Status of Research on the Outcomes of Welfare Reform, 2000. Results/Findings from Outcomes-Funded Projects

A substantial proportion of our FY 1998 and 1999 welfare outcomes funding went to partnering with states and counties through grants to study the experiences of people who left the TANF program ("leavers") or were diverted from welfare. While most of the studies funded with the targeted Policy Research funds are still on-going projects, including the state/county grants examining families who have been diverted from welfare, interim findings are available from many of the ASPE-funded grantees studying leavers and a few other long-term projects.

Most of the leavers grantees have released interim reports based on linked administrative data sets that tracked families who left welfare, and several grantees also have already released final reports with findings based on information gathered through surveys. These reports show that, despite some outliers, there is a surprising amount of consistency among findings across the studies overall with respect to employment, hours worked and earnings, income and poverty and returns to TANF. There is much less consistency with respect to the use of Medicaid and Food Stamps. Although many such studies exist, comparisons across studies beyond the ASPE-funded studies are somewhat problematic because of the many differences in study populations, time periods studied, sources of data, and research methodologies. Final reports, with more detailed findings, are expected by the Summer of 2001.

Notwithstanding differences in studies, findings from the state/local leavers studies are quite consistent, particularly in the areas of employment and recidivism.

• Employment. Administrative data indicate that about 55 percent of former TANF recipients (ranging from 46 to 64 percent across 11 studies) are working each quarter after exit from welfare. Over the 12-month period, some former recipients lost their jobs, while others found new employment, resulting in cumulative employment rates of approximately 80 percent, measured as those who were ever employed within the first 12 months of exit. The remaining 20 percent never worked during the year after exit.

  Self-reported survey data from five reports suggest somewhat higher rates of employment, both at time of interview (57 to 65 percent) and over the first year since exit (85 to 90 percent).

  In general, employed leavers tend to work full time and earn between $7 and $8 an hour. Quarterly mean earnings of employed leavers ranged from close to $2,200 to over $3,400 across ten studies in the first quarter after exit, with earnings rising steadily in every location over the course of the year following exit. Household income information (not including the Earned Income Tax Credit) from a few states suggests that over half of all leavers are "poor" (having total incomes below the official poverty threshold, which was $13,003 for a family of three in 1998).

• Recidivism. In general, about one in six leavers (ranging from 8 to 19 percent) are receiving TANF in the fourth quarter after exit. The proportion that ever returned to cash assistance at some point during the first 12 months after exit is higher, ranging from 18 to 35 percent.

State and local reports show a wider range of outcomes across the studies with respect to post-exit use of Food Stamps and Medicaid.

• Food Stamps. Between one-third and one-half of leavers receive Food Stamps in the first quarter after exit in most states, although participation by single-parent leavers was as low as 9 percent in one county. When measured over time, three of eight states show participation declines of ten percentage points or more. The other states report relatively flat participation rates.
• Medicaid. Rates of Medicaid enrollment vary considerably across states, with enrollment among single-parent leavers in the first quarter after exit ranging from 26 to 57 percent of adults. In studies reporting rates for both adults and children, slightly more children than adults (28 percent of children compared to 26 percent of adults in one study, and 62 percent of children and 55 percent of adults in another) are enrolled in Medicaid. Half the states show a decline in Medicaid participation over time of ten percentage points or more. Similar to Food Stamps, other states show flatter participation rates over time.

The five studies that have completed analysis of their survey data all include an analysis of the hardships experienced by former recipients and their families. In general, many leavers experience some hardship with respect to food, housing or medical problems after leaving welfare. For example, 24 to 44 percent of families report not having enough food six months after exit; 27 to 39 percent report being behind on their rent, and 8 to 31 percent report an inability to afford or get medical attention. Interestingly, hardships are reported at about the same rate for TANF participants as TANF leavers. Overall, approximately 60 percent of leavers report being better off after leaving TANF than before leaving TANF.

In addition, a number of entities have undertaken non-ASPE-funded projects to monitor the effects of pre- and post-welfare reform initiatives. Appendix L of the 2000 Green Book: Background Material and Data on Programs Within the Jurisdiction of the Committee on Ways and Means summarizes findings from primarily pre-PRWORA welfare programs and pilot projects that have tested numerous state and local welfare reform initiatives. These studies vary substantially in terms of study design, cohorts, administrative data linkages, research topics, and response rates. Finding from impact studies are included as well as findings from leavers studies. The 2000 Green Book can be found at http://waysmeans.house.gov/publica.htm.

