Experiences of Virginia Time Limit Families After Case Closure: An Interim Report

Transcription

1 Contract No.: CR MPR Reference No.: Experiences of Virginia Time Limit Families After Case Closure: An Interim Report Final Report January 2002 Anne Gordon Susanne James-Burdumy Renee Loeffler Barbara Guglielmo Carole Kuhns Submitted to: Virginia Department of Social Services 730 E. Broad Street Richmond, VA Under Subcontract to: Center for Public Administration and Policy Virginia Polytechnic Institute and State University 7054 Haycock Road Falls Church, VA Submitted by: Mathematica Policy Research, Inc. P.O. Box 2393 Princeton, NJ (609) Project Director: Anne Gordon

2

3 ACKNOWLEDGMENTS Many people at VDSS, Virginia Tech, and MPR contributed to this report. At VDSS, Carol Baron, Director of Research and Evaluation, had the initial vision for this study and has worked hard to make it successful and policy-relevant. Carol also provided thoughtful comments on several drafts of the report. Molly Sheahan worked with local offices to identify the sample frames for the surveys and obtain contact information. Mike Theis provided administrative data for the analysis and conducted checks of case records. MPR survey staff were responsible for collecting the data. As Survey Director, Lindsay Dobrzynski developed the instruments, trained and supervised interviewers, and managed the overall survey operation with skill and dedication. Phyllis Schulman supervised the field locators. For round two of the surveys (6-month interviews with cohort 2 and 18-month interviews with cohort 1), Amy Raduzycki, Rob Samardick, Robert Boyle, and Joyce Manowitz supervised the interviewers and locators in the phone center. We especially wish to acknowledge the teams who made trips from New Jersey to Virginia to do field locating of respondents. On the more technical side, Jennifer McNeill developed the 18-month CATI program, and Jennifer McNulty programmed the sample and created the data file, under supervision from Barbara Kolln. Xantha Burghardt coded the open-ended responses under supervision from Marianne Stevenson. Data analysis and report preparation were shared responsibilities of Virginia Tech and MPR. At Virginia Tech, Helen Barker helped develop analysis files, prepared tables, and assisted in editing and proofreading. At MPR, Ruo-Jiao Cao cleaned the data and developed the analysis files. Jane Dokko, Tim Novak, April Grady, and Lucy Lu programmed the tabulations for the MPR portions of the report. We built on work by Roberto Agodini and Amy Zambrowski on the previous report. In addition, Jackie Kauff, an MPR researcher currently at work on preparing the next report, contributed some refinements to the analysis for this report. Walt Corson provided internal review and useful comments. Walt Brower edited the report. Cathy Harper led the production team, with Bill Garrett taking over at the end. iii

4

5 CONTENTS Chapter Page EXECUTIVE SUMMARY...xvii I INTRODUCTION... 1 A. WELFARE REFORM IN VIRGINIA... 2 B. STUDY GOALS... 4 C. PLAN OF THE REPORT... 6 II SAMPLE AND DATA... 9 A. STUDY UNIVERSE AND SAMPLE... 9 B. SURVEY METHODS Six-Month Follow-Up Interviews Eighteen-Month Follow-Up Interviews Questions Asked in the 6- and 18-Month Interviews C. ADMINISTRATIVE DATA SOURCES III CHARACTERISTICS OF TIME LIMIT FAMILIES A. WHO REACHED THE TIME LIMIT? Demographic Characteristics TANF History VIEW Sanctions B. COMPARISON WITH VIEW CASES IN THE SAME AREAS C. TIME LIMIT KNOWLEDGE AND PLANS IV EMPLOYMENT A. WORK EXPERIENCE SINCE THE CASE CLOSED B. JOB CHARACTERISTICS v

6 CONTENTS (continued) Chapter Page C. TRANSPORTATION TO WORK D. COMPARISON OF EMPLOYED AND UNEMPLOYED RESPONDENTS V INCOME SOURCES AND TOTAL INCOME A. RECEIPT OF FOOD STAMPS B. WHY SOME DID NOT GET FOOD STAMPS C. OTHER PUBLIC ASSISTANCE D. CHILD SUPPORT E. EMPLOYMENT OF OTHER HOUSEHOLD MEMBERS F. TRENDS IN RESPONDENTS INCOME G. HOUSEHOLD INCOME AND POVERTY LEVELS AT THE 18-MONTH INTERVIEW H. COMPARISON OF EMPLOYED AND UNEMPLOYED RESPONDENTS VI INFORMAL AND COMMUNITY SUPPORT A. ASSISTANCE FROM FAMILY AND FRIENDS B. ASSISTANCE FROM COMMUNITY AGENCIES OR RELIGIOUS ORGANIZATIONS VII HOUSING AND HOUSEHOLD COMPOSITION A. HOUSING ARRANGEMENTS AND COSTS B. MOVES SINCE CASE CLOSED C. HOMELESSNESS vi

7 CONTENTS (continued) Chapter Page D. HOUSEHOLD SIZE AND COMPOSITION Household Composition Household Size Children No Longer in the Household VIII HEALTH ISSUES A. HEALTH INSURANCE COVERAGE Current Health Insurance Medicaid Coverage After Case Closure Trends in Insurance Coverage B. RESPONDENTS USUAL SOURCE OF CARE C. RESPONDENTS HEALTH Did Respondents Have Health Problems That Limited Work? Work Status and Income of Those with Health Limitations Health Problems of Household Members IX RESPONDENTS VIEWS OF THEIR SITUATION A. HOW DID RESPONDENTS ASSESS THEIR SITUATION? B. CHANGES IN OUTLOOK SINCE THE 6-MONTH INTERVIEW C. WHY SOME THOUGHT THINGS WERE BETTER D. WHY SOME THOUGHT THINGS WERE WORSE X CHILD CARE A. USE OF CHILD CARE ASSISTANCE, AS RECORDED IN ADMINISTRATIVE DATA B. USE OF CHILD CARE SUBSIDIES, AS REPORTED IN THE SURVEY vii

