Initial Synthesis Report of the Findings from ASPE's "Leavers" Grants

Prepared by

Gregory Acs and Pamela Loprest
The Urban Institute
2100 M Street, NW
Washington, DC 20037

---

This project was funded by the US Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation under task order number 8 under contract HHS-100-99-0003. The authors would like to thank Tracy Roberts for excellent research assistance. The views expressed in this report are those of the authors alone and do not necessarily reflect the views of The Urban Institute or its sponsors.

 

"

Abstract

This report summarizes results from eleven studies of former welfare recipients funded by the Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. The studies include administrative and survey data on the well-being of families who left welfare. Efforts were made to have these studies use comparable measures to facilitate cross-study comparisons.

This synthesis focuses primarily on economic well-being. It includes information on welfare leavers' employment and earnings, income and poverty status, public program participation, material hardships, and child care use. The results of the summary show a number of common findings across these study areas. Broadly speaking, among families leaving welfare, about three out of five work at any given time after exiting, and about three-quarters have worked at some point within a year of leaving welfare. When leavers work, they usually work full-time and earn $7-$8 dollars an hour. On average, working leavers make about $3,000 a quarter, and leavers' family incomes hover around the poverty line. A significant minority of leavers return to TANF in the year after initially exiting. Over one-third of leavers receive food stamps and about 2 in 5 have public health insurance coverage during the fourth quarter following their exit from welfare. Finally, many leavers experience hardships, such as not having enough food to eat, but in general they do not experience these hardships more frequently than when they were on welfare.

I. Introduction

Since passage of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) in 1996, ending "welfare as we know it" and replacing it with the Temporary Assistance to Needy Families (TANF) block grants to states, the federal cash assistance caseloads have dropped by 50 percent, from 4.4 million in August, 1996 to 2.2 million in June, 2000. There is concern at the federal, state, and local levels about the well-being of the unprecedented number of families that have left welfare:

• Are leavers holding down steady jobs?
• Do they have the financial wherewithal to provide for themselves and their children?
• Are they and their children suffering hardships from hunger to homelessness?

To help address these questions, the US Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation (ASPE) provided funding to 10 states and three large counties to conduct studies of families that have left the welfare rolls. Here, we review and synthesize selected findings from the 11 ASPE-funded leaver studies that have submitted interim or final reports as of October, 2000, as well as an early ASPE-funded study from Wisconsin.⁽¹⁾ This initial synthesis report will be followed by a more comprehensive report in Summer 2001 which will report on survey findings from additional sites not included here.

II. Leavers Studies in Context: Policy and Economic Environment and Characteristics of Welfare Leavers

Differences in states' welfare policies, their economies, and the demographic characteristics of recipients themselves all may affect the outcomes of welfare leavers. Before comparing and contrasting the status of welfare leavers across the twelve studies reviewed in this initial synthesis report, it is important to understand the environment in which these families make their decisions to leave welfare. Because there are so many factors contributing to the well-being of leavers it is difficult to ascribe differences in outcomes across studies to any specific difference in context. In this initial synthesis report, we cannot take all these contextual differences into account simultaneously; we only note them as they come to bear on comparisons across studies.

State TANF Policies

Under TANF block grants, states have substantial flexibility in determining the length of time families can receive cash assistance (time limits), the penalties for not complying with program rules (sanctions), and the generosity of cash grants as well as how benefits are reduced as a family moves from welfare to work. First, consider time limits. While none of the studies reviewed here include families whose benefits have already expired,⁽⁶⁾ families subject to shorter time limits may feel pressure to leave welfare sooner than families that are years away from exhausting their benefits. Conversely, leavers who have nearly exhausted their benefits may be more reluctant to return. Out of the 12 sites, seven focus on leavers who are subject to the federal 60 month (5 year) life time time limit on benefits as of 1997 (see Table II.1). Georgia and Cuyahoga county have shorter life time limits-48 and 36 months, respectively.⁽⁷⁾ And Arizona, Florida, and Illinois all have intermediate time limits, not only restricting the total number of months a family can receive benefits but also prohibiting a family from receiving their life time allotment over a single time period. For example, in Arizona, families can receive benefits for 24 months in any 60 month period.

Table II.1:
State Policies, 1997

State	Time Limit	Additional Time Limit	Initial Sanction	Maximum Sanction	Maximum Benefit	Earnings Disregards
Arizona	60 months	24 out of 60 months	25% for One month or until compliance, whichever is longer	Entire Benefit for one month or until compliance, whichever is longer	$347	$90 and 30% of remainder
District of Columbia	60 months		Adult Portion of benefit until compliance	Adult Portion of benefit for six months or until compliance, whichever is longer	$379	$100 and 50% of remainder
Florida	48 months	24 out of 60 months or 36 out of 72 months1	Entire Benefit until in compliance for 10 working days	Entire benefit for three months or until in compliance for 10 working days, whichever is longer	$303	$200 and 50% of the remainder
Georgia	48 months		25% until in compliance for 3 months	Entire Benefit permanently	$280	$120 and 33.3% of remainder for first 4 months, $120 next 8 months, $90 thereafter
Illinois	60 months	24 months followed by 24 months of ineligibility	50% until compliance	Entire Benefit for three months or until compliance, whichever is longer	$377	66.7%
Missouri	60 months		25% until compliance	25% for three months or until compliance, whichever is longer	$292	$120 and 33.3% of remainder for first 4 months, $120 next 8 months, $90 thereafter.
New York	60 months		Adult Portion of benefit until compliance	Pro-rata portion for six months	$577	$90 and 45% of remainder
Washington	60 months		Adult Portion of benefit until compliance	the greater of 40% or the adult portion until in compliance for 2 weeks	$546	50%
Wisconsin2	60 months		Minimum Wage Times the Number of Hours of Nonparticipation until compliance	Entire benefit permanently	$628	None
Cuyahoga Co.	36 months		Adult Portion of benefit for one month	Entire benefit for six months	$362	$250 and 50% of remainder for first 18 months
Los Angeles Co.	60 months		Adult Portion of benefit until compliance	Adult Portion of benefit for six months or until compliance, whichever is longer	$611	$225 dollars and 50% of remainder
San Mateo Co.	60 months		Adult Portion of benefit until compliance	Adult Portion of benefit for six months or until compliance, whichever is longer	$611	$225 dollars and 50% of remainder
1The 24 out of 60 month limit applies to non-exempt recipients who have received less than 36 months of assistance during the previous 60 months and are either over age 24 or under age 24 with a high school diploma/GED.  The 36 out 72 month limit applies to non-exempt recipients who 1). have received benefits for 36 of the previous 60 months or 2). are under age 24, have not completed high school/ GED, are not enrolled in a high school equivalency program, and have little or no work experience.
2All of the policies reported are for the W-2 Transition component.
Sources: See Appendix B for a complete listing of the leavers studies referenced.
Data reported from Urban Institute's Welfare Rules Database.  All data are reported as of 7/97.

Next, consider states' sanction policies. In general, states impose tiered sanctions, beginning with less severe sanctions at first and escalating penalties for repeated instances of non-compliance. The states conducting leaver studies reviewed here tend to have less severe sanction policies than other states. Note that leavers who were sanctioned off the rolls may have a particularly hard time after exiting welfare and may return to TANF at higher rates as some sanctioned leavers come back into compliance with program requirements and rejoin the TANF rolls. Table II.1 shows the initial and maximum sanction in each of the ten states covered by the leaver studies. Focusing on the most severe sanction, we see that five of the ASPE leaver studies (Arizona, Florida, Georgia, Illinois, and Cuyahoga county) are based in states that impose full-family sanctions, removing the adult unit head and the children from the TANF rolls. Further, Wisconsin not only imposes a full-family sanction, but it is also a lifetime sanction; families that reach this point can never come back into compliance and return to cash assistance in Wisconsin.

Finally, consider differences in TANF generosity. Table II.1 shows the maximum TANF benefit a family of three can receive and the earned income disregards prevailing in the ten states covered by the ASPE leaver studies. California -- the site of both the Los Angeles and San Mateo studies -- clearly has the most generous policies, with a high maximum benefit and large earnings disregards while Wisconsin has high benefits but no earnings disregards. Cuyahoga county and Florida have modest benefits but generous earnings disregards, and New York and Washington have high benefits and earnings disregards that are slightly more generous than average.

While we have not reviewed every aspect of states' TANF policies (for example, we have ignored work requirements and diversion policies ⁽⁸⁾), we can make some general observations about the policy context in which these 12 ASPE leaver studies are based. For example, California (LA and San Mateo studies), New York, and Washington generally pursued policies that would be expected to produce lower exit rates from welfare but higher incomes for those families that do leave. Conversely, Arizona and Georgia may move families off the welfare rolls faster but their leavers may face more difficulties. Finally, other studies are based in states that pursue a mix of policies that are likely to have offsetting effects on the outcomes of leavers-for example, Cuyahoga county has strict time limits and sanctions but very generous earnings disregards.

Economic Context

One would expect that when jobs are plentiful and wages are high, welfare leavers will generally fare better than during lean economic times. Table II.2 shows the 1997 unemployment rates and median incomes in the states in which the leaver studies we review were conducted. Note that we report state averages and some of the leaver studies focus more narrowly. For example, the economic conditions in Ohio may not necessarily reflect the conditions in Cuyahoga county.

Overall, we find that Wisconsin had both the lowest unemployment rate (3.7 percent) and the highest median income at $43,132. Florida had a low unemployment rate at 4.3 percent, but its median income is also among the lowest at $32,455. In contrast, the District of Columbia had both a relatively high unemployment rate (8.9 percent) and low median income ($32,382). Thus, it is clear that economic conditions vary considerably across the sites conducting the leaver studies reviewed here.

