Status Report on Research on the Outcomes of Welfare Reform, 2002

I . Introduction and Overview

Background

In its report for the FY 2002 Appropriation for the Department of Health and Human Services, the Conference Committee stipulated for the fifth year that research within the Office of the Secretary should be dedicated to study the outcomes of welfare reform:

"Within the funds available, $7,125,000 is to continue to study of the outcomes of welfare reform and to assess the impacts of policy changes on the low-income population. The conferees recommend that this effort include the collection and use of state-specific surveys, state and federal administration data, and data administratively linking the National Directory of New Hires, other child support enforcement data, TANF and Medicaid records together. These studies should focus on assessing the well-being of the low income population, developing and reporting reliable and comparable state-by-state measures of family hardship and well-being, the utilization of other support programs and the impact of child support enforcement efforts. These studies should continue to measure outcomes for a broad population of current, former and potential welfare recipients, as well as other special populations affected by state TANF policies. The conferees further expect these studies to analyze how the earnings of custodial and non-custodial parents who are, or have had children who are, current or former welfare recipients have changed over time and whether the pattern is significantly different among states. The conferees request a report on these topics to be submitted to the House and Senate Appropriations Committees by May 1, 2002." (H. Rept. 107-342, pages 115-116)

The following report has been prepared by the Office of the Secretary, Office of the Assistant Secretary for Planning and Evaluation (ASPE), in response to the requirement for a report. Unless otherwise noted, this report discusses only the welfare outcomes research agenda supported by the targeted funding. No attempt has been made to reflect the separately funded welfare research agenda of the Department's Administration for Children and Families (ACF) ⁽¹⁾ or ASPE's or the Department's health research agenda, except to the extent that some projects were supported jointly by welfare outcomes funding and funds from other sources.

The Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996 was a landmark event in our nation's welfare policy. The results of the reforms, based on a number of popular measures, have been dramatic. The number of families receiving Temporary Assistance for Needy Families (TANF) program cash benefits has declined substantially, marking the first-ever rapid caseload decline during an expanding economy. Employment among current and former welfare recipients has increased significantly, with the number of working recipients reaching all-time highs in fiscal years 1999 and 2000. In addition, the majority of mothers leaving welfare are employed after leaving. Earnings for welfare recipients continuing to receive cash assistance, and earnings for female-headed households in general, also have increased significantly. In addition, the child poverty rate continued to decline between 1996 and 2000, falling to its lowest rate in over 20 years.

Despite these gains, there is much to be done as we move to the next phase of welfare reform. States have had mixed success in fully engaging welfare recipients in work activities. While all states have met the overall work participation rates required by law, in 2000, in an average month, only about one-third of all families with an adult participated in work activities that were countable toward the state's participation rate. Substantial progress has been made nationwide in reducing teen births, but the proportion of births occurring outside of marriage remains relatively stable. Child poverty rates for African American and Hispanic children have also fallen dramatically during the past six years, although their poverty rates are still more than three times the rate for white, non-Hispanic children.

In addition, much remains to be learned as we move to the next phase of welfare reform. As a direct result of the dedication of research funds to study the outcomes of welfare reform, ASPE has contributed significantly to the scope, volume, and diversity of welfare reform research within the Department. "Welfare Outcomes" funding has enabled ASPE to sponsor or conduct a great number of studies designed to document trends in the low-income population (including both adults and children) as well as in the welfare recipient population. Our research agenda has also supported efforts to build state data capacity and data comparability through our support of monitoring studies, such as studies of families leaving welfare. Nonetheless, in its final report, the Panel on Data and Methods for Measuring the Effects of Changes in Social Welfare Programs convened by the National Research Council's Committee on National Statistics has identified some important data gaps and offered some conclusions and recommendations designed to build up the "science base" of welfare reform research. These are discussed in more detail in the Future Directions section later in this chapter.

General Strategies for Understanding the Outcomes of Welfare Reform

The dedication of research funds to studying welfare outcomes has enabled the Department to create a research, evaluation and data strategy designed to document the implementation of welfare reform and its effects, and to add to and enhance the information about welfare reform outcomes that is available to the Congress and other interested parties. Overall, our focus has been on creating an integrated picture of the low-income population, especially low-income families with children, combined with broader analyses of the economic condition, health and well-being, socio-demographic characteristics, and the social service needs of low-income individuals, families, and children. We believe this is consistent with both the Conference Committee's directives and with the far-reaching recommendations of the Panel on Data and Methods for Measuring the Effects of Changes in Social Welfare Programs.

There is a broad array of research about welfare reform being funded by the Department and other public and private sources. To optimize the potential that the targeted funds will increase the Department's understanding of the outcomes of welfare reform, ASPE has created, often with other funding partners involved, a portfolio of studies and strategies. Careful attention has been paid to identifying on-going research, evaluation, and data activities which could be enhanced or modified, identifying activities being funded or planned by other entities that might provide joint-funding opportunities, to identifying knowledge gaps, and to avoiding unnecessary duplication. We have used the targeted research funds to fully fund some projects, to fund specific portions of some larger studies, and to co-fund with other public and private agencies yet other projects. As a result, our research, evaluation and data activities cover a wide spectrum of welfare outcomes policy interests.

In keeping with the recommendations of the conferees, our research agenda over the past four years has covered a broad array of topics and approaches that complement other public and private efforts to assess the outcomes of welfare reform. We have funded or co-funded projects to study economic supports for poor families, children and youth, family formation, special populations and local service delivery issues, and cross-cutting topics. We have funded or co-funded competitive grant programs; projects to improve state data collection, comparability or capacity-building and analytic projects on welfare-related topics. We have funded or are funding projects that measure outcomes for welfare leavers, examine diversion practices, study the characteristics of the TANF caseload (or "stayers"), and measure family hardship and well-being including the utilization of other support programs. Projects also are in place to assess the effects of welfare reform on current, former and potential welfare recipients and other special populations (e.g., child-only cases, people with mental health and substance abuse problems and other disabilities, immigrant families) affected by state TANF policies. Our specific activities and plans in each of the areas recommended by the Conference Committee are summarized below.

Despite the breadth and scope of these efforts, from a research perspective our knowledge is still quite limited in many areas, and many factors can limit what research can accomplish. For example, there is wide variation in the design and application of policies across states, between local sites, and even from worker-to-worker. State policies and organizational structures continue to evolve and, in some cases, state responsibilities are further devolved. We know little about low-income families who do not become welfare recipients, and people who leave assistance can be difficult to track over time. Many variables other than welfare policies (such as the economy) affect the outcomes of welfare reform, and these variables often have confounding effects. Because of these factors, the ability of research, evaluation, and data to completely answer questions is always limited.

Future Directions

ASPE's research plan for the targeted welfare outcomes funds for FY 2002 is designed to meet the overall objective of creating an integrated picture of the low-income population, especially low-income families with children, as we move to the next phase of welfare reform. We continue our focus on broader analyses of the economic conditions, health and well-being, socio-demographic characteristics, and the social service needs of low-income individuals, families, and children. Our interest is to cover a wide spectrum of policy interests focusing on welfare outcomes, poverty, working families, supports for low-income populations, the hard-to-serve and other special populations, and effects on children. We envision a research agenda that:

• addresses a wide range of topics related to families and children, including economic and other supports for poor families,
• continues some of our earlier activities to promote, develop, and support state and local capacity for data collection and monitoring studies, by supporting state-level data collection efforts, administrative data linking, and the creation of public-use and restricted-access data files, and
• facilitates states' monitoring of outcomes for their own state and local populations. These activities include projects to build state data collection and research capacities, providing technical assistance to improve the quality of research results, ensuring more uniformity and comparability across studies, and synthesizing results across state and local level monitoring studies.

Implementing a broad-based research agenda that addresses this wide range of welfare outcomes research is one of ASPE's most important roles within the Department.

As in previous years, the welfare research, evaluation, and data projects planned for FY 2002 are designed to complement and enhance other efforts, both within and outside the federal government, to assess and monitor welfare outcomes without undue duplication. They include projects that address several of the Administration's priority themes, including encouraging work and self-sufficiency, promoting healthy marriage and strong families, improving child well-being, expanding state flexibility and accountability, expanding the use of faith-based and community-based organizations, and addressing the needs of special populations. Our planned projects also address the Conference Committee's research priorities for use of welfare outcomes funding. We are hopeful that these wide-ranging activities will not only increase our understanding of the outcomes of low-income families, but also contribute to the Department's ability to respond to questions about those outcomes. Specific projects planned for FY 2002 are described in Chapter III.

More broadly, we also continue to evaluate National Academy of Sciences recommendations to enhance the quality of welfare outcomes research. Conference report language accompanying the Fiscal Year 1998 and 1999 appropriations for welfare outcomes research funding included the recommendation that the Department "submit its research plan to the National Academy of Sciences (NAS) to provide further guidance on research design and recommend further research." Accordingly, we provided a total of over $1 million to the NAS in Fiscal Years 1998, 1999, 2000 and 2001 to convene an expert panel to evaluate current and future welfare reform research.

The aforementioned Panel on Data and Methods for Measuring the Effects of Changes in Social Welfare Programs' released an interim report, Evaluating Welfare Reform: A Framework and Review of Current Work, Interim Report, in September 1999. It provided a framework for conducting evaluations of welfare program changes, reviewed current Departmental efforts to evaluate these changes, and provided the Panel's initial conclusions and recommendations. The short-run recommendations for welfare evaluation strategies were consistent with the Department's approach to research on welfare reform, and many of the recommended steps were already being taken.

The Panel's final report, Evaluating Welfare Reform in an Era of Transition ⁽²⁾ , released in April 2001, highlighted some additional factors that need to be considered as the future direction of welfare research is contemplated. The report identified some important data gaps and offered numerous conclusions and recommendations with respect to defining research questions and outcomes of interest for measuring the effects of welfare reform, as well as the appropriate methods for answering those questions and the data needed to carry out these evaluations. It discussed research designs and methods for the study of welfare reform outcomes, and needed areas and topics of research. The report also addressed alternative federal and state data sources, the limitations of currently available data, and appropriate evaluation design and methods for analysis. In short, while applauding the Department for its broad-based welfare reform research agenda, it highlighted the need for further improvements and expansions in data collection, development of research questions, and methodological work to build up the "science base" of welfare reform research.

The Department has taken, or is taking, steps to address several of the Panel's recommendations. For example, our efforts to build capacity for conducting high-quality program evaluations at the state level and for conducting household surveys of low-income and welfare populations continue. We have committed resources to help improve national household survey questions to better measure program participation and benefit receipt. A project to improve the usefulness of state-level administrative data is planned. Improvements in state data reporting are included in the Administration's welfare reform proposal, as are changes to broaden the definition of assistance. State-specific data sets produced by each of the grantees studying welfare leavers have been made available for secondary data analyses of welfare outcomes measures. In addition, we have published a synthesis report that includes administrative data findings from all 15 of the ASPE-funded leavers studies. Study and consideration of other Panel conclusions and recommendations will continue.

Activities that Support Conferees' Recommendations

As we have noted, our overarching research objective in using targeted welfare outcomes funds has been and is to create an integrated picture of the low-income population, especially low-income families with children. We have devoted substantial resources to covering a wide spectrum of policy interests focusing on welfare outcomes, poverty, working families, supports for low-income populations, the hard-to-serve and other special populations, and effects on children. Welfare devolution and increased flexibility in the design and delivery of program benefits have created substantial new challenges for data collection and analysis to monitor welfare outcomes. To meet these challenges, new and better data are needed at the state and local level. Activities that support and enhance the collection, use, and linking of federal and state administrative data and state-specific surveys are critical to understanding the outcomes of welfare reform. Some examples of welfare outcomes-funded or proposed studies that support the research recommendations of the Conference Committee follow.

• Focus broadly on the low-income population, including current, former and potential welfare recipients, and other special populations affected by state TANF policies

We have funded grants to states and large counties to gather a variety of information about individuals and their families who apply to TANF (including those who are formally or informally diverted) and the degree to which TANF applicants receive, or are aware of their potential eligibility for, Medicaid, food stamps, and other programs and services that are important in helping low-income families make a successful transition to work. ⁽³⁾ We are funding projects in six states to study the characteristics of their TANF caseloads, with particular attention to the personal, family and community factors that may present barriers to employment. Employment programs for low-income parents with barriers to work will be tested under the long-term, multi-site demonstration and evaluation of programs for hard-to-employ low-income parents in order to identify effective strategies for promoting employment and family well-being and to determine the effects of such programs on employment, earnings, income, welfare dependence, family functioning, and the well-being of children. We also are supporting a National Governors Association project to build state and local capacity to provide work supports which help low-income working parents sustain employment and advance in the labor market. A separate project is assessing the coping mechanisms used by low-income families, by examining factors that may help or hinder a family's efforts to be self-sufficient, including formal and informal support services, social support networks, and time and money management.

In addition, we are supporting a supplement to a broader study of the implications of welfare reform for low-income families in three cities that is examining how work participation requirements and time limits are affecting the service utilization, health and development, support networks, parenting and child care arrangements of adults and children with disabilities. We also are providing continuing support for an evaluation of the effectiveness of a substance abuse research demonstration project that includes coordinating screening, referral and treatment with employment and training or vocational services. Another project is examining trends in the demand for emergency assistance services, such as homeless shelters and food banks, from the mid-1990's to 2000. In recognition of the fact that the highest concentration of populations on TANF are living in neighborhoods where incarceration rates are also the highest, we funded a project that is investigating the effects of incarceration on low-income children, families, and communities.

Our researcher-initiated grants program spanning Fiscal Years 1999, 2000 and 2001 funded a variety of proposals looking broadly at low-income individuals (both adults and children) and their families. For example, one project is analyzing three groups of low-income populations - those who leave welfare for work, those who remain on welfare, and non-working welfare leavers - to examine the characteristics and circumstances of people who leave welfare and are not working and how welfare "stayers" are different from the "leavers." We also funded a proposal to examine the determinants of successful employment and wage outcomes among low-income workers. A separate project is looking at the employment rates, earnings, income and other measures of well-being among low-income families who have not applied for welfare or who have applied and been denied or diverted. Another project is examining how the lives of adolescents in single-mother, low-income families who received welfare during the early years of TANF are affected by their mothers' employment, and whether these changes in adolescents' lives are specific to families receiving welfare or common among single-mother, low-income families more generally. We also funded a project documenting the incidence of child disability among low-income and welfare-recipient, single-mother families and the impact that children with disabilities have on their families' ability to exit welfare and poverty. Yet another project is examining the conditions and capabilities of vulnerable mothers and fathers in the first few years following enactment of PRWORA to evaluate the impact of TANF and child support policies.

Our FY 2002 research plan includes several proposed projects that focus broadly on the low-income population. For example, one study would focus on families who exit and re-enter welfare, particularly those who return for multiple times, in an attempt to identify the demographic characteristics and employment outcomes of "cyclers," whether their patterns of benefit receipt and cycling patterns have changed since PRWORA, and whether cyclers re-entering welfare for the third time are treated differently than other families by local offices. Another planned project would explore the coordination of TANF and one-stop employment centers and the unique challenges involved in serving welfare and ex-welfare clients, and the low-skilled in general, through one-stop centers designed to serve people at all income and skill levels. A separate planned project would attempt to better understand the different types of child-only cases, the service needs of these families, and how states are meeting those needs. These and other projects will help us achieve our goal of covering all important low-income population groups in our research programs, including current, former and potential welfare recipients.

• Collect and use state-specific surveys, state and federal administrative data and administratively-linked data to develop and report state-by-state measures of family hardship and well-being and the utilization of other support programs

ASPE has devoted substantial resources to supporting state-level data collection efforts and administrative data linking, making certain that national survey instruments are responsive to policy changes and needs, supporting and maintaining a wide range of Census Bureau data collection efforts, and supporting secondary analyses of state and national-level data to add to our understanding of the effects of welfare reform. We have also committed funds for several projects geared toward developing state-level data on hardship and/or program utilization.

One of our researcher-initiated grant projects is using state and county administrative data to study the demographics, welfare participation, employment retention, and post-exit earnings of five cohorts of welfare recipients in North Carolina, to examine welfare exits, employment stability, earnings mobility, recidivism, longer-term labor market outcomes, types of jobs recipients obtain and the range of wages for these jobs. Another is exploring the long-term utilization of food stamp and Medicaid benefits for two cohorts of welfare recipients who left cash assistance in Wisconsin. A separate study is examining the accuracy of self-reports of program participation in survey data by comparing self-reported program participation among Californians interviewed in the national Survey of Income and Program Participation (SIPP) with California administrative files of program participation. The New Jersey Department of Human Services is examining the extent to which former welfare recipients who leave welfare for work are likely to be eligible for Unemployment Insurance (UI), using a set of administrative and survey data from an ongoing evaluation. A study in Washington is estimating the effects of state-level policies (using a database of policy indicators in 11 program areas, including some transfer, tax, in-kind, and work support policies, over six years) on families' resources, resource packages, and experience of hardship.

We have provided support for an on-going project designed to match Social Security earnings records with samples of adult welfare recipients and non-recipients from Census surveys to help assess employment and earnings patterns and outcomes on the basis of baseline characteristics.

Our support for the administration of a welfare participation question in the State and Local Area Integrated Telephone Survey (SLAITS) has contributed to the fielding of a survey data element that, when combined with other data available from the survey, will permit the development of state-level estimates of the incidence of special health care needs of children of current and former welfare recipients, plus the health insurance status (including Medicaid and the State Children's Health Insurance Program, SCHIP) of current and former recipients. We also have provided funds to support the longitudinal New Immigrant Survey which will collect data on program utilization, health and economic status, and other characteristics among newly arriving low-income immigrant families in different states. ASPE's contribution will help ensure that comprehensive and relevant data are collected and analyzed and that the study focuses on how children in these families are faring. In addition, we have funded a project to advance our understanding of the value and limitations of measures of material hardship as a component of family well-being.

We have funded several projects which are utilizing existing administrative and survey data. For example, we have contributed support to Project on Devolution and Urban Change, a multi-disciplinary study of the implementation and impacts of welfare reform and welfare-to-work programs on low-income individuals, families and communities in four large urban areas, using, among other data sources, longitudinal administrative data for all families receiving AFDC/TANF or food stamps dating back to 1992. Under a separate task order we are analyzing survey data on the personal characteristics, potential barriers to employment, strengths and resources and preparation for employment of current TANF recipients in Illinois to explore the relationships of various factors contributing to employment outcomes. We also are supporting a small grant program for young scholars to encourage secondary analyses of new HHS-sponsored administrative and survey data sets to explore aspects of welfare reform. Finally, the competitive grants we awarded to states and large counties to examine welfare leavers, individuals and their families who apply to TANF (including those who are formally or informally diverted), and the characteristics of TANF caseloads, used or are using various combinations of state and federal administrative and survey data to report on a variety of welfare reform outcomes.

A project building on past ASPE-funded studies of welfare leavers and welfare applicants is proposed for FY 2002. This study would focus on the subgroup of TANF families who exit and re-enter welfare, particularly those who return for multiple times. Another proposed project would utilize five federally administered databases to provide state-by-state estimates of important welfare outcomes. Projects such as these support the overall goal of ensuring that good data are available to create an integrated picture of the low income population, especially families with children, in the wake of welfare reform.

• Include efforts to link the National Directory of New Hires, other child support enforcement data, TANF and Medicaid records together to assess the impact of child support enforcement efforts and analyze how the earnings of custodial and non-custodial parents have changed over time and whether the pattern is significantly different among states.

The Federal Parent Locator Service (FPLS), operated by ACF's Office of Child Support Enforcement (OCSE) is primarily a national system to help states locate non-custodial parents, alleged fathers, and custodial parents so they can establish and enforce child support obligations. The Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) expanded the data available in this system to include information on child support cases and wage and employment data, including the National Directory of New Hires.

PRWORA also authorized HHS to retain samples of this data and use them for research purposes likely to contribute to achieving the purposes of titles IV-A (Temporary Assistance for Needy Families) or IV-D (Child Support Enforcement and Paternity Establishment). The Department has awarded a contract to implement an FPLS research database, in conjunction with TANF and Medicaid data, that will provide information about the effects of welfare reform on TANF, Medicaid and child support participants, and help identify the significance of program interactions in increasing self-sufficiency among low-income families. The analysis from the database also will assist federal and state program oversight, research, statistical reporting, and the policy and evaluation purposes related to the TANF and child support enforcement programs and the low-income populations they serve. Currently these databases are not linked. While some limited cross-sectional matches are currently possible between TANF and the child support enforcement databases, there is no mechanism for creating a longitudinal database that would permit analysis of client and program interactions over time.

ASPE is working with OCSE, ACF's Office of Planning, Research and Evaluation (OPRE), and the Centers for Medicare & Medicaid Services' Office of Strategic Planning to develop this project, which builds on a feasibility study undertaken by ASPE and ACF in FY 1998 to examine research needs and database design issues. Departmental privacy, legal, and technical staff will be consulted, as well as appropriate outside experts. Assuming all the legal and technical issues can be resolved, this phase of the database project is expected to be completed by the end of 2003.

In its current form the expanded FPLS comprises one of the most comprehensive sources of employment information available anywhere, but does not include key program participation and demographic variables. Combined with other administrative data, such as the Medicaid Statistical Information System (MSIS), State Medicaid Research File (SMRF) and the TANF databases, it offers the opportunity to answer questions about the programs and populations of interest that previously we have been unable to answer. Such data matching could allow us to compare the TANF and/or Medicaid populations with similarly situated non-TANF and/or non-Medicaid families, compare families within and outside the child support enforcement system, and examine interactions between the three programs.