8 CONTENTS (continued) Chapter Page C. REASONS NO LONGER USING SUBSIDIES D. CHILD CARE ARRANGEMENTS E. CHILD CARE COSTS XI CHILD WELL-BEING A. CHARACTERISTICS OF CHILDREN B. INTERACTION WITH AND SUPPORT FROM NONCUSTODIAL PARENT C. CHILDREN S ACCESS TO HEALTH CARE D. CHILDREN S HEALTH STATUS XII A BROADER PERSPECTIVE: 6-MONTH RESULTS FOR COHORTS 1 AND A. EMPLOYMENT AND EARNINGS B. COMPARISONS OF EMPLOYED AND NOT EMPLOYED C. THE SAFETY NET FOOD STAMPS AND MEDICAID D. TRANSITIONAL BENEFITS CHILD CARE AND TRANSPORTATION E. OTHER RESOURCES Child Support Help from Family, Friends, and Community Agencies F. CHANGES IN INCOME AND OUTLOOK G. EXTENT OF HARDSHIP Homelessness or Housing Problems Lack of Health Insurance H. CHILDREN S EXPERIENCES viii

9 CONTENTS (continued) Page REFERENCES APPENDIX A: COMPARISON OF SURVEY RESPONDENTS AND NONRESPONDENTS... A.1 APPENDIX B: EMPLOYMENT AND EARNINGS: DETAILED TABLES OF 6-MONTH DATA FOR COHORTS 1 AND 2...B.1 APPENDIX C: INCOME SOURCES: DETAILED TABLES OF 6- MONTH DATA FOR COHORTS 1 AND 2...C.1 APPENDIX D: FAMILY AND CHILD WELL-BEING: DETAILED TABLES OF 6-MONTH DATA FOR COHORTS 1 AND 2... D.1 ix

10

11 TABLES Table Page II.1 SAMPLE SIZES FOR THE TIME LIMIT STUDY II.2 SURVEY RESPONSE RATES FOR 6-MONTH INTERVIEWS III.1 DEMOGRAPHIC CHARACTERISTICS OF TIME LIMIT CASES IN COHORTS 1 AND III.2 TANF HISTORY OF TIME LIMIT CASES III.3 COMPARISON OF TIME LIMIT CASES TO VIEW-MANDATORY CASES III.4 LEARNING ABOUT THE TIME LIMIT III.5 PLANS FOR COPING WITH THE TIME LIMIT III.6 RESPONDENTS PERCEPTIONS OF THE VIEW PROGRAM IV.1 WORK EXPERIENCE SINCE CASE CLOSED IV.2 JOB RETENTION IV.3 REASONS FOR NOT WORKING IV.4 IV.5 START OF FIRST JOB, AMONG THOSE WHO WORKED AFTER CASE CLOSED MOVEMENTS BETWEEN JOBS, AMONG THOSE WHO WORKED AFTER CASE CLOSURE IV.6 REASONS FOR JOB EXITS IV.7 IV.8 CHARACTERISTICS OF CURRENT OR MOST RECENT JOBS HELD BY TIME LIMIT RESPONDENTS CHANGES IN HOURS AND EARNINGS ON CURRENT OR MOST RECENT JOB IV.9 TRANSPORTATION TO WORK IV.10 FACTORS ASSOCIATED WITH BEING EMPLOYED AT THE 18-MONTH INTERVIEW xi

12 TABLES (continued) Table Page V.1 FOOD STAMP PROGRAM PARTICIPATION AND BENEFITS V.2 RESPONDENTS VIEWS ON WHY NOT RECEIVING FOOD STAMPS AT INTERVIEW V.3 PUBLIC ASSISTANCE BEFORE AND AFTER CASE CLOSURE V.4 CHILD SUPPORT RECEIVED BY TIME LIMIT FAMILIES V.5 CHILD SUPPORT BEFORE AND AFTER CASE CLOSURE V.6 EMPLOYMENT OF OTHER HOUSEHOLD MEMBERS V.7 RESPONDENTS MONTHLY INCOME BEFORE AND AFTER THE CASE CLOSED V.8 CHANGES IN RESPONDENTS MONTHLY INCOME SINCE TANF BENEFITS TERMINATED V.9 KNOWLEDGE AND USE OF THE EARNED INCOME TAX CREDIT (EITC) V.10 REPORTED TOTAL MONTHLY INCOME AT THE 18-MONTH INTERVIEW V.11 POVERTY STATUS V.12 SOURCES OF INCOME IN THE MONTH BEFORE THE 18-MONTH INTERVIEW, BY CURRENT EMPLOYMENT STATUS V.13 INCOME FROM SOURCES OTHER THAN EARNINGS IN THE MONTH BEFORE THE 18-MONTH INTERVIEW, BY CURRENT EMPLOYMENT STATUS V.14 TOTAL INCOME IN THE MONTH BEFORE THE 18-MONTH INTERVIEW, BY CURRENT EMPLOYMENT STATUS V.15 POVERTY STATUS IN THE MONTH BEFORE THE 18-MONTH INTERVIEW, BY CURRENT EMPLOYMENT STATUS VI.1 ASSISTANCE FROM FAMILY OR FRIENDS IN PAST MONTH VI.2 CHANGES IN ASSISTANCE FROM FAMILY OR FRIENDS xii

