Adoption Dynamics: The Impact of the Adoption and Safe Families Act (ASFA)


Issue Papers on Foster Care and Adoption

Adoption Dynamics: The Impact of the Adoption and Safe Families Act

by Fred H. Wulczyn Chapin Hall Center for Children at the University of Chicago August, 2002

This paper is available on the Internet at:

This issue paper has been prepared by Fred H. Wulczyn of the Chapin Hall Center for Children at the University of Chicago for the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services, under contract HHS-100-99-0007, delivery order #4. Additional funding has been provided by the Administration for Children and Families, U.S. Department of Health and Human Services and the Annie E. Casey Foundation. Analyses are based on the Multistate Foster Care Data Archive maintained by the Chapin Hall Center for Children. The opinions expressed in this paper are those of its authors and do not necessarily represent positions of the U.S. Department of Health and Human Services or of the Annie E. Casey Foundation.



In this briefing paper, adoptions from foster care are studied using data from the Multistate Foster Care Data Archive (the Archive). The Archive is a database containing administrative information regarding children in foster care from a number of states. Children included in the study were admitted to foster care for the first time between 1990 and 1999. The goal of the study is to understand what effects, if any, the federal Adoption and Safe Families Act (ASFA) has had on the proportion of children admitted to foster care who were later adopted and the time needed to complete those adoptions.(1) Although ASFA addresses issues pertaining to placement prevention, reunification and adoption, the specific provisions of ASFA that relate to the termination of parental rights and adoption are perhaps most central to the laws overarching purpose. These specific provisions include a tighter definition of reasonable efforts, a timeframe that clarifies when proceedings to terminate parental rights ought to commence, and incentives to states that increase the number of completed adoptions.

Using data from nine of the states included in the Multistate Foster Care Data Archive, the following questions are addressed:

  • Has there been an increase, since ASFA, in the percentage of foster care placements that end in adoption?
  • Has there been an observable decline in the time to adoption for the children who were in care at the time ASFA was passed?

Data from nine Archive states were used for the analyses presented in this issue brief: Alabama, Illinois, Maryland, Michigan, Missouri, New Jersey, New York, Ohio, and Wisconsin. These states were selected because the Archive includes information on all children in these states who first entered substitute care between 1990 and 1999 and provides information on exit destinations for all children who left care.(2)

Results in Brief

Results from the study support four main conclusions. They are:

  • It is too soon to say whether ASFA had an impact on the likelihood of adoption. The longitudinal data used for this study suggest that is still too soon to know whether ASFA increased the likelihood of adoption among children admitted since ASFA was passed. Historically, the likelihood of adoption peaks five to six years following entry into foster care. In the first years following foster care entry, relatively few children are adopted. For example, by December 31, 1999, only 7.2 percent of the children admitted in 1997 had been adopted. If we consider 1998 as the date when ASFA was implemented, we would expect that the full impact of the law will start to emerge in 2003.
  • During the mid-1990s (just prior to ASFAs passage), adoptions were not slowing down. Although data showing the likelihood of adoption and the time needed to complete adoptions were not available at the time, the perception during the early to mid-1990s was that adoptions were taking longer to complete than they had previously. In part, these perceptions prompted the provisions in ASFA that focused on expediting adoptions. The findings for the nine states reported in this study do not reveal a slow down in the adoption process during the pre-ASFA period.
  • The data presented reveal both pre- and post-ASFA effects. For children who were admitted to foster care prior to the passage of ASFA, the data indicate that the unit time likelihood of adoption (conditional probability) was increasing, even before the law changed. The positive increase in the probability of adoption continued after ASFA became law, although the post-ASFA data are still limited.
  • The evidence available suggests that the time to reunification has slowed. As the likelihood of adoption increased, the data point to a decline in the likelihood of reunification. These data suggest that an increased focus on adoption may have unintended consequences for reunification.

Results in Detail

Probability of adoption

The data in Table 1 indicate that through the end of calendar year 1999, 14 percent of the children admitted to foster care in 1990 left care because they were adopted. In 1991 and 1992, the likelihood of adoption was 14.9 percent, slightly higher than reported for 1990, even though the observation period was shorter. These data suggest that for children admitted early in the decade there was a slight increase in the probability of adoption when compared with 1990.

