United States General Accounting Office GAO Report to the Chairman, Committee on Finance, U.S. Senate, and the Chairman, Subcommittee on Human Resources, Committee on Ways and Means, House of Representatives April 1999 WELFARE REFORM Information on Former Recipients’ Status GAO/HEHS-99-48 United States GAO General Accounting Office Washington, D.C. 20548 Health, Education, and Human Services Division B-281749 April 28, 1999 The Honorable William V. Roth, Jr. Chairman, Committee on Finance United States Senate The Honorable Nancy Johnson Chairman, Subcommittee on Human Resources Committee on Ways and Means House of Representatives In recent years, states have reformed their welfare programs for needy families with children by strengthening and strongly enforcing work requirements for adults and imposing time limits on the receipt of cash assistance. During the same time period, the nation has experienced strong economic growth, and cash assistance caseloads have declined by 40 percent—from their peak of about 5 million families in 1994 to 3 million families as of June 1998. Many of the reforms, begun as demonstrations under the Aid to Families With Dependent Children (AFDC) program, were incorporated into federal welfare legislation—the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA). PRWORA ended the federal entitlement to assistance for eligible needy families with children under AFDC and created the Temporary Assistance for Needy Families (TANF) block grant, which makes $16.8 billion available to states each year through 2002 and is overseen by the Department of Health and Human Services (HHS) at the federal level. Specified goals of TANF include providing assistance to needy families so that children may be cared for in their own homes or in the homes of relatives; ending the dependence of needy parents on government benefits by promoting job preparation, work, and marriage; preventing and reducing the incidence of out-of-wedlock pregnancies; and encouraging the formation and maintenance of two-parent families. To help states achieve TANF goals, PRWORA gives the states increased flexibility over the design and implementation of their welfare programs; however, states are required to impose work requirements and enforce a 5-year lifetime limit on the receipt of federal assistance. These changes in welfare programs, designed to decrease dependency, combined with the dramatic declines in welfare caseloads, have generated interest among program administrators, state and local policymakers, welfare advocates, and the public in general about the condition of Page 1 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 families no longer receiving cash assistance under AFDC or TANF.1 Given the importance of knowing what has happened to the parents and children of families who have left welfare, and to better understand states’ progress in meeting TANF goals, you asked us to report on what is now known about families no longer receiving welfare. Because there are no federal requirements for states to report on the status of former welfare recipients,2 the only systematic data currently available on families who have left welfare come from research efforts initiated by states to meet their own information needs. As you requested, we (1) determined the extent to which states have reported information on the condition of families who have left welfare in the following key areas: economic status, family composition, and family and child well-being; (2) determined from generalizable state studies what is known about the status of former welfare families in the key areas; and (3) identified federal and state efforts to improve the usefulness of the data obtained through these state efforts. To do this work, we collected and examined reports—published by September 30, 19983—that were based on studies conducted or sponsored by states of families who left the AFDC or TANF rolls during or after 1995. We also spoke with state officials in the states that had published reports and with HHS officials, and we reviewed documents from several organizations that are monitoring states’ efforts to study former welfare families. We conducted our work between July 1998 and April 1999 in accordance with generally accepted government auditing standards. (See app. I for a more detailed discussion of how studies were identified, assessed, and compared.) Seventeen states have published information on the status of their families Results in Brief who have left welfare.4 Each of these states reported on the economic status of former welfare recipients, and the majority reported on family 1 For the purposes of this report, the term “welfare” refers to cash assistance received under AFDC or TANF. 2 PRWORA makes $1 billion available over 5 years to reward states that achieve high performance levels in meeting TANF goals. HHS has determined that the fiscal year 1999 High Performance Bonus will be based in part on job retention rates and the earnings gain rates of adults leaving welfare, outcomes that require states that choose to compete for the bonus to follow up with former welfare recipients who obtained employment. 3 For the states that had published reports by this date, we included additional information as it became available, including more recent reports and data. 4 These states are Idaho, Indiana, Iowa, Kentucky, Louisiana, Maryland, Michigan, Montana, New Jersey, New Mexico, Oklahoma, Pennsylvania, South Carolina, Tennessee, Washington, Wisconsin, and Wyoming. Page 2 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 composition and family and child well-being. The studies differed, however, in important ways, including the categories of families studied, geographic scope, the time during which families who had left welfare were tracked, and the extent to which the families for whom data were available are representative of all families in the sample. Taking these factors into account, we determined that studies from only 7 of the 17 states had enough information on a sample of families to generalize findings to most families who had left welfare in the state at the time of the study. These seven states’ studies reported that most of the adults in families remaining off the welfare rolls were employed at some time after leaving welfare. However, significant numbers of families also returned to the rolls. In the three studies that reported the information, from 19 to 30 percent of the families who left welfare returned to the rolls at some time during the follow-up period. Although the seven states’ studies generally had limited data on total household income, five reported that many families who had left welfare subsequently received noncash public assistance such as Medicaid and food stamps, indicating that families’ incomes were low enough to keep them eligible for these forms of government assistance. None of the studies reported on changes in family composition resulting from marriage or pregnancy after leaving welfare. Regarding measures of well-being, six states’ studies included data on homelessness or separation of children from their parents and reported no indication of increased incidence of these outcomes at the time of follow-up. Efforts are under way at both the federal and state levels to improve the usefulness of the data being collected to assess the status of former welfare families. Most states either are currently studying or plan to study former welfare families, and HHS has recently funded 14 projects to track and monitor families who have left welfare. The projects will cover families who leave welfare in 10 states, five counties in 2 other states, and the District of Columbia. These jurisdictions, which include three of the states whose studies are reviewed in this report, will receive technical assistance through HHS and from other states on developing their tracking efforts. State officials in many of the states whose studies we reviewed said they plan to continue studying former welfare families, in some cases with additional HHS support, but also on their own. Maryland, for example, plans to conduct a telephone interview of a sample of former welfare families to get more in-depth information on such items as the factors that helped families leave welfare, and Idaho is trying to locate additional Page 3 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 families who have left welfare in order to increase the ability of the data to support conclusions about the population of families leaving welfare in that state. In the future, these ongoing state efforts, many supported by HHS, should provide a more complete picture of the status of families who have left welfare. Under the AFDC program, many states received waivers from federal rules Background to strengthen work requirements for adults. In addition, some states began experiments with time limits on receiving cash assistance. Under TANF, states generally must impose work and other program requirements on most adults receiving aid and, when an adult does not comply, reduce a family’s benefit or, at state option, terminate the benefit entirely. Moreover, families receiving TANF face a lifetime limit of 5 years, or less at state option, of federal assistance. These reforms represent significant departures from previous state and federal policies for needy families with children and have been accompanied by large declines in the number of families receiving cash assistance, from an all-time high in 1994 of about 5 million families to just over 3 million as of June 1998. While numerous efforts are planned or under way to assess welfare reform nationally, currently little information is available on the status of families who have left welfare. Although families have always left welfare for a variety of reasons, including increased household income due to employment or marriage,5 once their cases were closed and the families no longer received assistance, they usually were not routinely tracked or monitored.6 However, in the new environment in which eligible needy families are no longer entitled to cash assistance and the emphasis is on moving families off welfare into employment, concern about the condition of families no longer receiving aid has increased. The Congress and others are interested in the employment status of former welfare recipients, changes in family composition resulting from marriage and pregnancy, and the overall well-being of these families and their children. While the Congress has earmarked $5 million for HHS to study the outcomes of welfare reform and has taken other steps to monitor the status of poor families as discussed below, states are not federally required to report on the condition of former welfare families. 5 See, for example, Mary Jo Bane and David T. Ellwood, The Dynamics of Dependence: The Routes to Self-Sufficiency (Cambridge, Mass.: Urban Systems Research and Engineering, 1983) and LaDonna Pavetti, The Dynamics of Welfare and Work: Exploring the Process by Which Women Work Their Way Off Welfare (Cambridge, Mass: Ph.D. dissertation, Harvard University, 1993). 