Can't buy me love? Subsidizing the care of related children

This paper uses a reform in Illinois that reduced the monthly subsidy offered to relatives asked to provide foster care as a plausibly exogenous change in the cost of caring for related children. Families offered a 30% lower wage were 15% less likely to provide care, with especially large declines for children who require mental health services, infants and teenagers. One innovation is a sample selection model that uses the foster care placement tendency of child protection investigators to predict entry into the sample-an instrument that should be unrelated to family characteristics due to a rotational assignment process that effectively randomizes investigators to families. Meanwhile, child health, education, and placement outcomes do not appear to suffer following the decline in the subsidy offer, consistent with similar quality levels among marginal kin and non-kin caregivers.