13 TABLES (continued) Table VI.3 VII.1 VII.2 Page ASSISTANCE FROM COMMUNITY AGENCIES OR RELIGIOUS ORGANIZATIONS COMPARISON OF HOUSING AT CASE CLOSURE AND AT 6- AND 18-MONTH INTERVIEWS MOVES AND CHANGE IN QUALITY OF HOUSING AMONG RESPONDENTS WHO MOVED VII.3 COMPARISON OF HOUSING AMONG THOSE WHO MOVED VII.4 VII.5 HOUSING AMONG THOSE WHO MOVED AND REPORTED A CHANGE FOR THE BETTER ASSESSMENT OF CURRENT LIVING CONDITIONS BY THOSE WHO MOVED AND REPORTED HOUSING TO BE BETTER VII.6 HOMELESSNESS VII.7 HOUSEHOLD COMPOSITION VII.8 HOUSEHOLD SIZE VIII.1 CURRENT HEALTH INSURANCE VIII.2 MEDICAID COVERAGE AFTER CASE CLOSED VIII.3 TRENDS IN INSURANCE COVERAGE OVER TIME VIII.4 RESPONDENT S USUAL SOURCE OF CARE, BY HEALTH INSURANCE STATUS VIII.5 HEALTH LIMITATIONS VIII.6 VIII.7 EMPLOYMENT AND OTHER SUPPORT AMONG RESPONDENTS WITH HEALTH LIMITATIONS HEALTH PROBLEMS OF OTHERS THAT AFFECT RESPONDENTS ABILITY TO WORK IX.1 RESPONDENTS OUTLOOK AT 6 MONTHS COMPARED TO 18 MONTHS xiii

14 TABLES (continued) Table IX.2 IX.3 Page REASONS WHY THINGS ARE BETTER FOR RESPONDENT SINCE CASE CLOSED REASONS WHY THINGS ARE WORSE FOR RESPONDENT SINCE CASE CLOSED X.1 SUBSIDIZED CHILD CARE PARTICIPATION AND PAYMENTS FOR FULL SAMPLE X.2 USE OF CHILD CARE SUBSIDIES X.3 USE OF CHILD CARE SUBSIDIES 18 MONTHS AFTER CASE CLOSURE X.4 REASON NO LONGER USING CHILD CARE SUBSIDIES X.5 CURRENT CHILD CARE ARRANGEMENTS, BY AGE OF CHILD X.6 NUMBER OF HOURS PER WEEK IN CHILD CARE, BY AGE X.7 OUT-OF-POCKET MONTHLY CHILD CARE EXPENSE FOR TIME LIMIT FAMILIES WITH CHILD CARE NEEDS XI.1 CHARACTERISTICS OF CHILDREN XI.2 NONCUSTODIAL PARENTS INTERACTION WITH AND SUPPORT FOR THEIR CHILDREN XI.3 SOURCE OF HEALTH CARE FOR CHILDREN XI.4 CHILDREN S HEALTH AND BEHAVIOR XII.1 EMPLOYMENT FROM CASE CLOSURE TO THE 6-MONTH INTERVIEW XII.2 FACTORS ASSOCIATED WITH BEING EMPLOYED AT THE 6- MONTH INTERVIEW xiv

15 FIGURES FIGURE PAGE IV.1 PERCENTAGE OF TIME LIMIT RESPONDENTS WORKING IV.2 AVERAGE HOURS WORKED PER WEEK BY TIME LIMIT RESPONDENTS IV.3 AVERAGE MONTHLY EARNINGS OF TIME LIMIT RESPONDENTS IV.4 AVERAGE HOURLY RATE OF PAY FOR TIME LIMIT RESPONDENTS IX.1 HOW ARE THINGS SINCE TANF CASE CLOSED IX.2 HOW ARE THINGS, BY CURRENT EMPLOYMENT STATUS xv

16

17 EXECUTIVE SUMMARY Most welfare recipients now face a time limit on their eligibility for cash assistance. The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 instituted a five-year lifetime limit on federal cash assistance for most recipients and permitted states, under the new Temporary Assistance for Needy Families (TANF) program, to set shorter time limits. Some states, including Virginia, had already begun to implement time limits under waivers. Because time-limited welfare is relatively new, policymakers and the public at large have been concerned about what happens to families who lose TANF benefits because of time limits. Because time limit policies vary widely, this question can only be answered state by state. In 1995, Virginia, as part of its welfare reforms, instituted a 24-month time limit on benefits under the Virginia Initiative for Employment not Welfare (VIEW). To provide reliable information on time limit families and what happens to them after reaching the time limit, the Virginia Department of Social Services (VDSS) contracted with Virginia Tech and Mathematica Policy Research, Inc. (MPR), for a longitudinal study. The study includes analysis of administrative data and of surveys of time limit families conducted about 6 and 18 months after their TANF cases closed. This is the second of four planned reports from the Virginia Time Limit Study. It presents 18 months of follow-up data on families whose TANF cases closed because of the time limit in early 1998 (cohort 1) and 6 months of follow-up data for a larger sample that includes families that reached the time limit in early 1998 and early 1999 (cohorts 1 and 2). Key findings include: In the 18 months after their cases closed, nearly all cohort 1 parents worked, and most worked for more than half of the follow-up period. Their jobs paid low wages, but average hours and wages increased over time. Cohort 1 survey respondents average incomes increased between the 6-month and 18-month interviews. At the 18-month interview, 26 percent of cohort 1 families reported that their total household income was above the poverty line. By the 18-month interview, 80 percent of cohort 1 respondents had health insurance coverage for their children, but less than one-third had coverage for themselves. Time limit families in both cohorts were long-term welfare recipients. The median length of time they had been receiving public assistance was over five years. They were similar to other VIEW cases in their areas in education and ethnicity. However, time limit family heads were older and had more children than heads of VIEW cases in general. xvii

