Skip to main content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Self-Sufficiency of Former Foster Youth in Wisconsin

Publication Date

Amy Dworsky and Mark E. Courtney
Institute for Research on Poverty
University of Wisconsin-Madison

December 2000

A report prepared for the
Office of the Assistant Secretary for Planning and Evaluation
U.S. Department of Health and Human Services

This report available on the Internet at:
http://aspe.hhs.gov/hsp/fosteryouthWI00/

Contents

Introduction

Previous research has raised concerns about the self-sufficiency of former foster youth who aged out of care (e.g., Barth, 1990; Cook, Fleishman & Grimes, 1991; Courtney, Piliavin, Grogan-Kaylor, & Nesmith, 1998; Festinger, 1983; Jones & Moses, 1984; Pettiford, 1981). In particular, several studies have shown that it is difficult for many former foster youth to maintain stable employment, and that the earnings of those who are employed are low. These studies have also found that a significant percentage of former foster youth received means-tested cash assistance or in-kind benefits such as Food Stamps at some point after their discharge from care.

With the exception of Pettiford (1981), all of the previous research has relied on interviews with former foster youth. These former foster youth were asked, among other things, about their employment, earnings and public assistance utilization since they were discharged from care. Although such studies can provide valuable information, that approach is very costly and sample attrition can be substantial.

More recently, researchers are beginning to examine the self-sufficiency of former foster youth using state administrative data.(1) This report presents the results of one such study which analyzes employment, earnings and public assistance receipt among former Wisconsin foster youth who exited out-of-home care between January 1, 1992 and December 31, 1998, and who were at least 17 years old at the time of their exit. The former foster youth were identified using the Substitute Care Module of the state's Human Services Reporting System (HSRS).(2)

The report is divided into five sections. Section 1 describes the characteristics of these youth, including information about their experiences in out-home-care. Section 2 examines the employment and earnings of these former foster youth during the first eight quarters after they exited out-of-home care. Section 3 examines their public assistance utilization during that same period. Section 4 examines total income when earnings from employment and public assistance benefits are combined. Section 5 discusses the results and their implications.

[ Go to Contents ]

Section 1
Demographic Characteristics and Out-of-Home Care
Experiences of Former Foster Youth

Between January 1, 1992 and December 31, 1998, 6274 foster youth age 17 and older exited out-of-home care in Wisconsin.(3) The data in Table 1 describe the demographic characteristics and out-of home cares experiences of these former foster youth. These former foster youth are disproportionately male, nearly three quarters are White, two-thirds were 17 years old at the time they were discharged, and one out of five were receiving services from Milwaukee County prior to exiting.(4)

Although almost 45% of these former foster youth had been adjudicated children in need of protective services (CHIPS), an even higher percentage had been adjudicated delinquent. This is consistent with two additional findings. First, three quarters of these former foster youth entered their most recent episode of out-of-home care at age 16 or 17. And second, the majority had remained in care less than 12 months during their most recent episode, and less than 5% had remained in care 5 years or more.

Nearly 60% of these former foster youth had experienced only one episode of out-of-home care; however, almost 20% had experienced three episodes or more. Regardless of the number of out-of-home care episodes these former foster youth had experienced, 40% had a cumulative length of stay of less than 12 months and another 65% had a cumulative length of stay of less than 2 years. Roughly three-quarters of these former foster youth had experienced only one placement during their most recent episode of out-of-home care, and two-thirds had experienced two or fewer different placements in total.

Although the majority of these former foster youth were discharged from foster homes, a significant minority were discharged from group homes or child care institutions (CCI's).(5) More than half were reunified or placed with relatives, 20% had aged out or been discharged to independent living and almost 13% were transferred to a state institution or other facility (e.g., hospitals, detention, jail).(6)

 

Table 1
Demographic Characteristics and Out-of Home Care Experiences of
1992 - 1998 Exit Cohorts of Former Foster Youth
(N = 6274)
  Number of Former
Foster Youth
Percent of Former
Foster Youth
Gender
Female 2673 42.6
Male 3601 57.4
Race/Ethnicity
African-American 1121 17.8
White 4610 73.5
Native American 206 3.3
Asian 105 1.7
Hispanic 191 3.0
Missing 41 0.7
County Providing Services
Milwaukee 1265 20.2
Other County 5009 79.8
Age at Exit
17 Years Old 4222 67.7
18 Years Old 1865 29.9
19 Years Old 150 2.4
Adjudicated Status (Most Recent Episode)
CHIPS - Abuse and/or Neglect 1106 17.6
CHIPS - Other 1704 27.2
Delinquent 2979 47.5
JIPS - Status Offender 228 3.6
Voluntary Placement 257 4.1
Age Entered Most Recent Episode
Ages 7 through 9 31 0.5
Ages 10 or 11 80 1.3
Ages 12 or 13 257 4.1
Age 14 or 15 1228 19.6
Ages 16 or 17 4678 74.6
Placement Type Prior to Exit
Foster Home 3650 58.2
Group Home 1514 24.1
Child Caring Institution 1110 17.7
Total Number of Episodes
1 3652 58.2
2 1477 23.5
3 619 9.9
4 291 4.6
5 115 1.8
6 or more 120 1.9
Number of Placements (Most Recent Episode)
1 4667 74.4
2 1084 17.3
3 321 5.1
4 107 1.7
5 or more 95 1.5
Total Number of Placements (All Episodes)
1 2731 43.5
2 1439 22.9
3 832 13.3
4 496 7.9
5 304 4.9
6 165 2.6
7 112 1.8
8 or more 195 3.1
Months in Out-of-Home Care (Most Recent Episode)
Less than 12 3410 54.4
12 to 23 1397 22.3
24 to 35 690 11.0
36 to 47 378 6.0
48 to 59 196 3.1
60 to 71 85 1.4
72 to 83 51 0.8
84 or more 67 1.1
Cumulative Months in Out-of-Home Care (All Episodes)
Less than 12 2471 39.2
12 to 23 1573 25.1
24 to 35 925 14.7
36 to 47 545 8.7
48 to 59 304 4.9
60 to 71 166 2.7
72 to 83 95 1.5
84 or more 195 3.1
Discharge Outcome
Reunified 3062 48.8
Placed with Relatives 258 4.1
Adopted 45 0.7
Reached Age of Majority or Completed Education 1770 28.2
Discharged to Independent Living 365 5.8
Ran Away 382 6.1
Transferred to State Institution or Other Facility 392 6.3

Our analysis of employment and earnings outcomes in Section 2 is based on data from the Unemployment Insurance file for the state of Wisconsin which identifies individuals using Social Security numbers, and only Social Security numbers. Although the Substitute Care Module of the state's Human Services Reporting System (HSRS) that we used to identify the former foster youth who fit the criteria for inclusion in our sample contains a field for Social Security number, this information is sometimes missing from the record. An SSN was present in the records of 4,316 of the 6,274 former foster youth who fit our criteria and absent from the records of 1958.(7) Although we had no a prior reason to expect systematic differences between those former foster youth for whom an SSN was present and those for whom it was not, any non-random differences could conceivably bias the results of our analysis.

Table 2 compares the demographic characteristics and out-of-home care experiences of former foster youth for whom an SSN was present in HSRS to those of former foster youth for whom an SSN was not present in HSRS. Although the two groups are similar in many respects, there are a number of potentially important differences between them. Compared to former foster youth for whom an SSN was not present, former foster youth for whom an SSN was present (1) were more likely to be African American and less likely to be White; (2) were more likely to have been receiving services from Milwaukee County; (3) were more likely to have been in care for a total of 12 months or longer; and (4) were more likely to have aged out or been discharged to independent living and less likely to have been reunified.(8) The implications of these differences are discussed in Section 5.

 