6 Families have been tracked in the past if they were involved in a program evaluation or other targeted research effort. Page 4 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 States’ greater responsibility for welfare programs under PRWORA has increased states’ need for information to support program management and decision-making, as well as to respond to the information requests from a variety of interested parties, such as service providers, advocacy groups, and the media. For example, some state legislatures are requiring state welfare agencies to report on outcomes from their reformed welfare programs, including the status of former welfare families. Consequently, many states have begun to track former welfare families.7 The data these states are reporting are the major source of information currently available on the condition of families who have left welfare. Only those families who actually become welfare recipients and then leave the rolls are included in most state tracking studies. However, the changes in welfare can also have the effect of decreasing the number of families coming onto the welfare rolls. For example, many states have diversion strategies designed to prevent families from coming onto the welfare rolls by providing a needed service, such as child care or transportation, providing a one-time cash payment to overcome a barrier to employment, or requiring that applicants conduct a job search before receiving cash assistance.8 As a result, a comprehensive assessment of the postreform status of poor families with children would include information on TANF-eligible families who did not become welfare recipients as well as former welfare recipients. To provide information on the postreform status of all low-income families, not just former welfare families, the U.S. Census Bureau at the direction of the Congress is conducting a longitudinal survey of a nationally representative sample of families, paying particular attention to eligibility and participation in welfare programs, employment, earnings, out-of-wedlock births, and adult and child well-being. Data from this survey, called the Survey of Program Dynamics, will help researchers and policymakers understand the impact of welfare reform on the well-being of low-income families and children by providing information on whether welfare recipients are finding jobs, what their earnings are, and what types of support they need to make the transition from welfare to work. In 7 In this report, we use the terms “track” and “tracking” to refer to efforts to collect information on families who have left welfare, regardless of whether data have been collected one time only or over time. 8 See Welfare Reform: States Are Restructuring Programs to Reduce Welfare Dependence (GAO/HEHS-98-109, June 18, 1998); Richard P. Nathan and Thomas L. Gais, Overview Report: Implementation of the Personal Responsibility Act of 1996 (Albany, N.Y.: Federalism Research Group, The Nelson A. Rockefeller Institute of Government, Oct. 1998); and Kathleen Maloy and others, A Description and Assessment of State Approaches to Diversion Programs and Activities Under Welfare Reform (Washington, D.C.: The George Washington University, Aug. 1998). Page 5 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 addition, the Urban Institute is conducting a multiyear project monitoring program changes and fiscal developments, along with changes in the well-being of children and families.9 As part of this project, the Urban Institute has surveyed nearly 50,000 people to obtain comprehensive information on the well-being of adults and children as welfare reform is being implemented in the states. A second survey is planned for 1999. Full results from the Census Bureau and Urban Institute surveys may not be available until the year 2000. In addition, a plethora of studies are under way that will be providing information in the future on various aspects of welfare reform.10 Seventeen states have collected data and reported on the status of some State Studies former welfare families in the key areas of economic status, family Reported Some composition, or family and child well-being. The state studies differed in Information on the important ways, such as categories of families tracked, the length of time families were tracked, and the sources of follow-up data. Some of the Status of Families studies presented no information on a substantial portion of the sample Who Have Left families, limiting the usefulness of these studies for drawing conclusions about the status of most former welfare families in the state. We Welfare determined that studies in 7 of the 17 states had enough data on a sample of families who had left welfare to generalize sample findings to the population of former welfare families from which the sample was drawn. Seventeen States Reported We identified a total of 18 state-sponsored or -conducted studies in 17 Some Information on states—2 studies in Wisconsin and 1 in each of the other states—that Economic Status, Family reported on the status of families who left welfare in 1995 or later. The reports contain a broad range of information on economic status, family Composition, or Family composition, and family and child well-being. Figure 1 summarizes the and Child Well-Being kinds of information reported in each of the 17 states and classifies the information according to the three major areas of interest. All of the studies reported information on economic status, all but one reported on family and child well-being, and most reported some information on family 9 The Urban Institute, a research organization located in the District of Columbia, is conducting a multiyear project designed to analyze the devolution of responsibility for social programs from the federal government to the states, focusing primarily on health care, income security, job training, and social services. Initial results from the 1997 National Survey of America’s Families are available at the Urban Institute’s Web site at www.urban.org. The survey is representative of the nonelderly population in the nation as a whole and in 13 states: Alabama, California, Colorado, Florida, Massachusetts, Michigan, Minnesota, Mississippi, New Jersey, New York, Texas, Washington, and Wisconsin. 10 For a list of completed and ongoing studies of welfare reform, see the Web site www.researchforum.org, created and maintained by the Research Forum on Children, Families, and the New Federalism, National Center for Children in Poverty, 154 Haven Avenue, New York, N.Y. 10032-1180. Page 6 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 composition. Overall, 15 of the 17 states reported information in all three areas. (App. II lists the 17 states and their study reports.) Page 7 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 Figure 1: Categories of Information Reported in State Studies of Former Welfare Families Page 8 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 a The Louisiana study included information on families leaving welfare in metropolitan New Orleans only. b Two studies were conducted in Wisconsin. c Idaho measured this variable, but the variable was not discussed in its report. d South Carolina reported separately on this variable in “Child Maltreatment Among Former Clients of the South Carolina Family Independence Program,” Oct. 1998. e Tennessee performed an additional study of child well-being for a nonrepresentative sample of children whose parents had left the state’s welfare program between Dec. 1996 and Feb. 1997 that includes information on emotional, economic, and motivational well-being. Studies Differed in Because states generally initiated tracking studies to meet their own Coverage, Timing, and information needs, the 18 studies in the 17 states differed in a number of Data Sources important ways, including the categories of families tracked, geographic coverage, the time periods covered, and the timing and frequency of follow-up. The studies also differed in the sources of data used for tracking families who had left welfare. Table 1 summarizes key information on the studies, including the categories of families studied, the time periods involved, the frequency of follow-up, the time between leaving and follow-up, and the method of data collection. Table 1: Key Ways in Which State Studies Differed Categories of families and time periods Follow-up Data collection involved Frequency Timing method(s) Idaho Families who left TANF Twice 6 to 12 and 13 to 24 Mail survey July to Dec. 1997 months after exit Indiana Families receiving AFDC Once 12 to 18 months after Telephone survey May 1995 to May 1996 enrollment who subsequently left AFDC Iowa AFDC families assigned to Once 8 to 12 months into Telephone survey, or who volunteered for the assignment in-person interviews, Limited Benefit Plana Nov. case studies, and 1994 to Apr. 1995 review of administrative data Kentucky Families who left TANF Once 1 to 11 months after Telephone survey Jan. to Nov. 1997 exit (continued) Page 9 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 Categories of families and time periods Follow-up Data collection involved Frequency Timing method(s) Louisiana Families in metropolitan Once 1 to 4 months after exit Telephone survey New Orleans who left TANF Jan. to Mar. 1998 Maryland Families who left TANF Quarterly Up to 12 months after Review of Oct. 1996 to Sept. 1997 exit administrative data Michigan Families whose AFDC Biannually 3 and 6 months after In-home interviews benefits were terminated exit and review of in Apr. 1996 because they administrative data did not comply with program rules Montana Families who received or Once Up to 22 months after Telephone survey and left AFDC or TANF Mar. (survey) or exit review of 1996 to Sept. 1997 monthly administrative data (administrative data) New Jersey Families whose TANF Once 1 to 2 months after exit Telephone survey benefits were terminated Jan. to Feb. 1998 because of failure to comply with program rules New Mexico b Families who left AFDC Once Mail survey July 1996 to June 1997 Oklahoma Families who left or were Once 2 to 18 months after Telephone survey denied TANF Oct. 1996 to exit or denial Nov. 1997 Pennsylvania Families who left TANF Once 1 week to 11 months Telephone survey Mar. 1997 to Jan. 1998 after exit South Carolinac Families with a household Once 9 to 14 months after Telephone survey and member required to seek exit i some in-person employment who left nterviews TANF July to Sept. 1997 and had not returned at time of follow-up (continued) Page 10 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 Categories of families and time periods Follow-up Data collection involved Frequency Timing method(s) Tennessee Families who lost TANF Once Approximately 3 Telephone survey benefits Jan. to Oct. 1997 months after exit because they did not comply with program rules and TANF families whose head was employed full- or part-time Feb. to Oct. 1997 Washingtonc Single-parent families who Once 2 to 