Some additional ASPE studies, supported in whole or in part by the targeted Policy Research funds, have been completed. Interim findings are also available from some other long-term projects. For example:

• Child-only cases. A study of child-only cases found that in 1999, 29 percent of TANF cases were child-only families in which only children received the benefits. Of those, about 40 percent consisted of children living with adult relatives who are not their parents, while parents were in the household in the remaining 60 percent of the cases, but they were ineligible for benefits for reasons such as sanctions, immigration status, or receipt of Supplemental Security Income (SSI) benefits.
• Immigrants. Three reports have been released to date from a study of welfare reform and the health and economic status of immigrants. One report found that the cash assistance and Medicaid approval rates declined by half for legal non-citizen families in Los Angeles between 1996 and 1998, with no corresponding decrease for citizen families. In addition, the cash assistance and Medicaid application rates for citizen children with non-citizen parents fell by almost half during the same period, while applications for children with citizen parents increased slightly. Another report found that welfare and immigration reform had more of a "chilling effect" on non-citizen use of cash assistance, Medicaid and Food Stamps than actual program eligibility changes. For example, between 1994 and 1997, a period when few legal immigrants would have been affected by benefit restrictions, the rate of decline within non-citizen households was more than double the rate of decline within citizen households. Finally, the third report found that one in ten children in America live in families where at least one parent is a non-citizen and one or more children are U.S. citizens, and that more children living in mixed status families have no health insurance when compared to all-citizen families.
• Representativeness of ASPE grantees. A study comparing ASPE grantee states to the nation as a whole found that the grantee states capture a diverse cross section of the U.S. experience, and that findings from these studies are helpful in representing the range of potential outcomes associated with welfare reform, although findings for the nation as a whole may differ in some dimension from findings in grantee states.
• Interim guidance from the National Academy of Sciences. The National Academy of Sciences Panel Study n Welfare Outcomes released its Interim Report providing guidance to the Department on welfare reform research and evaluation strategies. The short-run recommendations offered by the panel are consistent with the Department's research on welfare reform, and the report acknowledged that the Department is already taking most of the recommended steps.

The remainder of this chapter provides project-by-project summaries of the available results and findings from all outcomes-funded projects, including details on those studies highlighted above.

Conference Report language accompanying the targeted welfare outcomes research funding for Fiscal Years 1998 and 1999 urged the Department to submit its research plan to the National Academy of Sciences (NAS) to receive guidance on research design and recommendations for further research. Accordingly, we have provided over $1 million to the NAS to convene an expert panel to evaluate current and future welfare reform research. An interim report for this panel study, Evaluating Welfare Reform: A Framework and Review of Current Work, Interim Report, was released in September, 1999, and covers the NAS panel's early recommendations to the Department on welfare reform evaluation strategies. The final report is expected in late 2000.

The 30-month NAS Welfare Outcomes Panel Study was established to evaluate the design of current, proposed and future studies of the effects of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996. The panel's interim report provides a framework for conducting evaluations of welfare program changes, reviews current Departmental efforts to evaluate these changes, and provides the panel's initial conclusions and recommendations. The short-run recommendations for welfare evaluation strategies offered by the panel are consistent with the Department's research on welfare reform:

• set priorities on key welfare policy questions and concerns
• include a broad population of recipients, former recipients and potential recipients in welfare reform research
• place a high priority on improving the capabilities of data collection efforts
• work toward cross-state and within-state comparability of data and research on welfare program effects
• document and publish each State's TANF policies and changes to these policies on an ongoing basis
• encourage broader population coverage in existing leavers studies
• facilitate greater overall comparability in existing leavers studies

As noted in the body of the interim report, the Department is already taking most of the recommended steps. Our FY 1998, 1999 and 2000 studies cover a broad array of topics including the examination of diversion practices, an assessment of those leaving the TANF rolls, and participation in other programs such as Medicaid. Projects are also in place to study the effects of welfare reform on special populations (e.g., people with mental health and substance abuse problems and other disabilities, immigrant families) and the Department continues to provide leadership in national-level survey work. This research agenda complements other public and private efforts to assess the outcomes of welfare reform. The Department is also working to facilitate greater comparability in state and local level studies through grantee meetings, list-serve discussions, guidance to states and a technical assistance contractor. Furthermore, our strategy of providing grants to states and local TANF agencies is explicitly designed to increase state and local capacity for data collection efforts.

Welfare caseload numbers released in August showed that the number of welfare recipients has decreased from 14.1 million in January 1993 to 6.3 million in December 1999 - a drop of 56 percent, or 7.8 million. Nearly three-quarters of this overall decline has occurred since the welfare reform law (the Personal Responsibility and Work Opportunity Reconciliation Act) was enacted in 1996; 1999 caseloads were roughly half of what they were in 1996. 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?