18 Six-month interview findings for cohorts 1 and 2 combined were generally similar to previously published findings for cohort 1. However, employment rates were slightly lower and wage rates were higher in the broader sample, probably because it included more urban residents. BACKGROUND: VIRGINIA S TIME LIMIT In Virginia, the time limit is just one aspect of VIEW, which is mandatory for able-bodied TANF parents with no children under the age of 18 months. Key provisions of VIEW include: The signing of an Agreement of Personal Responsibility (APR) as a condition for receiving benefits Required job search for 90 days or until employed, followed by mandatory work either through regular employment or through participation in the Community Work Experience Program (CWEP), which involves work in a nonprofit or public setting in exchange for benefits Full family sanctions (complete loss of benefits) for noncompliance Generous earned income disregards, which allow families to continue to receive their full TANF grant for up to 24 months as long as their net earned income plus TANF benefits remains below the poverty line Supportive services, including subsidized child care, transportation assistance, and Medicaid, while on TANF and for one year after the TANF case closes A 24-month time limit on TANF benefits, followed by 12 months of eligibility for transitional benefits, and 24 months of ineligibility for TANF 1 All VIEW-mandatory cases are subject to the 24-month time limit. Of the 34,534 cases enrolled in VIEW between July 1995 and June 1998, however, only 5,673 cases, or 16 percent, had reached the time limit by the end of June Others have generally left TANF before reaching the time limit. The structure of VIEW implies that time limit families in Virginia are not necessarily like families that reach TANF time limits in other states. The Virginia time limit applies only to mandatory VIEW cases. After the first 90 days in VIEW, VIEW-mandatory cases must work at least 30 hours per week, take a CWEP position, or lose 100 percent of their TANF benefit. Months in which benefits are suspended because of failure to meet VIEW requirements still 1 Extensions are available under some circumstances, but VDSS reports that less than 3 percent of time limit families have received them. Families may reapply for TANF 24 months after they stop receiving benefits, including transitional benefits, so those who do not use transitional benefits may actually return to TANF earlier than those who use them. xviii

19 count toward the time limit unless the case head takes action to close the case. Thus, heads of cases that reach the time limit fall largely into two groups: (1) those who have been working and meeting VIEW requirements for some time; and (2) those who, because of a VIEW sanction, stopped receiving benefits before reaching the time limit, and then had their case officially closed upon reaching the time limit. Most are in the first group. SAMPLE AND DATA This report is based on two cohorts of Virginia TANF cases: (1) cases that closed because of the time limit between February 1, 1998, and June 30, 1998 (cohort 1); and (2) cases that closed because of the time limit between February 1, 1999, and June 30, 1999 (cohort 2). 2 Because of the staggered implementation of VIEW in Virginia, only a few parts of the state had cases that reached the time limit during the 1998 period. Most of these cases were in rural areas and small towns, or in relatively wealthy, suburban Northern Virginia. By early 1999, roughly half the state had cases reaching the time limit, including the cities of Richmond and Petersburg. Cohort 1 included 328 cases, and cohort 2 included 628 cases. Administrative data for all these cases were analyzed, and 6-month interviews were attempted with heads of all cases. For cohort 1, 18-month follow-up interviews were attempted with all respondents who completed the 6-month interview. 3 The number of cohort 1 cases that completed both 6- and 18-month interviews is 220, which is 67 percent of all cohort 1 cases. The number of cases in both cohorts that completed 6-month interviews is 751 (256 in cohort 1 and 495 in cohort 2), 79 percent of the two cohorts. Both interviews lasted from 30 to 40 minutes and were conducted using computer-assisted telephone interviewing. To increase response rates, field staff helped find some respondents, then provided cellular phones, if needed, for them to call MPR to complete the interview. STUDY GOALS The goals of the time limit study are to describe (1) who reached the VIEW time limit, and (2) how their lives changed after they lost their TANF benefits. An important caution is that this study is descriptive. It cannot show whether changes occur in people s lives because of the loss of TANF benefits or whether these changes would have come about anyway, because there is no control or comparison group to show what would have happened to these families without a time limit. 2 Because the number of time limit cases during these periods was small, all cases were included in the study; no sampling occurred. Cohort 2 includes a few cases that closed in July 1999, although they had been scheduled to close in June. 3 Later reports will present results from 18-month interviews with cohort 2 cases, as well as from 6 and 18-month interviews with a third cohort of cases that reached the time limit in early xix

20 FINDINGS WHO REACHED THE TIME LIMIT? The head of a typical time limit family was a single mother in her thirties. Most had one or two children, but over a third had three or more. More than two-thirds of the parents were African American, and a quarter were white. Forty-five percent of the sample had not completed high school or received a GED. 4 Parents in time limit families were similar in race, ethnicity, and education to all VIEW participants from the same communities, but they were somewhat older and had somewhat larger families. Most time limit families had been on public assistance for a long time. Taking into account periods off public assistance, more than half (58 percent) had received benefits for longer than five years. Most time limit parents complied with VIEW rules, but 32 percent had been sanctioned at least once for not meeting program requirements. 18-MONTH RESULTS FOR COHORT 1 The 18-month interviews with cohort 1 parents provide early indicators of long-term outcomes for time limit families. These data are among the first that follow families that reach time limits for more than a few months. Furthermore, the 18-month interview occurred after families had some time to adjust to life without TANF benefits, and also reflects a period after they had lost eligibility for transitional Medicaid, transportation, and child care assistance. 5 However, as noted above, time limit parents in cohort 1 were concentrated in rural and suburban parts of the state. Thus, in reviewing the 18-month follow-up results, readers should keep in mind that the experiences of cohort 1 families may not be representative of those in other parts of Virginia. HOW MUCH DID THEY WORK AFTER BENEFITS ENDED? Nearly all respondents had worked since leaving TANF, and many had worked steadily. Ninety-two percent of respondents worked at some point between case closure and the 18-month interview. On average, respondents worked in 73 percent of the follow-up months. 4 Characteristics of the time limit families are as recorded in TANF administrative data at the time of case closure. 5 Many children remained eligible for Medicaid based on poverty status. Most working families would also be eligible for other state child care assistance. xx