Using these databases, cross-sectional analysis will provide very important point-in-time estimations of the current IV-D, IV-A, and Medicaid policy environments and a snap-shot of the characteristics of various populations of interest to HHS. As noted above, some point-in-time analysis can be done with relatively simple cross-matches between existing data files. For example, FPLS data have been matched to some of the "welfare leavers" studies to provide data on employment and earnings. However, the planned longitudinal database would allow us to better address policy questions such as:

• What proportion of custodial parents and children are receiving benefits from other public assistance programs (e.g., TANF, Medicaid) over time? What is the relationship between IV-D enrollment and Medicaid and TANF receipt over time?
• Are TANF applicants and leavers enrolled in Medicaid over time? What are the characteristics of those enrolled and not enrolled?
• What proportion of TANF leavers are employed in the period immediately after they are identified as a "leaver?" How many have periods of unemployment and do they receive unemployment insurance? What are the characteristics of TANF leavers who become unemployed? What are the characteristics of TANF leavers who have frequent job changes?
• Does receipt of Medicaid and child support decrease the proportion of TANF leavers who return to TANF, or increase the length of time between TANF spells?
• Do TANF and Medicaid applicants, recipients, and leavers have child support orders? What are the characteristics of those with and without orders?
• Does receipt of Medicaid and/or TANF affect the amount of time it takes to move a IV-D case through the paternity and award establishment processes?
• What are the earnings patterns of non-custodial parents? What are the earnings patterns of custodial parents? How do these patterns affect program participation and collection of support?
• Are there different interactions across programs in individual states, and how do these different interactions relate to state policies and characteristics?

In addition to the linked micro-level administrative data, state macro-level policy, program and economic variable files also will be developed for use in conjunction with the database. These variables would include such things as annual state TANF benefit levels, Medicaid service packages, unemployment and not-in-the-labor-force rates, population poverty levels, program investments, paternity establishment rates and utilization of child support enforcement techniques. Such variables will allow us to examine state-by-state policy effects.

Our FY 2002 proposed research agenda includes a joint ASPE-ACF-CMS project to support the FPLS, TANF and Medicaid data warehouse to meet the on-going information needs of all the participating agencies. Another proposed project will link several federally administered databases (including TANF closed cases, the NDNH and quarterly earnings database) to provide state-by-state estimates of welfare outcomes for former TANF recipients. We also plan a project to examine how child support status (i.e., paternity established, order established, receipt of child support) interacts with TANF exit or re-entry, and coordination between the TANF and child support agencies. We anticipate that some questions could be answered through state administrative data matching, and others through site visits to states that demonstrate promising practices.

The Department is committed to advancing its welfare outcomes research agenda. The research, evaluation and data projects funded by the targeted Policy Research funds are critical to understanding the outcomes of welfare reform, and crucial to the Department's ability to respond to questions about those outcomes. The projects funded in FYs 1998, 1999, 2000 and 2001 cover a broad array of topics. We continue to provide leadership in national-level survey work and are working to facilitate greater comparability in state and local level studies and increase state and local capacity for data collection efforts.

Organization of the Report

Chapter II summarizes the results and findings we have received from projects funded in FYs 1998, 1999, 2000 and 2001 that have not been reported in previous reports. Chapter III provides an overview of our proposed FY 2002 research plan for welfare-outcomes funded projects. Chapter IV describes the current status and target completion dates of continuing projects funded by the targeted policy research appropriations in FYs 1998, 1999, 2000 and 2001. A synthesis of findings from the ASPE-funded grants to states and localities to study welfare applicants/diversion is included as Appendix A.

Endnotes

1.  Information on ACF's welfare related activities can be found at <www.acf.dhhs.gov/news/welfare/index.htm>. Information on ACF's welfare research activities can be found at <http://www.acf.dhhs.gov/programs/opre/rd&e.htm>.

2.  Copies of the Panel's interim and final reports, as well as its collection of methodological papers, Studies of Welfare Populations, Data Collection and Research Issues, are available through the National Academy of Sciences' website at <www.nas.edu/nap-cgi?Search=evaluating+welfare+reform>.

3.  See Appendix A for a synthesis of early findings from grants to states and localities to study welfare reform outcomes, with an emphasis on TANF applicants and diversion.

II. Findings/Results from Outcomes-Funded Projects

Findings from Completed Welfare Outcomes-Funded Studies

Synthesis of Welfare Outcomes Grants (2000) ⁽¹⁾

Final reports from most of the FY 1998 State Welfare Outcomes grantees have been released and research data sets are becoming available. Under contract with ASPE, the Urban Institute prepared a synthesis report that includes administrative data findings from all 15 of the ASPE-funded leavers studies. The report also includes more detailed findings based on information gathered through follow-up surveys of samples of former recipients from 12 of the 15 grantees. In addition, the researchers at the Urban Institute conducted secondary data analyses of welfare outcomes measures, drawing on the state-specific data sets produced by each of the grantees secured under the Technical Assistance on Researcher Access to Data Sets project. The final synthesis report, completed in December 2001, builds on these secondary data analyses of welfare outcomes measures and information in the grantees' written reports. The report, Final Synthesis Report of Findings from ASPE "Leavers" Grants, is available on ASPE's website at <http://aspe.hhs.gov/hsp/leavers99/synthesis02/>, under "Leavers and Diversion Studies." It includes information on welfare leavers' employment and earnings, public assistance program participation, income and poverty status, material hardships, and child well-being.

Some highlights of the report include:

Employment and Earnings

Encouraging families to move off welfare and into jobs is a goal of welfare reform. All fifteen studies collected some information about employment rates and earnings, wages, or employer-related benefits of families that left welfare. The major findings in this area across these studies are:

• About three-quarters of all leavers work at some point in the year after exiting TANF, on average, and about three out of five work at any given point in time. A little more than a third worked in all four quarters after exiting TANF.
• Mean earnings of welfare leavers are about $2,600 per quarter, according to administrative data. Most studies show a significant increase in quarterly earnings of at least $200 between the first and fourth quarter after exit.
• Working leavers' wages, averaging between $7 and $8 an hour, are comfortably above the federal minimum wage but are nevertheless low. Employed leavers work close to full-time, on average, at least 35 hours per week.
• About half of all working leavers are offered employer-sponsored health insurance through their jobs, but only about one-third actually have coverage. Some leavers receive other employer-sponsored benefits. In general, no more than half have paid sick leave or pension coverage. Paid vacations days are a bit more common.
• No single barrier to work consistently affects a majority of leavers; however, a substantial minority of leavers must overcome child care and health-related problems in order to work.
• Continuous leavers, those who did not return to TANF in the year after exit, are just as likely to have ever worked after exit as those who returned to TANF. However, continuous leavers are somewhat more likely to have worked all four quarters after exit than those who returned. Continuous leavers also have higher earnings than leavers in general.

Program Participation

Non-TANF government assistance can help families in their transition from welfare to work. However, some families return to TANF. The major findings across studies on returns to TANF and participation in other public assistance programs include:

• It is not uncommon for leavers to return to TANF--a quarter to a third of families who left welfare returned to TANF at some point in the first year after exit.
• About half of leaver families receive food stamps in the first quarter after exit and about two-thirds receive these benefits at some point in the year after exit.
• About three out of five leaver families have an adult enrolled in Medicaid in the first quarter after exit. Medicaid coverage of children is generally higher, ranging from 60 to 90 percent after exit.
• The percentage of leavers who receive food stamps and Medicaid at any point over the year after exit is significantly higher than the percentage receiving them in any of the individual quarters, suggesting a great deal of cycling on and off these programs.
• Participation in both food stamps and Medicaid is generally lower for continuous leaver families than those who return to TANF at some point in the year after exit.
• Food stamp and Medicaid program participation is generally higher for those who are not currently employed compared to those currently employed.
• Several studies also report on additional sources of government assistance, such as housing assistance, disability benefits, reduced-price lunches, WIC, fuel/energy assistance, unemployment compensation, and the Earned Income Tax Credit. The range of participation in these programs varies across studies.

Household Income

Household or family income is an important indicator of the well-being of welfare leavers. Although such information is difficult to gather, a subset of studies examine income levels, sources of income, and poverty. Results on income levels, sources of income, and poverty for the subset of studies that examine income are summarized below.

• Average monthly family income for leavers from all sources, including earnings, generally hovers near the poverty line.
• Across all leaver families, own earnings are the most important single source of income, and own earnings plus the earnings of other family members together comprise over three-quarters of leaver families' incomes on average.
• In the four studies that explicitly examine poverty rates of leaver families, on average, over half of leavers are poor. Two studies found that the majority of leavers have incomes below 185 percent of the federal poverty line.
• In the few studies that compare monthly income for subgroups, continuous leavers have considerably higher incomes than leavers in general. Employed leavers also have much greater monthly incomes than jobless leavers.

Material Hardship

A number of leavers studies go beyond earnings, employment, income, and program participation and examine the extent to which leavers experience material hardships such as hunger and housing problems and whether these hardships are different for families on and off welfare. Key findings include:

• A quarter or more leaver families experience food hardships at some point after exiting TANF - problems having enough money for food or having food last for the month - and similar percentages are experiencing trouble paying rent or utilities.
• Although some studies show that leavers experience higher levels of food and housing-related hardship after exit relative to when on TANF, other studies show that hardships decrease or remain the same after exit.
• With regard to medical hardship (being unable to access medical care), four studies found leavers were more likely to report being unable to afford health care for their families after exit as compared with before exit.
• Several studies report results on material hardship across subgroups of leavers, including workers and non-workers and those who left TANF due to sanctions or time limits.
• Most studies that compare material hardship across employment status find that leavers who are working have lower levels of food, housing, and health care-related problems.
• The available evidence on whether sanctioned and time-limited families experience greater material hardships than families who left welfare for other reasons is mixed.

Child Well-Being

Although virtually all families leaving welfare have children, it is difficult to assess child well-being from either administrative data or a single interview. Thus, leaver studies contain limited information about children's outcomes and well-being. For the studies reporting this information, findings on children's health insurance coverage, health status, behavior, interaction with child welfare services, and child care arrangements are summarized below.

• Reports of children in poor or fair health are generally low, ranging from 5 to 10 percent. However, one-tenth to one-quarter of leaver families have children without health insurance.
• Although the measures of child behavior are varied, most studies that compare behaviors pre- and post-exit find that the majority of leavers report child behavior is better after exit.
• Rates of interaction with child welfare services range from 1 to 13 percent, including reports of abuse/neglect and foster care services. There is little evidence on whether the percentage of families involved in child welfare services changed after exiting TANF.
• For child care, a substantial percentage of leaver families rely on parental care. For those using non-parental care, siblings and other relatives of the child are by far the most common sources of care for children.

Mental Health and Employment Among Welfare Recipients (2000)

As TANF policies move welfare recipients into the labor force, there is growing interest and concern about the barriers that may prevent recipients from gaining and keeping employment. Mental health problems are one such barrier. Under a task order, Mathematica Policy Research, Inc. profiled the efforts of four states (Florida, Oregon, Tennessee, and Utah) to address the mental health needs of welfare recipients. This study was designed to: (1) identify and provide detailed information about the design and structure of mental health services developed by state and local welfare offices to address the mental health needs of welfare recipients, (2) highlight service delivery options in designing and implementing these services, and (3) discuss the key implementation challenges and lessons learned in providing mental health services to welfare recipients. The researchers visited one rural and one urban site in each state. The sites provided services in a variety of ways, and the research did not collect any evidence that suggests that one model for providing mental health services is better than any other. Rather, the project was designed to draw attention to the key design and implementation choices made by policymakers and managers in providing mental health services for welfare recipients. The final report, Providing Mental Health Services to TANF Recipients: Program Design Choices and Implementation Challenges in Four States, was released in August 2001, and is available at <http://aspe.hhs.gov/hsp/TANF-MH01/>.

National Academy of Sciences Panel Study on Welfare Outcomes (1998, 1999, 2000 and 2001)

ASPE made its final installment of support for the National Academy of Sciences' (NAS) Panel on Data and Methods for Measuring the Effects of Changes in Social Welfare Programs. The purpose of convening this NAS Panel was 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, and to provide the Department of Health and Human Services (HHS) with unbiased scientific recommendations for studying the outcomes of recent changes in the welfare system. The panel's conclusions and recommendations on research questions and populations of interest, evaluation methods and issues, and data needs and issues were presented in its report, Evaluating Welfare Reform in an Era of Transition, which was disseminated in 2001. The Panel published a collection of methodological papers by welfare outcomes research experts, Studies of Welfare Populations, Data Collection and Research Issues, in January 2002. The papers included in the volume discuss the current state of knowledge for surveying low-income populations; preparation and use of and access to welfare program-relevant administrative data systems; and measuring important outcomes for welfare studies.

Access to Welfare Outcomes Data Sets (2000)

Through an interagency transfer, ASPE funds were used to support storage of certain data sets at the Research Data Center (RDC) of the National Center for Health Statistics (NCHS). These data sets include state-specific administrative and survey data from state and county grantees conducting Welfare Outcomes studies on former, current, and potential TANF recipients. As of March 2002, documented data sets on families who left TANF were available for further analysis from fifteen grantees and data sets from families applying for TANF from three grantees. A few additional data sets, primarily from studies of families applying for TANF, will be placed at the Research Data Center in the next few months. Information on procedures for gaining access to welfare outcomes data sets is posted on the "Data Files" section of the ASPE-sponsored "Leavers and Diversion" web page at <http://aspe.hhs.gov/hsp/leavers99/datafiles/>.

Grants to States and Localities to Enhance Studies of Welfare-Related Outcomes

The purpose of these grants is to enhance state-specific surveys of populations affected by welfare reform, by expanding or improving data collection activities. Grants to states are being used, for example, to add additional survey waves to measure longer-term outcomes, collect data to support greater sub-group analyses, and gather more detailed information on non-respondents. To be eligible, states had to have an existing survey that had been administered at least once, so the grants can facilitate real improvements, without paying for basic startup costs. Survey efforts needed to fill an important knowledge gap that could not be filled with states' existing data. The data cover a variety of welfare reform outcomes, such as measures of family hardship and well-being, barriers to employment, poverty status, and utilization of support programs. The surveys focus on various subsets of the low-income population including long-term welfare recipients, child-only cases, former recipients, potential recipients, welfare leavers with little or no reported income, and other special populations affected by state TANF policies.

Iowa (2000)

Iowa builds on an existing study of families leaving the Family Independence Program (FIP). The study was conducted by Mathematica Policy Research (MPR), and consisted of three components: one focusing on families who have left TANF and report very low incomes; a second focusing on longer-term outcomes for families that have left TANF; and a third focusing on non-respondents from their earlier survey. MPR has secured significant funding from foundations in addition to the ASPE grant for this project. All components have been completed.

In their study of families with very low incomes, MPR conducted in-depth interviews of 16 families reporting no more than $500 in total income per month, including those with no TANF and no employment, and those with low levels of TANF and/or employment. These interviews focused on possible income sources that were missed or incorrectly measured, coping strategies and family well-being. They found that while many families had two or more sources of income, their incomes varied significantly from month to month, often leaving gaps in their ability to meet regular monthly expenses such as rent. Coping strategies included negotiating partial payment on bill, pawning goods, doubling up on rent and finding additional sources of income such as recycling cans or babysitting. MPR found that the great variability in monthly income made it likely that traditional one-time phone interviews of leavers are limited in their ability to capture economic well-being of families. A link to the report, Living on Little: Case Studies of Iowa Families with Very Low Incomes is available on Mathematica's web site at <http://www.mathematica-mpr.com/PDFs/liveonlittle.pdf>.

MPR's second analysis of Iowa added a second wave to their existing survey of welfare leavers to observe outcomes two years after families left TANF. The analysis also incorporates administrative data to help track income and program participation. The study found that most family heads were working two years after leaving TANF, had above minimum wage jobs and were as well or better off than they were a year ago. However, their work was unstable and over half remained poor or near poor. While only about one fifth returned to TANF, closer to one half participated in Medicaid and over one third in food stamps. A link to the report, Iowa Families That Left TANF: How Are They Faring Two Years Later?, is available on Mathematica's web site at <http://www.mathematica-mpr.com/PDFs/iowatwoyears.pdf>.

MPR's final analysis under this grant involved interviews with approximately 47 non-respondent cases from their earlier survey of one-year outcomes, targeting a response rate of roughly 60 percent. Information from these interviews was used to assess the representativeness of survey data on welfare outcomes and the implications for interpreting findings, by examining whether non-respondents to that survey where significantly different from respondents, and whether their non-response significantly biased the survey's findings. Using intensive location efforts, the project found and interviewed the majority of TANF leavers who failed to respond to the original survey. The study reports that the non-respondents were somewhat worse off. For example, they were less likely to have health insurance, and more likely to have housing-related problems such as going without utilities or having to double up. But their outcomes were not sufficiently different to bias the original survey's results, such that even if they had been included in the original survey, the findings would not be significantly different.

Supporting Families After Welfare Reform: Access to Medicaid, S-CHIP and Food Stamps (1999)

ASPE, along with the ACF, the Centers for Medicare and Medicaid Services, and USDA, contributed funding to a major $5.9 million initiative of the Robert Wood Johnson Foundation to provide technical assistance and grants to states and large counties to improve their enrollment and redetermination processes for Medicaid, the State Children's Health Insurance Program (SCHIP), and food stamps. Under the Supporting Families program, the Robert Wood Johnson Foundation provided funding for assistance to states or counties to work on Medicaid and SCHIP, while federal funding provided assistance to work on food stamps, Medicaid, and SCHIP. The expert technical assistance included analysis of performance data, identification of the root causes of problems in their enrollment processes, and/or development of specific implementation plans to solve the problems and increase the participation rates in Medicaid, SCHIP, and food stamps. Information on the Robert Wood Johnson Foundation's overall initiative to solve problems in eligibility processes that make it difficult for low-income families to access and retain Medicaid, SCHIP or food stamps - particularly families moving from welfare to work - can be found under "Supporting Families after Welfare Reform" at <http://www.rwjf.org/>.

Federal funding supported a literature review and synthesis on the post-welfare reform drop in participation in the Medicaid and Food Stamp programs, including reasons underlying the changes in participation, and potential strategies for increasing participation among eligible families. The report, Access to and Participation in Medicaid and the Food Stamp Program - A Review of the Recent Literature, was released in March 2000, and can be found at <http://www.acf.dhhs.gov/programs/opre/med-fs.htm>. It includes findings from government- and privately-sponsored research projects, studies of participation in the Food Stamp Program (FSP) and Medicaid at the national and state level, studies of low-income families who have left welfare, reviews of research, and ongoing analysis and data collection efforts. Federal funding also supported visits to promising practices sites to identify practices that appear to enhance or facilitate participation in the Medicaid/SCHIP and Food Stamp programs by former TANF and low-income families. The final report, Promoting Medicaid and Food Stamp Participation: Establishing Eligibility Procedures that Support Participation and Meet Families' Needs, synthesizes findings across all promising practices and program improvement sites. The case study reports can be found at <http://www.mathematica-mpr.com/Publications/publications.asp>.

Low-Income/Low-Skilled Workers' Involvement in Nonstandard Employment (2000) (formerly Low-Income/Low-Skilled Workers' Involvement in the Temporary/Contingent Employment Sector)

The prevalence of alternative work arrangements such as temporary and on-call jobs in the U.S. work force has grown considerably in recent years. The growth is likely to have important implications for low-income workers, particularly since the advent of welfare reform and its emphasis on getting welfare recipients into jobs quickly. Alternative work arrangements, especially for those with limited work histories, might be expected to be a natural pathway to work for welfare recipients. However, little is known about the use of nonstandard work as a gateway into the work force for the low-income and the low-skilled. This project examined the role of nonstandard work arrangements in today's labor market, paying particular attention to the effect of such arrangements on low-income workers and those at risk of being on public assistance.

Findings from the October 2001 final report by the Urban Institute, which can be found at <http://aspe.hhs.gov/hsp/temp-workers01/>, include:

• Workers who are at risk of welfare dependency are more than twice as likely to be in alternative work arrangements as other workers.
• The number of industries utilizing temporary workers has increased, particularly among industries where the median education level of temporary workers is very high - suggesting that low-skilled workers will be increasingly less able to compete for these jobs.
• At-risk workers fare worse in nonstandard jobs than do others workers in such arrangements across a variety of dimensions: wages, incidence of part-time hours, job duration, and employer-provided benefits.
• Individuals who had a spell in temporary work had worse earnings and employment outcomes a year later than did similar individuals with a spell in standard employment. But if the employment outcomes of temporary workers are compared to those who had a spell of unemployment, rather than to workers in standard jobs, temporary workers fared much better.

Linking State TANF Policies to Outcomes:  A Preliminary Assessment (2000 and 2001)

In FY 2000, ASPE issued a task order to the Urban Institute to analyze and synthesize available information on state welfare and related support policies and assess which characteristics of state programs or background characteristics are most significant in predicting outcomes. As part of this project, Urban convened a technical work group (TWG) of researchers to make recommendations on which existing typologies were most promising and on directions for analysis. The TWG recommended that Urban develop new typologies, rather than modifying existing typologies.

