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/

How to Obtain a Printed Copy

Contents

Endnotes


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:

  1. What are the patterns of employment and the amount of earnings of youth aging out of foster care?
  2. How do these employment patterns compare with those of other youth of similar ages in low-income families?
  3. 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:

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.

Exhibit 1.
Description and Size of Study Populations
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 child’s 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.

Limitations

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 state’s 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.

Exhibit 2.
Sample Population Characteristics
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 child’s parent being in jail or prison.

Descriptive Results

Timing of employment earnings

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 child’s 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.

Exhibit 3a.
Experience of First Earnings in Illinois
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%

Exhibit 3b.
Experience of First Earnings in South Carolina
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%

Exhibit 3c.
Experience of First Earnings in California
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.

Exhibit 4.
Youth having first earnings after their 18th birthday
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 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 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

Exhibit 5c. Quarters in which youth had earned income for three groups in California: Aging Out, Reunification, and Low-Income Groups.

Mean Earnings

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.

Exhibit 6.
Mean Earnings per Child Per Quarter
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.
Exhibit 7a.
Logistic Regression Model of the Likelihood of Employment
During the First Eight Post-Exit Quarters for Foster Care Groups
  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.

Exhibit 7b.
Logistic Regression Model of the Likelihood of Employment
During the First Eight Post-Exit Quarters for All Comparison Groups
  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.

Exhibit 8a.
Estimates from OLS Regression of Estimated Wages
During First 2 Years After Age 18 for only Foster Care Groups
  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.

Exhibit 8b.
Estimates from OLS Regression of Estimated Wages During First 2 Years
After Age 18 for all Comparison Groups
  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.
Exhibit 9.
Estimate from OLS Regression of the Difference Between Average Earnings
in First and Second Year After Age 18
  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:

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 Children’s 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.


How to Obtain a Printed Copy

To obtain a printed copy of this report, send or fax the title and your mailing address to:

Human Services Policy, Room 404E
Assistant Secretary for Planning and Evaluation
200 Independence Av, SW
Washington, DC 20201
Fax:  (202) 690-6562


Where to?

Top of Page | Contents

Home Pages:
Human Services Policy (HSP)
Assistant Secretary for Planning and Evaluation (ASPE)
U.S. Department of Health and Human Services (HHS)

Last updated:  07/09/02