Employment Outcomes for Youth Aging Out of Foster Care
Final Report
Robert M. Goerge, Principal Investigator Lucy Bilaver, Bong Joo Lee Chapin Hall Center for Children at the University of Chicago Barbara Needell, Alan Brookhart, William Jackman Center for Social Services Research, University of California Berkeley
March, 2002
University of Chicago Chapin Hall Center for Children
This report is available on the Internet at:http://aspe.hhs.gov/hsp/fostercare-agingout02/
Contents
- Acknowledgements
- Executive Summary
- Introduction
- Study Population
- Data and Methods
- Results
- Summary of results
- Future research
- References
Executive Summary
Purpose and Background
There is a widespread belief that young people who age out of foster care near the time that they turn 18 are particularly vulnerable to poor economic and social outcomes as they enter adulthood. Over the past few years, significantly more attention has been paid to youth aging out of foster care and more concern expressed for their future prospects. The 1999 Foster Care Independence Act provides fiscal incentives to states for enhanced services to these youth. In addition, the Act requires states to evaluate their services to this population of young people, and has provided additional resources to do so.
The purpose of this report is to provide information on the employment outcomes of children exiting foster care near their eighteenth birthdays in California, Illinois, and South Carolina during the mid-1990s. We report when they begin to have earnings, in how many quarters over a 13-quarter time period they had earned income, and the amount of earned income they received over that time period. We compare these outcomes to those for youth who were reunified with their parents prior to their eighteenth birthday and to low-income youth.
This report addresses the following three primary research questions:
- What are the patterns of employment and the amount of earnings of youth aging out of foster care?
- How do these employment patterns compare with those of other youth of similar ages in low-income families?
- What are the sociodemographic characteristics and foster care service experiences that are related to the patterns of employment?
Summary of Findings
Youth aging out of foster care are underemployed. No more than 45 percent of the aging out youth have earnings in any of the three states during any one of the 13 quarters of the study. This is also the case for reunified youth. A slightly larger proportion of low-income youth has earnings, but never more than 50 percent.
Patterns of unemployment vary by state. About 30 percent of youth aging out of foster care in Illinois, 23 percent in California, and 14 percent in South Carolina had no earnings during the entire 13-quarter period.
Youth who do work begin to do so early. In all three states, youth were more likely to earn income for the first time during the four quarters prior to and the quarter of their eighteenth birthday than in the 2 years following. For youth who exited foster care by aging out, half in California and Illinois and two-thirds in South Carolina had earnings prior to their eighteenth birthday. Although the aging out group is more likely to work than the reunified group in South Carolina and California, there is no difference between the two groups in Illinois. In California and South Carolina, if youth did not work prior to exit, there was slightly more than a 50-50 chance that they would begin employment after exit. In Illinois, youth who did not have earnings prior to their eighteenth birthday had less than a 50 percent chance of beginning to work by the age of 20.
Youth aging out of foster care have mean earnings below the poverty level. Youth aging out of foster care earn significantly less than youth in any of the comparison groups both prior to and after their eighteenth birthday. Average quarterly earnings do grow significantly from the 4 quarters prior to the eighteenth birthdays to the 8 quarters after it. In each state, the average earnings increases roughly $500 per quarter. However, even with these increases, these youth average less than $6,000 per year in wages, which is substantially below the 1997 poverty level of $7,890 for a single individual.
Youth aging out of foster care progress more slowly in the labor market than other youth. In Illinois, low-income youth make a bigger increase in earnings from the first year to the second year after their eighteenth birthday than do either group of foster care youth. Low-income and aging-out youth in California see a larger increase in their earnings than reunified youth. There is no difference among the groups in South Carolina.
[Go To Contents]
Introduction
Purpose
The purpose of this report is to provide information on the employment outcomes of foster children exiting foster care at or around their eighteenth birthday in California, Illinois, and South Carolina. We report when they begin to have earnings, in how many quarters over a 13-quarter time period they had earned income, and the amount of their earned income over that time period. An important feature of this study is that we compare the results for youth aging out of foster care to youth that were reunified with their parents prior to their eighteenth birthday and to low-income youth.
This research is timely because of an increased level of attention to the well-being of this population of youth. The Foster Care Independence Act of 1999 (The Chaffee Act) provides incentives to states for enhanced services to these youth. This study provides a baseline against which the experiences of youth in the future and in other states can be compared.
Background and Policy Issues
Children are placed into the foster care system primarily because of abuse, neglect, uncontrollable behavior, or dependency. Foster care is intended to be a temporary service with a goal of reunifying children with their parents whenever possible. However, many children cannot be reunified, primarily because the courts and the child protective services system determine that they would be at continued risk of abuse or neglect or because their parents are simply not able to care for them. For an increasing proportion of these children, adoption or subsidized guardianship with kin are alternatives to reunification (Wulczyn, Brunner, and Goerge, 2000).
Unfortunately, a small proportion of children who enter the foster care system do not achieve a permanent status with their parents, kin, or adoptive parents and live in foster care until they reach the age of majority (18 years of age in all but a few states). These youth about 20,000 per year in the United States stay in foster care until they are emancipated after their eighteenth birthday. Although some of these youth return to their families after emancipation, many are completely without support from means other than government programs. In some states youth who are still in school may still receive room and board; however, this applies to a minority of youth.
Therefore, despite some additional support mandated by the Foster Care Independence Act, at the age of 18 emancipated foster children must seek independent means of support. Those youth who were employed prior to emancipation have some advantage due to their work experience and perhaps some savings. Those who are not working at the time of their emancipation must compete in a labor market that includes youth who have not had the disadvantage of being dependent on services designed to be temporary in nature and, until recently, not designed to be of direct benefit after leaving foster care.
The current policy situation
Independent Living Program and the Chaffee Bill
Prior to the passage of the Foster Care Independence Act in 1999, the Independent Living Program provided for services to youth until their eighteenth birthday. There were no special funds for youth transitioning out of foster care and states were not required to spend a portion of their funds on youth ages 18-21. With the passage of the Chaffee Act, the federal government effectively provided increased funding for most states Independent Living Programs, by requiring a 20 percent state match instead of no match for the first $45 million from the federal government and a 50 percent match on additional funds, which were previously not available. Of interest for this study, the law provides federal funds for states to provide services to ex-foster care youth ages 18-21, regardless of Title IV-E eligibility, for purposes of obtaining a high school diploma, career exploration, vocational training, and job placement and retention.
The benefits of the program offer the possibility of covering room and board, post-secondary educational assistance, and Medicaid coverage for these youth. From an employment perspective, these additional independent living program benefits will supplement the earned income that is usually inadequate to meet the financial needs of youth who are not being assisted by their families.
Evaluation at the Federal and State Levels
The new law also requires that the federal government engage in evaluation, technical assistance, performance management, and data collection. However, because these activities started after the passage of the law, there is little information on what happened to youth prior to the new program. Employment issues are explicitly discussed in the legislation and states are likely in the future to collect information on how well youth aging out of foster care do in the labor market. This study demonstrates one method of analyzing these outcomes using existing data sources.
Questions addressed in this report
This report addresses the following three primary research questions: What are the patterns of employment and the amount of earnings of youth aging out of foster care? Specifically, we analyze the likelihood of youth having earnings both prior to and after their eighteenth birthday, the amount of earnings during this period, and the change in earnings from the first to the second year after their eighteenth birthday.
How do these employment patterns compare with those of other youth of similar ages in low-income families? We compare these youth with similar populations of reunified youth and youth that were part of Aid to Families with Dependent Children and Temporary Assistance to Needy Families (AFDC/TANF) grants (our study period spans the transition between the AFDC and TANF programs). Comparing foster children to children who have been part of AFDC/TANF grants is a reasonable strategy because a large percentage of foster children come from poor families and the demographic profiles are often quite similar (U.S. DHHS, 2000 (1); U.S. DHHS, 2000 (2)).
What are the sociodemographic characteristics and foster care service experiences that are related to the patterns of employment? We examine the effect of race/ethnicity, gender, age at first placement (or AFDC/TANF entry), major urban region(s) (Cook County in Illinois, LA County in California and the MSA counties in South Carolina) versus balance of the state, type of placement, time in most recent episode of service, and the reason for foster care placement on the likelihood of having earnings and the amount of earnings.
Why these three states?
We chose these three states primarily because of the availability of longitudinal administrative data on foster children and AFDC/TANF recipients, and the availability of wage reporting data. At the outset of the project, we explored the participation of over a dozen states where, as a result of our work on the Multi-State Foster Care Data Archive, we knew that foster care data was available. We also were aware of other states that have AFDC/TANF and wage reporting data. However, when we pursued whether or not the link between the three data sources could be made and the data analyzed in a timely manner, we were left with only California, Illinois, and South Carolina.