21 More than 80 percent of respondents were working at least 30 hours per week in their current or most recent jobs. Half worked in service occupations. The current or most recent job paid $7.05 per hour, on average. WERE THEY MOVING TOWARD SELF-SUFFICIENCY? Employment, Earnings, and Child Support Although most respondents were working before their TANF case closed, slightly more respondents worked after. Sixty-two percent were working in the month the case closed, and 69 percent were working at the time of the 18-month interview. Among workers, hourly wages, hours worked, and earnings increased between the 6- and 18-month interviews. Current or most recent jobs held at the 18-month interview paid $7.05 per hour on average up from an average hourly wage of $6.26 at the 6- month interview. Average hours worked increased from 34 to 38 hours per week between the 6- and 18-month interviews. On average, monthly earnings of workers increased by 25 percent between the two interviews, from $935 to $1,170. The percentage of time limit families receiving child support increased over the follow-up period. Administrative data show that the proportion receiving child support increased from 23 percent to 37 percent in the 18 months after case closure. The average amount of child support collected by those receiving child support increased from $49 to $211 over the same period. Reliance on Food Stamps, Medicaid, and Subsidized Child Care The percentage of time limit families receiving food stamps declined after case closure. Administrative data show that the percentage of families receiving food stamps fell from 83 percent at case closure to 63 percent 18 months later. Medicaid continued to cover most children, but not their parents. As of the 18-month interview, about 75 percent of children in time limit families were covered by Medicaid. However, only 25 percent of respondents were still covered by Medicaid. According to administrative data, 18 months after case closure, the percentage of families receiving child care subsidies had fallen by two-thirds. In particular, 11 percent of all time limit families received a state-administered child care subsidy at 18 months, while 33 percent had received a subsidy in their final month on TANF. Many families did not need a subsidy because they did not use or need paid child care in order to work. xxi

22 Help from Family, Friends, and Community Agencies At the 18-month interview, 62 percent of respondents had received help from family or friends, and 24 percent had received help from community or religious organizations in the past month. HOW ARE FAMILIES DOING? Changes in Income and Outlook Respondents average income increased between case closure and the 18-month interview. 6 Between case closure and the 18-month interview, average monthly income increased from $973 to $1,071. The incomes of 45 percent of respondents increased between case closure and the 18-month interview, yet income fell for 41 percent of time limit respondents. 7 However, these figures may understate growth in household income, as more respondents lived with other earners as time went on. At the 18-month interview, one-fourth (26 percent) of time limit families reported incomes above the poverty line. At the same time, 74 percent of respondents reported incomes below the poverty line, and 26 percent reported incomes below 50 percent of the poverty line. Most families reported knowing about and using the Earned Income Tax Credit (EITC). Eighty percent of time limit families had heard of the EITC, an important source of income for low-income working families. Fifty-six percent had received the EITC. Most respondents felt their lives were the same or better, and their outlook had improved over time. Nearly three-quarters reported their situation was the same as or better at the 18-month interview than when they received TANF; 34 percent reported things were better and 39 percent that they were about the same. Only 27 percent reported they were worse off, down from 43 percent at the 6-month interview. Extent of Hardship Almost no respondents reported being homeless after case closure. Less than two percent of the sample reported ever having been homeless since their TANF case closed. 6 Income is calculated as the sum of respondents earnings, TANF, food stamps, child support, and other benefits, as reported in the surveys. 7 Income is considered unchanged if it increased or decreased by less than 10 percent. xxii

23 About two-thirds of parents and one-fifth of children lacked health insurance. Health coverage, especially for adults, is thus the major area in which most time limit families had become worse off over time. Changes in Household Structure Respondents were more likely to be living with a spouse by the 18-month interview. The percentage of households containing the respondent s spouse increased from 9 percent at the 6-month interview to 13 percent at the 18-month interview. WHAT WERE CHILDREN S SITUATIONS? Most children in time limit families were 5 to 12 years old, and most had one parent who lives elsewhere. Specifically, 65 percent of children were between the ages of 5 and 12, and 92 percent had one parent who lives elsewhere. Almost no children moved out of the household after case closure. Less than one percent of respondents minor children had moved to another household since the TANF case closed. Most children in time limit families (69 percent) used a private physician for their regular source of health care. Almost three-quarters (72 percent) had a well-child physician s visit in the last year. Most respondents with child care needs (93 percent) used only one nonschool child care arrangement, generally some type of formal care. No one had more than two such arrangements. In the year before the 18-month interview, 11 percent of children had regularly scheduled contact with their noncustodial parent. Over half of children never had contact with their noncustodial parent in that year. 6-MONTH RESULTS FOR A BROADER SAMPLE We analyzed data from the 6-month interviews with cohorts 1 and 2 combined to obtain a broader perspective on early outcomes for time limit families. Cohorts 1 and 2 together provide a sample that is both much larger and more representative of Virginia as a whole. Cohort 1 is disproportionately from rural areas and smaller towns, while cohort 2 covers a wider area of the state and includes the more urban areas of Richmond and Petersburg. Cohort 2 parents are also younger, more often African American, and somewhat less educated on average than cohort 1 parents. Overall, 6-month findings for cohorts 1 and 2 are similar to those for cohort 1 alone, as presented in Gordon et al. (1999). In both groups, most time limit parents worked both before and after reaching the time limit. Nonetheless, there are differences in employment rates and earnings. The broader sample worked less steadily than did the cohort 1 sample. As in the xxiii

