Understanding Adoption Subsidies: An Analysis of AFCARS Data. Subsidy Amount


In the first model of subsidy amount, which includes state-level variables (Model 2A), subsidy amount was positively related to the age of the child; the older the child, the larger the subsidy. Increased subsidies were also associated with being IV-E eligible, being male, and being adopted by non-relatives rather than by foster parents. The structure of the adopting family was also a significant factor; being adopted by an unmarried couple or a single female compared to a married couple was significantly related to the child's receiving a larger subsidy.

While the child's race/ethnicity was not a significant factor, being adopted by a Hispanic mother compared to a non-Hispanic white mother was significantly related to receiving a smaller subsidy. Children adopted by single females (but not by single males) received higher subsidies than those adopted by married couples, as did children adopted by unmarried couples. The latter category represents just over 1 percent of adoptions. Children adopted by relatives received smaller subsidies than those placed with foster parents, as did children with special needs.

The two state-level variables included in the model were significant predictors of subsidy amount. Children in states where the mean time in foster care prior to adoption was higher received higher subsidies. Children in states with higher FMAP rates, where the proportion of the subsidy paid with federal funds for IV-E eligible children is higher, had lower predicted subsidies. Since FMAP is inversely related to per capita income in the state, this finding indicates that less wealthy states offer lower subsidies, even with augmented federal support.

For Model 2B, which predicts the amount of the subsidy while controlling for state variation with the use of state dummy variables, most results were similar. However, subsidy amount was no longer significantly related to IV-E eligibility or special needs status. Significant positive or negative differences were found for 42 of the 48 dummy variables representing jurisdiction, indicating that unmeasured state-level factors also played an important role in the amount of the subsidy.

The differences between Models 2A and 2B suggest that the apparent disadvantage for children with special needs, seen in Model 2A, may reflect variations in the extent to which states classify children as having special needs. Table A-2 shows that the proportion of children classified as having special needs ranges from less than 50 percent to 100 percent.

View full report


"report.pdf" (pdf, 1.41Mb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®