Table II.2:
Economic Characteristics of States, 1997

State	Unemployment Rate (%)	Median Income ($)
Arizona	5.1	35,503
District of Columbia	8.9	32,382
Florida	4.3	32,455
Georgia	4.9	35,911
Illinois	5.2	40,094
Missouri	4.8	36,676
New York	6.3	34,783
Washington	6.4	37,458
Wisconsin	3.7	43,132
Cuyahoga Co.1	4.8	36,798
Los Angeles Co.1	7.8	38,976
San Mateo Co.1	7.8	38,976
United States	5.6	36,244
1 Unemployment rate is given for the entire state.
Source: "Interpreting TANF Leaver Studies: Comparing ASPE Grantee States to the Nation as a Whole."  Mathematica Policy Research, March 27, 2000.

Characteristics of Welfare Recipients

Differences in the personal characteristics of welfare recipients and welfare leavers also must be considered when comparing findings across leaver studies. Indeed, states whose welfare caseloads are markedly different may well produce leavers with very different characteristics. Thus part of any difference in outcomes across sites may be due to differences in leavers themselves. Further, states likely structure their welfare policies with their welfare populations in mind-for example, a state with a high proportion of high school drop outs may emphasize work readiness programs-and this too may affect the status of leavers.

Table II.3 highlights some differences in the characteristics of welfare recipients across the 12 study sites we examine here. First consider the ages of adults heading families on welfare. In 1997, the age distribution of case heads was fairly similar across our 12 study areas: in general about 6 percent of cases were headed by teenagers, more than 40 percent were headed by adults 20 to 29 years old, and more than 50 percent by adults age 30 or older. The exceptions are California (covering both LA and San Mateo counties) which has a disproportionate number of young and old case heads and Wisconsin which has a disproportionately low share of leavers age 30 and over.

Table II.3:
Characteristics of TANF Caseloads:
July - September 1997

	Age of Unit Head (%)			Number of Children1 (%)		Age of Youngest Child1 (%)			Educational Attainment1 (%)
State	< 20	20-29	30+	< 2	3+	<1	1-5	6+	< High School	High School +
Arizona	6	40	54	62	39	18	47	36	44	56
District of Columbia	6	42	53	76	24	14	49	37	55	45
Florida	7	43	50	74	27	14	45	41	55	45
Georgia	6	45	50	74	26	9	46	46	50	50
Illinois	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.
Missouri	5	45	49	74	26	12	47	42	47	53
New York	6	40	54	75	25	9	50	41	45	55
Washington	5	41	55	77	23	12	44	44	45	55
Wisconsin	8	52	39	68	32	17	46	37	74	27
California	12	29	60	75	25	11	55	34	50	50
Los Angeles Co.   San Mateo Co.
Ohio	8	48	44	75	25	15	46	39	44	56
Cuyahoga Co.
United States	8	40	52	74	26	12	48	40	47	53
1 Data are adjusted for the percentage of unknown responses.
Source: Data reported from ACF's National Emergency TANF Datafile as of 12/9/98.  Available on-line at http://www.acf.dhhs.gov/programs/opre/particip.table2.htm

Next, consider differences in the number and ages of children across our study areas. In general, three quarters of families on welfare have two or fewer children (see table II.3). The exception is Arizona, in which 38.5 percent of welfare families have three or more children. About 10 percent of families have an infant across our study locations, and about 40 percent have youngest children who are school age (6 or older). Arizona and Wisconsin are slight exceptions, with over 17 percent of welfare families containing an infant. And Arizona, Wisconsin, and California have a lower proportion of cases in which the youngest child is school-aged than elsewhere.

Finally, we examine educational difference among welfare recipients across our study areas and find only minor differences between 11 of our 12 study sites (table II.3).⁽⁹⁾ The share of welfare recipients with less than a high school degree ranges from a low of 44.2 percent in Arizona to a high of 54.0 percent in DC and Florida.

Overall, there are few significant differences in the characteristics of welfare caseloads across our locations. Nevertheless, it is important to keep in mind even small differences in caseloads, caseload declines, and the characteristics of leavers when comparing the status of leavers across studies.

III. Employment and Earnings

A central goal of welfare reform is moving families from welfare to work and, ultimately to self-sufficiency. Eleven of the ASPE-funded leaver studies included here use administrative data from their states' Unemployment Insurance systems to examine the employment and earnings of TANF leavers in the months and years following exit. In addition, five of the studies in this initial synthesis report also use surveys of leavers to examine employment, wage rates, and job characteristics of leavers. While surveys rely on self-reported information, they garner more detailed information than is available through administrative data systems.⁽¹⁰⁾

It is important to note that administrative data likely under represent the amount of work performed by TANF leavers. First, leavers who work across state lines or move to another state entirely will not appear in a state's UI system. Second, not all jobs are reported to a state's UI system-for example, leavers who are self-employed or who work in certain jobs in agriculture or in the federal government are not included in UI data systems. And finally, leavers who are domestic service workers (like nannies or housecleaners) may not appear in the UI system because their employers fail to report them. Thus, some of the leavers who appear to have "never worked" in administrative data may actually be bringing in earnings in some form. Further, some non-working leavers may well have a working spouse or partner.

Across the eleven studies using employment data from state UI systems, we find that slightly over half of all leavers work in the first quarter after exiting welfare. As table III.1 shows, the first quarter employment rates are tightly bunched, ranging from a low of 47 percent in Los Angeles to a high of 64 percent in Georgia.

Table III.1:
Employment of Single-Parent Welfare Leavers:
Administrative Data Findings

State/Study	Exit Cohort	Post-Exit Quarter (%)				Ever Worked After Exit (%)	Worked All Four Quarters (%)
		Q1	Q2	Q3	Q4
Arizona	1Q98	53	51	52	50	73	n.a.
Florida	2Q97	50	51	53	54	71	31
Georgia	1Q97	64	60	59	53	74	n.a.
Illinois	3Q97-4Q98	54	53	54	55	70	39
Missouri1	4Q96	58	58	59	58	n.a.	n.a.
New York	1Q97	50	49	48	48	62	40
Washington	4Q97	57	57	58	57	n.a.	n.a.
Wisconsin	July 1995- June 1996	63	62	61	62	75	n.a.
Cuyahoga Co.2	3Q96	59	54	56	57	72	40
Los Angeles Co.2	3Q96	47	46	46	47	n.a.	35
San Mateo Co.	1997	50	50	48	50	67	n.a.
1 Missouri reports employment data for all cases, not just for single parent cases.
2 Los Angeles and Cuyahoga counties require a leaver to have at least $100 in earnings to be considered working while others require only $1.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Over time, employment rates of leavers could rise as more leavers find jobs or they could fall as employed leavers encounter difficulties and lose their jobs. Interestingly, table III.1 shows that employment rates do not change much over time. Indeed, the median employment rate reported in these studies is 54 percent in both the first and fourth post-exit quarters. Wisconsin has the highest fourth quarter post-exit employment rate (62 percent) and New York, the lowest (48 percent).

While slightly over half of all leavers work in any given post-exit quarter, it is not uncommon for leavers to cycle in and out of jobs; consequently, the share of leavers who ever worked over the year after exit is considerably higher and the share who worked in all four quarters is considerably lower. Eight of the eleven studies report the share of leavers who worked in any of the four post-exit quarters. They find that about 70 percent of leavers worked in at least one quarter. In six out of eight studies over 70 percent worked (with a high of 75 percent in Wisconsin); 67 percent worked in San Mateo county and 62 percent worked in New York. Five of the eleven studies report the share of leavers who worked in each of the four post-exit quarters. They find between 35 and 40 percent of leavers worked in all four quarters.

Although there are some methodological differences between studies-for example, Los Angeles and Cuyahoga county require a leaver to have at least $100 in earnings to be considered working while others require only $1-these differences do not account for much of the meager variation across studies. Indeed, Cuyahoga county consistently reports high employment rates despite having a higher threshold for employment. Thus overall, these findings from administrative data suggest that the majority of welfare leavers work or have worked since exiting, but nearly three out of ten have never worked in the year following exit.

As of mid-October, 2000, five jurisdictions had submitted findings from surveys of TANF leavers where they asked the leavers themselves about their current employment status. The responses of leavers generally refer to employment about 6 months to a year after exit. Table III.2 compares these self-reported employment rates with fourth quarter post-exit employment rates computed from administrative data. The surveys consistently find higher employment rates than those reported in UI wage records; for the most part, they are about 7 percentage points higher. Illinois' survey presents some instructive information. In its administrative records, Illinois finds that 30 percent of leavers never worked over the first four post-exit quarters. In its survey, Illinois finds that only 15 percent of leavers say they have never worked in the 6 to 8 months since exiting TANF.

Table III.2:
Employment of Welfare Leavers:
Comparison of Administrative and Survey Findings

State/Study	Exit Cohort	Timing of Survey	Employment Rate (%)
			Survey Data	Administrative Data1
Arizona2	1Q98	12-18 months	57	50
District of Columbia	4Q98	~12 months	60	n.a.
Illinois2	Dec. 1998	6-8 months	62	55
Missouri	4Q98	26-34 months	65	58
Washington	Oct. 1998	6-8 months	59	57
1 Based on employment rate from the 4th post-exit quarter.
2 Employment data reported for single- parent cases.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Beyond employment, it is also important to examine the quality of the jobs held by TANF leavers. The most basic measure of job quality is how much the job pays. Table III.3 shows the quarterly earnings of employed TANF leavers based on UI wage records reported in ten leaver studies. These records include earnings information on all reported jobs a leaver has held during the quarter. The earnings of employed leavers during the first post-TANF exit quarter range from about $2,100 to about $3,500.⁽¹¹⁾ Over time, average earnings rise, especially in states with lower average earnings; during the fourth post-TANF exit quarter, earnings range from about $2,400 to about $3,600. Note that these figures represent total earnings over a three month period. The data do not provide information on the number of weeks or hours leavers actually worked to achieve their earnings.

