Gender-Specific Barriers to Self-Sufficiency among Former Supplemental Security Income Drug Addiction and Alcoholism Beneficiaries: Implications for Welfare-To-Work Programs and Services

Abstract

This study examines barriers to economic self-sufficiency among a panel of 219 former Supplemental Security Income (SSI) drug addiction and alcoholism (DA&A) recipients following elimination of DA&A as an eligibility category for SSI disability benefits. Study participants were comprehensively surveyed at six measurement points following the policy change. Generalized estimating equations were used to examine full-sample and gender-specific barriers to economic self-sufficiency. Results indicate that access to transportation, age, and time are the strongest predictors of achieving self-sufficiency for both men and women leaving the welfare system. Gender-specific barriers are also identified. Future research needs to assess the generalizability of these results to other public assistance recipients.

Prior to 1996, low-income individuals with severe substance abuse problems and an inability to acquire substantial gainful employment could qualify for public assistance benefits in the Supplemental Security Income (SSI) program. During the welfare reforms implemented that year, drug addiction and alcoholism (DA&A) were eliminated as qualifying impairments for disability benefits with the intent of removing substance abusers from the welfare rolls. Welfare reform legislators rationalized that former SSI DA&A beneficiaries with “legitimate” disabilities would requalify for SSI benefits under another disability category (Watkins, Podus, Lombardi, & Burnam, 2001). Unfortunately, only 35% of former SSI DA&A beneficiaries reacquired their SSI benefits (Lewin Group & Westat, 1998). The remaining 65% (over 110,000 individuals) were expected to become personally responsible, find employment, and achieve self-sufficiency.

The primary goals of the 1996 welfare reform movement were to reduce welfare dependence and motivate welfare-reliant individuals toward greater economic and material self-sufficiency (Danziger, Corcoran, Danziger, & Heflin, 2000; Daugherty & Barber, 2001; Julnes, Fan, & Hayashi, 2001). While the goal of reducing the number of people receiving public assistance (i.e., welfare dependence) was realized, in reality, these policy changes resulted in a substantial proportion of welfare leavers reporting increased economic and material hardship (Acs & Loprest, 2007); few were able to make any instrumental gains toward self-sufficiency (Long, 2001; Oliphant, 2000). As a special population of welfare leavers, many former SSI DA&A beneficiaries also reported compromised economic and material security following the loss of public assistance (Hogan, Unick, Speiglman, & Norris, 2008; Norris, Scott, Speiglman, & Green, 2003; Orwin, Campbell, Campbell, & Krupski, 2004). Hogan and associates (2008) found that nearly 50% of former SSI DA&A beneficiaries from Northern California who failed to requalify for SSI benefits under another disability category experienced no employment up to 3.5 years following the policy change; the average monthly income for this group was only $379. The purpose of the current study is to examine the factors that prevented this specific population of former public assistance recipients from meeting welfare reform policymakers’ goals and successfully transitioning out of the welfare system. Accurate identification of barriers to employment and self-sufficiency (e.g., lack of transportation, lack of job training or work experience, physical and mental health issues, substance abuse problems, and lack of access to child care) is important to policymakers for the efficient and successful provision of welfare-to-work programs, supports, and benefits.

In the current study, barriers to self-sufficiency for a panel of 219 former SSI DA&A beneficiaries from Northern California were examined following elimination of the DA&A category. Self-reported income and employment data were gathered at baseline (December, 1996), and 6-, 12-, 18-, 24-, and 42-month follow-up interviews to construct a liberal measure for self-sufficiency based on any paid employment and an average monthly income above the federal poverty line. Additional information was gathered on perceived barriers to self-sufficiency, including access to transportation, work history, lack of job training, physical and mental health, substance abuse, interpersonal violence, and amount of social support. Demographic information, such as age, ethnicity, marital status, and education, was also collected. General estimating equations were utilized to predict self-sufficiency over time for the entire study sample. Following this, the sample was disaggregated by gender to examine specific barriers to self-sufficiency among former male and female welfare recipients. The substantial male representation in the current study sample (54.8%) offered researchers a unique opportunity to examine gender-specific differences in barriers to self-sufficiency among a population of public assistance recipients transitioning out of the welfare system. Specifically, the research questions for this study were:

1. Which specific barriers to self-sufficiency had the most significant effects on former SSI DA&A beneficiaries?

2. Where there any gender-specific differences in these effects?

Literature Review

The barriers to employment and self-sufficiency identified for welfare leavers in previous research are numerous. The most prevalent predictors and correlates to transitioning from welfare to work include demographic characteristics such as age (Danziger & Lehman, 1996), ethnicity (Cheng, 2002), marital status (Julnes et al., 2001), education (Horowitz & Kerker, 2001), and number of children (Schmidt, Zabkiewicz, Jacobs, & Wiley, 2007); individual health characteristics are also prolific. These include physical and mental health status (Nam, 2005; Speiglman & Li, 2008), history of substance abuse (Jayakody, Danziger, & Pollack, 2000) and interpersonal violence (Riger, Staggs, & Schewe, 2004), and mental health and substance abuse treatment experiences (Metsch, Pereyra, Miles, & McCoy, 2003; Wickizer, Campbell, Krupski, & Stark, 2000). Employment characteristics such as job skills and work experience are prominent (Cheng, 2002); as are systemic or structural barriers including access to transportation (Cervero, Sandoval, & Landis, 2001) and child care (Dulmus, Rapp-Paglicci, Sarafin, Wodarski, & Feit, 2000). Social support from friends and family is also identified as a predictor to employment and self-sufficiency for former welfare beneficiaries (Brown & Riley, 2005). The vast majority of research on the effects of welfare reform, post-welfare-reform outcomes for welfare leavers, and transitional experiences from welfare to work has been conducted on single mothers leaving Temporary Assistance for Needy Families (TANF). In order to present a comprehensive profile of barriers to employment and self-sufficiency for welfare leavers, past research on gender-specific barriers, and information for an effective discussion, this literature review will present past research on both the SSI DA&A and TANF populations.