After 1992, the probability of adoption appears to decline. As of December 31, 1999, only 14.4 percent of children admitted in 1993 exited to adoption. However, these data clearly are time-censored  the observation period is too short to draw any inferences about the probability of adoption. More importantly, from the standpoint of understanding the impact of ASFA on the overall probability of adoption, the data indicate that far too little time has passed to determine whether changes consistent with the intent of ASFA have occurred. This is especially true for the 1998 and 1999 cohorts  the first full post-ASFA cohorts. The data indicate that for children entering placement after 1997, fewer than 3 percent of the children admitted had been adopted through the end of 1999. It is certain, however, that additional children from each of these cohorts will be adopted in subsequent years. Generally, five to six years must pass before most children who will be adopted from a given entry cohort have been discharged to an adoptive home.

Table 1: Number of Admissions and Percent Discharged to Adoption by Year of Admission AL, IL, MD, MI, MO, NJ, NY, OH, WI
Year of Admission Number of Admissions Percent Discharged to Adoption as of 12/31/99
1990 75,813 14.0%
1991 74,399 14.9%
1992 69,810 14.9%
1993 68,900 14.4%
1994 73,070 13.4%
1995 65,313 11.6%
1996 69,218 9.3%
1997 66,122 7.2%
1998 62,512 2.4%
1999 54,311 0.6%

Pre-ASFA and Post-ASFA Effects

In Table 2, the conditional probability of adoption is presented for each admission cohort. The conditional probability of adoption refers to the likelihood a child will be adopted in the next year given that the child started the year still in care. It pinpoints changes in the likelihood of adoption from year-to-year, with particular reference to the point in history when each cohort reaches the date when ASFA was implemented (the shaded area).

Despite perceptions that adoptions were taking longer than they should, these data suggest that the adoption process did not slow down during the 1990s. Reading down the columns in Table 2, the data reveal that a stable adoption process was in place between 1990 and 1994 for the first three years after admission into foster care. That is, during the first year in care, members of those three entry cohorts were equally likely to be adopted. The conditional probability of adoption for those years was .01, .03, and .07 during the first, second and third year following admission, respectively. The data also indicate that there was an increase in the conditional probability of adoption among children in care during the middle of the decade. The change can be seen by comparing the experiences of children admitted in 1993 to the children admitted in 1992. Among the children admitted in 1992, the conditional probability of adoption for those children who were still in care in 1995, four years after admission, was .12. For the 1993 cohort, the comparable figures were .13. One year later, the conditional probability for the 1992 cohort was .14, and .18 for children admitted in 1993.

Table 2 Conditional Probability of Adoption by Entry Cohort and Years Since Admission AL, IL, MD, MI, MO, NJ, NY, OH, WI
Years Since Admission
Year of Admission 1 2 3 4 5 6 7 8 9 10
1990 0.01 0.03 0.07 0.10 0.13 0.14 0.13 0.16 0.19 0.15
1991 0.01 0.03 0.07 0.11 0.14 0.14 0.17 0.22 0.17  
1992 0.01 0.03 0.07 0.12 0.14 0.18 0.22 0.20    
1993 0.01 0.03 0.07 0.13 0.18 0.21 0.22      
1994 0.01 0.03 0.07 0.14 0.20 0.23        
1995 0.02 0.03 0.09 0.19 0.23          
1996 0.02 0.04 0.13 0.19            
1997 0.04 0.06 0.13              
1998 0.02 0.05                
1999 0.01                  

If a trend can be detected in these data, the findings suggest that the adoption process was actually speeding up. Moreover, the data indicate the increase actually started before ASFA was passed. The experiences of children admitted in 1993 offer the strongest evidence for this conclusion. After 5 years in care, members of the 1993 had a .18 likelihood of adoption. That is, of the children admitted in 1993 who were still in care 5 years later, 18 percent were adopted during the next year. The comparable figure for children admitted in 1990 was .13. In short, the likelihood of adoption during the fifth year increased by more than one-third. The data for children admitted in 1995 and 1996 offer similar support, during the first and second year of placement. Finally, the shaded areas indicate when the members of each cohort reached the ASFA period. These data suggest that the likelihood of adoption during the year ASFA was passed was generally higher than it had been the year before.

