Donate
Independent, objective, nonpartisan research

R 903TMR

Authors

R 903TMR

Tagged with:

Publication PDFs

Database

This is the content currently stored in the post and postmeta tables.

View live version

object(Timber\Post)#3742 (44) { ["ImageClass"]=> string(12) "Timber\Image" ["PostClass"]=> string(11) "Timber\Post" ["TermClass"]=> string(11) "Timber\Term" ["object_type"]=> string(4) "post" ["custom"]=> array(5) { ["_wp_attached_file"]=> string(12) "R_903TMR.pdf" ["wpmf_size"]=> string(6) "672172" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(170158) "What Happens to Families When They Leave Welfare? ••• Thomas MaCurdy Grecia Marrufo Margaret O’Brien-Strain 2003 PUBLIC POLICY INSTITUTE OF CALIFORNIA Library of Congress Cataloging-in-Publication Data MaCurdy, Thomas E. What happens to families when they leave welfare? / Thomas MaCurdy, Grecia Marrufo, Margaret O'Brien-Strain. p. cm. Includes bibliographical references. ISBN: 1-58213-076-0 1. Welfare recipients—California—San Francisco Bay Area— Longitudinal studies. 2. California Work Opportunity and Responsibility to Kids (Program). I. Marrufo, Grecia. II. O'BrienStrain, Margaret. III. Title. HV99.S4M33 2003 362.5'4'097946—dc22 2003015550 Copyright © 2003 by Public Policy Institute of California All rights reserved San Francisco, CA Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source and the above copyright notice is included. PPIC does not take or support positions on any ballot measure or on any local, state, or federal legislation, nor does it endorse, support, or oppose any political parties or candidates for public office. Research publications reflect the views of the authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. Foreword A significant reduction in the nation’s welfare caseload has long been touted as the major success of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996. A PPIC report in 2002 documented that California reduced its caseload by larger numbers than any other state, although the percentage reduction was smaller than in most other states throughout the nation. A PPIC-sponsored research team—Thomas MaCurdy, Grecia Marrufo, and Margaret O’BrienStrain—has now taken a close look at a second measure of the success of welfare reform, how well families who have left welfare are doing. What Happens to Families When They Leave Welfare? concludes that, depending on your point of view, the glass is either half empty or half full. The researchers interviewed nearly 1,400 individuals in six Bay Area counties 5 to 16 months after they had departed the rolls. Over half of both one- and two-parent families had managed to earn enough income to raise them above the poverty level. However, two-parent families were faring much worse when it came to earning income sufficient to keep them out of poverty. One could argue that these families are so close to poverty wages that the slightest setback would land them back in poverty and in line at the welfare office. The glass-is-half-full folks would argue that at least these families are back in the labor force, establishing an employment record, and building a foundation for their families that is based on work, not welfare. The authors point out, however, that a remarkably large percentage of the families no longer on welfare were still eligible for such benefits as Food Stamps, Medi-Cal, the Earned Income Tax Credit, and day care but were not receiving these benefits either because they did not know about them or did not bother to apply. For families at or near the poverty line, these benefits are a societal safety net specifically designed to help those who are helping themselves. The authors strongly recommend that state policy be designed to reach out to these families iii and inform them of the benefit programs available. The long-term success of welfare reform may depend on how well these special benefit programs are funded and implemented. Otherwise, the state faces the prospect that large numbers of families no longer on welfare rolls may once again return for a full measure of the cash benefits. One concludes from reading this report that the welfare reform act of 1996 is doing better than many had predicted but that much remains to be done to see that the basic principle of “the end of welfare as we know it” can be maintained through this and future recessions. PPIC has now published nine reports to monitor this seminal change in welfare policy, and we will continue to carry out research on, and monitor the effects of, this new social policy well into the 21st century. David W. Lyon President and CEO Public Policy Institute of California iv Summary Between August 1996 and September 2001, the number of welfare recipients in California dropped by more than 1.4 million people. Most families leaving aid during this period left for the same reason families have always left aid: They found employment that increased their family income enough for them to lose eligibility for assistance. The booming economy was clearly an important factor in the caseload declines during this period. However, in these first years following welfare reform, it was feared that some families would leave aid—because of rigorous participation requirements, the specter of time limits, confusion about program rules (especially for immigrants), or merely the publicity about “the end of welfare as we know it”—without alternative sources of income. The same dramatic caseload declines seen by some as heralding the triumph of welfare reform were seen by others as an alarm bell regarding the well-being of former welfare recipients. Unfortunately, despite good information about the numbers leaving aid, we have only extremely limited information about the well-being of families after they leave the welfare rolls. To assist families in successfully leaving CalWORKs (California’s TANF1 program) and to prevent them from falling through the social safety net, policymakers must understand the circumstances of these “leavers” and how their circumstances change over time. To contribute to this understanding, this report addresses five basic questions: 1. How many families who have left CalWORKs remain off the program for extended periods of time? 2. What are the circumstances of these families, and do their circumstances improve over time? 3. What public assistance and services do leavers rely upon? ____________ 1Temporary Assistance for Needy Families. v 4. What problems and barriers do these leavers encounter while off CalWORKs? 5. Are there factors that determine which families who leave CalWORKs succeed in staying off aid (long-term leavers)? This report draws on extensive telephone surveys of welfare leavers conducted in six Bay Area counties (Alameda, Contra Costa, San Mateo, Santa Clara, Santa Cruz, and Sonoma) in the late 1990s to characterize the experiences of CalWORKs leavers over time. These surveys were conducted by The SPHERE Institute in cooperation with the six counties, with support from the U.S. Department of Health and Human Services and the William and Flora Hewlett Foundation. Key Findings Two waves of surveys give a picture of the circumstances of leavers 5 to 10 months after exit (depending on the county) and again 11 to 16 months after exit. In general, we distinguish between one-parent families and two-parent families. Our analysis presents snapshots of how well families are doing after they leave the CalWORKs program, describing their economic security, use of public assistance, and difficulties encountered in maintaining employment and coping with conditions adversely affecting their overall environment. During the study period, one in ten cases leaving aid was a two-parent family. These snapshots summarize the experiences of nonrecidivist leavers during the two postexit periods. Table S.1 presents an overview of selected circumstances for leavers about one year (11 to 16 months) after exiting CalWORKs. For longterm leavers, those families still off CalWORKs 11 to 16 months after exit, Table S.2 summarizes selected circumstances where families’ living conditions had improved or worsened between the two waves of interviews. In particular: • Monthly income for the average household was about $2,400, enough to bring 71 percent of one-parent leavers and 58 percent of two-parent leavers above poverty. Only 29 percent of one-parent leavers and 20 percent of two-parent leavers had income high enough to disqualify them for Medi-Cal. About 1 in 10 vi Table S.1 Selected Circumstances of Long-Term CalWORKs Leavers 11–16 Months Postexit Family Characteristics Average household income (all households) Household has earnings Respondent employed Average monthly earnings (households with earnings) Household above poverty Household receives Food Stamps Respondent uninsured Children uninsured Substandard housing Crowded housing Excessive rent burden Unstable childcare Child risk behaviors Household substance abuse Domestic violence One Parent Two Parents $2,411 $2,275 90% 77% $2,160 71% 9% 28% 18% 14% 17% 19% 17% 13% 16% 23% 94% 59% $2,160 58% 21% 29% 20% 28% 48% 33% 11% 7% 4% 11% SOURCES: The SPHERE Institute surveys of welfare leavers in six Bay Area counties, MEDS. Table S.2 Change in Selected Circumstances of Long-Term CalWORKs Leavers Between First and Second Postexit Survey Periods Family Characteristics Average household income Household has earnings Respondent employed Average household earnings Households above poverty Respondent health insurance Child health insurance Substandard housing Crowded housing Unstable childcare Child risk behaviors Substance abuse Domestic violence One Parent Two Parent — — Better Better — — Better Better Better — — Worse — Better Better Better Better Worse Worse Worse Worse — Better — SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. “—” indicates no significant change over time. vii households was very poor, with income below 70 percent of the poverty threshold. • Around 90 percent of families had at least one adult in the workforce. Workforce participation of respondents (almost always mothers) in one-parent families was quite high, around 77 percent. Mothers in two-parent families were somewhat less likely to work. However, for both one-parent and two-parent households, the surveys captured a significantly higher share of households with earnings than could be found through administrative data. • Income gains between the two survey waves were quite modest, only $60 to $70 per month on average. Half of the households experienced no income growth between the two periods. • Many families were eligible for but not receiving Food Stamps, often worth a substantial amount. According to administrative data, 21 percent of two-parent families and 9 percent of one-parent families received nonassistance food stamps 11 to 16 months after leaving aid. About 30 percent of one-parent families and over 40 percent of two-parent families who were eligible for Food Stamps were not receiving them. Our estimates of the benefits that eligible households would have received if they had applied for Food Stamps show that the median value of the forgone benefits was $212 per month for one-parent families and $262 per month for two-parent families. • About 50 percent of families reported Medi-Cal coverage 11 to 16 months after exit, down from about 60 percent in the 5 to 10 month period. However, administrative data show higher rates of MediCal enrollment, especially for two-parent families, so some families may not have been aware that they were still covered by Medi-Cal. About one in four survey respondents reported being uninsured, and about one in five children were reported uninsured, including about 10 percent of children who were eligible for Medi-Cal. • Almost half of two-parent families lived in crowded housing, onethird faced excessive rent burden, and one-quarter lived in substandard housing. Crowding and substandard housing were viii more common for long-term two-parent leavers than for those surveyed 5 to 10 months after exit, whereas housing conditions improved over time for one-parent leavers. • Although child care subsidies are available to working parents who leave welfare, more than one-quarter of one-parent leavers were unaware of these subsidies, as were one-third of two-parent leavers. About one-quarter of the respondents reported that obtaining child care was a significant barrier to full-time employment. In addition, the majority of respondents in working two-parent families and 40 percent of those in working one-parent families were unaware of the availability of the Earned Income Tax Credit (EITC). • Only about one out of ten families experienced incidences of child risk behaviors (e.g., expulsion from school, trouble with police, use of drugs, pregnancy), and such problems did not increase over time. Moreover, the percentage of respondents reporting substance abuse problems in the household declined among leavers the longer the time since their exit from CalWORKs. Across all outcome areas shown in Table S.2, it appears that circumstances generally improved more often than they got worse for leavers, particularly for one-parent families. We also find some areas of concern, however. Many long-term leavers still had very low incomes, and this share of very poor households saw little or no improvement during their time off CalWORKs. Considered broadly, leavers had a high level of participation in public assistance programs. Yet many families were not using benefits for which they were probably eligible, including Medi-Cal, child care, EITC, and Food Stamps. Taken together, families may be missing out on substantial support intended to help working poor families. In addition to the comparison of survey findings over time, we also assessed what factors predicted recidivism and other poor outcomes for families. On the topic of recidivism, we find some demographic characteristics at exit that are more associated with recidivism. Larger families were less likely to remain off welfare, as were families with younger children. Black families were more likely than white families to ix return to CalWORKs. Two-parent Hispanic families were more likely than other two-parent to successfully stay off CalWORKs. Speaking Spanish as a primary language was not a barrier to staying off aid. Families who returned to aid at 11 to 16 months after exit had lower income and lower earnings than other families 5 to 10 months after exit but primarily because they worked fewer hours. Recidivist families were also more likely to have still relied on public assistance 5 to 10 months after exit. Nevertheless, a large segment of long-term leavers are quite poor and face problems and barriers similar to those families who return to aid. Factors that had little effect on recidivism were sometimes predictive of other problems postexit. In a multivariate analysis of characteristics at exit associated with problems down the road for long-term one-parent leavers, such as crowded housing, living in poverty, lack of earnings, lack of health coverage, or child care problems, we find that Spanish-speaking families, bigger families, and families with no earnings at exit were more likely to encounter one or more problems by a year after exit. Families who had spent less than a year on aid were less likely to encounter at least one problem and particularly less likely to experience poverty or child care problems. Policy Implications From a policy perspective, our findings related to the take-up of public assistance benefits after exit from or denial of CalWORKs are important. Many leavers without health insurance or Food Stamps appeared to be eligible for these benefits. We estimate that many families were forgoing substantial benefits; 25 percent of one-parent families relinquished more than $300 per month in Food Stamps benefits alone. Similarly, there was often a lack of awareness of the availability of the Earned Income Tax Credit and child care subsidies. This points to the need for policies aimed at providing families with information about the availability of these post-CalWORKs benefits. Since it may be difficult to contact recipients after they leave CalWORKs, the dissemination of such information should begin while the families are still on aid. Similarly, this type of information could also be provided to CalWORKs applicants who are denied assistance but who could be eligible for other x types of assistance. Alternatively, presumptive eligibility for Food Stamps after families leave CalWORKs, allowed under the 2002 Farm Bill, might provide valuable support to working families who might not maintain these benefits otherwise. The results of our profiling analysis, where we identified recipient characteristics most likely to predict problems after exit from CalWORKs, could be useful in targeting supplemental assistance to these families, both while on CalWORKs and immediately after exit. Services designed to help recipients retain their jobs after leaving CalWORKs, for example, could be targeted more intensively toward leavers with these characteristics. In addition, families who have no earnings at the time of exit might be a flag of potential problems. Our findings on the discrepancy between administrative data and survey data that measure leavers’ earnings may also be quite important. Policymakers relying on administrative earnings from unemployment insurance files—the most common source of this information—should recognize that these data might significantly understate earnings. Finally, the question of why many leaver families in poverty choose not to go back on CalWORKs warrants further research. About onethird of our sample of long-term leavers had household incomes below the federal poverty level, and 11 percent were in the “very poor” category with monthly incomes of less than 70 percent of the poverty level. Many of the lessons learned from our analysis merely confirm conventional wisdom. Weak work experience constitutes an important determinant of whether families return to CalWORKs, and families with less child support have a greater chance of reentering the rolls as well. Still, many successful leavers look very much like families who go back to CalWORKs, and earnings are far from the determining factor for whether a family returns to cash assistance. Poor long-term leavers face problems and barriers similar to those families who return. Recidivists were only marginally more likely to indicate that they had been experiencing domestic violence, unstable child care, substance abuse, and depression; and, somewhat surprising, fewer reported that child care was a barrier to employment. After accounting for work and earnings experience, no particular condition appears as a prominent factor or trigger event for predicting who will be reentering CalWORKs. xi Contents Foreword ..................................... Summary..................................... Figures ...................................... Tables ....................................... Acknowledgments ............................... iii v xv xvii xxi 1. INTRODUCTION ........................... 2. DATA AND METHODOLOGY .................. Survey Methodology ........................... Administrative Data ........................... Categories of Leavers .......................... 1 5 5 7 8 3. THE BAY AREA IN CONTEXT .................. 4. ECONOMIC OUTCOMES ..................... Family Income, Earnings, and Employment ............ Growth in Income and Earnings for Long-Term Leavers .... Earnings Reported in Administrative and Survey Data ..... Barriers to Employment ........................ Summary .................................. 11 17 17 20 23 25 26 5. TAKE-UP OF ASSISTANCE AFTER LEAVING CalWORKs ................................ Postexit Participation in Public Assistance: Evidence from the Surveys ............................. Snapshots of Participation After Leaving CalWORKs .... Change in Participation by Long-Term Leavers over Time ................................. Snapshots and Participation Trends Using Administrative Data .................................. Take-Up of Food Stamps and EITC................. Food Stamps .............................. Earned Income Tax Credit ..................... 29 29 30 32 34 37 37 38 xiii Health Insurance Coverage ....................... 39 Summary .................................. 44 6. WELL-BEING MEASURES ..................... Housing Conditions ........................... Child Care ................................. Child and Respondent Well-Being .................. Summary .................................. 47 47 49 52 55 7. WHAT DISTINGUISHES LONG-TERM LEAVERS FROM RECIDIVIST LEAVERS? .................. Demographic Characteristics of Recidivists ............ Short-Term Leavers: Income and Work Effort .......... Short-Term Leavers: Public Assistance ............... Barriers to Employment and Knowledge of Assistance ..... What Distinguishes Poor Long-Term Leavers from Recidivist Leavers? ............................... Summary .................................. 57 57 61 63 65 65 68 8. WHICH FAMILIES ENCOUNTER MORE PROBLEMS WHEN LEAVING CalWORKs? .................. Methodology ............................... Findings .................................. Summary .................................. 69 69 71 78 9. CONCLUSION ............................. 79 Appendix: Data Sources and Methods .................. 83 Bibliography .................................. 91 About the Authors ............................... 93 Related PPIC Publications .......................... 95 xiv Figures 3.1. Share of Leavers Returning to CalWORKs by Months After Exit from CalWORKs, One-Parent Families Who Left Aid in 1998........................... 3.2. Share of Leavers Returning to CalWORKs by Months After Exit from CalWORKs, Two-Parent Families Who Left Aid in 1998........................... 3.3. Share of Leavers Receiving Nonassistance Food Stamps by Months After Exit, One-Parent Families Who Left Aid in 1998.............................. 3.4. Share of Leavers Receiving Nonassistance Food Stamps by Months After Exit, Two-Parent Families Who Left Aid in 1998.............................. 3.5. Share of Leavers Receiving Nonassistance Medi-Cal by Months After Exit, One-Parent Families Who Left Aid in 1998 ................................ 3.6. Share of Leavers Receiving Nonassistance Medi-Cal by Months After Exit, Two-Parent Families Who Left Aid in 1998 ................................ 4.1. Median Hourly Wage of Working Mothers ......... 4.2. Transportation as a Barrier to Full-Time Work ....... 6.1. Out-of-Pocket Child Care Expenses .............. 6.2. Awareness of Child Care Subsidies ............... 6.3. Child Risk Behaviors ........................ 6.4. Substance Abuse Reported by Respondents ......... 6.5. Domestic Violence Reported by Respondents ........ 7.1. Median Wage at 5–10 Months After Exit ........... 12 13 14 15 15 16 21 26 50 51 53 54 54 62 xv Tables S.1. Selected Circumstances of Long-Term CalWORKs Leavers 11–16 Months Postexit ................. S.2. Change in Selected Circumstances of Long-Term CalWORKs Leavers Between First and Second Postexit Survey Periods ............................ 2.1. Population, Sample Sizes, and Survey Response Counts .. 2.2. Demographic Characteristics of CalWORKs Leavers, by Percentage Share .......................... 4.1. Income Status of CalWORKs Leavers, from Survey Data .................................. 4.2. Family Earnings and Employment of CalWORKs Leavers, from Survey Data .................... 4.3. Mothers’ Employment Status, by Percentage Share..... 4.4. Growth in Income and Earnings for Long-Term CalWORKs Leavers, from Survey Data ............ 4.5. Family Earnings and Employment of CalWORKs Leavers, from Administrative Data ............... 4.6. Percentage Reporting Whether Child Care Creates a Problem for Full-Time Work .................. 5.1. Participation in Programs After Leaving CalWORKs, from Survey Data .......................... 5.2. Changes in Participation in Programs After Leaving CalWORKs, from Survey Data ................. 5.3. Participation in Programs After Leaving CalWORKs, from Administrative Data ..................... 5.4. Changes in Participation in Programs After Leaving CalWORKs, from Administrative Data ............ 5.5. Forgone Food Stamps Benefits of CalWORKs Leavers .. 5.6. Awareness and Use of the Earned Income Tax Credit Among CalWORKs Leavers, by Percentage Share ..... 5.7. Health Insurance Coverage of CalWORKs Leavers, by Percentage Share .......................... vii vii 6 10 18 19 20 22 23 25 30 33 34 35 38 39 40 xvii 5.8. Health Insurance Coverage of Children of CalWORKs Leavers, by Percentage Share ................... 5.9. Health Insurance Coverage of CalWORKs Leavers’ Children Eligible for Medi-Cal, by Percentage Share .... 5.10. Household Income Relative to Federal Poverty Level of Uninsured Respondents, by Percentage Share ........ 5.11. Type of Health Insurance Coverage Offered by Employers of CalWORKs Leavers, by Percentage Share .. 6.1. Housing Outcomes of CalWORKs Leavers, by Percentage Share .......................... 6.2. Primary Child Care Arrangements of CalWORKs Leavers, by Percentage Share ................... 6.3. Percentage of Time Child (Ages 5 to 13) in CalWORKs Leaver Family Was Left Unsupervised in Prior Month... 7.1. Demographic Characteristics of One-Parent CalWORKs Leavers, by Percentage Share ................... 7.2. Demographic Characteristics of Two-Parent CalWORKs Leavers, by Percentage Share ................... 7.3. Income Status of CalWORKs Leavers, One-Parent and Two-Parent Families Combined ................ 7.4. Participation in Programs After Leaving CalWORKs, One-Parent and Two-Parent Families Combined ...... 7.5. Logit Estimates on the Probability That Leavers Will Return to CalWORKs 6–12 Months Postexit ........ 7.6. Selected Circumstances of Recidivists and Poor LongTerm CalWORKs Leavers 11–16 Months Postexit..... 8.1. Relationship Between Characteristics at Exit and Outcomes at 11–16 Months Postexit, One-Parent Families ................................ 8.2. Relationship Between Characteristics at Exit and Outcomes 11–16 Months Postexit, Two-Parent Families ................................ A.1. Population and Sample Distributions, by County, Case, and Percentage Share ........................ A.2. Population, Samples, and Survey Response Counts Formulation of Weights ...................... A.3. Demographic Characteristics of All Leavers, by Grouping Variables and Percentage Share ................. 41 42 43 44 48 49 52 58 59 61 64 66 67 72 74 86 87 88 xviii A.4. Additional Demographic Characteristics of All Leavers, by Grouping Variables and Percentage Share......... 89 A.5. Counties Included in Each Region ............... 90 xix Acknowledgments We are grateful to the many individuals and organizations who contributed to this report and the earlier studies and surveys. Special thanks to Charles Lieberman, who was a major contributor to an earlier version of this report, and thanks as well to West Addison and Vanessa Lindler for their research support. Mark Gritz, David Mancuso, and Anne Moses all had important roles in individual studies included in this research. We received thoughtful comments from Gary Bjork, Maria Cancian, David Neumark, Deborah Reed, Lorien Rice, and Mary Tran. The surveys were funded through grants from the U.S. Department of Health and Human Services, Assistant Secretary for Planning and Evaluation, and the William and Flora Hewlett Foundation, with assistance from the San Mateo County Human Services Agency, the Contra Costa County Employment and Human Services Department, the Sonoma County Human Services Department, the Santa Clara County Social Services Agency, the Santa Cruz Human Resources Agency, and the Alameda County Social Services Agency. Supplemental support was provided by the Smith-Richardson Foundation, the Donner Foundation, and the National Institutes of Health (grant HD32055-02). Opinions expressed in this report are those of the authors and do not represent the views of the Public Policy Institute of California or any agency funding this research. xxi 1. Introduction Between August 1996 and September 2001, the number of welfare recipients in California dropped by more than 1.4 million people. Most families leaving aid during this period left for the same reason families have always left aid: They found employment that increased their family income enough for them to lose eligibility for assistance. The booming economy was clearly an important factor in the caseload declines during this period. However, in these first years following welfare reform, it was feared that some families would leave aid—because of rigorous participation requirements, the specter of time limits, confusion about program rules (especially among immigrants), or merely the publicity about “the end of welfare as we know it”—without alternative sources of income. The same dramatic caseload declines seen by some as heralding the triumph of welfare reform were seen by others as an alarm bell regarding the well-being of former welfare recipients. Unfortunately, despite good information about the numbers leaving aid, we have only extremely limited information about the well-being of families after they leave the welfare rolls. Administrative data from welfare systems allow us to track clients’ moves on and off public assistance programs, including take-up of noncash aid; and data from the unemployment insurance (UI) system provide quarterly earnings for employed individuals. Still, these data tell us little about the economic resources available to former welfare families or about their family circumstances, such as access to health care, housing conditions, and use of child care. To shed more light on these issues, the Assistant Secretary for Planning and Evaluation (ASPE) at the U.S. Department of Health and Human Services funded a series of studies of former welfare recipients to track outcomes for these families after they left aid in 1998 1 and 1999.1 Several of these studies, conducted by The SPHERE Institute, followed clients who left aid in the Bay Area counties of Alameda, Contra Costa, San Mateo, Santa Clara, and Santa Cruz. The William and Flora Hewlett Foundation funded an additional study of those leaving aid in Sonoma County in 1999. In each of these six Bay Area counties, former clients were interviewed twice within the first 18 months after they left aid. Together, the survey data collected on these “welfare leavers” provide a rich portrait of the family circumstances of these families and how these circumstances change over time. This report combines the survey data on leavers in these six counties, along with state- and county-level administrative data, to address five basic questions: 1. How many families who have left CalWORKs (California’s TANF2 program) remain off the program for extended periods of time? 2. What are the circumstances of these families, and do their circumstances improve over time? 3. What public assistance and services do leavers rely upon? 4. What problems and barriers do these leavers encounter while off CalWORKs? 5. Are there factors, besides income, that determine which families who leave CalWORKs succeed in staying off aid (long-term leavers)? The answers to these questions do not speak to the success or failure of welfare reform, because equivalent data are not available for prereform leavers. Nor is it possible to conduct very nuanced comparisons of these counties to other surveyed areas. Nevertheless, the answers to these questions may have powerful policy implications as California seeks to improve the CalWORKs program, especially as additional requirements are imposed by the reauthorization of the federal welfare reform legislation. ____________ 1These studies are summarized at http://www.aspe.hhs.gov/hsp/leavers99/ index.htm. 2Temporary Assistance for Needy Families. 2 The report is organized as follows. Chapter 2 describes the survey data and methodological approach. Because our surveys are all based in the Bay Area, Chapter 3 explores the degree to which experiences in the Bay Area can be expected to match those elsewhere in the state. The next three chapters then examine different aspects of the circumstances of leavers 5 to 10 months and 11 to 16 months after leaving aid. Chapter 4 considers the economic circumstances and employment outcomes for leavers, Chapter 5 looks at take-up of postexit assistance such as MediCal, Food Stamps, and the Earned Income Tax Credit (EITC), and Chapter 6 looks at other aspects of family well-being, ranging from food and housing insecurity to substance abuse and depression. These chapters focus largely on leavers who remain off aid. In Chapter 7, we turn to the issue of recidivism, looking at the degree to which barriers to employment determine which families return to aid between the two rounds of surveys. Chapter 8 then examines whether we can predict which families will encounter problems after leaving CalWORKs. Finally, Chapter 9 addresses the question of whether the findings for these six counties can be generalized to the rest of the state and suggests some policy implications. 3 2. Data and Methodology This study draws on two distinct types of data. The majority of results depend on findings from in-depth interviews conducted by telephone and in person with respondents from families who had left aid. Wherever possible, we have augmented the survey findings with administrative data available at the county or statewide level. In this chapter, we provide an overview of each type of data and then describe the basic subgroups of leavers considered throughout the analysis. Additional information on the data and methodology is provided in an appendix. Survey Methodology The analysis in this report relies on pooled data from three studies that surveyed CalWORKs leavers in six Bay Area counties: Alameda, Contra Costa, San Mateo, Santa Clara, Santa Cruz, and Sonoma.1 The surveys interviewed random samples of CalWORKs families that left aid in the following time periods: San Mateo, Santa Clara, and Santa Cruz: Contra Costa and Alameda: Sonoma: fourth quarter 1998 third quarter 1999 fourth quarter 1999 For the purposes of these surveys, a family was defined to have left aid if all members of the assistance unit left and remained off aid (in the original county or elsewhere in California) for two months.2 The sample included only cases that contained aided adults. The surveys were ____________ 1These surveys were conducted under the direction of The SPHERE Institute, with the support of the county social services agencies, ASPE, and the William and Flora Hewlett Foundation. See the bibliography for reports from the individual studies. 2It is important to note that none of these families left aid because of time limits. Since the time limit clock for CalWORKs started in January 1998, these families had substantial time left before time limits would be imposed. 5 conducted in English and Spanish; families who did not speak one of these languages were excluded from the sample. The surveys asked respondents about household composition, child well-being, child care, education and training, employment, income, food security, health insurance coverage, family well-being, and welfare experiences. Issues such as income and housing were addressed at the household level. Other questions, such as education and employment, were asked of the respondent, almost always the mother. Finally, questions about child well-being focused on a specific child in the household, chosen randomly to be the focal child. Each of the original three studies included at least two waves of surveys administered over a period of roughly 16 months after exit from CalWORKs.3 Depending on the survey, the timing of the first wave fell between 5 and 10 months after exit; the timing of the second wave fell between 11 and 16 months after exit. Interviews were attempted with all families in the sample in each wave, whether or not they had been successfully interviewed in the previous wave, and relocation efforts were more extensive in the later waves of the survey. For this reason, the number of completed interviews was actually higher in the second wave than in the first. The response rates for the total combined sample were 54 percent for the first set of interviews and 66 percent for the second wave.4 Table 2.1 shows the population of leavers, the original sample, Table 2.1 Population, Sample Sizes, and Survey Response Counts County Alameda Contra Costa San Mateo Santa Clara Santa Cruz Sonoma Total Population 1,082 964 301 1,013 238 309 3,907 Sample 238 304 170 149 184 309 1,354 First Wave 133 203 47 38 66 242 729 Second Wave 169 206 98 84 118 215 890 ____________ 3The San Mateo, Santa Clara, and Santa Cruz studies included three waves of surveys. 4Some of the analyses are limited only to those individuals who completed two rounds of interviews. 6 and the completed sample in each of the six counties. The survey results were all weighted to make the sample representative of the population of leavers from which the samples were drawn, with weights based on age, ethnicity/race, and family composition. Our weights also adjust for differential response to reflect the distribution of the population by county. Administrative Data In the absence of survey data such as those collected for these leavers studies, analysts must rely on administrative data to track outcomes for families who have left aid. The key sources of administrative data on former welfare recipients are the Medi-Cal Eligibility Determination System (MEDS), the Unemployment Insurance Base Wage File (UIBWF), and, and to a lesser extent, county data systems such as the Case Data System (CDS), which is used by our six survey counties but is not a statewide system. MEDS allows basic tracking of program participation in welfare, Medi-Cal, and Food Stamps by individuals and families. The UIBWF captures quarterly earnings for UI-covered employees. Finally, CDS captures data used to administer CalWORKs and other public assistance programs, including much of the same information in MEDS, but also data ranging from budget and grant calculation information to sanctions and exit reasons. Although the surveys used for this report provide a much more detailed portrait of family circumstances than can be deduced from administrative data, the administrative data have a number of advantages over survey data. In particular, these data are available over a long period of time for all program participants at very low cost. In contrast, the surveys are expensive to conduct and capture only a small fraction of participants. For these reasons, administrative data will continue to be a critical source for understanding the effects of welfare and other related programs. In light of this, we compare the survey findings to parallel information from administrative data in the handful of areas where the two overlap. 7 Categories of Leavers The analysis of the survey data cuts the data in two ways: by case type and by leaver status at the time of the survey waves. First, we distinguish between one-parent and two-parent CalWORKs cases. Oneparent cases are the stereotypical welfare cases, with aided single parents, usually mothers, and their children on the case. However, two-parent families are also eligible to receive CalWORKs, under only slightly different eligibility criteria but more rigorous participation requirements.5 At the end of 1999, two-parent families represented one out of every six CalWORKs cases with aided adults.6 Although twoparent families represent a minority of cases, previous research shows different caseload dynamics in two-parent families than in one-parent families. Moreover, not only does CalWORKs have different rules for one-parent and two-parent families, but California also no longer supports two-parent CalWORKs grants out of the federal block grant, because of the high federal participation requirements for two-parent families. For these reasons, policy choices around two-parent families may differ from those for the much larger group of one-parent cases. For each wave of the survey, we examine the circumstances of families off aid at the time of the surveys: 5 to 10 months after leaving aid and 11 to 16 months after leaving aid. By excluding in each period those who have returned to aid, we are focusing on the relatively successful leavers, at least as measured by welfare participation. However, because there is recidivism between these two “snapshots,” the sample of families is somewhat different at the two time periods. In the population of leavers from which our sample was drawn (and to which our sample is weighted), about 12 percent of families still off aid 7 months after exit had returned ____________ 5Two-parent families must have one parent incapacitated or an unemployed principal wage earner, where unemployment is defined as working fewer than 100 hours in the month before application. Two-parent families must participate in work activities (employment, job search, training, or education) for 35 hours per week, compared to 32 hours for one-parent families. 