Issue Papers on Foster Care and Adoption

Growth in the Adoption Population

by Fred H. Wulczyn and Kristin Brunner Hislop, Chapin Hall Center for Children at the University of Chicago March, 2002

This paper is available on the Internet at:http://aspe.hhs.gov/hsp/fostercare-issues02/adoption/index.htm

Contents

• Introduction
• Policy Context
• Research Questions
• Multistate Foster Care Data Archive
• Methodology
• Results
  • Results from Model I: Baseline Model
  • Results from Model II: Increased Adoptions
  • Results from Model III: Decreased Entries to Care
• Conclusions and Implications
• References
• Endnotes

Introduction

This paper explores the possible impact of recent federal policy changes on the future size of the adoption population--that is, children adopted from foster care. We present results from a simulation model that uses data from Chapin Hall's Multistate Foster Care Data Archive to project the size of the adoption population and the number of children in foster care for the next 20 years.⁽¹⁾ In particular, we examine the sensitivity of the projected size of the adoption population to changes in adoption rates and the number of children entering foster care in the future.⁽²⁾ We also consider how these changes affect the size of the adoption population in relation to the number of children in foster care. We find that in the next few years, most likely between 2004 and 2006, the number of children receiving adoption assistance will grow such that it exceeds the number of children in foster care. Further, the adoption population will continue to be larger than the foster care population well into the future, unless there are unusually dramatic changes in the number of foster care admissions. As a consequence, federal and state adoption assistance payments will approach total outlays for foster care board and maintenance within the next 10 to 15 years, depending on admissions to foster care over that period.

[Go to Contents]

Policy Context

Increased admissions to foster care in the mid to late 1980s created a backlog of children in foster care with the goal of adoption. The surge in admissions was particularly evident in the larger states. For example, between 1985 and 1990, the number of children in foster care increased by 63 percent in Illinois and 135 percent in New York (Wulczyn & Hislop, 2000).

The Adoption and Safe Families Act (ASFA), P.L. 105 89 of 1997, was designed to address that backlog in two primary ways: by creating mandated timeframes to help speed the process of adoption, and by creating financial incentives for states to increase the number of completed adoptions. To speed the process of adoption, ASFA requires states to take steps toward the termination of parental rights and the active recruitment of a potential adoptive family for children who have lived in foster care for 15 of the most recent 22 months. The law does provide for some exceptions to this rule if children are living with relatives, if services that would enable reunification have not been provided, or if some other compelling reason exists why termination of parental rights is not in the child's best interests. With regard to financial incentives for states, beginning in Federal Fiscal Year (FFY) 1999, states could receive $4,000 ($6,000 for special needs cases) for every completed foster care adoption in excess of a baseline number of adoptions. The baseline was determined using recent adoption levels. The incentive was initially capped at $20 million for all states combined for FFY 1998, but was later increased to $43 million for Federal Fiscal Year 1999 and beyond. As of September 2000, every state and the District of Columbia and Puerto Rico had qualified for these funds for one or both years of the program (DHHS, 2000). The number of reported adoptions has in fact increased by 79 percent, from 25,693 in 1995 to 46,072 in 1999, meaning that significant numbers of children have moved into permanent homes.⁽³⁾

The rising number of adoptions has both short- and long-term implications. In the short term, policy makers must determine what other changes, in addition to policy, will help improve rates of exit from foster care to permanent adoptive homes and then make the necessary resources available. Over the longer term, however, the growing number of adopted children creates new responsibilities and associated costs that must be considered. First, the significant increase in adoptions will carry increased costs associated with post-adoptive services. In addition, public officials will have to contend with a potential loss of foster home beds because most adoptions are by foster parents. Finally, adoption assistance payments will represent a growing financial responsibility for federal and state governments. According to The AFCARS Report (DHHS, April 2001), 88 percent of children who were adopted during Federal Fiscal Year 1999 were eligible for adoption assistance payments that will continue until the children reach age 18. As the number of adoptions increases, more children will receive subsidies, and to the extent that policies are successful in reducing the time to adoption, children will be adopted at a younger age and thus receive the subsidies for a longer period of time. Both of these trends will increase the amount of funding required to meet adoption assistance obligations, an increase that would be offset somewhat by lower foster care costs.