In an attempt 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 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 somewhat comparable studies were also funded in Wisconsin, through the Institute for Research on Poverty, and South Carolina, resulting in a total of 15 studies of former recipients funded in FY 1998. ⁽¹⁾

These states and counties represent a diverse range of state policies and underlying economic and demographic conditions. In addition, study designs varied across the sites. In most studies, researchers are using a combination of linked administrative data sets and surveys of former recipients to monitor the economic status and general well-being of families leaving welfare, but each grantee followed its own proposed study methodology. In general, the grantees' research questions can be grouped into eight broad research topic areas: employment and earnings, case closures and recidivism, other income supports, health insurance, child care, child well-being, barriers to self-sufficiency, insecurity/deprivation, and other topics.

ASPE took certain steps to promote comparability across the studies, including developing a consensus definition of "leavers" as those who remain off welfare for at least two months, encouraging researchers to report outcomes for "single-parent" leavers if different from the entire study population, and gaining agreement to exclude or disaggregate closures of "child-only" cases from the study population. Brief summaries of the projects can be found at <http://aspe.hhs.gov/hsp/leavers99/fy98.htm. A short paper synthesizing the findings to date of these ASPE-funded studies is being prepared by the Urban Institute and is expected to be released in the Fall, 2000. (When it is available, it will be posted at http://aspe.hhs.gov/hsp/index.htm.) A more comprehensive report is expected by Summer 2001.

As of July 2000, 11 ⁽²⁾ of the 15 studies initiated in FY 1998 had released reports of preliminary outcomes based on administrative data findings for an early cohort of families that left welfare as between 1995 and 1998. Findings from these studies focus principally on employment and earnings, returns to cash assistance, and program participation in Medicaid and food stamps. Four ⁽³⁾ of these 11 studies have also released reports containing both administrative data and survey data findings for a later cohort who left welfare after the transition from Aid to Families with Dependent Children (AFDC) to the Temporary Assistance for Needy Families (TANF) program. Preliminary survey findings supplement the administrative data in the areas of household income and experiences of material hardship of former recipients. While cross-site differences reflect a variety of conditions ⁽⁴⁾, it should also be noted that outcomes can vary significantly across various subgroups. Highlights of aggregate findings for "average" welfare leavers follow. ⁽⁵⁾ It should be noted that the findings are for single-parent leavers except where noted otherwise.

Employment and Earnings

• Employment

  Based on administrative data from ten sites for an early cohort, approximately 45 to 65 percent of former TANF recipients were working after leaving TANF. As shown in Table 1, most studies reported that 50 to 60 percent of leavers were employed in the first quarter after exit. Between 62 and 75 percent were ever employed in the 12-month period following exit, while 35 to 40 percent of leavers were employed in (but not necessarily continuously ⁽⁶⁾) all four quarters, indicating that many former recipients experienced spells of unemployment as well as employment. Results from four states reporting on two or more cohorts were mixed in terms of the stability of employment rates for later, in relation to earlier, cohorts - one state had higher employment rates for the later cohort, one state had lower employment rates, and two states had similar rates across the cohorts.

  Survey data from four sites found that self-reported employment rates were between one and ten percentage points higher than the rates based on administrative data. This discrepancy is not unexpected, considering that the administrative data findings are based on quarterly earnings records maintained by the state's unemployment insurance (UI) program, which do not cover earnings from self-employment, employment in the military or federal government, certain agricultural employment, and jobs across state boundaries.

• Earnings

  According to administrative data from ten study areas, mean quarterly earnings in the quarter immediately following exit from TANF ranged from about $2,200 to over $3,400 for those leavers with earnings. Table 2 shows that in every location earnings rose steadily over the course of the year following exit ⁽⁷⁾ to mean quarterly earnings in the $2,400 to $3,600 range. Data from three surveys not displayed on the table found that respondents who were employed reported average monthly earnings ranging from $958 (at 6 to 8 months after exit) to $1,101 (at 24 to 36 months after exit). The mean quarterly earnings observed in these studies suggest that annual earnings of employed welfare leavers were low, averaging no more than $10,000 to $13,000, despite the fact that many former recipients had jobs involving full-time work.

Table 1.

Table 2.

