Life After Welfare: Report of the Georgia Welfare Leavers Study

Prepared for: Georgia Department of Human Resources Division of Family and Children Services
January 2001

Prepared by:

E. Michael Foster, Ph.D., Principal Investigator

Dana K. Rickman, Project Manager

Applied Research Center Andrew Young School of Policy Studies
Georgia State University Atlanta, Georgia

Table of Contents

I. Executive Summary

1

II. Introduction

8

III. Methodology

10

IV. Results

19

1. Demographics

19

2. Why did they leave?

23

3. How are the women faring?

29

4. Which former recipients fare best?

39

5. Barriers to transition?

43

6. How are the children faring?

44

7. Comparisons to current Georgia recipients

53

8. Comparisons to leavers nationwide

55

V. Conclusion

58

References

62

I. Executive summary Funded by the Department of Human Resources, the Georgia State welfare leavers study
tracked families as they left Temporary Assistance to Needy Families (TANF).1 Using administrative data combined with the results of a telephone survey, the project monitored the impact of leaving welfare on the individuals and their families. The study includes both single-parent and child-only leavers as well as individuals who have returned to the rolls.
This report summarizes our findings. In our original proposal, we posed five sets of questions: 1) Why did the respondents leave? 2) How are the women faring? 3) Which group of former recipients are faring best? 4) What are the barriers to transitioning off TANF? 5) How are the children of former recipient faring?

Using data collected between June, 1999 and August 20002, we can answer these five

questions. Before doing so, however, we can report basic demographics for the women and families

in our study.

C

most leavers are single women and have finished high school.

97% of our respondents were female. A majority of respondents (61%) had never been married, while

1 The study also has received funding from the Assistant Secretary for Planning and Evaluation (ASPE) in
the Federal Department of Health and Human Services under Grant #98ASPE302A. This study is part of a
consortium of leavers.
2The response rate for the telephone survey approached 70%.

Life After Welfare (2/19/01)

p.1

only a small percentage (12%) were currently married. Moreover, 59% had finished high school or

received a GED, and another 19% reported having some college.

C

most respondents were African-American and the majority lived in urban or suburban

areas.

A full 84% of the respondents were African-American, and more than half (61%) lived in urban or

suburban areas.

C

most respondents had been on assistance for a year or longer.

Respondents did differ on length of time receiving TANF. Forty percent reported receiving TANF for

less than one year, and 60% reported receiving TANF for more than one year.

Question 1) Why did the respondents leave?

Most respondents left for employment, but a minority may have been confused about the terms of

welfare reform.



most of the respondents (73%) reported leaving TANF for employment .

An overwhelming majority (73%) of respondents left for employment. Among just single-parent cases,

that percentage rises to 81%. Nine percent reported leaving because the child left the household, and

only a very small percentage (2%) left for marriage. A majority (60%) of the respondents wanted to

leave, and 93% of the respondents were at least somewhat confident they will remain off.

C

some respondents were confused about welfare reform, which may contribute to the

closing of cases.

Eight percent of the single-parent cases and 13% of the child-only cases reported they exceeded the

Life After Welfare (2/19/01)

p.2

time limit. These figures illustrate some confusion about the terms of welfare participation. Single-parent cases will not begin to hit the time limit until January, 2001, and there is no time limit for child-only cases.

Question 2) How are the women faring?

Focusing on single parent cases, most leavers have joined the ranks of the working poor (or near

poor).



most leavers (69%) are working; however, among those working, full-time hours for

low earnings were the norm.

A full 69% of the respondents were working at the time of the survey. Of those working, 73%

reported working at least a forty hour work week. While most leavers are working, monthly earnings

were low. Sixty-five percent of employed leavers reported monthly earning below $1,000. Of those,

32% reported monthly earnings below $800. Only 18% reported earnings above $1,200 per month.

Unless these women are able to supplement their earnings with income from other sources, they and

their families are living in poverty.



child support payments were not a reliable source of income for mothers .

Child support income was not a reliable source of additional income for most leavers. The Child

Support figures indicate that the vast majority of women either had no judgement (31%) or received

nothing or less than they were owed (28%). Only 18% of all women were owed and received more

than $200 per month in child support.



most respondents remained off welfare, but many participated in other government

Life After Welfare (2/19/01)

p.3

programs.

Nine months after exit, only 15% of adults and 19% of the children had returned to TANF. Children in

child-only cases that close are more likely to return to welfare as a single-parent case. Between 3 and

9 months after exit, only 1% of child-only cases returned to welfare as a child-only case. After 9

month, 22% of former child-only cases were active single-parent cases. While most leavers remained

off the rolls, many were still participating in other government programs: 87% received free/reduced

priced lunch, 74% used Food Stamps, and 60% lived in public housing.



recipients worried about food, but most are able to make ends meet.

There was an obvious perceived worry about having enough to eat. Thirty-eight percent of the

respondents sometimes or often worried they may not have enough food to eat. However, 13%

reported running out of food before they had money to buy more. So while insecurities about the

availability of food was a concern, many of those families found ways to obtain the food they needed.

A majority of respondents (59%) said they relied on family members to make ends meet, and 43%

reported relying on friends.

3) Which group of former recipients are faring best?

Better educated leavers far substantially better than the less educated.



better educated leavers tend to do better than those without a high school diploma or

GED.

Forty-four percent of those without a high school diploma or a GED earned less than $800 per month

compared to 33% of high school graduates and 25% of those with some college. Thirty-three percent

Life After Welfare (2/19/01)

p.4

of those with some college earned more than $1,200 compared to only 11% of those without a high

school diploma or GED. Furthermore, those without a high school diploma or GED were twice as

likely to often worry food will run out.



education levels also impact the children living in the home .

Only 57% of children living in a home with a parent or primary care-giver that did not have a high

school diploma or GED had health insurance compared to nearly 80% of children whose parent or

primary care-giver had at least a high school diploma.



there are enormous differences between single-parent and child-only cases.

These differences are so great that the two groups should be tracked and examined separately.

Respondents in child-only cases were more likely to be married, have higher incomes, work fewer

hours, and rely less on Food Stamps. Children in child-only cases were more likely to be insured, stay

in day care for fewer hours each day and spend fewer days per week in a child-care arrangement.

4) What are the barriers to transitioning off TANF?

Report barriers among women having left are fairly low.



problems with transportation and other barriers affect only a minority of respondents.

Transportation was a barrier for some leavers. Nearly 1 in 5 respondents reported walking or relying

on rides from friends or family to get to work. Levels of reported domestic violence were fairly low

with only 6% report ever being abused by their spouse or partner. However, self-reported instances of

alcohol and drug abuse were slightly higher, 9% and 8%, respectively.

Life After Welfare (2/19/01)

p.5

5) How are the children of former recipient faring?

Children of leavers face many of the problems faced by all children in poor families, such as limited

involvement with absent parents and access to high-quality day care.



general health was good, but some children had physical disabilities or emotional

problems.

Nearly 80% of all children were reported as being in excellent or good health, but approximately 15%

of the children have health problems that limit their daily activity. Nearly 30% of the children were

reported by their mother or primary care-give as often or sometimes being unhappy, sad, or depressed.

Also, nearly 1 in 3 were reported as having trouble concentrating or have trouble getting along with

other children.



access to quality child care is still a concern.

Nearly 20% of the children under 12 were reported as staying either home alone or in the care of a

relative under the age of 13 as a "type of child care."



involvement with children by absent parents is very limited.

Ninety-seven percent of the children in this survey were not living with their father. Of those, 37% of

single-parent leavers reported the child's father was either dead, or `like-dead,' meaning the mother

had no contact with the father and did not know where he was living. A further 25% of children had

absolutely no contact with their absent parent. Moreover, nearly 40% of the children living in child-only

cases have no contact with either parent.

Where do we go from here? Continued research on leavers is clearly needed for at least two

reasons. First, this study was designed to provide a snapshot of how leavers fare shortly after leaving

Life After Welfare (2/19/01)

p.6

the rolls. Long term follow-up is needed to understand how these leavers fair over time. Second, as long-term recipients approach the four year time limit, one expects to see substantial shifts in the composition of leavers. How these families fare after leaving, compared to the families studied here, will be the true test of welfare reform.