[Go to Contents]

Research Questions

The purpose of this paper is to develop an understanding of the future size of the adoption population with a particular focus on how the factors that affect the number of adoptions, such as changes in adoption rates or the number of children entering foster care, will influence the size of the adoption population. Using data from several states included in the Multistate Foster Care Data Archive, we address the following questions:

• Using a variety of assumptions about exit rates and entries to foster care, what is the projected size of the adoption population on January 1, from 2000 through 2020?
• Using a variety of assumptions about exit rates and entries to foster care, what is the projected number of children in foster care on January 1, 2000 through 2020?
• How sensitive is the projected size of the adoption population to changes in adoption rates and number of children entering foster care?
• How do changes in the number of children entering foster care and exit rates from foster care affect the relationship between the size of the adoption population and the size of the foster care population?

[Go to Contents]

Multistate Foster Care Data Archive

The data used to model the adoption population projections come from the Multistate Foster Care Data Archive (the Archive) maintained by Chapin Hall Center for Children at the University of Chicago. The Archive is a database constructed from information drawn directly from the administrative databases that state agencies use to manage and operate their child welfare programs. The Archive currently maintains data from twelve states: Alabama, California, Illinois, Iowa, Maryland, Michigan, Missouri, New Jersey, New York, North Carolina, Ohio, and Wisconsin. At last count, the Archive included the placement records for over 1.3 million children placed into foster care.

For each child whose placement is recorded as part of the Archive, the following pieces of information are maintained: type of placement, date of entry, date of exit, and exit destination, as well as the child's gender, age, race/ethnicity, and county of residence. Using this information, we process the data to identify child spells that consist of a continuous period of time in care. A single spell may consist of multiple placements and a child may have more than one spell in care.

Although the Archive stores the same data elements for each of the states, and all states have provided data on activity through December 31, 1998, state data provided to the Archive differ in two ways. First, the time period covered by the data provided by the states varies. The Archive includes information from the late 1970s forward for Illinois, from the early to mid 1980s forward for Michigan, Missouri, New Jersey, and New York, from the late 1980s forward for Alabama, California, and Maryland, from 1990 forward for Ohio and Wisconsin and from 1995 forward for Iowa and North Carolina. In addition, seven states provide data on children who were already in care at the beginning of the reporting period, while five states provide entry-cohort data that consists of information on children who entered care for the first time during the time periods shown. The selection of states for any given analysis is a function of the questions being asked. In this instance, we chose to work with data from Illinois, Michigan, Missouri, New Jersey, and New York because these states provide data on all children in care beginning in the early 1980s. The historical data are critical because many of the children in the adoption population in 1999 were adopted during the preceding decade. No other source of data provides a basis for the analysis presented here.

[Go to Contents]

Methodology

The initial goal of our work was to estimate the number of children in adoption status on January 1 from 2000 to 2020. Given an adoption population on January 1, the change in the size of that population over the next 12 months is a result of both entries to and exits from the population. For the purposes of this work, we defined exits from the adoption population as the number of children who age out of adoption during a year.⁽⁴⁾ We made the assumption that children age out of adoption on their eighteenth birthday. Entries to the adoption population consist of new adoptions occurring during the year. Most new adoptions come from the population of children already in foster care on January 1, although a small number will be children who begin new spells in foster care during the year and exit quickly to adoption. In its most simplistic conceptual form, the equation used for our projections is as follows:

{Adoption Population on January 1} = [{Adoption population on prior January 1} + {new adoptions during the prior calendar year - children aging out during the prior calendar year}].

From this framework, we identified the following six key components that we would use to generate the projections:

1. Adoption population on January 1, 1999
2. Number of children in foster care on January 1, 1999
3. Rate of aging out from adoption
4. Number of new foster care spells started during a calendar year
5. Rates of exit for children in foster care exiting to adoption or some other exit destinations
6. Rates of exit for children starting new spells during a year and exiting to adoption or to other exit destinations by December 31

Data for the first two components, the beginning population counts of children in adoption and children in foster care, are both based on actual counts from the Archive. For the other components, we used the data in the Archive to examine the historical exit rates and entries to foster care from 1990 to 1998. Each of the components is discussed in more detail below.

