Status Report on Research on the Outcomes of Welfare Reform

Appendix B:
Findings from ASPE-Funded Leavers Studies

(Grants to States and Localities to Study Welfare Outcomes)

[ Main Page of Report | Contents of Report ]

Contents

• Employment and Earnings
• Program Participation
• Household Income and Family Well-Being
• Conclusion
• Other Outcomes Data

• Endnotes

Welfare caseloads have declined dramatically during the past several years. Overall, the welfare caseload has fallen by 8.4 million recipients, from 14.2 million recipients in 1994 to 5.8 million in June 2000, a drop of 59 percent. This is the largest welfare caseload decline in history. As the caseloads have fallen there has been widespread interest in the circumstances of recipients who have left welfare. How are they faring without cash assistance? Are they working? Are they moving out of poverty? To what extent do they return to welfare? To what extent do they continue to need and to receive assistance and supportive services through other programs?

To answer these questions, ASPE awarded approximately $2.9 million in grants to states and counties in FY 1998 to study the outcomes of welfare reform on individuals and families who leave the TANF program, who apply for cash welfare but are never enrolled because of non-financial eligibility requirements or diversion programs, and/or who appear to be eligible but are not enrolled. The 1998 grants were awarded to ten states and three large counties or consortia of counties (Arizona, the District of Columbia, Florida, Georgia, Illinois, Massachusetts, Missouri, New York, Washington, and Wisconsin; and Cuyahoga County, Ohio, Los Angeles County, California, and San Mateo, Santa Cruz, and Santa Clara Counties, California). Separate but comparable studies were also funded in Iowa (with FY 1999 funding) and South Carolina (in FY 1998 and 2000, as part of a longer-term project) resulting in a total of 15 studies with findings on former recipients as of spring 2001.⁽¹⁾

Following the devolution of welfare programs to the state level, ASPE chose a research strategy that combined local flexibility in study design with some national direction and coordination. Most of the projects used administrative data to track an early cohort of individuals who left welfare around 1996 or 1997. Projects also used a combination of administrative and survey data to track the economic status and general well-being of at least one cohort who left welfare one to two years later, after the transition from AFDC to the TANF program. Projects varied, however, in the number and types of administrative data sets examined and the design of the surveys of former recipients. Final survey sample sizes varied from 277 to over 3,500 cases, response rates ranged from 23 to 81 percent, and approximate time of interview varied from 6 to 30 months after exit, as shown in Table 1. All researchers were encouraged to collect data across multiple dimensions, including employment, program participation, economic status, family structure, child well-being, material hardship, barriers to employment, etc. Grantees designed their own survey instruments, however, which differed in wording and emphasis. While this diversity poses challenges for summarizing results nationally, it has allowed states to meet the demands of their elected officials and program administrators for timely information on families leaving their state's welfare program.

Although each study had its own methodology, ASPE took certain steps to promote comparability across the studies. Chief among these was developing consensus around a common definition of the "leaver" study population as "all cases that leave cash assistance for at least two months." This definition excludes cases that re-open within one or two months; such cases are more likely closed due to administrative "churning" than to true exits from welfare. In addition, through national meetings and an electronic list-serve, ASPE staff facilitated peer networking among researchers, promoted the use of nationally developed questions on topics such as food security and child well-being, and encouraged standardized reporting of certain administrative data outcomes.

As of March 2001, all 15 studies identified above had released preliminary reports based on administrative data findings, and 12 of the 15 also had released reports with more detailed findings from follow-up surveys. Highlights from these reports are presented below, with a focus on outcomes in employment and earnings, recidivism and program participation, and household income and family well-being.⁽²⁾ This summary stresses common findings for "average" welfare leavers in each jurisdiction, without analyzing how outcomes vary for different types of leavers (e.g., urban vs. rural, those who left due to earnings vs. sanctions). Findings are presented for all single-parent leavers in a state or county except where noted otherwise. Observed cross-state differences in outcomes reflect the diverse range of state policies and underlying economic and demographic conditions of the jurisdictions under study, as well as methodological differences in study design.⁽³⁾ A more comprehensive synthesis report, including analysis of how outcomes differ for various subgroups, is expected by Fall 2001.