Life After Welfare (2/19/01)

p.7

II. Introduction The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA)
replaced Aid to Families with Dependent Children (AFDC) program with the Temporary Assistance for Needy Families (TANF) block grant. Under the new state plan submitted by Georgia and approved by the federal government, welfare is no longer an entitlement but a program that provides temporary cash assistance. The Georgia Department of Human Resources' Division of Family and Child Services implemented the program according to these principles: 1) a central focus on work, 2) meeting the needs of children first, 3) linking benefits to personal responsibility, and 4) reducing teen pregnancy. In Georgia, cash assistance is limited to a maximum of four yearsone year less than the federal maximum lifetime benefit. Individuals who have been on the rolls continuously since the reform was enacted reached their lifetime limit in January, 2001.
Since the implementation of these changes, the TANF caseload in Georgia has clearly plummeted. Between January, 1997, and January 2000, the number of families receiving TANF in Georgia dropped by more than 63,000, representing nearly a 55% reduction in three years. What is much less clear is the impact of leaving welfare on these women, their children, and the communities in which they live. Funded by the Department of Human Resources,3 the Georgia State welfare leavers study monitors the impact of leaving on the individuals, their families, and their community. (For a fuller description of welfare reform in Georgia, please see technical appendix I, available under separate cover from the Applied Research Center, or at http://www/arc/gsu.edu.)

3The study has also received funding from the Assistant Secretary for Planning and Evaluation (ASPE) in the Federal Department of Health and Human Services. This study is part of a consortium of leavers studies.

Life After Welfare (2/19/01)

p.8

This project examines five key questions regarding former recipients and their families: 1) Why did the respondents leave? 2) How are the leavers faring? 3) Which group of former recipients are faring best? 4) What are the barriers to transitioning off TANF? 5) How are the children of former recipients faring?
This report will proceed in three major section. The first will outline the methodology of the study. It describes key characteristics of the study, including (1) definition of leavers; (2) use of interview and administrative data; and (3) a discussion of our efforts to locate respondents and an analysis of non-response. Using interview data, the second section will provide answers to the five key questions outlined above. The section also will compare Georgia leavers to those who have remained on cash assistance and to leavers nationwide. The final, third section will offer some conclusions and policy implications that can be gleaned from the Georgia leavers study.

Life After Welfare (2/19/01)

p.9

III. Methodology This section outlines the methodology of the study. First we will present the project design
including the definition of leavers and the use of administrative and interview data. Second, we will describe our efforts to locate the respondents and will provide an analysis of non-response. A fuller description of the study's design can be found in technical appendix II, available under separate cover from the Applied Research Center or at http://www.arc.gsu.edu.

Project Methods Definition of Leavers
A discussion of the methodology of any leavers study begins with the definition of leavers. This project defines leavers as cases not having received cash assistance for two consecutive months. The two-month limit excludes cases that leave welfare for a single month. Prior research has treated these individuals as having missed a payment but not as having moved off welfare. The definition of leavers is consistent with that used in studies in other states, including the other studies in the Assistant Secretary for Planning and Evaluation (ASPE) -funded consortium of leavers studies. Included as leavers are clients who continue to receive other TANF services, such transitional Medicaid, Food Stamps, or child care assistance. This report focuses on leavers that left welfare between January, 1999, and June 2000. Sample population
The study population in Georgia differs from that in other leaver studies in two ways. First, Georgia is among the few studies that include child-only cases in the study population. Child-only

Life After Welfare (2/19/01)

p.10

cases involve children who are receiving welfare but who are not living with a parent. These children might be living with a grandparent, aunt or uncle, or with a parent who is ineligible for TANF (e.g. an SSI recipient). While the child may not live with a parent, the care-givers income does not count towards determining TANF eligibility for the child as they are not included in the grant. Neither these children nor their care giver are subject to work requirements or time limits.
These cases are essential to understanding the full impact of welfare reform, especially on children. First, these children make up a substantial minority of children on TANF. In December 1999, child-only cases made up 47% of the active case load in Georgia, and 23% of cases leaving the rolls. Furthermore, some researchers, policy makers and advocates fear that welfare reform will stimulate the growth of child-only cases. In particular, they argue that reform gives parents an incentive to move their children in with relatives. Doubtless these changes in living arrangements have important implications for the child's well-being. For these reasons, we believe that researchers should not arbitrarily drop child-only cases from studies of either leavers or stayers.
A second difference between the Georgia State leavers study and those in some other states is that this study includes individuals who have returned to the rolls. One would expect these individuals differ systematically from individuals remaining off the rolls, and as a result, studies that exclude individuals who have returned offer an incomplete and misleading picture of how leavers are fairing. Furthermore, because states differ in the rate at which families return to the rolls, excluding those families makes it virtually impossible to compare results across states4.

4All studies funded by ASPE follow all leavers, including those who return to the rolls, provided they remain off cash assistance for a minimum of two months.

Life After Welfare (2/19/01)

p.11

Administrative data Like other leaver studies, the GSU study relies on interview data as well as various
administrative databases. For the latter, there are four key files: the closed case file, the TANF Federal Report file, the current recipiency file, and the Child Support Enforcement database.
The closed case file includes basic demographic information, including race, age, gender and relationship to other persons in the household and is used in several ways. First, these data serve as the sampling frame for the study. Initially, a sample of cases is drawn from which a study respondent is identified. For these individuals, the file provides the contact information with which the study begins to track potential respondents. Second, the file provides information on the family's use of welfare when they were receiving payments. This information includes case status, payments received, and case and client identification (ID) numbers. The two ID numbers are used to link these data to other data sources.
A second source of administrative data is the TANF Federal Report file. This file provides supplemental information describing the case when it closedthis information includes Food Stamp receipt, work eligibility status and work experience, reported earnings, and family structure at the time of case closure. Used in conjunction with the TANF Federal Report file is the third file--the current case file. This file is used to determine if and when individuals have returned to the welfare rolls.
A fourth and final source of administrative data is the Child Support Enforcement (CSE) database. This information is used to locate potential survey respondents and to determine whether they have an award and if they are currently receiving child support payments. Telephone Interview

Life After Welfare (2/19/01)

p.12

While informative, administrative data are somewhat limited in their scope. They provide no information on many key outcomes (e.g., a leaver's mental health or barriers to employment) and exclude some individuals, notably those no longer involved in any public programs. As a result, the project is conducting a telephone interview with approximately 200 leavers per month covering a wide range of topics. All respondents complete a core set of items concerning demographics, employment and economic status at the time of the interview, approximately 6 months after exit from welfare. Individuals also complete a randomly chosen module. The module topics are (a) sources of income and transportation; (b) child care arrangements; (c) mother's mental health and exposure to domestic violence; (d) parenting and home environment; and (e) understanding of welfare reform. Having study participants answer only part of the survey allows us to include a wide range of topics in the interview without overburdening respondents. Conventional analytical methods can easily allow for the resulting patterns of missing data.
Taken together, the administrative and survey data provide information on a range of key characteristics and outcomes. These include but are not limited to employment and earnings; health insurance; child care; child well-being; barriers to self-sufficiency; deprivation and insecurity; and attitudes toward and knowledge of TANF.

Locating Respondents and the Analysis of Non-response Locating Respondents and the Response Rate
Our analyses are based on individuals who left TANF between January, 1999 and April 2000.

Life After Welfare (2/19/01)

p.13

These individuals were interviewed between June, 1999 and October, 2000. Most interviews occurred from 4 to 6 months after exit from welfare. The individuals were interviewed approximately 5 to 6 months after leaving TANF.
While challenging, locating a large and representative sample of respondents for the survey is essential. In order to locate hard-to-find individuals, we have implemented a thorough tracking procedure that utilizes all available resources. This process represents a significant improvement over what was included in our original proposal. These methods include advance cover letters, a $25 incentive payment, a toll-free telephone number for call-ins, reverse directory look-up information, and data-matching with the CSE databases. Reflecting these improvements, our response rates rose dramatically between June, 1999 and October, 20005. It is important to note that as the response rate increased, the number of disadvantaged families responding to the survey increased. (For the full analysis of our response rate, see technical appendix III, available under separate cover from the Applied Research Center, or at http://www.arc.gsu.edu.)
Table 1 describes our response rate by cohort month. One can see that the increase in the response rate was driven by two factors. First, we increasingly were able to establish a working telephone number. One can see the valid sample size rose as a percentage of potential respondents from under 50% to over 80%. Second, one can see that the participation rate rose from just over half (53%) to more than 80% in the final months of the study.

5We completed interviews for the August, 2000 cohort in October, 2000.

Life After Welfare (2/19/01)

p.14

Table 1 Cohort6

Potential Respondents

Response Rate (%)

Valid Sample Size*

Participation Rate (%)**

June `99

600

26.0

293

53.2

August `99

600

34.1

370

55.1

September `99

600

37.3

413

54.2

October `99

600

52.6

355

89.0

December `99

600

50.6

362

83.9

February `00

450

54.6

246

87.2

March `00

450

63.6

286

87.5

April `00

450

67.8

323

94.4

May `00

450

66.9

336

89.6

June `00

450

70.1

378

83.3

August `00

400

69.3

328

84.1

* Excludes those where no working telephone number could be established ** Percent of valid sample size that is interviewed.