Therefore, under the revised task order, Urban Institute developed six typologies, each containing the variables that are expected to affect a specific outcome (such as recipient job entry or the poverty rate). The variables in these typologies, along with state level data on outcomes and contextual variables, have been entered into a database which will be available to researchers interested in studying the relationships between state TANF policies and the range of outcomes experienced by current and former welfare recipients and other low-income populations. This database will be available through the ASPE and Urban Institute websites in Summer 2002, along with a complete data dictionary, documentation of sources, and explanation of the reasoning behind the typologies. Urban has also conducted cluster and factor analysis on the recipient job entry typology as an example of what can be learned through these approaches.

Alternative Kinship Care Programs Set up Outside of TANF and Foster Care (2000)

Families in which a grandparent or another relative has taken over parental responsibilities make up approximately one-third of both the TANF and foster care caseloads. Neither of these service systems have been set up with such families in mind, and, in many ways, the services provided are an inadequate match with families' needs. Several states have set up separate kinship care assistance programs outside the traditional structures of both the child welfare and TANF systems. Under this project the Urban Institute profiled an emerging trend on the part of states and localities to develop "alternative" kinship care programs designed to meet the needs of kinship care families outside of traditional foster care or TANF programs. The final report, On Their Own Terms: Supporting Kinship Care Outside of TANF and Foster Care, describes the characteristics and service needs of kinship caregivers, the range and scope of alternative kinship care program models and services, lessons learned about designing and implementing alternative programs, and policy implications for child welfare, TANF/welfare, and other agencies. In-depth information on programs in seven sites ⁽²⁾ is also included. The report was released in September 2001 and is available at <http://aspe.hhs.gov/hsp/kincare01/>.

Welfare Reform and the Health and Economic Status of Immigrants and the Organizations that Serve Them (1998 and 1999)

Several government agencies (ACF, the Centers for Medicare and Medicaid Services (CMS) and ASPE in HHS; the Immigration and Naturalization Service in DOJ; and USDA's Food and Nutrition Service and Economic Research Service) and private foundations (the Ford, William and Flora Hewlett, and Andrew W. Mellon Foundations) awarded a grant to the Urban Institute in 1997 to study immigrant families in the context of welfare reform. ASPE contributed funds in 1998 and 1999. The project involved a large-scale study of immigrants and their communities in Los Angeles and New York City to deepen our understanding of the impact of changes in federal laws on immigrant families and children. The final report, How Are Immigrants Faring After Welfare Reform? Preliminary Evidence from Los Angeles and New York City, was released in March 2002.

This report analyzes survey data collected in late 1999 and early 2000 from 3,447 immigrant families in New York City and Los Angeles County. Major findings are summarized below; most are consistent with other recent research findings related to poverty rates and rates of health insurance among immigrants.

• More than half the members of immigrant families in Los Angeles and New York City are citizens, with about one-third being native, U.S.-born children.
• Immigrant families have relatively high rates of poverty, with recent immigrants being poorer than longer established immigrants.
• One-half of immigrant families in Los Angeles, and two-thirds in New York City, are limited English proficient (LEP), and LEP adults are nearly 3 times as likely to be poor as immigrants who speak English well.
• Immigrants tend to have lower incomes despite relatively high labor force attachment; labor force participation is about 80 percent for immigrant adults in Los Angeles and New York City, and 73 percent for low income immigrants.
• Similar to findings from other studies, immigrants in the two cities were less likely to have health insurance than citizens, primarily due to a lack of job-based coverage. Although immigrant adults are as likely as native citizens to be enrolled in Medicaid, they are 3 times more likely to be uninsured due to the lack of job-based insurance coverage.
• Citizen children in immigrant families are about twice as likely to have health insurance as immigrant children in those families. In general, children are more likely to be insured under the New York State SCHIP program than the California SCHIP program; the former has been in existence longer than the latter.
• Roughly one-third of immigrant families in Los Angeles and New York City are food insecure, which is about 3 times the level of food insecurity of native citizen families; LEP families are more likely to be food insecure than English proficient families (40 versus 21 percent in Los Angeles, and 36 versus 24 percent in New York City); about half of families where adults speak no English at all are food insecure.
• While participation of immigrant families in the Food Stamp Program declined significantly between 1996 - 2000 (about 50 percent), the reported reasons for reduced or terminated benefits were generally unrelated to immigrant status. The most prevalent reason given was an improvement in economic circumstances. Other reasons included changes to family composition, and simply choosing to stop receiving benefits.
• Even though large proportions of food insecure immigrant families do not receive food stamps (about four-fifths in both cities), the families that do receive benefits are those most in need; single parent families with children are more than twice as likely, and those with LEP adults three times as likely, to receive food stamps than other families, when controlling for poverty and immigration status.

This report and other project reports can be found at <http://aspe.hhs.gov/hsp/hspother.htm#immig>.

In a related effort, ASPE funds were used to support the Urban Institute's updating of the TRIM modeling program (used to simulate welfare caseload changes resulting from changes in various policy variables) to include parameters about immigrants, and as a subset, refugees and non-refugees, using 1995 data as a baseline. This updated model could be used to estimate the rates of participation in TANF, Medicaid, and food stamps by children, both citizen and immigrant, who live in immigrant- and citizen-headed households. (TRIM has subsequently been updated to include immigrant variables for later years, using other funding sources.)

Researcher Initiated Grants on Welfare Outcomes

ASPE has supported researcher-initiated proposals to study important questions related to the outcomes of welfare reform in FYs 1999, 2000 and 2001. Eight welfare outcomes research grants totaling $800,000 were awarded in FY 1999 to analyze a variety of information about low-income individuals (both adults and children) and their families, including their economic and non-economic well-being and their participation in government programs. Issues examined under those grants include caseload dynamics, the impact of spatial distribution of economic opportunities, health insurance and health care utilization, the use of food stamps, living arrangements, maternal and child health, domestic violence, and quality-of-life issues.

The FY 2000 grant program, conducted in cooperation with the Administration for Children and Families, focused on use of state and federal administrative data, and on current and former TANF recipients and other special populations affected by state TANF policies. Priority research interests included the composition of the caseload, patterns of government program use, sub-populations, non-working welfare leavers, sanctions, employment stability, marriage and family structure, TANF flexibility, barrier identification and service utilization, and entry effects and welfare dynamics. ASPE awarded approximately $1.3 million to 10 applicants. In general, ASPE funding supported research and secondary data analysis efforts that would be completed within 12 months covering a variety of information about adults, children, and families, including economic and non-economic well-being and participation in government programs. ACF awarded an additional $1.2 million in FY 2000 to support continuation of two of the projects beyond this first year and seven other longer-term projects involving primary data collection.

Some grants have been completed or have interim products. When available, final reports from the grantees will be posted on the ASPE website at <http://aspe.hhs.gov/hsp/>.

RAND: Entry, Exit and the Changing Composition of the Caseload (2000)

This project explored the role of the economy in explaining the welfare caseload declines. It examined the relative importance of changes in the rates of entry, exit, and re-entry in explaining the observed caseload declines using individual level data for California. The results suggested that adjustment is far from instantaneous. Changes in the entry rate were much more important than changes in the exit rate and re-entry rate for explaining changes in the welfare caseload itself.

Baruch College, City University of New York: Effects of Welfare Reform on Investments in Human Capital and Family Formation (2000)

This study investigated whether the behavior of teens and young adults ages 16 to 21 has changed as the result of welfare reform. Researchers used data from the 1979 and 1997 National Longitudinal Surveys of Youth (NLSY) to compare cohorts (both between and within) that entered these ages prior to and following welfare reform, describing differences in outcomes and behaviors such as high school completion, teenage and non-marital child bearing, employment and welfare receipt. The study found that rates of first-time entry onto welfare have declined considerably, primarily among older teens. The decline in the welfare caseload associated with reform is not solely a matter of an increase in the rate of leaving welfare or a decline in recidivism, but also has occurred because of a decline in the rate of initial entry. Teenage mothers were less likely to enter welfare in the late 1990s than in the late 1970s and early 1980s. The study also found some weak evidence that welfare reform may have reduced fertility among teens. The decline in welfare entry among teenage mothers has been accompanied by changes in their living arrangements. Compared to the pre-welfare reform cohort, teenagers who have had a non-marital birth are less likely to receive welfare, and those who are not on welfare are more likely to live with a parent.

University of Oregon: TANF and Household Savings (2000)

This project studied the impact of new savings incentives offered to participants in the TANF program. Researchers addressed the following questions: 1) Has saving increased among those low-income households who reside in states that have increased the liquid-asset and vehicle equity limits for program eligibility? 2) Has saving increased among those low-income households who reside in states that have introduced Individual Development Accounts? 3) What is the impact of time-limited benefits on household savings? 4) Are there differences by race, marital status, and poverty status in the response to the new saving incentives?

The study used data from the 1989, 1994, and 1999 wealth supplements of the Panel Study of Income Dynamics. A final report, released in June 2001, found that the near-term saving of the poor has responded modestly to several of the recent changes in welfare policy, but that in general the saving of the near-poor has not responded to welfare reform. Specifically, increasing asset limits has had a small positive effect on saving by the poor, while the imposition of time limits appears not to have motivated low-income households to save more.

University of Michigan: Mental Health, Substance Abuse, and Domestic Violence Service Utilization by Welfare Recipients (2000)

This project analyzed the impact that spatial proximity to social service providers and individual-level characteristics have on service utilization rates among welfare recipients in the three-county Detroit metropolitan area. This study showed that major depression and several limitations in physical functioning are related to lower probabilities of employment and few months of employment. The study also showed that severe abuse is prevalent for the sample and recent and persistent experiences are associated with welfare reliance without work, lower earnings, and great likelihood of material hardship and financial strain. Researchers used data from a survey of welfare recipients in the Detroit area.

Case Western Reserve University: The Effect of Job Accessibility and Neighborhood Characteristics on the Employment Stability of Welfare Leavers in an Urban Labor Market (2000)

This study examined the spatial patterns of employment and the effects of neighborhood conditions and job accessibility on the labor market experiences and employment stability of adult female case heads leaving welfare. The analysis focused on employment stability, earnings, and wages at approximately 6 and 13 months after individuals left TANF in the Cleveland metropolitan area. Labor market success was measured in a number of ways, including the level of earnings and wages, the growth of earnings and wages, and the probability of remaining employed. Results suggested that welfare leavers had less job access than the Cleveland area population in general. Relative to the available jobs in the labor market, their jobs were over-represented in inner city and minority employment locations. There were few effects on employment outcomes of job access.

University of Illinois: Young Mothers' Transitions On and Off TANF: How do Child Care Assistance, Job Training, and Social Supports Influence These Decisions? (2000)

This project identified the likelihood that young mothers entered, stayed on, or left TANF given the use and/or availability of child care, job training, and other social programs in their community. Data on three subgroups of young mothers (ages 18-24) who lived in the Chicago metropolitan area between January 1, 1997, and June 30, 2000, were analyzed. The major findings were that receiving Medicaid while not on TANF is related to a lower risk of starting to receive TANF. Mothers also are less likely to return to TANF after exiting if they get help paying for child care.

Washington University: The Impact of Welfare Reform in North Carolina: Exits, Employment, Earnings and Recidivism (2000) (formerly Employment, Earnings and Recidivism: How do Entrants to TANF Differ from Entrants to AFDC?)

The project used administrative data from North Carolina, focusing on five consecutive entry cohorts (those who began receiving AFDC/TANF payments in February, March or April of 1995, 1996, 1997, 1998, or 1999). Data on client demographics and welfare receipt were linked to employment data and to data on local labor markets. All analyses were broken out by race/ethnicity. The analyses found that:

• Later cohorts of entrants are younger, have less work experience immediately prior to entry, have lower levels of education, and are substantially more likely to have young children. Correspondingly, later cohorts were more likely to be experiencing their first spell of welfare receipt. These trends hold true across racial/ethnic groups.
• Family structure strongly affects exits from and return to welfare. Married couples exit welfare faster, and are less likely to return to welfare, even though married women have lower post-exit earnings than single ones. Families with a child less than one year old, are somewhat less likely to exit welfare than those without one, but very much more likely to return after having exited. Families with more children are not less likely to leave welfare, but are substantially more likely to return.
• There was a substantial increase over time in employment rates while receiving welfare. However, the percentage of exits that were work-related declined somewhat over the cohorts. Time off of welfare and increased work experience are both associated with earnings gains over time. However, most recipients' earnings were not enough to lift them and their families out of poverty, even four years after exit. Employment rates decline over time off welfare, even among those with previous work experience.
• Black women exited welfare more slowly than white or Hispanic women, and were more likely to return to the rolls. This is true even though they were more likely to work while receiving welfare, to have work-related exits from welfare, and to remain employed. Earnings and employment rates, as well as absolute gains in earnings over time are greater for black women than for white women. This is in part driven by higher rate of part-time work among white women. This is consistent with a hypothesis that the white women in this sample are more able to depend on income from other family members.
• Nearly 80 percent of all jobs obtained by welfare recipients were within the service and retail industries. Many of these jobs were part-time.

Low-Income Families: Coping as Parents and Workers (2000) (formerly How Low-Wage Working Families Cope as Parents and Workers)

Low-wage working families face multiple demands as workers and as parents. Besides working, low-income parents in both single- and two-parent families need time for training and education, navigating complex health and support services, parenting, and managing their children's needs. Some low-wage working parents also are providing care for family members who are elderly or have special needs, or must work nonstandard and irregular hours. This project, jointly sponsored by ASPE and ACF, is assessing coping mechanisms used by low-income families by examining a variety of factors that may help or hinder a family's efforts to be self sufficient, including formal and informal support services, social support networks, and time and money management. The project examines all low-income families as well as teen parents and those leaving TANF assistance and entering the labor force for the first time, and considers the effects of coping on children. A conference was held in November 2001, and research papers were presented at the conference.

The Feasibility of Replicating the Women's Employment Study (2000)

Widespread anecdotal evidence suggests that the welfare caseload is becoming increasingly harder to employ, and a number of surveys of adult welfare recipients have demonstrated that they have a higher prevalence of multiple barriers to employment than women at large. However, these studies are generally small and not representative, and the questions used to assess the prevalence of barriers differ from study to study, making cross comparisons difficult.

The purpose of this project was to build on experiences from the Women's Employment Study at the University of Michigan to review what we have learned to date and suggest how we might go about designing surveys that would provide data about multiple barriers to employment, including health, mental health, domestic violence, literacy, work skills, etc.

An ASPE-sponsored workshop was held in March 2001 to generate recommendations on how best to design a survey study of welfare recipients to understand the factors that encourage or, in contrast, hinder welfare recipients' transitions to employment and self-reliance. Participants included academic and policy researchers with a wide range of methodological and substantive interests, and research staff from HHS (including the National Institute of Mental Health) and the Michigan Family Independence Agency. The project's final report, released in February 2002, summarized the conference themes and included recommendations for a survey of welfare recipients. Some of the issues addressed include: designing the sample; reducing non-response, particularly non-response due to inability to locate sample members; costs and benefits of collecting administrative data along with survey data; ideas for measures of employability; conceptual and methodological issues related to studying domestic violence; measures of physical and mental health appropriate for welfare studies; and design issues for collecting data on substance abuse. The report, Designing Surveys of Welfare Populations: Report from the Workshop on Designing Surveys of Welfare Recipients, March 15-16, 2001, Ann Arbor, Michigan, will be posted at <http://aspe.hhs.gov/hsp/> when available.

Trends in the Economic Well-Being of Low-Income Americans (2000)

This project analyzed data from various sources, including the Current Population Survey (CPS), the Panel Study of Income Dynamics (PSID), and administrative data, to track trends in income, poverty and other economic measures, such as food security and access to health insurance. The resulting tabulations are being used to support internal analyses of income and poverty, children and their families, working-age adults, the elderly, and the impacts of public programs including outcomes of welfare reform.

Iowa State University Project: Considerations for Designing State-Surveys to Support Integration with National Survey Data (1999 and 2000) (formerly Support for Iowa State University SPD Project)

The goal of this research project was to investigate alternative systems for state-level surveys to support welfare reform assessments that can be linked to related national, regional and state data. The investigation focused on sampling design, questionnaire design, and local assessment of transportation needs in a rural state.

This project found that a dual-frame sample design - one that uses random digit dialing for a telephone survey and a list-based frame - can be implemented to target both welfare recipients and the general population to provide efficient estimates of the population of interest. In addition, it was found that local-specific topic modules could successfully be incorporated into a current national survey instrument. The researchers developed a transportation module and incorporated it into the instrument used for the national Survey of Program Dynamics to address information needs of rural communities in the state. A pilot study using the dual-frame sampling methodology and the modified instrument was successfully carried out.

Technical Assistance to Welfare Outcomes Grantees (2000) (incorporates Technical Assistance on Researcher Access to Data Sets, 1999)

States and counties that received FY 1998 and FY 1999 Welfare Outcomes grants have submitted research data sets that combine the state-specific administrative data they have collected on former, current, and potential TANF recipients and other special populations affected by state TANF policies, including diversion practices. Grantees were expected to submit the data sets to ASPE, and also to make them available for research purposes. To improve the quality and comparability of these data sets, and to ensure that the data are appropriately documented and accessible to outside researchers, ASPE modified and extended an earlier task order contract with ORC Macro in FY 2000 to provide technical assistance and coordination in the preparation of the data sets. ORC Macro helped coordinate the ASPE public use data file work group, and released a technical assistance guide for the grantees in Fall 2000 on procuring and documenting researcher-access data files. In addition, ORC Macro reviewed, analyzed, and edited (where necessary) those data files that were submitted by the grantees prior to the end of this contract in September 2001.

Research Technical Assistance on State Child Indicators Initiatives (1998, 1999, 2000 and 2001)

Funding was provided to researchers at the Chapin Hall Center for Children at the University of Chicago for technical assistance to states that received grants to promote child indicator work to monitor changes in welfare and other key policies. The technical assistance effort emphasized collaborative work among the states and peer-to-peer assistance efforts. Technical assistance was provided, for example, on conceptual and methodological issues in identifying and measuring appropriate sets of child health and well-being indicators within and across states; ways of creating or using survey and administrative data and of combining several data approaches; and ways to involve state policy makers who can help institutionalize data systems for measuring and tracking child indicators and establish procedures for using indicator information to inform policy deliberations. The reports and products developed by the states, as well as summaries of meetings at which technical assistance was provided to grantees, will be available primarily through the ASPE website (see <http://aspe.hhs.gov/hsp/cyp/child-ind98/>) in Summer 2002.

Interim Products from Continuing Welfare Outcomes-Funded Studies

Grants to States and Localities to Study Welfare Reform Outcomes, with an Emphasis on Diversion (1999)

One of the Congress's major objectives in providing welfare outcomes money to ASPE over the last several years is to measure outcomes for a broad population of low-income families, welfare recipients, former recipients, potential recipients, and other special populations affected by state TANF policies, including diversion practices ⁽³⁾. To this end, ASPE issued a request for applications from states and large counties in April 1999 with an emphasis on the study of applicants and potential applicants to the TANF program. ASPE awarded seven grants under this announcement, six of which specifically support state efforts to gather a variety of information about individuals and their families who apply to TANF, including those who are formally or informally diverted. In addition, several of the leavers studies funded in FY 1998 had significant applicant components to their projects.

ASPE is particularly interested in learning about the degree to which TANF applicants receive, or are aware of their potential eligibility for, Medicaid, food stamps, and other programs and services that are important in helping low-income families make a successful transition to work. Below are summaries of the grants provided to states and large counties in fiscal years 1998 and 1999 with a particular emphasis on TANF applicants and diversion from TANF. "Diversion" in this context is not limited to participation in formal diversion programs, but also includes "informal" divertees. These are usually defined as individuals who began the application process but were either deemed ineligible for non-monetary reasons, withdrew voluntarily after completing the process, or failed to complete the process for some other reason. These project updates are current as of March 2002.

Arizona (1999)

Arizona built on their FY 1998 study of leavers by looking at informal divertees and entrants to TANF. The study used a wide range of administrative data (including data on child care subsidies) to track second quarter 1999 divertees and entrants for 12 months, and included two waves of surveys of 400 individuals in both populations. Some of the subgroups on which the state focused included urban vs. rural applicants and applicants who are initially denied but eventually reapplied for TANF. Arizona collected administrative data from a number of different sources, including a data warehouse established as part of the FY 1998 ASPE leavers grant.

The final report for the project was released in October 2001, and a link to the report has been posted on ASPE's website. Both sets of applicants (divertees and entrants) showed dramatic increases in employment, income, and other indicators during the study period, and most of the families stated that they were better off at the time of the second interview (about 15 months after applying for TANF). However, about half of the respondents remained unemployed, and cited medical issues and preference to care for the child at home as the main reasons for not working.

Contra Costa County and Alameda County, California (1999)

Contra Costa and Alameda Counties are located in the East San Francisco Bay area of California and contain the cities of Oakland and Richmond. This project studied TANF leavers from both counties, as well as formal and informal divertees in Contra Costa County. Researchers at the SPHERE Institute were able to take advantage of these counties' Case Data System (CDS), which includes every TANF application that is initiated in the two counties. The CDS allowed SPHERE to uncover the reasons individuals were diverted from or left TANF, as well as make comparisons across the two counties. SPHERE used the CDS both to link all applicants with other administrative databases and to draw their survey sample of 850 leavers and 150 divertees from the third quarter of 1999. Comparisons between divertees, leavers and cases that transitioned to child-only cases, and between the two counties were made in the final report, which includes data from two rounds of surveys.

