State-Designated Special Needs, Post-Adoption Support, and State Fiscal Stress

Abstract

The Adoption and Safe Families Act of 2007 and its 2003 reauthorization offered bonuses to states that provide more children waiting in foster care with permanent families through adoption. Under ASFA, a majority of states increased the generosity of post-adoption financial support. Moreover, states increased the number and proportion of adopted children that received adoption assistance because the child had a special need. Administrative data indicate that states more diligently recorded the special needs of children, which was necessary to support increases in state requests for federal reimbursements and performance bonuses.

1. Introduction

Adoption is associated with better educational and psychological outcomes for children than long-term foster care (Hansen, 2006a; Triseliotis, 2002). In addition, governments spend about half as much to support a child who has been adopted as they do to support a similar child who remains in long-term foster care (Barth, Lee, Wildfire, & Guo, 2006). Since 1978, Congress has recognized the value of adoption and has promoted it as a means of providing permanency to waiting children (P.L. 95-266 Section 203). In 1980, Congress established the adoption assistance entitlement for adopted children with special needs (P.L. 96-272). More recently, the Adoption and Safe Families Act of 1997 and its 2003 reauthorization (P.L. 105-89 and P.L. 108-145) provided bonuses to states that increase adoptions from foster care.

To say that adoption is a value is not to say that it is inexpensive. Federal adoption assistance expenditures rose from less than $400,000 in fiscal year 1981 to $1.3 billion in 2002 (Dalberth, Gibbs, & Berkman, 2005; Scarcella, Bess, Zielewski, Warner, & Geen, 2004; U.S. House of Representatives, 2000). For fiscal 2007, federal adoption assistance spending is projected to be $2.044 billion, an increase of $161 million over 2006 (Child Welfare League of America [CWLA], 2006).

Although the federal share of adoption assistance is substantial, ranging from 50 to 76% (US Dept. of Health and Human Services [DHHS], 2006), some states find that they must divert funds from the Social Service Block Grant to cover their shares (CWLA, 2006). Fiscal stress led several states to put the brakes on adoption assistance spending in 2003–2005 (North American Council on Adoptable Children [NACAC], 2003; Eckholm, 2005). Oregon attempted a 7.5% across-the-board cut in adoption assistance payments, which was contested by adoptive parents.¹ Missouri’s attempt to institute means testing for receipt of adoption assistance was also contested by parents.² South Carolina reduced its reimbursement for up-front expenses by $1250 per child. Oklahoma tightened its definition of special needs and reduced post-adoption support for future adoptions. Delaware withdrew support for some psychological and medical treatments not otherwise covered by insurance. Kansas also made selected cuts.

Citing the value of adoption to waiting children and the importance of adoption assistance in allowing adoptive families to provide appropriate care for children who were victims of prior abuse and neglect, critics have warned that curtailing post-adoption support of adoptions from foster care is likely to result in fewer adoptions (NACAC, 2003; 2006a). The testimony of child advocates is supported by the statements of parents about the expenses they incur on behalf of their adopted children (Sedlack & Broadhurst, 1993; Children’s Rights, 2006) and by recent estimates of the statistical relationship between post-adoption support and the number of adoptions (Dalberth, et al., 2005; Hansen & Hansen, 2006; Hansen, forthcoming).

This paper describes how some states have found themselves in the contradictory position of promoting adoption while simultaneously cutting post-adoption support. The next sections outline federal adoption incentives and funding for post-adoption support, discuss the likely state response to the federal programs, and state the research question. The fourth section describes the data to be used and its limitations. The final sections document how states responded to federal incentives. I show that the proportion of adopted children who were recorded by states as having no special need declined while— at the same time—state claims for federal reimbursement of adoption and post-adoption expenses and subsidies rose. The coincidence of these trends indicates that states made a concerted effort to make greater use of federal matching funds for support of adoptive families as one strategy to increase adoptions under ASFA.