Previous Research
Two recent reviews of research on the well being of youth aging out of foster care state that much of the work has been on a small scale and not of a rigorous nature (GAO, 1999; Collins, 2001). However, the results show that youth aging out of foster care are generally ill prepared for self-sufficiency. A few of the studies stand out. A national evaluation by Westat (1991) found that a large percentage of youth aging out of foster care (46%) did so without a high school diploma, and 40 percent were dependent on the community through income assistance or Medicaid 2.5 to 4 years after leaving foster care. Researchers found that these youth were very similar to poor youth when compared to national census data. However, this study was based on a 50 percent response rate, which suggests that many of the youth whose outcomes were poorer may not have been found.
Courtney, et al. (1998) had greater success (a response rate of 83%) in finding Wisconsin youth 18 months after leaving foster care using state administrative data. They found that over 80% of the sample members report they have been employed at some time, with 57% stating they currently hold a job. They also found that 37 percent had not finished high school, 39 percent were unemployed, and 32 percent were receiving public assistance.
McMillen and Tucker (1999) found in Missouri that almost half of young people leaving care (45%) exit without jobs or a high school education, although many (64%) are considered to be making academic progress.
A recent analysis by Wulczyn and Hislop (2001) suggests that youth who are in foster care at the age of 16 do not really conform to the commonly held view that these youth have grown up in foster care and as a result are ill prepared for the transition into adulthood. They find that there are basically two types of youth in care at age 16. One group is composed of those teens who enter foster care close to their sixteenth birthday and exit within the next 12 to 18 months (before they turn 18) and the second, smaller group reaches the age of majority after a considerable period of time in care. This analysis does not suggest that youth who transition through foster care are any better prepared for independence than are those who spend a long period in care, but it does suggest how programs for these youth may be better planned and provided.
A recent study by Dworsky and Courtney (2000) tracked the employment and public assistance utilization of a cohort of youth in Wisconsin very similar to the cohorts in the three states of this study. Because the employment analyses were very similar to those done in this study, we discuss those results in combination with the results from this study. Regarding public assistance, they found that only a small minority of former foster youth had received AFDC/TANF cash assistance and/or Food Stamps at any time during the first 8 quarters after they were discharged from care. However, they found that there were significant race and regional effects, with African American youth and youth from Milwaukee being more likely to use AFDC/TANF or Food Stamps.
[Go To Contents]
Study Population
We define the study group population (the aging-out group) as youth who were emancipated from all types of foster care (aged out) and reached their eighteenth birthday during the study period. We use the following two groups as the comparison groups:
- Youth who were in all types of foster care and reunified with their families at any time after their fourteenth birthday and reached their eighteenth birthday during the study period (reunification group).
- Youth who exited an AFDC/TANF case at any time after their fourteenth birthday and reached their eighteenth birthday during the study period (low-income group).
The study period is 1996-1997 in Illinois and South Carolina, and 1995-1996 in California. These years were chosen because of data availability in each of the states (see discussion below). Exhibit 1 below summarizes the definition of each group and the size of each group in each state.
Group | Description | CA | IL | SC |
Aging Out group | youth who turned 18 during the study period and were emancipated from foster care in the year in which they turned 18 | 2,824 | 1,084 | 305 |
Reunification group | youth who were reunified at any time after their 14th birthday and before their 18th birthday and reached their 18th birthday in the study period | 3,138 | 1,504 | 773 |
Low-Income group | youth who were part of AFDC or TANF case after their 14th birthday and before their 18th birthday and reached their 18th birthday in the study period | 186,637 | 49,194 | 11,464 |
[Go To Contents]
Data and Methods
Data Description
Foster care data in three states
In each of the three states, we used the child welfare information systems to select the study populations who had aged out of foster care and who were reunified from foster care. The data that we used on each individual were those that were generally available across the three states. In particular, data on type of placement and reason for placement are not directly comparable across the states (and sometimes not within states because sub-state level field offices may be using different practices in recording information). However, since we did not combine data across the three states and only conduct within-state statistical analyses, we used these additional variables in our analyses. They provide important descriptors of the ways in which the states differ from one another.
Wage reporting data in three states
In each state, we accessed the Unemployment Insurance (UI) Wage Reporting data for each young person in the study. The federal government requires each state to collect this data, and it is collected in a uniform way across the three states. These data provide quarterly earnings for each job included in the UI system. These data cover most types of jobs, but exclude, most notably, federal and railroad jobs and personal services or consulting jobs (independent contractors), where the employer is not paying Unemployment Insurance (Scholz and Hotz, 1999). Thus, it is likely that some youth have jobs for which there is no UI wage record reported by the employer. The employment rate for youth has been shown to be as much as 10 percent less using UI data than when using survey data, with the greatest differences being for male youth (Kornfeld and Bloom, 1999).
AFDC/TANF data in three states
The data on the AFDC/TANF youth in this study come from the income maintenance program eligibility and tracking systems in each of the states. While most of the youth studied would have been AFDC recipients, it is possible that some youth would have been on TANF for a short period of time at the end of 1997 in South Carolina and Illinois when AFDC became TANF.
Linking of these three files in each of the states
In each of the three states, the three study populations were linked to their UI data through Social Security Numbers (SSNs) that were part of the childs AFDC/TANF, child welfare, and UI records. We attempted to use the same procedure to link in each of the three states to assure the greatest comparability. However, there were different percentages of missing SSNs in the foster care populations in each of the three states. In Illinois, 10.5 percent of the SSNs were missing, in South Carolina, 11.5 percent and in California, 19 percent. This very well could be a source of bias in the results if, for example, the youth with missing SSNs came from a particular geographic region.
This study has several important limitations. These include limitations inherent in the choice of study population, data sources, differences in how data is reported among the different states studied, and the fact that we have at our disposal limited variables.
The study populations examined in each state are select populations in that we have chosen to include only youth under age 18 although some stay in foster care longer. We made our choice of study population definition because of wage record data availability and our belief that our choice of study population is the most comparable across states. However, given the limitations of administrative data, we are unable to specify why some youth exit just after their eighteenth year and others stay in foster care longer. For example, we exclude youth who stay in foster care beyond their nineteenth birthday and we know very little about why each individual youth does stay beyond their nineteenth birthday. Indeed, in Illinois, we found that 18 year olds are not that much different than 19 year olds in their employment outcomes.
A second limitation is that unemployment insurance wage data includes information on most, but not all, employment. Information about informal and off the books employment is not captured, nor is military employment or employment out of state. These limitations may have caused us to underestimate employment somewhat. Methodological work by Kornfeld and Bloom (1999) finds that when compared with employment data collected through surveys of individuals (which have their own limitations) unemployment insurance data may substantially underestimate the amount of earnings, especially for youth with prior arrests. In comparison with survey data, unemployment insurance wage data usually produces estimates that are lower by about 10 14 percent, but with youth the discrepancy may be as high as 30 50 percent for some sub-populations (Hotz and Scholz, 2002). Discrepancies are less for employment rates and for employment of adults. While it is likely that our findings undercount employment, our earnings estimates for youth are so low that taking potential underestimates into account would not change our conclusions. In addition, our findings are generally in line with research on former foster care youth using survey methodologies (e.g. Courtney et al, 1998). Out of state employment is less likely to be a problem; current research tracking former forster care youth in Wisconsin is finding very little out of state mobility in this population (Mark Courtney, personal communication, February 21, 2002).
A third limitation is that the variables that are available to us across the three states are collected in different ways due to differences in state policies. Therefore, many of the differences across states may be due to how youth are classified in the administrative data, as well as due to the effects of state policies. It is difficult to disentangle these potential explanations. This limitation is the primary reason why we cannot make strong evaluative statements about youth doing better in one state or another.
A final limitation is the lack of information on characteristics of the youth that are not available in these data sources. The most obvious omission is their educational status. Knowing how many of these youth are still in school would allow us to better interpret the earnings information. Other data that would be useful would be data on which of the youth are parents receiving TANF and which of the youth may have been incarcerated.
[Go To Contents]
Results
Descriptive statistics of the study groups
Exhibit 1 above shows the size of each group in each state. In general, they represent the relative size of each states total youth population. However, South Carolina has roughly three times the number of children in the foster care groups proportionately relative to the low-income group as California, and twice that of Illinois. A youth in South Carolina is more likely to be in the aging out group than a youth in Illinois, who is more likely to be in the aging out group than a youth in California. In South Carolina, aging out youth are demographically more similar to AFDC/TANF youth than in the other two states.
The racial composition of the aging-out groups in the three states is different (Exhibit 2). Although in general they reflect the demographics of the state, African American youth are overrepresented and white youth are underrepresented. Hispanic youth are also underrepresented in Illinois and California, but not as much as white youth.