Refusal Rate (%)
1.7 0.5 2.2 0.9 2.8 7.5 2.5 2.7 1.3 0.4 2.3

Analysis of Non-response The response rate for this project is in the range of response rates among comparable projects7.
The quality of a study, however, depends not only on the response rate but on the extent to which respondents and non-respondents differ. The response rate could be rather high (80%), but the study

6 The cohorts are defined as the month the survey lab began to locate them for an interview. To make sure we obtained 200 completed interviews per month, we began by attempting to locate 600 individuals per monthly cohort. Because of our increased response rate, we reduced the pool of potential respondents to 450, and then to 400.
7The following response rates have been reported and are available on the ASPE web page: Arizona 72%, California 69%, New Mexico 72%, Ohio 70%, Texas 51%, Virginia 69%, and Washington State 72% ( see: http://aspe.os.dhhs.gov/hsp/leavers99/reports.htm)

Life After Welfare (2/19/01)

p.15

might be very misleading if the 20% who do not respond differ substantially from those who complete interviews. At the same time, a study with a much lower response rate might describe the population of leavers accurately if respondents and non-respondents are similar.8
In most policy studies, one knows little or nothing about non-respondentsby definition, those individuals failed to respond to requests for information. What distinguishes research on leavers, however, is the fact that the administrative data provide a great deal of information about individuals we are unable to interview. This information is extensive and invaluable. It allows us to compare individuals who do and do not respond across a range of relevant characteristics, including demographics, as well as past and current welfare receipt.
Using administrative data on all individuals in our study, the project examined such differences. We estimated statistical models using the roughly 5000 individuals we tried to contact to participate in the study. Table 2 compares those we were able to interview with those we were unable to locate. There are relatively few, if any, meaningful differences between the groups. Several of the differences are statistically significant or nearly so, but are small in practical terms. The former reflects the large sample size. It does appear that white leavers are somewhat under represented in our data. (For the full analysis of our response rate, see technical appendix III, available under separate cover from the Applied Research Center, or at http://www.arc.gsu.edu.)

8Of course, all else equal, a higher response rate is desirable because (1) the number of observations is greater, increasing statistical power; and (2) the potential bias caused by differences between respondent and nonrespondents is greater at higher levels of non-response. (If the response rate is 98%, then the potential bias is still rather small even if non-respondents are quite different.)

Life After Welfare (2/19/01)

p.16

Table 2 Race (p = .13)
Age (p = .11)
Length of Time on Welfare (p = .09)
Region Type (p = .19)
Total N

White Non-white
<18 18-25 26-35 36-44 45-62
>62 < 6 months 1-12 months
1-2 years >2 years
Urban Suburban Rural growth Rural decline

Respondents 16% 84% 0% 33% 33% 21% 10% 3% 16% 22% 30% 31% 36% 25% 26% 11% 2870

Non-respondents 23% 77% 1% 34% 37% 18% 8% 3% 17% 24% 30% 29% 35% 24% 30% 9% 2193

Item Non-response The study also suffered from a modest amount of item non-response, where survey respondents
either would not or could not answer particular questions. This was quite rare. By far, the most common instance where this problem occurred was in the case of earnings. About 15% of individuals who were employed or temporarily laid off were unable or unwilling to provide information on hourly wages or other information necessary to calculate monthly earnings. These data appeared to be missing completely at random, but we did consider the impact of this "missingness" on our findings. To allow

Life After Welfare (2/19/01)

p.17

for systematic differences among individuals with and without reported earnings, we used hot-deck imputation9 to impute missing earnings using respondents' demographic information. We found that the distribution of earnings did not change. The demographic variables used in the imputation were: age, race, education, and region. As shown in table 3, we were able to impute 58 observations using demographic information and the earnings distribution does not measurably change.

Table 3

N

Reported earnings

576

Imputed earnings

634

Mean $1,047 $1,066

SD $542 $585

9This is an advanced method to correct for the missing values where each missing value is replaced by a set of plausible values drawn from their predictive distribution. In any data set, a row of data with missing values in any of the variables is defined as a missing line of data, and a complete line is one where all the variables contain data. Our imputation method replaces the variables in the missing lines stochastically with the corresponding values of the complete lines using a Bayesian bootstrap method.

Life After Welfare (2/19/01)

p.18

IV. Results Our original proposal posed five sets of questions. Using interviews completed between June
1999 and October 2000, we will answer those questions. Before doing so, however, we briefly describe the demographic characteristics of our respondents. This section concludes by comparing key outcome measures of leavers in Georgia with stayers in Georgia and leavers nationwide. We provide these additional data to place our findings in a context. (For a review of findings from other ASPE leavers studies, see technical appendix IV, available under separate cover from the Applied Research Center, or at http://www.arc.gsu.edu.)

Basic Demographics

We begin by noting that 23% of our leavers are child-only cases. For these cases, the child's

primary care-giver served as the respondent. In most circumstances that was a grandparent (40%) an

aunt or uncle (20%), or

N=2870
Fig. 1 Type of Closed Case

other relative (22%).

23% 77%

Child-only Single-parent

Life After Welfare (2/19/01)

p.19

As noted in table 410, the average age of our respondent population is 34 years old. Twenty

percent are between 35 and 45, and almost 13% are over the age of 45. Only 2% are less than 18

years old.

When divided by case type, it is evident that care givers in child-only cases are significantly

older than parents in single-parent cases (45 vs. 29).

Table 4 Age Average Age (n=2868)

Age of Respondent
Less than18 18-25 25-35 35-45 Above 45

2% 31% 33% 20% 13%
34 years

Moreover, 61% of our leavers have never been married. Figure 2 describes marital status.

Approximately equal numbers of respondents are either married (12%) or divorced (11%). Only 8%

N=2858

Fig. 2 Marital Status

report cohabitating.

61%

4%

12%

4%

Married 11% Widowed
Divorced Separated Never Married
8% Cohabitating

10The number of observations varies by item due to item non-response or matrix sampling. A few reported outcomes, such as time receiving welfare, are from the Federal Report File from which we only have 6 months worth of information.

Life After Welfare (2/19/01)

p.20

When divided by case type, the most striking difference is among married respondents. Twenty-seven percent of care-givers in child-only cases are married, compared to only 7% of singleparent leavers.
Figure 3 describes the education level of the general population of respondents. A full 19% of leavers reported having some college education, and another 59% have a high school diploma or a GED. Only 1 in 4 (22%) leavers did not finish high school.

N= 2805

Fig. 3 Education Levels

19%

High school/GED

Lessthan HS

22%

59%

Some college

This distribution does not change by case type. Approximately 60% of both groups have at least a high school diploma or GED. However, a slightly higher percentage (20%) of child-only care givers have some college, compared to 17% of single-parent leavers.
Next, we look at where our respondent live. Fig 4 shows that more than half live in urban or

Life After Welfare (2/19/01)

p.21

suburban areas11. Only a small percentage (11%) live in a rural area with declining economic growth and population, a rate roughly comparable to the Georgia population as a whole (13%). (Bachtel, 1999). Twenty-eight percent of the leavers live in a rural area where the population and economy is growing. The distribution for child-only cases and single-parent cases was similar.

N= 2712 11%

Fig. 4 Region

28%

37%

Urban Suburban Rural Growth Rural Decline

24%

Table 5 provides a break down of the remaining demographic characteristics of our sample

11The report uses a classification system of the 159 counties in Georgia developed by Professor Doug Bachtel, a noted demographer at the University of Georgia and author of the Georgia County Guide. Variations in factors which characterize the quality of life in a particular geographic region such as income, employment, education, population migration, and housing are used as a basis for the classification of the counties. According to Dr. Bachtel's criteria, each county is classified into one of four categories: urban, suburban, rural growth, or rural decline. In summary, presented below is a brief description of the four areas within the classification. Of the 159 counties in Georgia, 7 are identified as urban, 35 are considered suburban, 77 counties are viewed as experiencing rural growth, and 40 counties are characterized as being in decline.
With populations over 50,000, characteristically, the urban counties represent the heart of Georgia's metropolitan urban centers. The suburban counties are, for the most part, metropolitan because a significant number of the residents living there commute to the urban areas to work. These areas generally are predominately white and affluent. Likewise, many residents in these areas possess a high degree of educational attainment and income level. Another group comprises those counties identified as "growing rural Georgia." While scattered across the state, these rural counties tend to be concentrated in the north. These areas are usually associated with having either scenic beauty or some type of landscape which makes them attractive places for tourism. Additionally, these areas are located near some regional growth center which contributes to the counties' economic development. Conversely, the counties identified as "declining rural Georgia," arguably, are the areas considered to be in the greatest peril. These counties are characterized as experiencing long term population loss, lack of employment opportunities, and low levels of supportive services. Historically, these areas have a legacy of low educational attainment and skill development. Thus, many of the residents in these counties are dependent on social welfare services (Bachtel, 1999).