We used data in the Archive to calculate the number of children in adoption status on January 1, 1999. This population count is composed of all children adopted prior to January 1, 1999 who had not yet reached the age of 18. The specific count was derived by looking at all admissions, exits to adoption, and the date of birth to determine when a child entered the adoption population and whether the child would have reached age 18 prior to January 1, 1999. Likewise, we were able to calculate the number of children in foster care on January 1, 1999 using data already in the Archive. This population count is composed of all children who had entered care prior to and had not yet exited care as of December 31, 1998.

The question of how many children age out from adoption during a given year depends on the age mix of children in adoption and the age at which they exit adoption status. A review of the average age at adoption for children adopted between 1990 and 1998 revealed that the average age at adoption remained relatively constant between 1990 and 1998 with children being adopted, on average, between 6 and 7 years of age. We also examined the percentage of children in the adoption population on January 1, 1990 - 1999 who would have aged out over the next 12 months. We found that between 1990 and 1999, the annual aging out rate increased from about 5 percent to 7 percent in Michigan, from 4 percent to 8 percent in Missouri, from 3 percent to 7 percent in New Jersey, and from 4 percent to 6 percent New York. The aging out rate decreased slightly in Illinois from 7 percent to 5 percent.

An important component of the model is the number of new foster care spells that begin each year. Over time, these admissions replenish the pool of potential adoptees. To understand the historical change in admissions, we examined the number of new foster care spells beginning in each of the states from 1990 to 1998. In Missouri and New Jersey admissions remained relatively constant between 1990 and 1998. Michigan has experienced a gradual increase in the number of new spells since 1992. Although New York experienced a decrease in new spells from about 26,000 in 1990 to about 19,000 in 1995, the number of new spells increased in 1996 to about 23,000 and remained constant through 1998. Illinois experienced an increase in new spells from 1990 through 1994, at which point the number of new spells decreased each year through 1998.

We approached the question of annual rates of exit from care from two angles. First, we considered how many children in care on January 1 exit to adoption or to some other destination during the next 12 months. These exits account for the bulk of the new entrants to the adoption population over the course of a year. Historical rates of exit to adoption were relatively constant in recent years for Michigan, Missouri, and New Jersey; however, Illinois and New York have experienced rapid increases in exits to adoption. Annual rates of exit to other destinations for children in care on January 1 have remained reasonably constant in recent years. Table 1 shows annual rates of exit for children who were in care on January 1, 1998. In 1998, rates of exit to adoption varied among the states from 6 percent in New Jersey to 14 percent in Michigan. In 1998, rates of exit to other destinations varied from 18 percent in Illinois to 30 percent in New Jersey.

We also examined the historical rates of exit for children who began new spells in care during a calendar year and exited to adoption or some other destination by the end of that same year. This is an important consideration when determining the number of children in foster care on the following January 1. In general, less than 1 percent of children who enter care during a calendar year are adopted by December 31. A larger percentage exit to other destinations including reunification by the end of the year. In 1998, these exit rates ranged from 19 percent in Illinois to 37 percent in New Jersey. The rates of exit to other destinations have remained constant during recent years.

After our review of the historical rates of exit from and entries to foster care, we ran three different projection models using three different sets of assumptions pertaining to future trends in foster care admissions and exit rates. First, we created a baseline model that projects January 1 adoption populations from 2000 through 2020 for all five states based on the assumptions that entries to foster care and rates of exit from foster care and adoption remain constant at recent levels into the future. Specifically, this model is based on the assumptions in Table 2.

Since our assumptions in this model do not include any future changes (i.e., improvements) in the adoption process that might be attributable to ASFA, our next step was to examine the sensitivity of the model to changes in the adoption rates or the number of entries to foster care. In order to examine what might happen if policy changes result in increased adoption rates, we created a second set of projections using all of the same baseline assumptions except that we increased the percentage of children adopted from foster care on January 1 by 5 percent each year for 5 years, from 1999 to 2003, and then held it constant for years beyond 2003. Table 3 summarizes these assumptions.