Sample Characteristics for AGE OUT | Sample Characteristics for REUN | Sample Characteristics AFDC | |||||||
Illinois | S.Carolina | California | Illinois | S.Carolina | California | Illinois | S.Carolina | California | |
Race | |||||||||
African-American | 56.8% | 58.7% | 32.2% | 38.9% | 41.5% | 24.6% | 56.1% | 79.8% | 18.5% |
Hispanic | 4.4% | N/A | 20.5% | 4.4% | N/A | 25.7% | 12.2% | 0.4% | 37.3% |
White | 37.0% | 40.3% | 42.7% | 55.5% | 56.5% | 45.2% | 30.2% | 18.8% | 31.1% |
Other | 1.8% | 1.0% | 4.3% | 1.2% | 1.9% | 4.5% | 1.5% | 1.0% | 12.7% |
Sex | |||||||||
Female | 48.7% | 58.7% | 60.8% | 53.9% | 67.7% | 63.4% | 50.5% | 52.2% | 51.8% |
Male | 51.3% | 41.3% | 39.2% | 46.1% | 32.2% | 36.6% | 49.5% | 47.8% | 48.2% |
Unknown | 0.0% | 0.0% | 0.0% | 0.0% | 0.1% | 0.0% | 0.0% | 0.0% | 0.0% |
Age | |||||||||
0 yrs | 2.8% | 0.0% | 0.7% | 1.5% | 0.0% | 0.1% | 0.0% | 0.0% | 0.0% |
1-5 yrs | 11.5% | 1.3% | 8.6% | 7.0% | 1.7% | 2.1% | 0.0% | 0.0% | 0.0% |
6-10 yrs | 20.0% | 4.6% | 33.8% | 12.5% | 6.6% | 16.1% | 60.9% | 47.3% | 42.0% |
11-15 yrs | 53.0% | 68.9% | 42.6% | 65.8% | 66.8% | 65.5% | 32.1% | 38.1% | 39.5% |
16+ yrs | 12.7% | 25.2% | 14.2% | 13.3% | 25.0% | 16.2% | 7.0% | 14.6% | 18.5% |
Region | |||||||||
Rural | 48.4% | 45.2% | 77.1% | 66.8% | 29.2% | 70.4% | 37.4% | 43.9% | 70.2% |
Urban* | 51.6% | 54.8% | 22.9% | 33.2% | 70.8% | 29.6% | 62.6% | 56.1% | 29.8% |
Type of Placement | |||||||||
HMR | 15.5% | 57.0% | 31.0% | 22.3% | 61.6% | 27.9% | N/A | N/A | N/A |
Trad. FC | 4.4% | 20.3% | 48.4% | 23.8% | 18.0% | 38.4% | N/A | N/A | N/A |
Inst | 6.4% | 12.1% | 16.6% | 30.9% | 15.9% | 23.3% | N/A | N/A | N/A |
Other | 73.7% | 10.5% | 4.0% | 23.0% | 4.5% | 10.3% | N/A | N/A | N/A |
Time in Most Recent Spell | |||||||||
1 or 1yr | 16.9% | 23.3% | 32.9% | 67.0% | 26.5% | 78.7% | 45.2% | 19.2% | 47.0% |
2-4 yrs | 47.0% | 13.1% | 30.7% | 23.5% | 14.6% | 14.0% | 24.2% | 33.9% | 23.1% |
5-10 yrs | 31.0% | 10.8% | 29.1% | 8.6% | 13.2% | 6.9% | 30.6% | 39.9% | 29.9% |
Over 10 yrs | 5.2% | 52.8% | 7.4% | 0.9% | 45.7% | 0.4% | 0.0% | 7.0% | 0.0% |
Reason for Foster Care Placement | |||||||||
Abuse | 12.5% | 36.7% | 28.7% | 12.8% | 36.7% | 34.5% | N/A | N/A | N/A |
Neglect | 41.7% | 36.4% | 61.1% | 32.5% | 36.4% | 48.8% | N/A | N/A | N/A |
Parent Child Problems | 7.1% | 3.6% | 0.0% | 13.4% | 3.6% | 0.0% | N/A | N/A | N/A |
Other | 38.7% | 23.3% | 7.7% | 41.3% | 23.3% | 15.9% | N/A | N/A | N/A |
Household Type | |||||||||
Single parent | N/A | N/A | N/A | N/A | N/A | N/A | 95.0% | 51.7% | N/A |
2 parent | N/A | N/A | N/A | N/A | N/A | N/A | 5.0% | 48.2% | N/A |
* Urban for Illinois is Cook County, for California, it is L.A. County and for South Carolina, it is the central and outlying counties of MSAs as defined by the Census Bureau. N/A Data not available. |
In all three states, white children represent the greatest portion of the reunified group of youth. In Illinois, the race distribution of the AFDC/TANF and aging-out groups are very similar. In South Carolina, almost 80 percent of the AFDC/TANF group was African American. In California, almost 40 percent of the AFDC/TANF group is Hispanic compared with 22 and 27 percent of the aging-out and reunification groups, respectively.
With respect to gender, California youth in the foster care groups are disproportionately female. The same was true of the reunified group in South Carolina but not the aging-out group. In Illinois, youth in the reunified group are also more likely to be female, but the aging-out group has slightly more males.
The distributions of age at initial placement of the aging out and reunification groups are quite different within California and Illinois. In California, for example, 42.6 percent of the aging out group entered care between the ages of 11 and 15 compared with 65.5 percent of the reunified group. In South Carolina, the age distributions of the two groups are very similar.
The type of out-of home-care placement that these youth exited from are vastly different across the states. In Illinois, nearly 74 percent of the aging-out youth were last served in living arrangements other than foster homes, kinship care, and institutions primarily independent living. The reunified group on the other hand is fairly evenly distributed across these four living arrangement categories. In both South Carolina and California, it was not an option for youth of this age to be in independent living arrangements. In South Carolina, the two foster care groups look much more similar with respect to their last type of living arrangement the vast majority was placed with relative foster parents. In California, there is a different pattern. Most often, children exited from traditional non-relative foster care homes. Slightly fewer were exiting from placements with relatives. Institutions and group homes were more commonly used by the reunified youth than by the aging-out youth (22.3% vs. 15.5%).
Examining the time spent in the most recent foster care spell reveals another important difference among the states and among the study groups. The aging-out groups tended to have been in out of home care longer than the reunified groups. In Illinois, this group had the longest length of stay (83.1% in the placement for more than 2 years) prior to discharge, followed by California (67.1%) and South Carolina (51.5%) respectively. In all three states, children in the reunified group tended to have very short stays in care. More than 60 percent of the children in the reunified groups had been in their most recent foster care spells less than 2 years.
With regard to the reason for foster care placement, neglect was the principal reason for foster care placement of the aging-out groups in Illinois (41.7%) and California (61.1%). The youth in the aging-out group in South Carolina were equally divided among neglect, abuse, or other reason categories. The other reason for placement includes children who enter the foster care system for dependency reasons a set of circumstances that are not maltreatment that prevent a child and parent from living together. This could include a child being an orphan or a childs parent being in jail or prison.
Descriptive Results
When youth first become employed relative to their foster care experience has implications for how child welfare agencies organize the provision of services to these youth. If employment prior to their eighteenth birthday were important for a childs post-foster care employment, providing youth with some kind of employment experience prior to exit might be a priority. We used data on when a youth had earnings in the period beginning 4 quarters prior to their eighteenth birthday up to 8 quarters after (13 quarters, including the quarter of their eighteenth birthday) to determine when he or she first worked relative to his or her foster care experience.
Youth with no income during the study period
About 30 percent of youth aging out in Illinois, 23 percent in California and 14 percent in South Carolina had no earnings during the entire 13-quarter period (Exhibits 3a-c, top panel). The value of comparison groups is that they allow us to determine whether these state-level differences reflect actual differences in the characteristics of aging-out youth, or differences across states in terms of employment opportunities. In Illinois, the aging-out group included a higher percentage of youth who had no income during the 13 quarters than the reunified group and the AFDC/TANF group. In both South Carolina and California, more of the aging-out group had earnings during the 13 quarters than either of the comparison groups.
Differences between the aging-out groups and AFDC/TANF groups in California, across racial, regional, and gender categories, were typically larger than those in Illinois and in South Carolina. In California, aging-out youth did much better than AFDC/TANF or reunified youth. In South Carolina, there were few differences between aging-out youth and AFDC/TANF youth. Both of these groups were more likely to be employed than reunified youth. In Illinois, a different pattern emerged, with reunified youth doing better than AFDC/TANF or aging-out youth.
Earnings prior to their eighteenth birthday
For youth who exited foster care through aging out, half in California and Illinois and two-thirds in South Carolina had earnings prior to their eighteenth birthday. African American youth were less likely than white youth to be employed prior to their eighteenth birthday in all three states. In California, the likelihood of employment for Hispanic aging-out youth was similar to that of white youth, while Hispanic youth in Illinois were more likely to be employed than African American youth and less likely than whites. Female youth and youth from non-urban areas were generally more likely to be employed prior to their eighteenth birthday than males or youth from the primary urban areas of each state, although this might be a function of different types of available jobs for males and females and the possibility that more females were captured in the data that we used.
