1. to assess the needs of new and transitioning welfare recipients;
2. to identify the constellation of personal, community, and programmatic barriers to self sufficiency and to successful departure from CalWORKs;
3. to determine what outcomes result for this population.

The criteria for individuals to be included in the study is:

1. age 18-592.
2. speak English, Spanish, Vietnamese, or Cambodian as their primary language
3. parent or a caregiver in a 1- or 2- parent household.
4. not a non-needy caregiver
5. no permanent disability
6. non-exempt from CalWORKs work requirements.

In addition, key informant interviews gathered information on program implementation from administrators and line staff.

Post Program Activities-
Support services used following leaving welfare.

Social/Support Services-
Assessment of utilization of mental/health, substance abuse, transportation, legal, financial, job retention and child care services.

To examine with multivariate methods the effects of barriers and supports on welfare and work outcomes.
To add to the literature an understanding of factors resulting in part-time vs. full-time work, work with and without welfare, and neither source of support, compared to reliance on welfare alone.

Each recipient was also questioned about one randomly selected child in his or her family for this interview.

* Wave 1 sample was stratified and a subset of respondents were selected for Wave 2 interviews. Stratification of sample for follow-up implies need for weighted analysis.

At each interview, representativeness compared to the original baseline sample was examined. Weights correct for distortions in Wave 2 sample introduced by stratification

• There were many differences between one- and two-parent cases.
• Two-parent respondents, compared to 0ne-parent respondents, were more likely to be married (68% vs. 14%), to identify themselves as Hispanic, Vietnamese, Cambodian or Other (72% vs. 44%), and to have been born outside the US (47% vs. 19%)
• Foreign-born respondents have more in common with foreign-born of another ethnicity than with second-generation persons of the same ethnicity
• Specific barriers to employment include physical health problems, serious mental health problems, alcohol and drug use problems, family violence, and learning problems
• Only 4% of the sample was assessed as having none of 17 potential barriers to work; 56% had four or more barriers
• Barrier prevalence differed by type: 34% had no health related barriers, 18% had no human capital barriers, 52% had no family responsibility barriers, and 47% had no logistic barriers
• Respondents from one- and two-parent cases had a similar number of barriers, about 4, but they were dealing with different barriers.
• 26% of the study participants have not worked 26 or more hours a week for at least 2 consecutive weeks within the last 4 years (44% for two-parent cases vs. 23%)
• Black respondents differed from other race/ethnic groups in several ways: more likely family history of welfare (71% vs. 46%), more likely no work history (16% vs. 11% for Whites), but also more likely to be looking for work*37% vs. 25%) or working full-time (36% vs. 25%), more likely to live in solo-adult household (45% vs. 39%), more likely to have a child with special needs (29% vs. 22%), and more likely to experience the social ills of involvement in crime (29% vs. 19%) and illegal drug use (15% vs. 9%), but less likely to report health problems (30% vs. 37%) or mental health problems (16% vs. 37%).
• Vietnamese and Cambodian respondents seem particularly unlikely to be able to meet work requirements and will need culturally and linguistically tailored support services.

• Progress through the CalWORKs program: Fewer than half the sample reported receiving employment counseling (42%) or participation in Job Club (35%) by the Wave 2 interview (Table 2-2). From the survey data alone, it is not clear whether additional respondents might also have actually received these services but did not recognize the terminology in the survey, or received these services at a location other than the CalWORKs offices and did not recognize them as CalWORKs services.
• There was considerable overlap between work and welfare. At the second interview, half of those working 32 or more hours per week were also receiving welfare benefits (Table 2-7). For many on welfare, “full-time” work appears insufficient to permit them to end reliance on cash welfare benefits.
• Twenty-two percent of the sample was exempt from working at one or both interviews. The percentage exempt declined from Baseline to Wave 2 (from 13% to 8%) (Table 2-6). Fewer than 1% of respondents said they were exempt from working at both waves, and their reasons were personal disability or care of a disabled family member.
• Foreign language welfare recipients may not be getting the English language skills training that they need to become self-sufficient. Of the 38 respondents interviewed in Spanish, Vietnamese or Cambodian, who also indicated that they had signed a welfare-to-work plan at Baseline, only seven said that ESL classes were part of their plan
• Fewer study participants were relying on public assistance at their Wave 2 interview. The average share of total household income from public assistance was 48% at Wave 2 compared to 60% at Baseline.
• The average total household income in the last 30 days increased 37% from $1780 ($21,360/year) at Baseline to $2441 ($29,292/year) at Wave 2, not controlling for inflation.
• Work makes a difference (Table 3-5). At Wave 2, working respondents had higher total household incomes than non-working respondents in the same welfare status, and the proportion in poverty was lowest for those working 32 or more hours per week.
• Among respondents working 32 or more hours per week, only 32% of those still on welfare and 47% of those off welfare had total household income above the lowest “basic monthly budget” estimated for California’s Central Valley.
• Barrier categories differed in prevalence and stability across time (Table 4-2). Over 70% of the sample had one or more family responsibility barriers at each wave, although affected persons changed over time. About two-thirds of the sample had one or more human capital barriers, and these tended to persist over time. About 63% of the sample had a logistic barrier, and transportation problems persisted over time. Over 40% of the sample had one or more health-related barriers.
• For some potential barriers, there was little change in prevalence from Wave 1 to Wave 2. Health-related barrier prevalence changed less than less than 3% in either direction for seven of 11 possible barriers across waves.
• There was also little change in prevalence of limited education, limited English fluency, lack of family child care, children requiring a lot of care, and temporary housing. The prevalence of learning disabilities appears to be similar to that in the general population.
• Prevalence of individual potential barriers in the San Joaqin sample is similar to those seen in CalWORKs populations in other regions of the state (Alameda, Stanislaus and Kern County) and TANF recipients elsewhere in the country (Michigan), varying from about 2% to 55%.
• Prevalence of the most common barriers to employment also significantly associated with not working 32 or more hours per week at Wave 2: (1) not having family child care (55%), (2) not owning a car (49%), (3) limited education (48%), (4) not having use of a car or driver’s license (46%), and (5) having fewer than five work skills (29%).
• At Wave 2, two less common barriers appeared to have an equally strong impact on work status: mental health conditions (12% of respondents), and criminal justice system involvement (10% of respondents). Fewer than 25% of individuals with these barriers were working at Wave 2, compared to about 46% working without such barriers.
• Number of barriers matters: Among respondents with no barriers, 88% were working, compared to only 30% of respondents with five or six potential barriers. Over 31% of respondents had five or more barriers at Wave 2.
• Supports for working: Over time, there was no increase in the 15% of respondents using government child care programs (Head Start, extended day school programs) or the 30% using formal day care or a sitter, but family-provided care increased from Wave 1 to Wave 2. At Wave 2, 85% of those using any child care were using family child care, many combining family care with other arrangements.
• Receipt of a rent subsidy may increase working. At Wave 2 the proportion of respondents working 32 or more hours with rent subsidies equals the proportion working without subsidies; however, from Wave 1 to Wave 2, working increased three-fold among those with a rent subsidy, compared to doubling of working among those without a subsidy.
• The CalWORKs program appears to detect only about 30% of respondents who self-identify problems with mental health, substance abuse, and family violence, and even fewer of those who do not self-identify needs for help. Failure to identify problems that compromise ability to find and retain employment costs recipients precious time.