Table 2
Demographic Characteristics and Out-of-Home Care Experiences of
Former Foster Youth by Presence of Social Security Numbers in HSRS
  Percent of Former Foster Youth
SSN Present in HSRS
(N = 4316)
SSN Not Present in HSRS
(N = 1958)
Gender
Female 43.1 41.6
Male 56.9 58.4
Race/Ethnicity
African-American 21.8 8.8
White 68.0 83.4
Native American 3.1 3.3
Asian 1.8 1.5
Hispanic 3.1 2.8
Missing 2.4 0.3
County Providing Services
Milwaukee 26.1 7.0
Other County 73.9 93.0
Age at Exit
17 Years Old 66.3 70.7
18 Years Old 30.7 28.2
19 Years Old 3.0 1.1
Adjudicated Status (Most Recent Episode)
CHIPS - Abuse and/or Neglect 19.4 13.8
CHIPS - Other 26.6 28.4
Delinquent 47.4 47.6
JIPS - Status Offender 3.6 5.0
Voluntary Placement 3.0 5.3
Age Entered Most Recent Episode
Ages 7 through 9 0.7 0.2
Ages 10 or 11 1.7 0.4
Ages 12 or 13 4.8 2.6
Age 14 or 15 20.7 17.2
Ages 16 or 17 72.2 79.7
Placement Type Prior to Exit
Foster Home 57.3 60.1
Group Home 23.9 24.7
Child Caring Institution 18.8 15.2
Total Number of Episodes
1 58.2 61.8
2 23.5 23.1
3 9.9 8.3
4 4.6 4.0
5 1.8 1.5
6 or more 1.9 1.4
Number of Placements (Most Recent Episode)
1 73.4 76.6
2 17.7 16.3
3 5.3 4.6
4 1.9 1.3
5 or more 1.6 1.2
Total Number of Placements (All Episodes)
1 43.5 47.6
2 22.9 24.0
3 13.3 11.8
4 7.9 6.7
5 4.9 4.5
6 2.6 2.2
7 1.8 1.4
8 or more 3.1 1.9
Months in Out-of-Home Care (Most Recent Episode)
Less than 12 56.9 60.4
12 to 23 17.7 22.5
24 to 35 10.6 8.6
36 to 47 6.2 4.9
48 to 59 3.3 2.2
60 to 71 1.8 0.8
72 to 83 1.3 0.3
84 or more 2.4 0.3
Cumulative Months in Out-of-Home Care (All Episodes)
Less than 12 35.4 48.3
12 to 23 25.1 24.9
24 to 35 15.9 12.2
36 to 47 9.3 7.4
48 to 59 5.3 3.8
60 to 71 3.0 1.8
72 to 83 2.0 0.5
84 or more 4.0 1.1
Discharge Outcome
Reunified 46.1 54.7
Placed with Relatives 4.0 4.3
Adopted 0.9 0.4
Reached Age of Majority or Completed Education 29.8 24.7
Discharged to Independent Living 6.2 4.9
Ran Away 6.3 5.7
Transferred to State Institution or Other Facility 6.7 5.3

[ Go to Contents ]

Section 2:
Employment and Earnings Outcomes of Former Foster Youth

Our analysis of the employment and earnings outcomes of former foster youth is based on data from the state of Wisconsin's Unemployment Insurance file for the years 1995 through 1999. Because the former foster youth in our sample exited out-of-home care between January 1, 1992 and December 31, 1998, the number of post-exit quarters for which we have UI data ranges from 20 quarters for youth who exited prior to 1995 to 4 quarters for youth who exited in the fourth quarter of 1998. Moreover, if we assume that earnings tend to be positively correlated with years of work experience, then earnings for the years 1995 through 1999 are likely to be higher among earlier as compared to later exit cohorts. To deal with these two limitations of our data, the majority of our analysis focuses on employment and earnings during the first eight quarters after the 1995 through 1997 exit cohorts were discharged from care.

We use two outcome measures in our analysis of these data. The first measure is the percentage of the first eight post-exit quarters during which former foster youth were employed. We compute this measure based on the number of quarters for which the UI data show non-zero earnings.(9) The second measure is total earnings for those eight quarters. We compute this measure by summing quarterly earnings from all employers across the first eight post-exit quarters.

Before presenting the results of our analyses, we first discuss a major limitation of this approach, namely, our exclusive reliance on data from Wisconsin's Unemployment Insurance file. Relying exclusively on UI data has several important implications for our analyses. On the one hand, this approach will underestimate employment and earnings among this population to the extent that former foster youth are working in jobs that are not "covered" under the state's Unemployment Insurance regulations.(10) Although state tax records might alleviate some of this problem, we would still have no information on unreported earnings from employment in the so-called "underground economy."(11) We will also underestimate employment and earnings to the extent that former foster youth have moved outside of Wisconsin and are working in another state. On the other hand, this approach will overestimate employment (but not earnings) to the extent that former foster youth work for only a few days or weeks out of an entire quarter. We could deal with this problem, at least in part, by imposing more stringent criteria than simply any non-zero earnings. For example, we could require that earnings be above some threshold in order for a quarter to count, and then examine how our findings vary depending on where the threshold is set.(12)

Employment: Descriptive Statistics

We begin with some descriptive statistics on the post-exit employment of former foster youth. Table 4A shows the number of quarters in which the 1995 through 1997 exit cohorts of former foster youth were employed, as indicated by nonzero earnings, during the first eight quarters after they were discharged from care. These former foster youth were fairly evenly divided between those who were employed in at least half of the eight quarters and those who were not. While one in five were never employed, and one in four were employed in more than zero but fewer than three quarters, 55% were employed in 4 quarters or more, including 15% who were employed in all eight.

 

Table 4A
Employment During the First Eight Post-Exit Quarters,
1995-1997 Exit Cohorts
Number of Quarters Employed N Percent
0 380 20.9
1 131 7.2
2 152 8.4
3 163 9.0
4 162 8.9
5 159 8.7
6 188 10.3
7 217 11.9
8 267 14.7
  1819 100.0

Table 4B shows the percentage of former foster youth employed in 0%, less than 50% and 50% or more of the first eight post-exit quarters broken down by gender, race/ethnicity, county (i.e., Milwaukee vs non-Milwaukee), adjudication status, placement type and discharge outcome. The percentage of quarters employed was generally higher (1) among female former foster youth than among male former foster youth; (2) among White former foster than among non-White former foster youth (especially African American and Native former foster youth); (3) among former foster youth receiving services from non-Milwaukee counties than among former foster youth receiving services from Milwaukee County; (4) among former foster youth who were discharged from foster homes or group homes than among former foster youth discharged from child caring institutions; and (5) among former foster youth who had been reunified, placed with relatives or adopted or who had aged out or been discharged to independent living than among former foster youth who had run away from care or who had been transferred to a state institution. In addition, although the percentage employed in at least four quarters does not vary much by adjudication status, former foster youth who were adjudicated status offenders (JIPS) or placed voluntarily were more likely to have been employed in zero quarters than former foster youth who were adjudicated delinquents or children in need of protection (CHIPS).

 

Table 4B
Employment During the First Eight Post-Exit Quarters:
1995-1997 Exit Cohorts
  Percent of Quarters Employed
  N 0 50% 50% +
Gender
Female 751 19.7 23.0 57.3
Male 1068 21.7 25.6 52.7
Race/Ethnicity
African American 423 33.1 30.0 36.9
White 1246 16.5 21.9 61.6
Native American 56 21.4 46.4 32.1
Asian 27 20.5 23.1 56.4
Hispanic 55 27.3 20.0 52.7
County Providing Services
Milwaukee 488 28.5 27.3 44.3
All Other Counties 1331 18.0 23.5 58.4
Adjudicated Status
CHIPS-Abuse/Neglect 374 22.2 25.1 52.7
CHIPS-Other 428 21.3 22.0 56.8
Delinquent 893 19.0 26.5 54.4
JIPS-Status Offender 60 28.3 16.7 55.0
Voluntary Placement 64 29.7 17.2 53.1
Placement Type
Foster Home 1041 18.2 23.7 58.1
Group Home 413 20.8 25.2 54.0
Child Caring Institution 365 28.8 26.0 45.2
Discharge Outcome
Reunification, Relative Placement or Adoption 948 19.2 23.8 57.0
Aged Out or Discharged to Independent Living 645 22.8 20.9 56.3
Ran Away 108 22.2 37.0 40.7
Transferred to Other State Institution 118 22.9 38.1 39.0

Earnings: Descriptive Statistics

Table 5A shows the total earnings during the first eight quarters after discharge from care among those former foster youth who were employed in at least one quarter. On average, former foster youth earned substantially less during their first eight quarters post-discharge than a full-time (i.e., 40 hours per week) minimum wage worker would have earned over the same period of time.(13)

 

Table 5A
Total Earnings During the First Eight Post-Exit Quarters
(If Employed in at Least One Quarter)
N 1439
Median 4478
Mean 7094
Standard Deviation 8138

Table 5B presents these same earnings data broken down by gender, race/ethnicity, county (i.e., Milwaukee vs non-Milwaukee), adjudicated status, placement type and discharge outcome. Median total earnings were generally higher (1) among female former foster youth than among male former foster youth; and (2) among non-Milwaukee former foster youth than among Milwaukee former foster youth. However, there was relatively little difference by either gender or county in terms of mean total earnings. Both median and mean total earnings were higher (1) among Asian and White former foster youth than among African American, Native American, or Hispanic former foster youth; (2) among former foster youth who had been placed as status offenders (JIPS) than among former foster youth placed for other reasons; (3) among former foster youth who had been discharged from foster homes than among former foster youth discharged from group homes or child caring institutions; and (4) among former foster youth who had aged out or been discharged to independent living or who had been reunified, placed with relatives or adopted than among former foster youth who had run away from care or who had been transferred to a state institution.