Program Participation

• Returns to TANF

  Families in the early cohort who left cash assistance between 1996 and 1998 did so before they were affected by either the federal or state time limits on benefit receipt. Thus they had the option of returning to cash assistance as needed. Administrative data from eight reports, displayed in Table 3, show that between five and 20 percent of welfare leavers returned to welfare within one quarter after exit. Over the next three months, an additional four to seven percent of leavers returned in most states, bringing the total population of former recipients receiving AFDC/TANF to between 10 and 28 percent at two quarters after exit. The recidivism rate increased only slightly over the next two quarters (with returns to welfare ranging from 12 to 29 percent at 12 months after exit), but the proportion who ever received assistance for at least one month during the first 12 months after exit was somewhat higher (ranging from 23 to 35 percent).

  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 on 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.

• Medicaid and Health Insurance

  Adults

Administrative data on Medicaid enrollment from the eight studies that reported on this measure showed dramatic variations across states, unlike the relatively consistent findings on employment, earnings and recidivism. As shown in Table 4, three months after exit, the Medicaid enrollment rate for the early cohort of adult leavers ranged from 24 to 76 percent, with most studies reporting rates between 35 and 60 percent. Participation in Medicaid declined over time in the year following exit in over half the states (dropping as much as 10 to 20 percentage points for adult leavers between the first and fourth quarters after exit), but remained relatively stable in three studies. Enrollment rates among "continuous leavers," i.e., those who do not return to cash assistance, were even lower, primarily because this group excludes leavers who return to TANF, most of whom are re-enrolled in Medicaid at the time of their return to cash assistance.

Survey data provides more complete information about the status of health insurance coverage. Data from the four studies reveal that one-third to one-half (33 to 53 percent) of adult leavers were covered by Medicaid at the time of interview, about one-fifth to one-third (20 to 34 percent)

Table 3.

were covered by employer or other health insurance, and about one-third (26 to 41 percent) of former adult recipients had no health insurance coverage. These studies suggest that there is little difference from state to state in private health insurance rates and that the difference in uninsurance rates among welfare leavers is due to Medicaid. State to state differences in Medicaid enrollment rates result from policy choices affecting Medicaid eligibility rules and

Table 4.

application procedures, and/or state and local administrative practices affecting families transitioning off welfare. Survey data from two states (discussed in the section on Material Hardship and displayed on Table 7 below) show the consequences of lack of health insurance coverage.

Children

As expected, children are covered under Medicaid (or the State Children's Health Insurance Program (SCHIP)) at considerably higher rates than adults. Survey data from four sites showed enrollment rates for children of leavers at six to 27 percentage points higher than rates for adults. One-half to two-thirds (51 to 67 percent) of leavers reported Medicaid/SCHIP coverage for their children. Similar to the adult rates, about one-fifth (20 to 23 percent) of leavers reported that their children had insurance coverage other than Medicaid. Unlike the adults, however, children were somewhat less likely than adults to have such insurance in the form of employer-sponsored coverage, and more likely to have it in the form of "other" insurance, including coverage through a non-custodial parent. Finally, between 11 and 29 percent of leavers reported that their children had no health insurance, although the majority of them should have been eligible for either Medicaid or SCHIP.

• Food Stamps

The percentage of leavers participating in the Food Stamp Program was lower than the percentage enrolled in Medicaid across almost all study sites. Administrative data from eight of nine studies reporting on food stamp use by the early cohort, shown in Table 5, found that roughly one-third to one-half (34 to 57 percent) of former cash assistance recipients received food stamps immediately after exit. Survey data from two studies had similar findings, reporting 33 to 50 percent participation six to eight months after exit. By one year after exit, food stamp participation rates had fallen to between 33 and 40 percent in seven of the nine studies. The percentage of leavers ever receiving food stamps in the first year after exit typically ranged from 57 to 67 percent. One study proved to be the exception to these trends - only 9 percent of leavers were participating in the first quarter after exit, and 14 percent in the fourth quarter. About 28 percent of leavers in that study ever received food stamps within one year after exit. The following section on Material Hardship and Table 7 below address the prevalence of food shortages reported in survey data from two states.

Poverty and Material Hardship

• Household Income and Poverty Status

  Information on total household incomes of welfare leavers is currently available only from the four studies with survey data. Across these studies, total monthly household income averaged from about $1,050 to over $1,400, with median household incomes even lower. See Table 6. Information on the distribution of income available from two studies showed that very few

Table 5.

leavers (2 to 4 percent) reported $0 in household income, while about half (43 to 54 percent) reported incomes of between $500 and $1,500 per month. In one study, eight percent of leavers had household incomes in excess of $3,000 per month two and a half years after exit from welfare.

Two state reports calculated a poverty rate for former recipients by multiplying monthly incomes by 12 and comparing these annual incomes to the federal poverty thresholds by family size. The estimated poverty rate of former recipients was the same in each state - 58 percent - despite the fact that one state measured poverty six to eight months after exit, and the other measured it two and a half years after exit.