In comparison to youth who were reunified after foster care and youth from AFDC/TANF cases, aging-out youth in South Carolina and California were more likely to be employed prior to exit. In Illinois, aging-out youth were less likely to be employed prior to exit. The statewide pattern was generally the case when looking at differences by race, gender, or region. For example, Hispanic aging-out youth in California were more likely to be employed prior to exit than reunified youth or youth exiting from AFDC/TANF cases.
Characteristics | Study Population | No Earnings | 4 quarters prior and quarter of 18th birthday | 8 quarters after 18th birthday | Total | No Earnings | 4 quarters prior and quarter of 18th birthday | 8 quarters after 18th birthday | Total |
Counts | Percentage | ||||||||
Total | AGE OUT | 316 | 547 | 221 | 1,084 | 29.2% | 50.5% | 20.4% | 100.0% |
REUN | 301 | 912 | 291 | 1,504 | 20.0% | 60.6% | 19.3% | 100.0% | |
AFDC | 12,444 | 26,169 | 10,581 | 49,194 | 25.3% | 53.2% | 21.5% | 100.0% | |
Race | |||||||||
African-American | AGE OUT | 252 | 231 | 133 | 616 | 40.9% | 37.5% | 21.6% | 100.0% |
REUN | 173 | 279 | 133 | 585 | 29.6% | 47.7% | 22.7% | 100.0% | |
AFDC | 8,266 | 12,826 | 6,498 | 27,590 | 30.0% | 46.5% | 23.6% | 100.0% | |
Hispanic | AGE OUT | 12 | 21 | 15 | 48 | 25.0% | 43.8% | 31.3% | 100.0% |
REUN | 10 | 40 | 16 | 66 | 15.2% | 60.6% | 24.2% | 100.0% | |
AFDC | 1,249 | 3,480 | 1,261 | 5,990 | 20.9% | 58.1% | 21.1% | 100.0% | |
White | AGE OUT | 48 | 284 | 69 | 401 | 12.0% | 70.8% | 17.2% | 100.0% |
REUN | 110 | 584 | 141 | 835 | 13.2% | 69.9% | 16.9% | 100.0% | |
AFDC | 2,698 | 9,495 | 2,665 | 14,858 | 18.2% | 63.9% | 17.9% | 100.0% | |
Other | AGE OUT | 4 | 11 | 4 | 19 | 21.1% | 57.9% | 21.1% | 100.0% |
REUN | 8 | 9 | 1 | 18 | 44.4% | 50.0% | 5.6% | 100.0% | |
AFDC | 231 | 368 | 157 | 756 | 30.6% | 48.7% | 20.8% | 100.0% | |
Sex | |||||||||
Female | AGE OUT | 117 | 299 | 112 | 528 | 22.2% | 56.6% | 21.2% | 100.0% |
REUN | 142 | 530 | 139 | 811 | 17.5% | 65.4% | 17.1% | 100.0% | |
AFDC | 5,269 | 14,322 | 5,249 | 4,840 | 21.2% | 57.7% | 21.1% | 100.0% | |
Male | AGE OUT | 199 | 248 | 109 | 556 | 35.8% | 44.6% | 19.6% | 100.0% |
REUN | 159 | 382 | 152 | 693 | 22.9% | 55.1% | 21.9% | 100.0% | |
AFDC | 7,175 | 11,847 | 5,332 | 24,354 | 29.5% | 48.6% | 21.9% | 100.0% | |
Region | |||||||||
Cook County | AGE OUT | 232 | 197 | 130 | 559 | 41.5% | 35.2% | 23.3% | 100.0% |
REUN | 136 | 253 | 111 | 500 | 27.2% | 50.6% | 22.2% | 100.0% | |
AFDC | 8,678 | 14,960 | 7,165 | 30,803 | 28.2% | 48.6% | 23.3% | 100.0% | |
Rural | AGE OUT | 84 | 350 | 91 | 525 | 16.0% | 66.7% | 17.3% | 100.0% |
REUN | 165 | 659 | 180 | 1,004 | 16.4% | 65.6% | 17.9% | 100.0% | |
AFDC | 3,766 | 11,209 | 3,416 | 8,391 | 20.5% | 60.9% | 18.6% | 100.0% |
Characteristics | Study Population | No Earnings | 4 quarters prior and quarter of 18th birthday | 8 quarters after 18th birthday | Total | No Earnings | 4 quarters prior and quarter of 18th birthday | 8 quarters after 18th birthday | Total |
Counts | Percentage | ||||||||
Total | AGE OUT | 44 | 203 | 58 | 305 | 14.4% | 66.6% | 19.0% | 100.0% |
REUN | 163 | 451 | 159 | 773 | 21.1% | 58.3% | 20.6% | 100.0% | |
AFDC | 1,935 | 6,467 | 3,062 | 11,464 | 16.9% | 56.4% | 26.7% | 100.0% | |
Race | |||||||||
African-American | AGE OUT | 29 | 114 | 36 | 179 | 16.2% | 63.7% | 20.1% | 100.0% |
REUN | 71 | 164 | 86 | 321 | 22.1% | 51.1% | 26.8% | 100.0% | |
AFDC | 1,549 | 5,009 | 2,592 | 9,150 | 16.9% | 54.7% | 28.3% | 100.0% | |
Hispanic | AGE OUT | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |
REUN | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | |
AFDC | 9 | 28 | 6 | 43 | 20.9% | 65.1% | 14.0% | 100.0% | |
White | AGE OUT | 15 | 88 | 20 | 123 | 12.2% | 71.5% | 16.3% | 100.0% |
REUN | 86 | 279 | 72 | 437 | 19.7% | 63.8% | 16.5% | 100.0% | |
AFDC | 339 | 1,371 | 445 | 2,155 | 15.7% | 63.6% | 20.6% | 100.0% | |
Other | AGE OUT | 0 | 1 | 2 | 3 | 0.0% | 33.3% | 66.7% | 100.0% |
REUN | 6 | 8 | 1 | 15 | 40.0% | 53.3% | 6.7% | 100.0% | |
AFDC | 38 | 59 | 19 | 116 | 32.8% | 50.9% | 16.4% | 100.0% | |
Sex | |||||||||
Female | AGE OUT | 21 | 124 | 34 | 179 | 14.1% | 67.2% | 18.8% | 100.0% |
REUN | 106 | 314 | 103 | 23 | 28.6% | 54.0% | 17.3% | 100.0% | |
AFDC | 882 | 3,487 | 1,614 | 5,983 | 14.7% | 58.3% | 27.0% | 100.0% | |
Male | AGE OUT | 23 | 79 | 24 | 126 | 19.9% | 62.5% | 17.6% | 100.0% |
REUN | 57 | 136 | 56 | 249 | 32.8% | 48.3% | 18.9% | 100.0% | |
AFDC | 1,053 | 2,980 | 1,448 | 5,481 | 19.2% | 54.4% | 26.4% | 100.0% | |
Region | |||||||||
Urban | AGE OUT | 21 | 111 | 35 | 167 | 15.6% | 64.5% | 19.9% | 100.0% |
REUN | 110 | 332 | 105 | 547 | 30.3% | 53.4% | 16.3% | 100.0% | |
AFDC | 985 | 3,962 | 1,481 | 6,428 | 15.3% | 61.6% | 23.0% | 100.0% | |
Rural | AGE OUT | 23 | 92 | 23 | 138 | 17.6% | 66.2% | 16.2% | 100.0% |
REUN | 53 | 119 | 54 | 226 | 29.1% | 49.0% | 21.9% | 100.0% | |
AFDC | 950 | 2,505 | 1,581 | 5,036 | 18.9% | 49.7% | 31.4% | 100.0% |
Characteristics | Study Population | No Earnings | 4 quarters prior and quarter of 18th birthday | 8 quarters after 18th birthday | Total | No Earnings | 4 quarters prior and quarter of 18th birthday | 8 quarters after 18th birthday | Total |
Counts | Percentage | ||||||||
Total | AGE OUT | 641 | 1,387 | 796 | 2,824 | 22.7% | 49.1% | 28.2% | 100.0% |
REUN | 980 | 1,368 | 790 | 3,138 | 31.2% | 43.6% | 25.2% | 100.0% | |
AFDC | 57,453 | 75,141 | 54,043 | 186,637 | 30.8% | 40.3% | 29.0% | 100.0% | |
Race | |||||||||
African-American | AGE OUT | 219 | 414 | 277 | 910 | 24.1% | 45.5% | 30.4% | 100.0% |
REUN | 286 | 276 | 211 | 773 | 37.0% | 35.7% | 27.3% | 100.0% | |
AFDC | 12,878 | 11,858 | 9,784 | 34,520 | 37.3% | 34.4% | 28.3% | 100.0% | |
Hispanic | AGE OUT | 125 | 291 | 164 | 580 | 21.6% | 50.2% | 28.3% | 100.0% |
REUN | 222 | 355 | 229 | 806 | 27.5% | 44.0% | 28.4% | 100.0% | |
AFDC | 17,835 | 29,627 | 22,172 | 69,634 | 25.6% | 42.5% | 31.8% | 100.0% | |
White | AGE OUT | 275 | 613 | 318 | 1,206 | 22.8% | 50.8% | 26.4% | 100.0% |
REUN | 432 | 669 | 317 | 1,418 | 30.5% | 47.2% | 22.4% | 100.0% | |
AFDC | 19,135 | 24,385 | 14,474 | 57,994 | 33.0% | 42.0% | 25.0% | 100.0% | |
Other | AGE OUT | 21 | 64 | 36 | 121 | 17.4% | 52.9% | 29.8% | 100.0% |
REUN | 40 | 67 | 33 | 140 | 28.6% | 47.9% | 23.6% | 100.0% | |
AFDC | 7,082 | 9,068 | 7,484 | 23,634 | 30.0% | 38.4% | 31.7% | 100.0% | |
Sex | |||||||||
Female | AGE OUT | 370 | 879 | 469 | 1,718 | 21.5% | 51.2% | 27.3% | 100.0% |
REUN | 612 | 912 | 467 | 1,991 | 30.7% | 45.8% | 23.5% | 100.0% | |
AFDC | 29,455 | 40,189 | 27,004 | 96,648 | 30.5% | 41.6% | 27.9% | 100.0% | |
Male | AGE OUT | 271 | 508 | 327 | 1,106 | 24.5% | 45.9% | 29.6% | 100.0% |
REUN | 368 | 456 | 323 | 1,147 | 32.1% | 39.8% | 28.2% | 100.0% | |