6Over one-fourth of all CalWORKs cases have no aided adults, because the parents are not eligible to receive aid or children are residing with caretaker relatives. These “child-only” cases are not subject to welfare-to-work rules, time limits, or other aspects of welfare reform. For this reason, we consider only cases where an aided adult recipient (and his or her children) has left welfare. 8 to aid by 14 months after exit. This group of “short-term leavers” is therefore included in results on leavers off at 5 to 10 months but excluded from results on leavers still off at 11 to 16 months (the “long-term leavers”), so that each snapshot presents the circumstances of the remaining leavers at the time. (In Chapter 7, we look explicitly at the issue of recidivism and the associated client characteristics.) Table 2.2 provides demographic characteristics for the leaver populations divided both by one- and two-parent families and by the two time periods. Since short-term leavers represent only a small share of all observed cases, the demographic characteristics are relatively unchanged from period to period. On the other hand, the two-parent families do differ demographically from the one-parent families. In the six survey counties, approximately one-third of one-parent families leaving aid were headed by blacks, about one-quarter were headed by Hispanics, and most others were white. In contrast, less than 10 percent of two-parent families leaving aid were headed by blacks, and nearly 40 percent were headed by Hispanics. The two-parent families were also twice as likely to be other ethnicities, predominantly Asian. Spanish speakers were also less common among one-parent families than among two-parent families. Two-parent families had more children on average and were also more likely to have very young children. Finally, two-parent families had fewer previous months on aid. More than half of one-parent families had been on aid for at least three years out of the last five. 9 Table 2.2 Demographic Characteristics of CalWORKs Leavers, by Percentage Share Demographic Characteristics Ethnicity Black Hispanic White Other Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Monthsa Monthsb Monthsa Monthsb 32 31 9 9 26 26 39 39 35 35 37 35 8 8 16 17 Primary language English 94 93 85 86 Spanish 6 7 15 14 Age of case head 16–25 26–35 36–45 46+ 33 31 26 25 36 36 38 39 25 26 29 29 77 8 8 Number of children 0–1 56 55 32 31 2 28 28 33 35 3+ 17 17 35 34 Age of youngest 0–2 3–5 6–11 12+ 32 30 40 40 25 25 24 24 24 25 21 21 19 19 15 15 Months on cash aid in preceding 5 years 1–12 20 19 26 25 13–36 25 25 28 27 37–60 55 56 47 48 Number of cases 2,709 2,396 329 288 NOTE: Columns may not sum to 100 because of rounding. aIncludes short-term and long-term leavers. bIncludes only long-term leavers. 10 3. The Bay Area in Context In assessing outcomes for families who left aid in the surveyed counties, a question that arises is to what degree are their experiences representative of the state. Unfortunately, the very absence of information that motivated the leavers’ studies makes it difficult to conduct in-depth comparisons between different groups of leavers. This chapter places the survey findings in a state context by comparing the Bay Area, where our surveys were conducted, to other regions of the state: Los Angeles County, other Southern California counties, the “Farm Belt” counties on the Central Coast and along the Central Valley, and the North and Mountain counties.1 However, in doing this, we are restricted to a minimal set of outcome measures available from statewide administrative data, notably recidivism rates and use of non-CalWORKs benefits, including take-up of nonassistance Medi-Cal (NAMC) and nonassistance Food Stamps (NAFS).2 The six counties included in our surveys are largely representative of the Bay Area, accounting for 78 percent of the population and 83 percent of CalWORKs cases at the time most of the samples were drawn in 1999. The economic conditions in these counties were also representative of the Bay Area, with an average unemployment rate of 3.1 percent in 1999. The Bay Area, however, was clearly not representative of the state as a whole, experiencing a much stronger economic boom ____________ 1Appendix Table A.5 lists the counties included in each of these regions. MaCurdy, Mancuso, and O’Brien-Strain (2000) provide an analysis of caseload trends across these regions. 2The term “nonassistance” refers to the receipt of these benefits outside the categorical eligibility provided through the CalWORKs program. Former CalWORKs recipients typically qualify for transitional Medi-Cal for up to two years after exit, and their children may continue to be eligible beyond this point. Former recipients, as well as other working poor families, qualify for nonassistance Food Stamps if their income is below 130 percent of the poverty threshold and they meet additional income and asset tests. 11 Percentage returning to CalWORKs than other regions of the state. However, on many outcomes for welfare leavers, the Bay Area was similar to other urban areas of the state, judging by analysis done by Gritz et al. (2001a). Using administrative data, we tracked welfare recidivism among families who left CalWORKs in 1998. As Figures 3.1 and 3.2 show for one-parent and two-parent cases respectively, welfare recidivism in the Bay Area was almost identical to recidivism in Southern California counties, except Los Angeles, which experienced lower caseload declines but also lower recidivism rates. For both one- and two-parent cases, the Bay Area’s recidivism rate fell in the middle of the five regions. CalWORKs recidivism was lowest in Los Angeles and highest in the two most rural regions. The greatest variation occurs among two-parent families, where the recidivism rate one year after exit ranges from about 8 percent in Los Angeles County to just over 20 percent in the North and Mountain region. The regional variation in recidivism in 1998 was greater than seen historically, although the rural North and Mountain and Farm Belt regions have long had higher recidivism rates than the 35 North and Mountain Farm Belt 30 Bay Area Southern California 25 Los Angeles 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.1—Share of Leavers Returning to CalWORKs by Months After Exit from CalWORKs, One-Parent Families Who Left Aid in 1998 12 Percentage returning to CalWORKs 35 North and Mountain Farm Belt 30 Bay Area Southern California 25 Los Angeles 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.2—Share of Leavers Returning to CalWORKs by Months After Exit from CalWORKs, Two-Parent Families Who Left Aid in 1998 urban parts of the state, in large part because of seasonal employment patterns.3 Among the urban regions, one possible explanation for the lower recidivism rate in Los Angeles may also explain the Bay Area leavers’ population. The Bay Area and Southern California (excluding Los Angeles) regions experienced substantially higher percentage caseload declines from the mid-1990s peak through 1998. If one accepts the premise that families with the fewest barriers to self-sufficiency would be the first families to leave aid, then the prolonged caseload decline before 1998 may have led to a Bay Area caseload in 1998 that had a larger share of families with significant barriers to self-sufficiency. Among 1998 leavers, then, the Bay Area leavers (as well as those in Southern California) may have included a segment of the caseload with more barriers than those leaving the caseload in Los Angeles, which could account for the differences in recidivism rates among the three regions. Take-up of noncash assistance after exit in the Bay Area generally fell between that of Los Angeles and other Southern California counties. For ____________ 3See Gritz et al. (2001a). 13 nonassistance Food Stamps, take-up was relatively low in the Bay Area, as seen in Figures 3.3 and 3.4. For both case types, take-up of NAFS was far higher in the rural parts of the state. Leavers in the rural areas may have had lower incomes than those in the Bay Area, because of the poorer economy in the rural regions, and therefore were more likely to be eligible for Food Stamps. Given the seasonal nature of work and welfare participation, leavers in these areas may also tend to be more aware of the availability of nonassistance Food Stamps than those in the Bay Area. For both case types, the Bay Area had the second highest percentage of leavers receiving nonassistance Medi-Cal, with the highest enrollment in Los Angeles County, as shown in Figures 3.5 and 3.6. We can hypothesize that the high take-up of nonassistance Medi-Cal may help explain the low recidivism rates in these regions. Enrollment in nonassistance Medi-Cal was highest in Los Angeles (where recidivism was the lowest) and lowest in the two rural regions (where recidivism was highest). Access to Medi-Cal may have played a role in reducing the incentives for leavers to return to CalWORKs. Percentage receiving nonassistance Food Stamps 40 35 North and Mountain Farm Belt 30 Southern California Bay Area 25 Los Angeles 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.3—Share of Leavers Receiving Nonassistance Food Stamps by Months After Exit, One-Parent Families Who Left Aid in 1998 14 Percentage receiving nonassistance Food Stamps 40 35 North and Mountain Farm Belt 30 Southern California Bay Area 25 Los Angeles 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.4—Share of Leavers Receiving Nonassistance Food Stamps by Months After Exit, Two-Parent Families Who Left Aid in 1998 100 90 80 70 60 50 40 Los Angeles 30 Bay Area 20 Southern California Farm Belt 10 North and Mountain 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.5—Share of Leavers Receiving Nonassistance Medi-Cal by Months After Exit, One-Parent Families Who Left Aid in 1998 Percentage receiving nonassistance Medi-Cal 15 Percentage receiving nonassistance Medi-Cal 100 90 80 70 60 50 40 Los Angeles 30 Bay Area Southern California 20 Farm Belt 10 North and Mountain 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.6—Share of Leavers Receiving Nonassistance Medi-Cal by Months After Exit, Two-Parent Families Who Left Aid in 1998 Comparison across regions suggests that findings for the Bay Area will be largely consistent with outcomes for other urban areas of the state, especially Southern California counties other than Los Angeles. Where there are differences, it is difficult to predict the direction of the effect. Given that the Bay Area and Southern California had experienced large caseload declines before the 1998–1999 period, it is possible that leavers in our survey period (those who remained after these declines) faced more barriers to self-sufficiency than leavers in Los Angeles, which had lower declines before this period. This hypothesis would fit with the higher recidivism rate in the Bay Area and Southern California, than in Los Angeles. In contrast, the findings for our six counties are unlikely to be representative of outcomes for the less urban areas of the state. 16 4. Economic Outcomes Economic stability is fundamental to the long-term prospects of former welfare recipients; thus, we begin our discussion of the experiences of Bay Area welfare leavers with an analysis of their economic outcomes. We consider three key outcomes: total family income from all sources, earnings, and employment. For each outcome, we look at snapshot portraits of leavers both in the 5 to 10 month and 11 to 16 month periods (including both short- and long-term leavers in the former period), and we compare the change in outcomes for long-term leavers between the two periods. First, we present the findings from our surveys; next, we compare selected findings to available administrative data; and finally, we compare the outcomes for those families who return to welfare to those who stay off welfare but remain poor. Family Income, Earnings, and Employment Monthly income—the financial resources available to families—is one of the most basic indicators of how well leavers are doing. Table 4.1 reports two different measures of family income at our two snapshot points for families remaining off welfare 5 to 10 months after exiting CalWORKs and 11 to 16 months after exit. The distribution of family income is shown in both absolute dollars and relative to the poverty level—a measure that adjusts for differences in family size. Drawing on the survey data, this table—and most subsequent tables in the chapter— reports results for one-parent families and two-parent families separately. By 5 to 10 months after leaving aid, the average family income for leavers hovered around $2,000 per month for both one- and two-parent families. For one-parent families, this income was sufficient to pull the majority above the poverty level, and one in four had income above the cutoff for Medi-Cal eligibility (185 percent of the federal poverty level). For the somewhat larger two-parent families, however, similar incomes in absolute terms translate into lower income relative to the poverty level. 17 Table 4.1 Income Status of CalWORKs Leavers, from Survey Data Income and Poverty Measures Distribution of monthly income, $ Mean Percentile 10 25 50 75 90 Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 1,942 600 1,092 1,700 2,461 3,693 2,411 800 1,200 1798 2,600 3,565 2,082 1,000 1,200 1,635 2,700 3,200 2,275 1,000 1,300 1,970 2,800 4,200 Income as share of poverty threshold, % 185 20 10 7 14 16 19 48 30 17 17 12 14 22 24 13 22 25 29 20 20 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. About 55 percent of two-parent families had income below the poverty threshold 5 to 10 months after leaving aid, compared to 36 percent of one-parent families. In terms of income, the greatest concern may be around the “very poor” families—those with incomes below 70 percent of the poverty level. One in five one-parent families was very poor 5 to 10 months after leaving aid; a smaller share of two-parent families fell into this category. If we look at leavers around a year after they left aid, we see that average income among one-parent families was about 25 percent higher than in the earlier period, at around $2,400. Among families off aid for 11 to 16 months, two-parent families had lower income than one-parent families. The majority of leavers off aid in the 11 to 16 month period had income above poverty for both one-parent and two-parent families, 18 and the share that was very poor fell among one-parent families. On the other hand, in the 11 to 16 month period, 14 percent of two-parent families remaining off aid were very poor. Earnings are the most important source of income for leavers. Around 90 percent of families have at least one adult in the workforce (that is, at least one member reports earnings greater than zero), as shown in Table 4.2. The mean earnings are around $1,800 to $1,900 in the first set of surveys, rising to around $2,200 in the second set. These calculations, as well as the distributions of monthly earnings, include only those families with earnings. Especially for two-parent families, it is clear that earnings make up nearly all of the total family income, particularly for those in the upper half of the income distribution. (In the next chapter, we review participation in public assistance, which makes up much of the other income for these families.) The survey asked the employment status of the respondent, virtually always the mother. For one-parent families, these mothers are the sole earners; more than 60 percent worked full-time and another 15 percent worked part-time (Table 4.3). Not surprisingly, mothers in two-parent families are less likely to be employed, but work was still common among Table 4.2 Family Earnings and Employment of CalWORKs Leavers, from Survey Data Earnings and Employment Measures Percentage working Distribution of monthly earnings, $ Mean Percentile 10 25 50 75 90 Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 88 90 94 93 1,799 600 1,000 1,500 2,200 3,500 2,160 750 1,100 1,600 2,500 3,400 1,947 775 1,200 1,700 2,700 3,200 2,160 50 1,150 1,800 2,750 4,200 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. 19 Table 4.3 Mothers’ Employment Status, by Percentage Share Employment Status One-parent families Not currently working Currently working part-time Currently working full-time Time After Leaving CalWORKs 5–10 Months 11–16 Months 22 23 15 12 63 65 Two-parent families Not currently working Currently working part-time Currently working full-time 49 5 46 41 7 52 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. these mothers, with around half employed full-time. The larger share of two-parent families with earnings is clearly driven by the employment of the fathers in these households. Still, by the 11 to 16 month period, we observe a larger share of mothers working in two-parent families. Are mothers in two-parent families more selective in accepting jobs? Figure 4.1 shows this is not the case; the median hourly wage of employed mothers is higher in one-parent families than in two-parent families. In fact, these findings suggest that mothers who rely on their own earnings to support the family are in higher-paying jobs. Growth in Income and Earnings for Long-Term Leavers Although the snapshot tables presented above provide an important description of the economic characteristics of “active” leavers at two different points after their cohort’s initial leave date, they do not tell us how the circumstances of a constant group of leavers change over time. As mentioned above, differences between the two snapshots are due to both the changing circumstances of long-term leavers and the removal of short-term leavers who returned to CalWORKs. To assess actual growth in earnings and income for leavers between these two periods, we turn to an analysis of long-term leavers only. Table 4.4 presents the net gains or 20 10.50 10.00 One parent Two parents Hourly wage, $ 9.50 9.00 8.50 8.00 7.50 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 4.1—Median Hourly Wage of Working Mothers losses in monthly income and earnings between the two surveys for the same respondents in the 5 to 10 month and 11 to 16 month postexit periods. In fact, income gains between the 5 to 10 month interviews and the 11 to 16 month interviews were very modest for long-term one-parent leavers. Average monthly income grew by only $61, and half of the leavers experienced no income growth at all. Twenty-five percent experienced monthly income losses greater than $114, and 10 percent had losses exceeding $275. An equal share saw their income grow, but the relative magnitude of these income changes was smaller. A slightly more optimistic pattern applies for two-parent leavers. Although half of two-parent leavers still experienced income losses, these losses were less than for one-parent families. Moreover, for those two-parent leavers enjoying gains in income, the gains were typically larger than for oneparent families. Reported changes in earnings were consistent in both direction and magnitude with changes in income for both one- and two-parent leavers. For both groups, a small percentage of families lost their jobs (moving 21 Table 4.4 Growth in Income and Earnings for Long-Term CalWORKs Leavers, from Survey Data Income and Earnings Measures Distribution of gains in monthly income, $ Average growth Percentile 10 25 50 75 90 Gains/Losses 5–16 Months After Leaving CalWORKs One Parent Two Parents 61 –275 –114 0 89 200 71 –75 –33 0 167 398 Distribution of gains in monthly earnings Percentage finding a job Percentage losing a job Average growth, $ Percentile 10 25 50 75 90 4 4 41 –275 –100 0 78 200 1 2 90 –33 –33 0 167 333 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. from positive to zero earnings) whereas other families gained jobs (moving from zero to positive earnings). These percentages were equal (or nearly equal), yielding little change in overall employment between the two periods. Thus, we find that the financial status of leavers remained quite stable after exit from CalWORKs: 75 percent of leavers experienced monthly income variations below $115 in absolute terms, whereas earnings changes were lower than $85 for 75 percent of leavers’ families (combining one-parent and two-parent leavers). The differences that show up in the two snapshots, therefore, are driven primarily by a 22 composition change, where the lowest income families are the most likely to return to aid. We take up the issue of recidivism in Chapter 7. Earnings Reported in Administrative and Survey Data Earnings is one aspect of well-being that we can assess looking at either survey or administrative data. For CalWORKs leavers employed in work covered by unemployment insurance, quarterly earnings are available in the UI Base Wage File. To construct family earnings using only administrative data, we match Social Security numbers for household members included in the assistance unit in the last month of aid to individuals’ earnings reported in the UI Base Wage File. For households with more than one adult on aid, we sum across individuals to get household earnings. Table 4.5 presents the percentage of leaver families with positive earnings and the distribution of their earnings constructed from the UI Base Wage File, in a structure comparable to that of Table 4.2. There are a number of reasons to expect the administrative data on earnings to differ from the survey reports. First, not all workers or all Table 4.5 Family Earnings and Employment of CalWORKs Leavers, from Administrative Data Earnings and Employment Measures Percentage working Distribution of monthly earnings, $ Mean Percentile 10 25 50 75 90 Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 71 75 80 81 1,423 341 809 1,450 1,922 2,556 1,574 367 802 1,491 2,187 2,938 2,392 3,049 886 1,213 2,033 3,087 3,812 1,334 1,986 2,692 3,439 5,840 SOURCE: UI Base Wage File, California Employment Development Division. 23 earnings may appear in the UI file. Self-employed workers are not required to report earnings, and earnings from casual or informal labor may be underreported or unreported. Second, our construction of earnings from the UI files relies on the Social Security numbers of household members at the time the household left aid. For one-parent families in particular, there may be household members (such as unmarried partners) who entered the household after it left aid or who were not included in the assistance unit. The earnings for these household members would not be captured in our administrative data match but are likely to be reported in the survey. Finally, there could be misreporting by the respondent. Relative to the survey, the administrative data vastly understate the share of households with earnings. For both one- and two-parent families, the percentage with earnings in the quarter including the survey month according to the UI Base Wage File is around 15 points lower than the percentage of households reporting positive earnings in our survey. The differences in the mean and distribution of earnings suggest that different factors are skewing the administrative data relative to the survey data for one- and two-parent families. For one-parent leavers, there is a consistent undercount of earnings in the UI Base Wage File compared to earnings in the respondents’ reports, although the medians are fairly similar. We suspect that the higher income reported in the survey is due to additional household members with earnings not included in the assistance unit, although these families may also have earnings that are not reported to the Employment Development Department. For twoparent families, we do not expect much change in the wages of adults included in the household. For these families, the fact that the UI Base Wage File shows higher earnings at every point in the distribution, but fewer families with earnings, suggests that the earnings not captured in the administrative data are disproportionately lower earnings. This would be consistent with a significant share of low-wage work in the informal labor market, which may fit with the larger share of nonEnglish-speaking immigrants in the two-parent caseload. These findings suggest that reliance on data from the UI Base Wage File will give a lower bound on the share of households with earnings. 24 The earnings captured in the administrative data are probably also a lower bound for one-parent families but may show an overly optimistic picture of typical earnings for two-parent families. Barriers to Employment Despite the relatively strong workforce participation among our surveyed CalWORKs leavers, many families experienced barriers to fulltime employment. The surveys asked whether and to what extent respondents viewed child care and transportation as barriers to full-time employment. As Table 4.6 shows, 28 percent of the one-parent respondents and 20 percent of the two-parent respondents in the 5 to 10 months postexit period reported that obtaining child care was a big problem. Among one-parent leavers who were still off welfare 11 to 16 months after exiting CalWORKs, only 20 percent reported child care barriers to employment. In contrast, 39 percent of the long-term twoparent leavers reported child care barriers. The higher reporting of barriers may reflect the larger families of two-parent leavers, or less access Table 4.6 Percentage Reporting Whether Child Care Creates a Problem for Full-Time Work One-parent families Not a problem A small problem A big problem Time After Leaving CalWORKs 5–10 months 11–16 months 56 59 17 21 28 20 Two-parent families Not a problem A small problem A big problem 70 46 10 15 20 39 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. 25 to child care assistance, which is generally available only when both parents are working at the same time. However, it appears that the additional potential earners in the household allow two-parent leavers to remain off aid without solving their child care problems. Transportation was also a barrier to full-time employment for about 16 percent of one-parent families and 12 percent of two-parent families in the 5 to 10 months postexit period, as Figure 4.2 shows. Furthermore, families who remained off aid for another two quarters (until 11 to 16 months) were not less likely to have this problem. The share of two-parent families reporting a transportation barrier was 16 percent among leavers who spent more time off aid and remained virtually constant for one-parent leavers. Percentage reporting transportation as problem 20 One parent Two parents 18 16 14 12 10 8 6 4 2 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 4.2—Transportation as a Barrier to Full-Time Work Summary Overall, the economic status of CalWORKs leavers is mixed. Although most families are not poor, two-parent families are generally less well-off because their incomes do not compensate for additional family members. Leavers who have stayed off CalWORKs for a year or 26 more have higher incomes and earnings than the group that was surveyed at 5 to 10 months, but actual growth in income/earnings for these longterm leavers, if it exists, is moderate. Another encouraging finding is that the work effort of these families is relatively high, especially for single mothers who must support their families alone, despite the child care and transportation barriers that persist for a minority of families. This workforce attachment probably contributes to the fact that these families remain off CalWORKs rather than returning to aid; but, still, onequarter of poor long-term leaver families appear to rely on income other than earnings—an issue we turn to in the next chapter. Our analysis also reveals that policymakers and program administrators must exercise caution when using administrative data to assess the earnings and work history of CalWORKs leavers, because they systematically miss certain types of employment, undercount earnings for one-parent families, and overcount earnings for two-parent families. 27 5. Take-Up of Assistance After Leaving CalWORKs Although they left CalWORKs, many of the families surveyed continued to rely on some form of public assistance or related meanstested programs. In this chapter, we examine patterns of use of public programs by one- and two-parent leavers. As with income and earnings, we look at both point-in-time snapshots of public assistance use at 5 to 10 months and at 11 to 16 months after exit and at changes in participation for the long-term leavers observed in both time periods. We also compare the survey findings to those generated from administrative data. After presenting these participation rates, we use our survey results to explore in more depth familiarity with and take-up of selected programs, including potential reasons for forgoing benefits. Postexit Participation in Public Assistance: Evidence from the Surveys Families who have left CalWORKs may be eligible for a variety of public assistance and related programs, including Food Stamps, MediCal, Supplemental Security Income/State Supplementary Payment (SSI/SSP), EITC, and Women, Infants, and Children (WIC) program vouchers. Eligibility rules differ from program to program, taking into account diverse factors such as income, assets, disability, age, and work activity. For example, families who have recently left CalWORKs are typically eligible for transitional coverage from the Medi-Cal program. Beyond this transitional period, children in households with income below 185 percent of the poverty level may remain eligible for Medi-Cal, depending on their age. Food Stamps has a complex set of tests for eligibility but requires that families have gross income below 130 percent of the poverty level. Low-income individuals who are aged, blind, or 29 disabled may qualify for SSI/SSP, and WIC provides vouchers for milk and other food products for pregnant and nursing women and young children. Finally, working families with children and income below approximately $22,000 (in 1999–2000) are eligible for EITC. Snapshots of Participation After Leaving CalWORKs According to our survey, more than 80 percent of leavers relied on at least one form of public assistance at both the 5 to 10 month and 11 to 16 month period. Table 5.1 summarizes reported participation in public assistance programs including nonassistance Food Stamps, nonassistance Table 5.1 Participation in Programs After Leaving CalWORKs, from Survey Data Income Support Programs Percentage receiving Any public assistancea Nonassistance Food Stamps Nonassistance Medi-Cal SSI EITC WIC vouchers Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 83 82 85 89 11 7 3 12 60 48 58 54 11 11 9 7 34 41 23 26 12 14 11 28 Distribution of Food Stamp recipient benefits, $ Mean Percentile 25 50 75 155 193 245 248 45 134 236 148 170 200 257 210 200 300 257 330 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. aIncludes EITC, Food Stamps, Medi-Cal, WIC, Refugee Assistance, SSI, Foster Child Payments, Workers’ Compensation, and Public Housing Assistance. 30 Medi-Cal, SSI/SSP, EITC, and WIC vouchers.1 Two-parent families were somewhat more likely to rely on some form of public assistance, especially those who remained off CalWORKs for 11 to 16 months. About 1 in 10 one-parent families reported receiving NAFS 5 to 10 months after exiting CalWORKs, and a slightly smaller percentage of leavers in the later period received them. In comparison, only 3 percent of two-parent leaver families reported Food Stamp receipt in the early period, but among long-term leavers in the second period, this figure was four times as high. Although these figures seem low, given the fact that we expect a higher percentage of leavers to have incomes low enough to qualify for NAFS, program participation is often underreported in surveys, and we believe that these results may be affected by that phenomenon. To test this hypothesis and determine the extent of underreporting, we look at administrative data on participation in the next section of this chapter. In subsequent sections, we explore further the issue of eligibility versus take-up for this and other programs. For those who reported receiving Food Stamps in the survey, the monthly value of these benefits is fairly substantial, as shown in the bottom part of Table 5.1. For one-parent families, the average monthly Food Stamp allotment equals $155 in the 5 to 10 month postexit period; and at 11 to 16 months after exit, this average is up to $193 for those who are still off welfare. Not surprisingly given the larger families, the benefits for two-parent families are higher at each point in the distribution. However, long-term leavers in one-parent families had higher benefits in the 11 to 16 month period than the cohort of leavers in the 5 to 10 month period—a result that does not hold for two-parent leavers. Families rely on nonassistance Medi-Cal more frequently than on nonassistance Food Stamps: 60 percent of one-parent families and 58 percent of two-parent families reported health insurance through this program 5 to 10 months after exiting CalWORKs, and the figures for the 11 to 16 month leavers were 48 percent and 54 percent, respectively. ____________ 1The term “nonassistance” for Food Stamps and Medi-Cal is used to distinguish families participating in these programs without participating in CalWORKs, as opposed to families categorically eligible for these programs because they receive CalWORKs. 31 This relatively high rate of Medi-Cal coverage is consistent with the use of transitional Medi-Cal. After Medi-Cal, EITC was the most commonly used program, although participation is low relative to the number of families reporting earnings. Unlike the other programs, EITC is administered by the Internal Revenue Service and provides a maximum refundable tax credit of around $4,000 for working families. In this program as well, participation rates are significantly lower for two-parent families, especially in the 11 to 16 month period when 41 percent of one-parent families received EITC, compared to 26 percent of two-parent families. One- and two-parent leavers participated in the WIC program at approximately the same rate 5 to 10 months after leaving welfare. For those who remained off welfare for 11 to 16 months, however, twoparent families were twice as likely (at 28 percent) as one-parent families to use WIC. Change in Participation by Long-Term Leavers over Time Focusing on long-term leavers only, we develop a richer picture of changes in program participation after leaving CalWORKs. Since families both enter and exit public assistance over time, we disaggregate the net change by calculating program entry and exit between the two surveys for the same respondents. For each assistance program, Table 5.2 presents the percentage of leavers enrolling and the percentage of leavers exiting between the first and second interview. There is, in fact, significant movement in and out of programs between the 5 to 10 month and 11 to 16 month period. Slightly more than one-third of one-parent families enrolled in or took up an additional benefit and a similar percentage left at least one program. Two-parent families, on the other hand, were more likely to move into additional programs (27 percent) than to leave (17 percent). The remainder of the table bolsters the argument that one-parent leavers became slightly less dependent on other forms of public assistance as they spent more time off aid, whereas two-parent leavers become more dependent. EITC was the program with by far the highest net enrollment rate (percentage enrolled minus percentage leaving), whereas Medi-Cal had the lowest net enrollment rate. The table shows that a 32 Table 5.2 Changes in Participation in Programs After Leaving CalWORKs, from Survey Data Percentage enrolling in an additional programa Leaving at least one programa Changes in Participation 5–16 Months After Leaving CalWORKs One Parent Two Parents 36 27 37 17 Percentage enrolling in nonassistance Food Stamps Leaving nonassistance Food Stamps 3 5 1 2 Percentage enrolling in Medi-Cal Leaving Medi-Cal 71 21 5 Percentage enrolling in SSI Leaving SSI 52 29 Percentage started using EITC Stopped using EITC 17 18 11 3 Percentage enrolling in WIC Leaving WIC 68 54 Distribution of gains in nonassistance Food Stamps, $ Average gain Percentile 25 50 75 –2 –5 00 00 00 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. aIncludes EITC, Food Stamps, Medi-Cal, WIC, Refugee Assistance, SSI, Foster Child Payments, Workers’ Compensation, and Public Housing. substantial percentage of one-parent leavers lost Medi-Cal coverage as they spent time off aid—14 percent—whereas the net loss for two-parent families was less, at least according to the surveys. 33 The survey results also show that changes in nonassistance Food Stamps benefits were negligible. Net participation rates decreased only slightly, and the actual benefit allotment did not change at all for at least 75 percent of leavers. This finding suggests that changes in the distribution of monthly benefits amounts observed in the 5 to 10 months and 11 to 16 months postexit periods (Table 5.1) reflect the different composition of the NAFS participants—short-term leavers dropping back into CalWORKs—rather than changes in circumstances for longterm leavers. For both NAFS and Medi-Cal, however, we will reexamine the participation trends using administrative data, below. Snapshots and Participation Trends Using Administrative Data Analyzing patterns in nonassistance Food Stamps and Medi-Cal participation using administrative data both strengthens the findings of the surveys and reveals additional trends that were masked by underreporting in the surveys. As discussed in Chapter 2, MEDS contains participation information for both Medi-Cal and NAFS. Tables 5.3 and 5.4 display the snapshots and trends in participation for these programs, and their figures can be directly compared to those in Tables 5.1 and 5.2. Table 5.3 Participation in Programs After Leaving CalWORKs, from Administrative Data Income Support Programs Percentage receiving nonassistance Food Stamps Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 11 9 7 21 Percentage receiving nonassistance Medi-Cal 71 53 74 73 SOURCE: MEDS. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 34 Table 5.4 Changes in Participation in Programs After Leaving CalWORKs, from Administrative Data Changes in Participation 5–16 Months After Leaving CALWORKs Income Support Progams One Parent Two Parents Percentage enrolling in nonassistance Food Stamps 4 10 Leaving nonassistance Food Stamps 3 2 Percentage enrolling in Medi-Cal Leaving Medi-Cal 24 22 13 SOURCE: MEDS. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. For one-parent families, the NAFS participation rates calculated from the surveys and MEDS are identical in the 5 to 10 months postexit period, and the MEDS rate is only marginally higher in the 11 to 16 months postexit period (Table 5.3). Two-parent families, on the other hand, appear more likely to underreport their participation, especially in the second period. Since we generally accept the MEDS data as correct, the true participation rate for two-parent leavers is 7 percent in the first period and 21 percent in the second, a difference of 4 percentage points and 9 percentage points, respectively, from the survey results. Thus, although participation in NAFS is still low, the administrative data reveal that long-term two-parent leavers are more likely than other groups to receive Food Stamps benefits. Moreover, using MEDS to focus on changes in participation for long-term leavers, we find that two-parent leavers were five times more likely to enroll in NAFS between the two periods than to drop benefits (Table 5.4). This finding is in stark contrast to the virtually nonexistent change in participation reported in the surveys, and it is consistent with other indicators of the increasing dependence of two-parent leavers. In comparison, the net increase in NAFS participation for one-parent long-term leavers was only 1 percent. Increases in NAFS take-up should not be viewed as a negative outcome. In fact, state and county social services agencies did additional outreach 35 to attempt to boost take-up of Food Stamps and other postexit benefits, because these benefits are viewed as an integral part of the safety net for working poor families and the drop in Food Stamps participation was an unintended side-effect of welfare reform. Turning to Medi-Cal, MEDS data reveal that nearly three-quarters of both one- and two-parent leavers had at least one family member enrolled in the program in the 5 to 10 month period after leaving CalWORKs (Table 5.3). This figure is substantially larger than that calculated from the surveys. As in the survey data, however, one-parent leavers in the 11 to 16 month period were less likely to participate in Medi-Cal than were leavers in the earlier period, whereas participation for two-parent leavers was the same in both periods. As for assessing the time trends for long-term leavers, the net loss of Medi-Cal coverage was even greater than the survey suggests. For one-parent leavers, the two data sources agree on the percentage leaving Medi-Cal, but according to MEDS only 2 percent, versus 7 percent in the survey, enrolled in MediCal (Table 5.4). For two-parent leavers, the figures for both enrollees and disenrollees were higher than reported in the surveys, but the net loss was still greater at 9 percent in the MEDS data. Given that we found relatively close agreement between the two data sources in the reporting of CalWORKs and Food Stamps enrollment, the differences in Medi-Cal suggest that many respondents may not have been aware of their enrollment in Medi-Cal. We can hypothesize that at least some of the discrepancy stems from the “Edwards Hold” problem: Medi-Cal coverage is automatically extended to CalWORKs leavers (under what is referred to as the “Edwards” category) pending redetermination of their eligibility to stay on Medi-Cal. During the time of our surveys, coverage under Edwards was extended to many recipients for relatively lengthy periods of time, as counties worked off their backlogs of former CalWORKs cases that needed redetermination under newly adopted eligibility rules.2 It is possible that many of these families were not aware that their Medi-Cal coverage had been extended, thereby contributing to the discrepancy between the administrative and survey ____________ 2For more on the evidence of the role of the Edwards Hold on Medi-Cal participation, see Lieberman and Mancuso (2001). 