• Finding in this report are based on multivariate analyses combining survey data on barriers and supports with administrative data on monthly welfare status and and quarterly earnings.
• Multinomial logistic regression explored predictors of six work and welfare outcomes: 1) Welfare only (comparison group), 2) neither work nor welfare, 3) welfare and work less than 32 hours per week, 4) welfare and work 32 or more hours per week, 5) work only less than 32 hours per week, and 6) work only 32 or more hours per week
• Number of barriers: The better outcomes were less likely for those who reported more barriers.
• Individual barriers: lack of transportation and child care problems were most likely to limit working or combining welfare and work
• Physical and mental health problems, alcohol or drug abuse, experiencing partner control or violence, and having low education are not strongly associated with working or combining work and welfare in our sample. While reporting language limitations and few work skills are not associated with all outcomes, they are associated with a lower odds of working full-time relative to receiving welfare alone.
• In simulations of a hypothetical recipient, removing barriers in general, and child care and transportation barriers in particular, increases the probability that she works full-time and decreases CalWORKs receipt in the absence of earnings. Removing barriers also lessens the probability that she is neither working nor receiving cash assistance. Even so, this hypothetical recipient in the best of circumstances has a 20-25% chance of working less than 32 hours per week either alone or in combination with CalWORKs, and about a 10% probability of receiving CalWORKs alone.
• The probability of receiving welfare alone in simulations drops from 26 to 10% as the number of reported barriers declines, while the probability of working full-time in combination with CalWORKs triples and the probability of working full-time as a welfare leaver grows more than five-fold. The two part-time work categories are much less affected; however, the probability of neither working nor receiving CalWORKs also drops nearly three-fold.

• Support for child care & transportation must be retained, and accessibility and utility to respondents must be substantially improved.
• Low uptake of child care subsidies, regardless of the reasons, means that young children of CalWORKs recipients are missing the opportunity to benefit from formal, structured child care. Child care supports may not be used by respondents due to difficulties in navigating the child care system and a shortage of desirable, affordable providers.
• Given the primacy of family-based child care in these models, it seems important to encourage creative solutions for supporting and funding family-based child care arrangements. TANF funding might be permitted to be used to encourage worksite child care as well as other options for managing non-standard work hours. Beyond TANF, it will be important to maintain funding for child care assistance for the working poor, and to assist low-wage former welfare recipients in transitioning to alternative sources of child care funding.
• Low uptake of transportation subsidies may be an indication that there is no substitute for owning a car. It seems obvious that no one needs a car more than a working single parent. Transportation subsidies appear to be inadequate support for dealing with limitations in public transportation and the complexity of travel needs. We recommend creative programs that support purchase and maintenance of private vehicles.
• High drop-out rates from educational activities may limit respondents’ ability to attain self-sufficiency, and may indicate the need for additional supports, or more consistently provided supports, for respondents pursuing education.
• Further research is needed on barriers that may limit or constrain, rather then prevent, working, and policy supports may need to be maintained for persons with such constraints.
• Given that every additional barrier reduces the likelihood of working, screening programs to identify individuals with physical and behavioral health barriers to employment, as well as human capital deficits may be advisable.
• A substantial proportion of welfare recipients may not be able to leave welfare or remain employed, at least with current CalWORKs program levels of support. Early screening and assessment may also facilitate decisions in favor of more intensive case management, appropriate exemptions from work requirements, or timely assistance in pursuing disability claims.
• Even those without barriers to employment appear likely to have periods of unemployment, and need for public assistance in some form. Policy options for those in unstable labor markets could include extended unemployment benefits, quick access to emergency funds for living expenses, and continued health insurance for unemployed adults.