4 months after exit Telephone survey and left TANF Apr. to July 1998 review of administrative data Wisconsin Single, female-headed Five times Quarterly for 5 Review of families who left AFDC quarters after family administrative data July 1995 to July 1996 left welfare Families who left Once 5 to 10 months after Telephone survey and AFDC/TANF Jan. to Mar. exit in-person interviews 1998 and did not return prior to survey Wyoming Families who left TANF Once 1 to 15 months after Telephone survey Dec. 1996 to Feb. 1998 exit a The Limited Benefit Plan was part of Iowa’s welfare reform initiated under waiver. It was a short-term alternative assistance program for AFDC recipients who did not comply with program rules and for some volunteers. Families in the Limited Benefit Plan received reduced cash benefits that were subsequently terminated for a fixed period of time, after which the family could reapply for benefits. b Information not provided in report. c South Carolina and Washington reported on groups of families who had left welfare earlier. We included the most recent sample in our summary. Source: GAO analysis of state studies. Fourteen of the studies reported data on a statewide sample of families who left welfare for a range of reasons, and one study reported on a sample of families who left welfare in the state’s major city. The remaining three studies focused primarily on families who left welfare because of an adult recipient’s failure to comply with program requirements. These three studies were conducted, at least in part, because of concerns about the potential impact on family well-being of the loss of the entire cash benefit, Page 11 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 rather than just a reduction in benefits, as was typically required under AFDC for noncompliance. None of the studies reported specifically on families who had left welfare because of time limits. While there is great interest in the status of these families, in most states few families have reached their time limits, and in states where they have, few families have lost benefits as a result of the time limits.11 However, as states’ programs mature and more families reach the federal 5-year time limit on TANF benefits or state-established time limits of shorter duration, more of these families will be included in the tracking studies.12 The studies also differed in the time period during which families left welfare and the length of time between the family’s exit and the study follow-up. The time at which states initiated a study of families who had left welfare depended in part upon when states’ reforms were implemented and when they needed information on the status of families affected by the reforms. The time periods of the 18 state studies ranged from as early as 1995 (before federal welfare reform) to as late as 1998 (after TANF was implemented in most states). The amount of time between leaving welfare and the follow-up also varied, ranging from 1 to 24 months. There were also differences in the frequency of follow-up. At least one state, Maryland, has been tracking families who have left welfare for a number of years and plans to track monthly samples of families for 2 years after they leave the rolls, whereas other states planned a one-time follow-up effort. In addition, the studies used different sources of data to locate and track families. The Maryland study and the first Wisconsin study relied solely on administrative data, while other states’ studies were based on surveys of the former recipients using in-home visits, the telephone, or the mail. Some states’ studies used both survey and administrative data. Administrative data are case-specific information from the files of various programs, services, or agencies, including state unemployment insurance, food stamps, Medicaid, child welfare, child support enforcement, and 11 As part of its overall evaluation of Florida’s Family Independence Program, the Manpower Demonstration Research Corporation conducted interviews with 25 families 6 months after the families lost benefits because of time limits. At least half of the 25 families reported that since their benefits had expired it had become more difficult to make ends meet. In general, however, there was no evidence of major changes in housing status or living arrangements, nor was there evidence that most respondents lacked the means to buy food. 12 The Manpower Demonstration Research Corporation is also conducting a cross-state study of welfare time limits in Florida, Vermont, and Wisconsin. Page 12 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 criminal and education agencies. Since administrative data are limited to data collected for program management purposes, they may not be as focused on the questions of interest as are the survey data. On the other hand, administrative data may be less expensive to collect and more accurate than the self-reported data and can more readily than a survey provide information on large numbers of individuals. Findings From Most State We determined that eight tracking studies, covering seven states, (1) were Studies Were Not designed to include most families who left welfare in the state at the time Generalizable to the of the study and (2) had sufficient data on the sample of families tracked for the sample to be considered representative of families studied.13 These Population of Former studies were designed to include families who left welfare for a range of Welfare Families in the reasons, although the studies varied in the specific category of families State covered. For example, the Maryland study included all families who had left welfare, while the South Carolina study included only families with a household member required to seek employment who subsequently left welfare and had not returned at the time of follow-up. Although none of the 18 studies were able to locate all families included in the samples to be tracked, eight studies had sufficient data on a sample of families to conclude that the sample represented the population from which it was taken. The nonresponse rates ranged from 15 to 88 percent for the state surveys. For the two studies using administrative data only, information about 8 percent and 18 percent of the families being tracked could not be found in the data being used. (See app. I for the proportion of families located in all 18 studies.) Missing information for some members of a sample raises concerns about the representativeness of the remaining sample and whether findings can be generalized to the population from which the sample was drawn. Families who left welfare and subsequently responded to a survey and families about whom information was available in administrative data may be different in important ways from families for whom no information is available; thus, results based on such families are not generalizable to the entire population of families who left welfare in a state. Some policymakers and researchers are concerned that families who do not answer surveys or whose current status is no longer reflected in administrative data might be worse off than families for whom there are 13 While the Iowa study had an 85-percent response rate, results could only be generalized to families assigned to Iowa’s Limited Benefit Plan, and not to families leaving welfare for other reasons. Page 13 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 data.14 While the families who were not located may have fared quite well in terms of employment or family formation, some missing families may be experiencing hardship. For the purpose of summarizing findings, we either included only those studies that had data on at least 70 percent of the sample of families from the population of interest in the state or included a nonresponse analysis that showed no important differences between respondents and nonrespondents. The seven states that we determined to have studies with results generalizable to their welfare populations are Indiana, Maryland, Oklahoma, South Carolina, Tennessee, Washington, and Wisconsin. We estimated that these seven states accounted for about 8 percent of the number of families who left welfare nationwide between October 1993 and June 1997. Figure 2 highlights these 7 states, along with the 10 other states that reported information on former welfare recipients. 14 In five states, follow-up was done by telephone with no provision for families without telephones. In one of these studies, only families for which the state had telephone numbers were included in the sample to be tracked. Since it is reasonable to assume that families without telephones could differ in important ways from families with telephones, findings based on a sample of families with telephones are unlikely to be the same as findings based on a sample that includes families that have no telephones. Page 14 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 Figure 2: States That Have Reported Data on Families Who Have Left Welfare Studies in the seven states had either (1) data on a high enough percentage of the sample to reasonably generalize the results to the population from which the sample was drawn or (2) an analysis showing that the Page 15 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 nonrespondents had some of the same key characteristics as the respondents, providing greater assurance that the results from the limited sample could be generalized to the population from which the sample was drawn. (See app. I for a more detailed discussion of our assessment.) Because the seven states’ studies differ in several ways, as discussed Most Adults in Former above, the results are not completely comparable across states. However, Welfare Families Were the studies provide an indication of the status of families who had left Employed at Some welfare in these states at the time of the studies and, to the extent that the results are consistent, suggest a pattern of what is happening to these Time After Leaving families. The studies had consistent findings on employment and earnings. Welfare; Little Else Is Most former welfare families had an adult who was or had been employed since leaving welfare. Although the studies indicated that former Known About Family recipients often worked at low-wage jobs, little information was available Well-Being on families’ total household incomes, which could include child support or earnings from a second worker. Some studies also reported that significant proportions of the families had returned to welfare. In general, the studies provided little information on family and child well-being. Adults Had Employment Employment rates ranged from 61 to 87 percent for adults in the families Rates of 61 to 87 Percent, who left welfare in the seven states; however, these employment rates but Little Is Known About were measured in different ways. Studies measuring employment at the time of follow-up reported employment rates from 61 to 71 percent. Household Income Studies measuring whether an adult in a family had ever