 

Table 5B
Total Earnings During the First Eight Post-Exit Quarters
(If Employed in at Least One Quarter)
  N Median Mean Standard Deviation
Gender
Female 603 4750 7106 8032
Male 836 4185 7086 8218
Race/Ethnicity
African American 283 2497 5631 8239
White 1041 5134 7562 8039
Native American 44 1890 5655 11059
Asian 31 7481 8146 6928
Hispanic 40 4185 6042 5507
County Providing Services
Milwaukee 349 3713 6814 8436
All Other Counties 1090 4734 7184 8042
Adjudicated Status
CHIPS-Abuse/Neglect 291 4964 7227 7375
CHIPS-Other 337 4585 7331 8663
Delinquent 723 3998 6805 8200
JIPS-Status Offender 43 7526 9600 9115
Voluntary Placement 45 4393 6725 6476
Placement Type
Foster Home 852 5098 7573 8017
Group Home 327 3876 6549 3213
Child Caring Institution 260 3203 6213 8498
Discharge Outcome
Reunification, Relative Placement or Adoption 766 4493 6888 8095
Aged Out or Discharged to Independent Living 498 5544 8418 8591
Ran Away 84 2031 3923 4109
Transferred to Other State Institution 91 2023 4520 7202

Employment and Earnings: Multivariate Analysis

Employment

We estimated a logistic regression model in which employment during the first eight quarters after discharge from out-of-home care was regressed on a set of covariates that included both demographic characteristics of the former foster youth (e.g., gender, race/ethnicity) and measures related to their experiences in out-of-home care (e.g., placement type, number of placements, length of stay in care).(14) Table 6 shows the parameter estimates and odds ratios for this model based on the 1995 through 1997 exit cohorts of former foster youth. An odds ratio greater than one means that, controlling for all of the other covariates in the model, the group with that characteristic is more likely to have been employed than the comparison group, whereas an odds ratio less than one means that the group with that characteristic is less likely to have been employed than the comparison group. Former foster youth who were African American or Hispanic were significantly less likely to have been employed than those who were White. Former foster youth who had been placed as either children or juveniles in need of protection were significantly more likely to have been employed than those who had been adjudicated delinquent. Former foster youth who had been discharged from foster homes were more likely to have been employed than those who had been discharged from child caring institutions. And finally, for every additional month former foster youth had been in care, their odds of having been employed decreased by slightly less than 1%.

 

Table 6
Logistic Model Estimating the Likelihood of Employment
During the First Eight Post-Exit Quarters +
Covariates Beta Odds Ratio
Intercept 0.71  
Gender
Male
Female
-.05 0.950
Race/Ethnicity
African American
Native American
Asian
Hispanic
White
-1.03***
-0.33
-0.37
-0.79*
0.358***
0.721
0.688
0.455*
County Providing Services
Milwaukee
All Other Counties
0.10 1.105
Adjudicated Status
CHIPS-Abuse/Neglect
CHIPS-Other
JIPS-Status Offender
Voluntary Placement
Delinquent
0.82*
0.70*
0.91**
-0.24
2.266*
2.013*
2.493**
0.783
Placement Type at Exit
Foster Home
Group Home
Child Caring Institution
0.87***
0.31
2.389***
1.362
Discharge Outcome
Aged Out/Discharged to Independent Living
Runaway
Transfer to State Institution
Reunification, Relative Placement or Adoption
-0.09
-0.12
-0.21
0.911
0.877
0.809
Total Number of Episodes -0.01 0.986
Total Number of Placements -0.01 0.995
Total Number of Months in Care -0.01*** 0.991***
Significance Levels: * p .05 ** p .01 *** p .001

+ Excluded categories italicized.

Earnings

We estimated an OLS regression model in which total earnings during the first eight quarters after discharge from out-of-home care were regressed on the same set of covariates that were used in the employment model. Table 7 shows the parameter estimates for this model based on the 1995 through 1997 exit cohorts of former foster youth who had been employed in at least one of the first eight post-exit quarters. Former foster youth who were African American had significantly lower average total earnings than those who were White. Former foster youth who had been placed voluntarily had significantly lower average total earnings than those who had been adjudicated delinquent. Former foster youth discharged from foster homes had significantly higher average total earnings than those discharged from child caring institutions. And finally, former foster youth who had aged out of care or been discharged to independent living had significantly higher average total earnings than those discharged to family, where former foster youth who either ran away or were transferred to another state institution had significantly lower average total earnings.

 

Table 7
OLS Model Estimating Total Earnings During the First Eight Post-Exit Quarters
(If Employed in at Least One Quarter) +
Covariates Beta
Intercept 8876.91
Gender
Male
Female
394.26
Race/Ethnicity
African American
Native American
Asian
Hispanic
White
-2958.25***
-1264.71
860.08
-1595.23
County Providing Services
Milwaukee
All Other Counties
431.17
Adjudicated Status
CHIPS-Abuse/Neglect
CHIPS-Other
JIPS-Status Offender
Voluntary Placement
Delinquent
-2306.11
-2018.79
-1754.45
-3463.71*
Placement Type at Exit
Foster Home
Group Home
Child Caring Institution
1448.92*
524.45
Discharge Outcome
Aged Out or Discharged to Independent Living
Ran Away
Transferred to Other State Institution
Reunification, Relative Placement or Adoption
1660.47***
-2501.32**
-2191.70*
Total Number of Episodes -193.31
Total Number of Placements -301.63
Total Number of Months in Care 12.83
Model Adjusted R-Square .0443
Significance Levels: * p .05 ** p .01 *** p .001

+ Excluded categories italicized.

As noted above, this report is based on UI data for the years 1995 through 1999. Although our analysis thus far has focused on employment and earnings for the first eight quarters post-exit, these data also allow us to examine whether earnings prior to discharge might account for a substantial proportion of the variance in post-discharge earnings, controlling for the demographic characteristics and out-of-home cares experiences of former foster youth. We examined this possibility using data for the 1996 through 1998 exit cohorts. Table 8A shows pre- and post-discharge employment and Table 8B shows total earnings for these former foster youth.

 

Table 8A
Employment During Last Four Pre-Exit and
First Four Post-Exit Quarters
  Pre-Exit Post-Exit
Number of Quarters Employed N Percent N Percent
0 726 39.4 555 30.1
1 292 15.9 264 14.3
2 271 14.7 262 14.2
3 227 12.3 290 15.7
4 326 17.7 471 25.6
  1842 100.0 1842 100.0

 

Table 8B
Total Earnings During Last Four Pre-Exit and
First Four Post-Exit Quarters
(If Employed in at Least One Quarter)
  Pre-Exit Post-Exit
N 1116 1287
Median 1215 2031
Mean 2070 3274
Standard Deviation 2788 3724

We estimated two OLS models in which total earnings for the first four quarters post-discharge were regressed on the same set of covariates used in the previous models. However, the second model also included total earnings for the four quarters immediately prior to discharge from care. Table 9 shows the parameter estimates for these two models based on the 1996 through 1998 exit cohorts. The parameter estimates for Model 1 indicate (1) that African American former foster youth had significantly lower total earnings than White former foster youth; (2) that former foster youth who had been reunified, placed with relatives or adopted had significantly lower total earnings than former foster youth who had aged out of care or been discharged to independent living, but significantly higher total earnings than former foster youth who had run away from care; and (3) that for every additional placement former foster youth experienced, total earnings decreased by $175. Despite the statistical significance of these covariates, this model explained little of the variance in total earnings.

The parameter estimates for Model 2 indicate (1) that every dollar in pre-exit earnings was associated with a statistically significant increase of 0.84 dollars in post-exit earnings; and (2) that even after controlling for pre-exit earnings, African American former foster youth had significantly lower total earnings than White former foster youth, and former foster youth who had been reunified, placed with relatives or adopted had significantly lower total earnings than former foster youth who had aged out of care or been discharged to independent living. Moreover, addition of pre-exit earnings markedly improved the model's explanatory power.

 

Table 9
Models Estimating Earnings During the First Four Post-Exit Quarters +
Covariates Beta
  Model 1 Model 2
Intercept 2725.44 2320.49
Gender
Male
Female
-47.69 13.12
Race/Ethnicity
African American
Native American
Asian
Hispanic
White
-1331.09***
35.85
591.96
102.09
-712.73*
568.71
-168.98
134.14
County Providing Services
Milwaukee
All Other Counties
359.59 148.87
Adjudicated Status
CHIPS-Abuse/Neglect
CHIPS-Other
JIPS-Status Offender
Voluntary Placement
Delinquent
244.47
-12.34
52.02
454.27
196.87
-116.57
-171.10
-18.29
Placement Type at Exit
Foster Home
Group Home
Child Caring Institution
551.57
121.05
-344.81
-440.26
Discharge Outcome
Aged Out or Discharged to Independent Living
Ran Away
Transferred to Other State Institution
Reunification, Relative Placement or Adoption
1393.62***
-895.96*
-64.02
708.07**
-598.60
97.62
Total Number of Episodes 75.30 -67.30
Total Number of Placements -175.33* -32.36
Total Number of Months in Care 5.51 -0.93
Earnings During Four Quarters Pre-Exit   0.84***
Model Adjusted R-Square .0580 .3696
Significance Levels: * p .05 ** p .01 *** p .001

+ Excluded categories italicized.

[ Go to Contents ]

Section 3:
Public Assistance Utilization

Our analysis of public assistance utilization by former foster youth is based on state administrative data from the Client Assistance for Re-employment and Economic Support (CARES) data collection system for the period January 1995 through June 2000. CARES contains monthly, client-specific information about participation in public assistance programs, including AFDC/TANF and Food Stamps.(15) To be consistent with our analysis of the employment and earnings data, we focus on public assistance utilization during the first eight quarters post-exit for the 1995 through 1997 exit cohorts of former foster youth.