Finally, given the increasing size of the adoption population even without improvements in adoption rates, we were interested in looking at how sensitive the size of the adoption population would be to a reduction in the number of children entering foster care. To do this we created a third set of projections that use all of the baseline assumptions, except that we decreased the number of new foster care spells per year by 5 percent each year for 5 years, from 1999 to 2003, and then held the number of new spells constant at the 2003 level for future years. Table 4 summarizes these assumptions.

[Go to Contents]

Results

We start by noting that in each of the five states the size of the adoption population grew steadily throughout the 1990s. The data presented in Figure 1 show the number of children in active adoption status for Illinois, Michigan, Missouri, New Jersey, and New York on January 1, from 1990 to 1999. All of the states show steady growth in the size of the adoption population throughout the1990s.⁽⁵⁾ Over the full 10-year period, the annual rate of change averaged about 6 percent in Missouri, 10 percent in Illinois and Michigan, and approximately 12 percent in New York and New Jersey. Illinois was the only state to reveal a sharp shift in the trend. For the period from 1990 through 1995, the annual rate of change in Illinois's adoption population averaged slightly less than 6 percent as compared to nearly 19 percent for the period from 1997 through 1999.

Figure 1. Number of Children in Adoption Population on January 1

Source:  Multistate Foster Care Data Archive, Chapin Hall Center for Children.

[Go to Contents]

Results from Model I: Baseline Model

Figure 2 summarizes the results from our initial Baseline Model, which holds exit from and entries to foster care constant at recent levels in the five states. These results suggest that we should expect the five-state adoption population to grow continuously through 2020, exceeding the size of the foster care population by about 2006. The number of children in the adoption status would grow from about 85,800 in 2000 to about 154,500 in 2020, an 83-percent increase. The number of children in foster care decreases at a much slower rate from 126,100 in 2000 to 114,200 in 2020, a 9-percent decrease. However, the foster care population does stabilize by 2009. If current trends hold true for an extended period, the adoption population would exceed the number of children in foster care by about 40,300 children by 2020.

Figure 2. Model I Baseline Estimates of Adoption and Foster Care Populations: 1990-2020, IL, MI, MO, NJ, NY

Source:  Multistate Foster Care Data Archive, Chapin Hall Center for Children.

Results from Model II: Increased Adoptions

In order to examine what might happen if policy changes were to succeed in increasing the rate of adoption, the model parameters were adjusted to reflect an increase in the annual percent of children adopted from foster care of 5 percent for each of 5 years. All other assumptions used in the baseline model are retained.Figure 3 summarizes the results from this model. Based on this model, we should again expect to see the five-state adoption population grow through 2020, but it will exceed the size of the foster care population in about 2004. In comparison to the initial model, the adoption population grows to almost 177,500 in 2020 compared to 154,500 in the baseline model. Thus, the projected increase in the rate of adoption increased the projected adoption population by approximately 23,000 children, or 15 percent. In addition, the gap between the number of children in foster care and the adoption population increases substantially in future years with the adoption population exceeding the number of children in care by over 71,000 children in 2020. The wider gap is, of course, attributed partly to a smaller number of children in foster care, a desirable outcome derived from achieving permanency for children through adoption. In Model II, the foster care population stabilizes at approximately 105,000 children, or about 9,000 fewer than projected in the baseline model.

Figure 3. Model II - Adoption and Foster Care Populations: 1990-2020, IL, MI, MO, NJ, NY

Source:  Multistate Foster Care Data Archive, Chapin Hall Center for Children

Results from Model III: Decreased Entries to Care

Given the rapid growth of the adoption population even without additional improvements to adoption rates, we wanted to examine how the comparative sizes of the foster care and adoption populations would respond to a decrease in the number of new foster care spells started each year. Figure 4 summarizes the results from a model that assumes a decline in the number of new foster care spells of 5 percent per year for 5 years. Under this set of assumptions, the five-state adoption population grows to 128,800 in 2020, about 25,700 fewer than projected with the baseline model. Although the growth of the adoption population is slower in this model, the model shows the adoption population exceeding the foster care population in 2004, even earlier than the initial model.

Figure 4. Model III - Adoption and In Care Populations: 1990-2020, IL, MI, MO, NJ, NY

Source:  Multistate Foster Care Data Archive, Chapin Hall Center for Children.