24 results for cohort 1 alone, 85 percent of the combined sample had worked at some point between case closure and the 6-month interview. However, a smaller proportion were working at each point in time. About 61 percent of respondents in the combined sample worked in the month the case closed, and 66 percent were working at the interview, seven months later on average. Among the cohort 1 sample, 63 percent were working in the month the case closed, and 71 percent were working at the 6-month interview. At the same time, despite their lower education levels, respondents in the combined cohort 1 and 2 sample were in jobs with higher hours and earnings, on average, than those in cohort 1. In the current or most recent job held as of the 6-month interview, respondents in the combined sample earned $6.50 per hour on average. Respondents in cohort 1 earned $5.99 per hour on average. The higher earnings of the broader sample (relative to cohort 1 alone) probably are due to the fact that the cohort 2 sample is more urban, and urban areas tend to have a higher cost of living and higher pay scales. 8 Other key findings from the 6-month data for cohorts 1 and 2 include: Based on survey data, average monthly income of respondents was slightly higher at the 6-month interview than before the case closed ($946 versus $899). Families had lost TANF benefits but more than filled the gap, on average, with increased income from earnings, food stamps, and child support. At the 6-month interview, 77 percent of respondents reported their children were covered by Medicaid, and 72 percent reported that they themselves were covered. Only about two percent of families reported being homeless since their case closed. In sum, in the broader sample as in cohort 1, we find most families were working, but incomes remained low. Future reports will provide longer-term findings for all three cohorts of Virginia time limit cases. 8 Data for cohort 2 are also from one year later than the cohort 1 data (and have not been adjusted for inflation); other factors may also be involved. xxiv

25 I. INTRODUCTION Welfare reform in Virginia includes a 24-month time limit on benefits for able-bodied parents with no children younger than 18 months. The time limit may affect a substantial proportion of eligible welfare cases. However, of the 34,534 cases enrolled in Virginia s welfare reform work program (the Virginia Initiative for Employment not Welfare [VIEW]) between July 1995 and June 1998, only 5,673 cases (16 percent) had reached the time limit by the end of June Because the 24-month time limit is a major departure from previous practice, policymakers and the public at large are greatly interested in its outcomes. Therefore, the Virginia Department of Social Services (VDSS) has contracted for a longitudinal study to obtain reliable information on time limit families and what happens to them. This second report covers the experiences of an early cohort of families that reached the time limit in early 1998, during approximately the first 18 months after their benefits ended. Administrative data on 328 cases are presented, along with results from 6- and 18-month followup interviews with 220 of these cases. In addition, this report describes the results from 6-month follow-up interviews with a larger sample, which includes families who reached the time limit in early 1999 (cohort 2) as well as those who reached it in early 1998 (cohort 1). Future reports will extend the follow-up period for cohort 2 to 18 months, and will also add data on a third cohort of cases who reached the time limit in the first half of The rest of this chapter provides background on welfare reform in Virginia. It then describes the questions the time limit study addresses, and lays out the plan of this report. 1 These are the most recent data available. Statistics were provided by Carol Baron of the Virginia Department of Social Services. 1

26 A. WELFARE REFORM IN VIRGINIA Virginia s multifaceted welfare reform program the Virginia Independence Program (VIP) has two distinct components. The first is changes made to Aid to Families with Dependent Children (AFDC) eligibility requirements in hopes of encouraging family responsibility. The second, VIEW, is one of the nation s strongest examples of a work first program, which emphasizes rapid movement of public assistance clients into jobs. Overall, VIP represents a substantial commitment to changing the culture of welfare both for program staff and for clients. The VIP eligibility requirements were implemented statewide on July 1, VIEW was implemented locality by locality from July 1995 to October In 1996, the federal Personal Responsibility and Work Opportunity Reconciliation Act replaced AFDC with a block grant for Temporary Assistance for Needy Families (TANF). Because Virginia had already shifted its AFDC program to a temporary assistance program with employment as its focus, the state implemented TANF in February 1997 with minimal modifications to VIP. The VIP eligibility requirements include: Stronger requirements for cooperation with child support enforcement A family cap on benefits for children born more than 10 months after assistance is authorized Age-appropriate immunizations for children Compliance with school attendance laws Determination of benefits for two-parent families using the same standards as for single-parent families VIEW is mandatory for able-bodied parents with no children under the age of 18 months. Key provisions of VIEW include: 2

27 The signing of an Agreement of Personal Responsibility as a condition for receiving benefits Required job search for 90 days or until employed, followed by mandatory work either through regular employment or participation in the Community Work Experience Program (CWEP), which involves work in a nonprofit or public setting in exchange for benefits Full family sanctions (complete loss of benefits) for noncompliance 2 Generous earned income disregards, which allow families to continue to receive their full TANF grant as long as their net earned income plus TANF benefits remains below the poverty line Supportive services, including subsidized child care, transportation assistance, and Medicaid, while on TANF and for one year after the TANF case closes A 24-month time limit on TANF benefits, followed by eligibility for transitional benefits for up to 12 months, and 24 months of ineligibility for all TANF services after transitional benefits end The structure of VIEW implies that time limit families in Virginia are not necessarily like families that reach TANF time limits in other states. The Virginia time limit applies only to mandatory VIEW cases. The proportion of Virginia s TANF caseload that is VIEW-mandatory has been declining as VIEW cases leave the rolls. After the first 90 days in VIEW, VIEWmandatory cases must work at least 30 hours per week, take a CWEP position, or lose 100 percent of their TANF benefit. Months in which benefits are suspended because of failure to meet VIEW requirements still count toward the time limit unless the case head takes action to close the case. Thus, heads of cases that reach the time limit will fall into two groups: (1) those who will have been working and meeting VIEW requirements for some time, 3 and (2) those who 2 Failure to sign the Agreement of Personal Responsibility results in case closure. Failure to comply with the job search or work requirements after signing the agreement results in a 100 percent sanction for a minimum period. During the sanction period, the months count toward the time limit unless the client chooses to close the case. 3 A few may have been in CWEP positions for part of their time in VIEW, but even those in CWEP will probably have worked in unsubsidized jobs for most of their 24 months. 3