We use two sets of outcome measures in our analysis of these CARES data. The first set of outcome measures contains three dichotomous variables indicating whether former foster youth received (1) AFDC/TANF cash assistance, (2) Food Stamps, and (3) AFDC/TANF cash assistance and/or Food Stamps during their first eight quarters after discharge from care.(16) Former foster youth were counted as public assistance recipients if and only if two conditions were met. First, the former foster youth received a non-zero AFDC/TANF cash grant and/or Food Stamps in at least one month during the first eight post-exit quarters.(17) Second, the former foster youth was coded as the "primary person" or case head. In other words, the former foster youth could not be dependents on the AFDC/TANF or Food Stamp case of a parent or other individual.(18)

The second set of outcome measures also contains three variables: (1) the total amount of AFDC/TANF cash assistance former foster youth received during the first eight post-exit quarters; (2) the total amount of Food Stamps former foster youth received during the first eight post-exit quarters; and (3) the total amount of AFDC/TANF cash assistance and/or Food Stamps former foster youth received during the first eight post-exit quarters. We computed the first two measures by summing monthly AFDC/TANF cash assistance and/or Food Stamp benefits, and the third measure by summing the first two.

An obvious limitation of this approach is that CARES does not include information about other means-tested government benefit programs which might be relevant to our sample of former foster youth, particularly SSI.(19) Another potential limitation is the absence of information about General Relief.(20) Individuals who are unemployable due to a temporary or permanent disability (including alcohol or other drug addictions) may be eligible for General Relief. However, only 31 of Wisconsin's 72 counties operate a General Relief program, and both eligibility criteria and benefit schedules vary across counties.(21) Finally, although CARES does include information about eligibility for Medical Assistance, the state's Medicaid program, information as to whether Medicaid payments were made for any services are recorded in another data system. Hence, one cannot determine from CARES whether an individual "received" Medicaid.(22)

Receipt of Public Assistance: Descriptive Statistics

We begin with some descriptive statistics on the receipt of public assistance by the 1995 through 1997 exit cohorts of former foster youth. Table 10A shows the percentage of these former foster youth who received (1) AFDC/TANF cash assistance; (2) Food Stamps; and (3) AFDC/TANF cash assistance and/or Food Stamps in at least one month during the first eight quarters after they were discharged from care. These data suggest that only a small minority of former foster youth received public assistance within that 24-month period.

 

Table 10A
Receipt of Public Assistance During the First Eight Post-Exit Quarters
  Received AFDC/TANF
Cash Assistance
Received Food Stamps Received Cash Assistance
and/or Food Stamps
N Percent N Percent N Percent
NO 1782 98.0 1720 94.6 1717 94.4
YES 37 2.0 99 5.4 102 5.6

Table 10B presents these same data broken down by gender, race/ethnicity, county (i.e., Milwaukee vs non-Milwaukee), adjudicated status, placement type and discharge outcome. Although relatively few former foster youth received public assistance, receipt of AFDC/TANF cash assistance and/or Food Stamps was more likely (1) among female former foster youth than among male former foster youth; (2) among African American former foster youth than among non-African American former foster youth; (3) among Milwaukee former foster youth than among non-Milwaukee former foster youth; (4) among former foster youth who had been placed in care as children in need of protection (CHIPS) than among former foster youth who had been placed for other reasons; (5) among former foster youth who were discharged from foster homes than among former foster youth discharged from group homes or child caring institutions; and (6) among former foster youth who ran away from care or who aged out or were discharged to independent living than among former foster youth who were reunified, placed with relatives or adopted or who were transferred to a state institution.(23)

 

Table 10B
Receipt of Public Assistance During the First Eight Post-Exit Quarters
  N Percent Received AFDC/TANF Cash Assistance Percent Received Food Stamps Percent Received Cash Assistance and/or Food Stamps
Gender
Female 751 4.9 11.6 12.0
Male 1068 0.0 1.1 1.1
Race/Ethnicity
African American 423 7.3 12.1 12.5
White 1246 0.4 3.5 3.6
Native American 56 0.0 0.0 0.0
Asian 39 0.0 2.6 2.6
Hispanic 55 1.8 5.5 5.5
County Providing Services
Milwaukee 488 5.7 9.8 9.8
All Other Counties 1331 0.7 3.8 4.1
Adjudicated Status
CHIPS-Abuse/Neglect 374 4.5 10.7 10.7
CHIPS-Other 428 3.7 7.7 8.4
Delinquent 893 0.5 2.2 2.2
JIPS-Status Offender 60 0.0 5.0 5.0
Voluntary Placement 64 0.0 4.7 4.7
Placement Type
Foster Home 1041 3.1 7.3 7.5
Group Home 413 0.5 3.1 3.4
Child Caring Institution 365 0.8 2.7 2.7
Discharge Outcome
Reunification, Relative Placement or Adoption 948 1.9 4.3 4.5
Aged Out or Discharged to Independent Living 645 2.6 7.1 7.3
Ran Away 108 1.8 9.3 9.3
Transferred to Other State Institution 118 0.0 1.7 1.7

Table 11A shows the total amount of (1) AFDC/TANF cash assistance; (2) Food Stamps; and (3) AFDC/TANF cash assistance and/or Food Stamps that former foster youth received during the first eight quarters after they were discharged from care among those former foster youth who received AFDC/TANF cash assistance and/or Food Stamps in at least one month during those eight quarters.

 

Table 11A
Total Public Assistance Benefits During the First Eight Post-Exit Quarters
(If Benefits Received in at Least One Quarter)
  AFDC/TANF Cash Assistance Food Stamps Cash Assistance and/or Food Stamps
N 37 99 102
Median 3873 668 797
Mean 4908 1256 3000
Standard Deviation 3958 1411 4404

Table 11B shows the total amount of AFDC/TANF cash assistance and/or Food Stamps former foster youth received during the first eight quarters after they were discharged from care broken down by gender, race/ethnicity, county (i.e., Milwaukee vs non-Milwaukee), adjudicated status, placement type and discharge outcome. Conditional upon receiving AFDC/TANF cash assistance and/or Food Stamps, there was considerable variation across different groups of former foster youth in terms of the total amount of AFDC/TANF cash assistance and/or Food Stamps that they received.(24) AFDC/TANF cash assistance and/or Food Stamps totals tended to be higher among female former foster youth than among male former foster youth; among African-American former foster youth than among White former foster youth; among Milwaukee former foster youth than among non-Milwaukee former foster youth; among former foster youth placed as children in need of protection(CHIPS) than among former foster youth who had been adjudicated delinquent; and among former foster youth who were discharged from foster homes than among former foster youth discharged from either group homes or child caring institutions. Although there was little difference in median total AFDC/TANF cash assistance and/or Food Stamps between former foster youth who had been discharged to live with family and former foster youth who had aged out of care or been discharged to independent living, the mean total was substantially higher among the latter, and both groups tended to have received higher amounts of AFDC/TANF cash assistance and/or Food Stamps than former foster youth who exited care by running away.

 

Table 11B
Total Public Assistance Benefits During the First Eight Post-Exit Quarters
(If Benefits Received in at Least One Quarter)
  N Median Mean Standard
Deviation
Gender
Female 90 999 3345 4581
Male 12 345 407 197
Race/Ethnicity
African American 53 1592 4861 5331
White 45 460 836 1093
Native American 0 0 0 0
Asian 1 51 51 0
Hispanic 3 1829 3557 3510
County Providing Services
Milwaukee 48 999 3345 4581
All Other Counties 54 534 1148 1763
Adjudicated Status
CHIPS-Abuse/Neglect 40 956 3756 4999
CHIPS-Other 36 1014 3212 4457
Delinquent 20 516 1735 3285
JIPS-Status Offender 3 875 884 627
Voluntary Placement 3 767 915 556
Placement Type
Foster Home 78 1042 3557 4791
Group Home 14 392 943 1918
Child Caring Institution 10 571 1537 2075
Discharge Outcome
Reunification, Relative Placement or Adoption 43 876 2391 3465
Aged Out or Discharged to Independent Living 47 875 4045 5307
Ran Away 10 599 1266 2256
Transferred to Other State Institution 2 182 182 185

Thus far, our analysis has been limited to public assistance utilization during the first eight quarters after former foster youth had been discharged from care. Table 12A shows what happens when the observation period for the 1995 to 1997 exit cohorts is extended beyond the first eight quarters through June 2000. Perhaps not surprisingly, extending the observation period increases the percentage of former foster youth who received AFDC/TANF cash assistance and/or Food Stamps after exiting out-of-home care.