[ Go to Contents ]

Employment and Earnings

Employment

Employment outcomes have been quite consistent across the 15 studies. Employment rates of former recipients ranged from 47 to 68 percent in the first quarter after exit according to administrative data (see Table 2). Moreover, employment rates remained fairly constant in the first year after exit in most study areas. This finding does not mean that the same 50 to 60 percent of leavers were employed every quarter. Some former recipients lost their jobs, while others found new employment, with the result that 62 to 90 percent of leavers had earnings at least once within the first four quarters after exit. Between 31 and 47 percent of leavers were employed in all four quarters (data not shown),⁽⁴⁾ according to the eight studies reporting this statistic.

Three of six jurisdictions analyzing employment across multiple cohorts found that recipients leaving welfare in 1998 had higher employment rates — by 5 to 10 percentage points — than those leaving in 1996 (data not shown). Two other two jurisdictions, however, found no change and one found a decrease in employment.

Administrative data do not capture all employment: quarterly earnings reported to the states' unemployment insurance (UI) programs do not capture earnings from self-employment, employment in the military or federal government, certain agricultural employment, and jobs across state boundaries. In fact, between 57 and 71 percent of former recipients reported working at time of interview. These self-reported employment rates from survey data were higher than the rates based on administrative data in all but one study (see Table 2). The vast majority of leavers - 85 to 92 percent - reported being employed at least once since exit. In addition, three studies found that the household employment rate (counting earnings of anyone in the household) was 9 to 15 percentage points higher than the individual rate for the leaver herself, or about 72 to 80 percent (data not shown).

Earnings

Median quarterly earnings of former recipients with jobs ranged from $1,900 in South Carolina to $3,400 in Washington, D.C. in the first quarter post-exit (see Table 3). In all reporting locations, quarterly earnings rose over the course of the year following exit.⁽⁵⁾ Median hourly wages, as reported in survey data from eight studies, ranged from $6.50 to $9.00 an hour. Former recipients with jobs worked an average of 33 to 39 hours per week; median hours averaged 40 hours per week.

In sum, the studies were consistent in finding that about three-fifths of leavers were working, generally 40 hours per week, but with relatively low wages and intermittent spells of unemployment. To what extent do families with these patterns of employment and earnings support themselves, and to what extent do they rely on government programs for support?

[ Go to Contents ]

Program Participation

Returns to TANF

According to data from 15 studies, between 3 and 21 percent of families leaving welfare returned to cash assistance within one quarter (see Table 4). Rates of welfare receipt rose to between 9 and 24 percent in the next quarter. Rates rose very slightly over the next six months, reaching 11 to 25 percent one year after exit. Because some people return to the rolls and then leave again, the proportion that ever returned within the first year after exit was higher, ranging from 17 to 38 percent.⁽⁶⁾

Survey data on returns to TANF are fairly similar to the administrative data. In addition, survey data also found that at least one half of those who returned to TANF did so for a job-related reason, such as job loss or decreases in work hours or wages. Other common reasons for returning to TANF included divorce or separation from partner, pregnancy or birth of a new child, re-compliance with program regulations, loss of other income, problems with child care, and problems with health or medical benefits.

Comparisons of early and later cohorts reveal no clear pattern of returns to welfare (data not shown). As compared with earlier cohorts, recidivism among 1998 leavers was higher in three states but lower in three others. No trend was apparent in two others.

Families leaving in the 1996 to 1999 period did so before they hit the five-year federal time limits on benefit receipt. Thus, most families had the option of returning to cash assistance as needed. Two studies, however, examined cohorts of 1999 leavers who were affected by state time limits of two years. Recidivism rates in these two states — Massachusetts and South Carolina — were lower than rates in other states, as shown in Table 4. Sub-group analysis in these two states indicates that families who left because of time limits were much less likely to be back on welfare at time of interview than other families; only 2 percent of the time-limited families in South Carolina and 8 percent in Massachusetts were back on welfare a year after exit.

Medicaid and Health Insurance

Although the majority of leavers remained off cash assistance, most continued to receive other government support. One of the most common supports was Medicaid, although rates of participation varied considerably across states. As shown in Table 5, between 42 and 80 percent of adult leavers were enrolled in Medicaid in the first quarter post-exit according to administrative data. In many areas, adult enrollment rates dropped 10 percentage points or more by the fourth quarter after exit. Medicaid coverage varied even more dramatically in survey data, ranging from 33 percent in Missouri (measured over 2 years after exit) to 81 percent in Massachusetts (measured slightly under a year after exit). A higher percentage of surveyed leavers — 51 to 83 percent — reported Medicaid coverage for their children, as shown in Table 6.

