The views expressed are those of the author and do not necessarily reflect those of the Urban Institute, its board, its sponsors, or other authors in the series.
Prepared for the Office of the Assistant Secretary for Planning and Evaluation and the Administration for Children and Families, Contract No. 100-95-0021 Task Order No. 6. The authors would like to thank Robert Lerman, Nancy Pindus and staff from U.S. DHHS for their insightful comments. They would also like to thank Daryl Dyer for his assistance in preparing the final report. Any opinions expressed in this report are those of the authors and do not necessarily represent the views of the Urban Institute or the U.S. Department of Health and Human Services.
Table of Contents
Executive Summary
Personal and Family Challenges to the Successful Transition From Welfare to Work
OVERVIEW
Time limits, increasing employment and training participation requirements, and strict economic sanctions for non-compliance with program expectations, all increase the importance of engaging a broader share of the AFDC caseload in employment or employment-related activities. Over time, these policy changes will require states and local welfare offices to develop strategies to engage welfare recipients who have traditionally been exempted from participation in education or training activities, as well as other long-term recipients, in welfare-to-work programs. This paper is the first of a series of papers that will examine the magnitude of this challenge and identify potential options for helping this group of families find and maintain employment.
In an attempt to identify the percentage of the caseload that is likely to need more assistance than traditional welfare-to-work programs generally provide, this paper examines the following questions:
(1) What are the types of personal and family challenges that may hinder a welfare recipient's movement toward productive and successful employment?
(2) What do we know about the prevalence of each of these potential barriers to employment among the welfare population?
(3) When the presence of multiple barriers is taken into account, what fraction of the current AFDC caseload is likely to experience at least one potential barrier to employment?
(4) What do we know about the relationship between the presence of these potential barriers to employment and actual work experience?
(5) What are the implications of these findings for the design and deliver of welfare-to-work programs?
To provide a comprehensive picture, we utilize existing literature, such as evaluations of welfare-to-work demonstration programs, informal surveys of direct service providers, analyses of large-scale national surveys, and case record reviews, which we supplement with a primary analysis of the National Longitudinal Survey of Youth (NLSY). Later papers will examine the characteristics and structure of programs already addressing these issues, develop a set of alternative intervention models to promote sustained employment among those who would otherwise fail to achieve employment goals, and design a multi-site demonstration project to test promising intervention strategies.
KEY FINDINGS
Welfare recipients, like many non-welfare families, experience a broad range of family and personal issues that make employment difficult.
Through an analysis of the literature, eight major personal and family challenges that may affect a recipient's transition from welfare to work were identified. These challenges include:
- physical disabilities and/or health limitations;
- mental health problems;
- health or behavioral problems of children;
- substance abuse;
- domestic violence;
- involvement with the child welfare system;
- housing instability, and;
- low basic skills and learning disabilities.
There is a wide range of estimates on the prevalence of specific potential barriers to employment among the welfare population.
Because there has been minimal attention to the role these issues play in the lives of welfare recipients, there is little documentation on the prevalence of these issues among the welfare population. The literature that does exist lacks standard definitions of the challenges identified above, resulting in a very wide range of estimates on the prevalence of these issues. For example, studies examining the prevalence of health conditions that limit the amount or kind of work a person can do range from 16.6 to 30.8 percent of the AFDC caseload. Similarly, studies examining the prevalence of substance abuse report estimates ranging from 6.6 to 37 percent of the AFDC caseload.
Most welfare recipients experience at least one potential barrier to employment, with low basic skills being the most common.
Analysis of data from the NLSY reveals that almost 90 percent of current recipients between the age of 27 and 35 experience one of five potential barriers to employment (low basic skills, substance abuse, a health limitation, depression, or a child with a chronic medical condition or serious disability). About half of all recipients experience a more serious form of one or more of these barriers (i.e., depression five to seven days a week; a health limitation that prevents work; concern that one is an alcoholic; repeated use of crack or cocaine; or extremely low basic skills). Low basic skills are much more common than the other potential barriers to employment included in this analysis. About one-third of recipients score in the bottom decile of the women's distribution of the Armed Forces Qualifying Test (AFQT), a test of basic skills that is highly correlated with employment and earnings, and another one-third score in the 10th to 25th percentiles. In comparison, 10 percent of recipients report not seeking work because of a medical problem; 13 percent report being depressed between five to seven days a week; five percent report a serious drinking problem and nine percent report heavy cocaine or crack use.
The majority of recipients who experience potential barriers to employment work, but do so intermittently.
In general, recent employment was almost as common among recipients who report potential barriers to employment as among those who do not report any barriers, making the existence of a potential barrier to employment a relatively poor indicator of whether a recipient will need additional assistance to find employment. The one exception to this pattern is for recipients with extremely low basic skills. Only 44 percent of recipients who scored in the bottom decile of the women's AFQT distribution reported working in the current or previous year, compared to 68 percent who scored above this level. About 57 percent of recipients reporting a potentially serious barrier, excluding extremely low skills, reported working, a percentage only slightly lower than the 62 percent of recipients who did not report such barriers.