Table III.3:
Mean Earnings of Employed Single-Parent Leavers:
Administrative Data Findings

State/Study	Exit Cohort	Post-Exit Earnings ($)
		Q1	Q2	Q3	Q4
Arizona	1Q98	2,211	2,354	2,695	2,511
Florida	2Q97	2,163	2,352	2,343	2,496
Georgia	1Q97	2,193	2,272	2,549	2,389
Illinois	3Q97-4Q98	2,663	2,746	2,846	2,959
Missouri1	4Q96	2,192	2,360	2,384	2,698
New York	1Q97	3,393	3,402	3,877	3,602
Washington	4Q97	2,678	2,906	2,975	3,275
Wisconsin	July 1995- June 1996	2,440	2,509	2,563	2,686
Cuyahoga Co.	3Q96	2,756	2,756	2,891	2,952
Los Angeles Co.	3Q96	3,414	3,387	3,521	3,576
San Mateo Co.	1997	3,124	3,407	3,457	3,647
1 Missouri reports earnings data for all cases, not just for single parent cases.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Surveys of welfare leavers obtain information on the hours worked and wages of leavers (see table III.4). All five surveys show that employed leavers work close to full-time on average, with mean weekly hours ranging from 35 to 39 and median hours (when reported) reaching 40. Average wages range from $7.52 to $8.74 an hour. If a leaver works 40 hours a week and earns $7.50 an hour, she would earn $3,900 in a quarter providing that she worked all thirteen weeks in the quarter. This is higher than the earnings reported in the UI wage records indicating that some leavers move in and out of jobs and are not continuously employed.

Table III.4:
Hours and Wages of Welfare Leavers:
Survey Findings

State/Study	Exit Cohort	Timing of Survey Post Exit	Hours Worked		Wage Rate
			Mean	Median	Mean	Median
Arizona1	1Q98	12-18 months	n.a.	n.a.	$7.52	n.a.
District of Columbia	4Q98	~12 months	36	40	$8.74	$8.13
Illinois	Dec. 1998	6-8 months	n.a.	37	$7.89	$7.41
Missouri	4Q98	26-34 months	39	40	n.a.	n.a.
Washington1	Oct. 1998	6-8 months	36	40	$7.70	$7.00
1 Earnings data reported for single-parent cases.
Source: See Appendix B for a complete listing of the leavers studies referenced.

In addition to monetary pay, employed leavers may receive non-cash employee benefits through their jobs. Three of the surveys ask explicitly about job-related benefits, and unlike employment and cash earnings, there is considerable variation across the studies. Table III.5 shows that the share of leavers with health insurance is 32 percent in the District of Columbia and 36.0 percent in Washington state. In Missouri, 53 percent of working leavers are offered employer-sponsored health insurance although some decline coverage. The share with paid vacations runs from a low of 31 percent (Washington state) to a high of 62 percent (DC), and the share with paid sick days ranges from 28 percent (Washington state) to 50 percent (DC). Finally, 46 percent of working leavers have retirement benefits in DC compared to 21 percent in Washington state.

Table III.5:
Employer Sponsored Benefits of Welfare Leavers:
Survey Findings

State/Study	Exit Cohort	Timing of Survey Post Exit	Percent of Leavers (%)
			Health Insurance	Paid Sick Leave	Paid Vacation	Pension
District of Columbia	4Q98	~12 months	32	50	62	46
Missouri	4Q98	26-34 months	53 2	40	52	n.a.
Washington1	Oct. 1998	6-8 months	36	28	31	21
1 Benefits reported for single-parent cases.
2 This is an offer rate, not a coverage rate.
Source: See Appendix B for a complete listing of the leavers studies referenced.

IV. Income and Poverty

Income and poverty status are important indicators of the well-being of welfare leavers. Yet few leaver studies fully examine the income of leavers. There are several important reasons for this. First, most leaver studies focus on the first year after leaving-very few leavers are likely to achieve economic security in such a short period of time. Indeed, most leavers will have low incomes. Second, income is very hard to measure accurately. Perhaps the most reliable source of income data are tax records; such records are highly confidential and are rarely available for research purposes. Most information on income comes from survey data, but to obtain fairly accurate income information, the survey generally must devote a great deal of time to ask about each possible income source and then obtain the amount. Few of the leaver study surveys could afford to devote this level of time to obtaining income data. Finally, income is only one measure of well-being, and many surveys ask about explicit hardships leavers face. We discuss this hardship information in a later section.

Four of the leaver studies reviewed here obtain data on income through surveys (Arizona, Illinois, Missouri, and Washington state). Table IV.1 shows that mean monthly incomes range from $964 in Illinois to $1,427 in Missouri. ⁽¹²⁾ It is important to note that Illinois and Washington report monthly income six to eight months after exit while Missouri examines income two to three years after exit. Further, Illinois' survey simply asks respondents to estimate their monthly incomes while Missouri's survey asks a detailed set of questions about income sources and amounts. Finally, Missouri includes two-parent families in its findings. Thus, these factors may help explain why Missouri leavers appear to have higher monthly incomes.

Table IV.1:
Income of Welfare Leavers:
Survey Findings

State/Study	Exit Cohort	Timing of Survey	Monthly Income1 ($)
			Mean	Median
Arizona2	1Q98	12-18 months	1,269 3	n.a.
Illinois2	Dec. 1998	6-8 months	964	800
Missouri	4Q98	26-34 months	1,427	1,166
Washington2	Oct. 1998	6-8 months	1,208	1,000
1 Income data are reported for cases in Arizona, households in Illinois and Missouri, and families in Washington.
2 Income data reported for single- parent cases.
3 Arizona reports income including food stamps; we present an adjusted version of income, reducing reported income by 7% because 7% of the average family income of welfare leavers in Arizona comes from food stamps.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Since about three out of five leavers are working, it is not surprising to find that 60 to 65 percent of leavers in the three surveys that ask about the sources of their income report having income from earnings (table IV.2). In addition, leavers may have access to the earned income of other household members. In Washington state, 21 percent of leavers have access to another's earnings, and in Missouri and Illinois, 86 and 80 percent of leavers, respectively, are in families in which someone is earning money (see table IV.2). Another source of income for leavers is child support; between 11 and 31 percent of leavers receive child support. Finally, leavers may receive support from public programs like SSI, and some may have returned to TANF. We examine the use of public programs later.

Table IV.2:
Leavers' Income from Private Sources:
Survey Findings

State/Study	Exit Cohort	Timing of Survey Post Exit	Percent of Leavers With Income From:
			Own Earnings	Other Earnings	Any Earnings	Child Support
District of Columbia1	4Q98	~12 months	60	n.a.	64	11
Illinois	Dec. 1998	6-8 months	63	n.a.	86	31
Missouri	4Q98	26-34 months	65	n.a.	80	22
Washington1	Oct. 1998	6-8 months	60	21	n.a.	23
1 Income data reported for single- parent cases.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Regardless of the sources of their income, the time period over which their income was measured, or the rigor with which they were asked about their incomes, all leaver studies that examine income show that leavers' incomes hover near the poverty line. Missouri and Washington explicitly examine the poverty status of leavers and both find that 58 percent of leavers have cash incomes that leave them in poverty (table IV.3). Note that Washington's study indicates that 42 percent of leavers are not poor 6 to 8 months after leaving welfare-the same share reported by Missouri over two years after its leavers exited welfare.

Table IV.3:
Poverty Among Welfare Leavers:
Survey Findings

State/Study	Exit Cohort	Timing of Survey Post Exit	% Poor	% Below 185% of Federal Poverty Level
Missouri	4Q98	26-34 months	58	89
Washington1	Oct. 1998	6-8 months	58	n.a.
1 Poverty data reported for single-parent cases.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Missouri's study also shows that almost nine out of every ten leavers have incomes below 185 percent of the federal poverty line. Low income leavers may be eligible for various in-kind public supports such as food stamps, WIC, child care assistance, and Medicaid or other public health insurance programs all of which can enhance their well-being and are not included in these poverty measures. In addition, low-income working leavers can receive cash support from the Earned Income Tax Credit, currently the largest program supporting low-income families in the US.

V. Program Participation

A goal of welfare reform is to reduce families' dependency on cash assistance and make that dependent period temporary. Part of the measure of success for those who exit TANF is the ability to move toward self-sufficiency in caring for their families. As families strive toward this goal, some government benefits remain in place to aid in the transition off cash assistance. These include food stamps, Medicaid, and child care assistance. Families who have low enough incomes and meet other program requirements continue to be eligible for these benefits. In this section we address the extent to which former TANF/AFDC recipients continue to receive these benefits over time.

For some who exit TANF, the transition is not permanent, and they return to the cash assistance roles. While there has always been a percentage of families who move on and off cash assistance, examining those returning to TANF takes on increased importance in light of the time limits on benefit receipt. While none of the leavers in the areas and cohorts we examine here have reached Federal or state time limits during the study time period, returning to TANF means further depleting their limited benefit time. In this section, we first examine returns to TANF and then examine participation in other government benefit programs.

Returns to TANF

Despite the relatively high employment rates shown earlier, in most of our study areas a sizeable minority of TANF exiters return to cash assistance in the first year after leaving. In eight out of eleven areas, more than 15 percent of former recipients were again receiving TANF at the end of the first year (table V.1).⁽¹³⁾ There is significant difference in the rate of returns across study areas. The highest rates according to administrative data were in Cuyahoga County, with 28.7 percent, and Missouri with 20.5 percent of leavers returning to cash assistance. San Mateo County had the lowest percent returning at 8.0 percent and Florida had the second lowest at 11.0 percent.

