In predicting whether a child received a subsidy in our model containing state-level explanatory variables (Model 1A), several factors were associated with increased likelihood of receiving a subsidy. Age was a significant factor: the older the child, the more likely he or she was to receive an adoption subsidy. The probability of receiving a subsidy was also greater for a child who was IV-E eligible, designated as special needs, or adopted by a single woman rather than a married couple. Children in states with longer mean times in foster care prior to adoption were more likely to receive subsidies.
In contrast, Hispanic children, and children who were adopted by Hispanic mothers (rather than white non-Hispanic mothers) were less likely to have received a subsidy. A child adopted by a relative or a non-relative who was not a foster parent was less likely to receive a subsidy than a child adopted by foster parents. The percentage of the subsidy provided by the federal government was not a significant determinant of subsidy receipt.
The second specification estimated whether a child received a subsidy, controlling for state variation by including a dummy variable representing each state (Model 1B). As in Model 1A, the child's age, IV-E eligibility, and special needs status were all positively and significantly related to subsidy receipt, while adoption by a non-relative (other than a foster parent) was negatively associated. However, in this specification, a child's and a mother's ethnicity, and the adoptive family's structure (other than it being unknown) were no longer significant determinants of whether a child received a subsidy. There were significant differences associated with state, compared to being from California (our comparison state). Significant positive or negative differences were found for 36 of 48 jurisdictions (46 states, District of Columbia, and Puerto Rico). State practice is clearly a major determinant of subsidy receipt.
The difference between Models 1A and 1B suggests that some of the distinctions seen in Model 1A (an apparent disadvantage for Hispanic children and Hispanic adoptive mothers, and a greater likelihood of subsidy receipt for single female adoptive mothers) may be an artifact of demographics and adoptive family characteristics in some states. Even after controlling for variations among states, age, IV-E eligibility, and special needs status are significantly associated with subsidy receipt. The latter two associations are unremarkable, since both IV-E eligibility and special needs status are required for federal adoption assistance. Although states are not bound by these requirements when determining state-funded subsidies, 74 percent of all subsidies for FY 2001 included federal funds, as seen in Table 3-6.