Demographic Barriers

Younger welfare recipients have more difficulty transitioning out of welfare (Danziger & Lehman, 1996; Schmidt et al., 2007), as well as an increased probability of returning to public assistance after leaving (Julnes et al., 2001). This outcome is generally attributed to the young age of most AFDC and TANF study samples. The significance of age on employment and self-sufficiency among former SSI DA&A beneficiaries is unknown; however, it was found in previous research on the current study sample that increased age and being male were associated with less income (Hogan et al., 2008). Similar findings have been reported for single (Danziger & Lehman, 1996) and never married welfare recipients (Julnes et al., 2001). Cheng (2002) found that former welfare recipients who were married had an increased likelihood of achieving “financial independence” by transferring from welfare dependency to autonomy (i.e., employment with income above the poverty line and no reliance on AFDC/TANF benefits). An overwhelming majority of former SSI DA&A beneficiaries were single or never married. In a sample of 519 former SSI DA&A beneficiaries from Northern California, only 9.1% (n = 47) reported being married at the time of the policy change (Hogan, 2005). In the United States, single adults are more than three times as likely to experience poverty as married couples (US Census Bureau, 2010).

In previous research on welfare leavers, ethnicity has not been a consistent predictor of employment or self-sufficiency. Utilizing US Department of Labor data, better welfare-to-work outcomes were identified for white study participants than Hispanics or African Americans (Cheng, 2002). In an event history analysis examining the effects of welfare reform legislation on a national sample of AFDC and TANF recipients, Cheng found that Hispanic and African American study participants were significantly more likely to remain dependent on welfare following the 1996 policy change. Conversely, research conducted on female welfare recipients in Michigan and Illinois found no significant relationship between employment and minority status (Danziger, Corcoran, Danziger, Heflin, et al., 2000; Riger et al., 2004). In previous research on former SSI DA&A beneficiaries and employment and income, ethnicity was either not modeled in the analyses (Campbell et al., 2003) or was not a significant factor (Hogan et al., 2008). Nationally, the SSI DA&A population was 38.0% white, 43.2% African American, and 18.8% some other ethnicity (Wittenburg, Stapleton, Tucker, & Harwood, 2003).

Lastly, among demographic variables, children have been associated with poorer employment outcomes among both male and female welfare recipients (Blumenberg, 2002). Whether it was the number of children (Danziger & Lehman, 1996; Schmidt et al., 2007), the presence of younger children (Danziger, Corcoran, Danziger, Heflin, et al., 2000), or the need for child care (Brooks & Buckner, 1996; Dulmus et al., 2000), children and child care demands have represented barriers to transitioning from welfare to work for many welfare recipients. Due to the unique demographic and eligibility-defining characteristics of the TANF and SSI populations, child and child care issues were significantly more substantial for former TANF recipients than SSI DA&A beneficiaries following welfare reform (Campbell et al., 2003).

Education and Employment-Related Barriers

Education has been a very prominent variable in post-welfare-reform outcome studies. Many studies have identified having a high school diploma or general equivalency diploma as a significant predictor of positive employment outcomes (see Horwitz & Kerker, 2001; Nam, 2005; and Wickizer et al., 2000, as examples). In a rare study comparing gender-specific barriers to employment among California welfare recipients, Blumenberg (2002) found the lack of a high school diploma was more negatively associated with employment for women than men. She also found that limited English-language skills was negatively related to employment for female welfare recipients; however, limited English-language skills actually indicated positive employment outcomes for male welfare recipients. Blumenberg hypothesized that differences among men and women related to education and language as barriers to employment “could reflect the disparate occupational distribution of low-income men and women. Female welfare participants typically find employment in the service and retail trades, jobs that are perhaps more dependent than male-dominated jobs on English-language proficiency and high school education” (p. 319). While language was not a problem for most former SSI DA&A beneficiaries, a substantial proportion failed to graduate from high school (Hogan, 2005). While having a high school diploma or its equivalent has been shown to increase income among the current study sample (Hogan et al., 2008), the differential gender-specific effects of education on employment or self-sufficiency has not been examined.

Welfare recipients are more likely than non-welfare recipients to face employment barriers resulting from work-specific human capital deficits (e.g., limited work experience and inadequate job skills) (Seefeldt, 2004). More work experience and a greater number of occupational skills have been associated with an increased likelihood of stable employment (Riger et al., 2004) and self-sufficiency among AFDC and TANF recipients (Cheng, 2002). Female welfare recipients with no prior work experience were less likely to leave welfare (Schmidt et al., 2007) and more likely to return to TANF after leaving (Julnes et al., 2001). Female welfare recipients with more than a year of prior work experience were more than twice as likely to be employed as other female welfare recipients (Kalil et al., 2001); those with full-time work experience (employment for more than 30 hours per week in the last three years) were even more likely to be currently employed (Speiglman & Li, 2008). Fewer job skills is a barrier to employment among welfare leavers; however, participation in job skills training does not necessarily increase the likelihood of finding employment (Riger et al., 2004).

SSI beneficiaries receiving disability benefits are defined by their inability to work. Consequently, extensive work histories among this population are unexpected. Another important difference between the SSI DA&A population that lost benefits and other populations of welfare leavers is that their discontinuation of public assistance was extremely abrupt and unaccompanied by any formal vocational or job skills training. Despite such circumstances, in past research, many former SSI DA&A beneficiaries have not identified inadequate job skills as a barrier to employment. In a two-year multi-site study of 1,764 former SSI DA&A beneficiaries, only 31% of those who lost benefits and were unable to requalify for SSI under another disability category identified needing help learning job skills 24 months following the policy change (Campbell et al., 2003).