The SPHERE Institute released the final report for this project in October 2001. They found that, overall, conditions improve in the period following TANF application for both leavers and the informally diverted, but that one year after applying for TANF, leavers were doing somewhat better than informally diverted families and much better than families that transitioned to child-only cases. Awareness and use of post-exit "transitional" benefits was not high. A link to the final report (Assessing the Family Circumstances of TANF Applicants and Leavers in Contra Costa and Alameda Counties, CA: Final Report) has been posted on ASPE's web site at <http://aspe.hhs.gov/hsp/leavers99/reports.htm>.

Texas (1999)

This project represented the combined efforts of the Texas Department of Human Services, the Texas Workforce Commission, and the University of Texas-Austin. It focused not only on informal divertees, but also on potential TANF applicants who are formally diverted by the state, either through a one-time lump sum payment or by redirection into work. The administrative data analysis incorporated a wide variety of sources, and tracked both applicants who were redirected into work or denied for non-financial reasons and participants in the lump-sum diversion program. In addition to the two waves of surveys conducted with applicants who were denied TANF for non-financial reasons, the state added leavers to both the survey and administrative samples.

The final report for the project was released by the Texas Department of Human Services in January 2002, and a link to the report has been posted on ASPE's web site. The state found that families diverted from TANF quickly resumed their prior levels of receipt of TANF, Medicaid and food stamps, as well as their prior level of earnings, suggesting that diversion policies in Texas have few long-term effects on diverted families. The majority of redirected applicants and leavers were employed in the period immediately following the study quarter.

Leavers Studies that also Examine Diverted Populations (Florida, San Mateo, and South Carolina) (1998)

Three FY 1998 grants that have a primary focus of studying outcomes for families leaving welfare also include research on families that were formally or informally diverted from entering TANF. These three grants are Florida, a consortium of California counties, and South Carolina.

The Florida study, undertaken by researchers at Florida State University, examines three groups of individuals from the second quarter of 1997: TANF leavers, individuals who began the application process but who either withdrew voluntarily after completing the process or failed to complete the process ("diverts"), and individuals who receive food stamps or Medicaid, have minor children, and have income and assets below the cash assistance limit but who do not receive cash assistance ("opt-nots"). Findings from Florida's final report, released in November 2000, indicated that the "diverts" and "opt-nots" look very similar to leavers in terms of employment rates. However, leavers appear to have slightly higher earnings and slightly lower use of government services than the other two groups.

The study in San Mateo, Santa Clara and Santa Cruz counties in California was funded primarily as a leavers study. However, because the administrative system in all three counties includes all applications, and not just those for individuals who receive TANF, researchers at the SPHERE Institute were able to also study informal divertees. Analysis of administrative data is being supplemented by surveys administered at six, 12 and 18 months after "case closure" (when either the applicant withdraws from the application process or the TANF recipient leaves the program). A draft report summarizing the first wave of survey data along with linked administrative data was released in December 2000.

A final report from SPHERE incorporating the second and third waves of survey data was released in November 2001. At eighteen months after the study quarter, conditions were improving for both leavers and informally diverted families. Trends in the awareness and use of transitional assistance were positive among leavers and divertees, but the findings also indicated room for further improvement. For example, the use of food stamps declined among one-parent leaver and informally diverted families, but the percentage of eligible families not receiving food stamps also declined.

The project in South Carolina also is focused primarily on leavers, but state researchers have also used food stamp records to identify families that appeared to be eligible for cash assistance but were not enrolled. Surveys were conducted with families who went on food stamps between October 1998 and March 1999 and who did not subsequently apply for TANF. The surveys took place at approximately one year and two years after entry into the food stamp caseload. Each of the families had dependent children and was eligible for TANF based on gross income, but did not enroll in the TANF program.

A report detailing the findings from Wave 1 of the survey was released by South Carolina and their contractor, MAXIMUS, in June 2001, and Wave 2 survey results subsequently were released in June 2002. Results of the study suggest that, in addition to families who divert from welfare due to time limits or work requirements, there may be a broader population of families who are not going on welfare even though they are eligible based on their level of income and number of dependent children. The final report detailing findings from Wave 3 of the survey and summarizing the previous reports is expected to be released in Fall 2002.

Describing State Initiatives to Promote and Support Marriage (2001)

One of the four purposes of TANF is to "end the dependence of needy parents on government benefits by promoting job preparation, work and marriage." An important step in advancing the policy agenda around supporting marriage is to build a greater knowledge base around existing state policies pertaining to marriage. This project, conducted by the Lewin Group, develops an inventory of state-level policies directly aimed at promoting and supporting marriage in the following areas: campaigns, commissions and proclamations; divorce laws and procedures; marriage and relationship preparation and education; state tax policies; state transfer policies; state vital statistics; marriage support and promotion; youth education and development; court initiatives; and specialty programs. The project is purely descriptive, documenting which policies exist across states without assessing policy effectiveness or promising practices. Information was gathered primarily through secondary sources of information such as printed reports and web searches. An interim report was released in Spring 2002 and is available on the ASPE website at <http://aspe.hhs.gov/hsp/marriage02/>. The final report, which will make minor revisions, is expected to be released in Summer 2002.

The Link Between Marriage and Low-Income Family Well-Being (2000) (formerly How Important is Marriage to Low-Income Family Well-Being?)

This project, a Researcher Initiated Grant on Welfare Outcomes awarded to the Urban Institute, began in September 2000 and is examining the interactions between family formation status and economic well-being to better understand the extent to which marriage is a protective factor against economic hardship, particularly among the disadvantaged population. The study examines various types of family formation, including single, married and cohabiting parents, and it looks at measures of poverty as well as material hardship. The project is producing four reports, based on various data sources.

The first report, completed in August 2001, is based on data from the National Survey of America's Families, and found that, overall, marriage reduced the likelihood that parents would experience poverty of material hardship. The report is titled Marriage as a Protective Factor Against Economic Hardship: An Initial Report. The second and third reports, Married and Unmarried Parenthood and the Economic Well-Being of Families: A Dynamic Analysis of a Recent Cohort and The Relationship Between Marriage and Other Family Structures and the Material Hardship Experienced by Households with Children: Evidence from the Survey of Income and Program Participation, respectively, are expected to be completed by Summer 2002. The final report, a literature review, also is anticipated to be completed by Summer 2002.

Evaluation of State Programs of Abstinence Education (2001)

The Balanced Budget Act of 1997 (BBA) authorized an evaluation of abstinence education programs that are funded through Title V Section 510 of the Social Security Act and administered by the Maternal and Child Health Bureau (MCHB) within the Health Resources and Services Administration (HRSA). The BBA set aside $3 million for each of fiscal years 1998 and 1999 for federally sponsored evaluations of abstinence education programs. ASPE was delegated responsibility for managing the evaluation, and competitively awarded a contract to Mathematica Policy Research, Inc. to carry out this work. The ASPE 2001 funds will allow the evaluation to follow adolescents for longer periods of time.

The Interim Report to Congress, The Title V Abstinence Education Program Evaluation: The First Four Years, released in April 2002, describes the progress of the evaluation and presents early implementation and operational findings, drawing most heavily on the experience of the 11 abstinence education programs that were selected to be in the evaluation. The report illustrates the wide range of abstinence education programs that are operating, and the innovative ways in which programs are using funds to promote abstinence as the healthiest choice for youth. A number of the initiatives are diverse, creative, and, in many instances, offer youth much more than a single message of abstinence. Youth respond positively to staff who show strong and unambiguous commitment to the program message, and programs that use an intensive set of youth development services to enhance and support the abstinence message are very well received. The report shows that addressing peer pressure is difficult, and many programs have struggled to address these issues and engage parents in this process. The report also describes some ways in which programs may partner with local schools to provide abstinence education, highlighting some of the challenges to creating and sustaining these partnerships. No programmatic impacts are presented in the Interim Report. Details and updated information about the evaluation can be found at Mathematica Policy Research's web site on abstinence-only education programs <http://www.mathematica-mpr.com/3rdlevel/abstinence.htm>.

Endnotes

1.  Dates in parentheses following the title of each project identify the Fiscal Year(s) in which the study was funded.

2.  The seven sites/programs included A Second Chance, Inc. (ASCI) in Pittsburgh, PA; Grandparents and Kinship Program in Denver, CO; Florida's statewide Relative Caregiver Program; the Kentucky Kinship Care Program; the Kentucky KinCare Project; the Kinship Support Network (KSN) in San Francisco; and a variety of ongoing kinship care initiatives in Oklahoma involving services to relative caregivers.

3.  Diversion programs include formal efforts to address the immediate needs of families seeking cash assistance in ways that avoid enrolling these families in TANF. Examples of formal diversion programs include lump sum payment programs, mandatory applicant job search programs and the exploration of alternative means of support.

III. Projects Proposed to Be Funded by the FY 2002 Appropriation to Study Welfare Outcomes

Introduction

The Office of Human Services Policy's (HSP) proposed Research and Evaluation Plan for FY 2002 includes a variety of welfare outcomes-funded research, evaluation, and data projects that will be critical to understanding the outcomes of low-income families and to the Department's ability to respond to questions about those outcomes. Many projects have been proposed and approved for funding; others are still in developmental stages. Some are stand-alone projects; others represent specific components of larger studies and/or enhance on-going research, evaluation, and data activities. Some projects are fully funded by the targeted welfare outcomes funds; others are co-funded with other partners. They have been developed with a view toward addressing knowledge gaps and complementing other research efforts.

Our proposed research agenda includes projects that address several of the Administration's priority themes, including encouraging work and self-sufficiency, promoting healthy marriage and strong families, improving child well-being, expanding state flexibility and accountability, expanding the use of faith-based and community-based organizations, and addressing the needs of special populations. Our research plan is loosely organized around the following broad topical areas:

• Working Toward Independence
• Rallying the Armies of Compassion
• Crosscutting Issues

Planning for our FY 2002 welfare outcomes-funded research agenda is not yet complete. The following narratives describe projects most likely to be funded by ASPE in FY 2002 from the targeted funding to study welfare outcomes. Some of these projects could be modified or cancelled; other welfare outcomes-funded projects could be added.

Working Toward Independence

The Enhanced Services for the Hard-to-Employ Demonstration and Evaluation Project

Employment programs for low-income parents with barriers to work will be tested under this large-scale project. The contractor, MDRC (in partnership with Urban Institute, the Lewin Group, and the California Institute of Mental Health) is conducting a multi-site, random assignment evaluation examining implementation issues, program designs, net impacts and costs and benefits of selected programs. It also will assess how the programs affect family functioning and child well-being. Program interventions will address a range of potential employment barriers including those having to do with physical health, mental health, human capital, and situational barriers. Some of the programs will focus on two generations, providing employment services to hard-to-employ adults while providing direct services to children or youth. The evaluation includes long-term follow up of the participating families. ASPE and our partner, ACF, first funded this project last year, and it will last about eight years.

Spending on Social Welfare Programs in Rich and Poor States

The purpose of this project is to examine the effects of fiscal capacity on state spending choices on programs to support low-income populations. Although a great deal of research has been conducted in the area of state fiscal capacity, little is currently known about how fiscal capacity directly affects state spending choices regarding social welfare programs. This project will focus on the effects of state fiscal capacity on the proportion of the each state's budget devoted to social service programs over time, as well as states' flexibility in spending for specific types of social service programs. Specifically, the project will explore whether choices in the allocation of state funds for low-income populations vary based on whether a state is "poor" or "rich." The project will include a two-part study of state spending on social services. The first part will use existing data sources to build a multivariate, 50-state model to examine social welfare spending choices made by each state. In the second part of the study, the contractor will gather additional information not provided by the 50-state model to build a separate, more detailed model that will analyze the difficult spending decisions across social welfare program that are faced by the very poorest states.

A Profile of Families Cycling On and Off Welfare

This study would focus on the subgroup of TANF families who exit and re-enter welfare, particularly those who return for multiple times. It builds on past ASPE-funded studies of welfare leavers and welfare applicants, which found that many leavers re-enter welfare, and similarly, that many applicants have prior welfare history. According to Moffit (2001), about 20 percent of the caseload have three or more spells over a ten-year period. This study would attempt to answer three research questions about such cyclers: (1) What are their demographic characteristics and employment outcomes, as compared with other comparison groups of welfare recipients or leavers? (2) How have patterns of benefit receipt and the phenomenon of cycling changed since PRWORA? (3) Do state or county offices treat families re-entering welfare for the third time differently than other families? To answer the first two questions, the contractor would review the literature and conduct analysis of existing administrative data bases that have already been linked to employment records, and preferably, also to survey data. Answering the third question would involve phone consultation with experts to identify states that treat cyclers differently from others, followed by phone interviews or possibly site visits with state and local officials with a handful of states.

Implementation of TANF Sanctions

Examining variations in sanction policies, practices and outcomes across offices is the objective of this study. Previous studies have found wide variation in sanctioning rates across sites (e.g., Urban Change) and even across offices within a single city (e.g., Legislative Audit Bureau study of W-2 in Milwaukee). The study would examine whether the welfare offices with different sanction policies and practices also experience differences in outcomes including: welfare exits and re-entries, participation in work activities, and family and child income and well-being. This project draws largely on administrative data, as supplemented by site visits in order to improve our understanding of how sanctions are implemented. Sites will be selected both to maximize variation and to leverage the contractor's previous knowledge of sites and experience with the administrative data sets.

Wage Progression and the Dynamics of the Low-Wage Labor Market

Given the strong work-focus of TANF and time limits on the receipt of federal TANF assistance, policy makers are interested in understanding the potential for advancement in the labor market among low-wage workers. Do individuals in low-wage jobs have opportunities to progress economically? What helps low-wage workers advance? Do low-wage workers need to change jobs and/or employers to make wage gains? Do individuals need to take multiple jobs to make significant economic gains? While past research has examined the dynamics of low-wage employment, wage progression, and unemployment in the 1980s, the same in-depth research has not been conducted after the passage of PRWORA. To help fill this gap, this project takes the newly released 1996 Panel of the Survey of Income and Program Participation (SIPP) to examine wage progression among low-wage workers by tracking the dynamics of low-wage employment over a four-year period from 1996 to 1999. The upcoming public release of the full longitudinal 1996 SIPP panel offers a valuable, rich data resource for examining these critical wage and labor market issues.

Serving TANF and Low-Income Populations through Workforce Investment Act One-Stop Centers

Exploring the coordination of TANF and one-stop employment centers and the capacity of one-stops to serve the low-income and low-skilled are the focuses of this study. The Workforce Investment Act (WIA) one-stop system is likely to become the nexus of support for ever larger numbers of low-income and low-skilled people seeking employment and employment-related services. DOL Welfare-to-Work grant activities are winding down, possibly further increasing pressure on one-stops to fill the gap in serving the hardest-to-employ. Anecdotal evidence indicates that some employers may be reluctant to participate in one-stops that serve a largely low-income or welfare client base, and that some one-stops are hesitant to take on special-needs clients such as those with substance abuse histories or learning disabilities. This project will build on earlier studies and assess the progress made in coordinating TANF and one-stop services. It will explore the unique challenges involved in serving welfare and ex-welfare clients and the low-skilled in general through one-stop centers designed to serve people at all income and skill levels. The study will involve discussions with and visits to TANF agencies, Workforce Investment Boards, employers, welfare and non-welfare one-stop clients, and others in selected sites.

Evaluation of NY Welfare Intervention for Substance Abusing TANF/GA Recipients

This project will launch a comprehensive evaluation of several innovative programs in New York City for welfare recipients with substance abuse and mental and physical health barriers to employment. It offers an unusual opportunity to systematically advance the knowledge-base about programs for the "hardest to employ" in the context of one of the largest and most complex service delivery systems in the country. The New York City project is a new effort to provide intensive case management and drug treatment services to TANF and General Assistance (GA) clients with alcohol and other drug problems. It is part of an evolving national evaluation of programs across the country that are targeted to welfare recipients and other low-income individuals who face significant behavioral/health barriers to finding and maintaining employment. Core funding for the evaluation is being provided by the Department through the Administration for Children and Families' Employment Retention and Advancement (ERA) project. While ACF's investment will provide for basic evaluation, the contractor, Manpower Demonstration Research Corporation (MDRC), and the site are interested in doing more intensive, in person follow ups with clients and with treatment providers. By providing funds, ASPE would be able to work with other funders (including the National Institute on Alcohol Abuse and Alcoholism (NIAAA) and the National Institute on Drug Abuse (NIDA)) to design enhancements to the core study. These enhancements would include collecting more data through in person interviews, additional questions, and over a longer follow-up period. The study will add to the knowledge base in several important areas. It will assess whether findings from an implementation study in New Jersey can be replicated in a complex urban system. It also will assess whether the findings can be extended beyond increased engagement in treatment to effects on substance abuse.

Study of Child Support and TANF Interaction

Examining the interaction between child support status and TANF exit or reentry, and coordination between the TANF and child support agencies is the focus of this study. Child support is an important component of family self-sufficiency. This project examines how child support status (i.e., paternity established, order established, receipt of child support) interacts with TANF exit or reentry, and coordination between TANF and child support agencies. Are child support enforcement services targeted to TANF recipients around the point of TANF exit? Because of child support distribution policies, after TANF exit many families will begin to receive all of the child support paid on their behalf for the first time. Do families know how much child support they will receive when they leave TANF? Are former TANF recipients getting the support paid on their behalf as soon as they exit TANF? Some of these questions could be answered through state administrative data matching and others through site visits to states that demonstrate best practices. The Office of the Inspector General (OIG) has done some preliminary work in this area that would inform this project and assist with site selection.

The Federal Parent Locator Service (FPLS), TANF, and Medicaid Database

The Federal Parent Locator Service (FPLS) is primarily a national data system to help states locate non-custodial parents, alleged fathers, and custodial parents so they can establish and enforce child support obligations. PRWORA expanded the data available in this system to include information on child support cases and wage and employment data on nearly all workers. The law also authorized HHS to retain samples of this data and use them for research likely to contribute to achieving the purposes of titles IV-A (TANF) or IV-D (Child Support Enforcement). Conference Committee language from the FY 2002 Appropriation for the Department of Health and Human Services recommends that Welfare Outcomes research funds be used in part for the collection and use of "data administratively linking that National Database of New Hires [which is part of the FPLS], other child support enforcement data, TANF, and Medicaid records together." The design and implementation of the Federal Parent Locator Service (FPLS), TANF and Medicaid data warehouse is a joint ASPE, ACF, and CMS project to monitor the progress of welfare reform. Child support data, TANF client data, and Medicaid client data will be merged into a research data warehouse to assist with federal and state program oversight, research, statistical reporting, policy and evaluation for the populations served by TANF, Child Support and Medicaid. One major advantage of the data base is that it will allow us to understand program participation and interactions over time. For example, information on child support, Medicaid participation, and earnings will help us understand individual and state differences in movement on and off TANF cash assistance and develop policies that appear to increase the potential for family self-sufficiency. Some of the difficulties in developing a state and nationally representative research data base that combines data from the various sources include: the various data bases are not programmed in compatible languages; the data cover different time frames; individuals have to be linked into related family units; and sampling schemes have to be constructed for each state. All of the program specific data bases are being maintained and/or were developed with contractor support. ASPE's FY 2002 contribution will help insure that we can fully respond to the Congressional intent of ensuring that other child support data is integrated into the research database to the full extent of their statutory authorizations.

Private Employers and Welfare Reform: Current Practices and their Implications for TANF

Over 10,000 companies participate in the Welfare-to-Work Partnership, and an array of companies are known for their ingenuity in recruiting, training, retaining and promoting TANF recipients. A number of case studies and pilot projects have examined innovative practices for successfully employing TANF recipients, and several recent surveys have examined particular industries and states. Nevertheless, we lack a national, overall portrait of the prevalence, variety, and limitations of such practices, such that basic questions have not been answered. Questions include what employers who rely on low-skill, entry-level workers do to successfully employ TANF recipients; why they do it; why many do not do it; what would enable them to increase and improve their efforts; and which segments of the TANF population present the greatest challenges. In order to begin to answer these questions, and to identify options for answering them more fully, this project will synthesize recent research and identify gaps in knowledge by reviewing existing studies, survey instruments and data sources, and consulting with experts and stakeholders. The final report will summarize the current state of knowledge in this area and its implications for TANF and for private employers, and will identify options for advancing this knowledge in order to observe the sensitivity of employer practices to changing labor market conditions, by conducting a followup survey or by revisiting the other data sources at a later time. In addition to having fundamental implications for TANF policy and administration, findings would allow the Department to better focus its technical assistance for private employers, labor market intermediaries and program administrators.

Welfare Closed Cases - job entry and earnings: Data Match TANF - FPLS: Data Matching to Assess Welfare Outcomes for the 50 States and the District of Columbia

This project will utilize five federally administrated databases to assess outcomes for former TANF recipients. The project will assess six outcomes: four are related to employment - job entry, earnings, earnings gains, job retention. By linking the various databases: ACF's TANF Database of closed TANF cases, National Directory of New Hires database (NDNH), Quarterly Earnings database, and possibly the Child Support Case Registry database, this project will provide state-by-state estimates of important welfare outcomes.