All four studies provided some information about the sources of household income. Earnings were the single largest income source, with the leaver's earnings accounting for 60 to 65 percent of household income across three studies, and earnings of the leaver and/or others in the household accounting for 80 to 86 percent of total household income. Other contributors to aggregate household income included TANF benefits, child support, and Supplemental Security Income (SSI) benefits.

• Material Hardship

Across the three states that measured food insecurity, approximately one-fourth of former welfare recipients reported that they did not have enough to eat, or they cut meal sizes or skipped meals in the period since leaving welfare. Table 7 shows that in two of the sites, food shortages were somewhat more common among leavers than on-going TANF recipients, and of about the same frequency before and after exit in the third site.

Former recipients also reported problems with housing arrangements, although somewhat less frequently than food shortages. The most common problems were loss of utilities (12 to 14 percent of leavers across three studies) and being forced to move (13 to 17 percent across two studies). Less often, former recipients were evicted or went to a homeless shelter. An estimated eight percent of former recipients in two states reported that their children were forced to live elsewhere, and three percent of leavers in one state reported that their children went into foster care. None of the states reported that separations of children from the family occurred more or less frequently after exit than while the families received welfare benefits.

Finally, between one-quarter and one-third (24 to 31 percent) of leavers in two sites reported that since leaving welfare they or someone in their household was unable to get needed medical attention because they could not afford it. This may reflect the fact that 40 and 36 percent of leavers survey respondents in those two states, respectively, reported being uninsured. However, 14 and 26 percent of these same respondents also reported an inability to get needed medical attention during the last six months of welfare receipt, when almost all respondents should have had health coverage through the Medicaid program.

Table 6.

Table 7.

ther 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 forthcoming synthesis report by the Urban Institute will be posted to the same web site. 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. As it becomes available, 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

The Conference Report accompanying the 1998 targeted appropriation specified that the Department assess the potential for using food stamp administrative data to look at welfare "leavers" and nonparticipants at the national level. Historically, almost all of the households that received AFDC/TANF also participated in the Food Stamp Program. As reported in "Characteristics and Financial Circumstances of TANF Recipients: July-September 1997," approximately 85 percent of TANF families received food stamp assistance during that period, which was consistent with previous levels under the AFDC program. (By contrast, households with AFDC/TANF constituted just 35 percent of all food stamp households in 1997 because of its broader base of eligibility, e.g., elderly people.) Since so many AFDC/TANF recipients also receive food stamps, the potential for using Food Stamp Program administrative data to examine the status of TANF recipients after they leave the TANF caseload was an important avenue to explore.

ASPE contracted with Mathematica Policy Research, Inc. (MPR) to conduct the study. However, after extensive discussions with USDA's Food and Nutrition Service concerning case information in the National Integrated Quality Control System (NIQCS) and state food stamp administrative records, it became clear that significant technical obstacles would need to be overcome in order to use food stamp administrative data to track welfare leavers and nonparticipants, and that assembling such data would not be feasible. The project was, therefore, re-focused.

The goal of the re-defined project was to explore whether grantee-states have similar characteristics to the nation as a whole, and thus the extent to which welfare outcomes findings from the grantee states (which are summarized at the beginning of this chapter) might be indicative of the national TANF experience. Using early data from the grantee leavers studies, the Current Population Survey, HHS administrative data and the FY1998 Food Stamp Quality Control program, MPR compared 15 study states to the rest of the nation in terms of caseload declines, employment and earnings outcomes, TANF policies and economic and demographic indicators.

In the March 2000 report, Interpreting TANF Leaver Studies: Comparing ASPE Grantee States to the Nation as a Whole, MPR found that the grantee states capture a diverse cross section of the U.S. experience, and thus findings from these studies are helpful in representing the range of potential outcomes associated with welfare reform. However, they also found some important differences and thus conclude that findings for the nation as a whole might differ from those in grantee states. For example, for the nation as a whole:

• TANF exits may be more likely to result from increased earnings,
• families may be less likely to be diverted from collecting benefits,
• families may be better off in terms of material well-being, and
• families may be more likely to continue receiving food stamps after leaving TANF.

Thus, it is possible that in some respects welfare outcomes for families in the nation as a whole may be somewhat more positive than for families in the grantee states.