Life After Welfare (2/19/01)

p.22

population and do not differ significantly by case type. Only a small percentage of our respondents

(3%) of the respondents were males. Eighty-four percent of the leavers are non-white, closely

resembling the percentage of non-white respondents in the active case load in Georgia (83%)

(Department of Human Resources, 1999). Another 30% received payments for more than 2 years. In

terms of household composition, nearly 70% of the households had 1 to 2 children, and 30% reported

having a child under 3 years of age in the home.

Table 5 Race (n=2868)

Leaver Demographics

Non-white

84%

White

16%

1 -2 Children

66%

Number of children in home 3 -4 Children

29%

(n=1955)

More than 4 children

5%

Households with children under 3 years old (n=1950)

30%

Child-only households

23%

(n=2870)

Male headed households (n=2870)

3%

Less than 6 months

19%

Short vs long term receipt 6-12 months

20%

(n=1919)

12-24 months

29%

More than 24 months

31%

Table 5 also describes prior welfare receipt by respondents. According to the Federal Report File, at the time they left the rolls, nearly 40% of the leavers received cash assistance for less than one year. Only 31% had received assistance for 2 years or more.
Having outlined the demographics of our entire sample, we now turn to the first question; why

Life After Welfare (2/19/01)

p.23

did the respondents leave?

Question #1 Why did the respondents leave? In this section we will explore why recipients left welfare, if they wanted to leave, and how confident they are they will remain off. We first present the results for our entire sample, then provide sub-groups analysis of how these reasons vary by race, education, monthly earnings, age, and case type (single-parent vs child-only cases). Why they left Figure 5 shows that the majority of leavers left for employment (73%). A small percentage (9%) reported they left because they exceeded the time limit. This finding indicates a misunderstanding on the part of the former recipient as the time limits do not take effect until January, 2001.

N=2697

Fig. 5 Reasons for Leaving

9% 2% 5%
9% 2%

73%

Employed Married Exceeded limit Too many rules Child left Other

Table 6 provides a breakdown of why the respondents left by sub-groups. In terms of race,

Life After Welfare (2/19/01)

p.24

reasons for leaving were relatively similar. However, the other groups provide some striking

comparisons. In terms of education, 81% of those with either a high school diploma or GED left for

employment compared to only 54% of those who had less than a high school education. Also, those

without a high school education or GED were most likely to believe their case had been closed because

of a time limit (19%). In terms of earnings, 88% of those who had monthly earnings of less than $800

per month left TANF for employment compared to 80% of those who earned more than $1,200. The

reasons vary by age as well. Approximately 80% of those under 35 years of age left for employment.

That percentage decreases as the respondents get older. Seventy percent of those age 35 to 45 left for

employment, and only 37% of those over 45 left for employment. The most common reason for leaving

in that age category is that there were no longer children in the household (42%).

Finally, comparisons by case type are also striking. Eighty-one percent of single-parent cases

closed due to employment, compared to only 45% of child-only cases. Predictably, 33% of the child-

only cases closed because the child was no longer living in the household.

Table 6

Sub-group Analysis of Why Respondents Left TANF

Employed Married Exceeded Too Child in Other

Time Many House

Limit Rules Left

Race

Non-white

75%

1%

9%

5%

8% 1%

(n=2697) White

63%

5%

11%

6%

13% 2%

High school/ 81%

2%

6%

3%

7% 1%

Education GED

(n=2650) Less than HS 54%

1%

19% 11% 13% 2%

Some college 73%

2%

8%

5%

9% 2%

<$800

88%

2%

5%

2%

2% 1%

Earnings $800-$1200 90%

1%

2%

2%

4% 0%

(n=1513) >$1,200

80%

1%

7%

2%

8% 1%

Life After Welfare (2/19/01)

p.25

Less than 18 78%

0%

9%

5%

2% 5%

18-25

83%

1%

7%

7%

1% 2%

Age

25-35

80%

3%

8%

5%

3% 1%

(n=2697) 35-45

70%

3%

12%

5%

9% 1%

Above 45

37%

1%

15%

3%

42% 2%

Case type Single-parent 81%

2%

8%

6%

2% 1%

(n=2666) Child-only

45%

3%

13%

4%

33% 2%

Desire to leave Table 7 shows that while a majority of respondents told the interviewers they wanted to leave,
that result varied by respondent type. Most striking is that those without a high school diploma or GED were more likely to want to leave compared to those with a diploma or GED (72% v. 53%). There is a 9 percentage point difference between single-parent cases who wanted to leave and child-only cases (62% v. 53%). Slightly more than half (57%) of those who earned less than $1,200 a month wanted to leave TANF. However, an overwhelming 72% of those who earned more than $1,200 a month wanted to leave TANF.

Life After Welfare (2/19/01)

p.26

Table 7

Sub-group Analysis of Whether

N=2805

Respondents Wanted to Leave TANF

Fig. 6 Confidence in StayingYeOs ff

Race

Non-white

61%

(n=2708)

White

5% H2%igh school/ GED

Educa1ti5o%n Less than HS

52%
E5x3t%remely c7o2n%fident

(n=2660)

Some college

S6o6m%ewhat

<$800

c5o7n%fident

Earnings

$800-$1200

N5o7t%so

(n=1587)

>$1,200

c7o2n%fident

Less than18 18-25 78%

N5o5t%at all c6o2n%fident

Age

25-35

62%

(n=2708)

35-45

58%

Above 45

50%

Household Single-parent

63%

type (n=2747)

Child-only

52%

Confidence in staying off Finally, figure 6 shows that 93% of the respondents are at least somewhat confident that they
will remain off TANF.

Life After Welfare (2/19/01)

p.27

Table 8 describes, different sub-groups of respondents' confidence that they will remain off

TANF. The confidence levels are relatively similar across sub-groups with the exception of education.

While more than 80% of high school graduates or those with a GED, or some college are extremely

confident they will remain off TANF, only 58% of those without a high school diploma or GED are

extremely confident.

Table 8

Sub-group Analysis - Confident Will Remain Off Welfare

Extremely Somewhat Not So Not At All

Confidant Confidant

High school/

82%

13%

3%

1%

Education GED

(n=2730) Less than HS

58%

24%

13%

6%

Some college

83%

12%

3%

2%

Race

Non-white

77%

15%

5%

3%

(n=2778) White

75%

15%

7%

3%

<$800

79%

15%

4%

2%

Earnings $800-$1200

85%

11%

2%

1%

(n=1527) >$1,200

87%

8%

3%

1%

Less than 18

75%

20%

4%

2%

18-25

79%

15%

5%

2%

Age

25-35

77%

16%

5%

2%

(n=2778) 35-45

75%

15%

6%

4%

Above 45

75%

15%

7%

4%

Household Single-parent

77%

15%

5%

2%

type (n=2745)

Child-only

75%

21%

6%

4%

Do leavers Understand the Rules?

Life After Welfare (2/19/01)

p.28

That some leavers indicated they left because of time limits is somewhat troublesome. This finding suggests that leavers (and stayers) may not fully understand welfare reform, however a majority of them do. Sixty-five percent of the respondents knew there was a lifetime time limit, and of those who knew, 79% said that it was 4 years. When asked if the household's benefits increase when a child is born after the mother has received TANF for more than 1 year, 62% said no. Finally, 65% reported that Medicaid does not automatically end when TANF ends. While these numbers are encouraging, there is still more than 30% of leavers that do not understand even the most basic of Georgia's welfare laws.
Question #2 How are the women faring? This section will focus only on the women in single-parent cases. We will present their employment rates, monthly earnings, receipt of child support payments, measures of food insecurity and hardship, and continuing government program use including recidivism rates. Employment As figure 7 shows, 69% of former recipients are working, and another 12% are actively looking for work. Only a small percentage (5%) report being disabled. (Figures on the extent of the disability will be presented.)

Life After Welfare (2/19/01)

p.29

N= 2132
N= 1422 Fig. 7 Employment Rates

F7%ig. 8 Self-Reported Work

5%

3% Hours
5% 6%

Working

1% 12%
3% 68%

21% 69%

Temp laid of L2o0okoinrglefossjobs R2e1ti-r3e9d D4is0abled HMouosreektphgan 40
Student

As shown in figure 8, 73% of those who report employment, report working at least 40 hours per week.

Life After Welfare (2/19/01)

p.30

Earnings Even though employment rates are high, those who are working reported low monthly earning.
Fig. 9 displays the distribution of earnings for those who are currently working and provided earnings information.12 As illustrated in fig. 9, 65% report monthly wages below $1,000 per month. Of those, 32% report wages below $800 per month. Only 18% report earning above $1,200 per month. This figure excludes individuals who reported they were not working (approximately 30%).