Project on Devolution and Urban Change

Our continuing support leverages data from a study of welfare reform in four large urban areas. The Urban Change project brings together data from an unusually wide array of sources: longitudinal administrative data for all families receiving AFDC/TANF or food stamps dating back to 1992, survey data, an implementation study, neighborhood indicators, an institutional study focusing on local service providers, and an ethnographic study of a limited number of families. In 2002-2003, MDRC will be primarily focused on writing the cross-component synthesis reports for Miami, Los Angeles and Philadelphia. (The Cleveland report is due out in Spring 2002.) Our contribution allows us to leverage the large amount of private money that has been invested in data collection toward questions of interest to HHS. Topics of interest include: the role of faith-based organizations in providing services in these communities, the use of sanctions as a mechanism to promote participation, and the characteristics of families that cycle on and off welfare.

Partners for Fragile Families Evaluation

Targeting young, never-married, non-custodial parents who may face obstacles to employment, the Partners for Fragile Families (PFF) demonstrations are designed to help the fathers in such families work with mothers in sharing legal, financial, and emotional responsibilities of parenthood. PFF demonstration projects in nine states were approved under HHS waivers and are currently underway. The projects test new ways for state-run child support enforcement programs and community-based organizations to work together to help young fathers obtain employment, make child support payments and learn parenting skills; and to help parents build stronger partnerships. The evaluation, which is being conducted by the Urban Institute, has three broad purposes: to increase knowledge about systems change; to build knowledge about program operations and delivery of services to fragile families; and to describe client behavior. Process and outcome evaluations are being conducted through interviews with service providers, including child support enforcement, community-based organizations, and partner agencies; and by analyzing client data and follow-up surveys. Case studies also are being conducted. ASPE and OCSE are sharing the cost of this five-year study; the current funding is year three of ASPE's four year funding commitment.

Analysis of Topics Related to Marriage: Program Supports for Married and Two-Parent Families

This project would seek to fill several important gaps in our knowledge of how state policies affect marriage and family formation outcomes. A key question in understanding the role of state policy in promoting two-parent and married-parent families is the degree to which these families access supports through public assistance programs, and the factors that might affect this. This project will examine participation of two-parent and married-parent families in TANF, food stamps, and Medicaid. This analysis would lend important support to an agenda undertaken by ACF that focuses primarily on developing marriage-strengthening programs, and we would work closely with ACF in implementing this analysis.

Child Outcomes in State Welfare Evaluation

With other federal and private funders, ASPE and ACF are working with states to improve measurement of child health and well-being outcomes in state welfare evaluations. Five states are using a common protocol to add child outcome measures to their welfare reform evaluations. Continuation funding is enabling states and their evaluators to receive research technical assistance on collecting survey data using the common core of instruments, using administrative data sources, and developing and coordinating data analysis and reporting strategies. The focus of the current phase of work is the production of a synthesis of the findings from the state evaluations. The research technical assistance is provided by the NICHD Research Network on Family and Child Well-Being under Child Trends' Leadership.

State Use and Reporting of TANF Administrative Data

The purpose of the project is to work with states to improve their capacity to use the administrative data they gather from the TANF program for program management and performance measurement. While every state and locality operating a TANF assistance program is mandated to collect a wide range of administrative data for program operations and mandated reporting requirements, many administrators fail to reap the full range of benefits from these data. The contractor, working closely with states through a series of working group meetings and other interactions, will develop a model or framework for using TANF administrative data for program management. This improved data capacity could then be "spun out" to other states that do not participate directly in the project. As a positive by-product of the project, there will be an improvement in the quality of the TANF administrative data collected by the states and reported to HHS. The final products of the study will include a CD-ROM that states can use to make their administrative data more usable, and a final report detailing the outcomes of the working group meetings.

Building Administrative Data Capacity Around Low-Wage Employment Outcomes

With the tremendous success achieved by welfare reform in moving large numbers of former recipients into jobs, attention is now turning to their ability to advance in the labor market, and the factors that affect those transitions. This study would analyze a unique administrative data base and produce a report describing the factors that are most likely to lead to positive employment transitions, or hinder those transitions, for low-wage workers and former public assistance recipients. These factors would include both personal characteristics of the workers, but also employer characteristics such as industry, firm size, turnover rates, capital investment, productivity, and location. With respect to location the analysis would show, for example, the nature of employment transitions in rural, urban or high poverty areas, or whether positive transitions are more likely among workers willing to commute from their central city neighborhoods to jobs in the suburbs. Existing studies on employment transitions based on survey data are limited in several ways. For example, CPS data can provide a current snapshot of the labor market, but are not useful for following workers' progress over time. Longitudinal surveys can follow workers over time, but sample sizes are often too small to produce reliable estimates for important subgroups, and are often outdated by a few years or more - a serious problem given frequently changing social policy and labor markets. Furthermore, most national surveys contain only information reported by the individual - they can analyze the effect of personal characteristics on employment, but can say little about what types of firms are more likely to generate positive transitions. This project using the Longitudinal Employer Household Dynamics data (LEHD) provides a unique opportunity to better understand the factors that affect job transitions and wage progression. The data are housed at the Census Bureau and are not publically available. The LEHD contains several administrative data files including records from the Census Bureau, UI wage records, and BLS establishment (company-based) records. The data cover the nine states of California, Illinois, Maryland, Florida, Texas, North Carolina, New Jersey, Pennsylvania and Minnesota, accounting for 60 percent of the U.S. labor market. Analysis using the LEHD has several advantages that address gaps described above. First, they are thought to be more accurate than survey data. Second, they contain information through 2001, capturing important changes in the U.S. labor market. Third, the LEHD contain millions of records and can produce reliable estimates for detailed subgroups. Fourth, the data are linked longitudinally to follow workers and the employers they work for over a period of about 10 years. Finally, they contain records not only for workers, but also for employers. The records are matched together using UI wage files, which contain both employee and employer identification numbers for nearly all workers.

Aggregated Data by State on the Web

Under this project, ASPE will work with ACF to enhance the web-based reporting system being developed for state reporting of aggregated caseload and expenditure data for the TANF program. The system will make recent data available more quickly and in a widely accessible form on the Internet. Data will be protected through a secure web site with passwords and other measures to ensure data integrity and security. The current web-based reporting system would be enhanced to produce national totals, graphs of time trends, and other tables and figures. The data will thus be more useful for program management and monitoring for both the states and federal government, especially since data would be available on a more timely basis than under current arrangements. In essence, the goal is to developing an enhanced, electronic version of the "flash report" that was used to provide monthly caseloads and expenditure data under the old AFDC program.

Rallying the Armies of Compassion

State and Local Contracting for Social Services Under Charitable Choice

Recent studies have identified considerable variation in state and local government officials' understanding and implementation of the Charitable Choice (CC) provisions of the 1996 welfare reform law, the Personal Responsibility and Work Opportunity Reconciliation Act. However, there has yet to be a close examination of the ways in which CC has affected social service procurement policies, practices, and decisions at the state and local level. This project will examine the varying ways in which CC is interpreted and implemented by state and local government officials in the four program areas affected - the TANF, Community Services Block Grant, Substance Abuse and Mental Health Services Administration, and the Department of Labor's Welfare-to-Work grants programs. In particular, the project will focus on those government officials - typically staff in state and local contracts offices - who serve as "gatekeepers" in the award and management of contracts and grants. The study also will examine those 19 states that have appointed "faith community liaisons" to oversee efforts to reach out to faith-based organizations and encourage them to enter into financial and other types of collaborations with the state. This focus will help identify the impact of faith community liaisons on state contracting decisions and outcomes.

Ongoing Support to the New Immigrant Survey

Study of the status of children in newly arriving immigrant families is ensured through our support for the New Immigrant Survey. The National Institutes of Health (NICHD and NIA), along with ASPE, the INS, and the National Science Foundation, have funded the New Immigrant Survey, a large, longitudinal survey of newly arriving immigrants beginning in 2002. ASPE's contribution will continue to help ensure that comprehensive and relevant data are collected and analyzed about program utilization, health and economic status, and other characteristics over time among newly arriving immigrant families in different states. In particular, ASPE's support will ensure that the study focuses on how children in these families are faring.

Support for the NGA Re-Entry Leadership Project

Effecting state systems change for prisoner re-entry through strategic planning and implementation is the objective of this project. Under this project the NGA would select a number of states to receive funding and technical assistance around the development and implementation of a strategic plan for prisoner re-entry. Plans would have to include coordination around health and human services issues for the returning prisoners and their children and families.

Study of Children in TANF Child-Only Cases with Relative Caregivers

The proportion of child-only TANF cases has continued to grow, representing over a third of the total TANF caseload by FY 2000 (782,000 cases). While the number of child-only cases reached its peak in FY1996, the subsequent decline of the child-only caseload was at a slower rate than the overall caseload. These are cases where only the children in a family/household are receiving TANF benefits, and the adults in the household are not eligible for TANF. The purposes of this study are to: (1) systematically review and analyze existing data and studies to describe the service needs and well-being of children who are receiving TANF assistance and being cared for by non-parent relatives who are ineligible for TANF; (2) describe whether states have established any goals or outcomes related to children in these child-only cases, and if so, what strategies are being employed to achieve those outcomes; and (3) better understand how such cases are handled at the local level, and the degree to which the service needs and the well-being of children in these families are taken into account.

Crosscutting Issues

Support for the Research Forum

ASPE and ACF support the maintenance and continued development of the database of the Research Forum on Children, Families and The New Federalism, hosted at the National Center for Children in Poverty. This database contains information on the large number of research projects on welfare reform, children and families. This will help ensure that researchers are aware of existing projects, and that new efforts are non-duplicative.

Support for the Panel Study of Income Dynamics

This project continues ASPE's on-going core support for the Panel Study of Income Dynamics (PSID). ASPE funds provide partial support for the continued collection and processing of longitudinal data relevant to research on economic factors and income support mechanisms, health, fertility, medical care, and disability affecting the poor and the elderly. ASPE funds will continue to support an expansion of the set of welfare related questions to assess the entry effects of recent reforms.

IV. Update on Continuing Projects Funded by the FY 1998, FY 1999, FY 2000 and FY 2001 Appropriations to Study Welfare Outcomes

Competitive Grant Programs

Continuation of 1998 Grants to States and Localities to Study Welfare Outcomes (1998 and 1999)

Some of the ASPE leavers grantees ⁽¹⁾ have findings that were not released in time to be included in the synthesis report (see Chapter II), or have not yet released a final report. These reports can be separated into two categories: those releasing their final reports from their studies funded in FY 1998, and those grantees that received additional funding in FY 1999 to extend the studies and administer a second or third wave of interviews, allowing analysis of longer-term outcomes for former recipients.

Studies Funded in FY 1998 (Wisconsin, Los Angeles County (CA), New York)

In November 2001, the Wisconsin Department of Workforce Development released its final report on individuals and families who left the Wisconsin Works (W-2) program between April and December 1998. According to findings from the survey, about 58 percent of leavers were employed at the time they were interviewed, and 82 percent had been employed at some point since exit. Those W-2 leavers who were employed worked for an average of just over 35 hours per week, at wages averaging about $7.95 per hour. Four out of five leavers were receiving Medicaid, and 63 percent were receiving food stamps, in the fourth quarter following their exit from W-2. The majority of respondents (86 percent) with children under 13 reported having some child care arrangements while at work, and less than a third reported that a child care problem had interfered with their work efforts. Further, among survey respondents, 60 percent reported feeling better about themselves since leaving W-2, while only 7 percent reported feeling worse about themselves.

The Los Angeles County (CA) project was conducted by the Manpower Demonstration Research Corporation (MDRC), and provided an addition to MDRC's Urban Change project with a focus on leavers. Cohorts were drawn from the third calendar quarters of 1996 and 1998. The project included analysis of ten years of full population administrative data developed for the Urban Change project, as well as a small mixed-mode sample survey drawn from a cohort surveyed between July and December 1999. This project and the Cuyahoga County, OH project, also by MDRC, effectively made up a two-site study that has allowed for comparisons while controlling for study design. Los Angeles also received funding from the Department of Housing and Urban Development to draw a sample of welfare leavers who received housing assistance, both through public housing and Section 8. MDRC and Los Angeles County have submitted a draft final report for this project to ASPE and it is expected to be released in Summer 2002.

New York's two-year study included both TANF exiters and individuals under sanction, whether or not they left assistance. The goals of the project included: determining the frequency of outcomes such as employment, job retention, use of transitional assistance and returns to assistance; identifying barriers to self-sufficiency; examining the effectiveness of sanction policies in changing behavior; and developing a longitudinal tracking capacity for welfare outcomes in New York City. The state used full population administrative data from TANF, SSI, food stamps, Medicaid, foster care, child support, and wage records to analyze cohorts from the first quarter of 1997 and the first or second quarter of 1999. The FY 1999 cohort involved a 900-case survey in the first quarter of 2000. The state released an interim report in July 1999 with administrative data findings on the first cohort of leavers. A revised version of this report was released in December 1999. A final report for this project has been submitted to ASPE and is expected to be released by the state in Summer 2002.

Studies Receiving FY99 Continuation Funding (Arizona, San Mateo County (CA), Missouri)

Three of the FY 1998 grantees - Arizona, Missouri, and a consortium of San Mateo, Santa Cruz, and Santa Clara counties - received additional funding in FY 1999 to extend their studies and administer a second or third wave of interviews, allowing analysis of longer-term outcomes for former recipients.

The Arizona study followed a large group of individuals who left cash assistance in the first three months of 1998 for two years. The state released its final report on the second year follow-up of leavers in October 2001. It found that many who left the rolls were self-sufficient after two years. Results from administrative data showed that fewer leavers utilized state services during the second year of the study compared the to the first year, and reported wages of employed leavers increased by 25 percent. However, those who were utilizing state services in the second year averaged more usage than they did in the first year. Surveys of the leavers showed that a majority of respondents felt their "general conditions" improved during both years of the study, though some particular indicators worsened in the second year.

Funding in FY 1999 allowed the SPHERE Institute to add a third survey at 18 months after exit and extend administrative data tracking of a cohort of individuals and families who left welfare in the fourth quarter of 1998 in San Mateo, Santa Clara, and Santa Cruz Counties, CA. The final report, released in November 2001, showed that conditions generally improved for leavers and informally diverted families over the 18-month follow-up period. In addition, trends in the awareness and use of transitional assistance were positive, but indicated room for future improvement. After 18 months, two-parent leaver families were somewhat better off than one-parent leaver families and families informally diverted from welfare.

The third grantee to receive continuation funding from ASPE in FY 1999, Missouri, has used the funding to more extensively track of their study cohort of those who left welfare in the fourth quarter of 1997. A final report on this follow-up study is expected from Missouri in Summer 2002.

Estimated Completion Date: varies by project

Grants to States and Localities to Study Welfare Reform Outcomes, with an Emphasis on Diversion (1999)

One of the Congress's major objectives in providing welfare outcomes money to ASPE over the last several years is to measure outcomes for a broad population of low-income families, welfare recipients, former recipients, potential recipients, and other special populations affected by state TANF policies, including diversion practices ⁽²⁾. To this end, ASPE issued a request for applications from states and large counties in April 1999 with an emphasis on the study of applicants and potential applicants to the TANF program. ASPE awarded seven grants under this announcement, six of which specifically support state efforts to gather a variety of information about individuals and their families who apply to TANF, including those who are formally or informally diverted. In addition, several of the leavers studies funded in FY 1998 had significant applicant components to their projects.

ASPE is particularly interested in learning about the degree to which TANF applicants receive, or are aware of their potential eligibility for, Medicaid, food stamps, and other programs and services that are important in helping low-income families make a successful transition to work. Below are summaries of the grants provided to states and large counties in fiscal years 1998 and 1999 with a particular emphasis on TANF applicants and diversion from TANF. "Diversion" in this context is not limited to participation in formal diversion programs, but also includes "informal" divertees. These are usually defined as individuals who began the application process but were either deemed ineligible for non-monetary reasons, withdrew voluntarily after completing the process, or failed to complete the process for some other reason. A synthesis of available findings is included in Appendix A.

New York (1999)

New York, which also received a FY 1998 leavers grant from ASPE, included divertees, all other denials, and entrants in their sample for this study. Their analysis has focused on comparing TANF applicants who were diverted with those who received cash assistance. Twenty-one local districts participated in the study, including New York City and other sites ranging from large urban to rural areas. In most districts, state researchers have used administrative data to track a March 2000 sample of divertees, denials, and entrants for 12 months after the application. The sample was drawn through intercept interviews with TANF applicants in each of the local districts using a methodology that allowed New York to include individuals who entered the TANF office with the intent to apply but who did not submit written applications. The state's contractor, ORC Macro, administered the survey to the sample of 864 families, evenly split between diverted applicants and entrants. The state is finishing up its analysis of the data and writing the final report.

Estimated Completion Date: Summer 2002

Washington (1999)

Following up on the leavers grant that they received in FY 1998, Washington studied formal and informal divertees and entrants. The state compared the experiences of individuals who participated in the state's Diversion Cash Assistance program, those who entered TANF, and those who were diverted and received assistance from neither program. They have analyzed administrative data for the full populations of each of these groups from the fourth quarters of 1997, 1998, and 1999, including data from up to 12 months prior to and 12 months after the selection quarter. The state has nearly completed administrative data collection.

The state also surveyed individuals who applied for TANF or Diversion Cash Assistance between July and October of 1999. The survey was administered between four and eight months after the time of application. The draft final report has been submitted to ASPE and is currently undergoing revisions within the state.

Estimated Completion Date: Summer 2002

Wisconsin (1998 and 2000)

This study of individuals applying for Wisconsin Works (W-2) assistance in Milwaukee has been undertaken by the Institute for Research on Poverty (IRP) at the University of Wisconsin at Madison. This portion of the study focuses on three subgroups of applicants: those who request assistance and subsequently participate in the W-2 program, those who request assistance but are determined to be ineligible for program participation, and those who request assistance, appear to be eligible, but do not participate in W-2. A six-month cohort of applicants is being tracked through a combination of linked administrative data (e.g., public assistance, quarterly earnings, child support, foster care, and mental health data) and three waves of surveys, the third being funded by ASPE in FY 2000. A report detailing findings from Wave 1 of the survey, administered at the time of application, was released in July 2001, and a link to the report is posted on the ASPE web site. A report on Wave 2 of the survey is expected in Summer 2002. The third wave of the survey is currently in the field; a descriptive report on this survey is expected in late fall 2002.

Estimated Completion Date: Winter 2003

Grants to States and Localities to Enhance Studies of Welfare-Related Outcomes

The purpose of these grants is to enhance state-specific surveys of populations affected by welfare reform, by expanding or improving data collection activities. Grants to states are being used, for example, to add additional survey waves to measure longer-term outcomes, collect data to support greater sub-group analyses, and gather more detailed information on non-respondents. To be eligible, states had to have an existing survey that had been administered at least once, so the grants can facilitate real improvements, without paying for basic startup costs. Survey efforts needed to fill an important knowledge gap that could not be filled with states' existing data. The data cover a variety of welfare reform outcomes, such as measures of family hardship and well-being, barriers to employment, poverty status, and utilization of support programs. The surveys focus on various subsets of the low-income population including long-term welfare recipients, child-only cases, former recipients, potential recipients, welfare leavers with little or no reported income, and other special populations affected by state TANF policies. The funded proposals include:

Alameda County, CA (2000)

Alameda County builds on its existing survey of current TANF recipients and TANF leavers who were interviewed at baseline and at 15 months. Under this project, researchers are conducting a 27- month follow-up survey and maintain the same detailed focus on health barriers to employment, including issues related to mental health and substance abuse. Researchers from the Public Health Institute are conducting in-person interviews and linking responses to the state's administrative data systems to gain information on demographics, earnings and program participation. The data gathering has been completed and analysis of the data are underway.

Estimated Completion Date: Summer 2002

Missouri (2000)

Missouri is building on its ASPE-funded study of former TANF recipients who left the rolls in 1996 and 1997, and is adding a cohort of recipients who have remained on TANF for at least 36 months. The study seeks to characterize and contrast the self-sufficiency outcomes and barriers for current and former TANF recipients, and to identify which factors are most predictive of successfully transitioning off welfare, as well as those characteristics most predictive of exhausting the time limit. Survey data is linked with administrative data on TANF, food stamps, child care, Medicaid, and some community-based assistance. Thus far, the contractor for the study, the Midwest Research Institute, has selected the sample of long-term recipients and has fielded the questionnaire on barriers to work. Data analysis is ongoing.

Estimated Completion Date: Summer 2002

San Mateo County, CA (2000)

This study uses both administrative and survey data to study child-only cases, including cases that have left TANF and those that remain on the rolls. The study seeks to better understand the characteristics and outcomes of these families, many of whom are headed by immigrant parents. The study takes place in San Mateo and Santa Clara counties, and is being conducted by the SPHERE Institute. Their study also draws on administrative data from county case files, wage records and Medicaid eligibility data. Thus far, researchers have developed their research design, coordinated with various county offices, fielded the survey instrument, and constructed the administrative data files. Data analysis is ongoing.

Estimated Completion Date: Summer 2002

Wisconsin (2000)

This project adds a third wave of interviews to the Institute for Research on Poverty's existing study of a cohort of TANF applicants in Milwaukee, Wisconsin. The findings from this wave will reflect outcomes for this cohort approximately two years after the baseline data were collected. The study is based on a sample of applicants, and the survey will contain results for those who have entered and subsequently left TANF, those still receiving TANF, and some who never received TANF. Adding a third wave to the applicant survey will support analysis of a significant number of cases who have reached the time limit in Wisconsin, and a significant number of cases who have cycled off and on the rolls. The researchers will examine a large number of outcomes related to employment, well-being and program participation. They anticipate that of the 1200 respondents from Wave 1, approximately 900 will complete interviews for Wave 3. The Wave 3 survey is currently in the field, and a descriptive report should be available in late fall.