36 data. Eventual awareness of their coverage may also explain why a higher percentage of long-term leavers, at least one-parent leavers, reported gaining coverage between the two survey periods than were recorded as actual enrollees in the administrative data. Take-Up of Food Stamps and EITC Although our survey results reveal that most CalWORKs leavers participate in at least one income-support program, participation in specific programs is still relatively low. It is especially low for Food Stamps (with only 10 to 20 percent of leavers participating) and EITC— in which approximately 25 to 40 percent participate but we expect nearly all leavers to be income-eligible. On the one hand, there is little reason for concern if families do not participate because their income exceeds eligibility thresholds. On the other hand, policymakers should be concerned if poor families do not receive all the assistance for which they are eligible. Moreover, although some stigma may influence participation in Food Stamps, it should have little effect on EITC participation. To gain a better understanding of the relationship between leavers’ well-being and their participation in these programs, we use survey results to assess take-up of benefits and the reasons for respondents’ nonparticipation. Food Stamps To examine the use of nonassistance Food Stamps, we determined which households appeared to be eligible for Food Stamps, judging by their income, household size, and other applicable factors. Thirty percent of one-parent leaver households were eligible but not receiving Food Stamps, as shown in Table 5.5. More than half of two-parent families appeared to be forgoing Food Stamps. Many of these leaver families had income low enough to remain eligible for CalWORKs. Without the CalWORKs grant, the forgone Food Stamps benefits are often quite high. Our estimates of the benefits that eligible households would have received if they had applied for Food Stamps show that the median value of the forgone benefits in the 5 to 10 month postexit 37 Table 5.5 Forgone Food Stamps Benefits of CalWORKs Leavers Income Support Programs Percentage eligible for but not receiving Food Stamps Potential distribution of monthly benefits , $ Mean Percentile 25 50 75 Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 30 28 54 43 256 214 145 117 230 212 329 305 239 286 175 149 211 262 257 366 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. period was $230 per month for one-parent families and $211 per month for two-parent families. These typical forgone benefits are actually higher than the average benefit reported by those who do participate, probably in large part because benefits received are underreported. By 11 to 16 months after exit, the share of two-parent families eligible but not receiving Food Stamps fell, although the average forgone benefits rose. The share of one-parent families forgoing Food Stamps was relatively unchanged between the two time periods, although the value of forgone benefits fell over time, as earnings rose. Earned Income Tax Credit A large fraction of families never heard of or used EITC, according to survey findings at the two postexit periods and summarized in Table 5.6. Among those families who had heard of EITC, only one out of two had used it (16 percent compared to 32 percent, not shown). Twoparent families were less often aware of EITC than one-parent families. For one-parent families, EITC take-up rates were higher for families off aid at 11–16 months: The share of families unaware of EITC decreased from 49 percent to 42 percent, whereas the share of families who had 38 Table 5.6 Awareness and Use of the Earned Income Tax Credit Among CalWORKs Leavers, by Percentage Share Awareness and Use One-parent families All families Never heard of, never used Heard of, never used Heard of and used Families with positive earnings Never heard of, never used Heard of, never used Heard of and used Time After Leaving CalWORKs 5–10 11–16 Months Months 49 42 18 17 33 41 46 41 17 16 36 43 Two-parent families All families Never heard of, never used Heard of, never used Heard of and used Families with positive earnings Never heard of, never used Heard of, never used Heard of and used 66 66 11 9 23 25 65 66 12 9 23 25 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. used EITC increased from 33 percent to 41 percent. However, little improvement in take-up occurred among two-parent families. Because virtually all two-parent families were working, restricting the analysis to those with at least one working member has no real effect on the results for two-parent families. Awareness of EITC was slightly higher among working one-parent families than nonworking one-parent families. Health Insurance Coverage Table 5.1 reported whether any family member participated in Medi-Cal. However, Medi-Cal eligibility applies to individuals, and 39 different members in the same family may or may not be eligible for Medi-Cal, especially after the transitional benefit period, where children may be eligible but not the adults, or younger children but not older children. Similarly, parents may receive employee health insurance coverage but not dependent coverage. For these reasons, we asked additional questions about health insurance coverage for the respondent and the children in each household. These results are reported in Tables 5.7 and 5.8. Table 5.7 summarizes respondents’ participation in various health insurance programs. Although nonassistance Medi-Cal is the most important source of insurance in the 5 to 10 month postexit period, for long-term leavers in the 11 to 16 month postexit period it is not the most common. For one-parent leavers, the adult uninsured rate is slightly higher for long-term leavers in the second period than for those who were off aid in the first period, but for two-parent leavers the opposite is true. For both family types, over one-third of long-term leavers had Table 5.7 Health Insurance Coverage of CalWORKs Leavers, by Percentage Share Type of Coverage One-parent families Not insured Medi-Cal Medicare Private Time After Leaving CalWORKs 5–10 11–16 Months Months 26 28 43 32 00 31 40 Two-parent families Not insured Medi-Cal Medicare Private 35 29 40 35 00 25 36 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 40 private insurance by the second period. Nevertheless, a significant share of the respondents remained uninsured long after leaving aid. Children are more likely to be insured than their mothers, as shown in Table 5.8. Children in two-parent families are only slightly more likely than children in one-parent families to be uninsured. Private insurance covers about one in four children—a rate lower than that for respondents. For long-term leavers in the second period, children in two-parent families had lower rates of private insurance and higher rates of Medi-Cal coverage than children in one-parent families or the more comprehensive group of leavers surveyed in the first period. Healthy Families (California’s health insurance program for children in lowincome families) was just beginning in this period, and coverage through Healthy Families was rarely reported for any of these children. One reason for the lack of insurance among children was that many who were eligible for Medi-Cal were not enrolled. According to the survey, only 62 to 64 percent of children eligible for Medi-Cal were Table 5.8 Health Insurance Coverage of Children of CalWORKs Leavers, by Percentage Share Type of Coverage One-parent families Not insured Medi-Cal Healthy Families Private/other government Time After Leaving CalWORKs 5–10 11–16 Months Months 16 18 53 51 21 28 30 Two-parent families Not insured Medi-Cal Healthy Families Private/other government 17 20 57 63 00 27 16 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Columns may not sum to 100 because of rounding. 41 enrolled 5 to 10 months after exiting CalWORKs (Table 5.9). Some of the children who were not enrolled in Medi-Cal were insured by other sources, including private health insurance, Healthy Families, and other government programs, but about 10 percent of children who were eligible for Medi-Cal remained uninsured. If parents sought acute or emergency medical care for these children, they would typically be enrolled in Medi-Cal or Healthy Families, given their income eligibility; the larger concern is if parents forgo important preventive care because they do not realize that the children qualify for these programs. In the first period, there was little difference in coverage patterns between oneand two-parent families. In the second period, however, the share of uninsured children who were eligible among one-parent families was 11 percent, whereas the share of uninsured children was only 6 percent for two-parent families remaining off welfare. In addition, children in twoparent families were more likely to be covered by nonassistance Medi-Cal Table 5.9 Health Insurance Coverage of CalWORKs Leavers’ Children Eligible for Medi-Cal, by Percentage Share Type of Coverage One-parent families Not insured Medi-Cal Healthy Families Private/other government Time After Leaving CalWORKs 5–10 11–16 Months Months 9 11 64 57 30 25 32 Two-parent families Not insured Medi-Cal Healthy Families Private/other government 10 6 62 79 00 28 15 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. 42 and less likely to have private insurance than were children in one-parent families. We might expect the uninsured respondents to have income too high to qualify for nonassistance Medi-Cal. Table 5.10 shows the percentage of uninsured respondents grouped by their household income relative to the federal poverty level. A substantial percentage of these uninsured leavers, especially two-parent families, had incomes below the poverty level, indicating that most were probably eligible for Medi-Cal coverage. The percentage was lower in the 11 to 16 month postexit period but still represented about one-third of the uninsured one-parent respondents and a larger share of two-parent respondents. Table 5.11 illustrates the role of employment in providing health coverage. About 45 percent of single mothers did not receive any coverage from their employers. This figure is higher among two-parent families—more than 50 percent of working mothers in two-parent Table 5.10 Household Income Relative to Federal Poverty Level of Uninsured Respondents, by Percentage Share Income as a Percentage of Poverty Level One-parent families 185 Time After Leaving CalWORKs 5–10 11–16 Months Months 30 13 12 16 11 21 19 18 28 32 Two-parent families 185 11 12 62 30 10 13 8 19 9 26 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 43 Table 5.11 Type of Health Insurance Coverage Offered by Employers of CalWORKs Leavers, by Percentage Share Type of Health Coverage One-parent families None Self only Family Time After Leaving CalWORKs 5–10 11–16 Months Months 45 42 55 50 52 Two-parent families None Self only Family 51 63 18 6 31 31 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. families did not receive any health coverage from their employers. Further, the share of working mothers in two-parent families not covered by their employers is higher in the 11 to 16 month period. Not surprisingly, mothers in two-parent families and their children are more likely to be covered by their partners. Summary Use of postexit assistance is common among the CalWORKs leavers we surveyed, although the rate of take-up for eligible families is well below 100 percent for the programs we tested. Participation is highest in nonassistance Medi-Cal, although some respondents are unaware that they have coverage. However, participation in this and most other programs declines over time, at least for one-parent families. Two-parent long-term leaver families are more likely to use assistance in general, and their participation increases in the later period—a finding that is 44 consistent with the fact that many struggle to raise their incomes above the poverty threshold (Chapter 4). Administrative data (when available) are the most accurate sources of information on program participation. However, these data cannot shed light on CalWORKs leavers who are eligible for assistance but do not use it. This disparity is particularly evident with respect to nonassistance Food Stamps: Although program participants collect rather substantial benefits (according to the surveys), over half of those who are eligible do not participate. The EITC also appears to be underutilized, according to the surveys, although the respondent may be unaware that her family is receiving the credit if she is not responsible for preparing the household’s taxes. Despite extensive Medi-Cal participation, a fraction of respondents and to a lesser extent their children are still uninsured, with many lacking employer-based coverage. Taken together, these findings indicate that although nonwelfare public assistance programs are important to CalWORKs leavers, many get by without them, either by choice or because of a lack of information. 45 6. Well-Being Measures The surveys allow us to capture a richer picture of the well-being of CalWORKs leavers, beyond just the economic measures. In this chapter, we highlight housing conditions, child care usage, and selected risk behaviors. Using our survey data, we present snapshot portraits of leavers after 5 to 10 and 11 to 16 months. We conclude this chapter with a discussion of the difference between welfare recidivists and poor longterm leavers. Housing Conditions Housing costs can be a real challenge for former CalWORKs families, as we see in Table 6.1. Common problems include substandard housing, excessive rent burden, and crowding.1 More than one-third of two-parent families face excessive rent burden and crowding. These problems are less common among one-parent families, but one-parent families are more likely to live in substandard housing. Housing assistance, either as public housing or housing vouchers, can help offset housing problems but is relatively rare, especially for the two-parent families. Twenty-eight percent of one-parent families but only 10 percent of two-parent families received assistance with housing. These results are consistent with prior information about these two types of families and relevant public policies. Whereas two-parent families have difficulty finding space for their relatively larger families but also allocate more of their monthly income to rent, one-parent families have more success securing housing subsidies (historically targeted to poor single____________ 1Substandard housing means that one or more of the following conditions are found: (1) a leaky roof or ceiling, (2) a toilet, water heater, or other plumbing that does not work, or (3) rats, mice, roaches, or other pests. House crowding exists if the ratio of household members to rooms (excluding bathrooms) is greater than one. Rent is excessive when it exceeds 50 percent of total household income. 47 Table 6.1 Housing Outcomes of CalWORKs Leavers, by Percentage Share Housing Outcomes One-parent families Substandard housing Excessive rent burden Crowding Housing assistance Time After Leaving CalWORKs 5–10 11–16 Months Months 27 14 20 19 23 17 28 29 Two-parent families Substandard housing Excessive rent burden Crowding Housing assistance 15 28 36 33 36 48 10 23 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. parent households) but this subsidized or other low-rent housing may be of lower quality. Comparing across the two time periods, we find that two-parent leavers who remain off CalWORKs for 11 to 16 months faced more housing problems than did two-parent families in the earlier survey period (which includes both long-term leavers and families who soon return to welfare). For example, 28 percent of two-parent families who had been off aid for 11 to 16 months lived in substandard housing, compared to 15 percent in the 5 to 10 month postexit period. Nearly half of two-parent families lived in crowded housing in the later period. However, those who remained off in the 11 to 16 month period were more likely to receive housing assistance: 23 percent compared to 10 percent in the earlier period. The high incidence of housing problems among long-term two-parent leavers is troubling, especially since compositional change between the two survey periods means that some of the 5 to 10 month leavers returned to welfare. Reported housing 48 problems for one-parent families were similar in the two survey periods, except that substandard housing was less prevalent. Child Care As discussed in Chapter 4, a substantial proportion of respondents indicated that child care was a barrier to obtaining full-time employment. Before policymakers can address this problem, however, it is important to know more about the types of child care welfare leavers use and how they are paying for it. Table 6.2 shows the distribution of various types of primary child care arrangements used by the respondents. Relatives were the most common child care providers in two-parent families, with about 72 percent of two-parent families relying on relatives for child care in the 5 to 10 month postexit period. Relatives were also the most frequent child care resource used by one-parent Table 6.2 Primary Child Care Arrangements of CalWORKs Leavers, by Percentage Share Child Care Arrangement One-parent families Head Start Formal day care Extended day care Adult relative Family day care Adult nonrelative Time After Leaving CalWORKs 5–10 11–16 Months Months 42 24 21 10 11 40 45 17 15 56 Two-parent families Head Start Formal day care Extended day care Adult relative Family day care Adult nonrelative 00 8 17 8 15 72 50 11 11 17 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 49 families, but other forms of nonrelative care were important as well. For example, in the 5 to 10 month time frame, 24 percent of one-parent families relied on formal day care (such as a child care center, preschool, or nursery school) and an additional 17 percent used family day care homes. In the later postexit period, two-parent families who remained off CalWORKs were less likely to use relative care and more likely to use formal and other nonrelative care than two-parent families in the earlier period. The distribution of providers for one-parent families was similar in both periods. One-third of one-parent families incurred out-of-pocket expenses for child care during the 5 to 10 month period (Figure 6.1). Two-parent families were less likely to incur these expenses (20 percent), consistent with the fact that two-parent families were more likely to use an adult relative for care, most of whom do not require payment. In the second period, when fewer two-parent leavers used relative care, they were more likely than in the earlier period to pay for child care (30 percent). However, the share with out-of-pocket expenses also increased for oneparent leavers, even though more one-parent families relied on adult relatives for care in the later period. Percentage paying out-of-pocket child care expenses 50 45 One parent Two parents 40 35 30 25 20 15 10 5 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.1—Out-of-Pocket Child Care Expenses 50 Working CalWORKs leavers qualify for subsidized child care, with payments made to the provider of their choice. However, as Figure 6.2 shows, a little over one-quarter of one-parent respondents in the 5 to 10 month postexit period indicated that they were not aware of government subsidies that would help pay for child care. Lack of awareness was more prevalent among two-parent families—34 percent did not know about subsidies. The proportion of two-parent families that were unaware of child care subsidies was even higher in the second postexit period, at nearly 45 percent, and we can assume that this result reflects the changing composition of leavers between the two snapshot surveys. This figure suggests that long-term leavers are more likely to be unaware of child care subsidies than short-term leavers, at least in two-parent families. In addition, some school-age children were not receiving care who could potentially need it. Table 6.3 shows the extent to which the respondents indicated that their children (ages 5 to 13) were left Percentage unaware of child care subsidies 50 45 One parent Two parents 40 35 30 25 20 15 10 5 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.2—Awareness of Child Care Subsidies 51 Table 6.3 Percentage of Time Child (Ages 5 to 13) in CalWORKs Leaver Family Was Left Unsupervised in Prior Month Hours Unsupervised per Month One-parent families 0 1–19 20+ Time After Leaving CalWORKs 5–10 11–16 Months Months 87 82 69 79 Two-parent families 0 1–19 20+ 95 94 45 11 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. unsupervised during the preceding month. Few two-parent families— but nearly 10 percent of one-parent families—indicated that this occurred 20 or more hours during that month. Child and Respondent Well-Being Our surveys included a variety of questions related to different aspects of the well-being of families in general and children in particular. Figure 6.3 shows the proportion of respondents reporting that the focal child in the family had engaged in risk behaviors in the preceding six months.2 Risk behaviors include being suspended or expelled from school, getting into trouble with police, having a problem with alcohol or drugs, doing something illegal to get money, dropping out of school, and getting pregnant or getting someone else pregnant. Over 12 percent of the one-parent respondents indicated that the child had engaged in such behaviors. In comparison, children in two-parent families were half as ____________ 2Focal children were randomly selected from all of the respondent’s children. 52 Percentage reporting child risk behaviors 20 18 One parent Two parents 16 14 12 10 8 6 4 2 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.3—Child Risk Behaviors likely to engage in such behaviors. The percentage of families reporting problem behaviors was essentially the same in both survey periods. Figure 6.4 shows the percentage of respondents who indicated that they or any other adults in their household had substance abuse problems. A household has a substance abuse problem if, according to the survey, people complained about the respondent’s use of alcohol or drugs, or they were having problems because of their alcohol or drug use, or any other adult in the household had a problem with alcohol or drugs. Similar to the pattern for child risk behaviors and supervision, one-parent families were more likely to report these problems. Contrary to child risk, however, longer-term leavers in the 11 to 16 month postexit period reported notably less substance abuse than CalWORKs leavers in the 5 to 10 month period. Figure 6.5 shows the percentage of respondents indicating the presence of domestic violence (physical or emotional abuse) in their household. In the first period, about 15 percent of one-parent respondents reported such occurrences, compared to 11 percent for two-parent families. In the second period, abuse affected 23 percent of one-parent families, but the 53 Percentage reporting substance abuse 16 14 One parent Two parents 12 10 8 6 4 2 0 5–10 11–15 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.4—Substance Abuse Reported by Respondents 24 22 One parent Two parents 20 18 16 14 12 10 8 6 4 2 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.5—Domestic Violence Reported by Respondents Percentage reporting domestic violence 54 abuse rate for two-parent families was basically unchanged. These findings may reflect the volatility of relationships and living arrangements for single parents. Summary Selected measures of the general well-being of CalWORKs leaver families indicate that challenges persist well after exiting cash aid. Specific problems with housing are different for one- and two-parent families, but they are consistent with other differences between the two populations; in particular, they provide more evidence that two-parent families are straining to get by even as they refrain from returning to welfare. The fraction of school-aged children of single parents who are left home alone is less than ideal, especially since one-quarter of single parents are unaware of child care subsidies for which they are most likely eligible. Risky behaviors by both children and adults affected a minority of families. These behaviors were more common among single-parent families, but we do not know how the incidence of these behaviors compares to that in the general population. 55 7. What Distinguishes LongTerm Leavers from Recidivist Leavers? Comparing the experiences of leavers who successfully remain off CalWORKs to those who return offers insight into the principal factors and trigger events that contribute to recidivism. To examine recidivism, we use administrative data for the entire leavers population in our six counties, considering the differences in demographic characteristics for short-term and long-term leavers and comparing these two groups to those who had returned to aid before the first wave of interviews. We then turn to survey data to look in more depth at whether factors identified in the 5 to 10 month interviews help predict which families will have returned to aid by the 11 to 16 month period. These factors help identify potential avenues for targeting policy resources to assist families in staying free of welfare. Demographic Characteristics of Recidivists To assess whether certain demographic characteristics are associated with recidivism, we divide the population of leavers into three categories—long-term, short-term, and temporary leavers. Long-term leavers are those who had not returned to CalWORKs by the time of our second wave of interviews—the group captured in our 11 to 16 month findings in previous chapters. In the administrative data, we define longterm leavers as those still on aid 14 months after exit. Our 5 to 10 month findings showed a snapshot that combined long-term leavers with those who left aid between the two survey waves. In the administrative data, these short-term leavers are defined as those who remained on aid at 7 months after exit but had returned to aid by 14 months. Finally, those families who left aid for 2 months but who had already returned by 57 month 7 (a group not included in our previous analysis) are defined as temporary leavers. For these three groups of leavers, Tables 7.1 and 7.2 summarize the demographic characteristics of one- and two-parent families, respectively. Using administrative data, these two tables show Table 7.1 Demographic Characteristics of One-Parent CalWORKs Leavers, by Percentage Share Demographic Characteristics Ethnicity Black Hispanic White Other All Leavers 34 26 33 8 Temporary Short-Term Long-Term Leavers Leavers Leavers 41 37 26 29 26 28 75 31 26 35 8 Primary language English Spanish 94 97 63 95 5 93 7 Age of respondent 16–25 26–35 36–45 46+ 34 38 36 35 24 23 64 43 36 17 4 31 36 26 7 Number of children 0–1 2 3+ 54 51 28 30 17 19 60 27 13 55 28 17 Age of youngest 0–2 3–5 6–11 12+ 33 39 26 26 24 24 17 11 41 27 19 13 30 25 25 19 Months on cash aid in preceding 5 years 1–12 13–36 37–60 18 25 57 14 24 62 19 28 53 19 25 56 Number of cases 3,473 764 313 NOTE: Columns may not sum to 100 because of rounding. 2,396 58 Table 7.2 Demographic Characteristics of Two-Parent CalWORKs Leavers, by Percentage Share Demographic Characteristics Ethnicity Black Hispanic White Other All Leavers 11 37 39 14 Temporary Short-Term Long-Term Leavers Leavers Leavers 18 10 9 30 34 39 44 49 35 8 7 17 Primary language English Spanish 87 92 80 86 13 8 20 14 Age of respondent 16–25 26–35 36–45 46+ 29 37 34 25 35 28 32 39 29 29 27 29 76 7 8 Number of children 0–1 28 17 41 31 2 32 28 20 35 3+ 39 54 39 34 Age of youngest 0–2 3–5 6–11 12+ 43 54 41 40 22 17 22 24 21 20 22 21 14 9 15 15 Months on cash aid in preceding 5 years 1–12 24 17 32 25 13–36 26 19 29 27 37–60 51 64 39 48 Number of cases 432 103 41 NOTE: Columns may not sum to 100 because of rounding. 288 the entire population of leavers (not just those included in the surveys), listing their family characteristics at the time they left aid. Sixty-nine percent (2,396 out of 3,473) of one-parent families and 66 percent (288 of 432) of two-parent families remained off aid for at 59 least 14 months. As we saw above, one-parent families who leave CalWORKs have fewer children than their two-parent counterparts and are more likely to be black, less likely to be Hispanic, and less likely to have Spanish as their primary language. In looking at the relationship of these and other demographic characteristics to the probability of remaining off aid, we see that for both one- and two-parent families the number and age of children and ethnicity are linked to the probability of remaining off aid. Larger families are less likely to remain off CalWORKs. This factor is especially pronounced for the two-parent cases, where families with three or more children make up 54 percent of two-parent temporary leavers and only 34 percent of two-parent longterm leavers. Families with older children have a higher probability of remaining off CalWORKs. For example, one-parent families whose youngest child is older than age 12 constitute only 11 percent of temporary leavers but about 19 percent of long-term leavers. In contrast, families with an infant (ages 0 to 2) present constitute 41 percent of temporary leavers (54 percent of two-parent and 39 percent of oneparent families) but only 31 percent of long-term leavers (40 percent of two-parent and 30 percent of one-parent). Black families are more likely than white families to return to CalWORKs, in both one-parent and two-parent families. Whereas blacks make up 39 percent of temporary leavers, combining one- and two-parent families, they make up only 28 percent of long-term leavers. However, two-parent families of Hispanic descent are more likely than their white counterparts to succeed in staying off CalWORKs. Hispanic two-parent families constitute only 30 percent of temporary leavers and make up 39 percent of long-term leavers. Overall, there is little relationship between success in staying off CalWORKs and having Spanish as the family’s primary language. For two-parent families, however, Spanish-speaking families represent almost double the share of long-term leavers than temporary leavers—stronger evidence that speaking Spanish did not constitute a barrier to remaining off CalWORKs. 60 Short-Term Leavers: Income and Work Effort Table 7.3 depicts the income and earnings status of short-term and long-term leavers during the 5 to 10 month period after last participating in CalWORKs. Although structured much like the tables in Chapter 4, here we separate out the short- and long-term leavers within the 5 to 10 Table 7.3 Income Status of CalWORKs Leavers, One-Parent and Two-Parent Families Combined Income and Earnings Measures Poverty status, % 185 Data Source Survey 5–10 Months After Leaving CalWORKs Short-Term Long-Term Leavers Leavers 42 16 10 22 16 17 23 20 9 25 Distribution of monthly income, $ Mean Percentile 10 25 50 75 90 Survey 1,383 509 665 1,143 2,000 2,993 2,012 700 1,200 1,717 2,546 3,900 Distribution of monthly earnings, $ Percentage working Mean Percentile 10 25 50 75 90 Survey 81 965 200 200 800 1,500 2,200 89 1,888 775 1,200 1,600 2,400 3,500 NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 61 month snapshot. At this first interview period, short-term leavers are distinctly poorer than their long-term counterparts. Over half of the short-term families have monthly incomes below the poverty level, with 42 percent classified as very poor, at less than 70 percent of poverty. This is not to say that poverty was uncommon among the long-term leavers. Although only 16 percent fell into the very poor income category, almost 40 percent still had incomes below the poverty level. On the other hand, 25 percent of long-term leavers had incomes exceeding 185 percent of the poverty level, compared to 9 percent of the short-term leavers. The distributions of income reveal that short-term leavers had uniformly lower monthly incomes than families who would remain off CalWORKs in the following months, with the average income for the short-term leavers reaching only about two-thirds that of the long-term leavers. Not surprisingly, most of this income differential reflects a shortfall in earnings for those families who would return to welfare in the next several months. Although their likelihood of working was similar to that of families who succeed in staying off welfare well beyond this period, they worked less and earned less. One might suspect that a major portion of this shortfall in earnings came from lower hourly wages, but Figure 7.1, which illustrates the wages of short-term and long-term 12 Short-term leavers Long-term leavers 10 Median hourly wage, $ 8 6 4 2 0 All One parent Two parents Figure 7.1—Median Wage at 5–10 Months After Exit 62 leavers, reveals trivial differences in wage rates. So, the principal reason for lower earnings can be attributed to fewer hours worked for families who would become recidivist. These findings support the view that leavers’ inability to maintain satisfactory economic conditions constituted an important factor motivating them to return to welfare. Short-term leavers clearly did worse as a group from an economic perspective than those families who succeeded in staying off welfare. On the other hand, many who did return were better off than some long-term leavers who did not, a topic discussed further below. Short-Term Leavers: Public Assistance To what extent were short-term leavers more dependent on public assistance programs before their reentry into CalWORKs than long-term leavers who did not return? Table 7.4 addresses this question, summarizing participation in income support programs available for families leaving CalWORKs, such as nonassistance Food Stamps and Medi-Cal, as well as alternative programs we considered in Chapter 4. For purposes of comparison, Table 7.4 shows Food Stamps and Medi-Cal receipt as recorded by both the surveys and MEDS. As with Table 7.3, Table 7.4 breaks down the findings for the 5 to 10 month period to distinguish between the short- and long-term leavers Most families relied on some form of public assistance after leaving CalWORKs, with nonassistance Medi-Cal being the most common form of involvement. However, Table 7.4 shows that short-term leavers had higher participation in all programs (except for child support). According to the survey findings, they were about four times more likely to collect Food Stamps and WIC, 6 percentage points more likely than long-term leavers to obtain EITC, and 34 percentage points more likely to have relied on Medi-Cal for health coverage. It is interesting to note that the administrative and survey data on Medi-Cal use provide almost identical results for short-term leavers, but many fewer of those leavers who would become long-term leavers reported Medi-Cal coverage in the earlier period, even though the administrative data show them as eligible. Of course, short-term leavers’ heavier reliance on public assistance in part merely reflected their weaker economic position. Also, consistent with 63 Table 7.4 Participation in Programs After Leaving CalWORKs, One-Parent and Two-Parent Families Combined Programs Percentage receiving Any public assistancea Nonassistance Food Stamps Data Source Survey MEDS Survey 5–10 Months After Leaving CalWORKs Short-Term Long-Term Leavers Leavers 93 83 33 8 34 7 Distribution of monthly support, $ Mean Percentile 25 75 195 145 200 45 200 200 Percentage receiving Nonassistance Medi-Cal SSI EITC WIC vouchers Child support MEDS Survey Surveyb Survey Survey Survey 80 82 22 37 37 16 70 58 10 31 10 19 Distribution of monthly support, $ Mean Percentile 25 75 224 356 100 125 214 475 NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. aIncludes EITC, Food Stamps, Medi-Cal, WIC, Refugee Assistance, SSI, Foster Child Payments, Workers’ Compensation, and Public Housing Assistance. bStatistics in this row are computed without survey weights. We do this to avoid overemphasis on several observations from a county wherein all observations happen to be on SSI. The difference between the participation estimates in this row is not statistically significant at conventional levels of confidence. popular beliefs, long-term leavers had greater amounts of child support income that offered more cushion for remaining off welfare. 64 Barriers to Employment and Knowledge of Assistance Given that recidivism is linked to weaker performance in employment activities, the natural question is whether there were potential barriers or other factors making work a less attractive option for short-term leavers. Chapter 4 showed that many respondents reported child care and transportation as barriers to full-time work. It also showed that a significant percentage of leavers were unaware of employment assistance programs such as EITC and child care subsidies. To quantify the effects of barriers to employment and lack of knowledge of employment-related programs on recidivism, we exploit multivariate probabilistic models. This approach tests the hypothesis that various barriers reported in the early period increased the probability that a family would return to CalWORKs in the next six months. The estimation results are summarized in Table 7.5. The first and second columns in the table present the coefficients and standard errors of the probabilistic model; the last column shows the effect of each factor on the probability that a family will return to CalWORKS shortly. Employment by either the respondent or any other member in the household has the highest effect on recidivism. The families of respondents who were employed at the time of the interview were 14 percentage points less likely to return to CalWORKS; an additional member employed reduces the probability by another 20 percent. Families that perceived child care as a barrier to full-time employment also had an increased likelihood of returning to aid between the two interview periods. However, paying out-of-pocket expenses for child care, reporting transportation barriers, or being unaware of the EITC was not predictive of recidivism. What Distinguishes Poor Long-Term Leavers from Recidivist Leavers? Although economic circumstances go a long way toward explaining which families return to CalWORKs, our analysis of income and earnings patterns reveals that a significant number of nonrecidivist leavers were relatively poor. Even in the 11 to 16 month postexit period, when 65 Table 7.5 Logit Estimates on the Probability That Leavers Will Return to CalWORKs 6–12 Months Postexit Variable Respondent is employed An additional member in the household is employed Child care is a barrier to full-time employment Coefficient –2.11** Standard Error 0.4008 Marginal Effect –0.14** –1.41** 0.3801 –0.20** 0.70* 0.3896 0.09* Family pays out-of-pocket child care expenses 0.52 0.4025 0.07 Job travel is a barrier to full-time employment 0.24 0.3603 0.03 Household is unaware of EITC 0.24 0.3474 0.03 Constant –1.70** 0.4314 Number of observations 557 Likelihood ratio 36.5 NOTES: Sample includes all one-parent and two-parent respondents in the first interview (5 to 10 months) who had not returned to CalWORKs at the time of the interview. Recidivists are those families that returned to CalWORKs sometime between their first interview and one year after exit. *Statistically significant at 90 percent. **Statistically significant at 95 percent. incomes had increased, 32 percent of all nonrecidivist families had household incomes below the federal poverty level, and 11 percent were below 70 percent of the poverty level. Most of these families were probably eligible to return to CalWORKs but had chosen not to do so. In an effort to better understand why some low-income leavers remained off aid while others returned, in Table 7.6 we compare several dimensions of income and well-being for poor long-term leavers (below 100 percent of the federal poverty guideline) and for families who were recidivist at the time of the interview in the 11 to 16 month period. This 66 Table 7.6 Selected Circumstances of Recidivists and Poor Long-Term CalWORKs Leavers 11–16 Months Postexit Household Characteristics Households with earnings Median household earnings, $ Respondent’s work activity Full-time Part-time Not working Barriers to full-time employment Child care Transportation Domestic violence Children uninsured Unstable child care Household substance abuse Depressed 3+ days weekly Recidivist Families 61 1,300 Long-Term Leavers’ Household Income Below 100% Poverty 76 900 22 45 14 12 64 43 27 40 22 22 30 26 4 19 28 23 97 41 34 NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. period corresponds to a time when long-term leavers have successfully remained off CalWORKs for one year or more and recidivist leavers have reentered the CalWORKs program for added income support. Recidivists include members of both the short-term and temporary leavers groups who may have reentered CalWORKs more than once and for extended periods prior to the 11 to 16 month interview. Table 7.6 reveals that a higher percentage of the poor long-term leavers had household earnings, 76 percent, compared to 61 percent of recidivist families. Moreover, a higher percentage of poor long-term leaver respondents were working and were doing so at full-time jobs (45 percent compared to 22 percent of recidivists). In light of this observation, it is perhaps surprising to see that the recidivist leavers had higher median household earnings. Although the low income of the long-term leavers shown in Table 7.6 is largely due to the subgroup we have selected, the income of the recidivist households suggests that 67 household members other than the respondent (that is, other than the case head who has returned to aid) are contributing to the household earnings. Thus, the respondent’s work activities may be one clue as to why some leavers chose to return to aid while others remained off aid. We also see some differences in the measures of family well-being, but again there is no clear pattern. The recidivists were only marginally more likely to indicate that they had been experiencing domestic violence, unstable child care, substance abuse, and depression. On the other hand, we also see that the recidivists were much less likely to report that child care was a barrier to employment and were no more likely to indicate that transportation was a problem. Summary Many of the lessons conveyed by the above findings merely confirm conventional wisdom: Weak employment experiences constitute an important determinant of whether families return to CalWORKs, and families who succeed off aid have greater amounts of child support income than families who return to aid. Recidivists face bleaker outcomes than long-term leavers long before returning to CalWORKS: At the 5 to 10 month period, recidivists are more likely to rely on other forms of public assistance and are also more likely to face poverty. Still, many families who remain off aid for periods longer than a year look very much like families who go back to CalWORKs, and earnings are far from the determining factor for whether a family returns to cash assistance. A large segment of the long-term leavers are quite poor and face problems and barriers similar to those of families who return. 