Adoption and Reunification

The final piece of analysis uses a proportional hazard model to compare adoption and reunification rates after also controlling for the attributes of children. Specific results from the multivariate models are presented in Appendix I. Figure 1 shows the results from three separate proportional hazards models. One model examines all exits; the second model considers only reunification; and the third model evaluates only adoptions. In each model, the probability of exit for each admission cohort is compared with the experiences of the 1990 cohort, the base year, after controlling for age, race, urbanicity, and placement type. The statistic reported in Figure 1 is the relative risk of exit. Rates above one imply faster movement when compared with the children admitted in 1990.(3)

With respect to all exits, the data suggest a relatively constant rate of exit when each subsequent cohort is compared with the 1990 group. However, adoption exits were more likely and reunification was less likely (on a unit time basis). According to the data in Figure 1, adoptions of children admitted in 1997 were 2.7 times more likely per unit time than was true for the children admitted in 1990, after controlling for the characteristics of the children. Further, the data indicate that steady progress was made from 1991 forward through 1997, but that progress relative to 1990 was diminished somewhat in 1998 and 1999, at least when compared with 1997.

Figure 1 Relative Rate of Exit by Exit Destination and Year of Admission AL, IL, MD, MI, MO, NJ, NY, OH, and WI

Figure 1. Relative Rate of Exit by Exit Destination and Year of Admission.

The data in Figure 1 and Table 2 together offer the most complete explanation of adoption dynamics relative to ASFA. It appears from Figure 1 that children admitted in 1991 had a slightly elevated chance of adoption overall when compared with 1990. The data from Table 2 suggest that this slight advantage is likely attributable to a small shift in the period specific probabilities observed for the 1991 cohort in the two post-ASFA years for which data are available. Specifically, the conditional probabilities for the 1990 and 1991 cohorts are nearly identical through the sixth year. Thereafter, for one year prior to the ASFA years and for the first ASFA year, the 1991 children were more likely to exit to adoption than were the 1990 children. A similar pattern describes the dynamics underlying the improved performance observed for the 1992, 1993, and 1994 cohorts. For example, in 1994, the probability of adoption during the first three years is identical to the observed probabilities that defined the 1990 cohort. In the fourth year, just before the ASFA effect should be observed, the conditional probability stands at .14 for the 1994 cohort and just .10 for the 1990 group. Finally, the data in Figure 1 further suggest that a decline in exit rates involving children admitted to foster care since ASFA become law (the 1998 and 1999 cohorts) may be underway. Relative to 1990, the children admitted in 1998 and 1999 were moving through the adoption process somewhat faster. However, relative to 1996 and 1997, the rate appears to have slowed somewhat. Although there have been relatively few adoptions involving children admitted to foster since ASFA was passed, the trend bears watching.

Summary and Discussion

These findings suggest several conclusions about the adoption process between 1990 and 1999. However, it is too soon to determine whether the probability of adoption has increased because of the Adoption and Safe Families Act. This analysis accounts for approximately 32,000 adoptions that were completed in either 1998 or 1999, just after ASFA became law. Of those adoptions, only 6 percent involved children who were admitted to foster care after 1997. Until more time elapses, it will be difficult to ascertain whether the likelihood of adoption has shifted in the direction anticipated by lawmakers. Firmer conclusions in this regard will not emerge until 2003 or 2004, after data from several post-ASFA cohorts can be studied.

Of the remaining adoptions that were tracked (94%), all the children involved were admitted to foster care before ASFA was passed. Understanding ASFAs impact requires an approach that takes into account other historical periods and multiple waves of foster care admissions. The specific findings do point to an increase in the rate of adoption per unit time. For children admitted between 1991 and 1993, there appears to have been both pre-ASFA and post-ASFA effects contributing to adoption speed. The pre-ASFA effects were identified for children who were adopted after 4 years or more in care, but before ASFA was passed. Among children who were still in care after ASFA passed, their quicker exit from the system suggests that ASFA did have the intended impact on the adoption backlog. The data also suggest that changes early in the adoption process started with cohorts admitted after 1994. Some of these changes were pre-ASFA effects; others were post-ASFA effects.