AFDC | 27,998 | 34,952 | 27,039 | 89,989 | 31.1% | 38.8% | 30.0% | 100.0% | |
Region | |||||||||
L.A. County | AGE OUT | 179 | 256 | 213 | 648 | 27.6% | 39.5% | 32.9% | 100.0% |
REUN | 346 | 306 | 276 | 928 | 37.3% | 33.0% | 29.7% | 100.0% | |
AFDC | 17,281 | 20,242 | 18,180 | 55,703 | 31.0% | 36.3% | 32.6% | 100.0% | |
Rural | AGE OUT | 462 | 1,131 | 583 | 2,176 | 21.2% | 52.0% | 26.8% | 100.0% |
REUN | 634 | 1,062 | 514 | 2,210 | 28.7% | 48.1% | 23.3% | 100.0% | |
AFDC | 40,172 | 54,899 | 35,863 | 130,934 | 30.7% | 41.9% | 27.4% | 100.0% |
Earnings after their eighteenth birthday
We examined earnings for those youth who first worked in the 8 quarters after their eighteenth birthday (Exhibit 4). In California and South Carolina, if youth did not work prior to exit, there was a slightly more than 50-50 chance that they would be begin employment after exit. In Illinois, youth who did not have earnings prior to their eighteenth birthday had less than a 50 percent chance of working by age 20. These findings suggest the potential importance of providing work-related services or experiences prior to exit.
State | Study Population | No Earnings | 4 quarters prior and quarter of 18th birthday | 8 quarters after 18th birthday | Percent of youth whose first employment is after 18 |
A | B | C | C/(A+C) | ||
Illinois | Age Out | 29.20% | 50.50% | 20.40% | 41.13% |
Reunified | 20.00% | 60.60% | 19.30% | 49.11% | |
AFDC | 25.30% | 53.20% | 21.50% | 45.94% | |
South Carolina | Age Out | 14.40% | 66.60% | 19.00% | 56.89% |
Reunified | 21.10% | 58.30% | 20.60% | 49.40% | |
AFDC | 16.90% | 56.40% | 26.70% | 61.24% | |
California | Age Out | 22.70% | 49.10% | 28.20% | 55.40% |
Reunified | 31.20% | 43.60% | 25.20% | 44.68% | |
AFDC | 30.80% | 40.30% | 29.00% | 48.49% |
Quarters in which youth had earned income
Exhibits 5 a-c shows the percentage of youth in each group who had earnings during each of the 13-quarter observation periods. The primary finding here is that in none of the three states in any quarter are there more than 45 percent of the aging-out youth who have earnings. This is also the case for reunified youth. For AFDC/TANF youth, a larger percentage of youth have earnings, but never more than 50 percent. Although some of the youth who do not have earnings recorded in the UI wage reporting data may have earned income from sources not captured in that data, it is unlikely that it is a major portion of those for whom we do not record earnings.
In California, the percentage of aging-out youth who had earnings grew steadily during the 13-quarter study period to a high of about 42 percent. In South Carolina, after increasing during the first 3 quarters, the percentage with earnings stayed relatively flat at around 40 percent for the remainder of the period, reaching a high of 44 percent in the last quarter. In Illinois, after growth in the percentage in the first 2 quarters, the percentage with earnings flattened at about 30 percent for the remainder of the 13 quarters.
In Illinois and California, the reunified and AFDC/TANF groups looked quite similar, growing steadily over the period to a high near 40 percent (+/- 2 points) in the final quarter. In South Carolina, the reunified group was more similar to the aging-out group, flattening out at 40 percent, and the AFDC/TANF group shows the strongest growth of any subgroup reaching a high of 49 percent by the end of the study period.
Exhibit 5a. Quarters in which youth had earned income for three groups in Illinois: Aging Out, Reunification, and Low-Income Groups
Exhibit 5b. Quarters in which youth had earned income for three groups in S. Carolina: Aging Out, Reunification, and Low-Income Groups
Exhibit 5c. Quarters in which youth had earned income for three groups in California: Aging Out, Reunification, and Low-Income Groups
The average earnings of youth are remarkably similar across states. In Exhibits 6, we only include those youth who had earnings. In general, when one looks across the three states at the average hourly wage levels for low-paying service sector jobs for the general population, earnings in California are generally 5 to 15 percent greater than Illinois and Illinois is 5-10 percent greater than South Carolina.(1) We would therefore expect that California youth would earn more than Illinois youth, who would earn more than South Carolina youth. This expectation is borne out. Aging-out youth earn about $300 per quarter (about 20%) more than Illinois and South Carolina youth for the entire 13-quarter period. However, earnings for Illinois and South Carolina aging-out youth are virtually the same. One could tentatively conclude from this that Illinois youth (who do have earnings) are probably earning the least relative to the mean earnings of the general population.
Average quarterly earnings do grow significantly from the 4 quarters prior to the eighteenth birthdays to the 8 quarters after it. In each state, the average earnings increase roughly $500 per quarter between the two periods. However, even with these increases, these youth average less than $6,000 per year in wages, which is substantially below the 1997 poverty level of $7,890 for a single individual.(2)
Youth aging out of foster care earn less than all of the youth in the comparison groups both prior to and after their eighteenth birthday. AFDC/TANF youth across the states have less variation on mean earningswith Illinois AFDC/TANF youth earning the most per quarter followed by California and South Carolina youth. The differential between aging out youth and AFDC/TANF youth is the greatest in Illinois, suggesting that Illinois aging out youth have the least success in obtaining employment in the formal labor market.
States | Study Populations | Mean Earnings Per Quarter | Mean Earnings Per Quarter Prior to 18th Birthday | Mean Earnings Per Quarter After 18th Birthday |
Illinois | AGE OUT | $1,089.04 | $719.15 | $1,233.17 |
REUN | $1,299.51 | $938.26 | $1,427.34 | |
AFDC | $1,560.43 | $1,038.48 | $1,733.22 | |
South Carolina | AGE OUT | $1,097.35 | $656.66 | $1,260.53 |
REUN | $1,310.18 | $874.40 | $1,459.20 | |
AFDC | $1,336.09 | $867.06 | $1,474.55 | |
California | AGE OUT | $1,363.93 | $925.34 | $1,558.85 |
REUN | $1,596.59 | $1,151.56 | $1,794.38 | |
AFDC | $1,486.85 | $1,002.56 | $1,702.09 |
Multivariate results
The multivariate analyses focus on the differences between the aging-out group and the two comparison groups in having earnings during the post-eighteenth birthday period and the amount of those earnings. In analyzing whether they have earnings after their eighteenth birthday, we include all youth; we only include those youth who have earnings in the analysis of the amount of those earnings.