28 may have stopped receiving benefits because of a VIEW sanction some months before their case was officially closed because of the time limit. Most people are expected to fall into the first group. B. STUDY GOALS The time limit study is one of five studies sponsored by VDSS to examine the implementation, outcomes, and impacts of welfare reform in Virginia. 4 VDSS contracted with the Virginia Polytechnic Institute and State University (Virginia Tech) and its subcontractor, Mathematica Policy Research, Inc. (MPR), to conduct the studies. Funding was provided by the U.S. Department of Health and Human Services, Administration for Children and Families, and by VDSS. The goals of the time limit study are to describe who reaches the time limit, what is happening to people in cases that reached the time limit 6 months and 18 months later, and how their lives changed since they lost their TANF benefits. An important caution is that this study is descriptive. It cannot show whether changes occur in people s lives because of the loss of TANF benefits or whether these changes would have happened anyway, because there is no control or comparison group to show what would have happened to these families without a time limit. The first report from this study (Gordon et al. 1999) covered the experiences of families that reached the time limit in early 1998 in approximately the first six months after their benefits ended. This second report updates the experiences of these cases to one year later, approximately 18 months after their benefits ended, based on a second follow-up interview and 4 The other four studies are (1) an early impact and outcomes analysis, based on cases randomly assigned to VIP/VIEW or to the old AFDC and JOBS programs; (2) an implementation study of VIP/VIEW; (3) an impact and implementation study of VIEW-PLUS, a new job retention program; and (4) a study of VIEW-exempt cases, focusing on child-only cases. 4

29 on administrative data. This report also provides six-month results for a broader sample of cases that includes both cohort 1 and cohort 2. This sample is more representative of Virginia as a whole and of cases that have reached the time limit to date. The research questions this report addresses are as follows: 1. Who reached the time limit and why? What are the characteristics of families that reached the time limit? How many were in compliance with VIEW for the full 24 months, and how many had been sanctioned? Did families have a plan for coping with the time limit? If so, were they able to follow their plan? 2. To what extent were heads of families who reached the time limit able to find jobs and stay employed? How many were working before benefits ended? How many worked after benefits ended? What types of jobs did they hold, and did these jobs become better over time? Were family heads able to increase their employment or earnings over time? 3. To what extent are time limit families moving toward self-sufficiency? - Do time limit families use transitional programs or other programs for which they may be eligible after their case closes, such as food stamps, Medicaid, child care assistance, and transportation assistance? Do they find other sources of income or assistance to replace transitional Medicaid, child care, or transportation assistance when their eligibility ends? - How much are they relying on help from family and friends? How much are they using community agencies? Does reliance on these sources decrease over time? 4. How are families doing? - How many families are having problems meeting their basic needs and those of their children? Indicators of problems meeting basic needs considered in this report include the following: (1) income had declined by 10 percent or more since the case closed, (2) household income was below 50 percent of the poverty level in the month before the interview, (3) the respondent reported being homeless since the case closed, (4) the respondent reported lacking health insurance. - How many families are doing better? Indicators that clients are doing better include (1) income up 10 percent or more since the case closed, and (2) household income above 130 percent of the poverty level. 5. What is the situation of children in families that have lost benefits? What types of child care arrangements do children experience? Are children being sent to live with relatives? Are they receiving child support? How many children have significant health or behavior problems? Do they have access to health care? 5

30 Results in this report are preliminary for several reasons. Cohort 1, the 1998 group of time limit cases for which we have 18 months of follow-up data, is small and is drawn from selected parts of the state. Because VIEW was implemented gradually, the only sites to have cases reach the time limit in early 1998 were those that implemented VIEW early, in late 1995 or early Thus, these cases are drawn from only 4 of Virginia s 18 Economic Development Districts (EDDs): (1) District 2, the Bristol/Galax area, a rural area in the southwest; (2) District 6, the large urban and suburban counties in Northern Virginia near Washington, DC; (3) District 9, the Lynchburg area, in the Piedmont region; and (4) District 7, the Culpeper area, which is also largely rural and was the first place to implement VIEW. Furthermore, these cases reached the time limit in close to the minimum time allowed so most would have been in VIEW for 24 consecutive months. Cohort 2 is drawn from a larger part of the state, but still includes cases from only half the state s EDDs. The data from cohorts 1 and 2 combined include cases from the urban areas of Richmond, Petersburg, and Lynchburg; the Northern Virginia suburbs of Washington, DC; and rural areas and smaller towns in various parts of the state (see Chapter III for the full distribution). However, about half the state, including most notably the Tidewater region, is not represented, because these localities implemented VIEW later and thus had no cases that had reached the time limit in early Cohort 3 will encompass cases from the entire state and will include cases that are among the first to reach the time limit in their areas, as well as cases from areas in which the time limit has been in place for several years. C. PLAN OF THE REPORT Chapter II describes the sample and data used in this study, and Chapter III describes the characteristics of time limit cases in the two cohorts. Chapters IV to XI analyze the 18-month interviews with cohort 1 cases and administrative data on selected outcomes over the 18 months 6

31 after the case closed. Chapter XII summarizes the six-month data on cohorts 1 and 2 combined, with emphasis on the changes since the first report on cohort 1 alone. Detailed tables from the analysis of the 6-month interviews are presented in the appendixes. 7