68 8. Which Families Encounter More Problems When Leaving CalWORKs? In Chapters 3 to 5, we found that families frequently encounter problems well after leaving CalWORKs. For example, in the period 11 to 16 months after exit, 29 percent of one-parent and 44 percent of twoparent families had income below poverty, 23 percent of mothers in oneparent families were unemployed, 18 to 20 percent of families had uninsured children, and 48 percent of two-parent families lived in crowded housing. In this chapter, we explore the extent to which program administrators can use known characteristics to predict the problems a family is likely to encounter after leaving aid. In other words, how are a family’s characteristics at exit associated with postexit measures of well-being? Identifying those characteristics will help policymakers design targeted programs to assist families leaving aid. Methodology For this analysis, we examine all families who have not returned to CalWORKs by the 11 to 16 month postexit period, trying to determine which nonrecidivists nevertheless face poor outcomes. We use the following five outcome indicators: • Crowded housing, • Income below 100 percent of the federal poverty guideline, • Not employed (no earnings in household), • No health insurance coverage for respondent or child, and • Child care is a barrier to full-time employment. In addition to examining each of these outcomes separately, we examine the probability of having at least one of these problems and of having at least two of these problems. 69 We exploit multivariate probabilistic models to quantify factors contributing to each of these outcomes. These models control for the effects of other variables when measuring the effect of a key variable on the probability that an outcome will occur. For example, we would like to know if a Hispanic family is exposed to unemployment more frequently than a black family. Using multivariate probabilistic models, we can estimate the effect of ethnicity on the probability that a family will be unemployed, after controlling for other variables known to affect employment such as number of children. Results are summarized in Tables 8.1 (for one-parent families) and 8.2 (for two-parent families). The columns of the tables present the results for each outcome, based on estimates from probabilistic models run for each outcome. The rows of the tables list the family characteristics considered in the analysis. These characteristics are grouped into seven categories: 1. Primary language spoken by the respondent, 2. Ethnicity of the head of the assistance unit, 3. Time on aid = number of months on aid in the previous five years, 4. Number of children, 5. Age of the youngest child when a family left CalWORKs, 6. Earnings at exit = earnings for the assistance unit in the quarter preceding exit according to the UI Wage Base File, and 7. Exit reason, according to CDS files. For each characteristic category, families are divided into mutually exclusive groups. These seven characteristics, by design, contain information that would be available to program administrators from administrative databases and could be used to target CalWORKs recipients who are most at risk of having problems after they leave the program. Table entries should be examined by category, relative to the reference group for each category. For example, under “Ethnicity,” the selected reference group is families whose head of household is white/ other. The findings then show the differences in probabilities that an outcome will occur for a certain group relative to the reference group, 70 controlling for the other characteristics in the model. Following our example, the “black” row presents differences in probabilities between blacks and white/other. To control for other characteristics, we use the estimates of the logit model to forecast the probability of a certain outcome, assuming that all individuals were black but leaving the remaining characteristics unchanged. Then, we calculate these probabilities assuming that all individuals were white/other, again leaving the remaining characteristics unchanged. The estimates presented in Tables 8.1 and 8.2 correspond to the mean of the differences of these two values over all families in the sample. Asterisks indicate which effects are estimated to differ from zero at standard confidence levels, with one asterisk indicating the 90 percent and two asterisks the 95 percent confidence level. Findings Starting with language and ethnicity, we find that one-parent families whose primary language is Spanish are more likely than Englishspeaking families to encounter at least two problems (Table 8.1). In particular, Spanish-speaking families are 26 percentage points more likely than English-speaking families to have income below poverty, despite having comparable chances of being employed. For two-parent families, the disparity between Spanish and English speakers with respect to poverty status is greater, at 31 percentage points (Table 8.2). However, our relatively small sample of two-parent families yields insignificant coefficients for the other outcomes. Among all families combined (not shown), lack of health care coverage is significantly more prevalent among Spanish speakers, although this result does not show up for each family type alone. Language and ethnicity play different roles for these families. Hispanic one-parent families are more likely than families of other ethnicities to encounter at least two problems, but the specific issues facing Hispanics are not identical to those facing Spanish speakers. Hispanic families do not have income below poverty more commonly than their counterparts, and the group as a whole is no more likely than its counterparts to lack health coverage. On the other hand, Hispanics more commonly live in crowded housing (by 19 percentage points), but Spanish-speaking families do not, suggesting that this problem does not 71 Table 8.1 Relationship Between Characteristics at Exit and Outcomes at 11–16 Months Postexit, One-Parent Families (marginal effect in percentage points) 72 Characteristics at Exit Primary language Spanish Englisha Ethnicity Black Latino/Hispanic White/othera Months on aid 0–12 13–36a 37–60 Number of children 1–2a 3+ Age of the youngest child 0–2 3–6 7+a Earnings at exit No earnings Positive earningsa Crowded Housing Income Below Poverty 11–16 Months Postexit No Earned Lacks Health Child Care a At Least One At Least Two Income Coverage Big Problem Problem Problems 0.01 0.26** –0.05 Ref Ref Ref 0.15 0.00 0.18** 0.25* Ref Ref Ref Ref 0.07 0.19** Ref 0.05 0.08 Ref 0.09* 0.00 Ref –0.03 0.04 Ref –0.07 0.03 Ref 0.06 0.11* Ref 0.01 0.14* Ref –0.03 Ref –0.01 –0.12** Ref 0.01 –0.05 Ref 0.01 0.11 –0.10** –0.01 –0.13** Ref Ref Ref Ref 0.04 0.05 0.08 0.00 Ref 0.24** Ref 0.11* Ref –0.06 Ref Ref Ref Ref 0.10 0.02 0.17** 0.16** 0.11** 0.00 –0.07** 0.00 –0.02 –0.03 0.04 0.07 –0.09* –0.02 0.06 –0.04 –0.08 –0.03 Ref Ref Ref Ref Ref Ref Ref 0.11 0.04 Ref Ref 0.11 Ref 0.12 0.13* 0.18** 0.16** Ref Ref Ref Ref Table 8.1 (continued) 73 Characteristics at Exit Administrative exit reason Failed to provide information Earnings Client request/noncooperation Othera Crowded Housing –0.14** –0.12** –0.15** Ref Income Below Poverty –0.05 –0.06 –0.12** Ref 11–16 Months Postexit No Earned Lacks Health Child Care a Income Coverage Big Problem –0.12** –0.01 –0.07 Ref 0.12* 0.02 –0.04 Ref –0.09 –0.05 –0.13** Ref SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. aReference category. (Other values shown are relative to this characteristic.) *Statistically significant at the 10 percent level. **Statistically significant at the 5 percent level At Least One Problem –0.03 0.02 –0.12 Ref At Least Two Problems –0.16** –0.19** –0.30** Ref Table 8.2 Relationship Between Characteristics at Exit and Outcomes 11–16 Months Postexit, Two-Parent Families (marginal effect in percentage points) 74 Characteristics at Exit Primary language Spanish Englisha Ethnicity Black Latino/Hispanic White/othera Months on aid 0–12 13–36a 37–60 Number of children 1–2a 3+ Age of the youngest child 0–2 3–6 7+a Earnings at exit No earnings Positive earningsa Crowded Housing –0.05 Ref Income Below Poverty 0.31* Ref 11–16 Months Postexit No Earned Lacks Health Child Care a Income Coverage Big Problem Insufficient data 0.21 0.33 Ref Ref At Least One At Least Two Problem Problems –0.13 Ref 0.26 Ref –0.30 0.11 Ref –0.15 –0.02 Ref 0.04 0.02 –0.45 0.04 0.01 0.11 0.08 0.14 Ref Ref Ref Ref 0.23 0.01 Ref Ref 0.15 0.11 0.01 –0.15 –0.03 0.10 Ref Ref Ref Ref 0.20 –0.33 0.03 0.11 Ref Ref 0.40 –0.22 Ref Ref Ref Ref –0.16 –0.14 0.02 0.13 0.42 0.11 0.29 0.13 Ref Ref –0.10 –0.25 Ref 0.24 0.49* Ref –0.04 –0.08 Ref 0.07 0.12 Ref 0.37 –0.26 Ref Ref 0.25 0.00 0.13** 0.20 Ref Ref Ref Ref Table 8.2 (continued) 75 Characteristics at Exit Administrative exit reason Failed to provide information Earnings Client request/noncooperation Othera Crowded Housing –0.04 –0.17 0.10 Ref Income Below Poverty –0.11 –0.03 –0.15 Ref No Earned Income Lacks Health Coverage –0.07 –0.43** –0.03 Ref SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. aReference category. (Other values shown are relative to this characteristic.) *Statistically significant at the 10 percent level. **Statistically significant at the 5 percent level. Child Care a Big Problem 0.22 0.27 –0.26** Ref At Least One At Least Two Problem Problems –0.08 –0.03 –0.01 Ref –0.09 –0.11 –0.05 Ref stem from the immigrant population. Neither Hispanics nor Spanishspeaking single-parent families are more likely than their counterparts to be unemployed, and none of the results for two-parent families are significant. Looking at black families, unemployment or lack of earned income is more prevalent in these families than in any other group. For example, one-parent black families are 9 percentage points more likely to have no earned income in the 11 to 16 month postexit period. Surprisingly, lower employment rates among single-parent black families do not translate into higher poverty rates or lower living standards when compared to other ethnic groups. Black families are as likely as white/other families to have income below poverty, suggesting that working black families have lower poverty rates than other groups. Also, blacks are not more likely than other ethnic groups to live in crowded housing. Time on aid before leaving CalWORKs is also linked to postexit outcomes for one-parent families but not for two-parent families. Not surprisingly, one-parent families (Table 8.1) who had relied on aid for less than 12 months in the last 5 years perform better than other families with higher dependency on aid in the past. For example, one-parent families with the lowest time on aid before exiting are 12 percentage points less likely to have income below poverty and 10 percentage points less likely to have child care as a barrier to employment. However, highly welfare-dependent families (families who spent more than 36 months out of the last 60 months on aid) are not more likely than moderate welfare users to have more problems (those with 13 to 36 months on aid). Comparing two-parent families by time on aid (Table 8.2) suggests that this is not a good indicator of the problems two-parent families will face after leaving aid. Single-parent families with three or more children are more likely to have at least two problems after leaving CalWORKs. Among these problems, crowded housing stands out as being the most prevalent for large families. One-parent families with three or more children are 24 percentage points more likely than other families to live in crowded housing (Table 8.1). In addition, they are 11 percentage points more likely to have incomes below the poverty threshold. The outcomes for 76 two-parent families are not significantly affected by the number of children in the family. One-parent families with young children are significantly more likely to live in crowded housing, regardless of the total number of children in the family. However, these families do not perform worse than other families in other areas. In fact, they are more likely to have earnings than families whose youngest child is over age six. For two-parent families with the youngest child between ages 3 to 6 years, child care needs are particularly prevalent. Tables 8.1 and 8.2 also compare families by their earnings at exit according to the UI Base Wage File. One-parent families who exit without earnings are 13 percentage points more likely to report that child care is a large barrier to full-time employment. They are also more likely to encounter at least two problems, although specific problems aside from child care are not independently significant. Looking at the magnitude of the estimated effects, however, it appears that crowded housing and lack of health care coverage are the largest problems for both one- and two-parent families. In fact, when we estimated the model on all families combined, both of these outcomes were significantly more likely among families without earnings at exit (not shown). It is interesting to note that additional factors beyond the absence of earnings at exit play a role in explaining poorer outcomes, since there were no significant differences in poverty or the probability of having no earned income 11 to 16 months postexit. Note also that since about two-thirds of families reporting no earnings in the UI file had positive earnings according to the survey (see Chapter 4), the group “No earnings” in Tables 8.1 and 8.2 includes all families with no earnings at exit as well as families working in sectors not covered by the UI file and families moving into larger households with other employed members. Finally, the last rows of Tables 8.1 and 8.2 compare families according to the exit reason as it is recorded in the administrative files. However, these results should be interpreted with caution, because many families simply failed to return forms or provide other information necessary to maintain eligibility. In fact, those families who simply left aid without providing an exit reason often outperformed other families. On one measure only—health coverage—families who left because of 77 failure to provide information had worse outcomes. Together, these findings suggest that families who exited without working closely with a caseworker typically had employment but may not have been aware of all the transitional benefits available, or—given that child care does not appear to be a greater barrier—simply chose to avoid the stigma of participating in public assistance. Summary The last two columns in Tables 8.1 and 8.2 summarize the experiences of leavers, measured by the number of problems encountered while off aid. The findings suggest that the best predictors of such problems among one-parent leavers are Hispanic ethnicity, three or more children in the family, Spanish as the primary language, and no earnings at exit according to the UI file. Except for earnings, characteristics at exit do little to predict the chance of two-parent families encountering at least one problem, although Spanish speakers were more likely to have income below poverty. This is likely due to our relatively small sample of twoparent families and the more complicated situations of these families, which might affect their postwelfare well-being. These characteristics could be used to develop a strategy to target postassistance support services for leavers. For example, programs related to health coverage could be targeted to Spanish-speaking families and families with no earnings according to the UI file. In the same way, housing assistance programs could be targeted to families with more than three children, families with young children, and Hispanic families. Finally, those who exited because of failure to provide information to maintain eligibility appear to have relatively strong outcomes in the 11 to 16 month period, but it might be helpful to provide additional outreach to these families to make sure that they understand the transitional benefits available, especially those allowing them to maintain health coverage. 78 9. Conclusion The survey data reviewed here provide a much richer picture of the experiences of leavers in six Bay Area counties than can be gleaned from administrative data. In fact, the administrative data, often used by policymakers and administrators to develop policies, have serious limitations in understanding outcomes for former recipients. Earnings data from the UI file, in particular, significantly understate both the level of employment and the amount of earnings in the households of welfare leavers. As far as the outcomes for leavers, we find aspects supporting both positive and negative views of their circumstances and prospects. On the positive side, a very high percentage of the households had someone with earnings, with this percentage remaining high throughout the period after leaving CalWORKs. Monthly income increased, albeit slightly, the longer leavers remained off CalWORKs, as did the percentage of households with incomes above the federal poverty level. About three out of four respondents, and four out of five children, had some form of health insurance coverage. Housing conditions also appeared to improve during the time away from CalWORKs, at least for the one-parent leavers. There was also fairly good news on other measures of well-being. Only about one out of ten leaver families experienced incidences of child risk behaviors (e.g., expulsion from school, trouble with police, use of drugs, and pregnancy), and such problems did not increase over time. Moreover, the percentage of respondents reporting substance abuse problems in the household declined among leavers the longer the time since their exit from CalWORKs. However, we also find some areas of concern. Many long-term leavers still had low incomes. Over one out of ten had monthly income falling below 70 percent of the poverty level, and this share of very poor households saw little or no improvement during time off CalWORKs. About half of the families saw their incomes and earnings decrease 79 between the first and second survey periods. Considered broadly, leavers had a high level of participation in public assistance programs. At the same time, many families eligible for nonassistance Food Stamps, nonassistance Medi-Cal, and EITC were not using these benefits. Similarly, about one-quarter of the respondents indicated that they were not aware of the availability of government subsidies that would help pay for child care for families that left welfare. From a policy perspective, our findings relating to the take-up of public assistance benefits after exit or benefit denial are important. Many leavers without health insurance or Food Stamps appeared to be eligible for these benefits. Similarly, there was often a lack of awareness of the availability of EITC and child care subsidies. These gaps point to the need for policies aimed at providing families with information about the availability of these post-CalWORKs benefits. Since it may be difficult to contact recipients after they leave CalWORKs, the dissemination of such information should begin while the families are still on aid. Similarly, this type of information could also be provided to CalWORKs applicants who are denied assistance but who could be eligible for other types of assistance. The results of our profiling analysis, where we identified recipient characteristics most likely to predict problems after exit from CalWORKs, could be useful in targeting supplemental assistance to these families, both while on CalWORKs and immediately after exit. Families exiting without earnings were the most likely to encounter problems, but other characteristics were also associated with later problems, such as having larger families or speaking Spanish as a primary language. Services designed to help recipients retain their jobs after leaving CalWORKs, for example, could be targeted more intensively toward leavers with these characteristics. Our findings on the discrepancy between administrative data and survey data that measure leavers’ earnings may also be important. Policymakers relying on administrative records—the most common source of this information—should recognize that these data may significantly understate earnings. Finally, the question of why many leaver families in poverty choose not to go back on CalWORKs warrants further research. About one- 80 third of our sample of long-term leavers had household incomes below the federal poverty level, and 11 percent were below 70 percent of the poverty level. In examining this question, we found that the survey respondents in the poor nonrecidivist leaver families had a higher level of work activity, but at the same time their household earnings were lower, on average, than their recidivist counterparts. This may have been due to differences in employment on the part of spouses or other household members—questions that were not specifically included in our surveys. 81 Appendix Data Sources and Methods This appendix describes data sources and methods used to compute the empirical findings presented in this report. The discussion starts by defining the study population and summarizing the datasets. It then presents the statistical approach. Finally, we discuss the construction of sample weights used to develop a representative sample of our population of study from the postexit surveys. Study Population The population for the study consists of CalWORKs assistance units, in six Bay Area counties, that left aid during the following time periods: • San Mateo, Santa Clara, and Santa Cruz Counties: fourth quarter 1998 • Contra Costa and Alameda Counties: third quarter 1999 • Sonoma County: fourth quarter 1999. To provide common reference points for characterizing the circumstances of leavers, we sorted the interviews according to the time elapsed since the families left CalWORKs and divided them into two periods of time: • 5 to 10 months (about 2 to 3 quarters) after leaving CalWORKs, and • 11 to 16 months (about 4 to 6 quarters) after leaving CalWORKs. Administrative Datasets This study uses county and state administrative data. We used county administrative data from each county’s Case Data System to identify the study populations. We also used county data to identify 83 family demographic characteristics and, in the case of leavers and the informally diverted, to identify the administrative reason for exit from or denial of cash aid. CDS information was also used to measure postexit and postdiversion receipt of public assistance within the three study counties. We used the state Medi-Cal Eligibility Data System to track receipt of public assistance elsewhere in California, outside the threecounty area, and we used the state Unemployment Insurance Base Wage File to track employment and earnings. The following list describes these datasets in more detail: • MEDS: This database includes monthly program participation for all Medi-Cal-eligible public assistance recipients as well as Food Stamps–only recipients (in recent years). Data are extracted from the system once a month to create a dataset that shows program eligibility month by month. MEDS also identifies county of aid receipt and individual demographic information, including ethnicity, age, gender, and language. Individuals are uniquely identified by Social Security number, and members of a TANF assistance unit are identified by a common case serial number. • UIBWF: Quarterly earnings records collected for the UIBWF are the key state-level data source we use to track employment and earnings. The UI system covers well over 90 percent of all workers in the state, with self-employed individuals constituting the largest category of excluded workers. • CDS: Each of the six counties in our study maintains its own CDS database, which it uses for maintaining current information on public assistance, Food Stamps, and Medi-Cal clients. CDS includes the same program participation and demographic information as MEDS but also contains many variables that are not found in MEDS, such as information on sanctions and reasons cases left public assistance. The counties sent to SPHERE extracts from CDS once a month, from which SPHERE constructed longitudinal files. Assistance units can be linked to MEDS by case serial numbers, and individuals can be linked to MEDS and the UIBWF by Social Security number. 84 Identifying the Study Population from CDS The population was identified using CDS data. To be considered to have left aid, all members of the assistance unit who were active or sanctioned must have left and remained off CalWORKs (off the original case and off any other case) for the next two months. The case had to have contained an aided (or sanctioned) adult. Thus, child-only cases were excluded from the population. Cases that left CalWORKs in a study county but moved to another county where they continued on aid were also excluded from the population; these intercounty transfers are identified in CDS by a specific exit-reason code. Postexit Surveys A central component of our study design is the implementation of a survey of a random sample of families in our study populations. Our survey contained topical modules with questions covering household composition, child well-being, child care, education and training, employment, income, food security, health insurance coverage, family well-being, and welfare experiences. The sampling frame for the survey consisted of all families in the study population, excluding those with a primary language other than English or Spanish. The sample is stratified by county; and for San Mateo, Santa Clara, and Santa Cruz Counties; the sample frame is also stratified by one-parent cases and two-parent cases. Table A.1 shows the distribution of our study population and sample by county and by case type. The interviewing was conducted primarily by telephone, although in some instances when attempts to contact the sample member by telephone were unsuccessful, an interviewer went to the sample member’s home and completed the interview in person. The interviewers were given contact information from CDS and MEDS, including addresses and telephone numbers, as well as names and Social Security numbers of adults in the assistance unit. Periodically, we gave the interviewers updated information from CDS and MEDS for sample members they had not yet located, and interviewers made use of public and private data sources to locate sample members. 85 Table A.1 Population and Sample Distributions, by County, Case, and Percentage Share Distribution of cases by county Alameda Contra Costa Santa Clara Santa Cruz San Mateo Sonoma Study Population Sample 28 18 24 22 26 11 6 14 8 12 8 23 Distribution of cases by type of case One-parent cases Two-parent cases 90 10 87 13 NOTE: Columns may not sum to 100 because of rounding. The survey respondent was almost always the mother or female caretaker of the children in the assistance unit. Some survey questions were asked about a “focal” child. The focal child was selected randomly from the children living with the respondent at the time of the first interview, and from those children for whom the respondent was responsible. Note that the focal child did not need to be the biological or adopted child of the respondent. Surveys were administered in two waves for each family, over a period of roughly 15 months after exit from CalWORKs. Table A.2 shows the population of leavers and the number of interviews completed in the 5 to 10 and 11 to 16 month postexit periods by county. The response rates for the total combined sample were 54 percent for the first set of interviews and 66 percent for the second wave.1 In addition to information on postexit aid use, employment, and earnings, MEDS and the UIBWF provide historical information on ____________ 1Not every case in our sample responded to both interviews. In some instances, we conducted separate analyses that linked only those individuals who responded to interviews in each period. 86 Table A.2 Population, Samples, and Survey Response Counts Formulation of Weights County Alameda Contra Costa San Mateo Santa Clara Santa Cruz Sonoma Total Population 1,082 964 301 1,013 238 309 3,907 Sample 238 304 170 149 184 309 1,354 First Wave 133 203 47 38 66 242 729 Second Wave 169 206 98 84 118 215 890 these outcomes that were used, along with other measures of family characteristics, to construct weights to adjust survey data for nonresponse. Constructing a representative sample of the population of leavers in the six counties included in this study required the computation of sample weights that adjusted for differences in sampling rates, nonresponse, and other sources of bias. To create these weights, we followed a two-stage approach. In the first stage, we computed a basic county weight equal to the inverse probability of selection, where the probability of selection is equal to the number of respondents in the county divided by the study population in the county. This basic weight adjusts sample observations for differences in sampling and response rates across counties. For the second stage, we applied iterative raking to balance a population’s distributions by ethnicity, age of the case head, and earning status at exit according to UIBWF. Iterative raking, used by the Census Bureau to adjust sample weights, adjusts the baseline take-up rate so that the predicted number of cases converges to the population benchmarks, while minimally affecting the baseline take-up rate. For each iteration of each grouping variable (called raking variable), an adjustment factor equal to the number of cases in the study population divided by the number of predicted totals using the weighted sample is computed and applied to the estimate of that cell. This procedure is 87 repeated until the estimated number of participants in each cell equals the study population controls. The following grouping variables were used: • Ethnicity: white, Hispanic, black, other, • Age of case head: 0 to 15, 16 to 25, 26 to 35, 36 to 45, 46+, • UIBWF earning status: positive earnings, no earnings. Table A.3 compares population and sample distributions after applying the constructed weights by grouping variable. Our calibration creates predicted distributions that exactly match those in the study population. This arises by construction, since our procedure minimizes discrepancies in this match. Table A.4 presents estimates of the distribution of subgroups of leavers for additional demographic characteristics. In contrast to results Table A.3 Demographic Characteristics of All Leavers, by Grouping Variables and Percentage Share Demographic Characteristics Ethnicity Black Hispanic White Other Leavers at 5–10 Monthsa (Population) 29 27 35 9 Leavers at 5–10 Monthsa (Sample) 29 27 35 9 Leavers at 11–16 Monthsb (Population) 28 27 35 9 Leavers at 11–16 Monthsb (Sample) 28 27 35 9 Age of case head 16–25 32 32 31 26–35 36 36 36 36–45 25 25 26 46+ 7 7 7 31 36 26 7 Earnings at exit No earnings 27 27 26 Positive earnings 73 73 74 26 74 Number of cases 3,038 541 2,684 NOTE: Columns may not sum to 100 because of rounding. aIncludes short-term and long-term leavers. bIncludes only long-term leavers. 575 88 Table A.4 Additional Demographic Characteristics of All Leavers, by Grouping Variables and Percentage Share Demographic Characteristics Primary language English Spanish Leavers at 5–10 Monthsa (Population) 93 7 Leavers at 5–10 Monthsa (Sample) 91 9 Leavers at 11–16 Monthsb (Population) 93 7 Number of children 0–1 2 3+ 53 28 19 53 30 18 52 29 19 Age of youngest 0–2 32 34 3–5 25 24 6–11 24 25 12+ 18 17 31 25 25 19 Months on cash aid in preceding 5 years 1–12 13–36 37–60 20 26 54 20 25 55 19 25 55 Number of cases 3,038 541 2,684 NOTE: Columns may not sum to 100 because of rounding. aIncludes short-term and long-term leavers. bIncludes only long-term leavers. Leavers at 11–16 Monthsb (Sample) 91 9 51 29 20 35 25 23 16 22 28 50 575 in Table A.3, our weighting mechanism does not automatically match these numbers, so the closeness of our estimates to the study population indicates that our weighting mechanism performs well in creating a representative sample of the study population. 89 Table A.5 Counties Included in Each Region Bay Area Alameda Contra Costa Marin Napa San Francisco San Mateo Santa Clara Santa Cruz Solano Sonoma Other Southern Los Angeles California Farm Belt Los Angeles Orange Colusa Riverside El Dorado San Bernardino Fresno San Diego Glenn Santa Barbara Imperial Ventura Kern Kings Madera Merced Monterey Placer Sacramento San Benito San Joaquin San Luis Obispo Stanislaus Sutter Tulare Yolo Yuba North and Mountain Alpine Amador Butte Calaveras Del Norte Humboldt Inyo Lake Lassen Mariposa Mendocino Modoc Mono Nevada Plumas Shasta Sierra Siskiyou Tehama Trinity Tuolumne 90 Bibliography Daly, Mary, “Labor Market Effects of Welfare Reform,” FRBSF Economic Letter 97-24, Federal Reserve Bank of San Francisco, San Francisco, California, August 29, 1997. Gritz, R. Mark, et al., Welfare Recidivism and Use of Non-Cash Aid by Welfare Leavers in California, The SPHERE Institute, report prepared for the U.S. Department of Health and Human Services (Assistant Secretary for Planning and Evaluation), Burlingame, California, March 30, 2001a. Gritz, R. Mark, David Mancuso, Charles J. Lieberman, and Vanessa L. Lindler, Assessing the Family Circumstances of TANF Applicants and Leavers in Contra Costa and Alameda Counties, The SPHERE Institute, report prepared for the U.S. Department of Health and Human Services (Assistant Secretary for Planning and Evaluation), the Contra Costa County Employment and Human Services Department, and the Alameda County Social Services Agency, October 26, 2001b. Lieberman, Charles, and David C. Mancuso, “Are California’s Welfare Leavers Staying Off Public Assistance? Welfare Recidivism and Use of Non-Cash Aid Since Welfare Reform,” California Policy Review, The SPHERE Institute, Burlingame, California, October 2001. MaCurdy, Thomas, David Mancuso, and Margaret O’Brien-Strain, The Rise and Fall of California’s Welfare Caseload: Types and Regions, 1980– 1999, Public Policy Institute of California, San Francisco, California, 2000. Mancuso, David C., Charles J. Lieberman, Vanessa L. Lindler, and Anne Moses, Examining Circumstances of Individuals and Families Who Leave TANF: Assessing the Validity of Administrative Data, Final 18Month Report, The SPHERE Institute, report prepared for the U.S. Department of Health and Human Services (Assistant Secretary for Planning and Evaluation), the U.S. Department of Housing and Urban Development, the San Mateo County Human Services Agency, 91 the Santa Clara County Social Services Agency, and the Santa Cruz County Human Resources Agency, November 30, 2001. Mancuso, David C., and Vanessa L. Lindler, Examining the Circumstances of Welfare Leavers and Sanctioned Families in Sonoma County, The SPHERE Institute, report prepared for the Sonoma County Human Services Department, June 29, 2001. 92 About the Authors THOMAS MACURDY Thomas MaCurdy is an adjunct fellow at PPIC, a professor of economics at Stanford University, and a senior fellow at the Hoover Institution. His principal research explores topics in the areas of income transfer programs, human resources, and labor markets. His recent studies investigate the consequences of governmental policies underlying welfare programs, unemployment compensation, Social Security, Medicare, Medicaid, and various forms of public assistance for lowincome populations. He holds an A.B. from the University of Washington and a Ph.D. in economics from the University of Chicago. GRECIA MARRUFO Grecia Marrufo is a research associate at The SPHERE Institute and a research associate at the Stanford Institute of Economic Policy Research. Her research has focused on the financing of social insurance programs and modeling the effect of labor regulations on such programs. She is also working on a U.S. Department of Agriculture project to evaluate alternative certification procedures in the Food Stamps program, focusing on the empirical simulation of changes in benefits arising from alternative reporting requirements. She holds a B.A. in economics from the Instituto Tecnológico Autónomo de México and a Ph.D. in economics from the University of Chicago. MARGARET O’BRIEN-STRAIN Margaret O’Brien-Strain is an adjunct fellow at PPIC and Director of Social Policy at The SPHERE Institute. She is interested in a broad range of social services issues and has written about child care, home visitation programs, performance measurement, and welfare reform in California. She holds a B.A. in economics from Swarthmore College and a Ph.D. in economics from Stanford. 93 Related PPIC Publications Employers and Welfare Recipients: The Effects of Welfare Reform in the Workplace Harry J. Holzer and Michael A. Stoll The Basic Skills of Welfare Recipients: Implications for Welfare Reform Hans P. Johnson and Sonya M. Tafoya Does California’s Welfare Policy Explain the Slower Decline of Its Caseload? Thomas E. MaCurdy, David C. Mancuso, and Margaret O’Brien-Strain Reform Reversed? The Restoration of Welfare Benefits to Immigrants in California Thomas MaCurdy and Margaret O’Brien-Strain The Rise and Fall of California’s Welfare Caseload: Types and Regions, 1980–1999 Thomas MaCurdy, David Mancuso, and Margaret O’Brien-Strain Who Will Be Affected by Welfare Reform in California? Thomas MaCurdy and Margaret O’Brien-Strain Federal Formula Grants and California: TANF and Welfare Programs Tim Ransdell PPIC publications may be ordered by phone or from our website (800) 232-5343 [mainland U.S.] (415) 291-4400 [Canada, Hawaii, overseas] www.ppic.org 95" } ["___content":protected]=> string(102) "

R 903TMR