been employed since leaving welfare reported employment rates from 63 to 87 percent. In the four studies reporting both employment measures, the percentage employed at some time since leaving welfare was considerably higher than the percentage reporting employment at the time of follow-up. (Table 2 summarizes employment and earnings data in seven states.) These employment rates generally exclude families who returned to welfare, which can be a substantial portion of the families who leave welfare. In the three studies for which such data were available, the percentage of the families who initially left welfare and then returned to the rolls ranged from 19 percent after 3 months in Maryland to 30 percent after 15 months in Wisconsin. Removing families who return to welfare from the employment rate calculations results in higher employment rates than Page 16 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 when they are included, since many former recipients who return to the welfare rolls are not employed.15 15 For example, if 2,000 families leave welfare and, at a 3-month follow-up, 1,000 are employed, 500 remain off welfare but are not employed, and 500 have returned to welfare and are not working, the 3-month employment rate is 50 percent for the entire 2,000 families and 67 percent if only the 1,500 families remaining off welfare are included in the calculation. Page 17 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 Table 2: Employment and Earnings Data From Studies in Seven States State and period Average number of during which families Employed at time Ever employed since Average hourly hours worked per Average earnings studied left welfarea of follow-up leaving welfare wage rateb week per quarterc Indiana, 1995-96d 64.3% 84.3% $6.34 32 $2,637 e Maryland, 1996-97 63.0f e e 2,384f e Oklahoma, 1996-97 64.5 6.51 34 2,877 South Carolina, 1997 61.8 85.6 6.45 36 3,019 g e Tennessee, 1997 61.0 5.67 37 2,727 Washington, 1998 71.0 87.0 8.09 36 3,786 h e e e Wisconsin, 1995-96 82.1 2,378i j Wisconsin, 1998 62.0 83.0 7.42 36 3,473i Note: Except where noted, these data include only families who did not return to welfare. a The year indicates the period during which the families studied left welfare. For more detailed information on the different time periods and frequency and length of follow-up in these studies, see table 1. b These figures represent the mean wage. While the mean wage tends to be higher than the median wage, we did not have the median wage for all studies. c For all studies except Maryland’s and the first Wisconsin study, we had to estimate quarterly earnings on the basis of reported average hourly wage and average number of hours worked per week. Because it is unlikely that all members of the sample worked all 13 weeks in a quarter, most of these estimates are likely to be somewhat higher than the actual average earnings per quarter. d Wage and earnings data for Indiana include those of recipients with earned income who were also on welfare. Because Indiana did report that average wage rates were significantly higher for former welfare recipients than for those combining work and welfare, the average wage rate for the combined groups may underestimate the wage rate for former recipients who are no longer on welfare. e Data were not available. f This figure also includes individuals who returned to welfare. g The Tennessee study reported separately for families who left welfare because of noncompliance and for those who were employed, whether on or off welfare. Employment rates presented here are for both groups, whereas wage data and earnings estimates are for the employed group only. h These data are based on a study using administrative data for families leaving welfare from July 1995 to July 1996. i Caution must be used in comparing these earnings figures because the earlier study used administrative data and the later one used survey responses. The administrative data may underestimate earnings because not all earnings were included. The survey data may be more inclusive of earnings but, because the data were self-reported, they could understate or overstate earnings. j These data are based on interviews with families who left welfare from Jan. 1998 to Mar. 1998. Page 18 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 While all eight studies reported some information on former recipients’ earnings or wages, the studies did not provide a complete story on hourly wages or number of hours worked.16 Average quarterly earnings for former recipients ranged from $2,378 to $3,786 in the studies that either reported quarterly earnings or for which we estimated quarterly earnings. Extrapolating these quarterly earnings to a year results in average annual earned incomes ranging from $9,512 to $15,144. These amounts of annual earned income are greater than the maximum annual amount of cash assistance and food stamps that a three-person family with no other income could have received in these states.17 However, if these earnings were the only source of income for the families after they left welfare, many of them would remain below the federal poverty level.18 The question of whether a family is economically better off after leaving welfare than when receiving cash assistance is quite complex. The answer depends on many factors, including the amount of the cash benefit while on welfare, which varies by state, family size, and earnings while on welfare; family earnings and other sources of income; and aid after leaving welfare, as well as any work-related expenses. For example, the 1995-96 Wisconsin study that tracked families for more than 15 months after they left welfare compared postwelfare earnings of these families to the maximum benefit they could have received under AFDC to see if families were economically better off after leaving welfare. The study found that whether postwelfare earnings exceeded the maximum AFDC benefit depended in part upon the number of children in the family. Postwelfare earnings exceeded the maximum AFDC cash benefit for 54 percent of the families with one child and for 41 percent of the families with three or more children. The study also noted that because some families combine welfare and work, the combination of the cash benefit and earnings could result in some families on welfare having more cash income than families with earnings only. The study showed that during their first year off welfare, less than half of the families had cash incomes higher than their incomes had been while on AFDC, including both benefits and earnings. 16 For another study of employment and earnings of families leaving welfare, see Sharon Parrott, Welfare Recipients Who Find Jobs: What Do We Know About Their Employment and Earnings? (Washington, D.C.: Center on Budget and Policy Priorities, Nov. 1998). 17 In these seven states, for a single-parent, three-person family with no income, the maximum annual amount of cash assistance and food stamps combined ranged from $6,000 in Tennessee to $9,744 in Washington, as of Jan. 1997. 18 For 1998, the federal poverty level for a family of three was $13,650. We estimated the average annual earnings from the studies of the seven states as follows: Indiana—$10,548; Maryland—$9,536; Oklahoma—$11,508; South Carolina—$12,076; Tennessee—$10,908; Washington—$15,144; and Wisconsin—$9,512 and $13,892. In two of the studies estimated earnings were above the federal poverty level. Page 19 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 While the tracking studies provide information on individuals’ earned incomes, much remains unknown about families’ total household income. For example, the studies generally do not provide information on whether others in a household have earnings or on other sources of household income, such as child support payments or financial assistance from relatives and friends. Moreover, most of the studies do not include comprehensive information on the receipt of other noncash benefits, such as food stamps, Medicaid, and child care or transportation assistance, or what employment-related expenses, including child care and transportation, households may have. Only three of the eight state studies had some information on household income. In the Oklahoma study, 57 percent of the former welfare families reported household incomes at or below the official poverty level. In the Indiana study, 57 percent of the families off welfare at follow-up reported monthly household income below $1,000. In contrast, Washington reported average total family income, including child support payments, equal to 130 percent of the federal poverty level for a family of three. According to the Washington study, 35 percent of the families who left welfare and had children received some child support, and 36 percent had at least one worker in the family other than the respondent to the survey. The 1995-96 Wisconsin study found that the proportion of families who had left and remained off welfare for at least 1 year who had earnings above the official poverty level varied by family size. While 35 percent of the families with one child and 24 percent of the families with two children had earnings above the poverty level, only 11 percent of the families with three or more children did. Although these studies do not provide a comprehensive picture of families’ financial situations, they consistently indicated that many of the families leaving welfare were employed at fairly low-paying jobs. Our recent report on TANF implementation in seven states and other studies indicate that many states and localities are providing support services, such as case management services and financial assistance with child care, to help former welfare recipients maintain their employment. Several states and localities have also undertaken efforts to help these low-wage workers upgrade their job skills to improve their job prospects.19 Moreover, the recently expanded earned income credit can increase the incomes of 19 See Rebecca Brown and others, Working Out of Poverty: Employment Retention and Career Advancement for Welfare Recipients (Washington, D.C.: National Governors’ Association and HHS, 1998); Mark Elliott, Don Spangler, and Kathy Yorkievitz, What Next After Work First? (Philadelphia: Public/Private Ventures, spring 1998); and Brandon Roberts and Jeffrey D. Padden, Welfare to Wages: Strategies to Assist the Private Sector to Employ Welfare Recipients (Chevy Chase, Md.: Brandon Roberts and Associates, Aug. 1998). Page 20 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 qualified low-income families by as much as $2,271 for families with one child and $3,756 for families with two or more children.20 Information on total household income and receipt of government supports is key to understanding the condition of former welfare recipients and the extent to which they continue to rely on government aid rather than becoming economically self-sufficient.21 The studies in five states reported on the extent to which former welfare families say they receive noncash public assistance.22 As shown in table 3, in these states, between 44 and 83 percent of the families who left welfare received Medicaid benefits, and between 31 and 60 percent received food stamps. The Wisconsin study that tracked families who left welfare between July 1995 and July 1996 for 15 months found significant decreases in the use of noncash public assistance over time. Forty-six percent of the former recipients who remained off welfare for at least 1 year received both Medicaid and food stamps in the first quarter after leaving welfare, and 28 percent received both in the fifth quarter after leaving cash assistance. Four studies had information on the receipt of child care subsidies. 