Table V.1:
Percent of Single-Parent Leavers Returning to TANF:
Administrative Data

State	Exit Cohort	Post- Exit Quarter (%)				Receipt Any Time in Year After Exit
		Q1	Q2	Q3	Q4
Arizona1	1Q98	5	13	17	16	28
District of Columbia1,2	4Q98	8	13	16	19	21
Florida1	2Q97	7	14	13	11	25
Georgia	1Q97	n.a.	n.a.	14	13	n.a.
Illinois1	3Q97-4Q98	16	19	18	16	29
Missouri2	4Q96	13	18	21	21	n.a.
New York	1Q97	n.a.	n.a.	n.a.	19	n.a.
Washington	4Q97	8	14	16	16	23
Wisconsin	July 1995-June 1996	14	19	19	17	28
Cuyahoga Co.	3Q96	20	28	30	29	35
San Mateo Co.1	1997	7	9	10	8	18
1 These studies report data for month after exit, not quarter.
2 These studies report results for all cases.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Despite some differences in return rates, in general, the pattern of returns to TANF is somewhat similar across study areas. The number of leavers who have returned to TANF after just one quarter, ranges from a low of 6.7 percent in San Mateo to 20.4 percent in Cuyahoga County.⁽¹⁴⁾ This percentage jumps up by the end of the second quarter and remains relatively steady for the rest of the year. While a few areas have slight declines in receipt over time from the second to the fourth quarter, in general over the first year we do not observe declining returns to TANF.

There are indications that at least in some areas, these steady numbers of TANF recipients masks a fair degree of "cycling" -- families returning to TANF and then leaving again. Eight of the eleven study areas report the percentage of families who ever received TANF in the year after initially exiting. These numbers range from 18 to 35 percent. A much higher percent of leavers ever returned to TANF over the course of the year than are on at the end of the year. This indicates a degree of movement on and then off again. For example, Arizona reports that 27.7 percent of leavers in the study cohort returned to TANF over the next year, although only 15.5 percent were on in the twelfth month after leaving. This means almost half of those who returned to TANF had left again.

One difference across the eleven study areas is the time period covered. At first it seems that earlier cohorts of leavers may have greater return rates because some of the highest returns are in areas with the earliest cohort periods, Cuyahoga County (3Q96) and Wisconsin (July 1995). However, some of the more recent study periods (District of Columbia in 4Q98 and Arizona in 1Q98) have higher returns than other study areas with earlier cohorts. In fact, the levels of returns are different across areas with similar study periods. This highlights that returns to welfare are determined by a number of factors, so cross-study comparison is difficult. However, within the few studies that report multiple cohorts, no clear pattern of returns over study cohorts is observed.

Finally, using the little information we have, it is interesting to examine how return rates vary by the time limit policy in these areas. As we noted earlier, shorter time limits may encourage families to leave earlier than in other areas, possibly families that are less "prepared" to make the transition. Illinois had a 24 month time limit during the study period and had high returns and San Mateo had a 60 month time limit with low returns.⁽¹⁵⁾

The previous results are all from administrative data sources. For a number of reasons this source has advantages over survey data for examining program participation. Survey data can have errors due to faulty respondent memory and misinterpretation of questions. However, survey data on program participation does have the advantage of capturing benefit receipt for those who have moved out-of-state and no longer appear in the original state's administrative data. Four studies report results on returns to TANF from their surveys of former recipients (table V.2). These results are generally comparable to the administrative data results for the same time period. They reinforce that while a significant percentage return to welfare, many who return exit again in the time period prior to the survey (which ranges from 6 to 30 months after exit). Of all the surveys, Missouri allows the longest-term picture of returns to TANF, two and a half years after the initial exit. At this time, 14 percent of leavers report they are back on TANF in Missouri.

Table V.2:
Percent of Leavers Returning to TANF:
Survey Data

State	Exit Cohort	Timing of Survey Post Exit	Since Exit (%)	At Time of Survey (%)
District of Columbia	4Q98	~12 months	25	19 2
Illinois	Dec. 1998	6-8 months	19	14
Missouri	4Q98	26-34 months	31	14
Washington1	Oct. 1998	6-8 months	n.a.	19
1 Single-parent cases.
2 Month prior to survey.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Food Stamp Participation

Food stamp benefits can add a significant amount to total income for a low-wage worker. Families with incomes below 130 percent of the federal poverty line can qualify for benefits. Generally, TANF recipients are also eligible for food stamps. Only three of our studies report food stamps receipt prior to leaving TANF, but all show over 80 percent participation.

TableV.3:
Percent of Single-Parent Leavers Receiving Food Stamps:
Administrative Data

State	Exit Cohort	Quarter Relative to Exit (%)						Receipt Any Time in Year After Exit
		Q-1	Q01	Q1	Q2	Q3	Q4
Arizona2	1Q98	83	45	39	39	38	35	67
District of Columbia2,3	4Q98	n.a.	70	36	38	37	38	n.a.
Florida2	2Q97	n.a.	79	45	41	38	35	67
Illinois2	3Q97-4Q98	85	n.a.	33	35	34	33	56
Missouri3	4Q96	n.a.	63	57	47	43	40	n.a.
New York	1Q97	n.a.	n.a.	n.a.	n.a.	n.a.	26	n.a.
Washington3,4	4Q97	90	91	46	42	40	36	n.a.
Wisconsin	July 1995-June 1996	n.a.	90	51	46	43	40	63
Cuyahoga Co.	3Q96	n.a.	n.a.	43	42	41	39	n.a.
San Mateo Co.2	1997	n.a.	100	9	11	11	10	24
1 Q0 represents the month of exit for studies reporting monthly data, while Q1 represents the third month after exit, Q2 the sixth month, etc.
2 Studies report data for month after exit, not quarter.
3 Studies report results for all cases.
4 In Washington, each adult in two- parent case is counted separately.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Food stamp receipt drops significantly after exiting TANF. In seven out of the eight study areas reporting food stamp benefit receipt in the quarter after exit, less than half of leavers are receiving food stamps. The actual percentages range from 57.2 percent in Missouri to 8.8 percent in San Mateo County.⁽¹⁶⁾ These low rates are at least in part due to some recipients no longer being eligible for food stamps. Unfortunately, without better income information we cannot assess to what extent this is the reason for the decline in receipt.

Over the first year after exit, five of the eight study areas show declines in food stamp receipt although only three are large. Those three areas, Florida, Missouri, and Washington, had the largest receipt rates in the first quarter after exit. By the end of the first year after exiting welfare, receipt of food stamps is between 33 and 40 percent in 6 out of 8 areas. Only San Mateo county and New York are exceptions.

These numbers conceal a more extensive decline in food stamp receipt among those who have permanently left TANF because they include those who have returned to TANF and are likely receiving food stamps. Three studies (District of Columbia, Illinois, and Florida) report food stamp receipt for continuous leavers, those who did not return to TANF (not shown in table). ⁽¹⁷⁾ The percentages of continuous leavers receiving food stamps one year after leaving TANF are considerably lower than the overall percentages receiving food stamps. For example, in Florida 35 percent of all leavers are receiving food stamps one year after leaving, compared with only 21 percent of continuous leavers. The pattern is similar in Illinois, 33 percent compared with 21 percent, and in District of Columbia, 38 percent compared with 23 percent.

A few studies report the percentage of leavers who received food stamps at any point over the year after exit. These numbers are significantly higher than the percentage receiving in any of the individual quarters. For example, 66.5 percent of Arizona leavers received food stamps at some point in the year after exiting TANF, but only between 35 and 39 percent are receiving benefits in any individual month. This suggests that while the receipt of food stamps seems fairly stable over time, there is actually a substantial degree of turnover with recipients leaving and entering the food stamp caseload.

There is information on food stamp receipt from survey data in four of the studies (table V.4). Where the timing of this information overlaps with that of administrative data, the results are similar. Missouri again affords the opportunity to observe program participation a fairly long time after leaving. In this study, 47 percent of leavers are receiving food stamps 30 months after exit, the same percent receiving food stamps in the second quarter after exit according to Missouri's administrative data.

Table V.4:
Percent of Leavers Receiving Food Stamps:
Survey Data

State	Exit Cohort	Time of Survey Post Exit	Since Exit (%)	At Time of Survey (%)
District of Columbia	4Q98	~12 months	55	41
Illinois	Dec. 1998	6-8 months	44	33
Missouri	4Q98	26-34 months	n.a.	47
Washington1	Oct. 1998	6-8 months	50	n.a.
1 Single-Parent cases.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Medicaid and Other Health Insurance

Another important benefit that can support the transition from welfare to work is public health insurance through the Medicaid program. ⁽¹⁸⁾ Like food stamps, families receiving TANF are generally eligible for this benefit. This is borne out in the high rates of receipt, above 92 percent, in the three studies reporting Medicaid coverage in the quarter prior to exiting TANF (table V.5). Most families exiting welfare through employment are eligible for Transitional Medical Assistance and most children in low-income families are eligible for Medicaid.

Table V.5:
Percent of Leavers Receiving Medicaid:
Administrative Data

State	Exit Cohort	Quarter Relative to Exit (%)						Ever Received Within Year After Exit
		Q-1	Q0	Q1	Q2	Q3	Q4
Adults
Arizona1,2	1Q98	96	55	50	45	42	37	72
District of Columbia2,3	4Q98	n.a.	98	42	47	47	48	n.a.
Florida2	2Q97	n.a.	100	55	52	48	45	74
Illinois2	3Q97-4Q98	93	42	57	52	47	40	69
Missouri3	4Q96	n.a.	n.a.	41	39	34	31	n.a.
New York	1Q97	n.a.	n.a.	n.a.	n.a.	n.a.	354	n.a.
Washington3,5	4Q97	93	99	53	49	46	43	n.a.
Wisconsin	July 1995-June 1996	n.a.	n.a.	76	69	66	63	82
Cuyahoga Co.	3Q96	n.a.	n.a.	41	42	40	38	n.a.
San Mateo Co.2	1997	n.a.	100	26	26	22	22	47
Children
Florida2	2Q97	n.a.	100	62	58	54	51	78
Missouri3	4Q96	n.a.	n.a.	81	86	97	87	n.a.
New York	1Q97	n.a.	n.a.	n.a.	n.a.	n.a.	344	n.a.
San Mateo Co.2	1997	n.a.	100	28	28	25	24	50
1 Only head of household enrollment considered.
2 Studies report data monthly, not quarterly.
3 Studies report results for all cases.
4 In New York, four quarters after exit, 45% of cases have any member with Medicaid.
5 All individuals are included, adults in two-parent households are counted separately.
Source: See Appendix B for a complete listing of the leavers studies referenced.