 

Table 12A
Receipt of Public Assistance Through June 2000
  Received AFDC/TANF
Cash Assistance
Received Food Stamps Received Cash Assistance
and/or Food Stamps
  Percent N Percent N Percent N
NO 88.3 1607 74.5 1355 74.0 1346
YES 11.7 212 25.5 464 26.0 473

Table 12B shows the same data broken down by gender, race/ethnicity, county (i.e., Milwaukee vs non-Milwaukee), adjudicated status, placement type and discharge outcome. The between-group differences are similar to those found in Table 11B, suggesting that differences in the likelihood of utilizing public assistance tend to be maintained over time.(25)

 

Table 12B
Receipt of Public Assistance Through June 2000
  N Percent Received
AFDC/TANF
Cash Assistance
Percent Received
Food Stamps
Percent Received
Cash Assistance
and/or Food Stamps
Gender
Female 751 27.2 48.3 49.3
Male 1068 0.4 9.5 9.6
Race/Ethnicity
African American 433 23.4 39.2 40.2
White 1246 7.5 21.4 21.8
Native American 56 8.9 19.6 21.4
Asian 39 5.1 15.4 15.4
Hispanic 55 15.5 25.5 25.5
County Providing Services
Milwaukee 488 20.7 35.7 36.0
All Other Counties 1331 8.0 21.8 22.3
Adjudicated Status
CHIPS-Abuse/Neglect 374 19.3 37.2 37.7
CHIPS-Other 428 15.9 32.2 33.2
Delinquent 893 6.6 18.4 18.7
JIPS-Status Offender 60 8.3 18.3 18.3
Voluntary Placement 64 6.3 18.8 18.8
Placement Type
Foster Home 1041 15.6 30.5 31.2
Group Home 413 6.1 18.6 18.6
Child Caring Institution 365 5.8 19.2 19.5
Discharge Outcome
Reunified, Placed with Relatives or Adopted 948 10.0 24.4 25.0
Aged Out or Discharged to Independent Living 645 14.4 29.3 29.6
Ran Away 108 14.8 29.6 30.6
Transferred to Other State Institution 118 3.4 10.2 10.2

Finally, although we do not yet have Unemployment Insurance data for the early exit cohorts in our larger sample (i.e., those former youth who exited care in 1992, 1993 and 1994), we do have CARES data for these early cohorts. Table 13 shows the percentage of former foster youth in each of those exit cohorts who received AFDC/TANF cash assistance and/or Food Stamp at any time through June 2000. In general, as years since discharge increases, so too does the likelihood of ever having received AFDC/TANF cash assistance and/or Food Stamps. While this relationship between years since discharge and the likelihood of public assistance utilization might have been expected, the implications of these data are disconcerting. In particular, they suggest that six to eight years after being discharged from care, a third or more former foster youth will have received AFDC/TANF cash assistance and/or Food Stamps.

 

Table 13
Receipt of Public Assistance from January 1992 Through June 2000 by Exit Cohort
Exit Cohort N Percent Received
AFDC/TANF
Cash Assistance
Percent Received
Food Stamps
Percent Received
Cash Assistance and/or Food Stamps
1992  644  16.0 36.0 36.7
1993 639 18.8 39.3 39.4
1994 595 17.8 33.1 33.3
1995 596 13.3 28.7 28.7
1996 604 12.4 25.8 26.8
1997 619 8.7 22.1 22.6
1998 619 7.6 19.9 20.5

Receipt of Public Assistance: Multivariate Analysis

We estimated a logistic regression model in which receipt of AFDC/TANF cash assistance and/or Food Stamps during the first eight quarters after discharge from out-of-home care was regressed on a set of covariates representing the demographic characteristics (e.g., gender, race/ethnicity) and out-of-home care experiences (e.g., placement type, number of placements, length of stay in care) of the former foster youth.(26) We limited the analysis to female former foster because so few male former foster youth were AFDC/TANF cash assistance and/or Food Stamp recipients. We also excluded the relatively small number of former foster youth who had been placed voluntarily. shows the parameter estimates and odds ratios for this model. An odds ratio greater than one means that, controlling for all of the other covariates in the model, the group with that characteristic is more likely to have received AFDC/TANF Table 14 cash assistance and/or Food Stamps than the comparison group, whereas an odds ratio less than one means that the group with that characteristic is less likely to have received AFDC/TANF cash assistance and/or Food Stamp than the comparison group. Former foster youth who were African-American were significantly more likely to have received AFDC/TANF cash assistance and/or Food Stamps than former foster youth who were White, and former foster youth who were placed in care as children in need of protection (CHIPS) for reasons other than abuse or neglect were significantly more likely to have received AFDC/TANF cash assistance and/or Food Stamps than former foster youth who were adjudicated delinquents or status offenders.

 

Table 14
Logit Model Predicting Receipt of Public Assistance
by Female Former Foster Youth During the First Eight Post-Exit Quarters
+
(N = 719)
Covariates Beta Odds Ratio
Intercept -3.714  
Race/Ethnicity
African American
Other
White
1.416***
-0.003
4.122***
0.977
County Providing Services
Milwaukee
All Other Counties
0.117 1.125
Adjudicated Status
CHIPS-Abuse/Neglect
CHIPS-Other
Delinquent or JIPS
0.547
0.683*
1.728
1.979*
Placement Type at Exit
Foster Home
Group Home
Child Caring Institution
0.420
0.343
1.521
1.409
Discharge Outcome
Reunification, Relative Placement or Adoption
Aged Out/Discharged to Independent Living
Runaway or Transfer to State Institution
-0.220
0.126
0.802
1.134
Total Number of Episodes 0.271 1.312
Total Number of Placements -0.037 0.964
Total Number of Months in Care -0.000 1.000
Significance Levels: * p .05 ** p .01 *** p .001

+ Excluded categories italicized.

We also estimated an OLS regression model in which the total amount of AFDC/TANF cash assistance and/or Food Stamps received during the first eight quarters after discharge from out-of-home care was regressed on the same set of covariates that were used in the preceding model. This analysis is limited to the 87 female former foster who had been AFDC/TANF cash assistance and/or Food Stamp recipients (and had not been placed in care voluntarily). Table 15 shows the parameter estimates for this model. The total amount of AFDC/TANF cash assistance and/or Food Stamp received by Milwaukee former foster youth was significantly higher than the total amount received by non-Milwaukee former foster youth. And while the total amount of AFDC/TANF cash assistance and/or Food Stamp received by African American former foster youth was higher than the total amount received by White former foster youth, the coefficient was only marginally significant, perhaps because of the small sample size.

 

Table 15
OLS Model Estimating Total Public Assistance Benefits Received by
Female Former Foster Youth During the First Eight Post-Exit Quarters
(If Benefits Received in at Least One Quarter) +

(N = 87)
Covariates Beta
Intercept -2771
Race/Ethnicity
African American
Other
White
2057
2552
County Providing Services
Milwaukee
All Other Counties
3852**
Adjudicated Status
CHIPS-Abuse/Neglect
CHIPS-Other
Delinquent or JIPS
-761
-625
Placement Type at Exit
Foster Home
Group Home
Child Caring Institution
3326
2800
Discharge Outcome
Reunification, Relative Placement or Adoption
Aged Out/Discharged to Independent Living
Runaway or Transfer to State Institution
874
-1055
Total Number of Episodes -918
Total Number of Placements 722
Total Number of Months in Care 4.49
Model Adjusted R-Square .2782
Significance Levels: * p .05 ** p .01 *** p .001

+ Excluded categories italicized.

[ Go to Contents ]

Section Four:
Total Income From Earnings and Public Assistance

Thus far, we have examined both total earnings from "covered" employment and total AFDC/TANF cash assistance and/or Food Stamps during the first eight quarters after former foster youth were discharged from care. In this section, we include both of these sources of income in our analysis.

We use three outcome measures in our analysis of these data. The first measure is total income from earnings and public assistance during the first eight post-exit quarters. We compute this measure by summing the total earnings and total AFDC/TANF cash assistance and/or Food Stamps that were used in the earlier analyses. Although we refer to this measure below as "total income," it should actually be thought of as a lower bound because there are several potential sources of income that it does not take into account, including benefits from other government programs (e.g., SSI), earnings from employment that is not "covered," or money from family and friends. The second and third measures are the percentage of total income from earnings and the percentage of total income from AFDC/TANF cash assistance and/or Food Stamps, respectively. We compute the second measure by dividing earnings by total income, and the third measure by dividing AFDC/TANF cash assistance and/or Food Stamps by total income.

Total Income: Descriptive Statistics

Table 16A shows the total income from earnings and AFDC/TANF cash assistance and/or Food Stamps during the first eight quarters after discharge among the 1995 through 1997 exit cohorts of former foster youth for (1) only those former foster youth whose total income was non-zero, and (2) all former foster youth, including those whose total income was zero. Not surprisingly, both median and mean total income are considerably higher when the analysis is limited to those with non-zero income.