[Go to Contents]

Conclusions and Implications

Our study of past adoption rates and adoption projections leads to two fundamental conclusions. First, the growing number of children in foster care throughout the 1990s demonstrates that the population of children adopted from foster care started growing well in advance of any potential impact of the Adoption and Safe Families Act. Moreover, our analysis suggests that this growth is likely to continue even without increases in adoption rates. Thus, we should expect that within the next 4 to 6 years, the total number of children in the population of adopted children will exceed the population of children in foster care (a crossover), at least in these five states. However, because these states represent 23 percent of the national foster care population and trends in other states mirror those described here, we expect the national data to track these trends rather closely.

Second, the results of the model indicate that the adoption population will surpass the foster care population, more or less without regard for how the underlying population dynamics change between 2000 and 2005. In other words, changes in the adoption rate and/or the number of admissions to foster care will affect only the eventual population levels, but not the crossover that is due to occur.

The reason why the crossover appears inevitable has to do with foster care admissions during the late 1980s and 1990s and the fact that once a child is adopted, the impact on long-term population dynamics stretches well into the future. Specifically, the residual duration (i.e., how long it will take before children will, on average, age out) among children who have been adopted is approximately 12 years, given an average age of 6 at the time of adoption. In historical context, the large number of children adopted in 1998 and 1999 will exert an influence on the underlying population dynamics in the years beyond 2010.

The implications of these trends are most easily understood in terms of services for children and their families following finalization and adoption assistance payments. Projected increases in the number of children adopted means that the demand for services to support families who have adopted children should grow commensurately. Unfortunately, very little is known about the number of adoptions that are displaced following finalization, the use of services by families with adopted children, or the cost of the services that are provided (Barth, 2001). Given the projected trends, research that addresses these concerns is absolutely critical if we hope to understand fully the costs and benefits of increasing the number of children adopted. In the meantime, this much is clear. If adoption displacement rates remain constant over time -- somewhere between 6 and 12 percent depending on the age of the child according to a study in Illinois (Goerge 1995) -- the population of children needing services will grow in direct proportion to the underlying adoption trends. A study by Goerge and colleagues suggests that the likelihood that adoptions will be displaced following finalization is influenced by the length of time a child has been in foster care. If this holds true in the future and if ASFA leads to earlier adoptions, then displacement rates might fall, but this effect will likely not be enough to offset the increased demand for services. In the end, states agencies along with the federal government will have to allocate greater resources to children adopted from foster care in order to ensure permanency and well-being.

With respect to the adoption assistance payments, the most recent data suggest that about 88 percent of the children adopted nationwide are eligible for assistance payments (HHS 2001) while approximately 89 percent of the children adopted in the five states used for the simulation were Title IV-E eligible, according to figures published by HHS (HHS 2001). Assuming these figures hold true in the future, then about 78 percent of the adoption population will be eligible for federally funded adoption assistance payments.⁽⁶⁾ When this figure is applied to the simulation results for the baseline model, the number of Title IV-E eligible children in the five states rises from 49,000 in 1997 to 120,000 by the year 2020, an increase of 145 percent. If no increase in the unweighted monthly assistance level is assumed, the total cost of assistance payments would rise from $235 million in 1997 to $576 million in 2020, again in just these five states.⁽⁷⁾ To place these estimates in national perspective, the costs presented here could be multiplied by a factor of 4 to 5 to construct a very rough cost estimate for the United States, assuming these five states account for about 23 percent of the total. Our calculation places the estimate of annual adoption assistance payments at $2.9 billion in 2020 (in constant dollars).

Yet another way to look at the long-term cost of current trends is to combine both the foster care and adoption populations. In 1990, our data indicate that adopted children accounted for 23 percent of the total number of foster and adopted children in the five states combined. If current trends continue, adopted children will account for nearly 60 percent of the total by 2020. Moreover, even though foster care caseloads may well drop over the next few years, the total population of children receiving either board and maintenance or adoption assistance payments will continue to grow rather substantially. Our baseline model estimates place the average annual rate of growth in the five states at 3 percent in the years between 2000 and 2020, with somewhat higher growth rates earlier in the period. Overall, the projected total population will grow by 26 percent between 2000 and 2020, despite the forecasted drop in the foster care caseload over the same time period. In short, federal and state expenditures for foster care and adoption assistance payments can be expected to grow even if foster care admissions fall between 2000 and 2005. The only real question, given admission and adoption rates, is the level to which payments will rise.