Estimated Completion Date: Fall 2002

South Carolina Welfare Outcomes Grant (1998, 2000 and 2001)

This project continued ASPE's support of a multi-year effort by South Carolina's Office of Budget and Control Board's Office of Research and Statistics to link administrative data and additional data from surveys of former welfare recipients. The funds provided through an ACF cooperative agreement allowed South Carolina to continue its contract for the expansion of the follow-up studies. Final reports on the first year follow-up on welfare leavers have been released ⁽³⁾, as well as first- and second-year reports focusing on those families diverted from cash assistance. A final 36-month follow-up report, summarizing previous reports on both leavers and divertees, is expected in the fall.

Estimated Completion Date: Fall 2002

State Studies of TANF Caseload (2001)

This project funds six states - California, Colorado, the District of Columbia, Maryland, Missouri, and South Carolina - to study the characteristics of their TANF caseloads. Particular attention will be given to the personal, family and community factors that may present barriers to employment. States will survey a sample of the current caseload by telephone, in order to gather information in such areas as physical and mental health, disability, substance abuse, and domestic violence, as well as information on demographics, work experience and income. To improve the comparability of survey data, states will use a standardized survey instrument (see project description below). Each state will supplement this common instrument with additional survey questions according to their particular interests, and will augment their survey data with administrative data in order to examine changes over time in recipient characteristics and program utilization. States will compare characteristics of recipients and barriers to employment across various subgroups, such as short-term and long-term recipients or employed and non-employed recipients.

Estimated Completion Date: Summer 2003

Researcher Initiated Grants on Welfare Outcomes (1999, 2000 and 2001)

ASPE continued to support researcher-initiated proposals to study important questions related to the outcomes of welfare reform in FY 2001. Background on the FY 1999 and 2000 grant programs is included in Chapter II. In FY 2001, ASPE awarded eleven grants in support of policy-relevant research to broaden our understanding of welfare reform outcomes. Nine ASPE grantees received approximately $1.2 million; the Administration for Children and Families funded two grants totaling over $200,000. The issues being addressed under the 2001 grants include barriers to service delivery, particularly for special populations; family formation; child and youth outcomes; maternal employment; the low-wage labor market; family economic security; measurement of welfare utilization; and effects of TANF time limits. On-going projects are described below. http://aspe.hhs.gov/hsp/

Manpower Demonstration Research Corporation (MDRC): An Analysis of Caseload Composition and the Non-Working Welfare Leavers (2000)

Researchers are examining three groups of low-income populations (those who leave welfare for work, those who remain on the welfare rolls, and non-working welfare leavers) to address the following questions: (1) In what ways are the families who remain on welfare different than the ones who have left? (2) What are the characteristics and circumstances of people who leave welfare and are not working? The project uses data from seven evaluation studies of welfare programs (six used random assignment) conducted by MDRC.

Estimated Completion Date: Summer 2002

Urban Institute: The Link Between Marriage and Low-Income Family Well-Being (2000) (formerly How Important is Marriage to Low-Income Family Well-Being?)

This project, which began in September 2000, is examining the interactions between family formation status and economic well-being to better understand the extent to which marriage is a protective factor against economic hardship, particularly among the disadvantaged population. The study examines various types of family formation, including single, married and cohabiting parents, and it looks at measures of poverty as well as material hardship. The project is producing four reports, based on various data sources. The first report has been completed (see Chapter II).

The second report is based on the National Longitudinal Survey of Youth, and explored in more depth the degree to which the protective role of marriage differs for more and less disadvantaged populations. Preliminary findings indicate that even among disadvantaged groups, marriage continues to significantly reduce the likelihood of being in poverty, although by some measures the effects were somewhat smaller among the disadvantaged population. The report, Married and Unmarried Parenthood and the Economic Well-Being of Families: A Dynamic Analysis of a Recent Cohort, is expected to be completed by Summer 2002.

The third report is based on data from the Survey of Income and Program Participation and focused more on measures of material hardship, including lack of phone service, housing inadequacy, food insecurity and utility shut-offs. Preliminary results indicate that, even after controlling for differences in income, married families tended to experience less material hardship than either cohabiting or single-parent families, and that this may be in part because married families have greater access to help from families, friends and communities. The report, The Relationship Between Marriage and Other Family Structures and the Material Hardship Experienced by Households with Children: Evidence from the Survey of Income and Program Participation, is expected to be completed by Summer 2002.

The final report will be a literature review of various studies that analyze the returns to marriage, particularly for disadvantaged families.

Estimated Completion Date: Summer 2002

Columbia University: Fragile Families and Welfare Reform (joint with ACF) (2000)

This study will describe the conditions and capabilities of vulnerable mothers and fathers in the first few years following enactment of PRWORA and begin an evaluation of the impact of TANF and child support policies. Specifically, researchers will document the composition of the actual and eligible welfare caseload, how unwed mothers are packaging various forms of support and government programs, and how well families are doing as a result of individual efforts and social policies. Researchers also will conduct subgroup analyses on teenage parents and immigrants. The Fragile Families and Child Well-being Study, a random sample of new unmarried mothers and fathers in 20 large cities across the United States, will be utilized. The first report under this grant, a baseline report that provides demographic and descriptive statistics as well as a comparison of mothers who are receiving public assistance to those who are not, is currently under review and is expected to be released in Summer 2002. Upcoming reports will use 12-month follow-up data to analyze length of assistance receipt, use of public and private supports, employment, changes in circumstances, and city and policy variations.

Estimated Completion Date: Summer 2003

UCLA/RAND: A Proposal to Examine the Reporting of Welfare Benefits in the SIPP Using Matched Administrative Records in California (joint with ACF) (2000)

This two-year study examines the accuracy of self-reports of program participation in survey data. In particular, researchers are comparing self-reported program participation among Californians interviewed in the Survey of Income and Program Participation (SIPP) with California administrative files of program participation for the same individuals. Researchers will document the degree of misreporting in a variety of programs, including AFDC/TANF, Medicaid, and food stamps, and investigate the implications of misreporting for conclusions about the dynamics of welfare participation.

Estimated Completion Date: September 2002

UCLA/University of Wisconsin: The Effects of the Work Pays Demonstration, EITC Expansions and the Business Cycle on the Labor Market Behavior of the California Caseload (2000)

This project will examine the effect of: 1) welfare changes, 2) the 1990 and 1993 expansions of the EITC, and 3) changes in the business cycle on three specific issues concerning the California welfare population. These issues include: 1) how do these factors contribute to the economic well-being of families; 2) how do they affect labor market and transfer program participation; and 3) how do they affect employment changes and earnings trajectories? The project will use California administrative data drawn from the welfare, unemployment insurance, and tax systems. This project has experienced delays in negotiating and securing the necessary data sharing agreements with the various agencies.

Estimated Completion Date: Fall 2002

University of Wisconsin: Toward Understanding the Longitudinal Health Insurance and Food Stamp Status of Short- and Long-Term Welfare "Leavers" (1999)

Subsidized medical insurance and food purchases through the Medicaid and Food Stamp programs potentially improve the health and economic well-being of low-income people, but only if eligible participants receive program benefits. Reports of low take-up rates and decreases in food stamp and Medicaid participation rates following passage of welfare reform legislation in 1996 raised concerns about the health care coverage and nutritional status of low-income people, particularly former recipients of cash welfare. This project explores the long-term utilization of food stamp and Medicaid benefits for two cohorts of welfare recipients who left cash assistance in Wisconsin. The first cohort consists of those who left cash welfare in 1995 (under welfare reform waivers); the second cohort consists of those who left welfare two years later, in 1997. The paper estimates both initial take-up rates (i.e., participation rates among those eligible immediately after exit from cash welfare) and participation rates for extended periods after leaving assistance. Preliminary findings show that participation eroded over time after exiting cash welfare in similar patterns for the two cohorts, but the overall level of take-up was substantially higher for the latter (1997) cohort of leavers.

Estimated Completion Date: Summer 2002

Child Trends: Maternal Employment and Adolescent Functioning During the Early Implementation of PRWORA: The Role of Parent-Adolescent Relationships and Adolescent Employment (2001)

This project will examine how the lives of adolescents in single-mother, low-income families who have received welfare during the early stages of PRWORA implementation are affected by their mothers' employment. Using data from the National Longitudinal Survey of Youth 1997 cohort, the project will investigate how maternal employment in these families relates to key aspects of adolescents' lives, including their autonomy, their relationships with their parents, and their own employment. This project will examine how these changes in adolescents' lives relate to their well-being. It also will investigate how relationships among these variables differ for various subgroups of adolescents, as well as whether they are indicative of a process occurring in single-mother, low-income families more generally, or are specific to families receiving welfare.

Estimated Completion Date: Fall 2003

Johns Hopkins University: A Study of TANF Non-Entrants (2001)

This project will use data from the Three-City Study, a study of low-income families in Boston, Chicago, and San Antonio. It will examine the employment rates, earnings, income, and other measures of well-being for families who have not applied for welfare or who have applied and been rejected or diverted. Specific questions from the Three-City Study survey regarding experiences with application and diversion will be used to document how common these events are. Longitudinal analyses will be conducted using two waves of survey data, examining how families fare after having been diverted or not having applied. The survey data will be supplemented by ethnographic, participant-observation data on families who have not entered TANF or who have applied and been rejected or diverted.

Estimated Completion Date: Fall 2003

New Jersey Department of Human Services: Are Former Welfare Recipients Likely to be Eligible for UI? Evidence from New Jersey (2001)

This project will examine the extent to which former welfare recipients who leave welfare for work are likely to be eligible for Unemployment Insurance (UI). Pre-welfare reform data has indicated that many prior welfare recipients who leave work are ineligible for UI. Using a set of administrative and survey data from the ongoing Work First New Jersey (WFNJ) evaluation, this study will focus on three main questions, in a post-welfare reform sample: 1) How many former welfare recipients who leave welfare for work are likely to be eligible for UI? 2) What are the actual UI experiences of welfare recipients who lose their jobs? and 3) What is the total extent of the safety net of these individuals, including welfare and UI amounts and durations?

Estimated Completion Date: Fall 2003

The SPHERE Institute: Implications for the Design of Federal Time Limit Rules: Who Will Hit TANF Time Limits in California? (2001)

This study is designed to: 1) estimate the number and proportion of California TANF cases likely to hit five-year federal time limits; 2) determine how families likely to hit the time limit differ from other TANF families; 3) estimate the number of working families likely to hit time limits; 4) determine how working families likely to hit time limits differ from working families that leave TANF, with a particular focus on differences in employment characteristics; and 5) consider the implications of these findings for the design of federal time limit rules, focusing on assessment of the current 20 percent hardship exemption and alternatives that account for employment behavior. Linked California statewide administrative data sources measuring TANF participation, recipient demographics and earnings, and employer characteristics will be used.

Estimated Completion Date: Summer 2002

University of Texas, Austin: Changes in Employment, Welfare Receipt, and Income as Predictors of Family and Child Care Contexts and Youth Risk and Resilient Behavior (2001)

This study will use longitudinal data from the New Hope evaluation (in Milwaukee, Wisconsin), a work-based, anti-poverty program, to investigate the relations of parents' employment, income, and welfare receipt to family organization, children's activities, and development of children and youth during the course of welfare reform in the late 1990's. The questions to be examined are: 1) Do the levels and changes in parents' employment, family income, and receipt of welfare for the period from 1995-2001 predict family organization, children's participation in child care and out-of school activities, and youth risk and resilient behavior? 2) Are effects of employment, income, and welfare on children and youth mediated by changes in family organization or child activities? 3) Are these relations different for children in middle childhood versus adolescents?

Estimated Completion Date: Fall 2003

University of Washington: State Policy Variation and the Economic Security of Low-Income Families in the Wake of Welfare Reform (2001)

The three principal investigators have developed a database of state policy indicators for 11 program areas, in each of six years. Programs in the database capture the components of a package of supportive assistance available to low-income families with children, including some transfer, tax, in-kind, and work support policies. The study will address two questions: 1) How did the packages of "family support policies" provided by states change between 1994 and 1999/2000? and 2) What were the consequences for the economic well-being of families with children? For this study, time series data will be used to describe state policy developments from 1994 to 1999/2000, and innovative analytic techniques will make it possible to estimate the effects of state-level policies on families' resources, resource packages, and experience of hardship.

Estimated Completion Date: Fall 2003

The Urban Institute: The Interactions of Workers and Firms In the Low-Wage Labor Market: Implications for Welfare Reform (2001)

This study will examine the determinants of successful employment and wage outcomes among low-income workers. Using longitudinal administrative data on workers and the firms that employ them, researchers will address the following questions: 1) For those low-income workers who ultimately succeed by gaining higher earnings, to what extent are they generated by returns to job retention as opposed to mobility across firms? 2) Do these pathways to higher earnings differ by age, gender, race/ethnicity, geographic location, and time period for the disadvantaged workers in question? and 3) What characteristics of employers contribute the most to these successful outcomes? Are there particular matches between worker and firm characteristics that are most helpful for achieving worker success? Are there particular transition patterns across employers that are most likely to generate such success? Researchers will use the Census Bureau's Longitudinal Employer-Household Dynamics data set, which includes an employer-worker data set based on state-level unemployment insurance data.

Estimated Completion Date: Fall 2003

Washington University: Children with Disabilities: Implications for the Transition from Welfare to Work (2001)

This project will examine aspects of the economic and psycho-social impacts a child with disabilities has on low-income and welfare-recipient, single-mother families. The project will document the incidence of child disability among low-income and welfare-recipient, single-mother families and examine the characteristics of families with disabled children. The impact that children with disabilities have on their families' ability to exit welfare and poverty, as well as the impact on adult labor supply in these families over time, will be examined.

Estimated Completion Date: Fall 2003

Wilder Research Center: Barriers, Service Delivery Issues, and Outcomes for Somali, Hmong, and American Indian Participants in the Minnesota Family Investment Program (2001)

This study will examine culturally specific characteristics and service delivery issues that affect employment and welfare use among Somali and Hmong immigrants and American Indians in Minnesota, groups with the lowest welfare exit rates. Researchers plan to collect in-depth, qualitative information directly from members of the cultural groups. A better understanding of culturally related variations in characteristics, needs, and experiences will facilitate planning for more effective services and policies at a time when members of these three groups constitute a disproportionate share of those who are likely to exhaust their 60-month TANF limits.

Estimated Completion Date: Fall 2003

Projects to Improve Data Collection, Comparability or Capacity-Building

Support to the New Immigrant Survey (2000 and 2001)

The New Immigrant Survey is a large, longitudinal survey of recently arriving immigrants beginning in 2000. The Immigration and Naturalization Service (INS) and the National Institute for Child Health and Human Development (NICHD of the National Institutes of Health) are the principal funders of the survey. ASPE has contributed to this effort and also has provided input to the planning of the study and the development of the pilot instruments. ASPE's contribution helps ensure that comprehensive and relevant data are collected and analyzed about program utilization and hardship and well-being over time among newly arriving low-income immigrant families in different states. In particular, ASPE's continued support will ensure that the study focuses on what is happening to children in these families under welfare reform.

Estimated Completion Date: 2005

State and Local Telephone Survey to Assess the Incidence of Children with Special Health Care Needs (2000)

The devolution of welfare to the states and increased flexibility poses substantial new challenges for data collection and analysis to monitor welfare outcomes. To meet these challenges new and better data are needed at the state and local level. This project supports the administration of a welfare participation question in the state level telephone survey - State and Local Area Integrated Telephone Survey (SLAITS) - sponsored by the Health Resources and Services Administration (HRSA) and carried out by the National Center for Health Statistics of the Centers for Disease Control and Prevention (CDC). The participation question has been cognitively tested and is the same as that asked on other national surveys (e.g., Current Population Survey, National Survey of Drug Abuse). This data element when combined with other data available from this survey will permit the development of state level estimates of the incidence of special health care needs among children of current and former welfare recipients, as well as the health insurance status (including Medicaid and SCHIP) of current and former recipients. The fielding of the survey is complete, and data are being readied for analysis. Following the initial analysis of the data, a public-release database will be released for analysis.

Estimated Completion Date: Summer 2002

Support for the Research Forum on Children, Families, and the New Federalism Database and Web Site (2000 and 2001)

This project, funded jointly by ACF and ASPE, supports the National Center for Children in Poverty (NCCP) Research Forum on Children, Families, and the New Federalism database and website. The website is designed to provide the most reliable information to key stakeholders, including researchers, policymakers, administrators, and practitioners concerning welfare reform interventions being tested; populations and geographic areas being assessed; research methods being used; major findings already available; and when future findings will be released. The data base and web site provide valuable information useful to Federal officials and other practitioners regarding research and demonstration initiatives related to welfare reform and the well-being of low-income children and families.

Estimated Completion Date: On-going activity

Use of Social Security Summary Earnings Records to Assess Welfare Reform Outcomes (1999 and 2000)

This project continues ASPE's support of a study to determine the prevalence of job-holding associated with a living wage in the post-1996 period for adults who received AFDC benefits in calendar year 1996. The sample of 1996 adult recipients will be drawn from the 1996 panel of the Survey of Income and Program Participation (SIPP), the Annual Demographic Supplement (ADS) to the March 1997 Current Population Survey and the 1997 base-line interview sample of the Survey of Program Dynamics. Post-1996 earnings activity will be documented using earnings records obtained from the Social Security Administration (SSA) administrative records matched to the samples for each of these surveys. Initial tracking of job holding and earnings levels via administrative records will be restricted to calendar years 1996, 1997 and possibly 1998. Job holding of female family heads with dependent children who were not receiving means-tested benefits also will be tracked to provide a broader context for interpreting the observed patterns among adult AFDC recipients. Employment and earnings outcomes will be differentiated both by baseline characteristics and earnings patterns established on the basis of the pre-1996 year-by-year lifetime earnings histories stemming from the SSA administrative records files. This project has been delayed because of difficulties in securing the necessary inter-agency data sharing agreements. All agreements are now in place, the data has been merged and the analysis phase is in progress. We expect to receive an initial report on welfare and former welfare recipients in Fall 2002.

Estimated Completion Date: December 2002

Project on Child Outcomes: Enhancing Measurement of Child Outcomes in State Welfare Evaluations and Other State Data Collections (2001)

With other federal and private funders, ASPE and ACF are working with states to improve measurement of child health and well-being outcomes in state welfare evaluations. Five states are using a common protocol to add child outcome measures to their welfare reform evaluations. Continuation funding is enabling states and their evaluators to receive research technical assistance on collecting survey data using the common core of instruments, using administrative data sources, and developing and coordinating data analysis and reporting strategies. The focus of the current phase of work is the production of a synthesis of the findings from the five state evaluations. The research technical assistance is provided by the NICHD Research Network on Family and Child Well-Being under Child Trends' Leadership.

Estimated Completion Date: Summer 2002

Survey Development for State Studies of TANF Caseload (2001)

This task order is for the development of a standardized telephone survey instrument to be used in ASPE's State Studies on TANF Caseload project (described above) and for technical assistance throughout this project. Under this task order, Mathematica Policy Research, Inc., has developed a survey instrument following review of existing instruments, such as the Women's Employment Survey (WES), the National Survey of America's Families (NSAF), the Survey of Program Dynamics (SPD), the Mathematica Policy Research Caseload Survey, and TANF "leavers" surveys from Missouri and Alameda County, CA. The survey instrument consists of a common set of questions to be administered by states participating in the above project, supplemented by additional questions according to each state's particular interests. Mathematica also will be responsible for gaining approval of the survey instrument from the Office of Management and Budget, and for providing a variety of technical assistance to participating states.

Estimated Completion Date: Winter 2003

A Study of Families on TANF in Illinois (2001) (formerly Analysis of Survey Data on the TANF Caseload in Illinois)

This task order is to assist ASPE in utilizing survey data on the TANF caseload in Illinois. The data were acquired by Mathematica Policy Research, Inc., in a project funded by the Packard and Casey Foundations. The data will include information on the personal characteristics and situations, potential barriers to employment, compensating strengths and resources, preparation for employment, and employment outcomes of current TANF recipients. Under the ASPE task order, Mathematica is analyzing this rich set of data, in order to explore the relationships of an array of factors to the employment outcomes of TANF recipients. In particular, the analysis will enhance ASPE's understanding of the extent to which "job readiness" or "human capital" - the skills, experiences, habits and attitudes that prepare individuals for employment - compensate for barriers to employment. Findings will inform TANF policy and administration, particularly regarding screening and assessment, job preparation strategies and the targeting of services. An interim report of preliminary findings is expected in July 2002, and a final report in January 2003.

Estimated Completion Date: Winter 2003

Small Grant Program on Use of HHS-Sponsored Data Sets (2001)

Over a dozen new HHS-sponsored administrative and survey databases have recently become available for researcher use. Secondary analyses of new databases will increase our understanding of the outcomes of welfare reform. These databases include data from the ASPE supported National Evaluation of Welfare to Work Strategies (NEWWS), as well as state and local welfare leaver studies. In this project, HSP is working with the Institute for Research on Poverty at the University of Wisconsin to provide dissertation and post-dissertation grants to young scholars to explore aspects of welfare reform using the HHS-sponsored data sets. The grants also will build capacity by encouraging younger scholars to undertake welfare-related research. Three grants were awarded in December 2001; additional grants are expected to be awarded in June 2002.