In the studies reporting Medicaid administrative data, however, we see relatively low rates of Medicaid coverage for both adults and children in the first months after exiting TANF. For adults, whom we expect to have lower Medicaid coverage than children, the results range from 26.4 percent in San Mateo County to 75.9 percent in Wisconsin. Five out of nine study areas reporting coverage in the first quarter (or third month) after leaving show less than half of adult leavers with Medicaid coverage.⁽¹⁹⁾ Only four studies report separate administrative data numbers for children. The coverage rates here vary quite a bit, with 61.6 percent coverage in Florida the third month after exit and 27.6 percent coverage in San Mateo County at the same point. However, in both of these areas, coverage for children is higher than for adults. New York is the only study reporting higher Medicaid coverage for adults than children (35 percent versus 34 percent) although the ranking reverses when considering single and two-parent families together.

Over time, the rate of Medicaid coverage declines for both adults and children in all study areas except the District of Columbia. As with food stamp receipt, the decline would likely be greater if we separated out those leavers who are receiving Medicaid after returning to TANF. The four studies reporting Medicaid use by continuous leavers show this more dramatic decline (not shown in table). Rates of Medicaid coverage one year after exiting TANF are substantially lower for continuous leavers than for all leavers in DC, Florida, Illinois, and Washington.⁽²⁰⁾ This is not unexpected since after six months, transitional Medicaid benefits are income-tested.

Also similar to food stamp receipt, the percentage of leavers who received Medicaid at any time over the year after exit is significantly higher than the percent receiving in any particular month or quarter. For example, in Florida, 68.8 percent of adults and 77.8 percent of children had received Medicaid at some point over the year, but only 40.0 percent and 50.8 percent of adults and children respectively were receiving benefits in the twelfth month after exit. This signifies a fair amount of turnover in Medicaid receipt.

In the area of health insurance coverage, survey data can add a great deal to our knowledge. It allows us to ascertain coverage by private sources as well as public and to discover the percentage of persons with no coverage at all. Five studies report survey data for insurance coverage of adults and children at the time of the survey (table V.6). The percentages for Medicaid are reported in the first column. This information roughly corresponds to the administrative findings where similar time periods are available, although survey reports of Medicaid tend to be higher than administrative reports.

Table V.6:
Percent of Leavers with Health Insurance by Coverage Type:
Survey Data

State	Exit Cohort	Timing of Survey Post Exit	Type of Coverage at Survey
			Medicaid	Employer1	Other	Uninsured
Adult
Arizona2	1Q98	12-18 months	39	15	5	40
District of Columbia	4Q98	~12 months	54	19	4	22
Illinois	Dec. 1998	6-8 months	473	214	n.a.	36
Missouri	4Q98	26-34 months	33	25	9	32
Washington2,5	Oct. 1998	6-8 months	53	13	12	26
Child
Arizona2	1Q98	12-18 months	51	12	8	26
District of Columbia	4Q98	~12 months	60	12	11	16
Illinois	Dec. 1998	6-8 months	533	234	n.a.	29
Missouri	4Q98	26-34 months	68	20	3	8
Washington2	Oct. 1998	6-8 months	67	9	10	13
1 Employer for adults includes own employer coverage for survey respondents. Spouse employer coverage, where reported separately (Missouri and Washington) is included in other.  For children, employer includes all employer coverage.
2 Single-parent cases
3 Illinois is the only area reporting Medicaid coverage since exit as well as at the time of survey.  Results since exit are 58.0% for adults and 62.0% for children.
4 Includes all private coverage.
5 Multiple responses allowed.
Source: See Appendix B for a complete listing of the leavers studies referenced.

The rates of uninsured among adults and children vary substantially across survey areas. For adults, the range is from 22.4 percent in the District of Columbia to 40.0 percent in Arizona. The range is due in part to the range in Medicaid coverage for adults across the states as well as, to a lesser extent, differences in private coverage. The lower rate of adult uninsurance in DC is a result of the somewhat higher rate of adult Medicaid coverage compared with other areas. Missouri has the lowest reported adult Medicaid coverage, 33 percent, but also has the highest rate of private (employer and other) coverage at 34 percent.

The rate of uninsurance within study area is always lower for children than adults, largely due to greater eligibility for Medicaid, but it still varies across area. For children, the range is from 8.0 percent in Missouri to 28.9 percent in Illinois. The low rate of uninsurance for children in Missouri is directly related to the high rate of Medicaid coverage, 68 percent, the highest of any of the areas. Employer-sponsored or other coverage for children is similar across study areas.

Other Sources of Public Support

There are a number of other sources of public support that can provide crucial assistance to families that have exited welfare. These include child care assistance, housing assistance through subsidies or public housing and Supplemental Security Income (SSI) program for persons with disabilities. A few areas asked about receipt of these benefits in their surveys, some asking about receipt at the time of the survey, some since exit and some for both time periods (table V.7).⁽²¹⁾  In addition, many working leavers will be eligible for the federal Earned Income Tax Credit (EITC) that supplements incomes of low-income workers.

Table V.7:
Percent of Leavers Receiving Other Publicly Funded Sources of Income:
Survey Data

State	Exit Cohort	Timing of Survey Post Exit	Timing1	Housing Assistance (%)	SSI (%)	EITC (%)
Arizona2	1Q98	12-18 months	TOS	18	n.a.	51
District of Columbia	4Q98	~12 months	TOS3	274	n.a.	n.a.
			SE	31	n.a.	n.a.
Illinois	Dec. 1998	6-8 months	SE	14	12	41
Missouri	4Q98	26-34 months	TOS	264	124	n.a.
			SE	n.a.	n.a.	n.a.
Washington2	Oct. 1998	6-8 months	TOS	n.a.	4 4	n.a.
			SE	19	n.a.	65
1 TOS= Time of Survey, SE= Since Exit.
2 Single-parent cases.
3 Month prior to survey.
4 Reports for month prior to survey.
Source: See Appendix B for a complete listing of the leavers studies referenced.

A significant minority of leavers are receiving housing assistance. At the time of the survey, the rates ranged from 18 percent of welfare leavers in Arizona to 27.4 percent of leavers in the District of Columbia (both about 12 months after exiting TANF). A smaller percentage of families received this assistance in Illinois, 13.6 percent in the 6 to 8 months after exiting.

In addition, between 4.0 and 12.0 percent of former recipients receive cash assistance from the SSI program at the time of the survey. Since this income is for persons with a disability that prevents them from working, some leavers who are not working may instead be relying on this income.

A final source of public support is the federal EITC. Working families with relatively low earnings are eligible to receive this credit from the federal government.⁽²²⁾ Three studies report how many leavers received this credit. Arizona reports 51 percent of leavers, Illinois reports 40.8 percent of leavers, and Washington reports 65 percent of leavers received the EITC. Arizona and Illinois also report that a higher percentage of leavers had heard of the EITC, 66 percent and 76 percent, respectively. Illinois probes further and finds that although three-quarters have heard of the EITC, only 47 percent say they know what it is, a percentage not much higher than those receiving the credit.

VI. Material Hardship and Well-Being

Income, earnings, and program participation are important parts of economic well-being, but they do not capture overall well-being. There are many other aspects of former recipients' lives that reflect how well these families are doing. Several studies have collected information in their surveys that reflect the extent to which former recipients are experiencing particular problems of material hardship. Because those receiving TANF benefits can also experience these problems, most studies compare results for leavers before and after exit.⁽²³⁾ The Washington state study has a different and perhaps stronger research design for comparing former and current recipients by making the comparison for leavers to a separate sample of families still on TANF. While the questions across these surveys are not identical so comparisons need to be made carefully, the addition of these measures significantly broadens our understanding of the well-being of welfare leavers.

One area of concern is the extent to which families who left welfare are having problems with the basic necessity of having enough food. The surveys use several measures to get at the extent and severity of families experiencing food problems (table VI.1). Several ask whether leavers experienced not having enough to eat or food not lasting until the end of the month. A fairly large number of leaver families report this experience in the time after exiting, ranging from 24.0 percent in Arizona to 45.5 percent in the District of Columbia. One action families might take in this situation is to cut the size of or skip meals. Illinois and District of Columbia both report that about a quarter of leaver families say they have taken one of these actions sometimes or often. Washington reports that 43 percent of leavers report having cut the size of meals and 27 percent have skipped meals. An indicator of even more severe problems is having a child skip meals. Missouri reports 3 percent of leaver families have taken this action and Washington reports 4 percent have done so. Finally, another indicator of need is families reaching out to food banks and shelters to help with emergency provision of food. Responses range from 7.2 percent of leaver families in Missouri receiving food assistance from a church or community group to 44 percent of leavers in Washington reporting receiving food from a food bank or shelter.

Table VI.1:
Leavers' Experience of Material Hardship:
Survey Data

	Percent of Leavers Reporting Hardships (%)
	Arizona1		District of Columbia		Illinois		Missouri	Washington1
	6 months pre-exit	6 months post-exit	While on TANF	Since exit2	6 months pre-exit	Since exit2	Any time in past month	Past 6 months on TANF	Past 6 months- Leavers
Food Problems
Not enough to eat/ Food didn't last	30	24	n.a.	46	51	44	26 3	n.a.	n.a.
Cut size of/ skipped meals sometimes or often	n.a.	n.a.	n.a.	25	24	25	n.a.	39 4	43 4
Child skipped meal	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	3	5 5	4 5
Went without food all day at least once	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	11	15
Received food from food bank/ shelter	29	21	n.a.	n.a.	15 6	12 6	7 7	35	44
Housing problems
Behind on rent/ housing costs	41	37	27	27	45	38	26 8	n.a.	n.a.
Forced to move because couldn't pay housing costs	21	17	8	6	15	13		3 9	7 9
Without place to live at least once	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	11	13
Stayed in homeless shelter	4	3	5	3	4	3	n.a.	2	1
Utilities cut off because of failure to pay	18	12	7 10	6 10	27	14	n.a.	12	12
Other
Children forced to live elsewhere	9	8	6	5	9	8	n.a.	n.a.	n.a.
Child in foster care	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	2	3
Unable to afford/ get medical attention	14	24	3	8	26	31	n.a.	n.a.	n.a.
1 Single-parent cases.
2 "Since exit" is between 6 and 8 months after exit for Illinois and approximately 12 months after exit for DC.
3 Actual question is "unable to buy enough food."
4 Percentages are for cut size of meals.  Washington asked "skipped meals" separately and found 22% for on TANF and 27% for leavers.
5 Includes households where both a child and an adult skipped meals.
6 Responses are for received food from a shelter.
7 Percentage represents those receiving food assistance from a church or community group.
8 Includes those unable to pay rent, mortgages, or utilities.
9 Percentage represents those who were evicted because of failure to pay rent.
10 Percentage represents those who went without electricity.  Separate numbers are available for water and heat in study.
Source: See Appendix B for a complete listing of the leavers studies referenced.