[Go to Contents]

References

Barth, R. (June 2001). Aftercare and Post-Adoption Services. Presented at Evidence Base for Child Welfare Policy meeting, Washington, D.C.

Goerge, R., Howard, E., and Yu, D. (1995). Adoption Disruption and Displacement: The Illinois Child Welfare System, 1976-94 . University of Chicago: Chapin Hall Center for Children. Discussion paper CS-35.

U. S. Department of Health and Human Services. (January 2000). AFCARS Report #2. [on-line]. Available: www.acf.dhhs.gov/programs/cb/publications/afcars/apr2001.htm.

U. S. Department of Health and Human Services. (September 20, 2000). Press Release: HHS Adoption Awards Bonuses and Grants. [on-line]. Available: www.hhs.gov/news/press/2000pres/20000920.html.

U. S. Department of Health and Human Services. (April 2001). AFCARS Report #5. [on-line]. Available: www.acf.dhhs.gov/programs/cb/publications/afcars/apr2001.htm.

U. S. Department of Health and Human Services. (June 18, 2001). Baseline for the Adoption Incentive Program for Adoptions Finalized in FY 2000. [on-line]. Available: www.acf.dhhs.gov/programs/cb/dis/tables/index.htm.

Wulczyn, F., Hislop, K. B. & Goerge, R. M. (2001). Foster Care Dynamics 1983-1998. Chicago: Chapin Hall Center for Children.

[Go to Contents]

Endnotes

1.  The term foster care is used here to refer to foster family care, relative care, and group (or congregate) care. For a fuller discussion of the underlying data please see Wulczyn, F., Hislop, K. B. & Goerge, R. M. (2001). Foster Care Dynamics 1983-1998. Chicago: Chapin Hall Center for Children.

2.  Throughout the paper we use the term adoption rates. In the context of ASFA and its impact on the number of adoptions, it is important to be clear about what exactly this term means. Generally, ASFA is intended to have two linked effects: an increase in the "speed of adoptions" (i.e., to decrease the time it takes to reach finalization) and an increase in the number of adoptions. The latter implies an increase in the fraction of children admitted to foster care who are eventually adopted. Whether this increase produces an increase in the number of adoptions depends in part on the number of admissions. Moreover, the speed of adoptions can increase without an increase in the proportion and vice versa. For our purposes, when we refer to an increase in the rate of adoption, we mean an increase the proportion of children in foster care who are adopted.

3.  www.acf.dhhs.gov/programs/cb/dis/tables/index.htm

4.  Exits from adoption also include a relatively small number of children whose adoptions are dissolved. We have not included this in our models because the Archive does not track adoption dissolutions, and we did not feel that there was sufficient information available elsewhere on the rate of dissolution and how it might differ by state and age at adoption.

5.  Even though the Archive contains data from the 1980s, it is likely that the counts of children in the adoption population as of January 1, 1990 are somewhat understated because children adopted in the late 1970s are not accounted for in the Archive or any other dataset. However, given the low levels of adoption during these years, the undercount is not, in our judgment, large enough to substantially alter the conclusions presented here.

6.  The 78 percent is calculated as 89 percent eligible for adoption assistance multiplied by 88 percent eligible for Federally funded assistance under Title IV-E. Nationally, the figure is somewhat lower, in part because states that are included in the simulation have higher rates of IV-E eligibility than most other states. See U. S. Department of Health and Human Services. (June 18, 2001). Baseline for the Adoption Incentive Program for Adoptions Finalized in FY 2000. [on-line]. Available: www.acf.dhhs.gov/programs/cb/dis/tables/index.htm.

7.  Each state used in the simulation provides assistance payments adjusted for the age of the child. Generally, older children receive a higher monthly payment. States also make allowances for special needs. New York also adjusts the payment level based on where in the state the family lives. For these rough calculations, payment levels are unweighted. Instead, a simple average payment is used. The source of the payment levels comes from the North American Council on Adoptable Children, State Profiles, found at: http://www.nacac.org/subsidy_stateprofiles.html.