Health-Related Barriers

Poor physical and mental health has been consistently associated with poor employment outcomes among former welfare recipients. Former welfare beneficiaries self-reporting fair or poor physical health (Nam, 2005) or chronic health issues (Kalil, Schweingruber, & Seefeldt, 2001; Riger et al., 2004) have faced significant challenges toward acquiring employment. Mothers self-reporting good physical health were more than three times more likely to be employed than other female welfare beneficiaries reporting health problems (Horwitz & Kerker, 2001). Blumenberg (2002) found health problems an even greater impediment to employment among male than female welfare recipients in California. Regarding mental health, welfare recipients report more mental health problems than non-welfare recipients (Stromwall, 2002). Major depression and related symptoms have been associated with decreased likelihood of employment (Danziger, Corcoran, Danziger, Heflin, et al., 2000; Kalil et al., 2001); mental health problems, in general, have been identified as a barrier to self-sufficiency among AFDC and TANF recipients (Jayakody et al., 2000; Speiglman & Li, 2008). Former SSI DA&A beneficiaries have reported substantial amounts of both physical and chronic mental health problems (Hogan, Speiglman, & Norris, 2010; McKay, McLellan, Durell, Ruetsch, & Alterman, 1998; Watkins, Podus, & Lombardi, 2001); however, physical and mental health problems have not been consistent predictors of unemployment among this population (Campbell et al., 2003). Campell and associates found regional inconsistencies in the significance of physical and mental health problems as barriers to employment among a national sample of former SSI DA&A beneficiaries who lost income and health care benefits following welfare reform legislation. Also, contrary to intuitive rationale, treatment for mental health problems has not necessarily been associated with better employment outcomes for welfare recipients (Riger et al., 2004; Wickizer et al., 2000).

Substance abuse has not been consistently validated as a barrier to employment among welfare recipients. While many studies have significantly associated alcohol and drug dependence with reduced rates of employment and welfare-to-work (see Danziger, Corcoran, Danziger, Heflin, et al., 2000; Jayakody et al., 2000; Morgenstern & Blanchard, 2006; Nam, 2005; Speiglman & Li, 2008, as examples), a substantial number of studies have found no direct relationship between substance use, abuse, and employment (see Brooks & Buckner, 1996; Riger et al., 2004; Schmidt et al., 2007, as examples). Former SSI DA&A beneficiaries have demonstrated higher levels of drug use than other welfare recipients (Metsch & Pollack, 2005; Speiglman, Norris, Kappagoda, Green, & Martinovich, 2003; Substance Abuse and Mental Health Services Administration, 2002); and those expressing a need for substance abuse treatment have indicated more difficulty finding employment (Campbell et al., 2003).

Substance abuse treatment has been associated with better welfare-to-work outcomes among many welfare recipients. Metsch and colleagues (2003) found that both treatment completion and length of time in treatment were associated with increased employment among a sample of 4,236 female AFDC and TANF recipients in Florida. In a qualitative study of substance-abusing welfare recipients from New Jersey, nearly all 58 study participants (96% female) agreed that post-welfare self-sufficiency would be unattainable without first receiving some type of substance abuse treatment (Bush & Kraft, 2001).

A final important health characteristic identified in the “barriers to employment” literature is experiencing domestic or intimate partner violence. Nam (2005), using data from the Women’s Employment Study, found that domestic violence accelerated non-employment-based welfare exits, as well as the return of women to welfare after leaving. Female welfare recipients reporting recent (past year or less) intimate partner violence experiences had poorer employment outcomes than other female welfare recipients; however, a lifetime history of intimate partner violence was not associated with decreased work stability (Riger et al., 2004). This acute characteristic of the relationship between interpersonal violence and self-sufficiency was reinforced in later research by Speiglman and Li (2008); they found that domestic violence and “partner control” (i.e., partner interference in employment or job-training) were both strongly associated with lack of current employment among California mothers in TANF child-only cases; however, neither variable was a factor in longer-term employment measures. Domestic abuse has been identified as a barrier to employment among former SSI DA&A beneficiaries as well. Logistic analysis of domestic abuse and seven other barriers to substantial gainful activity among this population failed to identify domestic abuse as a statistically significant predictor of poor employment outcomes (Campbell et al., 2003); however, this sample was not differentiated by gender.

Transportation-Related Barriers

Throughout the literature, one of the strongest predictors of successful employment following a welfare episode was access to transportation. In pre-welfare-reform research on AFDC women in California, there was a significant association between automobile ownership, income, and employment. Among a sample of 1,112 study particpants, automobile ownership translated into an additional 23 hours of work (monthly) and $152 in average monthly earnings (Ong, 1996). A similar sample of AFDC recipients demonstrated that close residential proximity to public transit was important in stimulating employment; however, car ownership was significantly more beneficial for transitioning from welfare to work (Cervero et al., 2001). In this study, a welfare recipient’s probability of obtaining employment increased “by a factor of 13 when an individual’s status switched from not owning to owning a car” (p. 57). Conversely, losing car ownership resulted in a decreased likelihood of successfully transitioning from welfare to work (Cervero et al., 2001). Post-welfare-reform outcomes for TANF recipients have been similar (Danziger, Corcoran, Danziger, & Heflin, 2000). Travel by public transit, as opposed to one’s own automobile, has been shown to be negatively associated with employment outcomes among both male and female welfare recipients (Blumenberg, 2002).

Lack of access to transportation, defined as needing help finding or paying for adequate transportation, was also found to be a barrier to employment among former SSI DA&A beneficiaries (Campbell et al., 2003). With the exception of Los Angeles, Campbell and associates reported former SSI DA&A beneficiaries experiencing transportation challenges were 2.0 to 4.5 times less likely to achieve substantial gainful activity (i.e., earning more than $500) than other former SSI DA&A beneficiaries in Seattle, Washington; Portland, Oregon; Chicago, Illinois; Detroit, Michigan; and Stockton, California.

Social Support-Related Barriers

Finally, social support has been associated with better employment outcomes among welfare recipients (Brown & Riley, 2005; Horwitz & Kerker, 2001). Informal assistance in such areas as money, child care, and transportation has played a predominant role in some welfare leavers achieving self-sufficiency (Julnes et al., 2001). The number of people available to provide informal social support was associated with work stability among Illinois mothers receiving TANF (Riger et al., 2004); however, Brown and Riley (2005) emphasized the importance of quality over quantity when assessing the value of social support among low-income women seeking employment.