In the three studies that compare food problems before and after exit, Arizona and Illinois generally find lower absolute numbers experiencing food hardship after exit than prior to exit.⁽²⁴⁾ However, Washington generally finds higher rates of food problems among leavers than among those who are on TANF. This difference across studies could be related to the fact that Washington is comparing leavers to a separate group of those on TANF, while the other studies compare the same group of people's experiences before and after leaving. Washington's results may be more persuasive, given they are not subject to problems with recall. However, it is important to note that the relative well-being of families on cash assistance will be affected by differences in state benefit levels.

Another area of material hardship is problems with housing and utilities. Again, the surveys use a number of different questions to get at the extent to which leavers are experiencing housing-related problems. One measure is whether the family has been behind on rent or housing costs. In all the studies that ask this question, over a quarter report having experienced this problem since exiting. Another more severe situation is having to move because of inability to pay housing costs. A smaller but substantial percentage of families report this, ranging from 5.7 percent in DC to 17.0 percent in Arizona. Approximately 3.0 percent of families in Arizona, Illinois, and DC report staying in homeless shelters after leaving TANF (Washington reports a lower 1.3 percent). In addition to problems with rent and places to stay, a number of families have had utilities cut off because of failure to pay. This ranges from 5.8 percent of families in DC having had electricity cut off, to 13.9 percent of families having some utility cut off since leaving TANF in Illinois.

Three out of four of the studies (Arizona, Illinois, and DC) find higher absolute percentages of families experiencing these housing-related problems before TANF than after exiting. Again, Washington finds that leavers are slightly worse off, being more likely to have been evicted or without a place to live than those on TANF.

Two other measures of hardship that were asked in several studies have to do with whether children needed to go live elsewhere because of financial problems and whether families were unable to afford or get medical attention. The percentage of families where children were forced to live elsewhere after the family left TANF ranges from 5.4 percent in DC to 8.2 percent in Illinois. The percentages experiencing this situation before leaving TANF is slightly higher. Washington does not include a similar measure but does report that 3 percent of leavers have a child in foster care at least once after TANF, while only 2 percent of those on TANF were in this situation.

The percentage of families unable to get medical attention varies more than the other measures reported, from 8.3 percent in DC to 30.5 percent in Illinois. Unlike the other measures, this is only one that is consistently higher for families after exiting TANF compared with while on TANF. This seems consistent with the significant declines in Medicaid coverage reported earlier.

A final measure of well-being is a more general question posed to families in three of the studies. The questions all ask families to compare their overall well-being since exiting TANF to the prior time period. Approximately one-fifth or less of families report they are worse off or much worse off after leaving TANF than before, 21 percent in Washington, 15 percent in Arizona, and 12.6 percent in Illinois (table VI.2). Less than a third are reporting they are doing about the same since leaving TANF. This leaves more than half of families saying they are better off since leaving welfare, with more than two-thirds of families saying they are better off in Arizona. These results are consistent with the general findings on the material hardship measures that although a number of families experience these problems after exiting, for the most part they experience them less than before leaving TANF. Even in Washington, which found higher rates of most material hardship measures among leavers than among those on TANF, 79 percent of families say they are the same or better off compared to before they left welfare.

Table VI.2:
Overall Current Well-Being Relative to Before Leaving TANF:
Survey Data

State	Much Better Off	Better Off	Same	Worse Off	Much Worse Off
Arizona1	31	37	16	12	3
Illinois2	n.a.	57	30	13	n.a.
Washington1	32	28	19	13	8
1 Single-parent cases.
2 Respondents were asked only whether "better off," "same," or "worse off."
Source: See Appendix B for a complete listing of the leavers studies referenced.

VII. Child Care

With work as a major goal of many states' welfare programs, the need for child care is an important consideration. Child care subsidies are generally available to employed TANF leavers, depending on income. However, use of subsidies depends on the type of care arrangement being used, knowledge of availability and eligibility for subsidies, and ease or difficulty of obtaining and using them. Concerns about the quality of care being received by children of working TANF leavers are also important, although related measures are generally beyond the scope of the surveys conducted.

Three surveys asked about the type of child care arrangements employed leavers use.⁽²⁵⁾ Since type of care differs by age group, particularly pre-school age, and post-school age, some studies show their results by age of child (table VII.1). A substantial percentage of families do not have a child care arrangement, other than a parent or regular schooling. The percentage is higher for older children: Illinois reports 64 percent of working families with children over 12 have no arrangement. Missouri reports 60 percent of working families with school-age children 6 to 13 have no child care, while Washington and Illinois report lower rates of 10 percent and 18 percent (for all children less than 13), respectively. The differences in rates across types of care could in part be due to the different categorizations of care made in each study. Missouri has higher rates of no child care, but lower rates of relative or sibling care than the other states.

TableVII.1:
Child Care Arrangements of Employed Leavers: Survey Data

Type of Care	State and Age of Children (%)
	Illinois			Missouri1		Washington2
	<6	6-12	>12	<5	6-13	<13
No Child Care Arrangement2	7	10	64	25	60	18
Relatives/ Siblings	54	53	21	31	18	34
Friends/ Neighbors	8	11	7	10	4	n.a.
Center/ Afterschool care/ Church or club	11	7	3	19	15	19
Preschool/ Head Start	2	n.a.	n.a.	2	n.a.	n.a.
Family Day Care/ Babysitter In-home3	14	14	n.a.	14	3	11
Other5	4	5	6	n.a.	n.a.	20 4
1 Missouri reports numbers for those with arrangements.  Calculations were made to convert numbers to be out of all employed leavers.
2 Single-parent cases.
3 Includes those with no caregiver other than parents or regular school.
4 Family day care and babysitter in-home are combined because several surveys did not ask separately about each type.
5 Other includes multiple arrangements, preschool/ Head Start, Child self-care, employer-sponsored care, and unspecified care in Washington.
Source: See Appendix B for a complete listing of the leavers studies referenced.

The percentage of working families using relative or sibling care is very high. In Illinois, for children under 6 and 6 to 12, the rates of this type of care are over 50 percent. For younger children in Missouri and Washington, the rates are around a third of working leavers. While some families pay relatives for care, this type of arrangement probably includes less paid care than the remaining categories. Among the other types of care used, the most used for pre-school children are center care and family day care or babysitter in the home.

Paying for child care is a critical issue for families leaving TANF for employment. Costs of child care can affect choice of arrangement. Two studies reported the percentage of employed leavers with child care arrangements who reported paying for child care. In Illinois and Missouri the share paying for child care are 44 percent and 40 percent, respectively. Both report this percentage varies with the age of the child.

At least in these two states, the percentage of leavers who might take up child care subsidies is a smaller subset of all leavers, those who have a child care arrangement and are paying for care. Given that many families have no child care arrangement, and less than half of families who have an arrangement are paying for care, it is not surprising that relatively small percentages of leaver families report using child care subsidies. This percentage varies from a low of 5.0 percent of working leavers with children in the District of Columbia to a high of 20 percent of all leavers in Washington.⁽²⁶⁾ Whether low usage of subsidies is the result of not needing these benefits, or simply not using them results from lack of knowledge of eligibility or difficulty in using them is not known.

Table VII.2:
Employed Leavers Paying for Child Care:
Survey Data

State/ Study	Percent of Employed Leavers with Childcare Arrangements Paying for Care	Percent of Employed Leavers Using Subsidies	Average Monthly Costs1
Arizona2	n.a.	15 3	n.a.
District of Columbia	n.a.	5/3 4	n.a.
Illinois	44 4	17 5	$211
Children <6	32	n.a.	n.a.
Children 6-12	22	n.a.	n.a.
Children >12	3	n.a.	n.a.
Missouri	40	14/36 6	$277
Children <5	38	n.a.	$221
Children 6-13	46	n.a.	$171
Washington2	n.a.	20 7	n.a.
1 Families' reports of payments for childcare, among those making payments.
2 Single-parent cases.
3 Percentage is from administrative data for 1st quarter after exit, and includes all leavers, not only employed leavers.
4 Percent of employed leavers receiving assistance from welfare office is 5.0%; receiving assistance from other private sources is 2.5%.
5 Percentage includes those working as well as those in job search, education, or training.
6 14% using subsidies at time of survey, 26-34 months after exit and 36% have used subsidies at some point since exit.
7 Percent of all leavers using subsidies at time of survey, 6-8 months after exit.
Source: See Appendix B for a complete listing of the leavers studies referenced.

VIII. Conclusions

Examining results from the twelve study areas summarized in this report, we find that although there are outliers in most measured outcomes, there are a surprising number of common findings. Broadly speaking, among families leaving welfare, about three out of five work at any given time after exiting and about three-quarters have worked at some point within a year of leaving welfare. When leavers work, they usually work full-time and earn $7-$8 dollars an hour. On average, working leavers make about $3,000 a quarter. And total family income of leavers hovers around the poverty line. A significant minority of leavers have returned to TANF after their initial exit. Over one-third of leavers receive food stamps and about 2 in 5 have public health insurance coverage in the fourth quarter following exit. Finally, many leavers experience hardships, such as not having enough food to eat, but in general they do not experience these hardships more frequently than when they were on welfare.