Lack of Medicaid enrollment is not necessarily a problem if leavers have health insurance through employment or other means. However, only 20 to 34 percent of adult leavers reported being covered by employer-sponsored or other insurance; somewhat fewer (7 to 28 percent) reported such coverage for their children. These figures reveal that, in most states, substantial numbers of former recipients and their children were without any health insurance. The percentage of adult leavers without insurance ranged from 7 to 45 percent; rates for children ranged from 8 to 33 percent, for Massachusetts and Florida, respectively. Data in Tables 5 and 6 indicate that lack of health insurance was more prevalent in states with low numbers of leavers enrolled in Medicaid. Survey data from six states (discussed in the section on Material Hardship and displayed in Table 10 below) show the consequences of lack of health insurance coverage.

Some of the state variation in Medicaid enrollment can be explained by differences in survey methodology (e.g., timing and wording of surveys) or in the linking and analysis of administrative data.⁽⁷⁾ Still, the observed cross-state variation is too wide to be solely attributable to measurement differences. Some variation in enrollment is likely to reflect differences in Medicaid eligibility (which is set by states) and in administrative practices, which vary across states and local areas.

Findings from the leavers studies and other research have prompted Federal and state initiatives to ensure that families leaving welfare are not incorrectly denied Medicaid benefits. In their leavers reports, several states mentioned changes in policies or procedures designed to increase Medicaid enrollment among future leaver cohorts. Early trends, between 1996 and 1998, show increased Medicaid enrollment in three jurisdictions, no change in one, and decreased enrollment in another.

Food Stamps and Other Program Participation

Participation in other forms of government assistance was also common, though generally at lower levels than for Medicaid. Participation rates of former recipients in the Food Stamp program, for example, ranged from 23 to 78 percent across 12 studies, with most finding that roughly one-third to one-half of AFDC/TANF leavers received food stamps immediately after exit (see Table 7). Similar rates were found in both administrative and survey data. Food stamp receipt declined in some states over time, but remained constant in others.

Other commonly received forms of government assistance included free- and reduced-price school lunches (43 to 87 percent of leavers), the federal Earned Income Tax Credit (32 to 65 percent of former recipients), housing assistance (16 to 60 percent of leavers), and Supplemental Security Income (2 to 12 percent of leavers), according to survey data from several surveys (data not shown). In addition between 11 and 35 percent of former recipients across seven studies reported receiving child support, often secured with help from the child support enforcement agency. As seen below, income from these sources can be an important component of household income.

[ Go to Contents ]

Household Income and Family Well-Being

Household Income and Poverty Status

Total household income is difficult to measure, particularly in leaver households. Paychecks can vary from month to month, and variations in unearned income and in household composition may generate added instability. Nevertheless, ASPE encouraged researchers to collect survey data on this critical measure of family well-being.

As shown in Table 8, average household cash income of former recipients ranged from $964 to $1440 per month across eight studies. When reported separately, median household incomes were about $200 lower. While not included in these cash income totals, food stamp benefits provided the average household with an additional $96 to $129 per month, according to three studies. (The value of the federal Earned Income Tax Credit also was not included in the cash income totals).

Six of the eight studies shown in Table 8 asked a series of detailed questions probing for income from various specific sources, while two (Illinois and the District of Columbia) simply asked for total household income. Consistent with past research, the surveys that asked multiple income-related questions uncovered higher levels of income than the other two surveys. This pattern of variation suggests that the lower incomes found in Illinois and the District of Columbia may reflect differences in income reporting rather than true differences in income.⁽⁸⁾

Five of the studies provided information about the sources of household income. Earnings were the largest income source: the leaver's own earnings made up 45 to 63 percent of total household income, while earnings of others in the household accounted for an additional 19 to 40 percent. Cash assistance from AFDC or TANF added another 3 to 8 percent. The final 9 to 20 percent of household income came from child support payments, Supplemental Security Income (SSI), and "other" income, including Social Security and survivors' benefits, veterans' benefits, workers' compensation, and financial assistance from others.

Four of the studies with comprehensive income questions also calculated the percentage of former recipients with household income below the federal poverty line. Estimated poverty rates ranged from 41 percent to 58 percent, depending in part on whether food stamps were included in measures of household income.⁽⁹⁾ Many leavers with household incomes at or above the poverty threshold were still close to poverty; the Iowa study found that 63 percent of leavers had income below 130 percent of the poverty threshold, in Cuyahoga 79 percent were below 150 percent of poverty and in Missouri 89 percent had cash incomes below 185 percent of the poverty threshold.