A substantial proportion of former SSI DA&A beneficiaries reported receiving no informal social support following welfare reform. In a sample of 412 former SSI DA&A beneficiaries from Northern California, 56% reported receiving no informal help (i.e., help from friends, family, or significant others) following the policy change (Hogan, 2005). The role that social support plays in employment, income, or self-sufficiency among this population has not been studied.

See Morris, Santhiveeran, and Lam (2006) for a content analysis of additional literature on employment and welfare recipients.

Methods

Study Sample

The study sample was selected from a population of SSI DA&A beneficiaries residing in four Northern California counties (Alameda, San Francisco, San Joaquin, and Santa Clara). Prior to elimination of the DA&A category, a sampling frame was developed by Maximus, the Social Security Administration’s regional substance abuse treatment referral and monitoring agency. The sampling frame constructed for this study by Maximus only contained SSI DA&A beneficiaries known to the agency; therefore, not every eligible SSI DA&A beneficiary in the four Northern California counties was available for selection. This is an important detail as these excluded individuals represented some of the most difficult and noncompliant SSI DA&A beneficiaries. Other criteria for inclusion in the study were age (between 21 and 59 years), not concurrently receiving SSDI benefits, and being an English or Spanish speaker. Study participants also could not be incarcerated or have a legal guardian at the time of the baseline interview.

Originally, 775 SSI DA&A beneficiaries were randomly selected from the Maximus sampling frame for the study. Eligible participants were initially contacted by the referral and monitoring agency. In a letter, Maximus explained the study and asked interested beneficiaries to call a toll-free number to schedule an interview. Eligible participants that did not respond to this invitation were contacted using information obtained from collateral sources, such as representative payees, families, and previous landlords. Ultimately, 519 (67.4%) completed baseline interviews.

Response rates for follow-up interviews were very good. The number of completed follow-up interviews at Waves 2, 3, 4, 5 and 6 were 494, 486, 479, 476, and 412, respectively. The 412 study participants interviewed at 42-month follow-up represent 87% of the study’s baseline participants not known to be deceased at Wave 6 (45 study participants were confirmed to have died over the lifetime of the study). From this panel of 412 study participants, former SSI DA&A beneficiaries that requalified for SSI benefits under another disability (physical or mental health) were removed (n = 193) resulting in a final study sample of 219 individuals who lost income and health care benefits as a result of the policy change. All results in the current study represent findings from these 219 former SSI DA&A beneficiaries.

Data Collection

For the collection of data, all study participants signed a consent form that explained the details of the research, the potential risks and benefits, the expectations of participating in the study, and the compensation arrangements. All study participants were informed that participation in the study was completely voluntary, that all information gathered was confidential and would not be shared with any Social Security Administration or county officials (except in aggregate form), and that no harm would come to a subject for participating (or not participating) in the research. Informed consent was not requested from anyone who appeared intoxicated or under the influence of drugs. Consequently, these individuals were not included in the study.

Once a former SSI DA&A beneficiary agreed to participate in the study, he or she was contacted by a trained field interviewer. Interviews were conducted in person, with administration of the survey lasting approximately 60–90 minutes. Study participants could be interviewed at a location of their choice. Following each interview, study participants were given $40. They were also asked to update contact information (i.e., phone number, address, and information regarding two friends or family members that could be contacted if the study participant was not reachable at their designated location). Following the baseline interview, attempts to contact study participants were made every three months to simply keep in touch or schedule a follow-up interview. Human subjects approval for collection of data was granted by the Institutional Review Board of the Public Health Institute (Oakland, California).

Instrumentation

The core survey instrument, originally developed for the “Multi-Site Study of the Termination of Supplemental Security Income Benefits for Drug Addicts and Alcoholics,” was comprised of 582 items taken from pre-existing survey instruments, such as the Addiction Severity Index (McLellan et al., 1992), the Global Appraisal of Individual Needs (Dennis et al., 1996), and the Community Childhood Hunger Identification Project food security questions (Wehler, Scott, & Anderson, 1996). The core survey covered a number of descriptive and functional domains, including demographic characteristics, alcohol and drug use, medical and psychiatric problems, mental health and substance abuse treatment, income and employment status, household and living arrangements, hunger and food acquisition, legal and criminal activity, and social service utilization. The 42-month follow-up (Wave 6) survey, developed specifically for the current study, contained some unique information related to work skills and welfare history, changes in family structure, hunger and coping for food, social support (i.e., help received from friends and family, as well as case managers and representative payees), and extent of involvement in a welfare-to-work program.

To test for reliability, an analysis of response coherence was conducted on baseline (Wave 1) data. High correlations were reported in the categories of SSI status (Pearson’s r = .94); legitimate employment and mental health (.95); casual work and physical health (.90); alcohol use (.88); and polysubstance use, including alcohol (.78). Lower coherence scores were produced relevant to drug use (.69), household size (.54), and reported money management (.16) (Swartz, Tonkin, & Baumohl, 2003, pp.111–112). Overall, the instrument had good internal consistency.

Variables

Self-sufficiency

A dichotomous variable for self-sufficiency was constructed based on a study participant’s past employment and level of average monthly income. A study participant was deemed to have had achieved the lowest level of self-sufficiency if 1) he or she indicated having worked at least one day in the month prior to the follow-up interview, and 2) he or she reported an average monthly income above that of the federal poverty line at the time of the follow-up interview. Study participants reporting receipt of public assistance in addition to employment were not excluded from self-sufficiency measures; receipt of TANF or GA (General Assistance) was controlled for in the analyses. The researchers felt this measure of self-sufficiency represented the most liberal interpretation of the construct.

