surely a lower bound since there are fewer convictions than offenses due to (i) the under-reporting of crimes to law enforcement and (ii) the low rate of clearance rates among reported crimes. Third, the benefits we consider include only the children’s subsequent reductions in crime and not the other benefits that we estimate (e.g., the reductions in crime for parents). Finally, as is common in studies in the economics of crime, the social cost of crime reductions excludes relatively minor crimes (e.g., traffic or drug offenses).
References
Anders, John, Andrew Barr, and Alexander Smith (2022). “The Effect of Early Childhood Education on Adult Criminality: Evidence from the 1960s through 1990s.” American Economic Journal: Economic Policy, Forthcoming. https://www.aeaweb.org/articles?id=10.1257/pol.20200660
Baron, Jason E. and Max Gross (2023). “Is There a Foster Care-To-Prison Pipeline? Evidence from Quasi-Randomly Assigned Investigators.” Working Paper. https://drive.google.com/file/d/1dKAv1MEt4he2F_gW8QRdb_fX9PtiRAii/view
Chalfin, Aaron (2015). “Economic Costs of Crime.” The Encyclopedia of Crime and Punishment, 1–12.
ChildTrends (2021). “Child Welfare Agency Spending in Michigan in SFY 2018.
Heckman, James J, Seong Hyeok Moon, Rodrigo Pinto, Peter A Savelyev, and Adam Yavitz. 2010. “The Rate of Return to the HighScope Perry Preschool Program.” Journal of Public Economics, 94(1-2): 114–128. DOI: https://doi.org/10.1016/j.jpubeco.2009.11.001
Hendren, Nathaniel and Ben Sprung-Keyser (2020). “A Unified Welfare Analysis of Government Policies.” The Quarterly Journal of Economics, 135(3): 1209–1318. DOI: https://doi.org/10.1093/qje/qjaa006
Policy Features
- Category
- Criminal Justice
- Sub-Category
- Child Health
- Beneficiary Type(s)
- Children and Youth
- Country of Implementation
- United States of America
- Empirical Method
- IV
- Research Type
- Primary
- Peer Reviewed
- No
- MVPF Publication Link
- drive.google.com/file/d/1dKAv1MEt4he2F_gW8QRdb_fX9PtiRAii/view?usp=sharing