" ["_permalink":protected]=> string(91) "https://www.ppic.org/publication/what-happens-to-families-when-they-leave-welfare/r_903tmr/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(8285) ["ID"]=> int(8285) ["post_author"]=> string(1) "1" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2017-05-20 02:36:15" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(3443) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(8) "R 903TMR" ["post_type"]=> string(10) "attachment" ["slug"]=> string(8) "r_903tmr" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(12) "R_903TMR.pdf" ["wpmf_size"]=> string(6) "672172" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(170158) "What Happens to Families When They Leave Welfare? ••• Thomas MaCurdy Grecia Marrufo Margaret O’Brien-Strain 2003 PUBLIC POLICY INSTITUTE OF CALIFORNIA Library of Congress Cataloging-in-Publication Data MaCurdy, Thomas E. What happens to families when they leave welfare? / Thomas MaCurdy, Grecia Marrufo, Margaret O'Brien-Strain. p. cm. Includes bibliographical references. ISBN: 1-58213-076-0 1. Welfare recipients—California—San Francisco Bay Area— Longitudinal studies. 2. California Work Opportunity and Responsibility to Kids (Program). I. Marrufo, Grecia. II. O'BrienStrain, Margaret. III. Title. HV99.S4M33 2003 362.5'4'097946—dc22 2003015550 Copyright © 2003 by Public Policy Institute of California All rights reserved San Francisco, CA Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source and the above copyright notice is included. PPIC does not take or support positions on any ballot measure or on any local, state, or federal legislation, nor does it endorse, support, or oppose any political parties or candidates for public office. Research publications reflect the views of the authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. Foreword A significant reduction in the nation’s welfare caseload has long been touted as the major success of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996. A PPIC report in 2002 documented that California reduced its caseload by larger numbers than any other state, although the percentage reduction was smaller than in most other states throughout the nation. A PPIC-sponsored research team—Thomas MaCurdy, Grecia Marrufo, and Margaret O’BrienStrain—has now taken a close look at a second measure of the success of welfare reform, how well families who have left welfare are doing. What Happens to Families When They Leave Welfare? concludes that, depending on your point of view, the glass is either half empty or half full. The researchers interviewed nearly 1,400 individuals in six Bay Area counties 5 to 16 months after they had departed the rolls. Over half of both one- and two-parent families had managed to earn enough income to raise them above the poverty level. However, two-parent families were faring much worse when it came to earning income sufficient to keep them out of poverty. One could argue that these families are so close to poverty wages that the slightest setback would land them back in poverty and in line at the welfare office. The glass-is-half-full folks would argue that at least these families are back in the labor force, establishing an employment record, and building a foundation for their families that is based on work, not welfare. The authors point out, however, that a remarkably large percentage of the families no longer on welfare were still eligible for such benefits as Food Stamps, Medi-Cal, the Earned Income Tax Credit, and day care but were not receiving these benefits either because they did not know about them or did not bother to apply. For families at or near the poverty line, these benefits are a societal safety net specifically designed to help those who are helping themselves. The authors strongly recommend that state policy be designed to reach out to these families iii and inform them of the benefit programs available. The long-term success of welfare reform may depend on how well these special benefit programs are funded and implemented. Otherwise, the state faces the prospect that large numbers of families no longer on welfare rolls may once again return for a full measure of the cash benefits. One concludes from reading this report that the welfare reform act of 1996 is doing better than many had predicted but that much remains to be done to see that the basic principle of “the end of welfare as we know it” can be maintained through this and future recessions. PPIC has now published nine reports to monitor this seminal change in welfare policy, and we will continue to carry out research on, and monitor the effects of, this new social policy well into the 21st century. David W. Lyon President and CEO Public Policy Institute of California iv Summary Between August 1996 and September 2001, the number of welfare recipients in California dropped by more than 1.4 million people. Most families leaving aid during this period left for the same reason families have always left aid: They found employment that increased their family income enough for them to lose eligibility for assistance. The booming economy was clearly an important factor in the caseload declines during this period. However, in these first years following welfare reform, it was feared that some families would leave aid—because of rigorous participation requirements, the specter of time limits, confusion about program rules (especially for immigrants), or merely the publicity about “the end of welfare as we know it”—without alternative sources of income. The same dramatic caseload declines seen by some as heralding the triumph of welfare reform were seen by others as an alarm bell regarding the well-being of former welfare recipients. Unfortunately, despite good information about the numbers leaving aid, we have only extremely limited information about the well-being of families after they leave the welfare rolls. To assist families in successfully leaving CalWORKs (California’s TANF1 program) and to prevent them from falling through the social safety net, policymakers must understand the circumstances of these “leavers” and how their circumstances change over time. To contribute to this understanding, this report addresses five basic questions: 1. How many families who have left CalWORKs remain off the program for extended periods of time? 2. What are the circumstances of these families, and do their circumstances improve over time? 3. What public assistance and services do leavers rely upon? ____________ 1Temporary Assistance for Needy Families. v 4. What problems and barriers do these leavers encounter while off CalWORKs? 5. Are there factors that determine which families who leave CalWORKs succeed in staying off aid (long-term leavers)? This report draws on extensive telephone surveys of welfare leavers conducted in six Bay Area counties (Alameda, Contra Costa, San Mateo, Santa Clara, Santa Cruz, and Sonoma) in the late 1990s to characterize the experiences of CalWORKs leavers over time. These surveys were conducted by The SPHERE Institute in cooperation with the six counties, with support from the U.S. Department of Health and Human Services and the William and Flora Hewlett Foundation. Key Findings Two waves of surveys give a picture of the circumstances of leavers 5 to 10 months after exit (depending on the county) and again 11 to 16 months after exit. In general, we distinguish between one-parent families and two-parent families. Our analysis presents snapshots of how well families are doing after they leave the CalWORKs program, describing their economic security, use of public assistance, and difficulties encountered in maintaining employment and coping with conditions adversely affecting their overall environment. During the study period, one in ten cases leaving aid was a two-parent family. These snapshots summarize the experiences of nonrecidivist leavers during the two postexit periods. Table S.1 presents an overview of selected circumstances for leavers about one year (11 to 16 months) after exiting CalWORKs. For longterm leavers, those families still off CalWORKs 11 to 16 months after exit, Table S.2 summarizes selected circumstances where families’ living conditions had improved or worsened between the two waves of interviews. In particular: • Monthly income for the average household was about $2,400, enough to bring 71 percent of one-parent leavers and 58 percent of two-parent leavers above poverty. Only 29 percent of one-parent leavers and 20 percent of two-parent leavers had income high enough to disqualify them for Medi-Cal. About 1 in 10 vi Table S.1 Selected Circumstances of Long-Term CalWORKs Leavers 11–16 Months Postexit Family Characteristics Average household income (all households) Household has earnings Respondent employed Average monthly earnings (households with earnings) Household above poverty Household receives Food Stamps Respondent uninsured Children uninsured Substandard housing Crowded housing Excessive rent burden Unstable childcare Child risk behaviors Household substance abuse Domestic violence One Parent Two Parents $2,411 $2,275 90% 77% $2,160 71% 9% 28% 18% 14% 17% 19% 17% 13% 16% 23% 94% 59% $2,160 58% 21% 29% 20% 28% 48% 33% 11% 7% 4% 11% SOURCES: The SPHERE Institute surveys of welfare leavers in six Bay Area counties, MEDS. Table S.2 Change in Selected Circumstances of Long-Term CalWORKs Leavers Between First and Second Postexit Survey Periods Family Characteristics Average household income Household has earnings Respondent employed Average household earnings Households above poverty Respondent health insurance Child health insurance Substandard housing Crowded housing Unstable childcare Child risk behaviors Substance abuse Domestic violence One Parent Two Parent — — Better Better — — Better Better Better — — Worse — Better Better Better Better Worse Worse Worse Worse — Better — SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. “—” indicates no significant change over time. vii households was very poor, with income below 70 percent of the poverty threshold. • Around 90 percent of families had at least one adult in the workforce. Workforce participation of respondents (almost always mothers) in one-parent families was quite high, around 77 percent. Mothers in two-parent families were somewhat less likely to work. However, for both one-parent and two-parent households, the surveys captured a significantly higher share of households with earnings than could be found through administrative data. • Income gains between the two survey waves were quite modest, only $60 to $70 per month on average. Half of the households experienced no income growth between the two periods. • Many families were eligible for but not receiving Food Stamps, often worth a substantial amount. According to administrative data, 21 percent of two-parent families and 9 percent of one-parent families received nonassistance food stamps 11 to 16 months after leaving aid. About 30 percent of one-parent families and over 40 percent of two-parent families who were eligible for Food Stamps were not receiving them. Our estimates of the benefits that eligible households would have received if they had applied for Food Stamps show that the median value of the forgone benefits was $212 per month for one-parent families and $262 per month for two-parent families. • About 50 percent of families reported Medi-Cal coverage 11 to 16 months after exit, down from about 60 percent in the 5 to 10 month period. However, administrative data show higher rates of MediCal enrollment, especially for two-parent families, so some families may not have been aware that they were still covered by Medi-Cal. About one in four survey respondents reported being uninsured, and about one in five children were reported uninsured, including about 10 percent of children who were eligible for Medi-Cal. • Almost half of two-parent families lived in crowded housing, onethird faced excessive rent burden, and one-quarter lived in substandard housing. Crowding and substandard housing were viii more common for long-term two-parent leavers than for those surveyed 5 to 10 months after exit, whereas housing conditions improved over time for one-parent leavers. • Although child care subsidies are available to working parents who leave welfare, more than one-quarter of one-parent leavers were unaware of these subsidies, as were one-third of two-parent leavers. About one-quarter of the respondents reported that obtaining child care was a significant barrier to full-time employment. In addition, the majority of respondents in working two-parent families and 40 percent of those in working one-parent families were unaware of the availability of the Earned Income Tax Credit (EITC). • Only about one out of ten families experienced incidences of child risk behaviors (e.g., expulsion from school, trouble with police, use of drugs, pregnancy), and such problems did not increase over time. Moreover, the percentage of respondents reporting substance abuse problems in the household declined among leavers the longer the time since their exit from CalWORKs. Across all outcome areas shown in Table S.2, it appears that circumstances generally improved more often than they got worse for leavers, particularly for one-parent families. We also find some areas of concern, however. Many long-term leavers still had very low incomes, and this share of very poor households saw little or no improvement during their time off CalWORKs. Considered broadly, leavers had a high level of participation in public assistance programs. Yet many families were not using benefits for which they were probably eligible, including Medi-Cal, child care, EITC, and Food Stamps. Taken together, families may be missing out on substantial support intended to help working poor families. In addition to the comparison of survey findings over time, we also assessed what factors predicted recidivism and other poor outcomes for families. On the topic of recidivism, we find some demographic characteristics at exit that are more associated with recidivism. Larger families were less likely to remain off welfare, as were families with younger children. Black families were more likely than white families to ix return to CalWORKs. Two-parent Hispanic families were more likely than other two-parent to successfully stay off CalWORKs. Speaking Spanish as a primary language was not a barrier to staying off aid. Families who returned to aid at 11 to 16 months after exit had lower income and lower earnings than other families 5 to 10 months after exit but primarily because they worked fewer hours. Recidivist families were also more likely to have still relied on public assistance 5 to 10 months after exit. Nevertheless, a large segment of long-term leavers are quite poor and face problems and barriers similar to those families who return to aid. Factors that had little effect on recidivism were sometimes predictive of other problems postexit. In a multivariate analysis of characteristics at exit associated with problems down the road for long-term one-parent leavers, such as crowded housing, living in poverty, lack of earnings, lack of health coverage, or child care problems, we find that Spanish-speaking families, bigger families, and families with no earnings at exit were more likely to encounter one or more problems by a year after exit. Families who had spent less than a year on aid were less likely to encounter at least one problem and particularly less likely to experience poverty or child care problems. Policy Implications From a policy perspective, our findings related to the take-up of public assistance benefits after exit from or denial of CalWORKs are important. Many leavers without health insurance or Food Stamps appeared to be eligible for these benefits. We estimate that many families were forgoing substantial benefits; 25 percent of one-parent families relinquished more than $300 per month in Food Stamps benefits alone. Similarly, there was often a lack of awareness of the availability of the Earned Income Tax Credit and child care subsidies. This points to the need for policies aimed at providing families with information about the availability of these post-CalWORKs benefits. Since it may be difficult to contact recipients after they leave CalWORKs, the dissemination of such information should begin while the families are still on aid. Similarly, this type of information could also be provided to CalWORKs applicants who are denied assistance but who could be eligible for other x types of assistance. Alternatively, presumptive eligibility for Food Stamps after families leave CalWORKs, allowed under the 2002 Farm Bill, might provide valuable support to working families who might not maintain these benefits otherwise. The results of our profiling analysis, where we identified recipient characteristics most likely to predict problems after exit from CalWORKs, could be useful in targeting supplemental assistance to these families, both while on CalWORKs and immediately after exit. Services designed to help recipients retain their jobs after leaving CalWORKs, for example, could be targeted more intensively toward leavers with these characteristics. In addition, families who have no earnings at the time of exit might be a flag of potential problems. Our findings on the discrepancy between administrative data and survey data that measure leavers’ earnings may also be quite important. Policymakers relying on administrative earnings from unemployment insurance files—the most common source of this information—should recognize that these data might significantly understate earnings. Finally, the question of why many leaver families in poverty choose not to go back on CalWORKs warrants further research. About onethird of our sample of long-term leavers had household incomes below the federal poverty level, and 11 percent were in the “very poor” category with monthly incomes of less than 70 percent of the poverty level. Many of the lessons learned from our analysis merely confirm conventional wisdom. Weak work experience constitutes an important determinant of whether families return to CalWORKs, and families with less child support have a greater chance of reentering the rolls as well. Still, many successful leavers look very much like families who go back to CalWORKs, and earnings are far from the determining factor for whether a family returns to cash assistance. Poor long-term leavers face problems and barriers similar to those families who return. Recidivists were only marginally more likely to indicate that they had been experiencing domestic violence, unstable child care, substance abuse, and depression; and, somewhat surprising, fewer reported that child care was a barrier to employment. After accounting for work and earnings experience, no particular condition appears as a prominent factor or trigger event for predicting who will be reentering CalWORKs. xi Contents Foreword ..................................... Summary..................................... Figures ...................................... Tables ....................................... Acknowledgments ............................... iii v xv xvii xxi 1. INTRODUCTION ........................... 2. DATA AND METHODOLOGY .................. Survey Methodology ........................... Administrative Data ........................... Categories of Leavers .......................... 1 5 5 7 8 3. THE BAY AREA IN CONTEXT .................. 4. ECONOMIC OUTCOMES ..................... Family Income, Earnings, and Employment ............ Growth in Income and Earnings for Long-Term Leavers .... Earnings Reported in Administrative and Survey Data ..... Barriers to Employment ........................ Summary .................................. 11 17 17 20 23 25 26 5. TAKE-UP OF ASSISTANCE AFTER LEAVING CalWORKs ................................ Postexit Participation in Public Assistance: Evidence from the Surveys ............................. Snapshots of Participation After Leaving CalWORKs .... Change in Participation by Long-Term Leavers over Time ................................. Snapshots and Participation Trends Using Administrative Data .................................. Take-Up of Food Stamps and EITC................. Food Stamps .............................. Earned Income Tax Credit ..................... 29 29 30 32 34 37 37 38 xiii Health Insurance Coverage ....................... 39 Summary .................................. 44 6. WELL-BEING MEASURES ..................... Housing Conditions ........................... Child Care ................................. Child and Respondent Well-Being .................. Summary .................................. 47 47 49 52 55 7. WHAT DISTINGUISHES LONG-TERM LEAVERS FROM RECIDIVIST LEAVERS? .................. Demographic Characteristics of Recidivists ............ Short-Term Leavers: Income and Work Effort .......... Short-Term Leavers: Public Assistance ............... Barriers to Employment and Knowledge of Assistance ..... What Distinguishes Poor Long-Term Leavers from Recidivist Leavers? ............................... Summary .................................. 57 57 61 63 65 65 68 8. WHICH FAMILIES ENCOUNTER MORE PROBLEMS WHEN LEAVING CalWORKs? .................. Methodology ............................... Findings .................................. Summary .................................. 69 69 71 78 9. CONCLUSION ............................. 79 Appendix: Data Sources and Methods .................. 83 Bibliography .................................. 91 About the Authors ............................... 93 Related PPIC Publications .......................... 95 xiv Figures 3.1. Share of Leavers Returning to CalWORKs by Months After Exit from CalWORKs, One-Parent Families Who Left Aid in 1998........................... 3.2. Share of Leavers Returning to CalWORKs by Months After Exit from CalWORKs, Two-Parent Families Who Left Aid in 1998........................... 3.3. Share of Leavers Receiving Nonassistance Food Stamps by Months After Exit, One-Parent Families Who Left Aid in 1998.............................. 3.4. Share of Leavers Receiving Nonassistance Food Stamps by Months After Exit, Two-Parent Families Who Left Aid in 1998.............................. 3.5. Share of Leavers Receiving Nonassistance Medi-Cal by Months After Exit, One-Parent Families Who Left Aid in 1998 ................................ 3.6. Share of Leavers Receiving Nonassistance Medi-Cal by Months After Exit, Two-Parent Families Who Left Aid in 1998 ................................ 4.1. Median Hourly Wage of Working Mothers ......... 4.2. Transportation as a Barrier to Full-Time Work ....... 6.1. Out-of-Pocket Child Care Expenses .............. 6.2. Awareness of Child Care Subsidies ............... 6.3. Child Risk Behaviors ........................ 6.4. Substance Abuse Reported by Respondents ......... 6.5. Domestic Violence Reported by Respondents ........ 7.1. Median Wage at 5–10 Months After Exit ........... 12 13 14 15 15 16 21 26 50 51 53 54 54 62 xv Tables S.1. Selected Circumstances of Long-Term CalWORKs Leavers 11–16 Months Postexit ................. S.2. Change in Selected Circumstances of Long-Term CalWORKs Leavers Between First and Second Postexit Survey Periods ............................ 2.1. Population, Sample Sizes, and Survey Response Counts .. 2.2. Demographic Characteristics of CalWORKs Leavers, by Percentage Share .......................... 4.1. Income Status of CalWORKs Leavers, from Survey Data .................................. 4.2. Family Earnings and Employment of CalWORKs Leavers, from Survey Data .................... 4.3. Mothers’ Employment Status, by Percentage Share..... 4.4. Growth in Income and Earnings for Long-Term CalWORKs Leavers, from Survey Data ............ 4.5. Family Earnings and Employment of CalWORKs Leavers, from Administrative Data ............... 4.6. Percentage Reporting Whether Child Care Creates a Problem for Full-Time Work .................. 5.1. Participation in Programs After Leaving CalWORKs, from Survey Data .......................... 5.2. Changes in Participation in Programs After Leaving CalWORKs, from Survey Data ................. 5.3. Participation in Programs After Leaving CalWORKs, from Administrative Data ..................... 5.4. Changes in Participation in Programs After Leaving CalWORKs, from Administrative Data ............ 5.5. Forgone Food Stamps Benefits of CalWORKs Leavers .. 5.6. Awareness and Use of the Earned Income Tax Credit Among CalWORKs Leavers, by Percentage Share ..... 5.7. Health Insurance Coverage of CalWORKs Leavers, by Percentage Share .......................... vii vii 6 10 18 19 20 22 23 25 30 33 34 35 38 39 40 xvii 5.8. Health Insurance Coverage of Children of CalWORKs Leavers, by Percentage Share ................... 5.9. Health Insurance Coverage of CalWORKs Leavers’ Children Eligible for Medi-Cal, by Percentage Share .... 5.10. Household Income Relative to Federal Poverty Level of Uninsured Respondents, by Percentage Share ........ 5.11. Type of Health Insurance Coverage Offered by Employers of CalWORKs Leavers, by Percentage Share .. 6.1. Housing Outcomes of CalWORKs Leavers, by Percentage Share .......................... 6.2. Primary Child Care Arrangements of CalWORKs Leavers, by Percentage Share ................... 6.3. Percentage of Time Child (Ages 5 to 13) in CalWORKs Leaver Family Was Left Unsupervised in Prior Month... 7.1. Demographic Characteristics of One-Parent CalWORKs Leavers, by Percentage Share ................... 7.2. Demographic Characteristics of Two-Parent CalWORKs Leavers, by Percentage Share ................... 7.3. Income Status of CalWORKs Leavers, One-Parent and Two-Parent Families Combined ................ 7.4. Participation in Programs After Leaving CalWORKs, One-Parent and Two-Parent Families Combined ...... 7.5. Logit Estimates on the Probability That Leavers Will Return to CalWORKs 6–12 Months Postexit ........ 7.6. Selected Circumstances of Recidivists and Poor LongTerm CalWORKs Leavers 11–16 Months Postexit..... 8.1. Relationship Between Characteristics at Exit and Outcomes at 11–16 Months Postexit, One-Parent Families ................................ 8.2. Relationship Between Characteristics at Exit and Outcomes 11–16 Months Postexit, Two-Parent Families ................................ A.1. Population and Sample Distributions, by County, Case, and Percentage Share ........................ A.2. Population, Samples, and Survey Response Counts Formulation of Weights ...................... A.3. Demographic Characteristics of All Leavers, by Grouping Variables and Percentage Share ................. 41 42 43 44 48 49 52 58 59 61 64 66 67 72 74 86 87 88 xviii A.4. Additional Demographic Characteristics of All Leavers, by Grouping Variables and Percentage Share......... 89 A.5. Counties Included in Each Region ............... 90 xix Acknowledgments We are grateful to the many individuals and organizations who contributed to this report and the earlier studies and surveys. Special thanks to Charles Lieberman, who was a major contributor to an earlier version of this report, and thanks as well to West Addison and Vanessa Lindler for their research support. Mark Gritz, David Mancuso, and Anne Moses all had important roles in individual studies included in this research. We received thoughtful comments from Gary Bjork, Maria Cancian, David Neumark, Deborah Reed, Lorien Rice, and Mary Tran. The surveys were funded through grants from the U.S. Department of Health and Human Services, Assistant Secretary for Planning and Evaluation, and the William and Flora Hewlett Foundation, with assistance from the San Mateo County Human Services Agency, the Contra Costa County Employment and Human Services Department, the Sonoma County Human Services Department, the Santa Clara County Social Services Agency, the Santa Cruz Human Resources Agency, and the Alameda County Social Services Agency. Supplemental support was provided by the Smith-Richardson Foundation, the Donner Foundation, and the National Institutes of Health (grant HD32055-02). Opinions expressed in this report are those of the authors and do not represent the views of the Public Policy Institute of California or any agency funding this research. xxi 1. Introduction Between August 1996 and September 2001, the number of welfare recipients in California dropped by more than 1.4 million people. Most families leaving aid during this period left for the same reason families have always left aid: They found employment that increased their family income enough for them to lose eligibility for assistance. The booming economy was clearly an important factor in the caseload declines during this period. However, in these first years following welfare reform, it was feared that some families would leave aid—because of rigorous participation requirements, the specter of time limits, confusion about program rules (especially among immigrants), or merely the publicity about “the end of welfare as we know it”—without alternative sources of income. The same dramatic caseload declines seen by some as heralding the triumph of welfare reform were seen by others as an alarm bell regarding the well-being of former welfare recipients. Unfortunately, despite good information about the numbers leaving aid, we have only extremely limited information about the well-being of families after they leave the welfare rolls. Administrative data from welfare systems allow us to track clients’ moves on and off public assistance programs, including take-up of noncash aid; and data from the unemployment insurance (UI) system provide quarterly earnings for employed individuals. Still, these data tell us little about the economic resources available to former welfare families or about their family circumstances, such as access to health care, housing conditions, and use of child care. To shed more light on these issues, the Assistant Secretary for Planning and Evaluation (ASPE) at the U.S. Department of Health and Human Services funded a series of studies of former welfare recipients to track outcomes for these families after they left aid in 1998 1 and 1999.1 Several of these studies, conducted by The SPHERE Institute, followed clients who left aid in the Bay Area counties of Alameda, Contra Costa, San Mateo, Santa Clara, and Santa Cruz. The William and Flora Hewlett Foundation funded an additional study of those leaving aid in Sonoma County in 1999. In each of these six Bay Area counties, former clients were interviewed twice within the first 18 months after they left aid. Together, the survey data collected on these “welfare leavers” provide a rich portrait of the family circumstances of these families and how these circumstances change over time. This report combines the survey data on leavers in these six counties, along with state- and county-level administrative data, to address five basic questions: 1. How many families who have left CalWORKs (California’s TANF2 program) remain off the program for extended periods of time? 2. What are the circumstances of these families, and do their circumstances improve over time? 3. What public assistance and services do leavers rely upon? 4. What problems and barriers do these leavers encounter while off CalWORKs? 5. Are there factors, besides income, that determine which families who leave CalWORKs succeed in staying off aid (long-term leavers)? The answers to these questions do not speak to the success or failure of welfare reform, because equivalent data are not available for prereform leavers. Nor is it possible to conduct very nuanced comparisons of these counties to other surveyed areas. Nevertheless, the answers to these questions may have powerful policy implications as California seeks to improve the CalWORKs program, especially as additional requirements are imposed by the reauthorization of the federal welfare reform legislation. ____________ 1These studies are summarized at http://www.aspe.hhs.gov/hsp/leavers99/ index.htm. 2Temporary Assistance for Needy Families. 2 The report is organized as follows. Chapter 2 describes the survey data and methodological approach. Because our surveys are all based in the Bay Area, Chapter 3 explores the degree to which experiences in the Bay Area can be expected to match those elsewhere in the state. The next three chapters then examine different aspects of the circumstances of leavers 5 to 10 months and 11 to 16 months after leaving aid. Chapter 4 considers the economic circumstances and employment outcomes for leavers, Chapter 5 looks at take-up of postexit assistance such as MediCal, Food Stamps, and the Earned Income Tax Credit (EITC), and Chapter 6 looks at other aspects of family well-being, ranging from food and housing insecurity to substance abuse and depression. These chapters focus largely on leavers who remain off aid. In Chapter 7, we turn to the issue of recidivism, looking at the degree to which barriers to employment determine which families return to aid between the two rounds of surveys. Chapter 8 then examines whether we can predict which families will encounter problems after leaving CalWORKs. Finally, Chapter 9 addresses the question of whether the findings for these six counties can be generalized to the rest of the state and suggests some policy implications. 3 2. Data and Methodology This study draws on two distinct types of data. The majority of results depend on findings from in-depth interviews conducted by telephone and in person with respondents from families who had left aid. Wherever possible, we have augmented the survey findings with administrative data available at the county or statewide level. In this chapter, we provide an overview of each type of data and then describe the basic subgroups of leavers considered throughout the analysis. Additional information on the data and methodology is provided in an appendix. Survey Methodology The analysis in this report relies on pooled data from three studies that surveyed CalWORKs leavers in six Bay Area counties: Alameda, Contra Costa, San Mateo, Santa Clara, Santa Cruz, and Sonoma.1 The surveys interviewed random samples of CalWORKs families that left aid in the following time periods: San Mateo, Santa Clara, and Santa Cruz: Contra Costa and Alameda: Sonoma: fourth quarter 1998 third quarter 1999 fourth quarter 1999 For the purposes of these surveys, a family was defined to have left aid if all members of the assistance unit left and remained off aid (in the original county or elsewhere in California) for two months.2 The sample included only cases that contained aided adults. The surveys were ____________ 1These surveys were conducted under the direction of The SPHERE Institute, with the support of the county social services agencies, ASPE, and the William and Flora Hewlett Foundation. See the bibliography for reports from the individual studies. 2It is important to note that none of these families left aid because of time limits. Since the time limit clock for CalWORKs started in January 1998, these families had substantial time left before time limits would be imposed. 5 conducted in English and Spanish; families who did not speak one of these languages were excluded from the sample. The surveys asked respondents about household composition, child well-being, child care, education and training, employment, income, food security, health insurance coverage, family well-being, and welfare experiences. Issues such as income and housing were addressed at the household level. Other questions, such as education and employment, were asked of the respondent, almost always the mother. Finally, questions about child well-being focused on a specific child in the household, chosen randomly to be the focal child. Each of the original three studies included at least two waves of surveys administered over a period of roughly 16 months after exit from CalWORKs.3 Depending on the survey, the timing of the first wave fell between 5 and 10 months after exit; the timing of the second wave fell between 11 and 16 months after exit. Interviews were attempted with all families in the sample in each wave, whether or not they had been successfully interviewed in the previous wave, and relocation efforts were more extensive in the later waves of the survey. For this reason, the number of completed interviews was actually higher in the second wave than in the first. The response rates for the total combined sample were 54 percent for the first set of interviews and 66 percent for the second wave.4 Table 2.1 shows the population of leavers, the original sample, Table 2.1 Population, Sample Sizes, and Survey Response Counts County Alameda Contra Costa San Mateo Santa Clara Santa Cruz Sonoma Total Population 1,082 964 301 1,013 238 309 3,907 Sample 238 304 170 149 184 309 1,354 First Wave 133 203 47 38 66 242 729 Second Wave 169 206 98 84 118 215 890 ____________ 3The San Mateo, Santa Clara, and Santa Cruz studies included three waves of surveys. 4Some of the analyses are limited only to those individuals who completed two rounds of interviews. 6 and the completed sample in each of the six counties. The survey results were all weighted to make the sample representative of the population of leavers from which the samples were drawn, with weights based on age, ethnicity/race, and family composition. Our weights also adjust for differential response to reflect the distribution of the population by county. Administrative Data In the absence of survey data such as those collected for these leavers studies, analysts must rely on administrative data to track outcomes for families who have left aid. The key sources of administrative data on former welfare recipients are the Medi-Cal Eligibility Determination System (MEDS), the Unemployment Insurance Base Wage File (UIBWF), and, and to a lesser extent, county data systems such as the Case Data System (CDS), which is used by our six survey counties but is not a statewide system. MEDS allows basic tracking of program participation in welfare, Medi-Cal, and Food Stamps by individuals and families. The UIBWF captures quarterly earnings for UI-covered employees. Finally, CDS captures data used to administer CalWORKs and other public assistance programs, including much of the same information in MEDS, but also data ranging from budget and grant calculation information to sanctions and exit reasons. Although the surveys used for this report provide a much more detailed portrait of family circumstances than can be deduced from administrative data, the administrative data have a number of advantages over survey data. In particular, these data are available over a long period of time for all program participants at very low cost. In contrast, the surveys are expensive to conduct and capture only a small fraction of participants. For these reasons, administrative data will continue to be a critical source for understanding the effects of welfare and other related programs. In light of this, we compare the survey findings to parallel information from administrative data in the handful of areas where the two overlap. 7 Categories of Leavers The analysis of the survey data cuts the data in two ways: by case type and by leaver status at the time of the survey waves. First, we distinguish between one-parent and two-parent CalWORKs cases. Oneparent cases are the stereotypical welfare cases, with aided single parents, usually mothers, and their children on the case. However, two-parent families are also eligible to receive CalWORKs, under only slightly different eligibility criteria but more rigorous participation requirements.5 At the end of 1999, two-parent families represented one out of every six CalWORKs cases with aided adults.6 Although twoparent families represent a minority of cases, previous research shows different caseload dynamics in two-parent families than in one-parent families. Moreover, not only does CalWORKs have different rules for one-parent and two-parent families, but California also no longer supports two-parent CalWORKs grants out of the federal block grant, because of the high federal participation requirements for two-parent families. For these reasons, policy choices around two-parent families may differ from those for the much larger group of one-parent cases. For each wave of the survey, we examine the circumstances of families off aid at the time of the surveys: 5 to 10 months after leaving aid and 11 to 16 months after leaving aid. By excluding in each period those who have returned to aid, we are focusing on the relatively successful leavers, at least as measured by welfare participation. However, because there is recidivism between these two “snapshots,” the sample of families is somewhat different at the two time periods. In the population of leavers from which our sample was drawn (and to which our sample is weighted), about 12 percent of families still off aid 7 months after exit had returned ____________ 5Two-parent families must have one parent incapacitated or an unemployed principal wage earner, where unemployment is defined as working fewer than 100 hours in the month before application. Two-parent families must participate in work activities (employment, job search, training, or education) for 35 hours per week, compared to 32 hours for one-parent families. 6Over one-fourth of all CalWORKs cases have no aided adults, because the parents are not eligible to receive aid or children are residing with caretaker relatives. These “child-only” cases are not subject to welfare-to-work rules, time limits, or other aspects of welfare reform. For this reason, we consider only cases where an aided adult recipient (and his or her children) has left welfare. 