Barriers

Descriptions of the mental health, substance abuse, and baseline income variables are detailed elsewhere (Hogan et al., 2008). Dichotomous variables (yes/no) were constructed for access to an automobile (i.e., transportation), a full-time work history of less than five years, any job training, any substance abuse treatment, any mental health treatment, any experience of interpersonal violence (past six months), and self-reported poor or fair physical health. As mentioned earlier, SSI beneficiaries are defined by their inability to work, so using work experience as a predictor variable could be problematic; however, nearly 40% of the current study sample reported positive work histories. No information was gathered regarding utilization of domestic violence or intimate partner violence (i.e., interpersonal violence) services. Social support was scaled using 12 yes/no questions related to receiving help with various tasks such as finding transportation, taking care of your children, learning job skills, and managing money matters (Cronbach’s alpha = 0.80 at baseline). Due to distribution issues, this variable was transformed using the square-root function.

Demographics

Demographic variables included age (in years), gender, ethnicity, marital status (married or not married), and education (high school graduate or equivalent, or less than high school). Ethnicity was categorized as Caucasian (reference), African American, Hispanic, and Other. In this study, former SSI DA&A beneficiaries averaged 2.1 children; however, the youngest child was already 10 years old at baseline. Subsequently, only seven study participants (3.2%) indicated child care as a barrier to employment at the time of the policy change (data not shown). As number of children and child care demands were not viable demographic barrier variables for the SSI DA&A population, they were excluded from analytical models.

Statistical Analysis

Generalized estimating equations were used to model self-sufficiency while accounting for the correlations between individual responses at the different waves using an exchangeable correlation matrix and robust standard errors. A logit link with binominal distribution was used for the models. Saturated and restricted models were estimated on the full sample and then separately for males and females. All models included the variables baseline income, education, age, ethnicity, gender, marital status, and other public assistance use. Other variables hypothesized to be related to self-sufficiency were included in the saturated models including physical health status, mental health symptoms, mental health treatment, alcohol and drug use, alcohol and drug treatment, interpersonal violence, job training, work history, transportation, and social support. Reduced models included only these other variables that had p-values less than 0.10. All statistical analyses were conducted using Stata 10.1 (StataCorp, 2008).

Results

Of the 219 study participants, 74.9% were of minority status and 42.0% had not graduated high school; the average age of the sample was 42.0 years (see Table 1). Other characteristics included, 79.5% reported at least one serious mental health symptom, 63.3% reported poor or fair physical health, 31.4% reported current illicit drug use, and one-third reported experiencing some type of interpersonal violence during the lifetime of the study. With regard to gender, women were more likely than men to have received some other type of public assistance (i.e., TANF or GA), had access to an automobile, received some type of substance abuse treatment or job training, and had more income at baseline; men were more likely than women to have had a full-time work history of five years or more (see Table 2). Table 3 displays rates of self-sufficiency following the policy change. Rates of self-sufficiency were initially low (7.3% at 6-month follow-up) but increased by the end of the study (21.5% at 42-month follow-up). There were no significant differences in rates of self-sufficiency among male and female study participants at any follow-up interview.

Table 1.

Baseline Demographics by Gender.

	Men	Women	Full Sample
Demographic Variables	n = 120	n = 99	n = 219
Age (in years) (SD)***	43.76 (7.57)	39.85 (7.71)	41.98 (7.86)
Ethnicity (%)
African American	42.5	38.4	40.6
Caucasian	23.3	27.3	25.1
Hispanic	25.0	20.2	22.8
Other	9.2	14.1	11.4
Marital status (%)
Married	13.2	8.3	11.0
Education (%)
High school graduate or GED	63.3	51.5	58.0
Income
Average monthly income in $100s (SD)*	8.27 (5.08)	9.59 (4.48)	8.87 (4.85)
Received other public assistance (Wave 2) (%)**	31.7	49.5	39.7

Comparisons made using Pearson’s chi-square and independent-samples t tests (two-tailed).

^*p < 0.05;

^**p < 0.01:

^***p < 0.001.

Table 2.

Barriers to Self-Sufficiency by Gender at 6-Month Follow-Up.

	Men	Women	Full Sample
Health-Related Barriers (%)	n = 120	n = 99	n = 219
Fair or poor physical health	60.5	66.7	63.3
Any mental health symptoms	76.3	83.3	79.5
Any mental health treatment	23.7	35.4	29.1
Heavy alcohol use	21.1	12.5	17.1
Current illicit drug use	33.3	29.2	31.4
Any substance abuse treatment*	21.9	34.4	27.6
Experienced interpersonal violence	28.1	39.6	33.3
Other Barriers (%)
No job training*	93.0	84.4	89.0
Limited work experience*	53.3	68.7	60.3
No access to transportation*	78.1	64.6	71.9
Limited social support (SD)	0.64 (0.79)	0.84 (0.92)	0.73 (0.86)

Comparisons made using Pearson’s chi-square and independent-samples t tests (two-tailed).

^*p < 0.05

Table 3.

Self-Sufficiency Percentages by Gender at Follow-Up, Waves 2–6.

	Men	Women	Full Sample
Follow-Up Interview	n = 120	n = 99	n = 219
Wave 2, 6-month follow-up	6.7	8.1	7.3
Wave 3, 12-month follow-up	13.3	16.2	14.6
Wave 4, 18-month follow-up	11.7	18.2	14.6
Wave 5, 24-month follow-up	17.5	15.2	16.4
Wave 6, 42-month follow-up	20.0	23.2	21.5

Comparisons made using Pearson’s chi-square.

No significant differences in rates of self-sufficiency by gender over the lifetime of the study.

The first models predict the odds of achieving self-sufficiency for the entire study sample (see Table 4). Overall, the odds of former SSI DA&A beneficiaries achieving self-sufficiency increased over time. In the restricted full-sample model the odds of being self-sufficient increased by 1.35 each year following the baseline interview, when controlling for other demographic and individual characteristics. Drug and heavy alcohol use, mental health treatment, job training, interpersonal violence and social support were not associated with a change in the odds of being self-sufficient. When examining outcomes for the entire study sample, demographic characteristics were generally not associated with an increased or decreased likelihood of achieving self-sufficiency. An exception to this was age. Each additional year of age among former SSI DA&A beneficiaries reduced the odds of being self-sufficient by five percent. Having more income prior to the policy change was also associated with better self-sufficiency outcomes.