 

Total 16A
Total Income from Earnings and AFDC/TANF Cash Assistance and/or Food Stamps
During the First Eight Post-Exit Quarters
Former Foster Youth with Nonzero Income All Former Foster Youth
N Median Mean Standard Deviation N Median Mean Standard Deviation
1462 4602 7192 8148 1819 2848 5781 7843

Table 16B shows the same data broken down by gender, race/ethnicity, county (i.e., Milwaukee vs non-Milwaukee), adjudicated status, placement type and discharge outcome. When the analysis is limited to former foster youth with non-zero total income, total income tends to be higher (1) among female than among male former foster youth; (2) among White or Asian than among African American, Native American or Hispanic former foster youth; (3) among non-Milwaukee than among Milwaukee former foster youth; (4) among former foster youth who had been adjudicated status offenders (JIPS) or children in need of protection (CHIPS) than among former foster youth who had been adjudicated delinquent or placed voluntarily; (5) among former foster youth who had been discharged from foster homes than among former foster youth discharged from group homes or institutions; and (6) among former foster youth who aged out or were discharged to independent living or former foster youth who were reunified, placed with relatives or adopted than among former foster youth who ran away from care or who were transferred to a state institution. When the analysis is extended to include former foster youth with zero total income, the relative differences in total income are essentially the same.(27)

 

Total 16B
Total Income from Earnings and AFDC/TANF Cash Assistance and/or Food Stamps
During the First Eight Post-Exit Quarters
  Former Foster Youth with Nonzero Total Income All Former Foster Youth
  N Median Mean SD N Median Mean SD
Gender
Female 621 5006 7384 8066 751 3269 6106 7849
Male 841 4141 7050 8210 1068 2565 5552 7835
Race
African American 297 2791 6233 8435 423 1073 4376 7619
White 1050 5108 7533 8020 1246 3786 6348 7856
Native American 44 1890 5655 11059 56 1402 4443 10054
Asian 31 7532 8148 6928 39 3689 6476 7000
Hispanic 40 4611 6309 5564 55 2373 4589 5514
County Providing Services
Milwaukee 365 3982 7183 8490 488 1580 5373 7977
All Other Counties 1097 4745 7195 8035 1331 3191 5930 7791
Adjudicated Status
CHIPS-Abuse/Neglect 301 5103 7487 7452 374 3213 6025 7314
CHIPS-Other 343 4903 7540 8687 428 3014 6042 8337
Delinquent 727 4002 6815 8186 893 2670 5548 7847
JIPS-Status Offender 45 6805 9233 9074 60 3473 6924 8812
Placement Type
Foster Home 868 5309 7753 8049 1041 3786 6464 7896
Group Home 330 3829 6530 8064 413 2348 5217 7668
Child Caring Institution 264 3203 6177 8441 365 1478 4468 7690
Discharge Outcome
Reunification, Relative Placement or Adoption 776 4619 6931 8056 948 3011 5674 7762
Aged Out or Discharged to Independent Living 509 5833 8609 8647 645 3664 6794 8446
Ran Away 86 2139 3979 4109 108 1484 3169 4001
Transferred to Other State Institution 91 2023 4524 7206 118 1159 3489 6602

Total Income: Multivariate Analysis

We estimated two OLS regression models in which total income from earnings and AFDC/TANF cash assistance and/or Food Stamps during the first eight quarters after discharge from care was regressed on the same set of covariates that was used in the earnings regression models. The only difference between the two models is that the first is estimated for the 1462 former foster youth whose total income was non-zero while the second is estimated for all 1819 former foster youth, including those whose total income was zero. Table 17 shows the parameter estimates for these models. Looking first at the model estimated for the 1462 former foster youth with non-zero total income, total income was significantly lower among former foster youth (1) who were African American compared to those who were White; (2) who had been placed voluntarily rather than those who had been adjudicated delinquent; and (3) who had run away from care or been transferred to a state institution rather than those who had been reunified, placed with relatives or adopted. Conversely, total income was significantly higher among former foster youth (1) who were discharged from foster homes compared to those who had been discharged from child caring institutions; and (2) who had aged out of care or been discharged to independent living compared to those who had been reunified, placed with relatives or adopted. The results are similar when the model is estimated for all 1819 former foster youth, with the exception that the coefficient for voluntary placement is only marginally significant.

 

Table 17
OLS Model Estimating Total Income from Earnings and AFDC/TANF Cash Assistance
and/or Food Stamps During the First Eight Post-Exit Quarters
  Model 1
N = 1461
Model 2
N = 1819
Covariates Beta
Intercept 8522 6366
Gender
Male
Female
111 -129
Race/Ethnicity
African American
Native American
Asian
Hispanic
White
-2379***
-1255
928
-1310
-2878***
-1273
187
-2006
County Providing Services
Milwaukee
All Other Counties
547 754
Adjudicated Status
CHIPS-Abuse/Neglect
CHIPS-Other
JIPS-Status Offender
Voluntary Placement
Delinquent
-2095
-1665
-1422
-3243
-893
-843
-344
-2804*
Placement Type at Exit
Foster Home
Group Home
Child Caring Institution
1607*
620
2191***
887
Discharge Outcome
Aged Out or Discharged to Independent Living
Ran Away
Transferred to Other State Institution
Reunification, Relative Placement or Adoption
1626**
-2575**
-2210*
1144**
-2154**
-2015**
Total Number of Episodes -228 -180
Total Number of Placements -259 -211
Total Number of Months in Care 13.8 -0.726
Model Adjusted R-Square .0400 .0433
Significance Levels: * p .05 ** p .01 *** p .001

+ Excluded categories italicized.

Table 18A shows the percentage of total income from earnings and the percentage of total income from AFDC/TANF cash assistance and/or Food Stamps during the first eight quarters after discharge among those former foster youth whose total income was nonzero. The most notable finding is that the median percentage of total income from earnings is 100% and the median percentage of total income from AFDC/TANF cash assistance and/or Food Stamps is 0%. This reflects the fact that while 79.1% of the former foster had at least some earnings from employment, only 5.6% received AFDC/TANF cash assistance and/or Food Stamps.

 

Table 18A
Percentage of Total Income from Earnings and AFDC/TANF Cash Assistance
and/or Food Stamps During the First Eight Post-Exit Quarters
(If Total Income > 0)

(N = 1462)
Percent of Income from Earnings Percent of Income from Cash Assistance and/or Food Stamps
Median Mean Median Mean
100.0 96.6 0.0 3.4

Table 18B shows the same data broken down by gender, race/ethnicity, county (i.e., Milwaukee vs non-Milwaukee), adjudicated status, placement type and discharge outcome. For all groups, the median percentage of total income from earnings is 100% and the median percentage of total income from AFDC/TANF cash assistance and/or Food Stamps is 0%. The mean percentage of total income from cash assistance and/or Food Stamps tends to be higher (1) among female than among male former foster youth; (2) among African American than among White, Asian, Native American or Hispanic former foster youth; (3) among Milwaukee than among non-Milwaukee former foster youth; (4) among former foster youth who had been adjudicated children in need of protection (CHIPS) or status offenders (JIPS) than among former foster youth who had been adjudicated delinquent or placed voluntarily; (5) among former foster youth who had been discharged from foster homes than among former foster youth discharged from group homes or institutions; and (6) among former foster youth who aged out or were discharged to independent living or former foster youth who were reunified, placed with relatives or adopted than among former foster youth who ran away from care or who were transferred to a state institution.

 

Table 18B
Percentage of Total Income from Earnings and AFDC/TANF Cash Assistance
and/or Food Stamps During the First Eight Post-Exit Quarters
(If Total Income > 0)
  Percent of Total Income from Earnings Percent of Total Income from Cash Assistance and/or Food Stamps
  N Median Mean Median Mean
Gender
Female 621 100.0 93.0 0.0 7.0
Male 841 100.0 99.3 0.0 0.7
Race
African American 297 100.0 89.1 0.0 10.9
White 1050 100.0 98.5 0.0 1.5
Native American 44 100.0 100.0 0.0 0.0
Asian 31 100.0 99.9 0.0 0.02
Hispanic 40 100.0 97.3 0.0 2.8
County Providing Services
Milwaukee 365 100.0 91.4 0.0 8.6
All Other Counties 1050 100.0 98.3 0.0 1.7
Adjudicated Status
CHIPS-Abuse/Neglect 301 100.0 93.2 0.0 6.8
CHIPS-Other 343 100.0 95.3 0.0 4.7
Delinquent 727 100.0 98.7 0.0 1.3
JIPS-Status Offender 45 100.0 97.1 0.0 4.6
Voluntary Placement 46 100.0 95.4 0.0 2.9
Placement Type
Foster Home 868 100.0 95.7 0.0 4.3
Group Home 330 100.0 98.4 0.0 1.6
Child Caring Institution 264 100.0 92.3 0.0 0.8
Discharge Outcome
Reunification, Relative Placement or Adoption 776 100.0 97.1 0.0 3.0
Aged Out or Discharged to Independent Living 509 100.0 95.6 0.0 4.4
Ran Away 86 100.0 95.0 0.0 2.0
Transferred to Other State Institution 91 100.0 99.9 0.0 0.04

[ Go to Contents ]

Section 5:
Discussion and Implications

The results presented above are based on an analysis of employment, earnings, and public assistance data from state administrative records for a sample of former foster youth who were at least 17 years old when they were discharged from care. Consistent with earlier studies, our findings indicate that a significant percentage of the 1995 through 1997 exit cohorts of former foster youth were either not employed at all (21%) or only sporadically employed (24%) during the first two years (eight quarters) after they were discharged from Wisconsin's out-of-home care system. Finding and/or maintaining stable employment may not be as much of a problem for former foster youth as these results suggest. In particular, former foster youth are counted as not employed if they were employed in jobs that are not "covered" or if they were employed, but living in another state. However, finding and/or maintaining stable employment may also be more difficult for former foster youth than these results seem to imply because former foster youth were counted as employed in any quarter for which earnings were reported, regardless of the number of hours that they worked.