Employment During the First 8 Post-Exit Quarters
Aging out and reunified groups
We employ logistic regression to understand the multivariate effects on the likelihood of employment during the 8 quarters after the youth turns 18 years old (Exhibit 7a). We compare our findings to those of Dworsky and Courtney in Wisconsin because they completed a very similar analysis. We control for race, gender, age at entry to foster care, reason for entry into foster care, and placement type at exit from foster care. With these controls, we find that the aging-out group is more likely to have earnings after their eighteenth birthday than the reunified group in South Carolina and California, with no difference in Illinois. Dworsky and Courtney (2001) found no difference in Wisconsin. The findings in Illinois and California are consistent with the descriptive analyses described above, but the multivariate findings differ from the descriptive findings for South Carolina.
The effect of race and ethnicity is quite different across the states. Hispanic youth are more likely than white and African American youth to work in California. African American youth are less likely to work than white and Hispanic youth in Illinois. In South Carolina, there is no race effect. Dworsky and Courtney found that African American and Hispanic youth were less likely to work than white youth in Wisconsin.
Males are less likely to work than females in Illinois. California and Illinois urban youth are less likely to work than non-urban youth. There are no gender or regional effects in South Carolina.
There were a few statistically significant effects of characteristics of the foster care experience. Youth who exited from kinship care compared to all other children (traditional foster care, group homes, or institutions and other types of placement, including independent living) are more likely to work in Illinois. Dworsky and Courtney found that youth exiting from traditional foster care were more likely to work than youth that exited from group homes and institutions. Youth in South Carolina who were placed because of parent-child conflicts were more likely to have earnings than youth who were placed for abuse or neglect and all other reasons. In California, the older the youth were at the time of initial placement in foster care, the more likely they were to have earnings.
Illinois | South Carolina | California | |||||||
Characteristics | Parameter Estimate | p-value | Adjusted Odds Ratio | Parameter Estimate | p-value | Adjusted Odds Ratio | Parameter Estimate | p-value | Adjusted Odds Ratio |
Intercept | 1.698 | *** | 1.877 | *** | -0.144 | 0.866 | |||
Gender | |||||||||
Male | -0.389 | *** | 0.678 | -0.250 | 0.779 | -0.010 | 0.990 | ||
Female | 1.000 | 1.000 | 1.000 | ||||||
Race/Ethnicity | |||||||||
Black | -0.930 | *** | 0.394 | -0.159 | 0.853 | -0.062 | 0.940 | ||
Hispanic | -0.059 | 0.943 | N/A | N/A | 0.209 | *** | 1.232 | ||
Other race/ethnicity | -0.891 | * | 0.410 | -0.711 | 0.491 | 0.197 | 1.218 | ||
White | 1.000 | 1.000 | 1.000 | ||||||
Age at Entry to Foster Care (continuous) | -0.009 | 0.991 | -0.020 | 0.980 | 0.047 | *** | 1.048 | ||
County Providing Service | |||||||||
Primary Urban County | -0.398 | *** | 0.672 | 0.162 | 1.176 | -0.244 | *** | 0.783 | |
All Other Counties | 1.000 | 1.000 | 1.000 | ||||||
Reason for Entry to Foster Care | |||||||||
Abuse | 0.067 | 1.069 | 0.066 | 1.068 | -0.026 | 0.974 | |||
Neglect | 1.000 | 1.000 | 1.000 | ||||||
Parent Child Problems | -0.073 | 1.178 | 2.016 | * | 1.104 | ||||
Other Reason | 0.164 | 0.930 | 0.099 | 7.511 | 0.258 | *** | 1.294 | ||
Placement Type at Exit | |||||||||
HMR | 0.386 | ** | 1.471 | -0.374 | 0.688 | 0.086 | 1.089 | ||
Traditional foster care | 0.027 | 0.973 | -0.312 | 0.732 | 0.076 | 1.079 | |||
Other placement | 0.058 | 0.944 | 0.041 | 1.042 | 0.053 | 1.055 | |||
Group Home or Institution | 1.000 | 1.000 | 1.000 | ||||||
Comparison Group | |||||||||
Age Out | 0.171 | 0.843 | 0.512 | ** | 1.669 | 0.502 | *** | 1.652 | |
Reunification | 1.000 | 1.000 | 1.000 | ||||||
N/A Data not available. * p.05 ** p.01 *** p.001 |
Aging-out, reunified and AFDC/TANF youth
We also model the likelihood of aging-out, reunified, and AFDC/TANF youth having earnings during the 8 quarters after their eighteenth birthday using logistic regression (Exhibit 7b). We do this in order to understand how the foster care groups compare to a group of low-income youth. In general, there is no pattern across the states in the likelihood of being employed after the eighteenth birthday. Since the results of these models are driven by the AFDC/TANF youth because of the relative size of these populations, the actual results for the foster care youth should be taken from the previous models. Nevertheless, these results are useful to see how foster youth compare to low-income youth.
In California, the aging-out group is more likely to be employed than both comparison groups. In South Carolina, the aging-out youth and AFDC/TANF youth are more likely to be employed than the reunified youth. In Illinois, the reunified youth are more likely to be employed than the aging-out youth and the AFDC/TANF youth.
Males are less likely to have earnings in Illinois and South Carolina. In California, there is no effect of gender. African American youth in California and Illinois are less likely to have earnings than white youth. Hispanic youth and those of other races are more likely to have earnings than white youth in California. The opposite is true in Illinois, with white youth being more likely to have earnings than youth of all other races and ethnicities. In South Carolina, youth who are not African American or white are less likely to be employed than these two racial groups, although this is a very small number of youth.
In all three states, the older youth are when they enter foster care or AFDC/TANF, the less likely they are to be employed. This suggests that those youth who are closer in time to the crisis that brought them to the program have more difficulties becoming employed, although in the previous models, this does not seem to be the case for foster care youth in California. If the youth live in either Los Angeles or Cook County, they will be less likely to have earnings than youth living in the balance of those states. In South Carolina, youth are more likely to have earnings if they live in an urban area.
Illinois | South Carolina | California | |||||||
Characteristics | Parameter Estimate | p-value | Adjusted Odds Ratio | Parameter Estimate | p-value | Adjusted Odds Ratio | Parameter Estimate | p-value | Adjusted Odds Ratio |
Intercept | 1.662 | *** | 2.106 | *** | 0.583 | *** | 1.792 | ||
Gender | |||||||||
Male | -0.408 | *** | 0.665 | -0.314 | *** | 0.731 | 0.007 | 1.007 | |
Female | 1.000 | 1.000 | 1.000 | ||||||
Race/Ethnicity | |||||||||
Black | -0.662 | *** | 0.516 | -0.108 | 0.898 | -0.186 | *** | 0.831 | |
Hispanic | -0.130 | *** | 0.878 | N/A | N/A | 0.372 | *** | 1.451 | |
Other race/ethnicity | -0.592 | *** | 0.553 | -0.881 | *** | 0.415 | 0.099 | *** | 1.104 |
White | 1.000 | 1.000 | 1.000 | ||||||
Age at Entry to Foster Care Service (continuous) | -0.020 | *** | 0.980 | -0.033 | *** | 0.967 | -0.005 | *** | 0.995 |
County Providing Service | |||||||||
Primary Urban County | -0.449 | *** | 0.638 | 0.259 | *** | 1.295 | -0.035 | *** | 0.965 |
All Other Counties | 1.000 | 1.000 | 1.000 | ||||||
Comparison Group | |||||||||
Age Out | -0.068 | 0.935 | 0.223 | 1.250 | 0.386 | *** | 1.472 | ||
Reunification | 0.201 | ** | 1.223 | -0.323 | *** | 0.724 | -0.029 | 0.972 | |
AFDC | 1.000 | 1.000 | 1.000 | ||||||
Age Out vs. Reunification | -0.117 | * | 0.765 | 0.237 | ** | 1.727 | 0.180 | *** | 1.515 |
N/A Data not available. * p.05 ** p.01 *** p.001 |
Total Earnings During the First 8 Quarters
Comparing the aging-out and reunified groups
We model the amount of earnings during the first 8 quarters after turning 18 using ordinary least squares regression. It is important to note that none of these models explain a great deal of the variation in earnings the highest R2 is for Illinois at 4.7 percent (Exhibit 8a). The variables that are available to us do not explain the variation well. In most research of this type, explaining 20-30 percent of the variation would be more satisfactory. By adding additional variables, such as earnings prior to the eighteenth birthday, we would increase the R2, but we would also include an endogenous variable that may bias our estimation of the other effects. Dworsky and Courtney (2001) have similar R2 statistics in their models.
We can compare the intercepts across the states because the covarites in each model are the same (i.e. the comparison categories for each covariate is the same across categories). The intercepts represent the mean earnings for the youth whose values for the explanatory variables are 0 (female, white, non-urban area, neglect, exiting from group homes or institutions, and having been reunified), while controlling for all of the variables in the model. The intercept is higher in South Carolina ($8,114) than in both Illinois ($7,166) and California ($7,123). This means that youth who are female, white, from non-primary urban areas, in care for neglect, who exit from a group home or institution, and are reunified have greater earnings in South Carolina than in Illinois and California.