In fiscal year 1997, ASPE funded the University of Wisconsin Institute for Research on Poverty to conduct an administrative data study of the outcomes of families who left AFDC in Wisconsin during 1995. The reports produced during the first study provided useful early results for the Department on the short-term economic and employment outcomes of women who left AFDC prior to both the enactment of TANF and the implementation of Wisconsin Works (also known as W-2, Wisconsin's replacement for AFDC). For example, almost half (48 percent) of AFDC recipients in 1995 left welfare for at least two consecutive months between August 1995 and July 1996, but 30 percent later returned. Those who left and did not return were not as poor as those who returned; however, most families who left had incomes below poverty and only a small fraction had incomes above 150 percent of poverty.

A follow-up study, funded in FY 1999, builds on the earlier study and includes an examination of longer-term economic outcomes. Based on analyses of linked state administrative data, this study examines the employment and earnings outcomes of mother-headed families who stopped receiving cash assistance for at least two months beginning in the last quarter of 1995 or 1997. The Interim Report, Before and After TANF: The Economic Well-Being of Women Leaving Welfare, was published in May 2000 (see http://www.ssc.wisc.edu/irp/pubs/sr77.pdf). In comparing women who left welfare under early welfare reforms (i.e., in 1995) with those who left under Wisconsin's post-TANF welfare program (i.e., in 1997), the main findings are as follows:

• While 1997 recipients appear to have more barriers to work (e.g., less education, more children, very young children) than those in 1995, the rate of exit is much higher in the second time period (i.e., post-TANF).
• Nearly 70 percent of those leaving welfare in both cohorts left for employment, a somewhat higher employment rate than those found in our other studies (which are summarized earlier in this chapter) of women leaving welfare under recent reforms.
• Over 80 percent of women in both cohorts had at least some earnings during the first year after exit.
• Women in the second cohort who worked have lower earnings, are somewhat more likely to have multiple employers throughout the year, and are somewhat less likely to be employed in all four quarters after exit compared to women in the first cohort. (The data do not distinguish whether the multiple employers represent job changes during the quarter or multiple jobs.)
• Those who left welfare in the last quarter of 1997 were somewhat less likely to return to welfare than those who left in the earlier cohort.
• Participation in the Food Stamp Program is much higher among women who left welfare in 1997 than among those who left in 1995. For example, 82 percent of women who left welfare in 1997 received food stamps at some point during the first year after exit compared to only 57 percent of the earlier cohort.

In addition to pre- and post-TANF cohort comparisons, the study also examined the longer-term outcomes (i.e., three years after exit) of those women who left welfare in 1995. Several main points emerge from this analysis:

• A majority of women who left welfare in 1995 (88 percent) worked at some point in the first three years after exit; however, the percent of women with at least some employment in a given year declined somewhat over the three-year period (from 81 percent in the first year to 77 percent in the third).
• Among those who worked, earnings increased each year; however, three years after exit most of these women have earnings below 150 percent of the poverty line.
• The proportion of families that would be classified as poor based on own earnings, estimated taxes and Earned Income Tax Credit (EITC) payments, the cash value

Since 1997, HHS, the Department of Agriculture and the Department of Justice have been funding the Urban Institute under a cooperative agreement to study the impact of recent changes in Federal laws on immigrant families and children and to profile immigrants with regard to their health, employment, economic circumstances and participation in government programs. The project includes an examination of existing data combined with intensive primary data collection in Los Angeles and New York - two cities that together account for one-fourth of all the immigrants in the United States. The field work began in spring 1999. Intensive interviews will also be conducted with public and private community organizations that serve immigrants, as well as focus groups with immigrants affected by the new laws. National profiles of immigrant populations are being developed using secondary data and these will be compared with natives. Local administrative data will be used to map out relevant local trends in program participation and, where possible, to develop neighborhood indicators of health and other trends.

Three reports have been released to date, and several additional reports will be completed, culminating with a final report in October 2001. All reports from this project will be posted at <http://aspe.hhs.gov/hsp/hspother.htm>.

The first report, Declining Immigrant Applications for Medi-Cal and Welfare Benefits in Los Angeles County, analyzed administrative data to compare citizen and non-citizen rates of applications and approvals under AFDC/TANF, SSI, General Assistance, Medicaid, and food stamps in Los Angeles County, between January 1996 and January 1998. Over this period, approved applications for legal non-citizen families under Medi-Cal (California's Medicaid program) and AFDC/TANF fell by 52 percent, while there was no decline in approved applications for U.S. citizens. The number of U.S. citizen children with non-citizen parents applying for AFDC/TANF and Medi-Cal declined by 48 percent during this period, while the number of U.S. citizen children with U.S. citizen parents increased by 6 percent. Fewer immigrant families appeared to be applying for assistance, although the program eligibility rules did not change during this period. Application denial rates in the county remained steady during the period examined. The full report can be found at <http://www.urban.org/immig/lacounty.html>.