20 The earned income credit is a refundable tax credit for qualified working people who have earned incomes below certain specified levels. 21 Research conducted in 1996 by Cancian and Meyer using the National Longitudinal Survey of Youth to trace welfare use, poverty status, and primary sources of income in the 5 years following an exit from welfare reports that although 40 percent of the AFDC exiters do not return to AFDC within the first 5 years, only 20 percent went the entire 5 years without using AFDC, food stamps, or Supplemental Security Income. See Daniel Meyer and Maria Cancian, Life After Welfare: The Economic Well-Being of Women and Children Following an Exit from AFDC (Madison, Wis.: Institute for Research on Poverty, discussion paper no. 1101-96, Aug. 1996). 22 While food stamp and Medicaid caseloads have declined recently, the reasons for these changes are not fully understood. We do not know, for example, how many former welfare families that are eligible for these benefits actually receive them. There is concern that some needy families no longer receiving cash assistance may be unaware that they are still entitled to Medicaid and food stamps. While a strong economy may also partially explain recent declines in food stamp and Medicaid caseloads, the relative contributions of this and other such factors to these trends are unclear. Page 21 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 Table 3: Percentages of Families Reporting Receipt of Public Assistance After Leaving Welfare Percentage receiving State and period during Supplemental Women, Infants, which families studied left Medicaid Security Income and Children Child care welfarea coverage Food stamps benefits benefits subsidies b Indiana, 1995-96 53 38 11 25 Oklahoma, 1996-97c 70 50 17 b 23d South Carolina, 1997 80 60 13 27 18 b b Washington, 1998 64 children; 44 45 38e adults b b b Wisconsin, 1995-96 83 in first quarter 49 in first quarter after leaving after leaving welfare; 56 in fifth welfare; 31 in fifth quarter after quarter after leaving leaving Wisconsin, 1998 71 49 17 38 17 Note: Unless specified, the percentage is of all families in the sample. Except for the earlier Wisconsin study, all data are from surveys in which respondents self-reported their receipt of aid. a For more detailed information on the different time periods and frequency and length of follow-up in these studies, see table 1. b This information was not reported. c Data include families that returned to work. d For respondents who are employed or in training and whose households include children, the percentage reporting the receipt of a child care subsidy increases to 34. e Families with children under 13 whose parent is at work. While receiving AFDC or TANF, families generally also receive Medicaid benefits to cover their health expenses. However, whether Medicaid benefits are retained after a family has left welfare depends on many factors, and health insurance coverage after leaving welfare varied in the states with these data.23 For example, about 9 percent of the children in families who left welfare in South Carolina, about 20 percent in Oklahoma, and 35 percent in Indiana did not have health insurance at the time of follow-up. For adults who left welfare in these states, 24 percent in Oklahoma, 32 percent in Washington, 48 percent in South Carolina, and 54 percent in Indiana did not have health insurance. 23 To ensure continued Medicaid coverage for low-income families, PRWORA generally preserves the Medicaid entitlement, setting eligibility standards at AFDC levels in effect on July 16, 1996. In addition, under certain conditions, families that leave TANF for employment may continue Medicaid for 12 months. Medicaid coverage is also available for many low-income children, even if their parents are not eligible. For more information, see Welfare Reform: Early Implications of Welfare Reform for Beneficiaries and States (GAO/HEHS-98-62, Feb. 24, 1998). Page 22 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 While much attention is paid to welfare recipients who become employed and stay off the rolls, there is also interest in how those who are not employed and have not returned to welfare are faring. The South Carolina and Wisconsin surveys asked nonworking former recipients what stopped them from working for pay. In both states, the most frequently mentioned reason was their own physical or mental illness, followed by the inability to find a job, lack of transportation, and lack of child care. The Wisconsin study attempted to determine how these families were supporting themselves. Of the 142 former recipients not currently working, 18 percent were living with employed spouses or partners. Sixty-five percent of the families of the remaining nonworking former recipients were receiving Social Security, state unemployment insurance, child support, or foster care payments; 23 percent were not receiving cash assistance but were receiving noncash assistance, such as free housing, rent subsidies, Medicaid, or food stamps. Studies in Seven States The studies in seven states provided limited information on the family Provided Limited composition and well-being of former welfare families and the children in Information on Family these families. Although a major goal of welfare reform was the promotion of two-parent families and the reduction of out-of-wedlock pregnancies, Composition and the the tracking studies report only minimal information on family Well-Being of Children and composition at the time of data collection, and no information on changes Families that may have occurred just before or after leaving welfare. The studies with surveys asked questions regarding family composition; however, these surveys did not provide information on changes in the number of children in a family, changes in marital status, or the formation of other two-parent families since a family left the welfare rolls. Further, beyond any inferences that could be drawn from the employment and earnings of parents, the studies provided little information on how former welfare children and families were doing relative to housing, health, education, food security, substance abuse, crime, and victimization.24 While some of the studies provided limited information on some of these factors, there are no comprehensive data on family and child well-being. Three studies—from Maryland, Oklahoma, and Washington—reported on the number of children in former recipient families that had ever been involved with child protective services. These studies found few cases in 24 These and other factors are considered indicators of well-being. Fig. 1 lists a number of these items reported in the state studies. Page 23 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 which children had been involved with child protective services since leaving welfare.25 For example, the Maryland study reviewed state data from its foster care program to determine the number of children placed in foster care after their families left welfare. This study reported that less than one-half of 1 percent of the children studied entered foster care after their families left cash assistance. Two studies, South Carolina’s as well as Wisconsin’s recent survey of families leaving welfare during the first quarter of 1998, asked former recipients to compare several aspects of their general well-being after leaving welfare with their situation when they were on welfare. Because Wisconsin used a modified version of the interview schedule developed in South Carolina, the data are comparable, even though the programs that the recipients experienced are not. Table 4 shows the results from the two states’ surveys. Former welfare recipients in both states were more likely to experience some deprivations after leaving welfare than while on welfare. At the same time, in South Carolina and Wisconsin, 76 and 68 percent, respectively, disagreed or strongly disagreed with the statement that “life was better when you were getting welfare.” Regarding housing status, an important aspect of well-being, the limited information from the studies did not suggest increased incidence of homelessness at the time of follow-up. 25 South Carolina, in separate analyses, compared the number of incidents of maltreatment reported to the Child Protective Services’ Central Registry for a sample of families who had left welfare with the number of incidents for families still on welfare; it also compared the number of incidents of maltreatment in a sample of former welfare families before and after leaving welfare. The differences were not statistically significant for either comparison. Page 24 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 Table 4: Recipients’ Comparisons of Deprivations While on and After Being South Carolinaa Wisconsinb on Welfare (percentage responding (percentage responding “yes”) “yes”) Question On welfare Off welfare On welfare Off welfare Did you ever get behind in rent or house payments? 13 15 30 37c Did you ever get behind on a utility bill? 16 18 49 47 Was there ever a time when you could not buy food? 6 9 22 32c Was there ever a time when you could not afford child care when needed in order to work? 11 9 22 33c Did somebody in your home ever get sick or hurt when you could not get medical care? 1 7c 8 11 Did you have to go to a homeless shelter? 