Despite the fact that the results on many measures are broadly similar, the variation in results on many specific measures are important. The study areas differ in a number of important ways including welfare policies in place, the state or local economic situation, and the characteristics of the caseload. Each of these will likely affect outcomes for leavers. While this report focuses on summarizing results across studies, relating outcome differences to each of these contextual factors is important. State and federal government can use this information to help decide what changes might be made to improve leavers' well-being. That said, figuring out the relationship between leaver outcomes and policy, economic conditions, and personal characteristics, and the importance of each is extremely difficult. Even with rigorous statistical modeling, the problem is complex and results can be ambiguous.

Beyond this, leaver studies are clearly important for state and local areas to understand what is happening to those no longer receiving welfare benefits. The studies here also provide a valuable baseline of information upon which to build. We would like to point out some of the unique features of specific studies summarized in this report and some new suggestions that could be incorporated into future leaver studies.

The first is additional use of subgroup analysis. Several studies included interesting subgroup analyses, not summarized in this report, such as substate geographic groupings, sanctioned/nonsanctioned cases, and one or two-parent cases. These breakdowns can provide a great deal of additional information on how outcomes for different leavers vary. Another important subgroup analysis to consider is outcomes for employed and nonemployed leavers. This can help us to understand how those who are not working are faring. Creative use of subgroup analysis can add a great deal of information to results and let us know whether there are groups that are not doing as well as others.

The second way to broaden these studies is to add in additional sources of administrative data beyond TANF, Medicaid, food stamps, and UI records. Some of the ASPE-funded leavers studies do this. While this requires both more time and money, it can add information particularly for areas unable to field additional surveys. Information on child support, child care and housing subsidies, and child welfare are a few of the measures reported on from other administrative sources.

Finally, states can build on these studies by repeating them for new cohorts of leavers or by following existing cohorts over time. Studying new cohorts allows comparison of whether the status of leavers is changing as policies become more fully implemented and time limits are reached. Reinterviewing or analyzing administrative data for the same cohort of leavers as time passes provides information on whether employment is becoming more stable, earnings are rising, and economic hardship is decreasing-in short, whether the well-being of leavers is improving over time.

Appendix A: Methodology

The studies summarized in this report use a variety of methods in obtaining their results. As part of the ASPE grant program, every attempt was made to encourage the use of comparable measures to enable comparison of results. However, many differences in methodology remain. This appendix attempts to give an overview of the differences. We do not attempt to connect differences in results to differences in methods. The primary areas of differences include the type of data being used, the cohort of exiters, and the specific types of cases included or excluded. The studies that have included surveys up to this point, have additional differences in response rates, timing of survey, and survey instrument. This appendix describes the extent and nature of these differences.

All of the studies focus on at least one cohort of families leaving the TANF/AFDC program. For the most part, the definition of who is a "welfare leaver" is comparable across studies. Most of the studies included as leavers those who received welfare in a specific month but did not receive any benefits for the following two months. The District of Columbia is an exception here, defining leavers as those who stopped receiving in a given month and did not receive at all in the next calendar month. Arizona reports results for both 1 month and 2 month leavers.

The broader difference in studies was the timing of the cohort(s) studied (table A1). Many include an early and late cohort, although some include multiple cohorts. Illinois includes six cohorts of information and a measure that combines information across all these cohorts. Six of the studies include a cohort from one of the quarters of 1996. Four of the studies include information from a cohort in 1998. Given the many changes occurring in welfare programs, caseloads, and the economy across these time periods, we expect that the particular timing of the cohort examined may make a difference in results. In our summary, we focus on the most recent complete cohort information available.

Appendix Table A1:
Methodology

State	Type of Study	Leaver Cohorts	Definition of leaver	Cases excluded	Reported Subsets
Arizona 1998	Survey and Administrative	1Q98	One month, two months	2 Parent Units, Child only cases, cases under tribal jurisdiction	Sanctioned, unsanctioned leavers
District of Columbia	Administrative and Survey	Two Cohorts: 4Q97, 4Q98	One month	Child Only	Continuous leavers
Florida1	Administrative	2Q97		Single Parent	Continuous leavers
Georgia1	Administrative	Four Cohorts: 1Q97, 2Q97, 3Q97, 4Q97	Two months	2 Parent Units, Child-only cases
Illinois	Survey and Administrative	Six cohorts: 3Q97, 4Q97, 1Q98, 2Q98, 3Q98, 4Q98; Survey: Dec. 1998	Two months	Child-only cases	Single Parents, Two Parents; Cook County, Downstate
Missouri	Administrative	4Q96	Two months	Child only	2
Missouri	Survey	4Q96	Two months	Child only	Kansas City, St. Louis
New York1	Administrative	1Q97	Two months	Child only	Single Parents, Two Parents
Washington	Administrative	Three cohorts: 4Q96, 4Q97, 4Q98	Two months		Leavers, on-TANF; continuous leavers
Washington	Survey	October 1998	Two months	Child Only	Single Parents, Two Parents; Leavers, on- TANF
Wisconsin	Administrative	July 1995-June 1996	Two months		Single Parents, Leavers, on-TANF; continuous leavers
Cuyahoga Co.1	Administrative	3rd Quarter 1996	Two months	Child-only cases, single women without children
Los Angeles Co.1	Administrative	3Q96	Two months
San Mateo Co.1	Administrative	Two Cohorts: 4Q96, 4Q97	Two months		Single Parents, Two Parents; Child Only; One or Two Parent Transitioning to Child Only
1 Jurisdictions with administrative data only will add survey data to their final reports (forthcoming).
2 Missouri will include a geographic subset in Section D of a forthcoming report.
Sources: See Appendix B for a complete listing of the leavers studies referenced.

Another methodological issue that may be important for interpreting results is the type of cases included and excluded in the study sample. Most of the studies exclude exiting child only cases completely or present results separately for non-child-only cases. In addition, a number of the studies exclude exiting two-parent cases or present results for single-parent cases separately. This allows comparisons on the subset of single-parent cases for most studies. The exceptions are District of Columbia and Missouri. However, the percentage of two-parent cases among the leaver population is less than 2 percent in the District of Columbia and less than 4 percent in Missouri.

Several of the studies present results on specific subgroups of leavers. Arizona presents results for sanctioned and non-sanctioned leavers separately. For their early cohort, Arizona presents separates urban and rural results. Illinois also presents a geographic break, reporting Cook County (Chicago) separately from downstate Illinois. Washington compares results for leavers with outcomes for current TANF recipients. Several studies report administrative program participation data for continuous leavers, that is, those who exit TANF and do not return in a specified time period (usually a year).

Administrative Data

All of the studies use some source of administrative data, at least information on welfare benefit recipiency (table A2). Some studies had access to rich databases of linked administrative data that included a wide variety of program participation much beyond welfare receipt. Given the scope of resources available, unless areas had on-going efforts to create and maintain multi-link/multi-program databases, reporting was more limited. The resources in time and money to link data accurately can be very high. In addition to program participation, most studies (with the exception of District of Columbia) used unemployment insurance administrative data to get information on employment and earnings.

Appendix Table A2:
Administrative Data

Study	Topics Reported in Study	Period of Follow-up After Exit
Arizona	Employment, TANF, Food Stamps, Medicaid, Childcare Subsidy, Child Support, Child Welfare1	One year
District of Columbia	TANF, Food Stamps, Medicaid	18 months
Florida	Employment, TANF, Food Stamps, Medicaid	7 quarters employment, 12 months program participation
Georgia	Employment, TANF2	One year
Illinois	Employment, TANF, Food Stamps, Medicaid, WIC, Childcare Subsidy, Family Case Management Services, Drug and Alcohol Treatment Services, EITC, Child Support, Child Welfare3	One year
Missouri	Employment, TANF, Food Stamps, Medicaid	Two years
New York	Employment, TANF, Food Stamps, Medicaid	One year
Washington	Employment, TANF, Food Stamps, Medicaid, Childcare Subsidy, Child Support Programs, Child Welfare3	Two years
Wisconsin	Employment, TANF, Food Stamps, Medicaid	15 months
Cuyahoga Co.	Employment, TANF, Food Stamps, Medicaid	One year
Los Angeles Co.	Employment2	One year
San Mateo Co.	Employment, TANF, Food Stamps, Medicaid	One year
1 Substantiated child protective service reports, out-of-home placements, and use of emergency services.
2 These are the topics covered to date in the current, preliminary reports.  Forthcoming reports will cover other topics.
3 Child abuse and neglect referrals and out-of-home placements.
Sources: See Appendix B for a complete listing of the leavers studies referenced.

Other than TANF, food stamps, Medicaid, and employment data, several studies reported additional information from administrative data sources. These include use of child care subsidies, receipt of child support payments, child welfare reports of abuse and neglect, and use of emergency assistance. The specific studies that have these additional data and report on it in their study are listed in the table.

The length of follow-up for any of these sources of administrative data varied greatly. Some studies reported as much as two years of follow-up data for certain cohorts. The majority reported one year of data. Illinois had a combined rolling cohort, including as much information for each of the six cohorts that they had up to the date of analysis.

Survey Data

Five of the twelve study areas had completed reports with results from survey data as of mid-October, 2000. There are many specific methodological differences across each survey: how the sample was selected, how potential respondents were located initially and follow-up location efforts, who conducted the survey, length of the survey and payment for completed surveys. All of these can affect a surveys response rates (the percent of all potential respondents who completed the survey) and potentially its results. We do not report on all these differences here, but focus on final sample size and response rate. Additional differences in the surveys include the exit cohort selected for the survey and the time after exit that the survey was conducted (table A3). One common factor to all the surveys is they included a mixed-mode of collection, that is, they used telephone surveys as well as in-person surveys.