8 to aid by 14 months after exit. This group of “short-term leavers” is therefore included in results on leavers off at 5 to 10 months but excluded from results on leavers still off at 11 to 16 months (the “long-term leavers”), so that each snapshot presents the circumstances of the remaining leavers at the time. (In Chapter 7, we look explicitly at the issue of recidivism and the associated client characteristics.) Table 2.2 provides demographic characteristics for the leaver populations divided both by one- and two-parent families and by the two time periods. Since short-term leavers represent only a small share of all observed cases, the demographic characteristics are relatively unchanged from period to period. On the other hand, the two-parent families do differ demographically from the one-parent families. In the six survey counties, approximately one-third of one-parent families leaving aid were headed by blacks, about one-quarter were headed by Hispanics, and most others were white. In contrast, less than 10 percent of two-parent families leaving aid were headed by blacks, and nearly 40 percent were headed by Hispanics. The two-parent families were also twice as likely to be other ethnicities, predominantly Asian. Spanish speakers were also less common among one-parent families than among two-parent families. Two-parent families had more children on average and were also more likely to have very young children. Finally, two-parent families had fewer previous months on aid. More than half of one-parent families had been on aid for at least three years out of the last five. 9 Table 2.2 Demographic Characteristics of CalWORKs Leavers, by Percentage Share Demographic Characteristics Ethnicity Black Hispanic White Other Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Monthsa Monthsb Monthsa Monthsb 32 31 9 9 26 26 39 39 35 35 37 35 8 8 16 17 Primary language English 94 93 85 86 Spanish 6 7 15 14 Age of case head 16–25 26–35 36–45 46+ 33 31 26 25 36 36 38 39 25 26 29 29 77 8 8 Number of children 0–1 56 55 32 31 2 28 28 33 35 3+ 17 17 35 34 Age of youngest 0–2 3–5 6–11 12+ 32 30 40 40 25 25 24 24 24 25 21 21 19 19 15 15 Months on cash aid in preceding 5 years 1–12 20 19 26 25 13–36 25 25 28 27 37–60 55 56 47 48 Number of cases 2,709 2,396 329 288 NOTE: Columns may not sum to 100 because of rounding. aIncludes short-term and long-term leavers. bIncludes only long-term leavers. 10 3. The Bay Area in Context In assessing outcomes for families who left aid in the surveyed counties, a question that arises is to what degree are their experiences representative of the state. Unfortunately, the very absence of information that motivated the leavers’ studies makes it difficult to conduct in-depth comparisons between different groups of leavers. This chapter places the survey findings in a state context by comparing the Bay Area, where our surveys were conducted, to other regions of the state: Los Angeles County, other Southern California counties, the “Farm Belt” counties on the Central Coast and along the Central Valley, and the North and Mountain counties.1 However, in doing this, we are restricted to a minimal set of outcome measures available from statewide administrative data, notably recidivism rates and use of non-CalWORKs benefits, including take-up of nonassistance Medi-Cal (NAMC) and nonassistance Food Stamps (NAFS).2 The six counties included in our surveys are largely representative of the Bay Area, accounting for 78 percent of the population and 83 percent of CalWORKs cases at the time most of the samples were drawn in 1999. The economic conditions in these counties were also representative of the Bay Area, with an average unemployment rate of 3.1 percent in 1999. The Bay Area, however, was clearly not representative of the state as a whole, experiencing a much stronger economic boom ____________ 1Appendix Table A.5 lists the counties included in each of these regions. MaCurdy, Mancuso, and O’Brien-Strain (2000) provide an analysis of caseload trends across these regions. 2The term “nonassistance” refers to the receipt of these benefits outside the categorical eligibility provided through the CalWORKs program. Former CalWORKs recipients typically qualify for transitional Medi-Cal for up to two years after exit, and their children may continue to be eligible beyond this point. Former recipients, as well as other working poor families, qualify for nonassistance Food Stamps if their income is below 130 percent of the poverty threshold and they meet additional income and asset tests. 11 Percentage returning to CalWORKs than other regions of the state. However, on many outcomes for welfare leavers, the Bay Area was similar to other urban areas of the state, judging by analysis done by Gritz et al. (2001a). Using administrative data, we tracked welfare recidivism among families who left CalWORKs in 1998. As Figures 3.1 and 3.2 show for one-parent and two-parent cases respectively, welfare recidivism in the Bay Area was almost identical to recidivism in Southern California counties, except Los Angeles, which experienced lower caseload declines but also lower recidivism rates. For both one- and two-parent cases, the Bay Area’s recidivism rate fell in the middle of the five regions. CalWORKs recidivism was lowest in Los Angeles and highest in the two most rural regions. The greatest variation occurs among two-parent families, where the recidivism rate one year after exit ranges from about 8 percent in Los Angeles County to just over 20 percent in the North and Mountain region. The regional variation in recidivism in 1998 was greater than seen historically, although the rural North and Mountain and Farm Belt regions have long had higher recidivism rates than the 35 North and Mountain Farm Belt 30 Bay Area Southern California 25 Los Angeles 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.1—Share of Leavers Returning to CalWORKs by Months After Exit from CalWORKs, One-Parent Families Who Left Aid in 1998 12 Percentage returning to CalWORKs 35 North and Mountain Farm Belt 30 Bay Area Southern California 25 Los Angeles 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.2—Share of Leavers Returning to CalWORKs by Months After Exit from CalWORKs, Two-Parent Families Who Left Aid in 1998 urban parts of the state, in large part because of seasonal employment patterns.3 Among the urban regions, one possible explanation for the lower recidivism rate in Los Angeles may also explain the Bay Area leavers’ population. The Bay Area and Southern California (excluding Los Angeles) regions experienced substantially higher percentage caseload declines from the mid-1990s peak through 1998. If one accepts the premise that families with the fewest barriers to self-sufficiency would be the first families to leave aid, then the prolonged caseload decline before 1998 may have led to a Bay Area caseload in 1998 that had a larger share of families with significant barriers to self-sufficiency. Among 1998 leavers, then, the Bay Area leavers (as well as those in Southern California) may have included a segment of the caseload with more barriers than those leaving the caseload in Los Angeles, which could account for the differences in recidivism rates among the three regions. Take-up of noncash assistance after exit in the Bay Area generally fell between that of Los Angeles and other Southern California counties. For ____________ 3See Gritz et al. (2001a). 13 nonassistance Food Stamps, take-up was relatively low in the Bay Area, as seen in Figures 3.3 and 3.4. For both case types, take-up of NAFS was far higher in the rural parts of the state. Leavers in the rural areas may have had lower incomes than those in the Bay Area, because of the poorer economy in the rural regions, and therefore were more likely to be eligible for Food Stamps. Given the seasonal nature of work and welfare participation, leavers in these areas may also tend to be more aware of the availability of nonassistance Food Stamps than those in the Bay Area. For both case types, the Bay Area had the second highest percentage of leavers receiving nonassistance Medi-Cal, with the highest enrollment in Los Angeles County, as shown in Figures 3.5 and 3.6. We can hypothesize that the high take-up of nonassistance Medi-Cal may help explain the low recidivism rates in these regions. Enrollment in nonassistance Medi-Cal was highest in Los Angeles (where recidivism was the lowest) and lowest in the two rural regions (where recidivism was highest). Access to Medi-Cal may have played a role in reducing the incentives for leavers to return to CalWORKs. Percentage receiving nonassistance Food Stamps 40 35 North and Mountain Farm Belt 30 Southern California Bay Area 25 Los Angeles 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.3—Share of Leavers Receiving Nonassistance Food Stamps by Months After Exit, One-Parent Families Who Left Aid in 1998 14 Percentage receiving nonassistance Food Stamps 40 35 North and Mountain Farm Belt 30 Southern California Bay Area 25 Los Angeles 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.4—Share of Leavers Receiving Nonassistance Food Stamps by Months After Exit, Two-Parent Families Who Left Aid in 1998 100 90 80 70 60 50 40 Los Angeles 30 Bay Area 20 Southern California Farm Belt 10 North and Mountain 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.5—Share of Leavers Receiving Nonassistance Medi-Cal by Months After Exit, One-Parent Families Who Left Aid in 1998 Percentage receiving nonassistance Medi-Cal 15 Percentage receiving nonassistance Medi-Cal 100 90 80 70 60 50 40 Los Angeles 30 Bay Area Southern California 20 Farm Belt 10 North and Mountain 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Months after exit Figure 3.6—Share of Leavers Receiving Nonassistance Medi-Cal by Months After Exit, Two-Parent Families Who Left Aid in 1998 Comparison across regions suggests that findings for the Bay Area will be largely consistent with outcomes for other urban areas of the state, especially Southern California counties other than Los Angeles. Where there are differences, it is difficult to predict the direction of the effect. Given that the Bay Area and Southern California had experienced large caseload declines before the 1998–1999 period, it is possible that leavers in our survey period (those who remained after these declines) faced more barriers to self-sufficiency than leavers in Los Angeles, which had lower declines before this period. This hypothesis would fit with the higher recidivism rate in the Bay Area and Southern California, than in Los Angeles. In contrast, the findings for our six counties are unlikely to be representative of outcomes for the less urban areas of the state. 16 4. Economic Outcomes Economic stability is fundamental to the long-term prospects of former welfare recipients; thus, we begin our discussion of the experiences of Bay Area welfare leavers with an analysis of their economic outcomes. We consider three key outcomes: total family income from all sources, earnings, and employment. For each outcome, we look at snapshot portraits of leavers both in the 5 to 10 month and 11 to 16 month periods (including both short- and long-term leavers in the former period), and we compare the change in outcomes for long-term leavers between the two periods. First, we present the findings from our surveys; next, we compare selected findings to available administrative data; and finally, we compare the outcomes for those families who return to welfare to those who stay off welfare but remain poor. Family Income, Earnings, and Employment Monthly income—the financial resources available to families—is one of the most basic indicators of how well leavers are doing. Table 4.1 reports two different measures of family income at our two snapshot points for families remaining off welfare 5 to 10 months after exiting CalWORKs and 11 to 16 months after exit. The distribution of family income is shown in both absolute dollars and relative to the poverty level—a measure that adjusts for differences in family size. Drawing on the survey data, this table—and most subsequent tables in the chapter— reports results for one-parent families and two-parent families separately. By 5 to 10 months after leaving aid, the average family income for leavers hovered around $2,000 per month for both one- and two-parent families. For one-parent families, this income was sufficient to pull the majority above the poverty level, and one in four had income above the cutoff for Medi-Cal eligibility (185 percent of the federal poverty level). For the somewhat larger two-parent families, however, similar incomes in absolute terms translate into lower income relative to the poverty level. 17 Table 4.1 Income Status of CalWORKs Leavers, from Survey Data Income and Poverty Measures Distribution of monthly income, $ Mean Percentile 10 25 50 75 90 Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 1,942 600 1,092 1,700 2,461 3,693 2,411 800 1,200 1798 2,600 3,565 2,082 1,000 1,200 1,635 2,700 3,200 2,275 1,000 1,300 1,970 2,800 4,200 Income as share of poverty threshold, % 185 20 10 7 14 16 19 48 30 17 17 12 14 22 24 13 22 25 29 20 20 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. About 55 percent of two-parent families had income below the poverty threshold 5 to 10 months after leaving aid, compared to 36 percent of one-parent families. In terms of income, the greatest concern may be around the “very poor” families—those with incomes below 70 percent of the poverty level. One in five one-parent families was very poor 5 to 10 months after leaving aid; a smaller share of two-parent families fell into this category. If we look at leavers around a year after they left aid, we see that average income among one-parent families was about 25 percent higher than in the earlier period, at around $2,400. Among families off aid for 11 to 16 months, two-parent families had lower income than one-parent families. The majority of leavers off aid in the 11 to 16 month period had income above poverty for both one-parent and two-parent families, 18 and the share that was very poor fell among one-parent families. On the other hand, in the 11 to 16 month period, 14 percent of two-parent families remaining off aid were very poor. Earnings are the most important source of income for leavers. Around 90 percent of families have at least one adult in the workforce (that is, at least one member reports earnings greater than zero), as shown in Table 4.2. The mean earnings are around $1,800 to $1,900 in the first set of surveys, rising to around $2,200 in the second set. These calculations, as well as the distributions of monthly earnings, include only those families with earnings. Especially for two-parent families, it is clear that earnings make up nearly all of the total family income, particularly for those in the upper half of the income distribution. (In the next chapter, we review participation in public assistance, which makes up much of the other income for these families.) The survey asked the employment status of the respondent, virtually always the mother. For one-parent families, these mothers are the sole earners; more than 60 percent worked full-time and another 15 percent worked part-time (Table 4.3). Not surprisingly, mothers in two-parent families are less likely to be employed, but work was still common among Table 4.2 Family Earnings and Employment of CalWORKs Leavers, from Survey Data Earnings and Employment Measures Percentage working Distribution of monthly earnings, $ Mean Percentile 10 25 50 75 90 Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 88 90 94 93 1,799 600 1,000 1,500 2,200 3,500 2,160 750 1,100 1,600 2,500 3,400 1,947 775 1,200 1,700 2,700 3,200 2,160 50 1,150 1,800 2,750 4,200 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. 19 Table 4.3 Mothers’ Employment Status, by Percentage Share Employment Status One-parent families Not currently working Currently working part-time Currently working full-time Time After Leaving CalWORKs 5–10 Months 11–16 Months 22 23 15 12 63 65 Two-parent families Not currently working Currently working part-time Currently working full-time 49 5 46 41 7 52 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. these mothers, with around half employed full-time. The larger share of two-parent families with earnings is clearly driven by the employment of the fathers in these households. Still, by the 11 to 16 month period, we observe a larger share of mothers working in two-parent families. Are mothers in two-parent families more selective in accepting jobs? Figure 4.1 shows this is not the case; the median hourly wage of employed mothers is higher in one-parent families than in two-parent families. In fact, these findings suggest that mothers who rely on their own earnings to support the family are in higher-paying jobs. Growth in Income and Earnings for Long-Term Leavers Although the snapshot tables presented above provide an important description of the economic characteristics of “active” leavers at two different points after their cohort’s initial leave date, they do not tell us how the circumstances of a constant group of leavers change over time. As mentioned above, differences between the two snapshots are due to both the changing circumstances of long-term leavers and the removal of short-term leavers who returned to CalWORKs. To assess actual growth in earnings and income for leavers between these two periods, we turn to an analysis of long-term leavers only. Table 4.4 presents the net gains or 20 10.50 10.00 One parent Two parents Hourly wage, $ 9.50 9.00 8.50 8.00 7.50 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 4.1—Median Hourly Wage of Working Mothers losses in monthly income and earnings between the two surveys for the same respondents in the 5 to 10 month and 11 to 16 month postexit periods. In fact, income gains between the 5 to 10 month interviews and the 11 to 16 month interviews were very modest for long-term one-parent leavers. Average monthly income grew by only $61, and half of the leavers experienced no income growth at all. Twenty-five percent experienced monthly income losses greater than $114, and 10 percent had losses exceeding $275. An equal share saw their income grow, but the relative magnitude of these income changes was smaller. A slightly more optimistic pattern applies for two-parent leavers. Although half of two-parent leavers still experienced income losses, these losses were less than for one-parent families. Moreover, for those two-parent leavers enjoying gains in income, the gains were typically larger than for oneparent families. Reported changes in earnings were consistent in both direction and magnitude with changes in income for both one- and two-parent leavers. For both groups, a small percentage of families lost their jobs (moving 21 Table 4.4 Growth in Income and Earnings for Long-Term CalWORKs Leavers, from Survey Data Income and Earnings Measures Distribution of gains in monthly income, $ Average growth Percentile 10 25 50 75 90 Gains/Losses 5–16 Months After Leaving CalWORKs One Parent Two Parents 61 –275 –114 0 89 200 71 –75 –33 0 167 398 Distribution of gains in monthly earnings Percentage finding a job Percentage losing a job Average growth, $ Percentile 10 25 50 75 90 4 4 41 –275 –100 0 78 200 1 2 90 –33 –33 0 167 333 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. from positive to zero earnings) whereas other families gained jobs (moving from zero to positive earnings). These percentages were equal (or nearly equal), yielding little change in overall employment between the two periods. Thus, we find that the financial status of leavers remained quite stable after exit from CalWORKs: 75 percent of leavers experienced monthly income variations below $115 in absolute terms, whereas earnings changes were lower than $85 for 75 percent of leavers’ families (combining one-parent and two-parent leavers). The differences that show up in the two snapshots, therefore, are driven primarily by a 22 composition change, where the lowest income families are the most likely to return to aid. We take up the issue of recidivism in Chapter 7. Earnings Reported in Administrative and Survey Data Earnings is one aspect of well-being that we can assess looking at either survey or administrative data. For CalWORKs leavers employed in work covered by unemployment insurance, quarterly earnings are available in the UI Base Wage File. To construct family earnings using only administrative data, we match Social Security numbers for household members included in the assistance unit in the last month of aid to individuals’ earnings reported in the UI Base Wage File. For households with more than one adult on aid, we sum across individuals to get household earnings. Table 4.5 presents the percentage of leaver families with positive earnings and the distribution of their earnings constructed from the UI Base Wage File, in a structure comparable to that of Table 4.2. There are a number of reasons to expect the administrative data on earnings to differ from the survey reports. First, not all workers or all Table 4.5 Family Earnings and Employment of CalWORKs Leavers, from Administrative Data Earnings and Employment Measures Percentage working Distribution of monthly earnings, $ Mean Percentile 10 25 50 75 90 Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 71 75 80 81 1,423 341 809 1,450 1,922 2,556 1,574 367 802 1,491 2,187 2,938 2,392 3,049 886 1,213 2,033 3,087 3,812 1,334 1,986 2,692 3,439 5,840 SOURCE: UI Base Wage File, California Employment Development Division. 23 earnings may appear in the UI file. Self-employed workers are not required to report earnings, and earnings from casual or informal labor may be underreported or unreported. Second, our construction of earnings from the UI files relies on the Social Security numbers of household members at the time the household left aid. For one-parent families in particular, there may be household members (such as unmarried partners) who entered the household after it left aid or who were not included in the assistance unit. The earnings for these household members would not be captured in our administrative data match but are likely to be reported in the survey. Finally, there could be misreporting by the respondent. Relative to the survey, the administrative data vastly understate the share of households with earnings. For both one- and two-parent families, the percentage with earnings in the quarter including the survey month according to the UI Base Wage File is around 15 points lower than the percentage of households reporting positive earnings in our survey. The differences in the mean and distribution of earnings suggest that different factors are skewing the administrative data relative to the survey data for one- and two-parent families. For one-parent leavers, there is a consistent undercount of earnings in the UI Base Wage File compared to earnings in the respondents’ reports, although the medians are fairly similar. We suspect that the higher income reported in the survey is due to additional household members with earnings not included in the assistance unit, although these families may also have earnings that are not reported to the Employment Development Department. For twoparent families, we do not expect much change in the wages of adults included in the household. For these families, the fact that the UI Base Wage File shows higher earnings at every point in the distribution, but fewer families with earnings, suggests that the earnings not captured in the administrative data are disproportionately lower earnings. This would be consistent with a significant share of low-wage work in the informal labor market, which may fit with the larger share of nonEnglish-speaking immigrants in the two-parent caseload. These findings suggest that reliance on data from the UI Base Wage File will give a lower bound on the share of households with earnings. 24 The earnings captured in the administrative data are probably also a lower bound for one-parent families but may show an overly optimistic picture of typical earnings for two-parent families. Barriers to Employment Despite the relatively strong workforce participation among our surveyed CalWORKs leavers, many families experienced barriers to fulltime employment. The surveys asked whether and to what extent respondents viewed child care and transportation as barriers to full-time employment. As Table 4.6 shows, 28 percent of the one-parent respondents and 20 percent of the two-parent respondents in the 5 to 10 months postexit period reported that obtaining child care was a big problem. Among one-parent leavers who were still off welfare 11 to 16 months after exiting CalWORKs, only 20 percent reported child care barriers to employment. In contrast, 39 percent of the long-term twoparent leavers reported child care barriers. The higher reporting of barriers may reflect the larger families of two-parent leavers, or less access Table 4.6 Percentage Reporting Whether Child Care Creates a Problem for Full-Time Work One-parent families Not a problem A small problem A big problem Time After Leaving CalWORKs 5–10 months 11–16 months 56 59 17 21 28 20 Two-parent families Not a problem A small problem A big problem 70 46 10 15 20 39 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. 25 to child care assistance, which is generally available only when both parents are working at the same time. However, it appears that the additional potential earners in the household allow two-parent leavers to remain off aid without solving their child care problems. Transportation was also a barrier to full-time employment for about 16 percent of one-parent families and 12 percent of two-parent families in the 5 to 10 months postexit period, as Figure 4.2 shows. Furthermore, families who remained off aid for another two quarters (until 11 to 16 months) were not less likely to have this problem. The share of two-parent families reporting a transportation barrier was 16 percent among leavers who spent more time off aid and remained virtually constant for one-parent leavers. Percentage reporting transportation as problem 20 One parent Two parents 18 16 14 12 10 8 6 4 2 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 4.2—Transportation as a Barrier to Full-Time Work Summary Overall, the economic status of CalWORKs leavers is mixed. Although most families are not poor, two-parent families are generally less well-off because their incomes do not compensate for additional family members. Leavers who have stayed off CalWORKs for a year or 26 more have higher incomes and earnings than the group that was surveyed at 5 to 10 months, but actual growth in income/earnings for these longterm leavers, if it exists, is moderate. Another encouraging finding is that the work effort of these families is relatively high, especially for single mothers who must support their families alone, despite the child care and transportation barriers that persist for a minority of families. This workforce attachment probably contributes to the fact that these families remain off CalWORKs rather than returning to aid; but, still, onequarter of poor long-term leaver families appear to rely on income other than earnings—an issue we turn to in the next chapter. Our analysis also reveals that policymakers and program administrators must exercise caution when using administrative data to assess the earnings and work history of CalWORKs leavers, because they systematically miss certain types of employment, undercount earnings for one-parent families, and overcount earnings for two-parent families. 27 5. Take-Up of Assistance After Leaving CalWORKs Although they left CalWORKs, many of the families surveyed continued to rely on some form of public assistance or related meanstested programs. In this chapter, we examine patterns of use of public programs by one- and two-parent leavers. As with income and earnings, we look at both point-in-time snapshots of public assistance use at 5 to 10 months and at 11 to 16 months after exit and at changes in participation for the long-term leavers observed in both time periods. We also compare the survey findings to those generated from administrative data. After presenting these participation rates, we use our survey results to explore in more depth familiarity with and take-up of selected programs, including potential reasons for forgoing benefits. Postexit Participation in Public Assistance: Evidence from the Surveys Families who have left CalWORKs may be eligible for a variety of public assistance and related programs, including Food Stamps, MediCal, Supplemental Security Income/State Supplementary Payment (SSI/SSP), EITC, and Women, Infants, and Children (WIC) program vouchers. Eligibility rules differ from program to program, taking into account diverse factors such as income, assets, disability, age, and work activity. For example, families who have recently left CalWORKs are typically eligible for transitional coverage from the Medi-Cal program. Beyond this transitional period, children in households with income below 185 percent of the poverty level may remain eligible for Medi-Cal, depending on their age. Food Stamps has a complex set of tests for eligibility but requires that families have gross income below 130 percent of the poverty level. Low-income individuals who are aged, blind, or 29 disabled may qualify for SSI/SSP, and WIC provides vouchers for milk and other food products for pregnant and nursing women and young children. Finally, working families with children and income below approximately $22,000 (in 1999–2000) are eligible for EITC. Snapshots of Participation After Leaving CalWORKs According to our survey, more than 80 percent of leavers relied on at least one form of public assistance at both the 5 to 10 month and 11 to 16 month period. Table 5.1 summarizes reported participation in public assistance programs including nonassistance Food Stamps, nonassistance Table 5.1 Participation in Programs After Leaving CalWORKs, from Survey Data Income Support Programs Percentage receiving Any public assistancea Nonassistance Food Stamps Nonassistance Medi-Cal SSI EITC WIC vouchers Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 83 82 85 89 11 7 3 12 60 48 58 54 11 11 9 7 34 41 23 26 12 14 11 28 Distribution of Food Stamp recipient benefits, $ Mean Percentile 25 50 75 155 193 245 248 45 134 236 148 170 200 257 210 200 300 257 330 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. aIncludes EITC, Food Stamps, Medi-Cal, WIC, Refugee Assistance, SSI, Foster Child Payments, Workers’ Compensation, and Public Housing Assistance. 30 Medi-Cal, SSI/SSP, EITC, and WIC vouchers.1 Two-parent families were somewhat more likely to rely on some form of public assistance, especially those who remained off CalWORKs for 11 to 16 months. About 1 in 10 one-parent families reported receiving NAFS 5 to 10 months after exiting CalWORKs, and a slightly smaller percentage of leavers in the later period received them. In comparison, only 3 percent of two-parent leaver families reported Food Stamp receipt in the early period, but among long-term leavers in the second period, this figure was four times as high. Although these figures seem low, given the fact that we expect a higher percentage of leavers to have incomes low enough to qualify for NAFS, program participation is often underreported in surveys, and we believe that these results may be affected by that phenomenon. To test this hypothesis and determine the extent of underreporting, we look at administrative data on participation in the next section of this chapter. In subsequent sections, we explore further the issue of eligibility versus take-up for this and other programs. For those who reported receiving Food Stamps in the survey, the monthly value of these benefits is fairly substantial, as shown in the bottom part of Table 5.1. For one-parent families, the average monthly Food Stamp allotment equals $155 in the 5 to 10 month postexit period; and at 11 to 16 months after exit, this average is up to $193 for those who are still off welfare. Not surprisingly given the larger families, the benefits for two-parent families are higher at each point in the distribution. However, long-term leavers in one-parent families had higher benefits in the 11 to 16 month period than the cohort of leavers in the 5 to 10 month period—a result that does not hold for two-parent leavers. Families rely on nonassistance Medi-Cal more frequently than on nonassistance Food Stamps: 60 percent of one-parent families and 58 percent of two-parent families reported health insurance through this program 5 to 10 months after exiting CalWORKs, and the figures for the 11 to 16 month leavers were 48 percent and 54 percent, respectively. ____________ 1The term “nonassistance” for Food Stamps and Medi-Cal is used to distinguish families participating in these programs without participating in CalWORKs, as opposed to families categorically eligible for these programs because they receive CalWORKs. 31 This relatively high rate of Medi-Cal coverage is consistent with the use of transitional Medi-Cal. After Medi-Cal, EITC was the most commonly used program, although participation is low relative to the number of families reporting earnings. Unlike the other programs, EITC is administered by the Internal Revenue Service and provides a maximum refundable tax credit of around $4,000 for working families. In this program as well, participation rates are significantly lower for two-parent families, especially in the 11 to 16 month period when 41 percent of one-parent families received EITC, compared to 26 percent of two-parent families. One- and two-parent leavers participated in the WIC program at approximately the same rate 5 to 10 months after leaving welfare. For those who remained off welfare for 11 to 16 months, however, twoparent families were twice as likely (at 28 percent) as one-parent families to use WIC. Change in Participation by Long-Term Leavers over Time Focusing on long-term leavers only, we develop a richer picture of changes in program participation after leaving CalWORKs. Since families both enter and exit public assistance over time, we disaggregate the net change by calculating program entry and exit between the two surveys for the same respondents. For each assistance program, Table 5.2 presents the percentage of leavers enrolling and the percentage of leavers exiting between the first and second interview. There is, in fact, significant movement in and out of programs between the 5 to 10 month and 11 to 16 month period. Slightly more than one-third of one-parent families enrolled in or took up an additional benefit and a similar percentage left at least one program. Two-parent families, on the other hand, were more likely to move into additional programs (27 percent) than to leave (17 percent). The remainder of the table bolsters the argument that one-parent leavers became slightly less dependent on other forms of public assistance as they spent more time off aid, whereas two-parent leavers become more dependent. EITC was the program with by far the highest net enrollment rate (percentage enrolled minus percentage leaving), whereas Medi-Cal had the lowest net enrollment rate. The table shows that a 32 Table 5.2 Changes in Participation in Programs After Leaving CalWORKs, from Survey Data Percentage enrolling in an additional programa Leaving at least one programa Changes in Participation 5–16 Months After Leaving CalWORKs One Parent Two Parents 36 27 37 17 Percentage enrolling in nonassistance Food Stamps Leaving nonassistance Food Stamps 3 5 1 2 Percentage enrolling in Medi-Cal Leaving Medi-Cal 71 21 5 Percentage enrolling in SSI Leaving SSI 52 29 Percentage started using EITC Stopped using EITC 17 18 11 3 Percentage enrolling in WIC Leaving WIC 68 54 Distribution of gains in nonassistance Food Stamps, $ Average gain Percentile 25 50 75 –2 –5 00 00 00 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. aIncludes EITC, Food Stamps, Medi-Cal, WIC, Refugee Assistance, SSI, Foster Child Payments, Workers’ Compensation, and Public Housing. substantial percentage of one-parent leavers lost Medi-Cal coverage as they spent time off aid—14 percent—whereas the net loss for two-parent families was less, at least according to the surveys. 33 The survey results also show that changes in nonassistance Food Stamps benefits were negligible. Net participation rates decreased only slightly, and the actual benefit allotment did not change at all for at least 75 percent of leavers. This finding suggests that changes in the distribution of monthly benefits amounts observed in the 5 to 10 months and 11 to 16 months postexit periods (Table 5.1) reflect the different composition of the NAFS participants—short-term leavers dropping back into CalWORKs—rather than changes in circumstances for longterm leavers. For both NAFS and Medi-Cal, however, we will reexamine the participation trends using administrative data, below. Snapshots and Participation Trends Using Administrative Data Analyzing patterns in nonassistance Food Stamps and Medi-Cal participation using administrative data both strengthens the findings of the surveys and reveals additional trends that were masked by underreporting in the surveys. As discussed in Chapter 2, MEDS contains participation information for both Medi-Cal and NAFS. Tables 5.3 and 5.4 display the snapshots and trends in participation for these programs, and their figures can be directly compared to those in Tables 5.1 and 5.2. Table 5.3 Participation in Programs After Leaving CalWORKs, from Administrative Data Income Support Programs Percentage receiving nonassistance Food Stamps Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 11 9 7 21 Percentage receiving nonassistance Medi-Cal 71 53 74 73 SOURCE: MEDS. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 34 Table 5.4 Changes in Participation in Programs After Leaving CalWORKs, from Administrative Data Changes in Participation 5–16 Months After Leaving CALWORKs Income Support Progams One Parent Two Parents Percentage enrolling in nonassistance Food Stamps 4 10 Leaving nonassistance Food Stamps 3 2 Percentage enrolling in Medi-Cal Leaving Medi-Cal 24 22 13 SOURCE: MEDS. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. For one-parent families, the NAFS participation rates calculated from the surveys and MEDS are identical in the 5 to 10 months postexit period, and the MEDS rate is only marginally higher in the 11 to 16 months postexit period (Table 5.3). Two-parent families, on the other hand, appear more likely to underreport their participation, especially in the second period. Since we generally accept the MEDS data as correct, the true participation rate for two-parent leavers is 7 percent in the first period and 21 percent in the second, a difference of 4 percentage points and 9 percentage points, respectively, from the survey results. Thus, although participation in NAFS is still low, the administrative data reveal that long-term two-parent leavers are more likely than other groups to receive Food Stamps benefits. Moreover, using MEDS to focus on changes in participation for long-term leavers, we find that two-parent leavers were five times more likely to enroll in NAFS between the two periods than to drop benefits (Table 5.4). This finding is in stark contrast to the virtually nonexistent change in participation reported in the surveys, and it is consistent with other indicators of the increasing dependence of two-parent leavers. In comparison, the net increase in NAFS participation for one-parent long-term leavers was only 1 percent. Increases in NAFS take-up should not be viewed as a negative outcome. In fact, state and county social services agencies did additional outreach 35 to attempt to boost take-up of Food Stamps and other postexit benefits, because these benefits are viewed as an integral part of the safety net for working poor families and the drop in Food Stamps participation was an unintended side-effect of welfare reform. Turning to Medi-Cal, MEDS data reveal that nearly three-quarters of both one- and two-parent leavers had at least one family member enrolled in the program in the 5 to 10 month period after leaving CalWORKs (Table 5.3). This figure is substantially larger than that calculated from the surveys. As in the survey data, however, one-parent leavers in the 11 to 16 month period were less likely to participate in Medi-Cal than were leavers in the earlier period, whereas participation for two-parent leavers was the same in both periods. As for assessing the time trends for long-term leavers, the net loss of Medi-Cal coverage was even greater than the survey suggests. For one-parent leavers, the two data sources agree on the percentage leaving Medi-Cal, but according to MEDS only 2 percent, versus 7 percent in the survey, enrolled in MediCal (Table 5.4). For two-parent leavers, the figures for both enrollees and disenrollees were higher than reported in the surveys, but the net loss was still greater at 9 percent in the MEDS data. Given that we found relatively close agreement between the two data sources in the reporting of CalWORKs and Food Stamps enrollment, the differences in Medi-Cal suggest that many respondents may not have been aware of their enrollment in Medi-Cal. We can hypothesize that at least some of the discrepancy stems from the “Edwards Hold” problem: Medi-Cal coverage is automatically extended to CalWORKs leavers (under what is referred to as the “Edwards” category) pending redetermination of their eligibility to stay on Medi-Cal. During the time of our surveys, coverage under Edwards was extended to many recipients for relatively lengthy periods of time, as counties worked off their backlogs of former CalWORKs cases that needed redetermination under newly adopted eligibility rules.2 It is possible that many of these families were not aware that their Medi-Cal coverage had been extended, thereby contributing to the discrepancy between the administrative and survey ____________ 2For more on the evidence of the role of the Edwards Hold on Medi-Cal participation, see Lieberman and Mancuso (2001). 