N=1308

Fig. 9 Self-Reported Monthly

Earnings
2%

9%

6%

7%

26%

<$500 $500-799 $800-$999

17%

$1,000-$1,200 $1,201-$1,499

33%

$1,500-$2,500 > $2,500

Many women also rely on child support payments each month. Of our sample of 2210 closed

12Monthly earnings were calculated using four variables. Respondents who replied that they were employed were then asked (1) what they were paid, (2) was that hourly, weekly, bi-weekly, monthly, or yearly, (3) was that before or after taxes, and (4) how many hours a week did they work. Therefore, the results presented here do not include those respondents who said they were not working, and are only the earnings of the respondent.

Life After Welfare (2/19/01)

p.31

single-parent cases, 95% were found in the Child Support Enforcement Database. However, of that 95%, 69% (1524 cases) had a judgement that required payment from the father. A total of 899 cases (59% of those who were eligible to receive payments) actually received the amount owed to them each month. As fig. 10 shows, even these payments are relatively low. More than half (56%) receive less than $150 per month.

N=899
Fig. 10 Level of Child Support Payments

6% 16%

32%

22%

24%

<$100 $101 - $150 $151 -$200 $201 - $300 > $300

Taken together, these figures indicate that the vast majority of women either had no judgement (31%) or received nothing or less than they were owed (28%). Only 18% of all women were owed and received more than $200 per month in child support payments. Hardship and deprivation measures
Considering the low income levels, we next explore levels of deprivation and need among leavers. We collected a series of items on food adequacy, and the resulting data are mixed. As figure 11 indicates, 46% of leavers state that they have enough of the types and kinds of food they want. An additional 41% report having enough food to eat but not always the kinds they would like.

Life After Welfare (2/19/01)

p.32

Obs. N=1507
Fig. 11 Availability of Food
12% 1%
46% 41%

Have enough of types and kinds
Have enough but not the kinds we like Sometimes don't have enough to eat Often don't have enough to eat

Obs. N=1369

Fig. 12 Worried Food Would

Run Out

10%

often true

sometimes

28%

true

62%

never true

Obs. N=1368
Fig. 13 Couldn't Afford Balanced Meal

4% 14%

often true

sometimes true
never true

82%

Moreover, as figure 12 indicates, 10% of the respondents often worried that their food would run out before they had the money to buy more. A total of 28% stated that they often or sometimes could not afford a balanced meal for themselves or their children.
The data indicate that worries about food adequacy are actually greater than food inadequacy. Thirty-eight percent of the respondents worried they may not have enough food to eat. However, 13% actually ran out of food before they had money to buy more. So while insecurities about the availability

Life After Welfare (2/19/01)

p.33

S

o

m

e

m

o

n

e

y

l

e

ft

of food is a concern, many of those families are finding a way to afford the food they need. In order to provide a better sense of the standard of living of those who have left TANF, we
asked a series of questions about material hardship. As figure 14 indicates, only 28% of the respondents report having some money left over at the end of the month after paying all their bills. Another 46% report having just enough money to cover their expenses each month.

Obs. N=989
Fig. 14 Money Left at the End of Month

26%

28%

46%

Some money left JUST enough NOT enough

As shown in table 9, 18% of respondents have been unable to pay their full rent or mortgage since they left TANF, and 22% have been unable to pay their full utility bill. Most revealing, however, is that more than half (59%) of single-parent leavers had to rely on family members to help them make ends meet.

Life After Welfare (2/19/01)

p.34

Table 9

Hardship Measures

Needed help from family member to make ends meet

59%

Needed help from friend to make ends meet

43%

Unable to pay full utility bill

22%

Telephone service disconnected

19%

Unable to pay mortgage/rent in full

18%

Didn't go to the dentist when needed

16%

Sometimes or often do not have enough food

13%

Not enough food to eat

12%

Gas or electricity turned off

12%

Didn't seek medical attention when needed

10%

Have been evicted since leaving TANF

4%

N= 792

Continuing use of government services We also asked former recipients about their continuing use of government programs. Fig. 15
notes that 87% of former recipients utilize the free or reduced price lunch program for their children. Seventy-four percent are still enrolled in Food Stamps. Approximately 76% report having insurance. Of those, 87% report still being on Medicaid. Finally, 60% still live in public housing or use vouchers.
These figures are generally high in absolute terms, but are low relative to the rates among respondents prior to leaving the rolls. Presumably, nearly 100% of respondents received Food Stamps while on the rolls. Some reduction in Food Stamp use is expected or even desirable as individuals work their way off the rolls. However, when we limit our focus to individuals earning $1,000 or less, we see that only 71% receive Food Stamps.

Life After Welfare (2/19/01)

p.35

100% 80% 60% 40%

Fig. 15 Government Program Use

87% 74%

87%

60%

Received free/reduced lunch n=1305
Food Stamp n=934

Live in public housing n=1523

20%

Medicaid n=1263

0%

Recidivism Table 10 compares recidivism rates for single-parent cases and child-only cases. The first
column shows individuals whose cases had been closed for 3 months. January 2000 leavers were matched against the April 2000 active case file, and February 2000 leavers were matched against the May 2000 active case file. To gauge recidivism rates six months after exit (column 2), October 1999 leavers were matched against the April 2000 active case file, November 1999 leavers were matched against the May 2000 active case file, and December 1999 leavers were matched against the June 2000 active case file. Finally, to gauge recidivism rates nine months after exit (column 3), July 1999 leavers were matched against the April 2000 active case file, and September 1999 leavers were matched against the June 2000 active case file.
The rows present adults and children separately. The rows are further sub-divided according

Life After Welfare (2/19/01)

p.36

to whether the closed case was single-parent or child-only. The recidivism process may depend on the type of case, and disaggregating the data in this way allows us to examine movement between childonly and single-parent cases. As discussed above, there is a concern that as adults remove themselves from the grant, children are shifted into child-only cases. Reading across the rows, one can see the distribution of recidivism rates for each type of leaver. One can draw several conclusions from the table. First, one can see that the vase majority of leavers stay off welfare. For example, six months after exit, only 14% of single-parent adult leavers had returned to TANF. A second lesson is that children return to the rolls at a much higher rate than adults. Three months after exit, 10% of children had returned, a rate double that for adults. Nine months after exit, only 15% of adults had returned to the rolls compared to 19% of the children. Third, there does appear to be movement between singleparent and child-only cases, even among children. Children in child-only cases that close are more likely to return to welfare as a single-parent case. Between 3 and 9 months after exit, only 1% of childonly cases returned to welfare as a child-only case. After 9 month, 22% of former child-only cases were active single-parent cases. These results are encouraging as there is not a lot of movement of children out of single-parent cases and into child-only cases.

Life After Welfare (2/19/01)

p.37

Table 10

3 months after exit

6 months after exit

9 months after exit

Characteristics of Leaver

No Cash Assist.

Adults

single-parent case 95%

child-only case 96%

Total 95%

Children

single-parent case 92%

child-only case 85%

Total 90%

ChildOnly
0% 1% 1%
0% 1% 1%

Single- No Cash Child-

Parent Assist.

Only

5%

86%

1%

3%

90%

7%

4%

87%

2%

8%

83%

0%

14%

77%

1%

9%

82%

1%

Single- No Cash Child-

Parent

Asst.

Only

13%

85%

1%

3%

86%

8%

11%

85%

3%

17%

83%

0%

22%

77%

1%

17%

81%

1%

SingleParent
14% 6% 12%
17% 22% 18%

Life After Welfare (2/19/01)

p.38

Question #3 Which recipients fare best? This section compares earnings and hardship measures across sub-groups of leavers. Comparisons are based on marital status, region, age of respondent, race, education level, length of time receiving TANF, family size, the presence of children under 3 years of age in the household, and child-only cases v. single-parent cases. Earnings measures Table 11 shows earnings measures by respondent type. A greater percentage of married respondents report their own monthly earnings above $1,200 (38%). Approximately 47% of widows report the lowest earnings of under $800, while 50% of women who have never been married, and 58% of separated women report earnings between $800 and $1,200 per month. Earnings also depend on the number of adults in the household. Twenty-eight percent of respondents who live in a household with 2 adults earn more than $1,200 per month compared to only 18% of those who are the only adult in the household. Background and demographic characteristics also influence earnings levels. Twenty-six percent of those living in suburban areas have earnings above $1,200. However, the distribution of earnings is relatively similar across region types for those earning under $800 per month. Eighty-eight percent of those under 18 years of age earn less than $1,200 per month. Twenty-three percent of those between 25 and 35 earn more than $1,200 per month. There is little variation between white respondents and African-American respondents across earnings categories. Not surprisingly, earnings vary by education level. Forty-four percent of those without a high school diploma or a GED earn less than $800 per month compared to 33% of high school graduates and 25% of those with some college. Thirty-three

Life After Welfare (2/19/01)

p.39

percent of those with some college earn more than $1,200 compared to only 11% of those without a high school diploma or GED. Those in child-only cases earn more than their single-parent counter parts. Sixty-seven percent of single-parent cases earn more than $800 per month compared to 74% of child-only cases. Furthermore, 27% of child-only cases earn more than $1,200 compared to 17% of singe-parent cases Hardship measures
To gauge hardship, we asked the respondents if they ever worried that food would run out before they had money to buy more. Table 12 presents the results by respondent type. Concerns about the availability of food varied little across respondent types. One source of variation was marital status. Nearly half (49%) of those who are cohabitating worry that food often or sometimes runs out compared to 37% of their married counterparts.
There are also some differences by education level. Those without a high school diploma or GED are twice as likely to often worry food will run out. Fifty-one percent never worry compared to 63% of those with a diploma or GED, and 70% of those with some college. There seems to be little relationship between earnings and how long they have been receiving welfare payments.

