Academic Articles Public Policy

“Child Protection and Child Outcomes” – Joseph Doyle

According to U.S. Department of Health and Human Services, over 1% of all children below the age of 12 were victims of maltreatment in 2004. Child abuse cases appear frequently on the news and it is truly a sad situation. Most people’s first reaction is that we need more stringent supervision of parents and the government should take kids away from abusive homes more frequently.

In the seminar I attended today, Joseph Doyle (“Child Protection and Child Outcomes: Measuring the Effects of Foster Care“) argues that in the case of Illinois, the government may be putting too many kids into foster care. Doyle has gotten access to the Illinois Department of Child and Family Services’ Child Abuse and Neglect Tracking System (CANTS) and has matched children in foster care with other data sources which track 1) delinquency, 2) teen pregnancy and 3) employment status and wages.

In his estimation, Doyle uses variation in an investigator’s propensity to send a child to foster care as an instrument for the likelihood a child is sent to foster care. More explicitly, the instrument is the investigator’s prior removal rate (the percentage of previous cases which he/she has sent to foster care. Since the cases are assigned in a cue to investigators (with the exception of children who are Spanish speakers) he has a quasi-experimental setup. He uses fixed effects as well for each (zip code*county*Spanish Speaking) cell.

Doyle finds that marginal child placed in foster care is 10%-20% more likely to be arrested, 10-20% more likely to become pregnant as a teenager, and 10% less likely to be working when they become an adult than the marginal child who was not placed in foster care. This does not mean that society should completely abandon the foster care system. Since severely abused children will be placed in foster care no matter which investigator is assigned and abuse free homes will never be assigned to foster care, Doyle’s coefficients only measure the impact of foster care on the marginal children. The above estimates represent a Local Average Treatment Effect (LATE). Abuse in foster care homes does occur and having the idealistic view that foster care is always a safe haven for these children may be na├»ve. The policy implication is that children should be assigned to foster care less frequently than is the current status quo in Illinois.