36 data. Eventual awareness of their coverage may also explain why a higher percentage of long-term leavers, at least one-parent leavers, reported gaining coverage between the two survey periods than were recorded as actual enrollees in the administrative data. Take-Up of Food Stamps and EITC Although our survey results reveal that most CalWORKs leavers participate in at least one income-support program, participation in specific programs is still relatively low. It is especially low for Food Stamps (with only 10 to 20 percent of leavers participating) and EITC— in which approximately 25 to 40 percent participate but we expect nearly all leavers to be income-eligible. On the one hand, there is little reason for concern if families do not participate because their income exceeds eligibility thresholds. On the other hand, policymakers should be concerned if poor families do not receive all the assistance for which they are eligible. Moreover, although some stigma may influence participation in Food Stamps, it should have little effect on EITC participation. To gain a better understanding of the relationship between leavers’ well-being and their participation in these programs, we use survey results to assess take-up of benefits and the reasons for respondents’ nonparticipation. Food Stamps To examine the use of nonassistance Food Stamps, we determined which households appeared to be eligible for Food Stamps, judging by their income, household size, and other applicable factors. Thirty percent of one-parent leaver households were eligible but not receiving Food Stamps, as shown in Table 5.5. More than half of two-parent families appeared to be forgoing Food Stamps. Many of these leaver families had income low enough to remain eligible for CalWORKs. Without the CalWORKs grant, the forgone Food Stamps benefits are often quite high. Our estimates of the benefits that eligible households would have received if they had applied for Food Stamps show that the median value of the forgone benefits in the 5 to 10 month postexit 37 Table 5.5 Forgone Food Stamps Benefits of CalWORKs Leavers Income Support Programs Percentage eligible for but not receiving Food Stamps Potential distribution of monthly benefits , $ Mean Percentile 25 50 75 Time After Leaving CalWORKs One Parent Two Parents 5–10 11–16 5–10 11–16 Months Months Months Months 30 28 54 43 256 214 145 117 230 212 329 305 239 286 175 149 211 262 257 366 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. period was $230 per month for one-parent families and $211 per month for two-parent families. These typical forgone benefits are actually higher than the average benefit reported by those who do participate, probably in large part because benefits received are underreported. By 11 to 16 months after exit, the share of two-parent families eligible but not receiving Food Stamps fell, although the average forgone benefits rose. The share of one-parent families forgoing Food Stamps was relatively unchanged between the two time periods, although the value of forgone benefits fell over time, as earnings rose. Earned Income Tax Credit A large fraction of families never heard of or used EITC, according to survey findings at the two postexit periods and summarized in Table 5.6. Among those families who had heard of EITC, only one out of two had used it (16 percent compared to 32 percent, not shown). Twoparent families were less often aware of EITC than one-parent families. For one-parent families, EITC take-up rates were higher for families off aid at 11–16 months: The share of families unaware of EITC decreased from 49 percent to 42 percent, whereas the share of families who had 38 Table 5.6 Awareness and Use of the Earned Income Tax Credit Among CalWORKs Leavers, by Percentage Share Awareness and Use One-parent families All families Never heard of, never used Heard of, never used Heard of and used Families with positive earnings Never heard of, never used Heard of, never used Heard of and used Time After Leaving CalWORKs 5–10 11–16 Months Months 49 42 18 17 33 41 46 41 17 16 36 43 Two-parent families All families Never heard of, never used Heard of, never used Heard of and used Families with positive earnings Never heard of, never used Heard of, never used Heard of and used 66 66 11 9 23 25 65 66 12 9 23 25 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. used EITC increased from 33 percent to 41 percent. However, little improvement in take-up occurred among two-parent families. Because virtually all two-parent families were working, restricting the analysis to those with at least one working member has no real effect on the results for two-parent families. Awareness of EITC was slightly higher among working one-parent families than nonworking one-parent families. Health Insurance Coverage Table 5.1 reported whether any family member participated in Medi-Cal. However, Medi-Cal eligibility applies to individuals, and 39 different members in the same family may or may not be eligible for Medi-Cal, especially after the transitional benefit period, where children may be eligible but not the adults, or younger children but not older children. Similarly, parents may receive employee health insurance coverage but not dependent coverage. For these reasons, we asked additional questions about health insurance coverage for the respondent and the children in each household. These results are reported in Tables 5.7 and 5.8. Table 5.7 summarizes respondents’ participation in various health insurance programs. Although nonassistance Medi-Cal is the most important source of insurance in the 5 to 10 month postexit period, for long-term leavers in the 11 to 16 month postexit period it is not the most common. For one-parent leavers, the adult uninsured rate is slightly higher for long-term leavers in the second period than for those who were off aid in the first period, but for two-parent leavers the opposite is true. For both family types, over one-third of long-term leavers had Table 5.7 Health Insurance Coverage of CalWORKs Leavers, by Percentage Share Type of Coverage One-parent families Not insured Medi-Cal Medicare Private Time After Leaving CalWORKs 5–10 11–16 Months Months 26 28 43 32 00 31 40 Two-parent families Not insured Medi-Cal Medicare Private 35 29 40 35 00 25 36 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 40 private insurance by the second period. Nevertheless, a significant share of the respondents remained uninsured long after leaving aid. Children are more likely to be insured than their mothers, as shown in Table 5.8. Children in two-parent families are only slightly more likely than children in one-parent families to be uninsured. Private insurance covers about one in four children—a rate lower than that for respondents. For long-term leavers in the second period, children in two-parent families had lower rates of private insurance and higher rates of Medi-Cal coverage than children in one-parent families or the more comprehensive group of leavers surveyed in the first period. Healthy Families (California’s health insurance program for children in lowincome families) was just beginning in this period, and coverage through Healthy Families was rarely reported for any of these children. One reason for the lack of insurance among children was that many who were eligible for Medi-Cal were not enrolled. According to the survey, only 62 to 64 percent of children eligible for Medi-Cal were Table 5.8 Health Insurance Coverage of Children of CalWORKs Leavers, by Percentage Share Type of Coverage One-parent families Not insured Medi-Cal Healthy Families Private/other government Time After Leaving CalWORKs 5–10 11–16 Months Months 16 18 53 51 21 28 30 Two-parent families Not insured Medi-Cal Healthy Families Private/other government 17 20 57 63 00 27 16 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Columns may not sum to 100 because of rounding. 41 enrolled 5 to 10 months after exiting CalWORKs (Table 5.9). Some of the children who were not enrolled in Medi-Cal were insured by other sources, including private health insurance, Healthy Families, and other government programs, but about 10 percent of children who were eligible for Medi-Cal remained uninsured. If parents sought acute or emergency medical care for these children, they would typically be enrolled in Medi-Cal or Healthy Families, given their income eligibility; the larger concern is if parents forgo important preventive care because they do not realize that the children qualify for these programs. In the first period, there was little difference in coverage patterns between oneand two-parent families. In the second period, however, the share of uninsured children who were eligible among one-parent families was 11 percent, whereas the share of uninsured children was only 6 percent for two-parent families remaining off welfare. In addition, children in twoparent families were more likely to be covered by nonassistance Medi-Cal Table 5.9 Health Insurance Coverage of CalWORKs Leavers’ Children Eligible for Medi-Cal, by Percentage Share Type of Coverage One-parent families Not insured Medi-Cal Healthy Families Private/other government Time After Leaving CalWORKs 5–10 11–16 Months Months 9 11 64 57 30 25 32 Two-parent families Not insured Medi-Cal Healthy Families Private/other government 10 6 62 79 00 28 15 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. 42 and less likely to have private insurance than were children in one-parent families. We might expect the uninsured respondents to have income too high to qualify for nonassistance Medi-Cal. Table 5.10 shows the percentage of uninsured respondents grouped by their household income relative to the federal poverty level. A substantial percentage of these uninsured leavers, especially two-parent families, had incomes below the poverty level, indicating that most were probably eligible for Medi-Cal coverage. The percentage was lower in the 11 to 16 month postexit period but still represented about one-third of the uninsured one-parent respondents and a larger share of two-parent respondents. Table 5.11 illustrates the role of employment in providing health coverage. About 45 percent of single mothers did not receive any coverage from their employers. This figure is higher among two-parent families—more than 50 percent of working mothers in two-parent Table 5.10 Household Income Relative to Federal Poverty Level of Uninsured Respondents, by Percentage Share Income as a Percentage of Poverty Level One-parent families 185 Time After Leaving CalWORKs 5–10 11–16 Months Months 30 13 12 16 11 21 19 18 28 32 Two-parent families 185 11 12 62 30 10 13 8 19 9 26 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 43 Table 5.11 Type of Health Insurance Coverage Offered by Employers of CalWORKs Leavers, by Percentage Share Type of Health Coverage One-parent families None Self only Family Time After Leaving CalWORKs 5–10 11–16 Months Months 45 42 55 50 52 Two-parent families None Self only Family 51 63 18 6 31 31 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. families did not receive any health coverage from their employers. Further, the share of working mothers in two-parent families not covered by their employers is higher in the 11 to 16 month period. Not surprisingly, mothers in two-parent families and their children are more likely to be covered by their partners. Summary Use of postexit assistance is common among the CalWORKs leavers we surveyed, although the rate of take-up for eligible families is well below 100 percent for the programs we tested. Participation is highest in nonassistance Medi-Cal, although some respondents are unaware that they have coverage. However, participation in this and most other programs declines over time, at least for one-parent families. Two-parent long-term leaver families are more likely to use assistance in general, and their participation increases in the later period—a finding that is 44 consistent with the fact that many struggle to raise their incomes above the poverty threshold (Chapter 4). Administrative data (when available) are the most accurate sources of information on program participation. However, these data cannot shed light on CalWORKs leavers who are eligible for assistance but do not use it. This disparity is particularly evident with respect to nonassistance Food Stamps: Although program participants collect rather substantial benefits (according to the surveys), over half of those who are eligible do not participate. The EITC also appears to be underutilized, according to the surveys, although the respondent may be unaware that her family is receiving the credit if she is not responsible for preparing the household’s taxes. Despite extensive Medi-Cal participation, a fraction of respondents and to a lesser extent their children are still uninsured, with many lacking employer-based coverage. Taken together, these findings indicate that although nonwelfare public assistance programs are important to CalWORKs leavers, many get by without them, either by choice or because of a lack of information. 45 6. Well-Being Measures The surveys allow us to capture a richer picture of the well-being of CalWORKs leavers, beyond just the economic measures. In this chapter, we highlight housing conditions, child care usage, and selected risk behaviors. Using our survey data, we present snapshot portraits of leavers after 5 to 10 and 11 to 16 months. We conclude this chapter with a discussion of the difference between welfare recidivists and poor longterm leavers. Housing Conditions Housing costs can be a real challenge for former CalWORKs families, as we see in Table 6.1. Common problems include substandard housing, excessive rent burden, and crowding.1 More than one-third of two-parent families face excessive rent burden and crowding. These problems are less common among one-parent families, but one-parent families are more likely to live in substandard housing. Housing assistance, either as public housing or housing vouchers, can help offset housing problems but is relatively rare, especially for the two-parent families. Twenty-eight percent of one-parent families but only 10 percent of two-parent families received assistance with housing. These results are consistent with prior information about these two types of families and relevant public policies. Whereas two-parent families have difficulty finding space for their relatively larger families but also allocate more of their monthly income to rent, one-parent families have more success securing housing subsidies (historically targeted to poor single____________ 1Substandard housing means that one or more of the following conditions are found: (1) a leaky roof or ceiling, (2) a toilet, water heater, or other plumbing that does not work, or (3) rats, mice, roaches, or other pests. House crowding exists if the ratio of household members to rooms (excluding bathrooms) is greater than one. Rent is excessive when it exceeds 50 percent of total household income. 47 Table 6.1 Housing Outcomes of CalWORKs Leavers, by Percentage Share Housing Outcomes One-parent families Substandard housing Excessive rent burden Crowding Housing assistance Time After Leaving CalWORKs 5–10 11–16 Months Months 27 14 20 19 23 17 28 29 Two-parent families Substandard housing Excessive rent burden Crowding Housing assistance 15 28 36 33 36 48 10 23 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTES: Column differences that are statistically significant at the 95 percent level are indicated in bold. Columns may not sum to 100 because of rounding. parent households) but this subsidized or other low-rent housing may be of lower quality. Comparing across the two time periods, we find that two-parent leavers who remain off CalWORKs for 11 to 16 months faced more housing problems than did two-parent families in the earlier survey period (which includes both long-term leavers and families who soon return to welfare). For example, 28 percent of two-parent families who had been off aid for 11 to 16 months lived in substandard housing, compared to 15 percent in the 5 to 10 month postexit period. Nearly half of two-parent families lived in crowded housing in the later period. However, those who remained off in the 11 to 16 month period were more likely to receive housing assistance: 23 percent compared to 10 percent in the earlier period. The high incidence of housing problems among long-term two-parent leavers is troubling, especially since compositional change between the two survey periods means that some of the 5 to 10 month leavers returned to welfare. Reported housing 48 problems for one-parent families were similar in the two survey periods, except that substandard housing was less prevalent. Child Care As discussed in Chapter 4, a substantial proportion of respondents indicated that child care was a barrier to obtaining full-time employment. Before policymakers can address this problem, however, it is important to know more about the types of child care welfare leavers use and how they are paying for it. Table 6.2 shows the distribution of various types of primary child care arrangements used by the respondents. Relatives were the most common child care providers in two-parent families, with about 72 percent of two-parent families relying on relatives for child care in the 5 to 10 month postexit period. Relatives were also the most frequent child care resource used by one-parent Table 6.2 Primary Child Care Arrangements of CalWORKs Leavers, by Percentage Share Child Care Arrangement One-parent families Head Start Formal day care Extended day care Adult relative Family day care Adult nonrelative Time After Leaving CalWORKs 5–10 11–16 Months Months 42 24 21 10 11 40 45 17 15 56 Two-parent families Head Start Formal day care Extended day care Adult relative Family day care Adult nonrelative 00 8 17 8 15 72 50 11 11 17 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 49 families, but other forms of nonrelative care were important as well. For example, in the 5 to 10 month time frame, 24 percent of one-parent families relied on formal day care (such as a child care center, preschool, or nursery school) and an additional 17 percent used family day care homes. In the later postexit period, two-parent families who remained off CalWORKs were less likely to use relative care and more likely to use formal and other nonrelative care than two-parent families in the earlier period. The distribution of providers for one-parent families was similar in both periods. One-third of one-parent families incurred out-of-pocket expenses for child care during the 5 to 10 month period (Figure 6.1). Two-parent families were less likely to incur these expenses (20 percent), consistent with the fact that two-parent families were more likely to use an adult relative for care, most of whom do not require payment. In the second period, when fewer two-parent leavers used relative care, they were more likely than in the earlier period to pay for child care (30 percent). However, the share with out-of-pocket expenses also increased for oneparent leavers, even though more one-parent families relied on adult relatives for care in the later period. Percentage paying out-of-pocket child care expenses 50 45 One parent Two parents 40 35 30 25 20 15 10 5 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.1—Out-of-Pocket Child Care Expenses 50 Working CalWORKs leavers qualify for subsidized child care, with payments made to the provider of their choice. However, as Figure 6.2 shows, a little over one-quarter of one-parent respondents in the 5 to 10 month postexit period indicated that they were not aware of government subsidies that would help pay for child care. Lack of awareness was more prevalent among two-parent families—34 percent did not know about subsidies. The proportion of two-parent families that were unaware of child care subsidies was even higher in the second postexit period, at nearly 45 percent, and we can assume that this result reflects the changing composition of leavers between the two snapshot surveys. This figure suggests that long-term leavers are more likely to be unaware of child care subsidies than short-term leavers, at least in two-parent families. In addition, some school-age children were not receiving care who could potentially need it. Table 6.3 shows the extent to which the respondents indicated that their children (ages 5 to 13) were left Percentage unaware of child care subsidies 50 45 One parent Two parents 40 35 30 25 20 15 10 5 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.2—Awareness of Child Care Subsidies 51 Table 6.3 Percentage of Time Child (Ages 5 to 13) in CalWORKs Leaver Family Was Left Unsupervised in Prior Month Hours Unsupervised per Month One-parent families 0 1–19 20+ Time After Leaving CalWORKs 5–10 11–16 Months Months 87 82 69 79 Two-parent families 0 1–19 20+ 95 94 45 11 SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. unsupervised during the preceding month. Few two-parent families— but nearly 10 percent of one-parent families—indicated that this occurred 20 or more hours during that month. Child and Respondent Well-Being Our surveys included a variety of questions related to different aspects of the well-being of families in general and children in particular. Figure 6.3 shows the proportion of respondents reporting that the focal child in the family had engaged in risk behaviors in the preceding six months.2 Risk behaviors include being suspended or expelled from school, getting into trouble with police, having a problem with alcohol or drugs, doing something illegal to get money, dropping out of school, and getting pregnant or getting someone else pregnant. Over 12 percent of the one-parent respondents indicated that the child had engaged in such behaviors. In comparison, children in two-parent families were half as ____________ 2Focal children were randomly selected from all of the respondent’s children. 52 Percentage reporting child risk behaviors 20 18 One parent Two parents 16 14 12 10 8 6 4 2 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.3—Child Risk Behaviors likely to engage in such behaviors. The percentage of families reporting problem behaviors was essentially the same in both survey periods. Figure 6.4 shows the percentage of respondents who indicated that they or any other adults in their household had substance abuse problems. A household has a substance abuse problem if, according to the survey, people complained about the respondent’s use of alcohol or drugs, or they were having problems because of their alcohol or drug use, or any other adult in the household had a problem with alcohol or drugs. Similar to the pattern for child risk behaviors and supervision, one-parent families were more likely to report these problems. Contrary to child risk, however, longer-term leavers in the 11 to 16 month postexit period reported notably less substance abuse than CalWORKs leavers in the 5 to 10 month period. Figure 6.5 shows the percentage of respondents indicating the presence of domestic violence (physical or emotional abuse) in their household. In the first period, about 15 percent of one-parent respondents reported such occurrences, compared to 11 percent for two-parent families. In the second period, abuse affected 23 percent of one-parent families, but the 53 Percentage reporting substance abuse 16 14 One parent Two parents 12 10 8 6 4 2 0 5–10 11–15 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.4—Substance Abuse Reported by Respondents 24 22 One parent Two parents 20 18 16 14 12 10 8 6 4 2 0 5–10 11–16 Months after exit SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. Figure 6.5—Domestic Violence Reported by Respondents Percentage reporting domestic violence 54 abuse rate for two-parent families was basically unchanged. These findings may reflect the volatility of relationships and living arrangements for single parents. Summary Selected measures of the general well-being of CalWORKs leaver families indicate that challenges persist well after exiting cash aid. Specific problems with housing are different for one- and two-parent families, but they are consistent with other differences between the two populations; in particular, they provide more evidence that two-parent families are straining to get by even as they refrain from returning to welfare. The fraction of school-aged children of single parents who are left home alone is less than ideal, especially since one-quarter of single parents are unaware of child care subsidies for which they are most likely eligible. Risky behaviors by both children and adults affected a minority of families. These behaviors were more common among single-parent families, but we do not know how the incidence of these behaviors compares to that in the general population. 55 7. What Distinguishes LongTerm Leavers from Recidivist Leavers? Comparing the experiences of leavers who successfully remain off CalWORKs to those who return offers insight into the principal factors and trigger events that contribute to recidivism. To examine recidivism, we use administrative data for the entire leavers population in our six counties, considering the differences in demographic characteristics for short-term and long-term leavers and comparing these two groups to those who had returned to aid before the first wave of interviews. We then turn to survey data to look in more depth at whether factors identified in the 5 to 10 month interviews help predict which families will have returned to aid by the 11 to 16 month period. These factors help identify potential avenues for targeting policy resources to assist families in staying free of welfare. Demographic Characteristics of Recidivists To assess whether certain demographic characteristics are associated with recidivism, we divide the population of leavers into three categories—long-term, short-term, and temporary leavers. Long-term leavers are those who had not returned to CalWORKs by the time of our second wave of interviews—the group captured in our 11 to 16 month findings in previous chapters. In the administrative data, we define longterm leavers as those still on aid 14 months after exit. Our 5 to 10 month findings showed a snapshot that combined long-term leavers with those who left aid between the two survey waves. In the administrative data, these short-term leavers are defined as those who remained on aid at 7 months after exit but had returned to aid by 14 months. Finally, those families who left aid for 2 months but who had already returned by 57 month 7 (a group not included in our previous analysis) are defined as temporary leavers. For these three groups of leavers, Tables 7.1 and 7.2 summarize the demographic characteristics of one- and two-parent families, respectively. Using administrative data, these two tables show Table 7.1 Demographic Characteristics of One-Parent CalWORKs Leavers, by Percentage Share Demographic Characteristics Ethnicity Black Hispanic White Other All Leavers 34 26 33 8 Temporary Short-Term Long-Term Leavers Leavers Leavers 41 37 26 29 26 28 75 31 26 35 8 Primary language English Spanish 94 97 63 95 5 93 7 Age of respondent 16–25 26–35 36–45 46+ 34 38 36 35 24 23 64 43 36 17 4 31 36 26 7 Number of children 0–1 2 3+ 54 51 28 30 17 19 60 27 13 55 28 17 Age of youngest 0–2 3–5 6–11 12+ 33 39 26 26 24 24 17 11 41 27 19 13 30 25 25 19 Months on cash aid in preceding 5 years 1–12 13–36 37–60 18 25 57 14 24 62 19 28 53 19 25 56 Number of cases 3,473 764 313 NOTE: Columns may not sum to 100 because of rounding. 2,396 58 Table 7.2 Demographic Characteristics of Two-Parent CalWORKs Leavers, by Percentage Share Demographic Characteristics Ethnicity Black Hispanic White Other All Leavers 11 37 39 14 Temporary Short-Term Long-Term Leavers Leavers Leavers 18 10 9 30 34 39 44 49 35 8 7 17 Primary language English Spanish 87 92 80 86 13 8 20 14 Age of respondent 16–25 26–35 36–45 46+ 29 37 34 25 35 28 32 39 29 29 27 29 76 7 8 Number of children 0–1 28 17 41 31 2 32 28 20 35 3+ 39 54 39 34 Age of youngest 0–2 3–5 6–11 12+ 43 54 41 40 22 17 22 24 21 20 22 21 14 9 15 15 Months on cash aid in preceding 5 years 1–12 24 17 32 25 13–36 26 19 29 27 37–60 51 64 39 48 Number of cases 432 103 41 NOTE: Columns may not sum to 100 because of rounding. 288 the entire population of leavers (not just those included in the surveys), listing their family characteristics at the time they left aid. Sixty-nine percent (2,396 out of 3,473) of one-parent families and 66 percent (288 of 432) of two-parent families remained off aid for at 59 least 14 months. As we saw above, one-parent families who leave CalWORKs have fewer children than their two-parent counterparts and are more likely to be black, less likely to be Hispanic, and less likely to have Spanish as their primary language. In looking at the relationship of these and other demographic characteristics to the probability of remaining off aid, we see that for both one- and two-parent families the number and age of children and ethnicity are linked to the probability of remaining off aid. Larger families are less likely to remain off CalWORKs. This factor is especially pronounced for the two-parent cases, where families with three or more children make up 54 percent of two-parent temporary leavers and only 34 percent of two-parent longterm leavers. Families with older children have a higher probability of remaining off CalWORKs. For example, one-parent families whose youngest child is older than age 12 constitute only 11 percent of temporary leavers but about 19 percent of long-term leavers. In contrast, families with an infant (ages 0 to 2) present constitute 41 percent of temporary leavers (54 percent of two-parent and 39 percent of oneparent families) but only 31 percent of long-term leavers (40 percent of two-parent and 30 percent of one-parent). Black families are more likely than white families to return to CalWORKs, in both one-parent and two-parent families. Whereas blacks make up 39 percent of temporary leavers, combining one- and two-parent families, they make up only 28 percent of long-term leavers. However, two-parent families of Hispanic descent are more likely than their white counterparts to succeed in staying off CalWORKs. Hispanic two-parent families constitute only 30 percent of temporary leavers and make up 39 percent of long-term leavers. Overall, there is little relationship between success in staying off CalWORKs and having Spanish as the family’s primary language. For two-parent families, however, Spanish-speaking families represent almost double the share of long-term leavers than temporary leavers—stronger evidence that speaking Spanish did not constitute a barrier to remaining off CalWORKs. 60 Short-Term Leavers: Income and Work Effort Table 7.3 depicts the income and earnings status of short-term and long-term leavers during the 5 to 10 month period after last participating in CalWORKs. Although structured much like the tables in Chapter 4, here we separate out the short- and long-term leavers within the 5 to 10 Table 7.3 Income Status of CalWORKs Leavers, One-Parent and Two-Parent Families Combined Income and Earnings Measures Poverty status, % 185 Data Source Survey 5–10 Months After Leaving CalWORKs Short-Term Long-Term Leavers Leavers 42 16 10 22 16 17 23 20 9 25 Distribution of monthly income, $ Mean Percentile 10 25 50 75 90 Survey 1,383 509 665 1,143 2,000 2,993 2,012 700 1,200 1,717 2,546 3,900 Distribution of monthly earnings, $ Percentage working Mean Percentile 10 25 50 75 90 Survey 81 965 200 200 800 1,500 2,200 89 1,888 775 1,200 1,600 2,400 3,500 NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. 61 month snapshot. At this first interview period, short-term leavers are distinctly poorer than their long-term counterparts. Over half of the short-term families have monthly incomes below the poverty level, with 42 percent classified as very poor, at less than 70 percent of poverty. This is not to say that poverty was uncommon among the long-term leavers. Although only 16 percent fell into the very poor income category, almost 40 percent still had incomes below the poverty level. On the other hand, 25 percent of long-term leavers had incomes exceeding 185 percent of the poverty level, compared to 9 percent of the short-term leavers. The distributions of income reveal that short-term leavers had uniformly lower monthly incomes than families who would remain off CalWORKs in the following months, with the average income for the short-term leavers reaching only about two-thirds that of the long-term leavers. Not surprisingly, most of this income differential reflects a shortfall in earnings for those families who would return to welfare in the next several months. Although their likelihood of working was similar to that of families who succeed in staying off welfare well beyond this period, they worked less and earned less. One might suspect that a major portion of this shortfall in earnings came from lower hourly wages, but Figure 7.1, which illustrates the wages of short-term and long-term 12 Short-term leavers Long-term leavers 10 Median hourly wage, $ 8 6 4 2 0 All One parent Two parents Figure 7.1—Median Wage at 5–10 Months After Exit 62 leavers, reveals trivial differences in wage rates. So, the principal reason for lower earnings can be attributed to fewer hours worked for families who would become recidivist. These findings support the view that leavers’ inability to maintain satisfactory economic conditions constituted an important factor motivating them to return to welfare. Short-term leavers clearly did worse as a group from an economic perspective than those families who succeeded in staying off welfare. On the other hand, many who did return were better off than some long-term leavers who did not, a topic discussed further below. Short-Term Leavers: Public Assistance To what extent were short-term leavers more dependent on public assistance programs before their reentry into CalWORKs than long-term leavers who did not return? Table 7.4 addresses this question, summarizing participation in income support programs available for families leaving CalWORKs, such as nonassistance Food Stamps and Medi-Cal, as well as alternative programs we considered in Chapter 4. For purposes of comparison, Table 7.4 shows Food Stamps and Medi-Cal receipt as recorded by both the surveys and MEDS. As with Table 7.3, Table 7.4 breaks down the findings for the 5 to 10 month period to distinguish between the short- and long-term leavers Most families relied on some form of public assistance after leaving CalWORKs, with nonassistance Medi-Cal being the most common form of involvement. However, Table 7.4 shows that short-term leavers had higher participation in all programs (except for child support). According to the survey findings, they were about four times more likely to collect Food Stamps and WIC, 6 percentage points more likely than long-term leavers to obtain EITC, and 34 percentage points more likely to have relied on Medi-Cal for health coverage. It is interesting to note that the administrative and survey data on Medi-Cal use provide almost identical results for short-term leavers, but many fewer of those leavers who would become long-term leavers reported Medi-Cal coverage in the earlier period, even though the administrative data show them as eligible. Of course, short-term leavers’ heavier reliance on public assistance in part merely reflected their weaker economic position. Also, consistent with 63 Table 7.4 Participation in Programs After Leaving CalWORKs, One-Parent and Two-Parent Families Combined Programs Percentage receiving Any public assistancea Nonassistance Food Stamps Data Source Survey MEDS Survey 5–10 Months After Leaving CalWORKs Short-Term Long-Term Leavers Leavers 93 83 33 8 34 7 Distribution of monthly support, $ Mean Percentile 25 75 195 145 200 45 200 200 Percentage receiving Nonassistance Medi-Cal SSI EITC WIC vouchers Child support MEDS Survey Surveyb Survey Survey Survey 80 82 22 37 37 16 70 58 10 31 10 19 Distribution of monthly support, $ Mean Percentile 25 75 224 356 100 125 214 475 NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. aIncludes EITC, Food Stamps, Medi-Cal, WIC, Refugee Assistance, SSI, Foster Child Payments, Workers’ Compensation, and Public Housing Assistance. bStatistics in this row are computed without survey weights. We do this to avoid overemphasis on several observations from a county wherein all observations happen to be on SSI. The difference between the participation estimates in this row is not statistically significant at conventional levels of confidence. popular beliefs, long-term leavers had greater amounts of child support income that offered more cushion for remaining off welfare. 64 Barriers to Employment and Knowledge of Assistance Given that recidivism is linked to weaker performance in employment activities, the natural question is whether there were potential barriers or other factors making work a less attractive option for short-term leavers. Chapter 4 showed that many respondents reported child care and transportation as barriers to full-time work. It also showed that a significant percentage of leavers were unaware of employment assistance programs such as EITC and child care subsidies. To quantify the effects of barriers to employment and lack of knowledge of employment-related programs on recidivism, we exploit multivariate probabilistic models. This approach tests the hypothesis that various barriers reported in the early period increased the probability that a family would return to CalWORKs in the next six months. The estimation results are summarized in Table 7.5. The first and second columns in the table present the coefficients and standard errors of the probabilistic model; the last column shows the effect of each factor on the probability that a family will return to CalWORKS shortly. Employment by either the respondent or any other member in the household has the highest effect on recidivism. The families of respondents who were employed at the time of the interview were 14 percentage points less likely to return to CalWORKS; an additional member employed reduces the probability by another 20 percent. Families that perceived child care as a barrier to full-time employment also had an increased likelihood of returning to aid between the two interview periods. However, paying out-of-pocket expenses for child care, reporting transportation barriers, or being unaware of the EITC was not predictive of recidivism. What Distinguishes Poor Long-Term Leavers from Recidivist Leavers? Although economic circumstances go a long way toward explaining which families return to CalWORKs, our analysis of income and earnings patterns reveals that a significant number of nonrecidivist leavers were relatively poor. Even in the 11 to 16 month postexit period, when 65 Table 7.5 Logit Estimates on the Probability That Leavers Will Return to CalWORKs 6–12 Months Postexit Variable Respondent is employed An additional member in the household is employed Child care is a barrier to full-time employment Coefficient –2.11** Standard Error 0.4008 Marginal Effect –0.14** –1.41** 0.3801 –0.20** 0.70* 0.3896 0.09* Family pays out-of-pocket child care expenses 0.52 0.4025 0.07 Job travel is a barrier to full-time employment 0.24 0.3603 0.03 Household is unaware of EITC 0.24 0.3474 0.03 Constant –1.70** 0.4314 Number of observations 557 Likelihood ratio 36.5 NOTES: Sample includes all one-parent and two-parent respondents in the first interview (5 to 10 months) who had not returned to CalWORKs at the time of the interview. Recidivists are those families that returned to CalWORKs sometime between their first interview and one year after exit. *Statistically significant at 90 percent. **Statistically significant at 95 percent. incomes had increased, 32 percent of all nonrecidivist families had household incomes below the federal poverty level, and 11 percent were below 70 percent of the poverty level. Most of these families were probably eligible to return to CalWORKs but had chosen not to do so. In an effort to better understand why some low-income leavers remained off aid while others returned, in Table 7.6 we compare several dimensions of income and well-being for poor long-term leavers (below 100 percent of the federal poverty guideline) and for families who were recidivist at the time of the interview in the 11 to 16 month period. This 66 Table 7.6 Selected Circumstances of Recidivists and Poor Long-Term CalWORKs Leavers 11–16 Months Postexit Household Characteristics Households with earnings Median household earnings, $ Respondent’s work activity Full-time Part-time Not working Barriers to full-time employment Child care Transportation Domestic violence Children uninsured Unstable child care Household substance abuse Depressed 3+ days weekly Recidivist Families 61 1,300 Long-Term Leavers’ Household Income Below 100% Poverty 76 900 22 45 14 12 64 43 27 40 22 22 30 26 4 19 28 23 97 41 34 NOTE: Column differences that are statistically significant at the 95 percent level are indicated in bold. period corresponds to a time when long-term leavers have successfully remained off CalWORKs for one year or more and recidivist leavers have reentered the CalWORKs program for added income support. Recidivists include members of both the short-term and temporary leavers groups who may have reentered CalWORKs more than once and for extended periods prior to the 11 to 16 month interview. Table 7.6 reveals that a higher percentage of the poor long-term leavers had household earnings, 76 percent, compared to 61 percent of recidivist families. Moreover, a higher percentage of poor long-term leaver respondents were working and were doing so at full-time jobs (45 percent compared to 22 percent of recidivists). In light of this observation, it is perhaps surprising to see that the recidivist leavers had higher median household earnings. Although the low income of the long-term leavers shown in Table 7.6 is largely due to the subgroup we have selected, the income of the recidivist households suggests that 67 household members other than the respondent (that is, other than the case head who has returned to aid) are contributing to the household earnings. Thus, the respondent’s work activities may be one clue as to why some leavers chose to return to aid while others remained off aid. We also see some differences in the measures of family well-being, but again there is no clear pattern. The recidivists were only marginally more likely to indicate that they had been experiencing domestic violence, unstable child care, substance abuse, and depression. On the other hand, we also see that the recidivists were much less likely to report that child care was a barrier to employment and were no more likely to indicate that transportation was a problem. Summary Many of the lessons conveyed by the above findings merely confirm conventional wisdom: Weak employment experiences constitute an important determinant of whether families return to CalWORKs, and families who succeed off aid have greater amounts of child support income than families who return to aid. Recidivists face bleaker outcomes than long-term leavers long before returning to CalWORKS: At the 5 to 10 month period, recidivists are more likely to rely on other forms of public assistance and are also more likely to face poverty. Still, many families who remain off aid for periods longer than a year look very much like families who go back to CalWORKs, and earnings are far from the determining factor for whether a family returns to cash assistance. A large segment of the long-term leavers are quite poor and face problems and barriers similar to those of families who return. 