Life After Welfare (2/19/01)

p.40

Table 11

Reported Monthly Earnings

<$800

$800-$1200

Married

21%

41%

Widowed Marital status Divorced

47%

41%

34%

36%

Separated

28%

58%

Never Married

34%

50%

Cohabitating

37%

40%

Number of 1 adult

32%

50%

adults in

2 adults

33%

39%

household

More than 2

30%

45%

Number of kids 1-2 kids

33%

47%

in household 3-4 kids

27%

57%

More than 4

43%

38%

Presence of very Kids over3

35%

46%

young kids

Kids under3

29%

51%

Less than 18

41%

47%

18-25

36%

49%

Age

25-35

30%

47%

35-45

34%

47%

Above 45

33%

48%

Race

Non-white

33%

48%

White

33%

42%

Case type

Single-parent

33%

50%

Child-only

26%

47%

Education

High school/GED

33%

50%

Less than HS

44%

45%

Some college

25%

42%

Length of receipt <6 months

39%

43%

6-12 months

33%

54%

12-24 months

28%

54%

>24 months

39%

45%

Urban

30%

50%

Region type Suburban

29%

45%

Rural Growth

37%

43%

Rural Decline

39%

53%

>$1,200 38%
12% 30% 14% 16% 22% 18% 28% 25% 19% 22% 19% 19% 20% 12% 15% 23% 19% 19% 19% 25% 17% 27% 17% 11% 33% 17% 13% 18% 16% 19% 26% 20% 9%

Life After Welfare (2/19/01)

p.41

Table 12

Worried That Food Would Run Out

often true sometimes true

Married

7%

30%

Widowed

12%

19%

Marital status Divorced

13%

30%

Separated

15%

29%

Never Married

10%

27%

Cohabitating

16%

33%

Number of 1 adult

10%

27%

adults in

2 adults

8%

33%

household

More than 2

16%

28%

Number of kids 1-2 kids

10%

29%

in household 3-4 kids

10%

30%

More than 4

10%

28%

Presence of very Kids over3

11%

29%

young kids

Kids under3

10%

27%

Less than18

17%

19%

18-25

10%

26%

Age

25-35

10%

29%

35-45

13%

31%

Above 45

8%

27%

Race

Non-white

11%

28%

White

10%

24%

Case type

Single-parent

11%

26%

Child-only

11%

26%

High school/GED

9%

27%

Education

Less than HS

15%

33%

Some college

8%

22%

Length of receipt <6 months

17%

29%

6-12 months

22%

21%

12-24 months

16%

33%

>24 months

17%

24%

Urban

12%

3%

Region type Suburban

9%

26%

Rural Growth

9%

28%

Rural Decline

11%

25%

never true 63% 69% 58% 55% 63% 51% 63% 59% 56% 63% 60% 62% 61% 62% 64% 64% 61% 56% 65% 62% 66% 58% 58% 63% 51% 70% 53% 57% 51% 60% 58% 65% 63% 65%

Life After Welfare (2/19/01)

p.42

Question #4 Barriers to a successful transition off welfare13 This section addresses the presence of barriers in the life of former recipients that may impede a sustained transition off welfare, such as domestic violence and drug and alcohol abuse, reported disabilities, and long welfare histories. A majority of individuals (60%) have access to a car to drive themselves to work. However, transportation is a barrier for some leavers. Eleven percent must rely on friends for rides and 9% must walk. Levels of reported domestic violence are fairly low with only 6% report ever being abused by their spouse or partner. However, reported instances of alcohol and drug abuse are slightly higher, 9% and 8%, respectively. Finally, there are relatively high instances of reported disability, especially among care-takers among child-only cases. (5% and 28% for single-parent and child-only respondents reported disability, respectively.) To gauge the extent of any disability, we asked individuals who reported themselves as disabled a series of questions about any limitations they faced. Half of those respondents (50%) report that they are unable to take care of their own shopping needs or are unable to drive or travel independently (51%). Twenty-seven percent report being unable to maintain their home independently or do not feel they are able to be responsible for their own medication. A full 34% are unable to handle their own financial matters, and 17% report being unable to operate a telephone without assistance.

13The questions concerning barriers to transition were each in one of five modules randomly assigned to the respondent. Therefore, only approximately one-fifth of our respondents were asked these questions.

Life After Welfare (2/19/01)

p.43

Table 13

Barriers To Transition

Transportation n=842

Drive

60%

MARTA

10%

Walk

9%

Depend on friends

11%

Domestic Violence n=1003

Abused by husband or partner

6%

Taken children to a shelter

5%

Drug/Alcohol Abuse n=657

Anyone in the house had a drinking problem in the

9%

past year

Anyone in the house had a drug problem in the past

8%

year

Reported disability n=345

Unable to operate telephone without help

17%

Unable to take care of shopping needs

50%

Unable to maintain house independently

27%

Unable to drive/ travel independently

51%

Can not be responsible for taking medications

22%

Unable to manage financial matters

34%

Question #5 How are the children faring? One of the main questions surrounding welfare reform involves its impact on children. This section examines the health and well-being of children of leavers as well as the environment shaping their development. The latter includes child care arrangements, involvement with absent parents, and residential stability. As discussed above, child-only and single-parent leavers differ in many ways, and as a result, we consider the circumstances of children in each separately. Also presented are subgroup analyses based on race, education level, region, and if the parent or care-giver

Life After Welfare (2/19/01)

p.44

wanted to leave TANF.14 Health care
Figure 16 examines general health outcomes, and the figure shows little difference between children in single-parent and child-only closed cases. In single-parent cases, 80% of the children are reported as being in very good or excellent health, a figure only slightly larger than that for child-only cases (72%).

Fig. 16 Child's Health

70% 60% 50% 40% 30% 20% 10%
0%
ExcellenVt erygood Good

Fair Poor

Single-parent n=1281
Child-only n=609

If we focus on children with health problems, we can see that child-only cases have higher instances of child disability. As indicated by table 14, children in child-only cases are nine percentage points more likely to have a health concern that limits their daily activity than a child in a single-parent closed case (19% v. 10%).

14The number of observations per variable changes due to the fact that these variables were in modules, and there were several filter questions and skip patterns. For example, the child-care module was randomly assigned only to respondents who had a child in the household under the age of 12.

Life After Welfare (2/19/01)

p.45

Our data also allow us to examine the emotional well-being and development of children in the

study. It was reported that 30% of children in child-only cases often or sometimes being unhappy, sad,

or depressed compared to 26% of children in single-parent cases. However, 33% of children in single-

parent cases often or sometime have trouble concentrating, compared to 26% of their child-only

counter parts. Finally, the children's ability to get along with other children is similar in child-only cases

and single-parent cases 31% in both groups are often or sometimes unable to get along with others.

Table 14

Children's Emotional Well-Being

and Development

Single-parent

Child-only

Health concerns limit activity

10%

19%

Often/sometimes does not get along

31%

31%

with others

Often/sometimes unable to concentrate

33%

26%

Often/sometimes unhappy, sad or

26%

30%

depressed

N

507

189

Children's developmental environment A key feature of a child's developmental environment is the child care arrangements his or her
care giver makes. In the child-care module, respondents with children under the age of 12 were asked "which of the following types of childcare are used for [identified child] on a regular basis, that is, at least once a week." Table 15 describes the child care arrangements respondents make. Since parents can identify multiple sources of care, each column does not total 100%. The table reveals significant differences between child-only and single-parent cases. Most commonly, children in child-only cases are cared for by a non-relative sitter (44%) or are enrolled in a Head Start center (33%). Like childonly cases, the largest portion of children in closed single-parent cases are cared for by a non-relative

Life After Welfare (2/19/01)

p.46

sitter (47%), followed by a Head Start program (42%). Among these children, however, the more

troublesome arrangements are utilized. Seventeen percent of the children are cared for by a relative

under the age of 13, and another 4% report the children stay home alone.

