1. ASSESSING WELFARE REFORM IN THE U.S.: FROM CASH ASSISTANCE TO LOW-WAGE EMPLOYMENT Sheldon Danziger, Henry J. Meyer Distinguished University Professor of Public Policy, University of Michigan APRIL 2006

2. Background Policy Context The 1996 federal welfare reform ended the 60-year entitlement to cash assistance for single mothers and mandated participation in work or work-related activities The law achieved several goals – 1.Welfare caseloads declined dramatically 2.More single mothers worked 3.Poverty fell 4.Many states used some funds previously spent on cash assistance for nonworkers to supplement low earnings and subsidize child care However, a significant minority of welfare recipients have personal problems that kept them from working steadily, even during the economic boom of the late 1990s. There has been an increase in the number who receive neither work nor wages.

3. Welfare Receipt, Women with Children, High School Degree or Less, 1975-2004, By Marital Status Fraction Using Welfare Year Source: Computations from March Current Population Surveys

4. Employment Rate, Women with Children, High School Degree or Less, 1975-2004, By Marital Status Fraction Working During Calendar Year Year Source: Computations from March Current Population Surveys

5. Official Poverty Rate, Women with Children, High School Degree or Less, 1975-2004, By Marital Status Fraction Poor Source: Computations from March Current Population Surveys Year

6. Women’s Employment Study Overview Longitudinal study of white non-Hispanic and African American women who received TANF (single parent case) in February 1997 Sample drawn from one urban Michigan county. Response rates: 1997 – 86% (753 women) 1998 – 93% (693 women) 1999 – 91% (632 women) 2001 – 91% (577 women) 2003 – 93% (536 women) No evidence that attrition bias is a problem.

7. Percent Working, Receiving FIP, and Receiving Food Assistance by Month: February 1997 – August 2004 (W5 non-SSI recipients only, N = 503)

8. Women without Work and Cash Welfare, 1997-2003 Notes: Non-respondents and respondents receiving SSI at each wave are not included. 1. Respondents without wages and without cash welfare in the month prior to the survey. 2. Loprest (2003) defines a woman as disconnected if she receives no wages and no cash welfare, has not worked recently, and is not living with a spouse/partner with earnings. 3. Definition 3 excludes women who are counted as disconnected in Definition 2 who lived with any other earners or lived in households where unemployment insurance/workers’ compensation was received in the month prior to the interview or worked in the three months prior to the interview.

9. Percentage of Months Worked, February 1997 to August 2003 (SSI recipients excluded) n = 503

10. Changes in Job Quality, 1997 – 2003 (in 1999 dollars) Good job: Full time (35+ hours/week), at least $7/hour and offers health benefits either immediately or after a trial period. OR, Full time, at least $8.50/hour and no health benefits. (Those who meet the criteria, but who are working part-time jobs voluntarily (would not take more hours if the employer offered) are considered to have good jobs).

11. Wave 1 N = 692 Wave 2 N = 585 Wave 3 N = 545 Wave 4 N = 494 Wave 5 N = 426 <$5.1522.5%8.6%8.8%6.9%11.7% $5.15 – 6.50 28.830.618.718.016.9 $6.51 – 7.50 17.118.520.917.216.7 $7.51 – 8.50 8.718.115.817.012.0 $8.51 – 9.50 5.18.214.111.99.6 $9.51 – 11.00 5.95.68.410.313.2 >$11.0112.010.413.218.620.0 Mean$7.59$7.84$11.08$10.21$8.83 Median$6.42$7.06$7.68$7.97$8.00 Distribution of Hourly Wage Rates, 2003$ (Respondents with positive wages)

12. Percent of Respondents with No Health Insurance by Work/Welfare Status (w5 Rs Only, n=536)

13. Monthly Poverty Rate, Current and Former Welfare Recipients, 1997-2003 Note: Income from all sources, including food stamps, and all household members. Net monthly income subtracts work related child care and transportation expenses and income and social security taxes Source: Women’s Employment Study, University of Michigan

14. Percent Poor Monthly Poverty Rate, Fall 2003, Current and Former Welfare Recipients, By Work Status and Marital Status Source: Women’s Employment Study, University of Michigan. All respondents received cash welfare in February 1997. In Fall 2003, only 19.6 percent received welfare. Sample size is 534. Poverty rate is measured as net household income which subtracts income and social security taxes and work-related and child care expenses

15. Research Questions 1.To what extent did welfare mothers have work limitations and other barriers when panel study began shortly after welfare reform? 2.How are barriers in 1997 correlated with the welfare to work transition, measured by accumulated months of employment and welfare receipt over 6½ year period? 3.How does persistence of barriers affect earnings and welfare receipt in 2003, controlling for many other factors?

16. Measures of Barriers Low work experience, job skills, work norms or criminal record No high school degree or learning/reading problems Met screening criteria for a mental health disorder (depression, post-traumatic stress disorder, general anxiety disorder, social phobia) Met criteria for alcohol dependence or used illegal drugs High health limitations and self-reported fair/poor health Child with physical, emotional or learning problems Experience with severe domestic violence in last year

17. Percent of Fall 2003 Respondents with problems in selected domains, Fall 1997 (n= 536) Low work experience, skills, work norms or criminal record 34.7% No high school degree or learning/reading problems44.8 Met screening criteria for mental health disorder36.0 Met criteria for alcohol dependence or used illegal drugs22.4 Health limitations and self-reported fair/poor health19.4 Child with physical, emotional or learning problems22.9 Experience with severe domestic violence in last year 16.0 No barriers16.4% Problems in any one domain24.4 Problems in any two domains27.6 Three domains16.8 Multiple barriers, four or more domains14.7 100 %

18. Persistence of Selected Barriers, 1997 – 2003 (Percent of Respondents)

19. Percent of Respondents who met Diagnostic Criteria for a Disorder at Least Once, 1997 – 2003 (Wave 5 R’s only)

20. Percent of Months Worked between February 1997 & August 2003 (79 months), By Number of Barriers at W1 (out of 17)

21. Measures of work and welfare outcomes  Percentage of months worked 1997-2003  Percentage of months received welfare, 1997-2003  Earnings last month, 2003  Received cash welfare last month, 2003

22. Multivariate regression models of welfare- to- work outcomes Dependent variables: % of months worked 1997-2003 % of months of TANF Monthly earnings in 2003 Receipt of TANF As a function of: ** 1997 work and educational limitations, health, mental health, child health, alcohol dependence, illegal hard drug use, or domestic violence problems. 1997 work and educational limitations Respondent met criteria for health, mental health, child health, alcohol/drugs, or domestic violence barrier never, in 1-2, or 3-5 waves ** Both sets of models control for maternal age, race, young children, married or cohabiting

23. Tobit Regression results predicting work and TANF, 1997- 2003

24. Regressions predicting 2003 monthly earnings and welfare receipt

25. Median earnings, Full-Time, Year Round Workers, 1960-2004 Source: U.S. Bureau of the Census

26.  “Make Work Pay”—increased aid for working poor - Raise minimum wage to about 40% of average wage and index it for inflation—about $6.75 now - Expand state child health insurance program (SCHIP) to cover parents—example: Badgercare in Wisconsin - Expanded earned income tax credit for single persons and childless couples - Make per capita child care credit in federal income tax refundable—poor who pay no federal income tax do not now receive it  “Transitional Jobs of Last Resort” - For those no longer entitled to cash welfare, this provides cash for those willing to work who cannot find jobs A modest antipoverty initiative to cut the U.S. poverty rate in half