2. Federal Adoption Incentives to States: A Brief History

Although states have the responsibility of arranging for the long-term care of children whose birth parents’ rights have been terminated, federal monies support about half of the work (Barth, et al., 2006; Dalberth et al., 2005; Hansen & Hansen, 2006; Scarcella et al., 2004). Through Title IV-E of the Social Security Act, the federal government pays half of the administrative cost of creating and maintaining adoptions and three-quarters of the training costs. The Tax Reform Act of 1986 (P.L. 99-514) subsidizes the up-front costs of the adoption of a child from foster care with up to $1,000 of matching funds. The Child Abuse Prevention, Adoption, and Family Services Act of 1988 (P.L. 102-295) provides funds for respite care and parent support groups to help prevent dissolution and disruption.

Adoption assistance payments are the primary vehicles for post-adoption support. Adoption assistance, authorized under the Adoption Assistance and Child Welfare Act of 1980 (P.L. 96-272), is a Title IV-E entitlement to qualified children, and is paid monthly to the adoptive family until the child reaches the age of maturity. A qualified child is one with a state-designated special need, which is a characteristics such as older age or disability that makes adoptive placement more difficult.³ The federal government matches state expenditures on adoption assistance at the same rate that it matches Medicaid expenditures. This rate is called the Federal Medical Assistance Percentage [FMAP]; the FMAP rate varies inversely with state income.

The Adoption and Safe Families Act of 1997 [ASFA] (P.L. 105-89) was a wide-ranging reform of foster care and adoption policy. Under the Adoption Incentive Program of ASFA, states could earn an annual performance bonus for increasing the number of adoptions from foster care. Adoptions over the state’s goal earned the state a bonus of up to $4,000 each. A super-bonus of up to $6,000 per adoption was paid for increasing adoptions of children with special needs. The Adoption Promotion Act of 2003 (P.L. 108-145) altered the super-bonus to award increases in adoptions of children aged nine and older.

3. Framework for Analysis

Recent studies have affirmed that higher adoption subsidies are associated with more adoptions from foster care (Hansen & Hansen, 2006; Hansen, forthcoming), and that adoptions increased under ASFA (Children and Family Research Center, 2003). However, the strategies used by states to increase adoptions from foster care have not been adequately studied.

To see how states might respond to the bonus program within the structure of federal funding for adoption, focus on the largest part of the adoption budget, adoption assistance payments.⁴ Suppose there are A adopted children and R children who qualify for federally-funded adoption assistance and whose families negotiate an adoption assistance payment with the child welfare authority (Hansen & Pollack, 2005).⁵ Then R/A=r is the recipiency rate—the fraction of adopted children who receive federally-funded adoption assistance. When the adoption assistance agreement is made, the state incurs a liability equal to the sum of the present value of the state’s share of monthly payments. This is, essentially, the value of an ordinary annuity: $P=(1-FMAP)(AAP)\left[\frac{{(1+i)}^m-1}{i}\right]$, where m the months between adoption and the age of maturity, i is the interest rate, and AAP is the monthly adoption assistance payment.⁶ If the state achieves an increase in adoptions from foster care so that it receives an adoption performance bonus payment totaling B dollars, then P is reduced by B/R.

Picture a state’s total adoption assistance expenditures as a rectangle with base R and height P, as shown in figure 1.⁷ The federal carrot of the bonus money may induce states to try to increase adoptions by committing post-adoption financial support to more families, in other words, by increasing recipients from R to R*). If the new adoption assistance commitments N=(R*-R)(P-B/R*) are larger than the savings provided by the bonus (R)(B/R*), then the state finds it must increase its adoption assistance budget as a consequence of its response to the federal incentive.⁸ This may what Missouri Governor Matt Blunt meant when he said that adoption subsidies were “spiraling out of control” (Eckholm, 2006).

Figure 1. The State Adoption Budget.

A performance bonus (B) only decreases the state adoption budget if the bonus is greater than new liabilities (N).

The research question is, therefore: did states in fact respond to federal incentives to promote adoptions by increasing recipients of adoption assistance or the recipiency rate? It would not be possible to address this question were it not for a relatively new source of administrative data on adoption.