2 1 5 3 a Based on a sample of 403 former welfare recipients. b Based on a sample of 375 former welfare recipients. c These differences are statistically significant at the .05 level. Source: South Carolina’s Survey of Former Family Independence Program Clients: Cases Closed During July Through Sept. 1997 and Wisconsin’s Survey of Those Leaving AFDC or W-2 Jan. to Mar. 1998, preliminary report. The number of states conducting or sponsoring studies that track the State and Federal status of families leaving welfare has increased in recent years, and state Efforts Are Under and federal efforts are under way to improve the usefulness of the data Way to Improve the being collected. Thirty-nine states, including the 17 we identified in this report, and the District of Columbia already are tracking or plan to track Usefulness of State families leaving welfare. In an attempt to improve the quality and Tracking Studies comparability of these studies, HHS has funded several states and other jurisdictions to conduct tracking studies and is providing them technical assistance in conducting these studies. State Tracking Efforts Are We have identified 39 states and the District of Columbia that are either Increasing in Number and planning to study former welfare families or are already doing so. Most of Expanding in Scope the 17 states that we discuss in this report are planning to continue their Page 25 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 tracking and to enhance their efforts, in some cases with federal funds and in other cases with state funds. Maryland, for example, plans to survey former recipients to get the kind of detailed information about families’ lives that is not available in the program data upon which state officials currently rely in their ongoing longitudinal studies. Among the topics to be covered are how former welfare families are able to make ends meet; what enabled them to leave the welfare system; and, in the cases of those who returned to welfare, what brought them back. Idaho is trying to locate families that did not respond to its survey. Maryland, Massachusetts, South Carolina, Wisconsin, and Mecklenburg County in North Carolina have received funds from HHS to support their efforts to link administrative data systems for purposes of studying the effects of welfare reform on other state and federal public assistance programs. A South Carolina official told us that by linking TANF data to state unemployment insurance data, the state was able to locate many of the families that did not respond to its survey. Federal Efforts Are To increase the usefulness of state tracking efforts in providing a more Helping States to Improve complete picture of the status of former welfare families, HHS is supporting the Usefulness of the Data some states and counties with funds and technical assistance. As part of its overall strategy to evaluate welfare reform, HHS has awarded grants to Being Collected on 14 projects covering 16 jurisdictions—10 states, five counties in 2 other Families Leaving Welfare states, and the District of Columbia—to support efforts to track, through administrative data, surveys, or other methods, former TANF recipients’ work transitions and receipt of other benefits, including supportive services.26 Each of these tracking efforts plans to collect information on one or more of the following: families diverted from welfare, eligible families who do not apply for benefits, and families who have left welfare. All 14 grantees will collect both administrative data and survey data on former recipients. (See app. III for information on the 14 studies.) In addition, HHS submitted its overall research plan for evaluating welfare reform to the National Academy of Sciences for guidance on research design and recommendations for further research.27 The National Academy has convened a panel of experts on program evaluation methods, survey design, administrative record analysis, state database 26 In May 1998, HHS announced the availability of funds and requested proposals from states for research into the status of individuals and families who leave the TANF program and eligible families who are diverted or who fail to enroll. Of the proposals received, 14 awards were made. 27 HHS has funded other related research efforts, including assessments of the impacts of welfare reform on children and immigrants. Page 26 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 development and analysis, and welfare policy and evaluation to review data needs and methods. One of the panel’s first activities was to conduct a workshop to review and assess the HHS grantees’ study methodologies. The workshop provided a forum in which representatives from states and counties that had been awarded grants were able to talk with experts about their planned tracking studies. The panel plans a 30-month study on a broad range of issues related to evaluating welfare, with an interim report to be ready in June 1999. HHS expects to use data from the 14 funded projects to generate a picture of what is happening to families exiting welfare and families diverted from ever entering welfare. In recognition of the need for high-quality research and comparable findings, HHS is providing technical assistance to the states directly, and through the National Academy panel, and is encouraging grantees to share information with one another. The National Academy panel is providing advice on issues of data quality and comparability as well as policy relevance. Initially, the 14 grantees have agreed to work toward increasing comparability across studies by using a common definition of welfare “leaver.”28 They have also agreed to clarify which studies will be tracking only families with adults and which will also track welfare cases that only include children. Finally, with the encouragement of HHS and the National Academy, the grantees will be sharing common approaches to studying such areas as insecurity and deprivation, child well-being, and changes in household composition.29 While we were able to learn about the status of former welfare recipients Observations in several states, we could conclude little about the status of most families that have left welfare nationwide. However, the limited information on economic status of the families being tracked indicates that many families who leave welfare find jobs that are low-paying. The low wages of these jobs emphasize the importance that income supports, such as subsidized medical and child care and the earned income credit, can assume in these families’ total financial resources. As we noted in our earlier report on TANF implementation in seven states, federal and state policies and programs for assisting low-income working families are likely to play a 28 The grantees have agreed to define a “leaver” as a family that was off welfare for 2 months or longer. 29 In addition, HHS has funded five states—Connecticut, Florida, Indiana, Iowa, and Minnesota—to assess the effects of different welfare reform approaches on child well-being. The state agencies and research organizations involved in these projects are attempting to ensure compatibility of outcomes and measures to promote the ability to compare outcomes among states. Page 27 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 critical role in helping these families remain off cash assistance and move toward economic self-sufficiency. But much remains unknown about most families leaving welfare nationwide. In our attempt to describe the condition of former welfare families, we were constrained by the data currently available from these early state tracking studies. More specifically, the high nonresponse rates in many state studies limit the usefulness of the results because generalizations cannot be made to all families of interest. Because those families who do not respond to surveys, or who may not show up in administrative data for other programs, may be the ones at greatest risk of negative outcomes, some policymakers and program officials are particularly concerned about not having enough information to determine the status of these families. In addition, for policymakers to better understand whether states are making progress in meeting the goals of TANF, more comprehensive information is needed on household income; receipt of government assistance; and changes in family composition, including increases in the number of two-parent families and additional births, especially to teens. And, finally, the data often are not comparable among the states. Consequently, even if each state had collected generalizable data on a comprehensive range of topics, it would often be difficult to generate a national picture from such studies. More comparable data would also be useful to individual states that want to understand how former welfare families fare in their states as compared with those in other states. In addition, comparable data among the states could help policymakers and program administrators at all levels of government identify promising approaches and practices for assisting low-income families. The limited nature of the information currently available emphasizes the importance of additional state efforts, such as those funded by HHS. The ongoing state efforts promise to provide a more complete picture in the future. Many more states have tracking studies in progress or planned and efforts are under way at the state and federal levels to improve the usefulness of these efforts. As HHS continues to work with states to support their efforts to collect data on families who have left welfare, it has an opportunity to help states develop more generalizable, comprehensive, and comparable data. We obtained comments on a draft of this report from HHS, which stated Agency Comments that the report provides useful information on the status of former welfare Page 28 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 recipients and the varied efforts being made by states to follow up on the impacts of welfare reform. HHS also noted, however, how difficult it was to glean general results from such varied studies. It expressed concern that, while the report appropriately discusses the issues that preclude generalizing findings in many of the studies to the state level, the report does not address other factors, such as differences in the definitions of the populations studied and in states’ economic conditions, which make it difficult to report general results from any of the studies. HHS further suggested that rather than attempting to find areas of comparability in the studies, we should focus on the crucial differences between studies and emphasize the contribution that the HHS-funded state studies of families leaving welfare will make. Finally, HHS had concerns about our reliance upon studies with low response rates. We agree with HHS about the difficulties involved in discussing general results from the varied studies. We did not suggest, however, that results from the eight studies could be generalized beyond the states from which the study samples were drawn. We also pointed out that much remains unknown about most families who have left welfare nationwide. In addition, we were not attempting to report program impacts, which would require controlling for other factors that could affect family status, such as varying economic conditions. Rather, we focused on what is currently known about the status of former welfare families given the extent to which a particular state study was generalizable to the study population within the state. We also agree with HHS’ concern regarding low response rates, and this was the reason most of the studies were determined to not be generalizable and no data from them were included in the report. Finally, the report had already noted that we believe the HHS-funded state efforts will make an important contribution toward improving the usefulness of future studies and increase understanding of the condition of former welfare families. To address HHS’ concerns, we revised the