Table 15

Child Care Arrangements

Single-parent

Child-only

Relative under 13

17%

22%

Relative over 13

3%

11%

Non-relative sitter

47%

44%

Care in relative homes

5%

0%

Care in non-relative homes

3%

0%

Head start center

42%

33%

PreK/nursery program

5%

0%

Child stays alone

4%

0%

N

729

285

A second feature of the child's development environment is the availability of health insurance.

The number of insured children varies minimally between single-parent and child-only cases. Both are

approximately 90%. However, as noted in figure 17, children of single-parent leavers are more likely

to rely on Medicaid (92%) than child-only cases (87%).

Table 16

Children With Health Insurance

Single-parent cases

89%

Child-only cases

92%

Life After Welfare (2/19/01)

p.47

Fig. 17 Types of Insurance

100% 80% 60% 40% 20% 0% Medicaid Employer

Own Peachcare

Single-parent n=1144
Child-only n=586

A third feature of a child's development environment involves residential mobility. Single-

parent-cases are more mobile than child-only cases, though both are relatively stable. (Again, the

columns for single-parent and child-only are calculated separately and sum to 100%.) Only 21% of

single-parent cases moved once or twice in the past year, compared to 13% of child-only cases.

Among those who did move, for both case types, the most common reason was for more space

(approximately 32%) or for a better neighborhood (27% and 23%). Approximately 20% of both case

types moved because they were buying a home. Given the difficulties with locating families that move

frequently, we suspect that our findings offer a somewhat optimistic view of residential stability.

Table 17

Number of Moves in Past Year

Single-parent Child-only

Never

75%

85%

Once

15%

9%

Twice

6%

4%

More than twice

3%

3%

Life After Welfare (2/19/01)

p.48

Fig. 18 Reason for Move

40% 30% 20% 10%
0% Single-parent n=438

Child-only n=73

Morespace/betterplace Buy home LessSpace/rent better nhood

N

1144

586

Because so few of the children live with both parents, involvement with absent parents is a key feature of the developmental environment. For this reason, we asked a subset of respondents a series of questions about absent parent involvement. For the single-parent cases, the absent parent asked about was the father. Individuals who were randomly assigned the absent parent module were asked if the father was still alive and lived outside the home. For the child-only cases, the absent parent referred

Life After Welfare (2/19/01)

p.49

to was the parent who put the child in the care of the respondent, this was always the mother. Tables 18 and 19 show a very low level of absent parent involvement in children's lives for
both case types. Ninety-seven percent of the children in this survey are not living with their father. Of those, in 37% of single-parent leavers report the child's father is either dead, or `like-dead,' meaning the mother had no contact with the father and did not know where he was living. A further 25% of children had absolutely no contact with their absent parent. While the target parent differs, absent parent involvement is somewhat greater among child-only cases. For only 42% is the child's mother deal, "like dead" or never seen by the child This figure is somewhat lower than that for single-parent cases (62%). However, one should remember that the figure for the child-only cases applies to the child's mother. One can only presume that these children have no contact with their father. In that case, the figures in table 18 imply that two in five children have no contact with either parent.

Table 18
Absent Parent Dead/or "like dead" Never About once a year Several times a year 1-3 times a month About once a week Several times a week N Sub-group comparisons

Absent Parent Visits

Single-parent

Child-only

37%

16%

25%

26%

6%

8%

8%

23%

10%

12%

5%

1%

10%

14%

697

185

For the sub-group comparisons, we combine both single-parent cases and child-only cases.

The first three outcomes (issues concerning residential mobility and health insurance) were asked in the

Life After Welfare (2/19/01)

p.50

main body of the survey, so all respondents answered each of those questions. As presented in table 19, racial comparisons reveals few differences. Approximately 90% of both groups have moved no more than once in the past year. Insurance coverage and involvement with absent parent are very similar as well.
However, table 20 reveals that child outcomes do vary by respondent education levels. Higher education levels indicate more stability. Approximately 90% of those with at least a high school diploma or GED have moved no more than once in the past year. That figure is about twenty percentage points higher than those without a high school diploma or GED. Moreover, only 57% of children whose parents or care-givers do not have a diploma or GED are insured. The number of insured children increases by approximately 25 percentage points if child's parents or care-givers finished high school, earned their GED, or have some college.
Also, children who live with parents who wanted to leave TANF differ from those living with care givers who did not. Table 21 shows those who wanted to leave TANF were more likely to have never moved or moved only once compared to those who not did want to leave (89% v. 92%). Also, those who did not want to leave TANF were 4 percentage points less likely to retain health insurance for their children than those who wanted to leave (86% v. 90%). There is little to no variation by education level.15

15If anything, this suggests that our figures understate the difference between more and less educated respondents. One can only presume that the less educated moved more often. But since we only located people who moved infrequently, our figures may offer somewhat optimistic view of all families and particularly so for the less educated. Moreover, there is little difference in health insurance. This is one instance where the presence of a public policyMedicaid and Chipeliminates the difference between the more and less educated.

Life After Welfare (2/19/01)

p.51

Table 19
Never moved One move within a year Child insured
N Father16 dead, `like-dead', or never sees the child
N

Race Categories

Non-white

White

75%

75%

16%

15%

74%

73%

2303

48%

39%

5 580

Table 20

Education Categories

HS/GED

No HS/GED Some College

Never moved

82%

58%

69%

One move within a year

12%

22%

21%

Child insured

82%

57%

73%

N

2166

Father dead, `like-dead', or never sees the

27%

39%

18%

child

N

498

Table 21
Never moved One move within a year Child insured
N Father dead, `like-dead', or never sees the child
N

Wanted to Leave TANF

Yes

No

66%

88%

22%

6%

67%

87%

2228

53%

28%

579

16Mother for child-only case.

Life After Welfare (2/19/01)

p.52

Comparability Figures To put our data in context, we compared these figures first to stayers in Georgia, and second, to national figures on leavers and stayers. Comparability to stayers in Georgia Table 22 compares leavers and stayers in Georgia. (Figures on the former are provided by Risler and Nackerud, 1999). Given the large differences between child-only and single-parent cases, figures are presented for the two groups separately.17 The first three rows of the table present basic demographics for leavers and stayers. One can see that for both single-parent and child-only cases, leavers and stayers are of similar age and marital status. However, a slightly higher percentage of leavers are white. When considering background characteristics, both single-parent and child-only leavers and stayers have approximately the same number of children. However, age at first birth is greater for child-only leavers (24.5) compared to child-only stayers (18.8). The greatest difference between leavers and stayers is education level. Among single-parent cases, 24% of leavers did not have a high-school diploma or a GED compared to 45% of stayers. The difference between child-only leavers and stayers are more noteworthy. Only 20% of the respondents for child-only leavers had neither a high-school diploma nor a GED compared to 52% of respondents for child-only stayers. For both child-only and single-parent cases, more disadvantaged are accumulating on the TANF rolls.

17The time periods are somewhat inconsistent, however previous analyses of stayers and leavers for the same time period produce similar results.

Life After Welfare (2/19/01)

p.53

Not surprisingly, employment is much higher for leavers than stayers. Employment rates for single-parent leavers are approximately thirty percentage points higher than stayers. Among child-only leavers, 42% report being employed compared to only 27% of single-parent stayers. Considering home ownership, single-parent stayers are the least likely to own their own home (5%). Child-only stayers are the most likely to own their own home (48%). Finally, there are differences between health insurance for children, especially for single-parent leavers. While 76% of single-parent leavers have health insurance for their children, a full 96% of children in single-parent stayers are insured.
While differences exist, a generally optimistic attitude about ending welfare use prevails across leavers and stayers in single-parent cases. An overwhelming majority of stayers (79%) and leavers (76%) felt extremely confident that they will either get off welfare or remain off welfare in the future.

Life After Welfare (2/19/01)

p.54

Table 22

Child-Only

Single-Parent

Characteristic

Stayers Leavers Stayers Leavers

Average age (years)

43.5

43.2

29.3

29.2

Race white

29%

28%

17%

20%

Marital Status never married

29%

36%

70%

68%

Average number of kids in household

2.3

2.1

2.3

2.2

Age at first birth (years)

18.8

24.5

18.0

17.2

% who did not graduate HS or earn a GED

52%

20%

45%

24%

Monthly earnings below $1000

*

51%

63%

65%

% employed

28%

42%

39%

67%

% extremely confident they will get off/ remain

27%

72%

79%

76%

off welfare

% of children with health insurance

88%

82%

96%

76%

% who own their own home

48%

34%

5%

14%

% who "sometimes or often do not have enough to eat"

*

11%

5%

13%

* Because this question was in a module, the sample sizes for the child-only leavers are small, and we do not present them here.

