State-Level Earned Income Tax Credits and Child Welfare Outcomes

Background: There is a well-documented association between poverty and child maltreatment. Preliminary evidence also suggests a causal relationship between receipt of income support and reduced involvement with the child protection system. At this point, however, there is a dearth of research regarding what, if any, relationship exists between child maltreatment and the United States’ largest income support program for working people, the Earned Income Tax Credit (EITC) program. The federal EITC program pays nearly $60 billion in benefits to more than 25 million recipients each year. In addition, as of 2014, 27 states had created supplementary tax credit programs. The credits available through these state programs vary by state and have fluctuated over time as tax credit programs have been introduced, adjusted, and discontinued. As a result, low-income workers within and across states receive differing payouts over time. This study will exploit this variation in order to examine whether changes in state EITC programs are associated with state-level child protective services (CPS) outcomes.

Methods: This study will use a panel data strategy, including data from each state for years 2004-2012. The study’s analysis will consist of ordinary least squares regression with state and year fixed effects. The state fixed effects will control for the many state-specific factors that may affect child welfare outcomes, insofar as such factors remain constant throughout the study period. Dependent variables regarding CPS outcomes will be drawn from the National Child Abuse and Neglect Data System and the Adoption and Foster Care Analysis and Reporting System, and aggregated at the state level. These databases contain state-submitted records of child maltreatment investigations and placements in out-of-home care (OHC). Independent variables include each state EITC as a percentage of the federal credit, a set of economic covariates such as the unemployment and poverty rates and average household income, and a set of demographic covariates including state child population and state racial/ethnic composition.

Results: Preliminary results show that state EITC generosity is differentially associated with physical abuse and neglect. Models estimate that a 10 percentage-point increase in the proportion of the federal EITC available through the state EITC is associated with a roughly 7% increase in the neglect rate, but not associated with the physical abuse, substantiation, or investigation rates. In the models concerning OHC rates, a 10-point increase in EITC programs is associated with approximately a 7% decline in OHC entry, and a small decrease in the share of OHC entries attributed to inadequate housing. There was no association between increases in state EITC generosity and the share of OHC entries due to physical abuse or neglect.

Conclusions and Implications: These findings indicate that changes in state EITCs are differentially associated with various child welfare outcomes, suggesting that different types of child maltreatment – abuse and neglect here – react differently to changes in income support programs. This study further suggests that policymakers who are considering interventions to limit the incidence of CPS involvement may not find effective tools in work-conditioned income support programs such as the EITC.