A second report, Trends in Noncitizens' and Citizens' Use of Public Benefits Following Welfare Reform: 1994-1997, analyzed Current Population Survey (CPS) data, comparing participation rates for U.S. citizens and non-citizens in cash assistance, food stamps and Medicaid programs. Nationally, overall use of public benefits by the total population declined between 1994 and 1997. However, during this period, the rate of decline within non-citizen households (35 percent) was more than double the rate of decline within citizen households (14 percent). During this time period, relatively few legal immigrants would have been affected by the benefit restrictions directed primarily to newer arrivals. Thus it appears that the steeper drop in overall non-citizen use of cash assistance benefits, food stamps, and Medicaid, reflects the "chilling effect" of welfare and immigration reform more than it does actual program eligibility changes.

Other evidence of the chilling effects is that refugees, who were given greater benefit protections under welfare reform than other immigrants, experienced participation declines at a rate (33 percent) about equal to declines among all non-citizens. Welfare use in non-citizen households with children declined at similar rates (36 percent). Overall, non-citizen households, who in 1994 were 9 percent of all households receiving welfare, comprised 23 percent of the drop in welfare caseloads between 1994 and 1997. The report can be found at: <http://www.urban.org/immig/trends.html>.

The third report is an analysis of the composition of immigrant households using 1998 CPS data. The report, All Under One Roof: Mixed-Status Families in an Era of Reform, found that "mixed status" families (i.e., families where at least one parent is a non-citizen and one or more children are U.S. citizens) are surprisingly prevalent in the United States. One in 10 children in America live in such families. As one might expect, these families are more prevalent in places where immigrants are concentrated. More than a quarter (27 percent) of all California families and 14 percent of all New York families are mixed status. Nearly half (47 percent) of all children in Los Angeles and more than a quarter (27 percent) of all children in New York City live in mixed-status families. Compared to other families, more mixed status families are low-income, and more children living in these families have no health insurance. The report can be found at <http://www.urban.org/immig/all_under.html>.

In 1999, 29 percent of TANF cases were child-only families in which only children received the benefits. We know from earlier data (1998) that in about 40 percent of these families the children are living with adult relatives who are not their parents (and benefits are paid only on behalf of the children under their care), while in the remaining 60 percent of the cases, parents are in the household, but ineligible for benefits for reasons such as sanctions, immigration status, or receipt of Supplemental Security Income (SSI) benefits. Before the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) was enacted, the numbers of child-only cases had been increasing steadily. Since then, the absolute numbers of such cases have declined, although the latest 1999 TANF caseload data show an increase in numbers over the previous year. As a proportion of the total TANF caseload, child-only cases have increased steadily. Unlike other TANF cases, child-only cases are exempt from many of the major requirements of the program, including time limits and work requirements, and the children in these families are likely to remain dependent on TANF for longer periods of time than other families.

In 1998, ASPE contracted with The Lewin Group to analyze national trends on child-only cases and, because the prevalence of child-only cases varies widely among states, to look at local characteristics and program practices in three states - California, Florida, and Missouri - and three counties in these states. The study obtained more detailed information about the characteristics of child-only families from administrative data and case files. State and county-level policy staff and caseworkers were interviewed to examine policy and programmatic issues.

The study found varying policy choices being made about sanctions, treatment of aliens, treatment of relative caregivers and treatment of SSI income. Although these policies affect child-only cases, the effects are not often considered in local decision making processes.

Most (two-thirds) non-parental caregivers were grandmothers, and they were substantially older than adults in regular TANF cases. The average age of non-parental caregivers in the three counties visited was 53. Non-parental cases have relatively higher incomes than parental child-only cases, but over a third of the non-parental cases qualified for food stamps. The primary reasons that children come to reside with non-parental caregivers are desertion, substance abuse, incarceration, child abuse and neglect. The report, Understanding the AFDC/TANF Child-Only Caseload: Policies, Composition, and Characteristics in Three States, <http://aspe.hhs.gov/hsp/child-only-caseload00/index.htm

One of the findings from the Lewin study was that all three states have created alternative programs for relative caregiver families that offer higher payments than TANF programs and have additional requirements for eligibility. These programs generally have less stringent eligibility requirements than foster care. To examine this issue further, ASPE has funded a follow-up project in FY 2000 to identify the types of alternative kinship care programs states have designed to serve relative caregivers outside of the traditional foster care or TANF systems.