4. Data

The success of state and federal policy in the late 1990s to promote adoption is apparent in the doubling of adoptions with state agency involvement, from about 25,700 in 1995 to 51,000 in 1999 to just over 52,000 in 2004 and 2005 (US DHHS, 2007).⁹ In addition to providing an incentive for states to increase adoptions, the Adoption Incentive Program of ASFA provided an incentive for data collection. To qualify for bonus payments, states had to document increases in adoptions. Effectively states were required to improve compliance with a 1993 federal rule on submission of data to the Children’s Bureau (Maza, 2000). The Adoption and Foster Care Analysis and Reporting System [AFCARS] includes information on each adoption finalized after state agency involvement. The Children’s Bureau publishes tabulations of some of the data on the Web and in an annual outcomes report (e.g., US DHHS, 2002; 2007). The data used to produce the outcomes report form the basis for the public use version of the data.

The AFCARS adoption data include a categorical variable intended to capture whether an adopted child has a special need that might qualify him or her for federal adoption assistance. The categories of special needs bases that are defined in the AFCARS data set include (1) being of minority race, (2) being of older age, (3) belonging to a sibling group, (4) having a diagnosed medical condition or disability, and (5) having an other (unspecified) special need. AFCARS allows only one special need to be assigned to each child for reporting purposes, even though many children, in fact, have several qualifying special needs. If the state has recorded no special need for the child, the AFCARS record reads not applicable.

By federal definition, special needs are characteristics of a child that can make adoption more difficult. Each state has the latitude to set its own criteria for special needs designation (NACAC, 2006b) and to specify a priority for classifying children if they meet multiple criteria. Some caseworkers may report the one special needs criterion that is the easiest to document. Additionally, some states or individual social workers may specify may use the other category differently. For these reasons, the data regarding special needs are not entirely comparable across states.

AFCARS includes additional detail on diagnosed conditions such as mental retardation, visual or hearing impairment, physical disability, emotional disturbance, and other (again, unspecified) diagnoses. These are referred to as disabilities to distinguish them from state-defined special needs. Disabilities are not mutually exclusive categories; the AFCARS record may indicate that a child has multiple disabilities.

3.1. Limitations of the Data

The Children’s Bureau puts little faith in the AFCARS data for 1995 through 1997.¹⁰ Relatively few states were in compliance with the federal rules on AFCARS or the ASFA requirement for submission of data before fiscal year 1998. For example, in 1995, only 31 states submitted some adoption data to AFCARS. Moreover, the data submitted were incomplete.

About 13% of adoption records for fiscal years 1996 through 2003 contain incomplete data on special needs. For fiscal year 1996, over 19% of adoption records are missing information on special needs basis. For 1997, about 11% of records were missing special needs data. The disabilities variables are similarly incomplete for 1996, but fewer than 10% of cases have missing values for disabilities for 1997.¹¹

The completeness of both special needs and disabilities data improves markedly beginning in fiscal year 1998. Invalid observations of special needs fell to about 3% of records in 1998 and to about 2% in 2003. Invalid observations of disabilities ranged from 0.4% to 6% of observations.

AFCARS also records whether a state requested federal reimbursement of up-front on on-g0ing adoption support. For fiscal year 1996, 10 percent of adoption cases reported in AFCARS included invalid observations of the query on whether the state had claimed Title IV-E reimbursement. Of reported adoptions in 1997, 5.6 percent were missing information on whether IV-E reimbursement was claimed. The AFCARS data on Title IV-E claims are nearly complete for fiscal year 1998 and subsequent years. Only 0.1 to 0.3 percent of observations for 1998 through 2003 were incomplete.

Although the first three years of data are suspect, AFCARS still represents the only source of case-level data on adoptions with state agency involvement that is reasonably consistent in format across states and over time. Moreover, at least at the state level, the AFCARS count for fiscal year 1996 is highly correlated with data reported through other sources (Hansen & Hansen, 2006). The next section describes the special needs and disabilities of adopted children recorded in the AFCARS adoption records compares the AFCARS to what is known about the disabilities of children in the general population.