report to place greater emphasis on the studies’ differences by moving detailed information on the studies’ varied populations, time periods, and methodologies from the appendix to the body of the report. We also added an additional caveat to the discussion about employment and earnings information, pointing out the lack of complete comparability among the studies. In addition, where results from several states were displayed, we added information on time periods and references to more detailed information on the studies’ populations and methodologies. HHS also made technical comments, which Page 29 GAO/HEHS-99-48 Status of Former Welfare Recipients B-281749 we incorporated where appropriate. (See app. IV for the text of HHS’ comments.) We also provided copies of the draft report to the 17 states whose studies we reviewed and to an expert on welfare reform issues. We incorporated their technical comments where appropriate. As agreed with your offices, unless you publicly announce its contents earlier, we plan no further distribution of this report until 30 days from the date of this report. At that time, we will send copies to the Honorable Donna E. Shalala, Secretary of the Department of Health and Human Services; state TANF directors; and other interested parties. We will also make copies available upon request. If you or your staff have any questions about this report, please contact me on (202) 512-7215. Other staff who contributed to this report are listed in appendix V. Cynthia M. Fagnoni Director, Education, Workforce, and Income Security Issues Page 30 GAO/HEHS-99-48 Status of Former Welfare Recipients Page 31 GAO/HEHS-99-48 Status of Former Welfare Recipients Contents Letter 1 Appendix I 34 Scope and Methodology Appendix II 38 Reports From States’ Studies of Families Who Left Welfare Appendix III 41 Information on Selected Grants Awarded by the Department of Health and Human Services to Study Families Leaving Welfare Appendix IV 44 Comments From the Department of Health and Human Services Appendix V 46 Major Contributors to This Report Related GAO Products 48 Tables Table 1: Key Ways in Which State Studies Differed 9 Page 32 GAO/HEHS-99-48 Status of Former Welfare Recipients Contents Table 2: Employment and Earnings Data From Studies in Seven 18 States Table 3: Percentages of Families Reporting Receipt of Public 22 Assistance After Leaving Welfare Table 4: Recipients’ Comparisons of Deprivations While on and 25 After Being on Welfare Table I.1: Response Rates for State Studies 35 Table III.1: Selected Information on Grants Awarded 41 Figures Figure 1: Categories of Information Reported in State Studies of 8 Former Welfare Families Figure 2: States That Have Reported Data on Families Who Have 15 Left Welfare Figure III.1: Administrative Data Sources Grantees Planned to 43 Use Abbreviations AFDC Aid to Families With Dependent Children HHS Department of Health and Human Services PRWORA Personal Responsibility and Work Opportunity Reconciliation Act of 1996 TANF Temporary Assistance for Needy Families Page 33 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix I Scope and Methodology This appendix provides more detail on how we (1) identified state studies of families who had left welfare, (2) assessed the extent to which the studies could support statewide generalizations of results, and (3) summarized the findings on employment and income from the studies with results generalizable to each state’s welfare population. To obtain information to answer this request, we searched for state studies Identifying and of families who left welfare during or after 1995 that had published results Assessing the Studies by September 30, 1998. We began with a list of state tracking efforts prepared by a joint effort of the staffs of the National Conference of State Legislatures, the National Governors’ Association, and the American Public Human Services Association.30 In addition, we talked to representatives of the 10 states with the largest welfare caseloads (California, Florida, Illinois, Michigan, New Jersey, New York, Ohio, Pennsylvania, Texas, and Washington) and asked if they had any studies of former welfare families for which results had been published. We also talked to welfare experts and asked them if they knew of any ongoing studies of former welfare recipients. Finally, we talked to representatives of those states that we identified as having published reports on the basis of their tracking efforts and asked what additional plans, if any, they had for tracking former welfare families and if they had updated their work. If a state had more recent information available, we included it in our analysis when possible, in some cases using a report published after September 30, 1998. Through this process, we identified 18 separate tracking efforts in 17 states.31 The 18 studies varied in degree of data completeness and statewide generalizability. We were interested in summarizing results that could reasonably be generalized to most families who left welfare in the state at the time of the study. We considered a study to be of an acceptable level of statewide generalizability if the study successfully obtained data on at least 70 percent of the sample of families for which it sought such data, or if a nonresponse analysis of the data showed that were no important differences between families represented in the data and those 30 This list appears in a document called “State Efforts to Track and Follow Up on Welfare Recipients” and is updated periodically as additional studies are identified. The list we used was dated July 30, 1998. For the most recent list of state efforts to track welfare recipients, see the Web site www.aphsa.org. 31 Wisconsin had two studies. The second surveyed a cohort of welfare recipients that left the welfare rolls between Jan. and Mar. 1998. The results of this survey were issued on Jan. 13, 1999. Page 34 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix I Scope and Methodology missing from the data.32 Except for this assessment, we did not independently verify the data included in the state studies. Using this assessment, we identified eight studies representing seven states. The proportion of families that responded to surveys or for whom data were located in administrative databases in each study is shown in table I.1. Table I.1: Response Rates for State Studies Number of Response rate State Population of interest respondentsa (percentage) Idaho Families who left TANF July to Dec. 1997 447 17 Indiana Families receiving AFDC May 1995 to May 1996 who subsequently left AFDC 1,589 71 Iowa AFDC families assigned to or who volunteered for the Limited Benefit Plan Nov. 1994 to Apr. 1995 137 85 Kentucky Families who left TANF Jan. to Nov. 1997 560 17 Louisiana Families in metropolitan New Orleans who left TANF Jan. to Mar. 1998 349 17 Maryland Families who left TANF Oct. 1996 to Sept. 1997 2,156 82b Michigan Families whose AFDC benefits were terminated in Apr. 1996 because they did not comply with program rules 67 53 Montana Families who received or left AFDC or TANF Mar. c 1996 to Sept. 1997 208 New Jersey Families whose TANF benefits were terminated Jan. to Feb. 1998 because of failure to comply with program rules 453 45 New Mexico Families who left AFDC July 1996 to June 1997 617 12 Oklahoma Families who left or were denied TANF Oct. 1996 to Nov. 1997 292 53 (continued) 32 For other purposes, a different criterion may be more appropriate. For example, some program administrators believe that a major effort should be made to locate all families in the sample, especially if the purpose of a study is to ensure that families have not been adversely affected by program changes. Page 35 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix I Scope and Methodology Number of Response rate State Population of interest respondentsa (percentage) Pennsylvania Families who left TANF Mar. 1997 to Jan. 1998 169 47 d South Carolina Families with a household member required to seek employment who left TANF July to Sept. 1997 and had not returned at time of follow-up 403 76 Tennessee Families who lost TANF benefits Jan. to Oct. 1997 because they did not comply with program rules, and TANF families whose head was employed full- or part-time Feb. to Oct. 1997 2,436 51 Washingtond Single-parent families who left TANF Apr. to July 1998 592 52 Wisconsin Single, female-headed families who left AFDC July 1995 to July 1996 54,518 92b Wisconsin Families who left AFDC/TANF Jan. to Mar. 1998 and did not return prior to survey 375 69 Wyoming Families who left TANF Dec. 1996 to Feb. 1998 200 32 a These are the families actually located for whom data are reported in the study. b Represents the percentage of families about whom information was located in program administrative data. c Data not available. d Both South Carolina and Washington reported on groups of families who left welfare earlier. We included the most recent sample in our summary. Source: GAO analysis of state studies. We summarized the results for the eight studies that we considered could Summarizing Data reasonably be generalized to the state level in the three major areas of Related to interest: economic status, family composition, and family and child Employment and well-being. This effort was constrained by the different sources of data used by the state studies, different categories of families tracked, and Income different questions asked of respondents in the surveys. The area for which each state study had somewhat comparable data was economic Page 36 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix I Scope and Methodology status. The state studies with surveys generally asked employment status at the time of follow-up, hourly wage rate, number of hours worked, and whether the respondent had ever worked since leaving welfare. The two studies relying on state unemployment insurance program data to track employment could report only whether an individual had worked at some point during a 3-month period and total earnings for the period. Further, since state unemployment insurance programs do not cover some categories of employed individuals, program data would not have information on these individuals. For example, self-employed individuals and certain agricultural workers are generally not covered. The two studies reporting average quarterly earnings based on state unemployment insurance program data did not report average hourly wage or average number of hours worked. In order to make earnings data comparable among the eight studies, we estimated average quarterly earnings for the other studies for which we had, or could calculate, average hourly wage rates and average number of hours worked. We multiplied the average hourly wage in each of these studies by the study’s reported average number of hours worked in a week and multiplied by 13 to estimate a quarterly wage. This enabled us to compare estimated quarterly earnings with the reported quarterly earnings. We also had to make some adjustment to ensure that the employment rates