Estimated Completion Date: Reports from the first grants are expected in early 2003; reports from the second round of grants are expected in late 2003.

Measures of Material Hardship (2001) (formerly Measures of Family Hardship)

Although a number of national and state surveys have begun gathering measures of material hardship (e.g., utility cutoffs, inability to get needed medical attention, food insecurity, evictions), it is hard to respond to Congressional interest in gathering information on a state-by-state basis, given the small sample size of most national surveys and the lack of comparability across state surveys. The purpose of this project is to advance understanding of the value and limitations of measures of material hardship as a component of family well-being. The initial phase of this project has been completed, including the convening in February 2002 of a one-day roundtable on measuring material hardship. Over the coming months, the contractor, Abt Associates, will produce a final report summarizing the one-day meeting.

Estimated Completion Date: December 2002

Panel Study of Income Dynamics: Core Support and Expanded Sample for Child Supplement (2001)

This project continued ASPE's on-going core support for the Panel Study of Income Dynamics (PSID). ASPE funds provide partial support for the continued collection and processing of longitudinal data relevant to research on economic factors and income support mechanisms, health, fertility, medical care, and disability affecting the poor and the elderly. In particular, ASPE funds continued to support an expansion of the set of welfare related questions to assess the entry effects of recent reforms. In addition, these funds supported an expanded sample of low-income families with children for core data collection activities related to the Child Supplement, which collects a host of information on the well-being of children. Data collected as part of this effort will be made publicly available for analysis by the grantee. Core support for the PSID comes from the National Science Foundation ($2 million), the National Institute on Aging, and the Departments of Agriculture and Housing and Urban Development, along with other funders.

Estimated Completion Date: On-going activity

Improving State Child and Youth Indicators of Health and Well-Being (2001)

Working with the Chapin Hall Center for Children at the University of Chicago and Child Trends, ASPE is preparing a technical assistance conference for states that are interested in developing and monitoring indicators of youth health and well-being as changes occur in welfare and other key policies. The contractors have created inventories of national and state efforts to improve youth measures, and are working with states to design and conduct a meeting on youth indicators, develop dissemination materials, and identify the best opportunities for HHS to assist states in the future. The meeting was held in Spring 2002, and materials are being/will be disseminated during major meetings, such as the National Youth Summit in June 2002, and the annual meeting of the National Association of Welfare Research and Statistics scheduled for August 2002.

Estimated Completion Date: Summer 2002

Analytic Projects on Welfare-Related Topics

Transition Events in the Dynamics of Poverty (2000)

This project is studying the events associated with people entering and exiting poverty. The project will document the likelihood of entering and exiting poverty for various groups, as well as the extent to which various transition events or combinations of events account for entries and exits from poverty. This project also will examine whether poverty rates are declining because fewer people are entering poverty or because more people are exiting poverty. The product will be a report based on data from the Panel Study of Income Dynamics (PSID) and the Survey of Income and Program Participation (SIPP) that analyzes the events that increase individuals' likelihood of entering and exiting poverty, and the dynamics behind changes in the poverty rate over time.

Estimated Completion Date: Summer 2002

Quick Turnaround Analyses under Welfare Reform (2000)

The TANF program must be reauthorized before the end of Fiscal Year 2002, as must the Child Care and Development Block Grant, the Food Stamp Program, and several other programs. We anticipated that, as in the period before enactment of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA), ASPE would be called upon to contribute to the reauthorization discussions by providing analyses of policy issues and options, especially those affecting low-income children and families. Under this project, ASPE awarded a task order contract to the Urban Institute to have the Institute perform very quick analyses of existing data sets, such as the Current Population Survey (CPS), the Survey of Income and Program Participation (SIPP), data from the Urban Institute's National Survey of America's Families (NSAF), administrative data on the TANF program, state policy data, and others. Data analyses could provide information about the characteristics of children and families across a wide range of policy relevant topics, including transitions to employment, child and family well-being, child poverty, interactions with food stamps, Medicaid/SCHIP and SSI, the impact of state policy changes made as a result of the flexibility of the TANF law, or other questions identified during the reauthorization process. Each question would be expected to result in a deliverable of a memorandum with detailed tables.

Estimated Completion Date: Fall 2002

Implementation of Welfare Reform at the Local Level: Implications for Special Populations (2000)

With the implementation of welfare reform, state and local agencies have established a variety of rules and procedures governing enrollment in TANF and Medicaid. As authority for welfare policy has devolved to state and, oftentimes, local levels, local agencies and caseworkers may have more discretion over how individual cases are handled. This study is examining the implications for special populations, particularly individuals of different backgrounds and limited English language abilities. The results of this project will provide additional information about the effects of program policy and implementation at different levels on program utilization by these special populations. The project consists of detailed case studies that examine agency policies and practices, as well as caseworker training and discretion, in six metropolitan areas, which are in the process of being identified.

Estimated Completion Date: Summer 2002

Understanding the Declines in Teen Birth Rate (2000)

PRWORA describes several outcomes of concern related to teen pregnancy, including an increased likelihood of dependence on public assistance, and reducing teen pregnancy is viewed as an important aspect of promoting self-sufficiency and family well-being within the context of welfare reform. Yet, as teen birth rates have fallen at an unprecedented rate since 1991, there is a debate regarding factors that have contributed to this decline. This project will use data from the National Survey of Family Growth to describe data on sexual activity, partner characteristics, and contraceptive use for women surveyed in 1995 who were teens at any time during the study period. This data will be used to create simulation models that may clarify which factors are associated with changes in teen pregnancy and births and how possible future changes in these factors might affect teen pregnancy and birth rates. Using monthly event history data, the study will observe trends in behavior between 1991 and 1995. Trend information will be presented for multiple population subgroups, including by race/ethnicity, age, and parity (whether or not they had a prior teen birth). Estimates based on event history data are being created. Work is beginning on setting up a base model to assess factors associated with a continued decline in the teen birth rate.

Estimated Completion Date: Fall 2002

From Prisons to Home: The Effect of Incarceration on Children, Families, and Low-Income Communities (2000)

A majority of incarcerated men and women are parents, and the impact of incarceration appears to be greatest in poor, minority, urban communities. The toll on children, families, and communities of this incarceration and return of inmates has begun to be recognized, together with the growing realization that families served by TANF and other Department of Health and Human Services programs are families who also are more likely to experience the effects of incarceration. The prison to home project commissioned 11 papers which were presented at a policy conference in January 2002. About 300 representatives from the criminal justice, health, and human services fields attended, including federal officials, state officials, program operators and researchers. Among the important preliminary findings was that the highest concentration of populations on TANF are living in neighborhoods where incarceration rates also are the highest. This is particularly true in neighborhoods with large numbers of children (in one such community, 54 percent of the children were receiving public assistance). The papers from the conference will be posted on the ASPE website when available. A final report on the project is expected in Summer 2002.

Estimated Completion Date: Summer 2002

Learning from State Corrections and Human Services Collaborations: From Prison to Home - Part II (2001)

As part of ASPE's work examining how families and children are affected by the high rates of incarceration in some low-income communities, a symposium was held in November 2001, bringing together the collaborating agency heads and other key stakeholders from selected sites. This symposium provided essential information about the issues for and choices made by states as they implement cross-cutting strategies that involve criminal justice, welfare and health systems. Many of the incarcerated and formerly incarcerated individuals and their families are current or former welfare recipients. Key areas explored at the symposium included resource needs, identification of barriers, partnership building, and implementation lessons. National organizations, such as the National Governors Association and National Conference of State Legislatures, and other federal and private sector representatives also attended the symposium. Information from the symposium was used to plan and implement the January 2002 ASPE policy conference "From Prison to Home," and will be incorporated into the final report on that conference (see above).

Estimated Completion Date: Summer 2002

Poor Families with Infants and Toddlers (2000)

The purpose of this project, co-sponsored by ASPE and the David and Lucile Packard Foundation, is to identify and provide information about promising federal, state and community-based strategies which are believed to be useful in promoting the overall well-being of infants and toddlers and their families. The focus is on poor families and post-PRWORA changes in policies and programs. The highlighted strategies have been grouped into the four core areas of family life on which the project is focused - economic security, basic supports, nurturing relationships, and special supports for vulnerable children and families. Many strategies touch on more than one core area. Representatives of the strategies met in a workshop in March 2002 with government staff at various levels to share information. The resulting product will be a sourcebook of profiles of programs/initiatives targeted toward poor infants and toddlers and their families. Information about the profiled strategies will be presented in a user-friendly way in order to be valuable to a broad audience. The sourcebook will be made available through ASPE's website.

Estimated Completion Date: Summer 2002

Enhancement of the Study of Trends in Emergency Assistance Related to TANF (2000) (incorporates Trends in the Demand for Assistance Services, 1999)

The 1999 funded project, Trends in the Demand for Assistance Services, jointly funded with the Office of Program Systems (PS) within ASPE, examined the trends in the demand for emergency assistance services, such as homeless shelters and food banks, from the mid-1990's to 2000. One study, Welfare Reform and Emergency Service Use in Communities with Rapidly Declining Caseloads, was conducted by the SPHERE Institute in San Mateo County, California. The second study, Trends in Demand for Emergency Services in Massachusetts Post Welfare Reform, 1995-2000, was conducted by the University of Massachusetts and focused on the entire state. The studies found both parallels and differences in the patterns of emergency assistance. Among the parallels were a substantial decline in the welfare rolls, substantial increases in the cost of affordable housing, increased use of emergency rental assistance by poor families, and increases in the use of shared living space among families that used emergency food assistance.

Differences in findings between the two grants also were noted. In the California study, welfare reform did not appear to be associated with significant increases in the number of residents relying on food assistance or emergency shelter. Food assistance appeared to show declines during the study period, particularly among homeless recipients. Between 1994 and 1998, the overall number of homeless persons was unchanged, but the single homeless adults increased while homeless children decreased.

In the Massachusetts study, while more parents were working, the researchers believe that poor families were poorer (earnings increases did not offset drops in TANF and food stamp benefits). The researchers found increased use of emergency shelter and food assistance programs by poor families during the study period. The researchers noted housing affordability problems - defined as the loss of affordable housing stock, a tripling in the turn-back of Section 8 vouchers, and rent increases - as significant contributors to the increased demands for emergency assistance.

To add more locations to this analysis, the Office of Program Systems funded an Intra-Agency Agreement with the Substance Abuse and Mental Health Services Administration (SAMHSA) during FY 2000. SAMHSA's Center for Mental Health Services added funding to the Agreement. SAMHSA awarded a contract to the Gallup Organization that added Spokane, Washington and Philadelphia, Pennsylvania to the analysis. The sites used time frames and emergency assistance programs that were consistent with the above grants. The contract was being jointly monitored by ASPE and SAMHSA. The Gallup Organization has submitted a draft final report that is undergoing program review. A Summer 2002 release of the findings from the four sites is anticipated.

Estimated Completion Date: Summer 2002

Welfare Reform and Its Implications for Persons with Disabilities (1998)

This project is a supplement to an ongoing four-year study of the implications of welfare reform for low-income families living in Boston, Chicago, and San Antonio (a summary of the study is available at <http://www.jhu.edu/~welfare>). The broader study is being undertaken by researchers at Johns Hopkins University, Pennsylvania State University, Harvard University, University of Chicago, and the University of Texas-Austin. Funds are being used to explore how welfare reform is affecting the lives of a particularly vulnerable subset of the welfare population - adults and children with disabilities. This will be accomplished by conducting longitudinal case studies of families with members with disabilities receiving TANF and through a broader survey effort. The purpose of the data collection efforts is to better understand how recent work participation requirements and time limits under welfare reform are affecting service utilization, family member's health and development, support networks, parenting, and child care arrangements.

ASPE and the Administration for Children and Families are the primary funders of the disability component of the study. The National Institute of Child Health and Human Development is the primary funder of the broader study, along with several private foundations. Andrew Cherlin, Johns Hopkins University, is the principal investigator. The first round of interviews for the main survey were conducted from March 1999 until December 1999. Fieldwork in the ethnography component began in fall 1999 and is ongoing. Results from the project will be available over the course of the study. Listed below are the current publications from the broader study; results from the ethnographies focused on disability will be available later.

Accomplishments to date:

• The Characteristics of Families Remaining on Welfare, March 2002 - Policy Brief 02-2
• The Characteristics of Families Remaining on Welfare, Working Paper 02-02
• Welfare Reform: What About the Children?, January 2002 - Policy Brief 02-1
• A Closer Look at Changes in Children's Living Arrangements in Low-Income Families, Working Paper 02-01
• Child Care Choices in the Era of Welfare Reform: Quality, Choices, and Preferences, December 2001 - Policy Brief 01-4
• Public Assistance Receipt Among Native-Born Children of Immigrants, October 2001 - Policy Brief 01-3
• Health Insurance Coverage for Children and Their Caregivers in Low-Income Urban Neighborhoods, July 2001 - Policy Brief 01-2

Estimated Completion Date: 2003

Temporary Assistance for Needy Families: Time Limits (2001)

In late 2001, the first families reached the 60-month federal time limit on receipt of TANF benefits. ASPE and ACF jointly funded the Manpower Demonstration Research Corporation (MDRC) to study the early experiences of states in implementing both the federal time limit and shorter state time limits, in order to be able to answer questions about these policies that come up as part of the debate regarding TANF reauthorization. Based on a survey of state TANF administrators, the report will examine the policies that states have adopted regarding time limits, exemptions, and extensions; the number of cases that have reached state or federal time limits; and whether they have been terminated as a result. The report will then draw upon site visits to examine the implementation of time limits in greater detail. Finally, the report will summarize the existing literature on the effects of time limits on welfare receipt, employment, income and other measures of adult and child well-being, and on the outcomes for families terminated from welfare due to time limits.

Estimated Completion Date: Summer 2002

Welfare Reform and Children (2001)

While some experimental studies of welfare to work programs have measured outcomes for children, broader discussions of welfare reform are rarely framed around issues for children. Under this project, Mathematica Policy Research (and subcontractor Child Trends) will summarize the current body of research on welfare programs and children (either all children or a salient subgroup) and identify paths for future research. Topics under consideration include: characteristics of children in welfare-dependent families, family structure, child care, and child welfare.

Estimated Completion Date: Fall 2002

Synthesis of Literature on Family Composition and Resource Sharing (2001)

Welfare reform's effects on family formation and composition as well as how such changes can affect the financial resources and material well-being of members of low-income households are issues of strong interest to policy makers. Through a literature review, The Urban Institute is documenting what is currently known about the effects of welfare policies on family formation and resource sharing. This synthesis will provide an overview of existing research across disciplines (e.g., anthropology, sociology and economics). The study also will summarize key data sources on this issue (e.g., SIPP, CPS, NLSY, SPD, National Survey of America's Families, National Survey of Family Growth) and their strengths and weaknesses for investigating family composition and household resource and well-being issues.

Estimated Completion Date: Summer 2002

Follow-up Work on Minnesota Family Investment Program Evaluation (2001)

Recently released findings based upon a small sample from the Manpower Demonstration Research Corporation's (MDRC) evaluation of the Minnesota Family Investment Program (MFIP) suggest that the program had significant positive impacts on the marital stability of two-parent families. ASPE is funding a follow-up study of all two-parent recipient families in the sample to determine if the robustness of the original findings can be replicated in the larger sample. This work is a necessary first step before a decision is made about conducting a longer-term follow-up analysis of well-being outcomes for these families. If the original findings hold up for the larger sample, the balance of the task order funds will be used to develop (but not field) survey instruments for measuring well-being for the full sample.

Estimated Completion Date: Fall 2002

Case Studies of State Initiatives to Reduce Out-of-Wedlock Births (2001)

The reduction of nonmarital births is an important focus of TANF. This is highlighted in the goals of TANF and several provisions including the Bonus to Reward Reduction in Illegitimacy Ratio. This project examines recent state experiences in their efforts to reduce nonmarital births, including the role the Bonus has played in those efforts. This project gathers information from a sample of states to learn more about what efforts states have pursued in reducing nonmarital births and the challenges they have faced. The information has been gathered through panel discussions with stakeholders in a sample of states and from existing surveys of state efforts.

Estimated Completion Date: Fall 2002

Other Projects

Project on Devolution and Urban Change (2000 and 2001)

This five-year project (which is primarily foundation-funded) is a multi-disciplinary study by the Manpower Demonstration Research Corporation (MDRC) of the implementation and impacts of welfare reform and welfare-to-work programs on low-income individuals, families and communities in four large urban areas - Cleveland, Philadelphia, Los Angeles, and Miami. The project brings together data from an unusually wide array of sources - longitudinal administrative data for all families receiving AFDC/TANF or food stamps dating back to 1992, survey data, an implementation study, neighborhood indicators, an institutional study focusing on local service providers, and an ethnographic study of a limited number of families. In addition to ASPE, federal funding partners include ACF and the Economic Research Service at USDA. The federal contribution to this project leverages a substantial investment by foundations, which are funding the majority of the over $20 million project cost.

Accomplishments to date:

Is Work Enough? The Experiences of Current and Former Welfare Mothers Who Work, November 2001.

This report examined the work experiences of current and former welfare recipients in four urban areas, comparing those with the most stable recent employment histories to those with less stable employment. The researchers found high levels of full-time work and employment stability among these women, but even the women with high employment stability still worked in low-wage jobs and experienced multiple material hardships.

Readying Welfare Recipients for Work: Lessons from Four Big Cities as They Implement Welfare Reform, March 2002.

This report described how the welfare-to-work services provided to clients have changed in four urban areas since TANF, examining services and activities, the role of case management, participation rates, and spending patterns. The paper also looks at the effects of the funding provided under the formula grant component of the Welfare to Work Grants.

(These and other Urban Change reports are available at <http://www.mdrc.org/>)

Estimated Completion Date: September 2003

Continuation of New Jersey Welfare/Substance Abuse Evaluation (1998 and 2000)

In FY 1998 we began funding, in partnership with ACF, a three year grant to support the evaluation of a New Jersey initiative which aims to improve employment and family outcomes for TANF recipients with substance abuse problems through substance abuse treatment, intensive case management and supportive services. This evaluation is providing important information about the effectiveness of a type of intervention several states are experimenting with to move substance abusing welfare clients toward self-sufficiency. The intervention New Jersey is implementing includes screening of welfare recipients for substance abuse problems, treatment referral mechanisms with enhanced case management, and substance abuse treatment coordinated with employment and training or vocational services. The evaluation, using a random assignment model, compares two models for providing such services, looking at outcomes in several domains including employment and family self-sufficiency, substance use and associated behaviors, child development and family functioning, and child welfare involvement. The intervention being evaluated is intended to improve the post-welfare prospects of TANF recipients with substance abuse problems. The evaluation is being conducted in two New Jersey counties (Essex and Atlantic). ASPE and ACF have provided support for this project. Other aspects of the evaluation are being funded by the Annie E. Casey Foundation and the Department's National Institute on Drug Abuse.

Estimated Completion Date: Random assignment of clients to the intervention models began in mid-1999. Research Notes on the effectiveness of two approaches to screening and assessment of substance abuse in welfare settings, and on the initial rates of treatment engagement and retention for program participants versus the control group were published in January 2001. A report entitled Barriers to Employability Among Women on TANF with a Substance Abuse Problem has been completed. Reports on project implementation and outcomes will be published later in 2002.

Los Angeles Family and Neighborhood Survey (L.A. FANS) (1999)

The Los Angeles Family and Neighborhood Survey (L.A. FANS) is a longitudinal study by RAND of children, families and neighborhoods in Los Angeles County, funded primarily by the National Institutes of Health, with ASPE and Los Angeles County providing supplemental funding for the first wave of data collection. L.A. FANS includes a representative sample of 65 neighborhoods (census tracts) throughout Los Angeles County, with an oversample of poor neighborhoods. In each neighborhood, interviews are conducted with a total of 40 to 50 randomly-chosen households; households with children (0 to 17) are oversampled. Extensive information is collected on household socioeconomic status, health care utilization, immigration, and other characteristics. In addition, L.A. FANS collects a detailed two-year month-by-month calendar of changes in employment, unemployment, health insurance coverage for adults and children (by type and reason for changes), and program participation (TANF, SSI, GA, food stamps). Information also is collected on the characteristics and available health services in each sampled neighborhood.

Fieldwork for the first wave of the study was conducted between April 2000 and December 2001. Interviews were completed with adults and children in a total of 3,171 households. Because of recent immigration trends and the oversample of poor communities, the L.A. FANS wave 1 sample includes a substantial number of first and second generation immigrants. Reflecting the demography of Los Angeles, a majority of respondents are Latino. Interviews were conducted in both English and Spanish by native speakers of each language. Analysis of the wave 1 data will be carried out during the next year and regular reports of findings will be released as available on RAND's website at <http://www.lasurvey.rand.org>. Public use data files also will be released for other researchers' analyses of welfare reform, health disparities, insurance coverage, and health care utilization. In addition, RAND is working with the Los Angeles County Commission on Children and Families, the Los Angeles County Children's Planning Council, and other organizations to insure that data and results are available to local and state policymakers. RAND is currently seeking funding for a second wave planned for 2004.