National Comparability Figures

Another way to put the experiences of Georgia leavers in a context is to compare them to

experiences of leavers nationwide. For comparison data, we turned to the National Survey of

America's Families (NSAF). The study collects data on low-income families in 13 states. When

weighted, these data are representative of the non-institutionalized, civilian population of persons under

age 65.18 Recent analysis of these data (Loprest and Zedlewski, 1999) compared welfare leavers and

18For details on the NSAF, see http://newfederalism.urban.org/nsaf/

Life After Welfare (2/19/01)

p.55

current welfare recipients.19 Table 23 summarizes findings from the GSU Leavers study and the NSAF. Using the latter, we
describe both leavers and current welfare recipients. First, leavers in Georgia (column A) resemble both leavers (column B) and stayers (Column C)
in other states in terms of their gender and age. Georgia leavers differ in that they are more likely to have never been married (61% v. 44%), and more likely to be non-white (78% v. 36%). Leavers in Georgia are also somewhat better educated in that they are more likely to have completed high school or a GED (59% v. 37%). However, leavers in the national study are more likely to have some college (20% v. 27%).
In terms of their economic status, leavers from Georgia fare a bit better. Sixty-nine percent of leavers in Georgia report being employed, compared to 61% of other leavers.20 Comparisons of food insecurity also suggest Georgia leavers are faring better. In Georgia, only 38% of leavers report that they sometimes or often "worried that food would run out before we got money to buy more". This is a much lower figure than for leavers nation wide (50%).

19In terms of leavers, The findings presented in Loprest and Zedlewski 1999, consider leavers individuals who received welfare between 1995-1997 and were not receiving cash assistance at the time of the NSAF interview in 1997. The current recipients were receiving cash assistance payments at the time fo the NSAF interview in 1997.
20This difference is likely not meaningful because of methodological differences in how employment is measured. The numbers do indicate, however, that employment is roughly the same across studies.

Life After Welfare (2/19/01)

p.56

Table 23

A. GSU Leavers Study

Female

96%

Age

18-25

31%

26-35

33%

Race

White

22%

Non-White

78%

Never married

61%

3 or more children

33%

Less than HS

21%

Education

GED or HS Diploma

59%

Some College

20%

Economic

Employed

69%

Status

Food Inadequacy**

14%

* NSAF data are from Loprest and Zedlewski 1999.

NSAF*

B. Former Recipients

D. Current Recipients

93%

96%

30%

30%

44%

39%

52%

41%

36%

34%

44%

31%

33%

41%

28%

40%

37%

35%

27%

20%

61%

21%

50%

61%

** Often or sometimes "worried that food would run out before we got money to buy more"

Life After Welfare (2/19/01)

p.57

V. Conclusion Our findings provide a variety of insights into life after welfare. We find that most single-parent
leavers left for employment. Rates of employment are substantially higher than those for open cases. This difference reflects both the fact that leavers went to work as well as preexisting differences between leavers and stayers. Comparisons between the two groups (Table 22) suggest that leavers are substantially more educated than stayers.
The forces driving child-only cases to close were not apparent at first. One possibility is that economic growth lead these families to leave the rolls. Our data bear this out to some extent. Four in ten left for employment. Another reason is perhaps more obvious. Roughly one-third of child-only cases closed when the only child in the household moved out. Also included in this group are instances where the only child in the household turned 18 and became ineligible.
We also find that many but not all leavers are well informed about welfare reform. Most know that there is a lifetime limit and that the limit is four years. There is somewhat more confusion about other programs, but most know that Medicaid coverage does not end when the family leaves the rolls. Most also know that Georgia does not increase benefits when a welfare recipient gives birth while on the rolls.
On the other hand, there is some confusion about reform, at least among some sub-groups. One in five low-educated single-parent leavers say they left because of the time limits. This could either mean that they were conserving eligibility or that they believe they have exceeded the limit already. The latter implies some confusion about the terms of welfare reform. One in 7 (13%) of child-only leavers indicate that they left because of time limits, which do not apply to those cases.

Life After Welfare (2/19/01)

p.58

While most leavers are working, earnings levels are low, especially in rural areas. As a result, families piece together resources from various sources. Many leavers still participate in government programs, including Food Stamps. Many families rely on friends or family for support. Many also receive child support, but the amount of money involved is very modest. As a result, hardship is still common. One in five report having their telephone service disconnected. Two in five report that they often or sometimes worried that their food would run out.
Our figures also identify substantial diversity among leavers. This diversity is most apparent in terms of education. Individuals with a high-school degree or greater are roughly 50% more likely to identify employment as their reason for leaving the rolls than those with less than a high school degree. Barely half of the latter left the rolls for employment. This difference is also apparent in their confidence about remaining off the welfare rolls. One in five leavers (19%) with less than a high school degree were not so or not confident at all that they would remain off the rolls. This is nearly five times the rate for individuals with a high-school degree (4%). We also find some regional diversity. In areas of rural decline, earnings are especially low.
We also find substantial differences between single-parent and child-only leavers. The latter, for example, are far less likely to leave the rolls for employment. We also examined the movement between the two types of cases. In general, we find little movement from single-parent cases to childonly cases. It does not appear that the closure of single-parent cases leads to child-only cases. If anything, the flow of children from child-only cases to single-parent cases is greater. This flow might involve children moving back in with their parents or having a child of their own.
The children of welfare leavers appear to be faring reasonably well. They generally are in good

Life After Welfare (2/19/01)

p.59

health. A substantial minority, however, have a health concern that limits their activity. This rate is particularly high for children in child-only cases (19%). The vast majority of children are insured, but the rates are lower than for stayers. Residential mobility is fairly low, suggesting a reasonably stable home environment. There are, however, troublesome features of their developmental environment. These children have little involvement with absent fathers. As many as two in five children in child-only cases have no contact with either parent. Child-care arrangements also leave something to be desired. One in five children ages 12 and under are either left alone or with another relative age 12 or younger.
On the whole, therefore, leavers have much in common with the working poor, a finding that is consistent with other leavers studies (See appendix IV.) Comparisons with stayers suggest that leavers are faring better in some ways, including employment. In other ways, leavers are faring worse than stayers. In terms of insurance for their children or food adequacy, they appear worse off.
The differences between leavers and stayers must be interpreted in light of the fact that leavers are better educated. As a result, one suspects that they fared better than stayers even before leaving the rolls. The true nature of this difference is difficult to ascertain. To some extent, the advantaged position of better educated leavers suggests that schools provide useful skills, which is fairly obvious. On the other hand, differences across education levels probably capture more than the impact of years of schooling. Individuals who have more schooling likely are better motivated and have greater ability. For that reason, providing the less educated with opportunities to complete their schooling will only partially close the gap between the more or less educated leavers.
The difference between more and less educated leavers, therefore, represents a sort of upper limit to what one might accomplish by providing skills and training to the less educated. However, the

Life After Welfare (2/19/01)

p.60

difference between the two groups is so large in so many areas, better education and training may be essential to helping individuals remaining on the rolls to make a successful transition off welfare.
Where do we go from here? Clearly, continued research on leavers is needed. Our data provide only a quick snapshot of how leavers fare shortly after leaving the rolls. Longer-term followup is needed to understand how the leavers fare over time. Furthermore, the least educated and most disadvantaged remain on the welfare rolls. As the state approaches its time limit, one expects to see substantial shifts in the composition of leavers. How these families fare after leaving the rolls will be the true test of welfare reform.
Further research is also needed on the child-only cases. Additional information is needed on the households in which these children currently live. We chose to interview the adult who had been the head of the child-only case rather than the head of the household where the child now lives. As a result, our information on the child's current circumstances is somewhat limited. Moreover, additional information is needed on how the child-only cases interact with the foster care system.
Finally additional research is also needed on the stayers. The state's study of stayers is now over one year old, and it seems clear the composition of stayers has changed in that time. More current data on stayers would facilitate the interpretation of the data on leavers.

Life After Welfare (2/19/01)

p.61

References
Bachtel, Douglas C, ed. 1999. The Georgia County Guide. 17th ed. Athens: University of Georgia, Center for Agribusiness and Economic Development.
Department of Human Resources: Georgia Division of Family and Children Services. Descriptive Date: State Fiscal Year 1999. Evaluation and Reporting Section: Atlanta, Ga.
Loprest, Pamela and Sheila R. Zedlewski 1999. "Current and Former Welfare Recipients: How Do They Differ?" The Urban Institute Report 99-17.
National Survey of American Families [On-line]. Available http://newfederalism.urban.org/nsaf/.
Risler, Edward, and Larry Nackerud. 1999. "The Georgia Welfare Reform Research Project: The Remaining TANF Recipients." Athens: University of Georgia, School of Social Work.

Life After Welfare (2/19/01)

p.62