were for comparable categories of families. Although some of the studies reported employment rates only for adults who left welfare and were still off the rolls at the time of follow-up, others included all families who left the rolls during the study period—even those who had returned to welfare at the time of follow-up. To estimate comparable employment rates, we removed from the calculation data on the individuals who returned to the rolls and assumed that those who returned to the rolls were not employed. Page 37 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix II Reports From States’ Studies of Families Who Left Welfare Project Self-Reliance TAFI Participant Closure Study (II), Idaho Idaho Department of Health and Welfare, spring 1998. The Indiana Welfare Reform Evaluation: Assessing Program Indiana Implementation and Early Impacts on Cash Assistance, Abt Associates, Inc., Aug. 1997. The Indiana Welfare Reform Evaluation: Who Is On and Who Is Off? Comparing Characteristics and Outcomes for Current and Former TANF Recipients, Abt Associates, Inc., Sept. 1997. The Indiana Welfare Reform Evaluation: Program Implementation and Economic Impacts After Two Years, Abt Associates, Inc., and The Urban Institute, Nov. 1998. Iowa’s Limited Benefit Plan: Summary Report, Mathematica Policy Iowa Research, Inc., and the Institute for Social and Economic Development, May 1997. A Study of Well-Being Visits to Families on Iowa’s Limited Benefit Plan, Mathematica Policy Research, Inc., June 1998. From Welfare to Work: Welfare Reform in Kentucky, Welfare Reform Kentucky Evaluation No. 1, Center for Policy Research and Evaluation, Urban Studies Institute, University of Louisville, Jan. 1998. Exiting Welfare: The Experiences of Families in Metro New Orleans, Louisiana School of Social Work, Southern University at New Orleans, June 1998. Life After Welfare: An Interim Report, University of Maryland School of Maryland Social Work, Sept. 1997. Life After Welfare: Second Interim Report, University of Maryland School of Social Work, Mar. 1998. Page 38 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix II Reports From States’ Studies of Families Who Left Welfare A Study of AFDC Case Closures Due to JOBS Sanctions April 1996, Michigan Michigan Family Independence Agency, May 1997. Montana’s Welfare Reform Project: Families Achieving Independence in Montana Montana FAIM, February 1998 Update, Montana Department of Public Health & Human Services, Feb. 12, 1998. WFNJ (TANF) Sanction Survey, New Jersey Department of Human Services, New Jersey July 2, 1998. Survey of the New Mexico Closed-Case AFDC Recipients July 1996 to New Mexico June 1997, Final Report, University of New Mexico, Sept. 1997. Family Health & Well-Being in Oklahoma: An Exploratory Analysis of TANF Oklahoma Cases Closed and Denied October 1996 to November 1997, Oklahoma Department of Human Services, Sept. 1998. TANF Closed-Case Telephone Survey, Pennsylvania Department of Public Pennsylvania Welfare, Mar. 1998. Survey of Former Family Independence Program Clients: Cases Closed South Carolina During January Through March 1997, South Carolina Department of Social Services, Division of Program Quality Assurance, Mar. 3, 1998. Survey of Former Family Independence Program Clients: Cases Closed During July Through September 1997, South Carolina Department of Social Services, Division of Program Quality Assurance, Oct. 9, 1998. Summary of Surveys of Welfare Recipients Employed or Sanctioned for Tennessee Non-Compliance, University of Memphis, Mar. 1998. Washington’s TANF Single Parent Families Shortly After Welfare: Survey of Washington Families Which Exited TANF Between December 7 and March 1998, Washington DSHS Economic Services Administration, July 1998. Page 39 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix II Reports From States’ Studies of Families Who Left Welfare Washington’s TANF Single Parent Families After Welfare, Washington DSHS Economic Services Administration, Jan. 1999. Post-Exit Earnings and Benefit Receipt Among Those Who Left AFDC in Wisconsin Wisconsin, Institute for Research on Poverty, University of Wisconsin-Madison, Aug. 17, 1998. Post-Exit Earnings and Benefit Receipt Among Those Who Left AFDC in Wisconsin, Institute for Research on Poverty, University of Wisconsin-Madison, Oct. 30, 1998. Survey of Those Leaving AFDC or W-2 January to March 1998, Preliminary Report, State of Wisconsin, Department of Workforce Development, Jan. 13, 1999. A Survey of Former POWER Recipients (Personal Opportunities With Wyoming Employment Responsibilities), Western Management Services, LLC, for Wyoming Department of Family Services, May 1998. Page 40 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix III Information on Selected Grants Awarded by the Department of Health and Human Services to Study Families Leaving Welfare This appendix provides information on the study methodologies planned by the jurisdictions receiving grants from HHS to study families who have left welfare. Table III.1: Selected Information on Grants Awarded Timing of survey Award (months after Grantee amount Population of interest exit) Arizona $249,824 Families leaving welfare 12 Oct. to Dec. 1996 and Jan. to Mar. 1998 California—Los Angeles 250,000 Families leaving welfare 12 County Oct. to Dec. 1996 and Jan. to Mar. 1998 California—San Mateo, 160,270 Families leaving or diverted 6 and 12 Santa Clara, and Santa from welfare Oct. to Dec. Cruz counties 1996 and July to Dec. 1998 District of Columbia 249,749 Families leaving welfare 6 July to Sept. 1997 and Oct. to Dec. 1998 Florida 274,719 Families leaving or diverted 18 to 20 from welfare Apr. to June 1997; families receiving food stamps or Medicaid who were eligible for, but not receiving, welfare Apr. to June 1997 Georgia 246,660 Families leaving welfare 6 Jan. to Oct. 1997 and July 1998 to June 2001 Illinois 250,000 Families leaving welfare 1 to 9 Oct. 1997 to June 1998 and Jan. to Mar. 1999 Massachusetts 206,294 Families leaving welfare 3, 6, 9, and 12 Jan. to June 1997 and Dec. 1998 to Feb. 1999 Missouri 250,000 Families leaving welfare 15 Oct. to Dec. 1996 and Oct. to Dec. 1997 New York 80,476 Families leaving welfare 12 Nov. 1997 to Mar. 1998 and Jan. to Mar. 1999 Ohio—Cuyahoga County 250,000 Families leaving welfare 12 Oct. to Dec. 1996 and Jan. to Mar. 1998 South Carolina 200,000 Families leaving welfare 6 and 9 Jan. to Mar 1997 and Jan. to Mar. 1999 (continued) Page 41 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix III Information on Selected Grants Awarded by the Department of Health and Human Services to Study Families Leaving Welfare Timing of survey Award (months after Grantee amount Population of interest exit) Washington 244,965 Families leaving welfare 6 to 9 Oct. to Dec. 1996, Oct. to Dec. 1997, and Oct. to Dec. 1998 Wisconsin 204,200 Families leaving welfare 5 to 10 Oct. to Dec. 1996 and Jan. to Dec. 1998; families applying for welfare in Milwaukee Oct. 1998 to Mar. 1999 Source: Office of the Assistant Secretary of Planning and Evaluation, HHS. Page 42 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix III Information on Selected Grants Awarded by the Department of Health and Human Services to Study Families Leaving Welfare Figure III.1: Administrative Data Sources Grantees Planned to Use a This category includes other data sources, such as sources of tax, welfare-to-work, and health information. Page 43 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix IV Comments From the Department of Health and Human Services Page 44 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix IV Comments From the Department of Health and Human Services Page 45 GAO/HEHS-99-48 Status of Former Welfare Recipients Appendix V Major Contributors to This Report Gale C. Harris, Assistant Director, (202) 512-7235 Margaret Boeckmann, Evaluator-in-Charge, (202) 512-6992 Emily Loriso Regina Santucci Jay Smale Page 46 GAO/HEHS-99-48 Status of Former Welfare Recipients Page 47 GAO/HEHS-99-48 Status of Former Welfare Recipients Related GAO Products Welfare Reform: States’ Experiences in Providing Employment Assistance to TANF Clients (GAO/HEHS-99-22, Feb. 26, 1999). Domestic Violence: Prevalence and Implications for Employment Among Welfare Recipients (GAO/HEHS-99-12, Nov. 24, 1998). Welfare Reform: Early Fiscal Effects of the TANF Block Grant (GAO/AIMD-98-137, Aug. 18, 1998). Welfare Reform: Child Support an Uncertain Income Supplement for Families Leaving Welfare (GAO/HEHS-98-168, Aug. 3, 1998). Welfare Reform: Many States Continue Some Federal or State Benefits for Immigrants (GAO/HEHS-98-132, July 31, 1998). Welfare Reform: Changes Will Further Shape the Roles of Housing Agencies and HUD (GAO/RCED-98-148, June 25, 1998). Welfare Reform: States Are Restructuring Programs to Reduce Welfare Dependence (GAO/HEHS-98-109, June 18, 1998). Welfare Reform: Transportation’s Role in Moving From Welfare to Work (GAO/RCED-98-161, May 29, 1998). Medicaid: Early Implications of Welfare Reform for Beneficiaries and States (GAO/HEHS-98-62, Feb. 24, 1998). Welfare Reform: States’ Efforts to Expand Child Care Programs (GAO/HEHS-98-97, Jan. 13, 1998). Welfare Reform: States’ Early Experiences With Benefit Termination (GAO/HEHS-97-74, May 15, 1997). Welfare Waivers Implementation: States Work to Change Welfare Culture, Community Involvement, and Service Delivery (GAO/HEHS-96-105, July 2, 1996). (116019) Page 48 GAO/HEHS-99-48 Status of Former Welfare Recipients Ordering Information The first copy of each GAO report and testimony is free. Additional copies are $2 each. Orders should be sent to the following address, accompanied by a check or money order made out to the Superintendent of Documents, when necessary. VISA and MasterCard credit cards are accepted, also. Orders for 100 or more copies to be mailed to a single address are discounted 25 percent. Orders by mail: U.S. General Accounting Office P.O. Box 37050 Washington, DC 20013 or visit: Room 1100 700 4th St. NW (corner of 4th and G Sts. NW) U.S. General Accounting Office Washington, DC Orders may also be placed by calling (202) 512-6000 or by using fax number (202) 512-6061, or TDD (202) 512-2537. Each day, GAO issues a list of newly available reports and testimony. To receive facsimile copies of the daily list or any list from the past 30 days, please call (202) 512-6000 using a touchtone phone. A recorded menu will provide information on how to obtain these lists. For information on how to access GAO reports on the INTERNET, send an e-mail message with "info" in the body to: firstname.lastname@example.org or visit GAO’s World Wide Web Home Page at: http://www.gao.gov PRINTED ON RECYCLED PAPER United States Bulk Rate General Accounting Office Postage & Fees Paid Washington, D.C. 20548-0001 GAO Permit No. G100 Official Business Penalty for Private Use $300 Address Correction Requested
Welfare Reform: Information on Former Recipients' Status
Published by the Government Accountability Office on 1999-04-28.
Below is a raw (and likely hideous) rendition of the original report. (PDF)