Though these poverty rates are quite high, one study (Washington) reported an even higher poverty rate — 83 percent — among a sample of recipients remaining on welfare for six months. Mean and median household incomes of ongoing recipients also were lower (data are not shown, but were $890 and $642, respectively) than those of former recipients. While the Washington study does not track the same group of people over time, it provides some evidence that economic status improves after exit from welfare.

Family Well-Being and Material Hardship

Partly because of the challenges of measuring income, most leavers surveys also asked directly about family well-being and material hardships resulting from not having enough money. Although surveys varied in wording, they generally asked about hardships related to food shortages, housing problems, and medical hardship.

Between one-eighth and one-half of leavers (13 to 52 percent) of leavers reported some level of food hardship, according to the 12 studies with survey data (see Table 9). Rates of food insecurity ranged from 32 to 46 percent among the three states that measured food insecurity through a standardized six-item questionnaire. About half of these families, or 16 to 26 percent of all leaver families, were classified as food insecure with hunger.⁽¹⁰⁾ Other states, while not using the six-item standardized scale, found similar evidence of food insecurity: 20 to 43 percent of respondents said that adults in the household cut the size of meals or skipped meals and 13 to 52 percent exhibited other signs of food insecurity, such as not being able to buy enough food. Very few leavers reported that children in their households skipped meals (3 to 5 percent, according to two studies).

Former recipients also reported problems with housing arrangements, although somewhat less frequently than food shortages. As shown in Table 10, the most common problems were falling behind in rent or mortgage (18 to 38 percent across seven studies), loss of utilities (12 to 36 percent of leavers across seven studies) and being forced to move (13 to 32 percent across four studies). Less often, former recipients were evicted (4 to 7 percent), went to a homeless shelter (1 to 7 percent, except one study reported 17 percent), or reported that their children were forced to live elsewhere (3 to 8 percent, except one study reported 19 percent).⁽¹¹⁾

Studies were split as to whether housing and food shortages were greater before or after exit; some found more hardship after exit, some found less hardship after exit, and some showed little difference. None of the leavers studies reported a significant change in use of homeless shelters before and after exit, or in experiences with separations of children from the family.

Six studies also examined medical hardship, and found that between 8 percent and 31 percent of leavers in the six sites reported that they or someone in their household was unable to get needed medical attention since leaving welfare because they could not afford it. Studies consistently reported more difficulty getting needed medical care after exit than while on welfare.

Finally, when directly asked about overall economic well-being or standard of living, 46 to 68 percent of families in five states reported they were better off financially after exit; 16 to 32 percent said they were the same, and 13 to 30 percent said they were worse off (see Table 11).

[ Go to Contents ]

Conclusion

In sum, findings across the 15 studies showed that about three-fifths of leavers were working, generally 40 hours per week. Former recipients experienced intermittent spells of unemployment and financial hardship, however, and about one-fourth to one-third returned to welfare at least once in the first year after exit in most states studied. Although quarterly earnings rose over time, total household incomes remained fairly low, averaging about $1,400 or less per month. Access to health insurance and food stamps appeared problematic for some recipients, and there also were reports of food shortages and inability to get needed medical attention. Evidence was mixed as to whether material hardships were greater before or after exit; families generally reported that they are better off overall after leaving welfare.

[ Go to Contents ]

Other Outcomes Data

The descriptive statistics highlighted above provide some important insights into the outcomes and well-being of individuals and families leaving welfare. However, they do not represent the sum total of the rich administrative and survey data collected by states and counties under the ASPE-funded grants. Links to most of the individual state and county reports can be found at <http://aspe.hhs.gov/hsp/leavers99/reports.htm>. The initial synthesis report by the Urban Institute is posted on the same web site, at <http://aspe.hhs.gov/hsp/leavers99/synthesis01/index.htm>. In addition, ASPE is working collaboratively with the grantees and a technical assistance contractor to make the grantees' welfare outcomes data files available to researchers for secondary analyses. Information on how to secure access to these data files can be found on the ASPE-sponsored web page on Welfare Leavers and Diversion Studies at <http://aspe.hhs.gov/hsp/leavers99/index.htm>.