There are significant differences between aging-out youth and reunified groups in Illinois and California, where the aging-out group earned from $783 (CA) to $1,213 (IL) less during the 8-quarter period than the reunified youth. In Wisconsin, Dworsky and Courtney (2001) found that the aging-out group earned more than the reunified group.
African Americans earn less than white youth in all states, from just over $1,000 less in California to nearly $3,000 less in South Carolina and Illinois during the 8 quarters. Dworsky and Courtney (2001) also found that African American youth earned less than all other groups in Wisconsin. Whites earn less than Hispanic youth in Illinois. Males earn more than females in South Carolina and Illinois. In California, youth earn less in LA County than in the rest of the state, with no significant geographic differences in the other states.
Only in Illinois is there an effect of reason for placement and type of placement. All youth who entered care for reasons other than neglect earned less money. Children in other placements in Illinois (primarily independent living) earn less than youth placed in group homes or institutions. These are primarily those youth who age out.
Illinois | South Carolina | California | ||||
Characteristics | Parameter Estimate | p-value | Parameter Estimate | p-value | Parameter Estimate | p-value |
Intercept | 7166.436 | *** | 8114.970 | *** | 7123.297 | *** |
Gender | ||||||
Male | 554.798 | *** | 744.124 | *** | -43.315 | |
Female | ||||||
Race/Ethnicity | ||||||
Black | -2982.326 | *** | -2825.665 | *** | -1129.999 | ** |
Hispanic | 1976.339 | *** | N/A | N/A | 207.384 | |
Other | -484.784 | 5964.594 | *** | 1312.910 | ||
White | ||||||
Age at Entry to Foster Care (continuous) | 206.889 | *** | 31.204 | 83.129 | ||
County Providing Service | ||||||
Primary Urban County | -23.780 | -99.922 | -1293.188 | ** | ||
All Other Counties | ||||||
Reason for Entry to Foster Care | ||||||
Abuse | -1616.898 | * | -538.658 | 224.504 | ||
Neglect | ||||||
Parent Child Problems | -2694.522 | *** | 795.721 | |||
Other Reason | -1276.982 | ** | 871.409 | 559.869 | ||
Placement Type at Exit | ||||||
HMR | -486.785 | -637.576 | 1165.773 | ** | ||
Traditional foster care | -777.462 | -961.400 | 302.319 | |||
Other placement | -1546.329 | ** | -377.229 | 36.612 | ||
Group Home or Institution | ||||||
Comparison Group | ||||||
Age Out | -1213.854 | * | -986.619 | -783.188 | ** | |
Reunification | ||||||
R-square | 0.047 | 0.026 | 0.011 | |||
N/A Data not available. * p.05 ** p.01 *** p.001 |
Comparing the aging out, reunified and AFDC/TANF youth
When we add the AFDC/TANF youth to the models in the previous section, we see many similarities. Both groups of foster care youth earn less than AFDC/TANF youth in all three states, except for reunified youth in California who earn more than AFDC/TANF youth (Exhibit 8b). For the 8-quarter period, California aging-out youth earn $478 less than AFDC/TANF youth; Illinois aging-out youth earning $3,767 less than the Illinois AFDC/TANF group. Aging-out youth have the lowest earnings in all three states, when controlling for the other covariates.
African American youth earn the least relative to all other racial/ethnic groups in all three states. Hispanic youth in Illinois and California earn more than white youth. The older youth are when they begin a foster care or AFDC/TANF episode, the more they earn in Illinois and California. Males earn more in all states. There is no urban effect in these models.
Illinois | South Carolina | California | ||||
Characteristics | Parameter Estimate | p-value | Parameter Estimate | p-value | Parameter Estimate | p-value |
Intercept | 7176.841 | *** | 8613.367 | *** | 7183.134 | *** |
Gender | ||||||
Male | 555.038 | *** | 744.778 | *** | 1190.683 | *** |
Female | ||||||
Race/Ethnicity | ||||||
Black | -3007.306 | *** | -2829.643 | *** | -2501.277 | *** |
Hispanic | 1941.960 | *** | N/A | N/A | 297.332 | *** |
Other | -527.505 | 5953.123 | *** | -892.953 | ||
White | ||||||
Age at Entry to Service (continuous) | 209.410 | *** | 31..352 | 83.620 | *** | |
County Providing Service | ||||||
Primary Urban County | 834.455 | -99.437 | -43.159 | |||
All Other Counties | ||||||
Comparison Group | ||||||
Age Out | -3767.202 | *** | -2004.178 | *** | -478.422 | ** |
Reunification | -2681.694 | *** | -1166.424 | ** | 142.961 | * |
AFDC | ||||||
R-square | 0.048 | 0.026 | 0.017 | |||
N/A Data not available. * p.05 ** p.01 *** p.001 |
Difference between average earnings in the first and second year after turning 18
In order to determine whether there was a major difference between earnings during the first and second year after the eighteenth birthday, we conducted similar OLS regressions for 4 quarters as we did for 8 quarters discussed above. We found no substantive differences across the study populations or the states. Therefore, we do not report those results here.
However, it is important to analyze changes in earnings from the first to second year after turning 18 in order to understand how these youth progressed in the labor market (Exhibit 9). In all three states, on average, youth earned more in their second year, with significant differences among sub-groups. The mean increase in California, as represented by the intercept, is four times as large as that in South Carolina and more than six times as large as Illinois. The importance of this is that California youth make a very large jump in earnings between their first and second years after turning 18.
AFDC/TANF youth have a larger increase in earnings than both aging-out and reunified youth in Illinois. The same is true in South Carolina and California, but the differences are not significant. In general, this analysis suggests that foster care youth do not progress in the labor market as quickly as AFDC/TANF youth.
African American youth make fewer gains than white youth, who make fewer gains than Hispanic youth in California and Illinois. Males have a larger increase than females. There is no consistent urban effect.
Illinois | South Carolina | California | ||||
Characteristics | Parameter Estimate | p-value | Parameter Estimate | p-value | Parameter Estimate | p-value |
Intercept | 259.464 | *** | 481.825 | *** | 1849.196 | *** |
Gender | ||||||
Male | 39.529 | ** | 61.754 | ** | 447.198 | *** |
Female | ||||||
Race/Ethnicity | ||||||
Black | -70.253 | *** | -73.910 | ** | -393.636 | *** |
Hispanic | 175.934 | *** | N/A | N/A | 230.640 | *** |
Other | 64.130 | 475.307 | *** | 62.815 | ||
White | ||||||
Age at Entry to Service (continuous) | 11.180 | *** | -6.157 | 21.401 | *** | |
County Providing Service | ||||||
Primary Urban County | 79.339 | -58.717 | ** | 136.696 | *** | |
All Other Counties | ||||||
Comparison Group | ||||||
Age Out | -191.355 | *** | -99.052 | -38.190 | ||
Reunification | -179.292 | *** | -68.393 | -321.477 | * | |
AFDC | ||||||
R-square | 0.007 | 0.005 | 0.005 | |||
N/A Data not available. * p.05 ** p.01 *** p.001 |
[Go To Contents]
Summary of Results
These analyses show clearly that youth aging out of foster care have very low levels of employment and earnings. Fewer than half of youth aging out of foster care have earnings in any given quarter, many have no earnings at all during the three year study period, and those who are employed earn very little. Specific employment rates vary substantially among the three state studied. In addition, whether youth aging out of foster care look better or worse on employment measures when compared to youth reunified with their families and youth on welfare is inconsistent.
Initiation of Employment
The percent of youth aging out of foster care who had earnings at any point from four quarters prior to their eighteenth birthday to 8 quarters after varied dramatically by state. About 30 percent of youth aging out in Illinois, 23 percent in California, and 14 percent in South Carolina had no earnings during the entire 13-quarter period.
In none of the three states in any of the 13 quarters are there more than 45 percent of the aging-out youth who have earnings. This is also the case for reunified youth. For AFDC/TANF youth, there is a larger percentage of youth who have earnings, but never more than 50 percent.
In all three states, youth were more likely to earn income for the first time during the 4 quarters prior to and the quarter of their eighteenth birthday than in the 2 years afterward. For youth who exited foster care through aging out, half in California and Illinois and two-thirds in South Carolina had earnings prior to their eighteenth birthday. The aging-out group was more likely to work than the reunified group in South Carolina and California, and there was no difference in Illinois.
In California and South Carolina, if youth did not work prior to age 18, there was slightly more than a 50-50 chance that they would begin employment after age 18. In Illinois, youth who did not have earnings prior to their eighteenth birthday were unlikely to begin earning income after their exit from foster care during our study period.
Earnings
Youth aging out of foster care earn less than all of the youth in the comparison groups, both prior to and after their eighteenth birthday. Average quarterly earnings do grow significantly from the 4 quarters prior to the eighteenth birthday to the 8 quarters after it. In each state, the average earnings increase roughly $500 per quarter. However, even with these increases, these youth average less than $6,000 per year in wages, which is significantly below the 1997 poverty level of $7,890 for a single individual. The multivariate analyses confirm these findings.