Finally, the data also point to a possible slow down in the two post-ASFA cohorts. These findings are preliminary, but warrant further study in light of specific ASFA provisions. The presence of both pre- and post-ASFA effects invites questions about what other factors might account for the pre-ASFA increase in the speed of adoptions. The strongest pre-ASFA effects were observed in the mid-1990s and involved cohorts admitted during that period as well as children who were still in care at that time having been admitted earlier in the decade. Two explanations come to mind. First, the findings are based on data from states that have very different adoption dynamics (Wulczyn, Brunner, & Goerge, 2000). Individual state differences are an underlying source of variation that might account for the observed pre-ASFA effect. New York and Illinois, in particular, are two states that addressed adoption issues in advance of ASFA.

Second, ASFA was but one of several federal attempts during the 1990s to change adoption practice. The most notable federal legislation affecting adoptions other than ASFA is the Multi-ethnic Placement Act of 1994 (MEPA) and its subsequent revision in 1996. It is not possible to conclude from these analyses that MEPA accounts for the observed pre-ASFA effects because race-specific changes in the time to adoption were not examined.

ASFAs provisions most directly affect adoption outcomes. However, impacts on one type of foster care exit may affect other types of exit as well. This research was designed to ascertain whether ASFA had the intended effect on the adoption process. Children who were in care when ASFA was passed did appear to move to adoption more quickly, especially those children who may have been part of the adoption backlog. But the data also point to a slow down in the reunification process (see Figure 1). Although it cannot be said from these data that reunification was less likely, given the broad goals and objectives of ASFA specifically and public policy more generally, a slower rate of reunification would be an unintended consequence. If reunification of children has been adversely affected by efforts to promote more timely adoptions, the systemic factors that influence that trade-off will have to be confronted.

Appendix I:

Multivariate Results Presented in Figure 1 Results from Cox Proportional Hazard Models (Risk Ratios) (AL, IL, MD, MI, MO, NJ, NY, OH, and WI)
Characteristic All exits Reunification Adoption
Year of admission  
1990 1.000 1.000 1.000
1991 1.003* 0.965 1.056
1992 1.01* 0.946 1.162
1993 1.022 0.937 1.259
1994 1.013 0.899 1.354
1995 1.036 0.906 1.547
1996 1.04 0.893 1.841
1997 1.055 0.867 2.532
1998 1.013 0.852 1.652
1999 0.997* 0.842 1.155
Spell sequence  
First spell 1.000 1.000 1.000
Second or higher 1.048 0.783 1.037
Balance of state 1.000 1.000 1.000
Primary urban area 0.739 0.641 0.633
Caucasian 1.000 1.000 1.000
African American 0.897 0.792 0.638
Hispanic 1.084 1.034 0.689
Other 1.028 0.942 0.75
Age at first admission  
Under one 1.000 1.000 1.000
age 1-5 1.104 1.574 0.503
age 6-12 1.11 1.579 0.248
age 13-17 1.664 1.792 0.08
Type of placement  
Foster boarding home 1.000 1.000 1.000
Congregate care 1.28 1.225 0.211
Kinship care 0.708 0.735 0.595
Mixed care types 0.987* 0.788 0.224
* Not statistically significant (p > .05). Shaded category is comparison group. Separate results (coefficients) for states not displayed.


1.  The specific focus is on public agency adoptions. Private adoptions and international adoptions are not included in the analysis.

2.  These states account for approximately 55% of the children in foster care nationwide. Although the states in the Archive are diverse, they are not necessarily representative of the states not included. Thus, a more representative sample of states might yield slightly different results.

3.  It is important to stress that the comparisons made in Figure 1 are over time by exit type and not within year by exit type. For example, the graph indicates that the adoptions happened much faster per unit time for children admitted in 1997 when compared with what happened to the children admitted in 1990. The graph should not be interpreted to mean that in 1997, adoptions happened much faster than reunification.