[ Go to Contents ]

Endnotes

1.  In addition to funding the Iowa leavers study in FY 1999, ASPE funded leavers studies in Texas and in Contra Costa and Alameda Counties, California, as well as several applicant/diversion studies. Also in FY 1999, ASPE awarded an additional $837,000 for continuations and extensions of several of the FY 1998 leavers projects. In addition, $1.236 million was awarded in FY 2000 to enhance some existing studies of welfare-related outcomes. In all, ASPE has committed over $5 million to state and county grantees to study welfare outcomes.

2.  Findings presented here are based on an Initial Synthesis Report of the Findings From ASPE's "Leavers" Grants (prepared by the Urban Institute and available at <http://aspe.hhs.gov/hsp/leavers99/synthesis01/index.htm>; "A Cross-State Examination of Families Leaving Welfare:  Findings from the ASPE-Funded Leavers Studies," prepared by ASPE staff and available at <http://aspe.hhs.gov/hsp/leavers99/cross-state00/index.htm>; and ASPE staff analyses of reports submitted between November 2000 and March 2001.

3.  Cross-state comparisons are affected by a variety of factors, ranging from state sanction policies, maximum benefit levels and earnings disregard policies, to survey sample sizes, time of interview and response rates. They are also affected by the underlying economic, social and demographic conditions of the study sites. Some observed differences also reflect methodological issues, including questionnaire design or population under study. Brief summaries of the projects can be found at <http://aspe.hhs.gov/hsp/leavers99/fy98.htm>. Information on comparing survey instruments can be found at <http://aspe.hhs.gov/hsp/leavers99/cross.htm#comparing>.

4.  These individuals may not have been employed in every month, however, since UI records are based on quarterly earnings, reflecting any covered employment during that quarter.

5.  Data from the UI system are limited to aggregate quarterly earnings, without underlying information about hourly wages or hours worked in a quarter. Therefore, the data do not indicate whether increased earnings are due to wage rate increases or more hours of work. Also, since leavers without earnings in the quarter are excluded when calculating mean earnings, the earnings increases could also be due to low earners dropping out of the labor market.

6.  Recidivism rates would be higher if the studies had included those who exited for less than two months. Also note that recidivism was generally lower in studies that measured it on a monthly basis than in those that observed welfare receipt over a quarterly (three-month) period.

7.  The potential for measurement variation can be seen in the fact that two studies — Missouri and San Mateo — more than doubled their initial enrollment rates from administrative data. In both cases, researchers re-analyzed administrative data and classified additional eligibility codes as "Medicaid" enrollment, after noting large discrepancies between administrative and survey data. Earlier syntheses of findings from leavers grantees showed a wider range in Medicaid enrollment rates, based on the initial reports by Missouri and San Mateo.

8.  In fact, both quarterly earnings and hourly wages of leavers in Illinois and the District of Columbia were comparable or higher than those in the other regions.

9.  Poverty rates were 41 percent in Iowa (counting cash and food stamps), 47 percent in Iowa (counting cash income only), 57 percent in Cuyahoga County (counting cash and food stamps), and 58 percent in both Missouri and Washington (counting cash only). The official poverty measure does not include food stamps; food stamp benefits are included, however, in alternative poverty measures recommended by a panel from the National Academy of Sciences. The Panel on Poverty and Family Assistance also recommended that poverty measures take into account the effects of other non-cash benefits, taxes (such as the EITC) and work expenses.

10.  The six-item scale is an abbreviated version of a broader 18-question scale developed by the U. S. Department of Agriculture. Families that answer yes to two or more questions on the six-item scale are considered "food insecure" and those that answer yes to five or more questions are considered "food insecure with hunger." National estimates of food insecurity, based on the 18-item scale, indicate that 37 percent of families below the poverty threshold were food insecure in 1999, including 12 percent who were food insecure with hunger (U. S. Department of Agriculture, Household Food Security in the United States, 1999). Estimates from the six-item scale are generally comparable with those from the broader scale.

11.  The atypically high rates of homelessness (19 percent) and removals of children (17 percent) were from the same study, Florida. These results should be viewed with caution, because figures were imputed for the 73 percent of the sample that could not be located by telephone. The raw, unadjusted percentages, reported in an appendix, were closer to those reported by other states (4 percent homeless and 8 percent with children living elsewhere).

Where to?

Top of Page
Contents
Main Page of Report | Contents of Report

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