32

33 II. SAMPLE AND DATA To learn about time limit families, we selected samples, conducted interviews, analyzed survey data, and also analyzed available administrative data on these samples. Here, we present the data collection methods used in the 6-month and 18-month follow-up interviews with time limit parents, and describe the topics covered in the interviews. A. STUDY UNIVERSE AND SAMPLE The universe for this study is Virginia time limit cases that closed because they reached the 24-month time limit in selected periods. Cohort 1 includes all Virginia TANF cases that closed because they had reached the 24-month time limit between February 1, 1998, and June 30, As noted in Chapter I, because of the staggered implementation of VIEW in Virginia, in only a few parts of the state did cases reach the time limit during this period. Cohort 2 includes all Virginia TANF cases that closed because of the time limit between February 1, 1999, and June 30, More cases reached the time limit in this period, and the areas they come from include roughly half the state. A period one year after the selection of cohort 1 was chosen to save on data collection costs, since the field periods for the 6-month followup for cohort 2 and the 18-month followup of cohort 1 then overlapped. In addition, this period was chosen to balance the need for timely information on time limit cases and the desire to wait until more of the state would have cases reaching the time limit. 2 1 Cases that reached the time limit but were granted hardship extensions were excluded from the study universe; less than three percent of time limit cases through June 2000 have received hardship extensions, according to data provided by VDSS. 2 A third cohort of time limit cases was selected among those who reached the time limit between February 1, 2000, and June 30, Data on 6- and 18-month follow-up interviews with these cases will be presented in future reports. In February through June 2000, TANF 9

34 Identifying time limit cases in cohorts 1 and 2 was a complex process that involved several steps. VDSS evaluation staff initially identified VIEW cases that were in their 23rd month of benefits and attempted to obtain enhanced contact information on these cases. 3 Several months later, before data collection began, VDSS staff checked administrative records to determine whether in fact the case had closed because of the time limit. 4 Cases that had not closed for this reason were dropped. 5 During and after the survey period, a few more cases were dropped from the study universe, because further scrutiny of administrative records indicated (1) that they had closed before using up their 24 months of eligibility, (2) that they had received an extension, or (3) that they were otherwise not part of the study universe. 6 After final cleaning, the universe of cases reaching the time limit during February through June 1998 (cohort 1, or the 1998 cohort) includes 328 cases. The universe of cases reaching the time limit in the corresponding period in 1999 (cohort 2) includes 628 cases. All these cases are included in the analyses of available administrative data (Table II.1). Six-month follow-up (continued) clients throughout the state were reaching the time limit, and there were more cases than needed to reach the sample sizes targeted for the surveys. Thus, for the first time in 2000, we selected a random sample of the cases reaching the time limit during the selection period, instead of including all cases. 3 VDSS notified caseworkers to update the contact information on these cases when they conducted exit interviews. In addition, caseworkers were asked to obtain information on a relative or friend who would know how to reach the case head if possible. Caseworkers submitted the contact information to VDSS evaluation staff on a standard form, and VDSS created a sample database and forwarded it and the forms to MPR. 4 Records were checked for a closure date within the study period. 5 Some cases had closed in the 23rd month and had thus not reached the time limit. Other cases became exempt or were granted extensions. 6 For cohort 2, we kept a small number of cases that were scheduled to close in June 1999 but had actually closed in July. 10

35 TABLE II.1 SAMPLE SIZES FOR THE TIME LIMIT STUDY Survey Data Administrative Data 6-Month Interview 18-Month Interview Cohort Cohort NA Total NA NA = not available for this report. 11

36 interviews were attempted with all these cases, while 18-month interviews for cohort 1 were attempted with all who had completed a 6-month interview. B. SURVEY METHODS MPR conducted the 6- and 18-month follow-up surveys with time limit families using computer-assisted telephone interviewing. The 6-month interviews took 35 to 40 minutes on average, and the 18-month interviews took about 30 minutes. 1. Six-Month Follow-Up Interviews Each month from August to December 1998, MPR began to interview time limit cases that had closed six months earlier. Thus, interviews with cases that closed in February began in August, interviews with cases that closed in March began in September, and so forth. A similar process was used for the cohort 2 cases in For cases not reached right away, we continued interview attempts until we completed an interview or reached the end of the survey period (early February 1999 for cohort 1 and the end of February 2000 for cohort 2). Most cases were interviewed from 6 to 8 months after the case closed (72 percent of completes), but the remaining completed cases were interviewed from 9 to 14 months after the case closed. Survey operations proceeded in several stages. A week before interviews were scheduled to begin with each part of the sample, we sent an advance letter to the respondents in that group, explaining the study and offering a $10 incentive for completing the interview. The letter asked respondents to call MPR s toll-free number at their convenience to complete the interview. Those who did not call or were not reached by the interviewers within a few weeks were referred to MPR s locating department, which used methods such as address corrections, reverse phone directories, and checks with on-line databases from credit bureaus to find updated addresses or phone numbers. In addition, VDSS staff checked their data systems for updated information 12

37 several times during the field period and found some new addresses or phone numbers. To reach respondents who did not have phones or whose phone number was not available, MPR locating staff sent letters to each new address identified, requesting that the respondent call MPR s tollfree number. If we believed an address to be current but found no phone number, we sent additional letters sent regularly, with content and formats that varied to get the respondent s attention. After phone center staff had tried for several weeks to obtain interviews through phone and letter contacts, we assigned cases not reached to field staff if likely addresses had been identified. Field staff tried to locate the respondent in person. When respondents were not home, the locator left a note and, in some cases, checked with neighbors to ensure that the address was correct. When respondents were found, the field locator asked them to call a toll-free number at MPR from an available phone or provided them with a cellular phone to use. The field locators were able to pay the respondents immediately if they completed the interview. Between August 1998 and February 1999, 256 interviews with eligible respondents in cohort 1 were completed (Table II.2). Another 13 cases were not interviewed because of language barriers, as limited resources precluded interviewing non-english speakers. One respondent was deceased, and one was too impaired to do the interview. Thus, the response rate was 82 percent among those eligible to complete the interview, or 78 percent of the full universe. Between August 1999 and February 2000, 495 interviews with eligible respondents in cohort 2 were completed. Another 12 cases were not interviewed because of language barriers. Thus, the response rate was 80 percent among those eligible to complete the interview, or 79 percent of the full universe. Overall, 751 six-month interviews were completed with the 956 families in cohorts 1 and 2, for a response rate of 79 percent. In general, this high response rate implies that respondents 13