This project promoted state efforts to develop and monitor indicators of the health and well-being of children as welfare reform and other significant policy changes occur. Capacity building was the fundamental goal of the project. It has been very successful in getting states to look seriously at their capacity to assess trends in children's well-being, especially for low-income populations, and to make improvements in capacity. ASPE, with additional support from ACF and the David and Lucile Packard Foundation, provided grants of approximately $50,000 each to fourteen states in both FYs 1998 and 1999. Each state was required to form a partnership of state government agencies including, at a minimum, the agencies and councils with responsibilities for children issues and services and the state welfare agency. Many states also included universities, community government agencies, or other partners. A brief description of each state's project is available at <http://aspe.hhs.gov/hsp/cyp/cindicators.htm>. States were provided with opportunities to work with one another, experts in various fields, and federal staff. Working from different starting points, states have designed and carried out a variety of activities, released a variety of reports on indicators, and disseminated other products. All states have indicated specific plans to continue their indicator efforts beyond the project period. Technical assistance has been provided to the grantees by the Chapin Hall Center for Children at the University of Chicago, which plans a final project meeting for Summer 2001. Products from the project will be finalized and disseminated by the technical assistance contractor.

ASPE undertook three related projects using the newly available source of welfare reform data collected through the Emergency TANF Data Report (ETDR): (1) working with the Administration for Children and Families' Office of Planning, Research, and Evaluation, to improve the collection, storage, and release of these data; (2) doing cross-sectional analyses using all data from all states; and (3) assessing the feasibility of creating a longitudinal database using data from individual states. First we had to resolve technical issues associated with cheaply storing and transferring the large amounts of data required for a comparative analysis of AFDC and TANF, and then build an archive of TANF data for researchers, which is expected to be available on the Internet. Cross-sectional analyses compared AFDC and TANF using a "before vs. after" welfare reform notion, analyzing similarities and differences in the caseloads. Finally, we explored the research potential of developing a longitudinal data set, preferably combined with additional linked data sets for more in-depth analyses. Longitudinal analyses in this early phase of administrative data reporting focused on short-term outcomes, including changes in work participation, child care use, Medicaid, and household composition (especially to child-only cases and conversion of two-parent to one-parent cases). This feasibility work lays the groundwork for creation of state longitudinal databases.

This project organized into a data archive the Characteristics Surveys of AFDC recipients (which were carried out every two years from 1967 through 1979) and the annual AFDC Quality Control administrative data (collected from 1982 through 1996). These data facilitate comparisons between the TANF program and its predecessor AFDC program to analyze changes in the characteristics and size of the caseload over time. The Urban Institute, the contractor, made the data variable names and definitions consistent across years, as well as cleaned, documented, and published these data on the Internet for use by researchers wishing to compare the new world of welfare reform with the old world of AFDC, among other uses. The ASCII (plain text) data files, documentation of files and variables, and commands to create a SAS data set for each year are available at: <http://afdc.urban.org/>.

1. In FY 1999, an additional $837,000 was awarded for continuations and extensions of several of the FY 1998 projects. Also funded in FY 1999 were leavers studies in Iowa, Texas, and Contra Costa County, California, as well as several applicant/diversion studies.

2. Studies with administrative data findings that are included in this summary are from Arizona, the District of Columbia, Georgia, Illinois, Missouri, New York, Washington, Wisconsin, Cuyahoga County, OH, Los Angeles County, CA, and San Mateo County, CA. Links to most of the reports can be found at <http://aspe.hhs.gov/hsp/leavers99/reports.htm>.

3. Survey data included in the summary are from Arizona, Illinois, Missouri, and Washington.

4. 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. In addition, since survey findings reported here are from only four sites, they should not be presumed to be representative of the experiences of welfare leavers across the country. Differences in survey findings are also affected by differences in the questionnaires. See <http://aspe.hhs.gov/hsp/leavers99/cross.htm#comparing for information about comparing survey instruments.

5. The findings reported here and the accompanying tables are taken from a paper by prepared by ASPE staff, "A Cross-State Examination of Families Leaving Welfare: Findings from the ASPE-Funded Leavers Studies," which was presented at the 40^th Annual Workshop of the National Association for Welfare Research and Statistics (NAWRS) in August 2000.

6. Since UI records are based on quarterly earnings, reflecting any covered employment during that quarter. Individuals do not need to be employed in every month of the quarter for an earnings record to be generated.

7. The UI system collects data on aggregate quarterly earnings, without providing underlying information about actual wages, hours worked, or months worked in a quarter. Therefore, the data do not permit us to determine whether increased earnings are due to wage rate increases or increased hours of work. In addition, it should be noted that, 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.