We found a number of factors that seem to be related to employment among our sample of former foster. Both our univariate and multivariate analysis suggest that non-White former foster youth were significantly less likely to have been employed than those who were White; that former foster youth who had been discharged from foster homes were more likely to have been employed than those who had been discharged from child caring institutions. Interestingly, although our univariate analysis seemed to indicate that former foster youth who were adjudicated status offenders (JIPS) were less likely to have been employed than former foster youth who were adjudicated delinquents or children in need of protection (CHIPS), our multivariate analysis shows that, controlling for other factors, former foster youth who had been adjudicated status offenders (or children in need of protection) were more likely to have been employed than those who had been adjudicated delinquent.

Again, consistent with what previous research has found, the earnings of former foster youth were low. In particular, during the first eight quarters after they were discharged from care, total earnings among the former foster youth in our sample were, on average, substantially lower than what a full-time minimum wage worker would have earned over the same period of time.

Total earnings were related to several factors. Both our univariate and multivariate analysis indicate that African American former foster youth earned significantly less than White former foster youth, that former foster youth who had been discharged from foster homes earned significantly more than those discharged from child caring institutions, and that former foster youth who had run away or been transferred to a state institution earned significantly less than those who had been reunified, placed with relatives or adopted. Although our univariate analysis showed little differences between the total earnings of former foster youth who had aged out or been discharged to independent living and those who had been reunified, placed with relatives or adopted, the results of our multivariate analysis suggest that former foster youth who had aged out of care or been discharged to independent living earned significantly more than those had been reunified, placed with relatives or adopted. Finally, although our univariate analysis indicated that former foster youth who had been placed as status offenders tended to earn more than those who had been for other reasons, there was no significant differences between the total earnings of former foster youth who had been adjudicated status offenders and those who had been adjudicated delinquent once other factors were controlled for.

Although we identified several factors related to post-discharge earnings, very little of the variance in total earnings was explained by the model that we estimated. However, the results of our analysis of total earnings among the 1996 through 1998 exit cohorts of former foster youth suggests that adding pre-discharge earnings to the model improves the model's explanatory power substantially. In other words, employment while in foster care appears to be by far the best predictor of post-discharge employment.

Our initial analysis of the public assistance data seemed to indicate 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 eight quarters after they were discharged from care. Nevertheless, both our univariate and multivariate analysis suggest that African-American former foster youth were more likely to have been AFDC/TANF cash assistance and/or Food Stamps recipients than White former foster youth, and that former foster youth who were placed in care as children in need of protection (CHIPS) for reasons other than abuse or neglect were more likely to have been AFDC/TANF cash assistance and/or Food Stamps recipients than former foster youth who were adjudicated delinquents or status offenders.

Both our both our univariate and multivariate analysis suggest that among the relatively small percentage of former foster youth who had been recipients of AFDC/TANF cash assistance and/or Food Stamps, Milwaukee former foster youth received larger amounts of AFDC/TANF cash assistance and/or Food Stamps during the first eight quarters after they were discharged from care than non-Milwaukee former foster youth. Although our univariate analysis suggests that AFDC/TANF cash assistance and/or Food Stamps totals tended to be higher among African American former foster youth than among White former foster youth, the difference was not statistically significant in the multivariate analysis.

Although less than 6% of the 1995 through 1997 exit cohorts of former foster youth had received AFDC/TANF cash assistance and/or Food Stamps at some point during the first eight quarters after they were discharged from care, 26% had received AFDC/TANF cash assistance and/or Food Stamps as of June 2000. However, despite this increase in the percentage of former foster youth who had received AFDC/TANF cash assistance and/or Food Stamps, differences in the likelihood of having been a recipient tend to be maintained over time. That is, the former foster youth who were most likely to have received AFDC/TANF cash assistance and/or Food Stamps at some point during the first eight quarters after they were discharged from care were also the most likely to have received AFDC/TANF cash assistance and/or Food Stamps at some point between the time they were discharged from care and June 2000.

This increase in the likelihood of having received AFDC/TANF cash assistance and/or Food Stamps is further illustrated by the percentage of former foster youth in each exit cohort from 1992 through 1998 who had received AFDC/TANF cash assistance and/or Food Stamps at some point between the time they exited care and June 2000. With the exception of former foster youth in the 1992 exit cohort, who were less likely to have been AFDC/TANF cash assistance and/or Food Stamp recipients than former foster youth in the 1993 exit cohort (but more likely to have been AFDC/TANF cash assistance and/or Food Stamp recipients than former foster youth in the 1994 exit cohort), the percentage of former foster who had received AFDC/TANF cash assistance and/or Food Stamps by June 2000 increased with each additional post-discharge year.

If this trend were to continue, at least one third of all the former foster youth in the 1995 through 1997 exit cohorts will have received AFDC/TANF cash assistance and/or Food Stamps within six to eight years after being discharged from care. Clearly, the data necessary to evaluate this possibility are not yet available. Nonetheless, it will be interesting to see whether the percentage of the former foster youth in these exit cohorts continues to increase as years since discharge increases. This question is especially intriguing given the dramatic reduction in both cash assistance and Food Stamp caseloads that Wisconsin has experienced in recent years.

By almost any standard, total income from earnings, AFDC/TANF cash assistance and Food Stamps for the first eight quarters after discharge from care was very low among our sample of 1819 former foster youth. Median total income and mean total income for this two year period were $2848 and $5781, respectively- well below the poverty line for a one-person family in 1995, which is the earliest year in which the former foster youth were discharged from care.(28) Even when the analysis is limited to the 1416 former foster youth who were employed and/or recipients of public assistance at some point during their first eight post-discharge quarters, both median total income ($4,602) and mean total income ($7192) for this two year period were still below the 1995 federal poverty line for a one-person family.(29)

Both our univariate and multivariate analyses seemed to indicate that total income was lower (1) among African American former foster youth than among White former foster youth and (2) among former foster youth who had run away from care or who had been transferred to a state institution than among those who had been reunified, placed with relatives or adopted, higher. Conversely, total income was higher (1) among former foster youth who were discharged from foster homes than among those who had been discharged from child caring institutions; and (2) who had aged out of care or been discharged to independent living than among those who had been reunified, placed with relatives or adopted. These differences exist regardless of whether former foster youth who were never employed and who were never recipients of public assistance during their first eight quarters after they were discharged from care are included in the analysis or not.

Earnings from employment were the primary source of income among the 1462 former foster youth in our sample for whom our measure of total income was greater than zero; on average, earnings from employment comprised 96.6% of total income whereas public assistance comprised only 3.4%. This reflects the fact that 1439 of these former foster youth had a least some earnings from employment while only 102 had received AFDC/TANF cash assistance and/or Food Stamps. These two groups were not mutually exclusive, however, only a small minority (N = 79) of the former foster youth in our sample had income from both earnings and public assistance during the first eight quarters after they were discharged from care.

As alluded to above, our measure of total income suffers from at least two important limitations. First, it only includes earnings from "covered" employment, AFDC/TANF cash assistance, and Food Stamps. The former foster youth in our sample may have had other sources of income during the first eight quarters after they were discharged from care, such as earnings from "uncovered" employment, other government programs such as SSI, or family and friends. Second, it is based on state administrative data. The former foster youth in our sample may have moved outside of Wisconsin during the first eight quarters after they were discharged from care. This might explain why our measure of total income was zero for nearly 20% (N = 357) of the former foster in our sample. It is also possible that some of these former foster youth were being supported by a parent, other relative, spouse or partner with whom they were living.

Perhaps the most serious limitation of our analysis is the fact that we restricted our sample to former foster youth for whom HSRS provided a valid Social Security number. As a result of this restriction, 31.2% of the 1992 through 1998 exit cohorts of former foster youth who were at least 17 years old when they were discharged from care were excluded from our sample. The exclusion of these former foster youth would not be a problem if it could be ascertained that there were no systematic differences between those former foster youth whose SSN was present and those whose SSN was missing. However, as our comparison made evident, the two groups were different in a number of respects; SSN's were more likely to be missing among former foster youth who are White, who were receiving services from non-Milwaukee counties, who were in care for a total of less than 12 months, and who were reunified than among former foster youth who are African American, who were receiving services from Milwaukee County, who were in care for 12 months or longer, and who aged out or were discharged to independent living.

While the implications of these differences for the results of our analysis are difficult to discern, there is some indication that the former foster youth missing SSN's are a somewhat more "advantaged" group. For example, White former foster youth were more likely to be have been employed and, if employed, had significantly higher earnings than African American former foster youth. They were also less likely to have received AFDC/TANF cash assistance and/or Food Stamps, and if they were recipients, received a significantly smaller amount of public assistance than African American former foster youth. There is a similar pattern of differences between former foster youth who were receiving services from non-Milwaukee counties and those who were receiving services from Milwaukee County, although only the earnings coefficient was statistically significant. However, the implications associated with differences in discharge outcome and length of stay in care are not as clear.(30)

Much of our analysis has focused on the variation in employment, earnings and public assistance receipt among former foster youth with different demographic characteristics and/or foster care experiences. The result of this analysis suggests that some of these demographic characteristics and/or foster care experiences are related to subsequent outcomes. Although we had initially planned to examine the outcomes of former foster youth during the pre- and post-TANF periods, it was not possible to do this given the limited number of pre-TANF years for which we had UI data.(31)

Rather than limiting our analysis to former foster youth who aged out of care or who were discharged to independent living, we broadened our sample to include all former foster youth who exited care between 1992 and 1998 and were at least 17 years old at the time they exited. We then compared the employment, earnings and public assistance receipt among former foster youth with different discharge outcomes.(32)

Our results suggest that former foster youth who had aged out of care or been discharged to independent living earned significantly more during the first eight quarters after they were discharged from care than those who had been reunified placed with relatives or adopted; as a result, their total income was also significantly higher. These findings are consistent with what one would expect if former foster youth who had aged out of care or been discharged to independent living were, in fact, living on their own and having to support themselves, while those who had been reunified placed with relatives or adopted were being supported by the families with whom they were living. We also found that former foster youth who had run away or been transferred to an institution were employed in significantly fewer quarters and earned significantly less during the first eight quarters after they were discharged from care than those who had been reunified, placed with relatives or adopted; again, their lower earnings were reflected in significantly lower total income as well.