4. The Special Needs of Adopted Children

Little is known about the special needs and disabilities of adopted children in general. The federal Census of 2000 was the first census to collect data on adopted son/daughter separate from natural born son/daughter and stepson/stepdaughter (Kreider, 2003). The adopted category includes all kinds of adoption—adoption of kin, of stepchildren, adoption through private and public agencies, domestic and international adoption, and independent and informal adoption. The census enumerated 2.1 million adopted children, who represented about 8% of all sons and daughters of householders in 2000. About 1.8% of all households (817,000 households) contained only adopted children. Another 1.8% contained both adopted children and birth children.

The disabilities of children under five were not recorded by the census. For children aged five to 17, householders were twice as likely to report at least one disability among their adopted children as compared to their other children. About 5.2% of birth children were reported to have a disability, while 11.8% of adopted children were reported to have a disability. Householders reported that just less than 1% of birth children had a sensory, a physical, or a self care disabilities, while they reported about 1.5% of adopted children had disabilities that fell into one of those categories. About 4% of birth children and 10.4% of adopted children were reported to have a mental disability, including learning disabilities and difficulties concentrating.

Rates of disability among children adopted after foster care appear to be much higher than in the population of all adopted children. Among all AFCARS adoptions records for 1996 through 2003, only about 46,000 adoptions with state agency involvement (14.6% of valid observations) were finalized for children who did not have state-designated special need.¹² About 33,500 adopted children (10.6% of valid observations) had the special need basis of belonging to minority race. Older age was the special need of almost 75,300 adopted children (almost 24% of valid observations). Needing to be placed with one or more siblings was the special need of about 53,700 adopted children (17% of valid observations). Over 64,000 adopted children had a medical condition or physical disability (about 20% of valid observations). About 10,600 children (13.6%) had a special need that fell outside these categories.¹³

Figure 2 shows the trend in special needs basis for all reported adoptions in fiscal years 1996–2003. Some changes in the designation of special needs between 1997 and 1998 are notable. The fraction of adopted children who were reported to have no state-designated special needs dropped by more than 50%, from about 32% in 1997 to 14% in 1998 and 1999. The fraction of adopted children for whom age constituted the main special need more than doubled, from 10% in 1997 to 26% in 1998.¹⁴ The fraction of children for whom race was a special need declined from 17% in 1997 to 10% in 1998.¹⁵ The fraction of children who were adopted with siblings also increased, but not as dramatically.¹⁶

Figure 2. Special Needs Bases of Adopted Children.

Source: Computation of author from AFCARS Public Use Data.

Adopted children who had a special need because of a medical condition or disability increased from 14 to 24% of adoptions. Less than 8% of the adoptions with state agency involvement from 1996 through 2003 were placements of children with physical disabilities, visual or hearing impairments, or mental retardation, as shown in table 1. The percentage of adopted children who were mentally retarded fluctuated between the bounds of 1.5 and 3.5%. The percentage of adopted children with a physical or sensory disability climbed from 3.3% (just over 400 children) in 1996 to almost 5.3% (1,600 children) in 2003.¹⁷ Children with a physical or sensory disability or mental retardation, however, account for only about one quarter of children whose primary special need was a medical condition or disability.¹⁸

Table 1.

Prevalence of Disabilities by Special Needs Basis (percent of all valid observations)

	No Special Need	Racial or Ethnic origin	Age	Sibling Group member	Medical condition or disabilities	Other	Total
Mental Retardation	0.19	0.27	0.33	0.58	11.42	0.48	2.67
Physically Disabled	0.19	0.57	0.23	0.51	12.09	0.44	2.80
Visually or Hearing Impaired	0.12	0.20	0.16	0.26	5.19	0.29	1.25
Emotionally Disturbed	0.63	2.60	1.59	3.02	41.51	1.66	10.17
Other Diagnosed Condition	2.37	6.12	1.41	3.71	62.12	7.22	15.81

Source: Computation of author from AFCARS Public Use Adoption Data.