Table 4.

Odds Ratios of Achieving Self-Sufficiency, Full-Sample and Gender Models.

	Full Sample		Women		Men
Variables [95% CI]	Saturated Model	Restricted Model	Saturated Model	Restricted Model	Saturated Model	Restricted Model
Years from Baseline	1.39*** [1.18, 1.62]	1.39*** [1.18, 1.58]	1.50** [1.13, 1.98]	1.50** [1.14, 1.97]	1.37** [1.11, 1.69]	1.36** [1.10, 1.68]
Age (in years)	0.95** [0.92, 0.99]	0.95** [0.92, 0.99]	0.95 [0.89, 1.02]	0.94* [0.88, 1.00]	0.95* [0.90, 0.99]	0.95* [0.90, 0.99]
African American	0.56 [0.28, 1.10]	0.58 [0.29, 1.15]	1.29 [0.38, 4.31]	1.16 [0.38, 3.57]	0.32* [0.13, 0.79]	0.33* [0.13, 0.83]
Hispanic	0.63 [0.31, 1.32]	0.71 [0.35, 1.45]	1.15 [0.28, 4.70]	1.07 [0.32, 3.59]	0.44 [0.16, 1.16]	0.49 [0.19, 1.29]
Other	0.15** [0.05, 0.48]	0.15** [0.05, 0.47]	0.08 [0.01, 1.06]	0.07* [0.01, 0.89]	0.24* [0.08, 0.74]	0.23* [0.07, 0.74]
Married	0.96 [0.50, 1.83]	0.93 [0.49, 1.78]	1.80 [0.60, 5.39]	1.54 [0.54, 4.37]	0.46 [0.08, 0.74]	0.51 [0.07, 0.74]
High school graduate	1.51 [0.82, 2.78]	1.49 [0.84, 2.63]	3.70* [1.30, 10.51]	3.50* [1.29, 9.49]	0.70 [0.35, 1.38]	0.70 [0.36, 1.38]
Baseline income	1.08** [1.03, 1.14]	1.08** [1.03, 1.12]	1.11 [0.99, 1.24]	1.10 [0.99, 1.23]	1.05 [1.00, 1.11]	1.04 [1.00, 1.10]
Other public assistance	0.51* [0.26, 0.99]	0.55 [0.29, 1.05]	0.55 [0.19, 1.59]	0.56 [0.21, 1.53]	0.45 [0.17, 1.21]	0.56 [0.22, 1.42]
Poor physical health	0.57** [0.37, 0.86]	0.56** [0.37, 0.86]	0.56 [0.30, 1.05]	0.54 [0.29, 1.00]	0.57* [0.32, 0.99]	0.60 [0.35, 1.02]
Any mental health symptoms	0.45** [0.28, 0.73]	0.40*** [0.26, 0.61]	0.52 [0.20, 1.31]		0.35*** [0.19, 0.65]	0.35*** [0.21, 0.59]
Any mental health treatment	1.07 [0.63, 1.84]		1.02 [0.47, 2.19]		1.21 [0.52, 2.83]
Heavy alcohol use	1.38 [0.76, 2.49]		2.18 [0.76, 6.27]		1.42 [0.70, 2.90]
Current drug use	1.52 [0.81, 2.86]		1.38 [0.51, 3.71]		1.50 [0.60, 3.74]
Any AOD treatment	0.45** [0.25, 0.79]	0.42** [0.23, 0.76]	0.39* [0.15, 1.00]	0.33* [0.11, 1.00]	0.57 [0.28, 1.19]	0.58 [0.28, 1.19]
Interpersonal violence	0.62 [0.35, 1.07]		0.39* [0.17, 0.88]	0.38** [0.19, 0.78]	0.68 [0.31, 1.49]
Job training	1.65 [0.89, 3.04]		1.87 [0.83, 4.21]		1.41 [0.60, 3.33]
More work experience	2.48** [1.33, 4.64]	2.58** [1.40, 4.76]	2.60 [0.76, 8.93]	2.24 [0.76, 6.57]	2.85** [1.31, 6.21]	2.75* [1.22, 6.20]
Transportation	3.27*** [2.10, 5.11]	3.35*** [2.19, 5.14]	3.43*** [1.77, 6.63]	3.42*** [1.72, 6.81]	3.70*** [1.98, 6.91]	3.54*** [1.91, 6.56]
Social support	0.82 [0.64, 1.05]		1.06 [0.78, 1.45]		0.70 [0.46, 1.06]	0.68 [ 0.45, 1.04]

^*p < 0.05;

^**p < 0.01;

^***p < 0.001.

Reference group: Caucasian (ethnicity).

For the full sample, mental health symptoms and poor to fair physical health were associated with sharp reductions in the odds of achieving self-sufficiency (OR = 0.40 and 0.56, respectively), as was substance abuse treatment (OR = 0.42). Not surprisingly, a stronger work history was associated with better self-sufficiency outcomes (OR = 2.58). For the full sample of former SSI DA&A beneficiaries, automobile availability was the strongest predictor of achieving self-sufficiency (OR = 3.35).

The same overall model was estimated separately for men and women. Both groups reported higher odds of self-sufficiency as time progressed following the policy change. Over the lifetime of the study, the odds of achieving self-sufficiency for men and women increased annually by 1.36 and 1.50, respectively. The interaction between gender and time was not statistically significant. Lack of transportation and age also remained significant barriers to self-sufficiency when differentiated by gender; health status (i.e., reporting fair or poor physical health) became marginally significant (p < .10) for each group.