Because the discharge outcomes of these former foster youth were not randomly assigned, our findings cannot be interpreted as evidence that the differences in employment, earnings and public assistance receipt that we observed were caused by the different discharge outcomes they experienced. Rather, a more likely explanation is that the discharge outcomes former foster youth experienced were endogenous to a combination of observable and unobservable individual, family, worker, and agency characteristics.

[ Go to Contents ]

References

Barth, R. P. (1990). On their own: The experiences of youth after foster care. Child and Adolescent Social Work Journal, 7, 419-440.

Cook, R., Fleishman, E., & Grimes, V. (1991). A national evaluation of title IV-E foster care independent living programs for youth: Phase 2. Final Report, Volume 1. Rockville, MD: Westat, Inc.

Courtney, M. E., Piliavin, I., Grogan-Kaylor, A., & Nesmith, A. (1998).Foster youth transitions to adulthood: Outcomes 12 to 18 months after leaving out-of-home care. Madison, WI: Institute for Research on Poverty.

Festinger, T. (1983).No one ever asked us: A postscript to foster care. New York: Columbia University Press.

Gallagher, L., Uccell, C., Pierce, A., & Reidy, E. (1999). State general assistance programs: 1998. Washington, DC: Urban Institute.

Holcomb, P., Flores, K., Herbig, C., Tumlin, K., Botsko, C., Kaye, L., & Seefeldt, K. (1998).Income support and social services for low-income people in Wisconsin. Washington, DC: Urban Institute.

Jones, M. & Moses, B. (1984).West Virginia's former foster children: Their experiences in care and their lives as young adults. New York: Child Welfare League of America..

Pettiford, P. (1981).Foster care and welfare dependency: A research note. New York: Human Resource Administration, Office of Policy and Program Development.

[ Go to Contents ]

 

Endnotes

1.  The analysis of administrative data has its own limitations. We discuss the limitations most relevant to our analysis in the concluding section of this report.

2.  The Human Services Reporting System (HSRS) is a state-wide data collection system that includes both client-specific and summary information about the social and mental health services provided by state and/or county agencies.

3.  This does not include the 21 youth age 17 or older who were transferred to a private child welfare agency, or the seven youth age 17 or older who died while in out-of-home care.

4.  The race/ethnicity of former foster youth who had received services from Milwaukee County is very different from the race/ethnicity of former youth who received services from the state's other counties. Of former foster youth who received services from Milwaukee County, 60% are African American and 34% are White. Conversely, of former foster youth who received services from other counties, only 7% are African American and 83% are White.

5.  Although we had hoped to distinguish between relative and non-relative foster home placements, our preliminary analysis leads us to believe that the HSRS variable used to indicate this distinction is not reliable.

6.  Once we have UI data for former foster youth who were 16 years old when they exited care, we will be able to compare the employment and earnings outcomes of former foster youth who aged out or were discharged to independent living to those of former foster youth who were reunified or placed with relatives at least one year prior to the time at which they would have aged out had they remained in care.

7.  This excludes those cases where an SSN was present but invalid (e.g, '999999999' or '123456789').

8.  Differences #1 and #2 are consistent with the fact that the majority of former foster youth who were receiving services from Milwaukee are African American while the majority of former foster youth who were receiving services from other counties are White.

9.  If former foster youth exited in first half of the quarter (i.e., the 1st of Month 1 through the 15th of Month 2), that quarter is treated as the first post-exit quarter. If former foster youth exited in second half of the quarter (i.e., the 16th of Month 1 through the 30th/31st of Month 3), the following quarter is treated as the first post-exit quarter.

10.  Approximately 91% of Wisconsin workers are employed in jobs that are covered under the state's UI program.

11.  Yet another important limitation of our approach is that the UI data provide no information about hourly wages or number of hours employed.

12.  One of the authors will pursue this approach in subsequent analyses of these data.

13.  Given the current minimum wage, a full-time minimum wage worker would have earned a total of $21,424 over a period of eight quarters — three times what the former foster youth earned on average. However, even at the lower minimum wage rate that was in effect prior to August 1996 (i.e., $4.25), the full-time minimum wage worker would have earned a total of $17,680-still more than twice the average earnings of the former foster youth.

14.  The dependent variable is coded as 1 if former foster youth were employed in at least one of the first eight post-exit quarters and 0 if they were not.

15.  Wisconsin's TANF program is Wisconsin Works, or more commonly, W-2. It was first implemented in September 1997. AFDC was not completely phased out until March 1998.

16.  If former foster youth exited in the first half of the quarter (i.e., the 1st of Month 1 through the 15th of Month 2), that quarter is treated as the first post-exit quarter. If former foster youth exited in the second half of the quarter (i.e., the 16th of Month 1 through the 30th/31st of Month 3), the following quarter is treated as the first post-exit quarter.

17.  Not all W-2 participants are eligible for cash assistance. Eligibility for cash assistance is limited to those assigned to the program's two lowest employment tiers (i.e., W-2 Transition and Community Service Job) and to those who have given birth to a child within the past 12 weeks. Other W-2 participants are eligible for services and non-cash benefits such as child care subsidies. Therefore, not all former foster youth who participated in W-2 would be included in our measure of cash assistance receipt.

18.  Because CARES includes only State of Wisconsin data, this approach will underestimate public assistance utilization to the extent that former foster youth were living in and receiving benefits from another state.

19.  One of the authors of this paper will soon have access to SSI data for these former foster youth.

20. In the past, all Wisconsin counties were legally required to provide both cash and medical assistance to eligible, very low income single adults through the state's mandatory General Relief program. In the 1995 - 1997 biennial budget, that program was replaced by the Relief Block Grant program, which counties have the option to participate in or not (Gallagher, Uccello, Pierce & Reidy, 1999; Holcomb, Flores, Herbig, Tumlin, Botsko, Kaye, & Seefeldt, 1998).

21.  For example, in Dane County (Madison), the most populous county with a non-medical (i.e., cash) GR program, the maximum program benefit is $247 per month (Gallagher, Uccello, Pierce & Reidy, 1999).

22.  Moreover, even if one did have data on Medicaid payments, it is not clear how this information would be incorporated into an income measure.

23.  It is not surprising that all of the AFDC/TANF cash assistance recipients were female given that AFDC/TANF eligibility was limited to those former foster youth who were pregnant or caring for dependent children. However, female former foster youth were also far more likely than males to have received Food Stamps, for which individuals without children may be eligible.

24.  It is possible that differences in the total amount of AFDC/TANF cash assistance and/or Food Stamps that former foster youth received are the result of differences in the number of quarters for which AFDC/TANF cash assistance and/or Food Stamps were nonzero. However, one finds similar differences when amount of AFDC/TANF cash assistance and/or Food Stamps per quarter is used.

25.  The one dissimilarity between Table 12B and Table 10B is that former foster youth who had aged out or been discharged to independent living were the most likely to have received AFDC/TANF cash assistance were no longer the most likely to have received AFDC/TANF cash assistance once the observation period was extended.

26.  The dependent variable is coded as 1 if the youth received AFDC/TANF cash assistance and/or Food Stamps in at least one of the first eight post-exit quarters and 0 if the youth did not.

27.  The one exception is that median total income among former foster youth who had been placed voluntarily is no longer lower than median total income among former foster youth who had been adjudicated children in need of protection (CHIPS).

28.  The federal poverty line for a single individual was $7,470 in 1995, $7,740 in 1996, $7,890 in 1997, $8,050 in 1998, $8,240 in 1999, and $8,350 in 2000.

29.  As noted above, our measure of total income for the first eight post-exit quarters, which includes only earnings and AFDC/TANF cash assistance, is incomplete in that it does not take several potentially important sources of income into account. Nevertheless, it is the best indicator of self-sufficiency that our data allow us to compute.

30.  There are statistical techniques for dealing with this problem of sample selection bias. One of the authors will be using these techniques in future analyses of these data.

31.  We expect to have UI data for the years 1992 through 1994 in the near future.

32.  This is somewhat different than the analysis we had initially proposed, which was to compare the outcomes of former foster youth who would have aged out between 1992 and 1998 had they not been returned home to those of former foster youth who did age out of care during this same period. However, because we included former foster youth who had been reunified, placed with relatives or adopted in our sample, we were able to make a very similar comparison.

Populations
Youth
Location- & Geography-Based Data
State Data