Appendix Table A3:
Survey Information

Study	Cohort Date	Mode	Sample Size	Response Rate	Timing of Survey After Exit
Arizona	1Q98	Phone, in person	821	72%	12-18 months
District of Columbia	4Q98	Phone, in person	277	61%	12 months
Illinois	Dec 1998	Phone, in person	514	51%	6-8 months
Missouri	4Q96	Phone, in person	878	75%	26-34 months
Washington	Oct 1998	Phone, in person	987	72%	6-8 months
Sources: See Appendix B for a complete listing of the leavers studies referenced.

The final sample size of the survey information varied from 987 in Washington to 277 in the District of Columbia. The response rates varied somewhat as well. Three of the five had response rates over 70 percent. The District of Columbia had a response rate of 61 percent while Illinois's was 51 percent.

Four out of five of the surveys focus on exit cohorts in 1998, late 1998 for three (Illinois, Washington, and District of Columbia). Missouri surveys the earliest exit cohort, fourth quarter of 1996, but is therefore able to survey respondents at a longer follow-up period, 30 months after initially exiting TANF. Their response rate does not seem to have suffered as a result of the long time between leaving welfare and being interviewed, relative to the other surveys. The remaining surveys interview leavers at from six to eight months after leaving (Illinois and Washington) or about 12 months after leaving (Arizona and District of Columbia). Results in this summary report try to point out the differences in timing of cohorts and time between survey and exiting.

Appendix B: ASPE Funded Welfare Outcomes Reports Included in this Review

Arizona

Westra, K. & Routley J. (January 2000).  Arizona Cash Assistance Exit Study: First Quarter 1998 Final Report.  Arizona Department of Economic Security.

District of Columbia

Acs, G. & Loprest, P. (October 2000).  The Status of TANF Leavers in the District of Columbia: Final Report.  The Urban Institute.

Florida

Crew, R., Eyerman. J., Graham, J., & McMillan N.  Appendix 3, Commonly Reported Administrative Data, from forthcoming report on Tracking the Outcomes of Welfare Reform in Florida for Three Groups of People.

Georgia

Foster, E.M. (April 1999).  "Amended Quarterly Progress Report: Outcomes for Single Parent Leavers by Cohort Quarter."  Georgia State University.  Data are preliminary.

Illinois

Julnes, G., Halter, A., Anderson, S., Frost-Kumpf, L., Schuldt, R., Staskon, F., and Ferrara, B. (forthcoming).  Illinois Study of Former TANF Clients, Final Report.  Institute for Public Affairs, University of Illinois at Springfield and School of Social Work, University of Illinois at Urbana-Champaign.

Missouri

Midwest Research Institute (June 2000), Chapter 1- Employment and Earnings of Former AFDC Recipients in Missouri, Chapter 2- Household Income and Poverty, Chapter 3- The Continuing Use of Assistance by Former Missouri AFDC Recipients, and Chapter 4- The Continuing Use of Assistance by Former Missouri AFDC Recipients.  Interim Reports for the Missouri Department of Social Services.

Ryan, S., Theilbar, M., Choi, S., Qu J., Deng M., and Ellebracht, L.  (April 1999).  Preliminary Outcomes for 1996 Fourth Quarter AFDC Leavers: First Interim Report.  University of Missouri-Columbia.

New York

Rockefeller Institute, New York State Office of Temporary and Disability Assistance, and the New York State Department of Labor (December 1999).  After Welfare: A Study of Work and Benefit Use After Case Closing.  Revised Interim Report

Washington

Ahn J., Fogarty, D., Kraley, S., Lai, F., and Deppman, L.  (February 2000).  A Study of Washington State TANF Departures and Welfare Reform.  Welfare Reform and Findings from Administrative Data.  Final Report.  Washington Department of Social and Health Services.

Du, J., with Fogarty, D., Hopps, D., and Hu, J.  ( February 2000).  A Study of Washington State TANF Leavers and TANF Recipients.  Findings from the April- June 1999 Telephone Survey.  Final Report.  Washington Department of Social and Health Services.

Wisconsin

Cancian, M., Haveman, R., Kaplan, T., and Wolfe, B.  (January 1999).  Post- Exit Earnings and Benefit Receipt Among Those Who Left AFDC in Wisconsin.  Institute for Research on Poverty, University of Wisconsin- Madison.

Cuyahoga County

Coulton, C. & Verma, N.  (May 1999).  Employment and Return to Public Assistance Among Single, Female Headed Families Leaving AFDC in Third Quarter, 1996, Cuyahoga County, Ohio.  Prepared for Cuyahoga Work and Training.

Los Angeles County

Verma, N. & Goldman, B.  (January 2000).  "Los Angeles County Post-TANF Tracking Project: Quarterly Progress Report."  Manpower Demonstration Research Corporation.  Data are Preliminary

San Mateo County

Moses, A. & Macuso, D.C.  (May 1999).  Examining Circumstances of Individuals and Families Who Leave TANF: Assessing the Validity of Administrative Data.  SPHERE Institute.

Many of these reports can be accessed at: http://aspe.hhs.gov/hsp/leavers99/reports.htm

Endnotes

1.  The Wisconsin study uses a slightly different definition of leavers than the later ASPE-funded studies. The results reported here are based on supplementary work done by the authors of the Wisconsin study which makes the findings more comparable with those reported in subsequent ASPE-funded work.

2.  Many of these studies are reviewed in Loprest and Brauner (1999), GAO (1999), Acs and Loprest (2000), Isaacs and Lyon (2000), and DHHS/ASPE (2000).

3.  As of October, 2000, findings are not yet available from Massachusetts, but are expected soon. A fourteenth leaver study in South Carolina funded under a different mechanism, will have findings in 2001.

4.  Most studies require a family to remain off welfare for two months to be considered a leaver. Arizona uses a one month requirement throughout its study but presents supplementary findings using a two month definition. In this synthesis report, we focus on the two month findings from Arizona for comparability. DC requires a leaver to remain off welfare one full calendar month.

5.  Most studies examine families that left welfare during a given three month period. Wisconsin's leaver cohort is made up of leavers from a twelve-month period.

6.  Massachusetts' leaver study is not reviewed here because it was not complete by October 2000; however, a portion of welfare recipients in the state have reached their interim time limits and Massachusetts report concentrates heavily on time limited v. other TANF leavers. The Massachusetts report will be included in our final report in Summer 2001.

7.  The cohort of leavers from Cuyahoga county, Ohio, which we examine in this report (3Q96) pre-dates the imposition of the 36 month time limit.

8.  Diversion policies may have particularly strong effects on return rates.

9.  For Wisconsin, 73.5 percent of case heads whose educational attainment is reported have less than 12 years of schooling.

10.  Throughout the report, we include administrative results for Single-Parent families and note the few exceptions. Because three out of five surveys primarily report results for all cases together, our survey findings are for all cases except where otherwise noted.

11.  Note that we report nominal monetary values. While inflation was very low during the late 1990s, a two-year difference between studies can represent about a five percent difference in purchasing power.

12.  Illinois and Missouri report monthly household income; Washington reports family income, and Arizona presents case income.

13.  Some studies report quarterly information (receipt at some point within a three-month quarter) and some present monthly information (receipt in particular month). Studies reporting quarterly information will report a higher percentage than if they reported a monthly number. For example, the New York study shows that 17 percent of original exiters were receiving TANF in the twelfth month after leaving, but 19 percent received at some point in the fourth quarter after leaving. However, across studies we do not observe any pattern of difference between monthly and quarterly reports. Table V.1 indicates whether numbers reported are monthly or quarterly.

14.  Recall that most of these studies require that a family remain off welfare for at least two months to be considered a leaver. This common leaver definition may contribute the similar, low return rates across sites.

15.  Cuyahoga county has a 36 month time limit but it was not implemented until 10/97, after the study period. California's time limit was implemented in 1/98.

16.  Despite reporting 100 percent receipt of food stamps in the month of exit, San Mateo County has very low percentages of food stamp receipt across the entire post-exit period. This is due to the relatively low rates of returning to TANF (most of whom would be eligible for Food stamps) as well as very low rates of food stamp receipt among families not receiving cash assistance. Note that high maximum benefit levels and high earnings disregards policies in California may result in higher incomes (and less eligibility for food stamps) among families who leave welfare in California as compared with other states.

17.  Wisconsin also presents results for continuous leavers; however, its findings for all leavers and continuous leavers cannot be compared in the same way as we do for other sites because Wisconsin uses a slightly different leaver definition for the two groups.

18.  A number of states also have programs that extend public coverage to children at higher income levels. While these sometimes go by names other than Medicaid, we are including them under Medicaid here.

19.  Washington is reporting results for all individuals, so it may also have lower Medicaid coverage for adults alone.

20.  Just as in the case of food stamps, Wisconsin presents Medicaid findings for continuous leavers but they cannot be compared to all leavers because of differences in the definition of leaver across the two groups.

21.  Leavers may also receive assistance through the school lunch program, the Women, Infants, and Children (WIC) nutritional program, unemployment insurance, Social Security, and foster care programs. These sources of support are not consistently reported in the survey-based studies reviewed here and are excluded from this initial synthesis report. Child care assistance is discussed later in this report.

22.  As of 1998, three of our study areas had state EITCs as well, District of Columbia, Illinois, and New York.

23.  Arizona, Illinois, and Washington use a 6-month window for their questions. Missouri asks about the past month, and the District of Columbia asks about the time while on TANF and the time since exiting TANF, about 12 months.

24.  We do not discuss here whether these differences are statistically significant, which is reported in some studies and not others.

25.  District of Columbia also asks about child care arrangements used by employed leavers, but because the results are not broken out by age in the report, making comparable categorizations difficult, the results are excluded here. Arizona reports child care arrangements of all leavers, but does not separately report arrangements for employed leavers, so these results are also excluded.

26.  In District of Columbia an additional 2.5 percent of leavers report getting assistance from other private sources which could include government funds given to community or church groups. Missouri reports that 36 percent of leavers used child care subsidies at some point over the 30 months since exit.