The special needs category of medical condition or disability includes children with disabilities of emotional disturbance or another diagnosed condition. Figure 3 shows trends in these two categories of disability. The percentage of adopted children with an emotional disturbance increased from about 5.5% in fiscal years 1996 and 1997 to 8.3% in 1998 and to 12.7% in 2003.¹⁹ For comparison, the National Center for Education Statistics [NCES] (2004) reports that emotionally disturbed students were about 1% of the student population throughout the period. The percentage of adopted children with some other disabling condition, which could include learning disability or psychiatric diagnosis such as attention deficit/hyperactivity disorder, a mood disorder, post-traumatic stress disorder, or a learning disability has increased from an average of 9% for 1996–1997 to 13.7% in 1998 and 18.2% in 2003.²⁰ Among all pre-school to secondary school-age children, about 6.0% had a learning disability during the period (NCES, 2004).

Figure 3. Disabilities of Adopted Children.

Source: Computation of author from AFCARS Public Use Adoption Data.

5. Special Needs Designation as a Response to Federal Incentives

While it is clear that the number of adopted children recorded as having a designated special need or diagnosed disability increased, the changes in policy or practice produced the increase are not transparent. This section uses descriptive statistics from AFCARS to consider the likelihood of three reasons for the increase:

1. Special needs designations increased because children in foster care received more or better services, which resulted in more diagnoses.

2. Special needs designations increased because of better record keeping by the states.

3. Special needs designations increased because states actively sought additional federal funds to support the adoption of children with special needs.

These reasons are not mutually exclusive; they may coexist.

The first possibility is that specific diagnoses rose in the population of adopted children because the children got better physical, mental health, and educational services, and the services are well-covered by Medicaid. Children in out-of-home care do get more health care, on average, than other children on Medicaid (Rubin et al., 2004), so greater awareness of mental, emotional, and learning problems among foster and pre-adoptive parents may increase the diagnosis rate. On the other hand, only 43% of jurisdictions routinely provide mental health and developmental examinations, in addition to physical exams, for children entering out-of-home care (Leslie et al., 2003; Rosenfeld et al., 1997), so the mental health diagnoses seems least likely to grow without an explicit change in policy, which contradicts the trends discussed in the previous section.

The second possibility is that record-keeping improved in case reporting or in case management. If better record-keeping was the main cause of the change in special needs and disabilities designations, we would expect that states that experienced a decline in cases with no recorded special need would have experienced an increase in most of the other special needs categories. Few states, however, have such a clear pattern. The best case for better record keeping is New Mexico: A downward trend in cases in which special needs was not applicable was accompanied by an upward trend in each of the other special needs categories.²¹

The third reason for increasing special needs may be that after 1998 states began to use special needs designation strategically to increase the financial incentives for families to adopt. There is strong evidence in AFCARS that states have promoted the adoption of children by more actively seeking post-adoption support through Title IV-E: Increases in requests for reimbursement through Title IV-E are coincident with decreases in the number of children with special needs recorded as not applicable. Figure 4 shows the national trend in the proportion of adoptions for which Title IV-E reimbursement was claimed, together with the trend in adoptions with no recorded special need. Title IV-E claims jumped from less than 50% of cases in 1997 to 70% of cases in 1998, while cases with no special needs fell from 30% in 1997 to less than 15% in 1998.

Figure 4. Title IV-E Claims Rose as “No Special Needs” Cases Fell.

Source: Computation of author from AFCARS Public Use Adoption Data.

Several states showed dramatic drops in the percentage of children adopted with no special need. Declines of 50 percentage points or more occurred in Kentucky, Delaware, Minnesota, Pennsylvania, Texas and Vermont. Thirteen additional states had declines of 10 percentage points or more. In some states the decline in “no special need” adoptions happened all at once (Colorado, Delaware, Iowa, Maryland, Massachusetts, Minnesota, Montana, New York, Tennessee, Texas, and Wisconsin), while in other states the decline was steady (Arizona, Arkansas, Kansas, Kentucky, North Dakota, Oregon, and Pennsylvania). Smaller, but notable, changes occurred in a few states (Alabama, Iowa, Louisiana, and Puerto Rico).²² Figure 5 shows the coincidence in 11 states between the decline in children with no special need and an increase in requests for reimbursement.