68 8. Which Families Encounter More Problems When Leaving CalWORKs? In Chapters 3 to 5, we found that families frequently encounter problems well after leaving CalWORKs. For example, in the period 11 to 16 months after exit, 29 percent of one-parent and 44 percent of twoparent families had income below poverty, 23 percent of mothers in oneparent families were unemployed, 18 to 20 percent of families had uninsured children, and 48 percent of two-parent families lived in crowded housing. In this chapter, we explore the extent to which program administrators can use known characteristics to predict the problems a family is likely to encounter after leaving aid. In other words, how are a family’s characteristics at exit associated with postexit measures of well-being? Identifying those characteristics will help policymakers design targeted programs to assist families leaving aid. Methodology For this analysis, we examine all families who have not returned to CalWORKs by the 11 to 16 month postexit period, trying to determine which nonrecidivists nevertheless face poor outcomes. We use the following five outcome indicators: • Crowded housing, • Income below 100 percent of the federal poverty guideline, • Not employed (no earnings in household), • No health insurance coverage for respondent or child, and • Child care is a barrier to full-time employment. In addition to examining each of these outcomes separately, we examine the probability of having at least one of these problems and of having at least two of these problems. 69 We exploit multivariate probabilistic models to quantify factors contributing to each of these outcomes. These models control for the effects of other variables when measuring the effect of a key variable on the probability that an outcome will occur. For example, we would like to know if a Hispanic family is exposed to unemployment more frequently than a black family. Using multivariate probabilistic models, we can estimate the effect of ethnicity on the probability that a family will be unemployed, after controlling for other variables known to affect employment such as number of children. Results are summarized in Tables 8.1 (for one-parent families) and 8.2 (for two-parent families). The columns of the tables present the results for each outcome, based on estimates from probabilistic models run for each outcome. The rows of the tables list the family characteristics considered in the analysis. These characteristics are grouped into seven categories: 1. Primary language spoken by the respondent, 2. Ethnicity of the head of the assistance unit, 3. Time on aid = number of months on aid in the previous five years, 4. Number of children, 5. Age of the youngest child when a family left CalWORKs, 6. Earnings at exit = earnings for the assistance unit in the quarter preceding exit according to the UI Wage Base File, and 7. Exit reason, according to CDS files. For each characteristic category, families are divided into mutually exclusive groups. These seven characteristics, by design, contain information that would be available to program administrators from administrative databases and could be used to target CalWORKs recipients who are most at risk of having problems after they leave the program. Table entries should be examined by category, relative to the reference group for each category. For example, under “Ethnicity,” the selected reference group is families whose head of household is white/ other. The findings then show the differences in probabilities that an outcome will occur for a certain group relative to the reference group, 70 controlling for the other characteristics in the model. Following our example, the “black” row presents differences in probabilities between blacks and white/other. To control for other characteristics, we use the estimates of the logit model to forecast the probability of a certain outcome, assuming that all individuals were black but leaving the remaining characteristics unchanged. Then, we calculate these probabilities assuming that all individuals were white/other, again leaving the remaining characteristics unchanged. The estimates presented in Tables 8.1 and 8.2 correspond to the mean of the differences of these two values over all families in the sample. Asterisks indicate which effects are estimated to differ from zero at standard confidence levels, with one asterisk indicating the 90 percent and two asterisks the 95 percent confidence level. Findings Starting with language and ethnicity, we find that one-parent families whose primary language is Spanish are more likely than Englishspeaking families to encounter at least two problems (Table 8.1). In particular, Spanish-speaking families are 26 percentage points more likely than English-speaking families to have income below poverty, despite having comparable chances of being employed. For two-parent families, the disparity between Spanish and English speakers with respect to poverty status is greater, at 31 percentage points (Table 8.2). However, our relatively small sample of two-parent families yields insignificant coefficients for the other outcomes. Among all families combined (not shown), lack of health care coverage is significantly more prevalent among Spanish speakers, although this result does not show up for each family type alone. Language and ethnicity play different roles for these families. Hispanic one-parent families are more likely than families of other ethnicities to encounter at least two problems, but the specific issues facing Hispanics are not identical to those facing Spanish speakers. Hispanic families do not have income below poverty more commonly than their counterparts, and the group as a whole is no more likely than its counterparts to lack health coverage. On the other hand, Hispanics more commonly live in crowded housing (by 19 percentage points), but Spanish-speaking families do not, suggesting that this problem does not 71 Table 8.1 Relationship Between Characteristics at Exit and Outcomes at 11–16 Months Postexit, One-Parent Families (marginal effect in percentage points) 72 Characteristics at Exit Primary language Spanish Englisha Ethnicity Black Latino/Hispanic White/othera Months on aid 0–12 13–36a 37–60 Number of children 1–2a 3+ Age of the youngest child 0–2 3–6 7+a Earnings at exit No earnings Positive earningsa Crowded Housing Income Below Poverty 11–16 Months Postexit No Earned Lacks Health Child Care a At Least One At Least Two Income Coverage Big Problem Problem Problems 0.01 0.26** –0.05 Ref Ref Ref 0.15 0.00 0.18** 0.25* Ref Ref Ref Ref 0.07 0.19** Ref 0.05 0.08 Ref 0.09* 0.00 Ref –0.03 0.04 Ref –0.07 0.03 Ref 0.06 0.11* Ref 0.01 0.14* Ref –0.03 Ref –0.01 –0.12** Ref 0.01 –0.05 Ref 0.01 0.11 –0.10** –0.01 –0.13** Ref Ref Ref Ref 0.04 0.05 0.08 0.00 Ref 0.24** Ref 0.11* Ref –0.06 Ref Ref Ref Ref 0.10 0.02 0.17** 0.16** 0.11** 0.00 –0.07** 0.00 –0.02 –0.03 0.04 0.07 –0.09* –0.02 0.06 –0.04 –0.08 –0.03 Ref Ref Ref Ref Ref Ref Ref 0.11 0.04 Ref Ref 0.11 Ref 0.12 0.13* 0.18** 0.16** Ref Ref Ref Ref Table 8.1 (continued) 73 Characteristics at Exit Administrative exit reason Failed to provide information Earnings Client request/noncooperation Othera Crowded Housing –0.14** –0.12** –0.15** Ref Income Below Poverty –0.05 –0.06 –0.12** Ref 11–16 Months Postexit No Earned Lacks Health Child Care a Income Coverage Big Problem –0.12** –0.01 –0.07 Ref 0.12* 0.02 –0.04 Ref –0.09 –0.05 –0.13** Ref SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. aReference category. (Other values shown are relative to this characteristic.) *Statistically significant at the 10 percent level. **Statistically significant at the 5 percent level At Least One Problem –0.03 0.02 –0.12 Ref At Least Two Problems –0.16** –0.19** –0.30** Ref Table 8.2 Relationship Between Characteristics at Exit and Outcomes 11–16 Months Postexit, Two-Parent Families (marginal effect in percentage points) 74 Characteristics at Exit Primary language Spanish Englisha Ethnicity Black Latino/Hispanic White/othera Months on aid 0–12 13–36a 37–60 Number of children 1–2a 3+ Age of the youngest child 0–2 3–6 7+a Earnings at exit No earnings Positive earningsa Crowded Housing –0.05 Ref Income Below Poverty 0.31* Ref 11–16 Months Postexit No Earned Lacks Health Child Care a Income Coverage Big Problem Insufficient data 0.21 0.33 Ref Ref At Least One At Least Two Problem Problems –0.13 Ref 0.26 Ref –0.30 0.11 Ref –0.15 –0.02 Ref 0.04 0.02 –0.45 0.04 0.01 0.11 0.08 0.14 Ref Ref Ref Ref 0.23 0.01 Ref Ref 0.15 0.11 0.01 –0.15 –0.03 0.10 Ref Ref Ref Ref 0.20 –0.33 0.03 0.11 Ref Ref 0.40 –0.22 Ref Ref Ref Ref –0.16 –0.14 0.02 0.13 0.42 0.11 0.29 0.13 Ref Ref –0.10 –0.25 Ref 0.24 0.49* Ref –0.04 –0.08 Ref 0.07 0.12 Ref 0.37 –0.26 Ref Ref 0.25 0.00 0.13** 0.20 Ref Ref Ref Ref Table 8.2 (continued) 75 Characteristics at Exit Administrative exit reason Failed to provide information Earnings Client request/noncooperation Othera Crowded Housing –0.04 –0.17 0.10 Ref Income Below Poverty –0.11 –0.03 –0.15 Ref No Earned Income Lacks Health Coverage –0.07 –0.43** –0.03 Ref SOURCE: The SPHERE Institute surveys of welfare leavers in six Bay Area counties. aReference category. (Other values shown are relative to this characteristic.) *Statistically significant at the 10 percent level. **Statistically significant at the 5 percent level. Child Care a Big Problem 0.22 0.27 –0.26** Ref At Least One At Least Two Problem Problems –0.08 –0.03 –0.01 Ref –0.09 –0.11 –0.05 Ref stem from the immigrant population. Neither Hispanics nor Spanishspeaking single-parent families are more likely than their counterparts to be unemployed, and none of the results for two-parent families are significant. Looking at black families, unemployment or lack of earned income is more prevalent in these families than in any other group. For example, one-parent black families are 9 percentage points more likely to have no earned income in the 11 to 16 month postexit period. Surprisingly, lower employment rates among single-parent black families do not translate into higher poverty rates or lower living standards when compared to other ethnic groups. Black families are as likely as white/other families to have income below poverty, suggesting that working black families have lower poverty rates than other groups. Also, blacks are not more likely than other ethnic groups to live in crowded housing. Time on aid before leaving CalWORKs is also linked to postexit outcomes for one-parent families but not for two-parent families. Not surprisingly, one-parent families (Table 8.1) who had relied on aid for less than 12 months in the last 5 years perform better than other families with higher dependency on aid in the past. For example, one-parent families with the lowest time on aid before exiting are 12 percentage points less likely to have income below poverty and 10 percentage points less likely to have child care as a barrier to employment. However, highly welfare-dependent families (families who spent more than 36 months out of the last 60 months on aid) are not more likely than moderate welfare users to have more problems (those with 13 to 36 months on aid). Comparing two-parent families by time on aid (Table 8.2) suggests that this is not a good indicator of the problems two-parent families will face after leaving aid. Single-parent families with three or more children are more likely to have at least two problems after leaving CalWORKs. Among these problems, crowded housing stands out as being the most prevalent for large families. One-parent families with three or more children are 24 percentage points more likely than other families to live in crowded housing (Table 8.1). In addition, they are 11 percentage points more likely to have incomes below the poverty threshold. The outcomes for 76 two-parent families are not significantly affected by the number of children in the family. One-parent families with young children are significantly more likely to live in crowded housing, regardless of the total number of children in the family. However, these families do not perform worse than other families in other areas. In fact, they are more likely to have earnings than families whose youngest child is over age six. For two-parent families with the youngest child between ages 3 to 6 years, child care needs are particularly prevalent. Tables 8.1 and 8.2 also compare families by their earnings at exit according to the UI Base Wage File. One-parent families who exit without earnings are 13 percentage points more likely to report that child care is a large barrier to full-time employment. They are also more likely to encounter at least two problems, although specific problems aside from child care are not independently significant. Looking at the magnitude of the estimated effects, however, it appears that crowded housing and lack of health care coverage are the largest problems for both one- and two-parent families. In fact, when we estimated the model on all families combined, both of these outcomes were significantly more likely among families without earnings at exit (not shown). It is interesting to note that additional factors beyond the absence of earnings at exit play a role in explaining poorer outcomes, since there were no significant differences in poverty or the probability of having no earned income 11 to 16 months postexit. Note also that since about two-thirds of families reporting no earnings in the UI file had positive earnings according to the survey (see Chapter 4), the group “No earnings” in Tables 8.1 and 8.2 includes all families with no earnings at exit as well as families working in sectors not covered by the UI file and families moving into larger households with other employed members. Finally, the last rows of Tables 8.1 and 8.2 compare families according to the exit reason as it is recorded in the administrative files. However, these results should be interpreted with caution, because many families simply failed to return forms or provide other information necessary to maintain eligibility. In fact, those families who simply left aid without providing an exit reason often outperformed other families. On one measure only—health coverage—families who left because of 77 failure to provide information had worse outcomes. Together, these findings suggest that families who exited without working closely with a caseworker typically had employment but may not have been aware of all the transitional benefits available, or—given that child care does not appear to be a greater barrier—simply chose to avoid the stigma of participating in public assistance. Summary The last two columns in Tables 8.1 and 8.2 summarize the experiences of leavers, measured by the number of problems encountered while off aid. The findings suggest that the best predictors of such problems among one-parent leavers are Hispanic ethnicity, three or more children in the family, Spanish as the primary language, and no earnings at exit according to the UI file. Except for earnings, characteristics at exit do little to predict the chance of two-parent families encountering at least one problem, although Spanish speakers were more likely to have income below poverty. This is likely due to our relatively small sample of twoparent families and the more complicated situations of these families, which might affect their postwelfare well-being. These characteristics could be used to develop a strategy to target postassistance support services for leavers. For example, programs related to health coverage could be targeted to Spanish-speaking families and families with no earnings according to the UI file. In the same way, housing assistance programs could be targeted to families with more than three children, families with young children, and Hispanic families. Finally, those who exited because of failure to provide information to maintain eligibility appear to have relatively strong outcomes in the 11 to 16 month period, but it might be helpful to provide additional outreach to these families to make sure that they understand the transitional benefits available, especially those allowing them to maintain health coverage. 78 9. Conclusion The survey data reviewed here provide a much richer picture of the experiences of leavers in six Bay Area counties than can be gleaned from administrative data. In fact, the administrative data, often used by policymakers and administrators to develop policies, have serious limitations in understanding outcomes for former recipients. Earnings data from the UI file, in particular, significantly understate both the level of employment and the amount of earnings in the households of welfare leavers. As far as the outcomes for leavers, we find aspects supporting both positive and negative views of their circumstances and prospects. On the positive side, a very high percentage of the households had someone with earnings, with this percentage remaining high throughout the period after leaving CalWORKs. Monthly income increased, albeit slightly, the longer leavers remained off CalWORKs, as did the percentage of households with incomes above the federal poverty level. About three out of four respondents, and four out of five children, had some form of health insurance coverage. Housing conditions also appeared to improve during the time away from CalWORKs, at least for the one-parent leavers. There was also fairly good news on other measures of well-being. Only about one out of ten leaver families experienced incidences of child risk behaviors (e.g., expulsion from school, trouble with police, use of drugs, and pregnancy), and such problems did not increase over time. Moreover, the percentage of respondents reporting substance abuse problems in the household declined among leavers the longer the time since their exit from CalWORKs. However, we also find some areas of concern. Many long-term leavers still had low incomes. Over one out of ten had monthly income falling below 70 percent of the poverty level, and this share of very poor households saw little or no improvement during time off CalWORKs. About half of the families saw their incomes and earnings decrease 79 between the first and second survey periods. Considered broadly, leavers had a high level of participation in public assistance programs. At the same time, many families eligible for nonassistance Food Stamps, nonassistance Medi-Cal, and EITC were not using these benefits. Similarly, about one-quarter of the respondents indicated that they were not aware of the availability of government subsidies that would help pay for child care for families that left welfare. From a policy perspective, our findings relating to the take-up of public assistance benefits after exit or benefit denial are important. Many leavers without health insurance or Food Stamps appeared to be eligible for these benefits. Similarly, there was often a lack of awareness of the availability of EITC and child care subsidies. These gaps point to the need for policies aimed at providing families with information about the availability of these post-CalWORKs benefits. Since it may be difficult to contact recipients after they leave CalWORKs, the dissemination of such information should begin while the families are still on aid. Similarly, this type of information could also be provided to CalWORKs applicants who are denied assistance but who could be eligible for other types of assistance. The results of our profiling analysis, where we identified recipient characteristics most likely to predict problems after exit from CalWORKs, could be useful in targeting supplemental assistance to these families, both while on CalWORKs and immediately after exit. Families exiting without earnings were the most likely to encounter problems, but other characteristics were also associated with later problems, such as having larger families or speaking Spanish as a primary language. Services designed to help recipients retain their jobs after leaving CalWORKs, for example, could be targeted more intensively toward leavers with these characteristics. Our findings on the discrepancy between administrative data and survey data that measure leavers’ earnings may also be important. Policymakers relying on administrative records—the most common source of this information—should recognize that these data may significantly understate earnings. Finally, the question of why many leaver families in poverty choose not to go back on CalWORKs warrants further research. About one- 80 third of our sample of long-term leavers had household incomes below the federal poverty level, and 11 percent were below 70 percent of the poverty level. In examining this question, we found that the survey respondents in the poor nonrecidivist leaver families had a higher level of work activity, but at the same time their household earnings were lower, on average, than their recidivist counterparts. This may have been due to differences in employment on the part of spouses or other household members—questions that were not specifically included in our surveys. 81 Appendix Data Sources and Methods This appendix describes data sources and methods used to compute the empirical findings presented in this report. The discussion starts by defining the study population and summarizing the datasets. It then presents the statistical approach. Finally, we discuss the construction of sample weights used to develop a representative sample of our population of study from the postexit surveys. Study Population The population for the study consists of CalWORKs assistance units, in six Bay Area counties, that left aid during the following time periods: • San Mateo, Santa Clara, and Santa Cruz Counties: fourth quarter 1998 • Contra Costa and Alameda Counties: third quarter 1999 • Sonoma County: fourth quarter 1999. To provide common reference points for characterizing the circumstances of leavers, we sorted the interviews according to the time elapsed since the families left CalWORKs and divided them into two periods of time: • 5 to 10 months (about 2 to 3 quarters) after leaving CalWORKs, and • 11 to 16 months (about 4 to 6 quarters) after leaving CalWORKs. Administrative Datasets This study uses county and state administrative data. We used county administrative data from each county’s Case Data System to identify the study populations. We also used county data to identify 83 family demographic characteristics and, in the case of leavers and the informally diverted, to identify the administrative reason for exit from or denial of cash aid. CDS information was also used to measure postexit and postdiversion receipt of public assistance within the three study counties. We used the state Medi-Cal Eligibility Data System to track receipt of public assistance elsewhere in California, outside the threecounty area, and we used the state Unemployment Insurance Base Wage File to track employment and earnings. The following list describes these datasets in more detail: • MEDS: This database includes monthly program participation for all Medi-Cal-eligible public assistance recipients as well as Food Stamps–only recipients (in recent years). Data are extracted from the system once a month to create a dataset that shows program eligibility month by month. MEDS also identifies county of aid receipt and individual demographic information, including ethnicity, age, gender, and language. Individuals are uniquely identified by Social Security number, and members of a TANF assistance unit are identified by a common case serial number. • UIBWF: Quarterly earnings records collected for the UIBWF are the key state-level data source we use to track employment and earnings. The UI system covers well over 90 percent of all workers in the state, with self-employed individuals constituting the largest category of excluded workers. • CDS: Each of the six counties in our study maintains its own CDS database, which it uses for maintaining current information on public assistance, Food Stamps, and Medi-Cal clients. CDS includes the same program participation and demographic information as MEDS but also contains many variables that are not found in MEDS, such as information on sanctions and reasons cases left public assistance. The counties sent to SPHERE extracts from CDS once a month, from which SPHERE constructed longitudinal files. Assistance units can be linked to MEDS by case serial numbers, and individuals can be linked to MEDS and the UIBWF by Social Security number. 84 Identifying the Study Population from CDS The population was identified using CDS data. To be considered to have left aid, all members of the assistance unit who were active or sanctioned must have left and remained off CalWORKs (off the original case and off any other case) for the next two months. The case had to have contained an aided (or sanctioned) adult. Thus, child-only cases were excluded from the population. Cases that left CalWORKs in a study county but moved to another county where they continued on aid were also excluded from the population; these intercounty transfers are identified in CDS by a specific exit-reason code. Postexit Surveys A central component of our study design is the implementation of a survey of a random sample of families in our study populations. Our survey contained topical modules with questions covering household composition, child well-being, child care, education and training, employment, income, food security, health insurance coverage, family well-being, and welfare experiences. The sampling frame for the survey consisted of all families in the study population, excluding those with a primary language other than English or Spanish. The sample is stratified by county; and for San Mateo, Santa Clara, and Santa Cruz Counties; the sample frame is also stratified by one-parent cases and two-parent cases. Table A.1 shows the distribution of our study population and sample by county and by case type. The interviewing was conducted primarily by telephone, although in some instances when attempts to contact the sample member by telephone were unsuccessful, an interviewer went to the sample member’s home and completed the interview in person. The interviewers were given contact information from CDS and MEDS, including addresses and telephone numbers, as well as names and Social Security numbers of adults in the assistance unit. Periodically, we gave the interviewers updated information from CDS and MEDS for sample members they had not yet located, and interviewers made use of public and private data sources to locate sample members. 85 Table A.1 Population and Sample Distributions, by County, Case, and Percentage Share Distribution of cases by county Alameda Contra Costa Santa Clara Santa Cruz San Mateo Sonoma Study Population Sample 28 18 24 22 26 11 6 14 8 12 8 23 Distribution of cases by type of case One-parent cases Two-parent cases 90 10 87 13 NOTE: Columns may not sum to 100 because of rounding. The survey respondent was almost always the mother or female caretaker of the children in the assistance unit. Some survey questions were asked about a “focal” child. The focal child was selected randomly from the children living with the respondent at the time of the first interview, and from those children for whom the respondent was responsible. Note that the focal child did not need to be the biological or adopted child of the respondent. Surveys were administered in two waves for each family, over a period of roughly 15 months after exit from CalWORKs. Table A.2 shows the population of leavers and the number of interviews completed in the 5 to 10 and 11 to 16 month postexit periods by county. The response rates for the total combined sample were 54 percent for the first set of interviews and 66 percent for the second wave.1 In addition to information on postexit aid use, employment, and earnings, MEDS and the UIBWF provide historical information on ____________ 1Not every case in our sample responded to both interviews. In some instances, we conducted separate analyses that linked only those individuals who responded to interviews in each period. 86 Table A.2 Population, Samples, and Survey Response Counts Formulation of Weights County Alameda Contra Costa San Mateo Santa Clara Santa Cruz Sonoma Total Population 1,082 964 301 1,013 238 309 3,907 Sample 238 304 170 149 184 309 1,354 First Wave 133 203 47 38 66 242 729 Second Wave 169 206 98 84 118 215 890 these outcomes that were used, along with other measures of family characteristics, to construct weights to adjust survey data for nonresponse. Constructing a representative sample of the population of leavers in the six counties included in this study required the computation of sample weights that adjusted for differences in sampling rates, nonresponse, and other sources of bias. To create these weights, we followed a two-stage approach. In the first stage, we computed a basic county weight equal to the inverse probability of selection, where the probability of selection is equal to the number of respondents in the county divided by the study population in the county. This basic weight adjusts sample observations for differences in sampling and response rates across counties. For the second stage, we applied iterative raking to balance a population’s distributions by ethnicity, age of the case head, and earning status at exit according to UIBWF. Iterative raking, used by the Census Bureau to adjust sample weights, adjusts the baseline take-up rate so that the predicted number of cases converges to the population benchmarks, while minimally affecting the baseline take-up rate. For each iteration of each grouping variable (called raking variable), an adjustment factor equal to the number of cases in the study population divided by the number of predicted totals using the weighted sample is computed and applied to the estimate of that cell. This procedure is 87 repeated until the estimated number of participants in each cell equals the study population controls. The following grouping variables were used: • Ethnicity: white, Hispanic, black, other, • Age of case head: 0 to 15, 16 to 25, 26 to 35, 36 to 45, 46+, • UIBWF earning status: positive earnings, no earnings. Table A.3 compares population and sample distributions after applying the constructed weights by grouping variable. Our calibration creates predicted distributions that exactly match those in the study population. This arises by construction, since our procedure minimizes discrepancies in this match. Table A.4 presents estimates of the distribution of subgroups of leavers for additional demographic characteristics. In contrast to results Table A.3 Demographic Characteristics of All Leavers, by Grouping Variables and Percentage Share Demographic Characteristics Ethnicity Black Hispanic White Other Leavers at 5–10 Monthsa (Population) 29 27 35 9 Leavers at 5–10 Monthsa (Sample) 29 27 35 9 Leavers at 11–16 Monthsb (Population) 28 27 35 9 Leavers at 11–16 Monthsb (Sample) 28 27 35 9 Age of case head 16–25 32 32 31 26–35 36 36 36 36–45 25 25 26 46+ 7 7 7 31 36 26 7 Earnings at exit No earnings 27 27 26 Positive earnings 73 73 74 26 74 Number of cases 3,038 541 2,684 NOTE: Columns may not sum to 100 because of rounding. aIncludes short-term and long-term leavers. bIncludes only long-term leavers. 575 88 Table A.4 Additional Demographic Characteristics of All Leavers, by Grouping Variables and Percentage Share Demographic Characteristics Primary language English Spanish Leavers at 5–10 Monthsa (Population) 93 7 Leavers at 5–10 Monthsa (Sample) 91 9 Leavers at 11–16 Monthsb (Population) 93 7 Number of children 0–1 2 3+ 53 28 19 53 30 18 52 29 19 Age of youngest 0–2 32 34 3–5 25 24 6–11 24 25 12+ 18 17 31 25 25 19 Months on cash aid in preceding 5 years 1–12 13–36 37–60 20 26 54 20 25 55 19 25 55 Number of cases 3,038 541 2,684 NOTE: Columns may not sum to 100 because of rounding. aIncludes short-term and long-term leavers. bIncludes only long-term leavers. Leavers at 11–16 Monthsb (Sample) 91 9 51 29 20 35 25 23 16 22 28 50 575 in Table A.3, our weighting mechanism does not automatically match these numbers, so the closeness of our estimates to the study population indicates that our weighting mechanism performs well in creating a representative sample of the study population. 89 Table A.5 Counties Included in Each Region Bay Area Alameda Contra Costa Marin Napa San Francisco San Mateo Santa Clara Santa Cruz Solano Sonoma Other Southern Los Angeles California Farm Belt Los Angeles Orange Colusa Riverside El Dorado San Bernardino Fresno San Diego Glenn Santa Barbara Imperial Ventura Kern Kings Madera Merced Monterey Placer Sacramento San Benito San Joaquin San Luis Obispo Stanislaus Sutter Tulare Yolo Yuba North and Mountain Alpine Amador Butte Calaveras Del Norte Humboldt Inyo Lake Lassen Mariposa Mendocino Modoc Mono Nevada Plumas Shasta Sierra Siskiyou Tehama Trinity Tuolumne 90 Bibliography Daly, Mary, “Labor Market Effects of Welfare Reform,” FRBSF Economic Letter 97-24, Federal Reserve Bank of San Francisco, San Francisco, California, August 29, 1997. Gritz, R. Mark, et al., Welfare Recidivism and Use of Non-Cash Aid by Welfare Leavers in California, The SPHERE Institute, report prepared for the U.S. Department of Health and Human Services (Assistant Secretary for Planning and Evaluation), Burlingame, California, March 30, 2001a. Gritz, R. Mark, David Mancuso, Charles J. Lieberman, and Vanessa L. Lindler, Assessing the Family Circumstances of TANF Applicants and Leavers in Contra Costa and Alameda Counties, The SPHERE Institute, report prepared for the U.S. Department of Health and Human Services (Assistant Secretary for Planning and Evaluation), the Contra Costa County Employment and Human Services Department, and the Alameda County Social Services Agency, October 26, 2001b. Lieberman, Charles, and David C. Mancuso, “Are California’s Welfare Leavers Staying Off Public Assistance? Welfare Recidivism and Use of Non-Cash Aid Since Welfare Reform,” California Policy Review, The SPHERE Institute, Burlingame, California, October 2001. MaCurdy, Thomas, David Mancuso, and Margaret O’Brien-Strain, The Rise and Fall of California’s Welfare Caseload: Types and Regions, 1980– 1999, Public Policy Institute of California, San Francisco, California, 2000. Mancuso, David C., Charles J. Lieberman, Vanessa L. Lindler, and Anne Moses, Examining Circumstances of Individuals and Families Who Leave TANF: Assessing the Validity of Administrative Data, Final 18Month Report, The SPHERE Institute, report prepared for the U.S. Department of Health and Human Services (Assistant Secretary for Planning and Evaluation), the U.S. Department of Housing and Urban Development, the San Mateo County Human Services Agency, 91 the Santa Clara County Social Services Agency, and the Santa Cruz County Human Resources Agency, November 30, 2001. Mancuso, David C., and Vanessa L. Lindler, Examining the Circumstances of Welfare Leavers and Sanctioned Families in Sonoma County, The SPHERE Institute, report prepared for the Sonoma County Human Services Department, June 29, 2001. 92 About the Authors THOMAS MACURDY Thomas MaCurdy is an adjunct fellow at PPIC, a professor of economics at Stanford University, and a senior fellow at the Hoover Institution. His principal research explores topics in the areas of income transfer programs, human resources, and labor markets. His recent studies investigate the consequences of governmental policies underlying welfare programs, unemployment compensation, Social Security, Medicare, Medicaid, and various forms of public assistance for lowincome populations. He holds an A.B. from the University of Washington and a Ph.D. in economics from the University of Chicago. GRECIA MARRUFO Grecia Marrufo is a research associate at The SPHERE Institute and a research associate at the Stanford Institute of Economic Policy Research. Her research has focused on the financing of social insurance programs and modeling the effect of labor regulations on such programs. She is also working on a U.S. Department of Agriculture project to evaluate alternative certification procedures in the Food Stamps program, focusing on the empirical simulation of changes in benefits arising from alternative reporting requirements. She holds a B.A. in economics from the Instituto Tecnológico Autónomo de México and a Ph.D. in economics from the University of Chicago. MARGARET O’BRIEN-STRAIN Margaret O’Brien-Strain is an adjunct fellow at PPIC and Director of Social Policy at The SPHERE Institute. She is interested in a broad range of social services issues and has written about child care, home visitation programs, performance measurement, and welfare reform in California. She holds a B.A. in economics from Swarthmore College and a Ph.D. in economics from Stanford. 93 Related PPIC Publications Employers and Welfare Recipients: The Effects of Welfare Reform in the Workplace Harry J. Holzer and Michael A. Stoll The Basic Skills of Welfare Recipients: Implications for Welfare Reform Hans P. Johnson and Sonya M. Tafoya Does California’s Welfare Policy Explain the Slower Decline of Its Caseload? Thomas E. MaCurdy, David C. Mancuso, and Margaret O’Brien-Strain Reform Reversed? The Restoration of Welfare Benefits to Immigrants in California Thomas MaCurdy and Margaret O’Brien-Strain The Rise and Fall of California’s Welfare Caseload: Types and Regions, 1980–1999 Thomas MaCurdy, David Mancuso, and Margaret O’Brien-Strain Who Will Be Affected by Welfare Reform in California? Thomas MaCurdy and Margaret O’Brien-Strain Federal Formula Grants and California: TANF and Welfare Programs Tim Ransdell PPIC publications may be ordered by phone or from our website (800) 232-5343 [mainland U.S.] (415) 291-4400 [Canada, Hawaii, overseas] www.ppic.org 95" ["post_date_gmt"]=> string(19) "2017-05-20 09:36:15" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(8) "r_903tmr" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2017-05-20 02:36:15" ["post_modified_gmt"]=> string(19) "2017-05-20 09:36:15" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(50) "http://148.62.4.17/wp-content/uploads/R_903TMR.pdf" ["menu_order"]=> int(0) ["post_mime_type"]=> string(15) "application/pdf" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" ["status"]=> string(7) "inherit" ["attachment_authors"]=> bool(false) }