For women, a high school diploma or equivalent resulted in significantly higher odds of obtaining self-sufficiency (OR = 3.50); this was not true for men. In this study, a high school diploma or equivalent actually decreased the odds of self-sufficiency for men (OR = 0.70), although this result was not statistically significant. The interaction between gender and high school graduation in the full-sample restricted model was statistically significant (p = 0.03; data not shown). For men, reporting at least one serious mental health symptom resulted in lower odds of achieving self-sufficiency; for women, attending an alcohol or drug treatment program significantly decreased the likelihood of achieving a positive outcome. In the gender-specific models, work history proved to be more important for men; women had significantly lower odds of obtaining self-sufficiency if they reported experiencing interpersonal violence (OR = 0.37).

Discussion

The SSI DA&A population is a historically unique group of former public assistance recipients; however, their post-welfare-reform experiences are, to some degree, similar to other populations of welfare leavers. Obviously, the SSI DA&A and TANF populations are qualitatively different in many respects; the SSI DA&A population is predominantly male, older, and has relatively few children (Wittenburg et al., 2003), while the TANF population is comprised of 90% women, all of whom have children (Administration for Children & Families, 2009). Despite these differences, the two groups have reported similar employment and quality-of-life experiences following welfare reform. For both groups, welfare reform resulted in decreased welfare dependence and increased employment; however, the number of former SSI and TANF beneficiaries reporting less poverty, more income, or better life circumstances remains minimal (Acs & Loprest, 2007; Campbell, Baumohl, & Hunt, 2003; Hogan et al., 2008; Johnson, Rector, & Abramovitz, 2007). Can any of the findings from the current study of SSI DA&A beneficiaries be generalized to other public assistance recipients?

Interestingly, one segment of the TANF population that begins to converge toward the SSI DA&A population on a number of characteristics is the hard-to-serve, or hardship, group of TANF recipients who are either nearing their cumulative lifetime benefit limit or have been exempted from welfare-to-work activities by their respective states. For example, in a longitudinal study (3.5 years) of TANF recipients in New Mexico following welfare reform, 22% of the study sample failed to leave welfare during the lifetime of the study (Richardson, Schoenfeld, LaFever, & Larsh, 2002). This group was characterized by the researchers as more difficult to serve than other study participants. Characteristics of this group included 40% did not complete high school and 49% were never married; among the current study sample of former SSI DA&A beneficiaries, at baseline, 42% did not complete high school and 47% were never married (data not shown). Many of the New Mexico TANF study particpants still on public assistance reported limited work histories; for study participants aged 40 and older, 45% reported never working or not working in the past two years (Richardson et al., 2002). Among the current sample of former SSI DA&A beneficiaries, approximately 60% reported working full-time for less than five years.

Some health areas were also similar between the two groups of welfare leavers. Among New Mexico TANF study particpants who were still on welfare and not working 3.5 years since beginning the study, 23% stated that the most important factor in not working was physical health problems; 74% considered the condition permanent (Richardson et al., 2002). Among the current sample of former SSI DA&A beneficiaries, 22% self-reported poor physical health at the 42-month follow-up (data not shown). Among New Mexico TANF study particpants still on public assistance, 44% reported being depressed for most of the past year and 30% aged 35 and older reported being treated for depression; among the current study sample of former SSI DA&A beneficiaries, 44% reported experiencing serious depression in the past six months (42-month follow-up; data not shown) and 29% reported receiving mental health treatment. So, while the differences between SSI and TANF recipients in general are great in a number of areas, we can see some similarities between the SSI DA&A population and hardship TANF cases, particularly older TANF participants who have been on public assistance for some time and have been characterized as more difficult to serve. The following discussion focuses on the current study’s findings for SSI DA&A beneficiaries as well as some relationships to current conditions for TANF recipients. Possibly, some of these findings can be generalized to TANF hardship cases populated by older recipients.

General Barriers to Self-Sufficiency

Understanding specific barriers to employment and self-sufficiency among welfare recipients transitioning from welfare to work is important for policymakers and program administrators. It is critical in a fiscal environment characterized by conservative spending and selective provision allocation that social welfare service structures provide welfare leavers with the most effective programs and services possible. The most significant barrier to self-sufficiency for both men and women identified in this study was access to an automobile. Public assistance programs do make allowances for transportation and the value of a private vehicle; however, most vehicle exclusion policies are inadequate and insufficient. In 2005, the federal government, to its credit, did remove all restrictions on the value of an automobile owned by SSI beneficiaries (Social Security Administration, 2010). Unfortunately, states have not followed suit for TANF recipients. In California, the TANF program (California Work Opportunity and Responsibility to Kids) currently allows recipients to own a vehicle worth not more than $4,650 (California Department of Social Services, 2009). Sadly, this is the same vehicle exclusion limit mandated by the original 1997 welfare reform legislation in that state (California Assembly Bill 1542). Since the inception of welfare reform in California, state policy related to welfare recipients and private transportation has neither been amended nor adjusted. Given the evidence-based importance of private mobility for welfare recipients to successfully transition from welfare to work, this appears to be neglectful policymaking. Basing vehicle exclusion limits on fair market values from nearly 15 years ago decreases the probability of welfare recipients having access to sound and reliable transportation. Allowing for ownership of a reliable vehicle should be a policy goal rather than an indication of some type of welfare fraud or malfeasance on the part of recipients. Parity for SSI and TANF beneficiaries on this eligibility criterion would be optimal; however, simply raising the value limit of personal transportation for TANF recipients to a realistic level would improve employment outcomes among welfare leavers. More intensive programs designed for welfare recipients to acquire and maintain private transportation, such as no-interest automobile loans, government-based transportation subsidies, automobile maintenance training, and subsidized or low-cost car insurance would also be beneficial.