Estimated Completion Date: 2003

Demonstration and Evaluation of Enhanced Services for the Hard-to-Employ (2001)

State and local agencies are making substantial investments through TANF and other sources to help low-income families with demonstrated difficulty entering and sustaining employment. There is a significant amount of activity and a variety of approaches being used to help low-income parents address or cope with the personal and family problems that interfere with their employment stability. ACF and ASPE are supporting a long-term, multi-site evaluation of programs working with hard-to-employ low-income parents in order to identify effective strategies for promoting employment and family well-being and to determine the effects of such programs on employment, earnings, income, welfare dependence, family functioning, and the well-being of children. The Manpower Demonstration Research Corporation has been chosen as the contractor to design and conduct a multi-site evaluation that studies the implementation issues, net impact, and benefit-costs of selected programs. During the first year, the MDRC will assist HHS in identifying and recruiting programs with potential for evaluation and will assist selected programs in strengthening or expanding services to meet requisite conditions for rigorous evaluation.

Estimated Completion Date: 2009

Support for Working Families and Their Children (2001)

This project helps support a National Governors Association (NGA) project to build state and local capacity to provide work supports which help low-income working parents sustain employment and advance in the labor market, as well as increase positive family functioning and child well-being. In October 2001, NGA convened a roundtable of federal and state policymakers, program administrators, and researchers with subject area expertise to develop clear goals for serving low-income working families. NGA has identified programs that are currently providing supports for low-income workers and their families, and is conducting site visits in order to identify lessons learned from their experiences. These lessons will be incorporated into a State Policymakers Guide to Developing a Policy Agenda for Low-Income Working Families, and NGA will convene one or more conferences for state and local officials. NGA also will provide customized technical assistance to three states that wish to develop these models further. ACF is the lead agency on this project, and USDA/ERS also is providing funding for this project.

Estimated Completion Date: February 2003

Evaluation of State Programs of Abstinence Education (2001)

ASPE/HSP continues to manage an existing $6 million multi-year contract with Mathematica Policy Research, Inc. (MPR) to conduct a Congressionally mandated evaluation of selected programs under Title V, State Abstinence Education Program. This large and complex, rigorous evaluation is taking an empirical look at the differential effectiveness of several types of abstinence programs. It will measure the success of different program models in altering adolescent attitudes and intentions about premarital sex, reducing sexual activity among teens, convincing adolescents who have had sex to become abstinent, and lowering exposure to sexually transmitted diseases and nonmarital births. The 2001 ASPE funds will allow the evaluation to follow adolescents for longer periods of time and to coordinate with the newly mandated community-based abstinence education evaluation (see below). Highlights from the interim early implementation report to Congress are included in Chapter 2. A report on the success of programs in achieving their short term goals of changing knowledge, attitudes and near-term behavioral choices is expected in Winter 2003. The final impact findings will be available in 2005. Details and updated information about the evaluation can be found at Mathematica Policy Research's web site on abstinence-only education programs <http://www.mathematica-mpr.com/3rdlevel/abstinence.htm>.

Estimated Completion Date: 2005

In FY 2001, a new community based abstinence education grant program was created in the Health Resources and Services Administration (HRSA). These community grants provide support to public and private entities for the development and implementation of abstinence education programs in communities. ASPE has been given responsibility for the Community Based Abstinence Education Program Evaluation activities. This separate but related project, to be conducted by Abt Associates, will explore the programmatic and evaluative information that currently exists in the area of abstinence-only education and related fields, and develop design options for evaluation activities. This work will be coordinated with the ongoing evaluation of the state formula grant program activities.

Case Studies on Privatization of Service Delivery and Performance-Based Contracting (2001)

With the devolution of responsibility for welfare programs under TANF, many state and local governments have turned to non-profit, and increasingly for-profit, organizations to carry out human services functions that traditionally have been provided by the public sector. As a result, there has been an increase in the prevalence of performance-based contracts, as state and local governments attempt to maintain accountability while testing the effectiveness of these non-traditional service providers. Under this project, Mathematica Policy Research is undertaking case studies of six state and local human services agencies that have privatized services funded under the TANF block grant. The sites selected for participation in this study include: Delaware, Wisconsin, Hennepin County, MN, Palm Beach County, FL, San Diego County, CA, and Lower Rio Grande Valley, TX. Agency administrators and front-line workers are being interviewed to determine the types of services that have been privatized and the performance measures used by government entities in their relationships with private organizations. Site visits began in March 2002 and will take place throughout the spring and summer of this year. A final report will describe emerging positive trends and problem areas, in both the services delivered and the types of contracts used. In addition, a review of the literature in the area of privatization of welfare services has been completed, and will be posted on the ASPE web site at: <http://aspe.hhs.gov/hsp/>.

Estimated Completion Date: January 2003

Endnote

1.  In 1998 grants were awarded to ten states and three large counties or consortia of counties (Arizona, the District of Columbia, Florida, Georgia, Illinois, Massachusetts, Missouri, New York, Washington, and Wisconsin; and Cuyahoga County, Ohio, Los Angeles County, California, and San Mateo, Santa Cruz, and Santa Clara Counties, California). Separate but comparable studies were also funded in Iowa (with FY 1999 funding) and South Carolina (in FY 1998 and 2000, as part of a longer-term project) resulting in a total of 15 studies. Three grantees - Arizona, Missouri, and a consortium of San Mateo, Santa Cruz, and Santa Clara counties - received additional funding in FY 1999 to extend the studies and administer a second or third wave of interviews, allowing analysis of longer-term outcomes for former recipients. Descriptions of the funded projects and links to available reports can be found at <http://aspe.hhs.gov/hsp/leavers99/index.htm#background>.

2.  Diversion programs include formal efforts to address the immediate needs of families seeking cash assistance in ways that avoid enrolling these families in TANF. Examples of formal diversion programs include lump sum payment programs, mandatory applicant job search programs and the exploration of alternative means of support.

3.  Although funded separately, findings from South Carolina's study of welfare leavers were incorporated with the Findings from ASPE-Funded Leavers Studies (Grants to States and Localities to Study Welfare Outcomes), which are summarized in Chapter II.

Appendix. Grants to States and Localities to Study Welfare Reform Outcomes, with an Emphasis on TANF Applicants and Diversion: Synthesis of Findings

Grants to States and Localities to Study Welfare Reform Outcomes, with an Emphasis on TANF Applicants and Diversion (1998 and 1999): Synthesis of Findings

Since the passage of the Personal Responsibility and Work Reconciliation Act of 1996 (PRWORA), there has been a significant drop in the federal cash assistance caseloads. During this period, a large number of families have left welfare, leading to great interest among researchers and policymakers in studying the circumstances of welfare "leavers." However, much of the caseload decline may be due to changes in entries onto welfare, so it is also important for researchers to look at individuals and families who have applied for cash assistance.

To this end, ASPE funded ten different studies that have either focused exclusively on applicants and/or potential applicants to the Temporary Assistance for Needy Families (TANF) program or included an applicant component within the study. Six of these ten studies were funded through a request for applications from states and large counties in April 1999 with an emphasis on the study of applicants and potential applicants to the TANF program, including those who were formally or informally diverted from TANF.

The ten ASPE grantees with an applicant component to their projects included: Arizona, Florida, Illinois, New York, South Carolina, Texas, Washington, Milwaukee County in Wisconsin, and two consortia of counties in the Bay Area of California (Contra Costa and Alameda Counties and San Mateo, Santa Clara, and Santa Cruz Counties). Eight of these grantees have released final reports detailing their findings, allowing us to develop this brief synthesis of the studies of TANF applicants and diversion. New York and Washington are expected to release final reports in Summer 2002. In addition, researchers at the Institute for Research on Poverty (IRP) at the University of Wisconsin-Madison will be releasing the Wave Two survey report of the applicant study in Milwaukee County in Summer 2002.

As with the ASPE-funded studies of welfare "leavers," each of the grantees include both administrative and survey data, with some of the surveys including multiple waves. ASPE provided technical assistance to grantees while allowing them to determine the cohorts to be studied, the administrative data measures to be used, and the surveys to be administered. This synthesis of ASPE's applicant and diversion studies has many of the same inherent difficulties as previous syntheses of ASPE's leavers studies. These include variations in the time period studied by the grantees, as well as the period of time between the application period and the administration of the survey. Further, each survey was developed by a separate set of researchers in each state or locality, leading to differences in measured outcomes.

In addition to these issues, it became apparent from the beginning of the project that it would much more difficult to facilitate cross-study comparisons across the applicant grantees than it was for the "leavers" grantees. In both sets of studies, grantees had significant discretion over their research questions and construction of their surveys. However, while studies of TANF leavers used a common definition of leaver (as those who left TANF and remained off the rolls for at least two months), it was not possible to use a common definition of "applicant" or "diverted" populations because of the significant differences in states' application procedures and availability of administrative data on applications that are only partially completed.

While the definitions of the populations of applicants and divertees vary significantly across grantees, they can be divided into five general categories, as described below and summarized in Table 1:

1. Participating in the state or county's formal diversion program. Many states have implemented formal programs to divert TANF applicants from entering cash assistance. The most well-known of these programs are those that pay grantees a one-time lump-sum payment in lieu of enrolling them in the TANF cash assistance program. Several of the states and counties receiving grants in FY 1998 and FY 1999 to study diversion have implemented a formal diversion program. However, only one of the ASPE grantees included here, Texas, decided to study formal diversion programs, specifically participants in the state's one-time lump-sum payment program. Others did not study this population, partially because of low participation in the programs.
2. Begin the application process but fail to complete it. Many states have the ability, either by administrative data or through intercept interviews within the welfare office, to begin collecting data on individuals as soon as they enter the welfare office and provide contact information to an intake worker. If researchers are able to begin tracking applicants at this point, then those individuals who do not complete their applications for whatever reason are often considered diverted applicants. This was a popular method for defining divertees among the ASPE grantees included in this synthesis; six of the eight chose to include this population of applicants in their sample of diverted applicants.
3. Complete the application process and determined to be eligible for TANF but do not enroll in TANF. Some individuals may complete their application process but decide not to enroll in TANF, even after being found eligible to receive cash benefits. At this point, the completed application is often already in many states' administrative data systems, making tracking of these individuals possible. Four of the ASPE grantees that have released findings included these individuals in their sample.
4. Ineligible for TANF for non-financial reasons. Often TANF applicants meet the income eligibility requirements of the TANF program but are denied benefits for other reasons. These non-financial reasons for denial may include: refusal or inability to participate in required job search or job readiness programs, lack of cooperation with the child support enforcement agency, and failure to meet particular requirements set in place by states or localities. Five of the ASPE grantees included here targeted these individuals as diverted applicants. There may have been some overlap between this population and those who completed their application but did not enroll for unknown reasons.
5. Receiving food stamps and/or Medicaid and appear to be eligible for TANF but do not receive TANF. Two of the grantees included in this synthesis - Florida and South Carolina - used administrative data from the Medicaid and/or Food Stamp programs to help formulate a population of diverted applicants. While the two grantees approached the task differently, the basic theory was to use these data files to find a proxy for TANF divertees. If an individual received Medicaid or food stamps, appeared to be income eligible for TANF, and had not received TANF for a set period of time, then a reasonable assumption was made that they were informally diverted from receiving TANF. While there are many complicating issues surrounding this method, it is one way to try and identify individuals who were diverted from TANF before ever entering the welfare office to apply.

Despite the differences in the definitions of the study populations among the ASPE grantees, some themes have emerged across the studies. Below are cross-state findings in a number of number of different areas. First, there are data on the characteristics of applicants across the various study area. Next, outcomes of the applicants are synthesized in the areas of employment and earnings, participation in government programs, and eventual receipt of TANF cash assistance. Finally, the applicants' experiences with the TANF application process in their state or county are cited.

Characteristics of Applicants

Each of the studies contains information on the demographic characteristics of applicants in the study area, including their age, race/ethnicity, educational attainment, marital status, and the number of children in the household. These characteristics varied greatly across the study areas, and no conclusions can be made regarding the effects of varying demographic characteristics on outcomes for applicants. However, the information is useful for descriptive purposes.

• The majority of case heads were in the 18-35 age range. However, some studies had a significant number of older applicants (24 percent over age 35 in Contra Costa County, 19 percent over age 35 in Arizona, and 24 percent over age 40 in South Carolina).
• The race/ethnicity of the applicant population varied along with the natural demographics of each of the various study areas. The percentage of applicants/divertees who were African-American ranged from 6 percent in Arizona to 76 percent in Milwaukee County, while the percentage of Latino/Hispanic applicants ranged from 11 percent in Milwaukee County to 50 percent in Contra Costa County. The percentage of TANF applicants who were white ranged from 10 percent in Milwaukee County to 51 percent in Arizona.
• A few of the grantees collected information on the educational attainment of the primary caretaker of the applicant unit. These grantees found that between 37 and 58 percent of the TANF applicants/divertees had less than a high school education, while 34 to 54 percent had either a high school diploma or General Equivalency Degree (GED).
• In the few areas that collected data on marital status of the applicants, the largest percentage of respondents (44 to 80 percent) were never-married single parents. In addition, a significant percentage of respondents were divorced or separated. Fewer than one-fifth of the applicants were married at the time of the surveys.
• Most of the studies asked the respondents about the number of dependent children living in the household. In each study area, between one-third and one-half of respondents had only one children living in the household. The percentage of applicants with three or more children in the household ranged from 18 percent in South Carolina to 38 percent in Milwaukee County.

Employment and Earnings

Each of the grantees reported findings on the employment rates of applicants and divertees in the period preceding and following their application for TANF. Most of the studies included both administrative data from the Unemployment Insurance (UI) data system and the results of survey questions about the employment of applicants. In general, the data show either a steady employment rate or slight decline in employment during the period just prior to application, then a small but constant increase in employment in the months following diversion from TANF.

Grantees also used both UI administrative data and survey data to report on the earnings of TANF applicants. In general, there was a significant dip in earnings in the period immediately prior to application/diversion. This drop is very likely associated with applicants' decision to apply for TANF. The period in which the application/diversion took place was followed by a noticeable increase in earnings over the following year.

• Three of the grantees reporting administrative data on employment in the quarter of application to or diversion from TANF report rates similar to those of the ASPE "leavers" studies - around 60 percent. Three report somewhat lower employment rates of between 31 and 45 percent. Some possible reasons for the differences in employment rates include policy differences within the individual states, difference in the definition of diversion, and varying definitions of the quarter of exit.
• Grantees reporting survey data found slightly lower employment rates, particularly at the time of application. For example, 12 percent of respondents to a survey at the time of application in Milwaukee County said that they were currently employed, while 28 percent in Arizona reported being employed in a similar survey. One explanation for the lower employment rates from survey data might be that administrative data report employment at any time within a quarter, and the surveys show employment at one point in time. Thus, an individual might lose her job and choose to apply for TANF, but still be recorded by administrative data as employed in that quarter.
• As mentioned above, employment rates generally increased in the period following diversion from TANF, according to both administrative and survey data. Administrative data showed a steady but small increase (less than 5 percentage points) in employment across most grantees in the year following diversion. The exception was Florida, where the increase in the employment rate between the study quarter and 15 months after diversion was 14 percentage points (from 38 percent to 52 percent). Increases in employment were also shown in survey data findings reported in Arizona, Illinois, and San Mateo County.
• Administrative data from each of the studies showed a sudden drop in applicants' earnings in the quarter of TANF application, followed by a steady gain in earnings over the year following diversion. Mean earnings in the quarter before application ranged from about $1,800 to $2,500 in the non-California areas - this range fell to about $1,500 to $2,000 in the application quarter. Earnings then steadily increased over the next year; in the fourth quarter after application, the range of earnings was approximately $2,200 to $3,000. Earnings in the five counties in California were higher than those in the other study areas, possibly due to higher standards of living in the Bay Area of California. ⁽¹⁾
• Of those TANF applicants who were employed in the period following application/diversion, many worked in the service or clerical field . Median hourly earnings varied greatly across studies but generally were above the minimum wage. Those individuals who were not employed generally cited illness or injury, child care, or a preference to stay home with the child as a major reason for their unemployment.

Participation in Medicaid and Food Stamps

Government programs such as Medicaid and food stamps are important supports for low-income families not receiving cash assistance as they move toward self-sufficiency. Each of the ASPE grantees studying TANF applicants and diversion collected administrative and survey data on the receipt of these programs. While the level to which applicants were enrolled in these programs varied across the different sites, the data show that program receipt declined across each of the study areas in the year following TANF application/diversion.

• State administrative data from five of the applicant studies reveal that between 15 and 35 percent of the population receive food stamps in the first quarter after application for TANF. In three of the study areas, receipt of food stamps fell gradually over the next three quarters. ⁽²⁾
• Survey data from the two California studies show a decline in food stamp receipt between surveys administered at six and twelve months after diversion. However, Illinois survey data show an increase in receipt of food stamps between the time of application and the administration of the survey six to nine months later (from 49 to 66 percent).
• Five applicant studies used administrative data to determine the number of TANF applicants and divertees enrolled in the Medicaid program. Medicaid enrollment among family heads in the fourth quarter following application/diversion ranged from 23 percent to 56 percent. South Carolina was the only site in which more than half of the applicants received Medicaid in any one of the four quarters following their application.
• Survey data collected by a few of the grantees also showed varying levels of enrollment in the Medicaid program. Twelve months after diversion from TANF, 41 percent of those interviewed in Contra Costa County were receiving Medicaid benefits, compared to 55 percent of respondents in Arizona.
• Researchers in Milwaukee County, Wisconsin interviewed individuals at the time they applied to the TANF program. They found that 61 percent of the applicants surveyed were receiving food stamps in the month prior to applying for TANF, while 77 percent of the respondents were enrolled in Medicaid in the preceding month.

Returns to TANF

While those applicants who are diverted do not, by definition, receive TANF at the time of application, many eventually reapply for TANF and receive cash assistance. In fact, those studies that also followed TANF entrants show that equal numbers of those who were originally diverted from TANF cash assistance and those who entered the program in the application quarter are receiving TANF cash assistance at 12 months after the study quarter.

• The six studies that used information from their state or county's TANF administrative data base to report findings in the area of TANF recidivism had a wide range of findings. The percentage of applicants or divertees who were receiving TANF three months after applying and being diverted from the program varied from 22 percent in Contra Costa County to virtually none in South Carolina. A reason for this discrepancy might be South Carolina's definition of their diverted population, described above.
• Three studies asked applicants/divertees about receipt of TANF in their surveys. At twelve months after diversion from the cash assistance rolls, receipt of TANF ranged from 13 percent to 36 percent.

Experience with the TANF Application Process

Participants in four of the studies - Arizona, Illinois, South Carolina, and Contra Costa County - reported findings about applicants' experiences with the TANF application process in their state or county. As these questions were all asked in a slightly different way, it is difficult to synthesize these findings across the studies.

• Both Arizona and Contra Costa County were able to discern from administrative data the reasons for denial of TANF application among so-called "diverted" applicants. In both locations, nearly half of the applications were denied because of a failure to complete the interview process. Other reasons included a failure to comply with program rules or a voluntary withdrawal of the application by the applicant herself.
• Contra Costa County compared its administrative data on the reason for denial with the results of a survey question in which applicants were asked to give the reason they thought their application was denied. Two out of five respondents said their application was denied due to "administrative hassles", while another 19 percent listed employment as the reason for not finishing the application process.
• Arizona found that, among diverted applicants, 39 percent did not complete the application process because of issues related to income, such as finding a job or making too much money. However, 32 percent of respondents cited issues related to the application process (too many hassles or having to provide too much documentation) as the major reason they did not complete the process.
• In Illinois, 20 percent of those respondents whose applications were denied or who withdrew their applicants felt that they did not go onto TANF because they had too much income or assets, while another 19 percent did not provide the paperwork that was required. The majority of these individuals reported receiving job search and job assistance services while their application was being considered.
• South Carolina asked those who were receiving food stamps and were income-eligible for TANF, but not receiving TANF, why they had not gone on welfare. The majority of this population - 59 percent - said that they only wanted food stamps and did not want to be on welfare. An additional 16 percent said they did not know that people could get cash assistance, while another 12 percent said they had not applied for TANF because they did not think they would be eligible.

Overall, the population of TANF applicants and divertees is difficult to define, and thus a hard population for whom to determine outcomes. However, some generalizations can be made across the ASPE-funded studies of TANF applicants and diversion. First, there appears to be a significant dip in earnings just prior to application to TANF - this dip and the decision to apply for TANF might be related somewhat. The drop in the employment rate during this time is much less pronounced than the dip in earnings. In addition, there is moderate use of food stamps and Medicaid among this population, decreasing gradually in the year following the application to or diversion from TANF. Finally, a fair number of application reapply for TANF and eventually receive cash assistance in the year following the initial diversion.

Table 1:
Definitions of Applicant/Diversion Populations Studied by ASPE Grantees Included in this Synthesis

Study population	Arizona	Florida	Illinois	South Carolina	Texas	Milwaukee Co.	Contra Costa Co.	San Mateo Co.	Total
Formally diverted					X				1
Failed to complete application	X		X		X	X	X	X	6
Completed but not enrolled				X		X	X	X	4
Ineligible for non-financial reasons			X		X	X	X	X	5
Appear to be eligible		X		X					2

Endnotes

1.  The range of earnings in the two California study areas in the fourth quarter after application/diversion was approximately $4,100 to $6,200.

2.  South Carolina is not included in this range, as food stamp receipt there was significantly higher than in the other study areas. This is likely due to South Carolina's definition of their diverted population as families who entered food stamps in a particular quarter, were income eligible for TANF, but did not receive TANF cash assistance.