In Illinois, AFDC/TANF youth make a bigger increase from the first year to the second year after their eighteenth birthday than all foster care youth. AFDC/TANF and aging-out youth in California make a larger increase than reunified youth. There is no difference among the groups in South Carolina.
Multivariate Findings
In the multivariate analysis, we find that the aging-out group is more likely to have earnings after their eighteenth birthdays than the reunified group in South Carolina and California, with no difference in Illinois.
There are significant differences between aging out youth and reunified youth groups in Illinois and California, where the aging-out group earned from $783 (CA) to $1,213 (IL) less for the 2-year period than the reunified youth.
Comparison with Current Population Survey Employment Data
The youth analyzed in this report represent a sub-group of the American workforce for which there is little information. Not only are the employment patterns of young people exiting foster care seldom studied, but the employment patterns of youth in general are not often the focus of national statistics. There are two sources of national data that we used to compare with rates observed among the study populations. First, from routine Current Population Survey results, one can identify the civilian employment-population ratio among youth ages 16-19. This statistic is seasonally adjusted and represents the proportion of the population that is employed. Calculated on a monthly basis, the rates since 1996 have ranged from 43-45 percent. (We would expect monthly employment statistics to be somewhat lower than quarterly statistics, since an individual only had to have earnings at anytime during the quarter, rather than at anytime in a month.) Compared to the quarterly percentage of study population youth who worked between their seventeenth and nineteenth birthdays, we see that only the AFDC/TANF group in Illinois and California approach these averages as they near the end of the fourth quarter after their eighteenth birthday. Youth aging out of foster care and youth reunified with their families from foster care work less than their agemates do in the general population.
A second source of data also comes from the Current Population Survey. In a report on trends in youth employment among youth ages 15-17, CPS data was used to calculate the percent of youth employed during the school year and the summer separately. Using specially tabulated Illinois data as a comparison, we found that 16 percent of the foster care group was employed compared with 24.7 percent of youth in general. During the summer, the difference was even greater, with 19.4 percent of Illinois foster children age 15-17 employed compared with 33.8 percent of youth in general. Both comparisons of the results of this study with CPS data show that foster children work less than the nation's youth overall.
Usefulness of These Methods
The two principal ways of learning about how youth fare after leaving foster care are (1) to ask youth themselves through survey research; and (2) to analyze their interactions with government programs using administrative data. To date, most research on outcomes for youth aging out of foster care has been of the former type. While valuable and rich in detail, such studies are difficult and expensive to conduct. The current study was intended, in part, to test the feasibility and utility of using administrative data to examine one key outcome of interest: employment.
The results obtained from unemployment insurance wage data generally agree with those obtained through surveys. There is reason to believe that coverage issues in UI data, particularly the lack of information on informal employment, cause us to underestimate total wages somewhat. Most studies have found this underestimate to be 10 to 14 percent, although with some populations, especially male youth with arrest records, larger discrepancies have been reported (Hotz and Scholz, 2002). Survey data, however, also are problematic when sporadic or short term employment are involved. But since wage reporting data collection is standard practice, it can be used over time to develop reliable trend information, even if the estimates are somewhat low. In addition, such undercounting is likely to be similar across comparison groups and therefore unlikely to affect relative income and employment rates.
Overall we believe that unemployment insurance data represent a useful complement to survey research on outcomes for youth who have aged out of foster care. While the results reported here may underrepresent income to some extent, our findings are consistent with survey based research on this population. In addition, the earnings of former foster care youth are so low that we would remain concerned about their employment status even if we have missed substantial income. Finally, employment rates are unlikely to be seriously compromised by the underreporting of income through UI data.
[Go To Contents]
Future Research
This analysis only addresses one issue for youth exiting foster care when youth have earnings. There are a number of additional questions that need to be addressed before the field has a complete picture of the challenges that these youth face, and then, to understand what programs might help improve outcomes. Some of these questions are:
- What is the educational achievement of youth aging out of foster care?
- How many are entering the armed forces?
- What is the participation in welfare programs for youth aging out?
- What is the criminal history of youth aging out of foster care?
- How do educational achievement and participation in welfare programs interact with employment to affect the well-being of the youth?
- How do special needs affect the employment experiences of youth?
Some of these questions can be addressed through the use of administrative data in specific jurisdictions. Currently, however, only the question of participation in welfare programs can be addressed in a comparable way in multiple states. Educational achievement and special needs data are not readily available to be linked to foster care data in many states. Either these data have to be developed, or we must continue to rely on smaller, survey-based studies or evaluations to understand the outcomes for these youth.
[Go To Contents]
References
Collins, Mary Elizabeth. (2001). Transition to Adulthood for Vulnerable Youths: A Review of Research and Implications for Policy. Social Service Review 75, 271-291.
Cook, Ronna. (1994). Are We Helping Foster Care Youth Prepare for Their Future? Children and Youth Services Review 16, 213-229.
Courtney, Mark, Irving Piliavin, and Andrew Grogan-Kaylor. (1998). The Wisconsin Study of Youth Aging Out of Out-of-Home Care: A Portrait of Children about to Leave Care. (See an article covering this material in the Nov/Dec 2001 issue of Child Welfare)
Dworsky, Amy and Courtney, Mark. (2001). Self-sufficiency of former foster youth in Wisconsin: Analysis of Unemployment Insurance Wage Data and Public Assistance Data. Institute for Research on Poverty Special Report Series. (University of Wisconsin-Madison, SR #81).
Hotz, V. Joseph and Scholz, John Karl. (2002). Measuring employment and income for low-Income populations with administrative and survey data. In Ver Ploeg, Michele, Moffitt, Robert A. and Citro, Constance, Studies of Welfare Populations, Data Collection and Research Issues. Washington, DC: National Academy Press.
Kornfeld, Robert and Bloom, Howard. (1999). Measuring program impacts on earnings and employment: Do unemployment insurance wage reports from employers agree with surveys of individuals. Journal of Labor Economics 17 (January), 168-197.
McMillen, J. Curtis, Gregory B. Rideout, Rachel H. Fisher, and Jayne Tucker. (1997). Independent-Living Services: The Views of Former Foster Youth. Families in Society: The Journal of Contemporary Human Services 78 (5), 471-79.
Scholz, K and Hotz, J. (1999). Measuring Employment Outcomes with Administrative and Survey Data. National Research Council Panel on Data and Methods for Measuring the Effects of Changes in Social Welfare Programs Workshop on Data Collection on Low Income and Welfare Populations, December 16-17, 1999.
United States General Accounting Office. (1999). Foster Care: Effectiveness of Independent Living Services Unknown. (General Accounting Office, Report no. GAO/HEHS-00-13). Washington, D.C.: U.S. General Accounting Office.
U.S. Department of Health and Human Services. (2000). Dynamics of Childrens Movement Among the AFDC/TANF, Medicaid, and Foster Care Programs Prior to Welfare Reform: 1995 1996. Office of the Assistant Secretary for Planning and Evaluation. (1)
U.S. Department of Health and Human Services. (2000). Health Care Conditions, Utilization and Expenditures of Children in Foster Care. Office of the Assistant Secretary for Planning and Evaluation. (2)
Westat, Inc. (1991). A National Evaluation of Title IV-E Foster Care Independent Living Programs for Youth. Washington, D.C.: Department of Health and Human Services.
Wulczyn, Fred and Kristen Brunner Hislop. (2001). Children in Substitute Care at Age 16: Selected findings from the Multistate Foster Care Data Archive. (Unpublished manuscript) Chapin Hall Center for Children. University of Chicago, April 2, 2001.
Wulczyn, Fred, Robert Goerge, and Kristen Brunner Hislop. (1999). Foster Care Dynamics: An Eleven State Report from the Multistate Foster Care Archive. Chicago: Chapin Hall Center for Children.
[Go To Contents]
Endnotes
(1) U.S. Department of Labor (2001) 2000 State Occupational Employment and Wage Estimates http://www.bls.gov/oes/2000/oessrcst.htm (December 3, 2001)
(2) http://aspe.hhs.gov/poverty/97poverty.htm
Acknowledgements
We would like to thank all of the state agencies that supported us through supplying data and substantive information for this report. This includes: Mark Testa at the Illinois Department of Children and Family Services, Dave Gruenenfelder at the Illinois Department of Human Services, Marilyn Edelhoch at the South Carolina Department of Social Services, Diana Tester and David Patterson at the South Carolina Budget and Control Board Office of Research and Statistics and the California Department of Social Services.
We would also like to thank Laura Radel, our Project Officer at the Office of the Assistant Secretary for Planning and Evaluation at the U.S. Department of Health and Human Services, for her substantive support and patience during this project.