Former SSI DA&A beneficiaries, by definition, should have been barred from employment by their alcohol and drug disabilities; however, in the models tested, substance abuse (i.e., heavy alcohol and illicit drug use) was not a barrier to self-sufficiency for this population. Programmatically, another interesting outcome was that neither substance abuse nor mental health treatment were beneficial in achieving self-sufficiency, despite high levels of self-reported alcohol, drug use, and mental health problems. This may be an indication of ineffective services, or it may represent the fragile nature and severity of health issues among the study population. Former SSI DA&A beneficiaries who failed to requalify for SSI benefits under another disability category and yet required treatment services may have had more difficulty finding employment as a result of the severity of their illnesses. This would help explain women who participated in substance abuse treatment having a significantly decreased likelihood of achieving self-sufficiency. Continuous measures of health severity, as predictors of welfare-to-work outcomes, may be an area of future research among welfare recipients.

Another non-factor in achieving self-sufficiency among the current study population was job skills training; however, as a welfare-to-work service, these findings are inconclusive. In eliminating drug addiction and alcoholism as qualifying impairments for disability benefits in the SSI program, Congress never legislated for the provision of any job skills training for former SSI DA&A beneficiaries. As a result, those who lost benefits were forced to find employment with the job skills they had; less than 11% of the current study population reported any job skills training up to 3.5 years following the policy change. Typically, welfare-to-work programs designed to promote self-sufficiency do provide opportunities for job skills training and job search assistance; this was not an option for the SSI DA&A population.

Gender-Specific Barriers to Self-Sufficiency

Some barriers to successfully transitioning from public assistance to self-sufficiency are different for men and women. The most marked among these is education. A high school diploma or equivalent is associated with significantly greater likelihood of employment for female welfare recipients; it is essentially of no consequence for most men transitioning from welfare to work. This outcome for former SSI DA&A beneficiaries replicates Blumenberg’s findings (2002) on gender, education, employment and TANF recipients in California. The impact of education on employment may be a consequence of the type of employment welfare recipients acquire after leaving public assistance. Most welfare leavers acquire low-wage jobs in service industries (Brauner & Loprest, 1999). This is true for both men and women (Blumenberg, 2002). However, following the service industries, women tend to work in sales, including restaurants and bars, clerical, and health-related occupations; men tend to gravitate toward more labor-intensive industries, such as construction, agriculture, maintenance and repair. These secondary occupations for women may require more education than those occupied by men. Conversely, work experience benefited men more than women. While over 30% of women in the study sample reported full-time work histories of five years or more, such experience failed to translate to higher levels of self-sufficiency. The reason for this is unclear.

Another gender-specific barrier for SSI DA&A women that differentially affects achieving self-sufficiency is experiencing recent interpersonal violence. This finding is also consistent with previous research on TANF recipients in California (Speiglman & Li, 2008). It is not surprising that acute trauma and interference from abusive intimate partners would hinder women from leaving welfare and acquiring employment. From a programmatic perspective, these gender-specific findings have case management and service provision implications for men and women transitioning out of welfare.

Conclusion

Limitations and Strengths

There were some limitations to this study. Causal ambiguity was a concern for some relationships. For example, the longitudinal analysis could not differentiate specifically when access to an automobile occurred in relation to acquiring employment. While automobile availability was highly predictive of self-sufficiency in the models tested, controls for when each event occurred were not employed. Measurement fidelity was another minor concern of researchers. The outcome variable, self-sufficiency, was constructed from self-reported employment and income data. Any inconsistencies or inaccuracies in self-reported data could compromise the validity of this construct. Finally, despite making a case for similarities between SSI DA&A and hardship TANF cases populated by older recipients, generalizing results from this study to other public assistance beneficiaries should be made with caution.

This study did have some strengths, also. The longitudinal design and comprehensive survey instrument allowed for strong analyses of relationships between variables. Such a high response rate over six data points for such a vulnerable and disadvantaged study population is fairly remarkable. The high representation of male study particpants was another unique advantage of this study. Few studies on former public assistance recipients have had such a substantial male sample size allowing for such strong analytical gender comparisons. Lastly, this study builds on previous research examining income (Hogan et al., 2008) and employment (Campbell et al., 2003) outcomes among former SSI DA&A beneficiaries. Hogan and associates (2008) demonstrated that a significant proportion of former SSI DA&A beneficiaries who failed to requalify for SSI benefits under another disability category suffered increased economic hardship following the policy change; however, they failed to answer the question as to why? Campbell and others (2003) examined barriers and employment outcomes among former SSI DA&A beneficiaries up to 24 months following the policy change. The current study combines income and employment into a single outcome variable (self-sufficiency), extends the lifetime of the study to 42 months, and improves on previous statistical models to provide greater insight into the post-welfare-reform experiences of the SSI DA&A population.

Implications for Policy and Welfare-To-Work Programs

While women constitute an overwhelming majority of adult public assistance recipients and, consequently, welfare leavers, approximately 10% of this population is male (Administration for Children & Families, 2009). This represents a significant number of potential welfare leavers who may face unique barriers to employment and self-sufficiency. Designing welfare-to-work programs and services that efficiently optimize the potential of low-income individuals to acquire employment and achieve self-sufficiency is the goal of social welfare administrators and policymakers. Examining gender-specific barriers to these goals indicates that individually-based interventions intended to increase the human capital of welfare recipients, such as education, job skills training, and treatment services, may have limited effectiveness for low-income men transitioning from welfare to work. Structural or systemic interventions, such as improved access to transportation, child care, and workfare (i.e., public job creation) would be more beneficial for this population. While the American Recovery and Reinvestment Act of 2009 does have a provision for subsidized employment for TANF recipients (Section 2101), this provision only scratches the surface of creating greater access to employment among welfare recipients. A program of publicly created jobs giving priority to welfare recipients would benefit all public assistance beneficiaries but particularly those recipients who are not aided by education or generic job skills training services. Federally subsidized work programs could also benefit SSI recipients desiring to work. This strategy has historical precedence during times of high unemployment and economic depression; a federal job creation program dedicated to placing public assistance recipients in meaningful and adequately paying employment would significantly ameliorate the stress created by inadequate public assistance benefits and punitive social welfare policies, particularly for older and hard-to-serve public assistance recipients transitioning out of the welfare system toward personal responsibility and self-sufficiency.