Figure 5. Example Changes in No Special Needs and IV-E Claims by State.

Source: Computation of author from AFCARS Public Use Adoption Data.

Table 2 summarizes the likelihood that Title IV-E reimbursement was claimed on behalf of the child, given the reported special needs basis of the child. The recording of a special needs basis was correlated with request for Title IV-E funds, but not perfectly correlated. For fiscal years 1998–2003, IV-E funds were claimed in an average of about 70% of cases in which a special need basis is indicated. Claims for Title IV-E reimbursement were less often made in fiscal years 1996 and 1997; for the categories medical condition and other the rate of IV-E claims is as low as 44%.²³

Table 2.

Title IV-E Reimbursement Claims by Special Need of Child (percent of valid observations)

	Not applicable	Racial or ethnic origin	Age	Sibling group member	Medical condition or disabilities	Other	Average
1996	20.6	54.6	59.6	64.7	65.0	45.1	51.6
1997	22.0	62.8	62.1	49.3	44.2	39.5	46.7
1998	44.9	69.4	83.4	79.4	78.2	61.1	69.4
1999	48.9	75.6	85.4	80.7	80.4	65.7	72.8
2000	47.4	80.4	85.1	82.0	78.8	67.4	73.5
2001	41.1	85.0	82.0	78.9	78.9	70.9	72.8
2002	40.1	77.3	84.4	77.1	77.6	69.5	71.0
2003	36.5	75.5	78.0	69.6	73.4	67.0	66.7
Average	37.9	74.2	82.1	76.0	75.8	65.0	69.0

Source: Computation of author from AFCARS Public Use Adoption Data.

Taken together, the trends in specific special needs designations indicate that there was an emphasis on documenting specific special needs to support claims for Title IV-E reimbursement. Table 2 shows that Title IV-E claims on behalf of children whose primary special need was older age rose beginning in fiscal year 1998. The share of Title IV-E claims made on behalf of children with other special needs and on behalf of children who were part of sibling groups also rose. These two groups together comprised just over 20% of Title IV-E claims in 1996, but half of claims by 2003. Claims made on behalf of children for whom racial background was a special need fell. While the share of cases without a recorded special need for which Title IV-E reimbursement was claimed went up, indicating that there are still weaknesses in the collection of the data, the share cases with no special need in total claims went down.

6. Conclusion

In the 1990s, when states began in earnest to promote the adoption of children from foster care, they had a number of policy options at their disposal. They could try to increase adoptions by making the public more aware of the need for adoption from foster care (Macomber, et al., 2005; Wilson, Katz, & Geen, 2005). They could increase adoptions by encouraging the adoption option among adults with on-going relationships to children in care (Festinger, 2001; Boots & Geen, 2003). Finally, they could encourage adoption by providing post-adoption support for more children.

The AFCARS data imply that states used special needs designation, and the promise of the federally-funded child-based entitlement that special needs designation represents, to promote adoption. There was a dramatic increase in the number and percent of adopted children who were deemed to have a special need according to state definitions. Simultaneously, claims for federal reimbursement of adoption costs and on-going adoption assistance support increased. Claims by states for federal support of adopted children with most special needs increased, but so did state spending.

The knee-jerk reaction of states to fiscal stress has been to try to cut adoption spending without considering the wisdom of the cuts. Future research on adoption policy must address the question of whether states’ spending on adoption is justified. Is the substantial cost of permanency through adoption worth the programs that must be cut in order to fund it? State legislatures should consider this question in broad terms: Do adoptions have net benefits to society that are as high as sports stadiums? In light of research on the savings to government from adoption (Barth et al., 2006; Sedlack & Broadhurst, 1993), and considering the evidence that expenditures on improving the environments of children have significant future payoffs (Carniero & Heckman, 2003), an initial impression is that the rate of return to adoption is likely to be substantial.