While employment is common among welfare recipients who face most potential barriers to employment, continuous employment is not. Only eleven percent of recipients reporting a serious barrier to employment who worked reported being employed for a full year, compared to 27 percent of those who did not report such a barrier.
The one-quarter of the AFDC caseload who report a potentially serious barrier to employment and have no recent attachment to the labor force are likely to need more assistance than most traditional welfare-to-work programs provide to succeed in the labor market. Almost twice as many may need additional assistance to maintain steady employment.
While many recipients with serious barriers to employment work, albeit intermittently, a modest fraction have no recent connection to the labor market. This group of recipients, who account for about 25 percent of the AFDC caseload, are likely to need more assistance than traditional welfare-to-work programs generally provide to succeed in the labor market. When one accounts for the fact that the vast majority of recipients with potentially serious barriers to employment who work do so only intermittently, the percentage of the caseload likely to need additional assistance to fare well on their own over the long-term, especially within a time-limited welfare system, increases dramatically —to 51 percent.
Introduction
Over the last quarter-century, welfare policy has shifted from emphasizing income security to stressing work and ultimately self-sufficiency. In 1967, the Work Incentive Program formally introduced the idea that single mothers who did not have responsibility for caring for pre-school children should be expected to work. The Family Support Act of 1988 (FSA) required states to actively engage an increasing proportion of the AFDC caseload in employment or education and training activities through the Job Opportunities and Basic Skills (JOBS) training program. Changes already taking place in many states under federal waivers and proposed national reforms advance these efforts further.
Time limits, increasing employment and training participation requirements and strict economic sanctions for non-compliance with program expectations, all increase the importance of engaging a broader share of the AFDC caseload in employment or employment-related activities. Over time, these policy changes will require states and local welfare offices to develop strategies to engage welfare recipients who traditionally have been exempted from participation, as well as other long-term recipients, in welfare-to-work programs.1 Because states have only been required to serve a modest percentage of all recipients, limited knowledge exists regarding the potential of traditionally exempt recipients to become employed, earn sufficient wages, and sustain employment in order to exit welfare permanently. When faced with stringent work requirements, some families will likely seek and maintain employment on their own. Others will be ready to work but unable to find jobs at their current skill level. These families may need extra training, education or work experience in order to compete in an increasingly competitive labor market. Still others will need substantial assistance to make the transition from welfare to work.
Families on AFDC, like other families, face a variety of circumstances which make employment difficult. Transportation and child care have already been acknowledged by the Family Support Act as logistical barriers which prevent some welfare recipients from working. The low wages that welfare recipients can command in the labor market make it difficult for many former recipients to cover these additional expenses associated with working. While there is room for improvement in the availability and the design and delivery of services to deal with these two problems, most states at least have established procedures for addressing them. In contrast, very few states have examined how other personal and family challenges may affect the transition from welfare to work. Nor have they identified the types of supports families who confront these challenges may need to succeed without ongoing assistance from the government.
This paper is the first of a series that will address these issues. This paper answers the following five questions about the challenges welfare to work programs will need to address as an increasing share of the AFDC caseload becomes subject to time limits and more stringent work requirements:
(1) What are the types of personal and family challenges that may hinder a welfare recipient's movement toward productive and successful employment?
(2) What do we know about the prevalence of each of these potential barriers to employment among the welfare population?
(3) When the presence of multiple barriers is taken into account, what fraction of the current AFDC caseload is likely to experience at least one potential barrier to employment?
(4) What do we know about the relationship between the presence of these potential barriers and actual work experience?
(5) What are the implications of these findings for the design and delivery of welfare- to-work programs?
To provide a comprehensive picture we utilize existing literature, such as evaluations of welfare-to-work demonstration programs, informal surveys of direct service providers, analyses of large-scale national surveys, and case record reviews, which we supplement with a primary analysis of the National Longitudinal Survey of Youth (NLSY). Later papers will examine the characteristics and structure of programs already addressing these issues, develop a set of alternative intervention models to promote sustained employment among those who would otherwise fail to achieve employment goals, and design a multi-site demonstration project to test promising intervention strategies.
I. Personal and Family Problems That May Hinder the Transition to Employment
Case managers working in welfare-to-work programs identify a variety of personal and family challenges that often make it difficult for families to leave welfare for work, including physical and mental health conditions, child health or behavioral problems, chemical dependency, family violence, housing instability, and low basic skills or learning disabilities.
These challenges may affect the transition to employment in a variety of ways. For example, a welfare recipient whose child has a chronic health problem may have the skills to succeed in the labor market, but may be unable to find employment that has enough sick leave or flexibility to allow her to attend to her child's special needs. Alternatively, the fear of leaving a child with special needs in a new environment may create more stress for the recipient than she is capable of handling, making it impossible for her to take even the smallest steps towards self-sufficiency without constant encouragement and support.
Mental health problems or substance use problems may prevent recipients from being able to undertake the tasks necessary to find employment or recipients dealing with these issues may lack the self-confidence they need to take on new challenges. Others may be able to find employment, but may not be able to sustain it over the long term. In their most severe forms, these problems may be so debilitating that it is impossible for a recipient to search for employment or participate in an education or training program until medical treatment is obtained.
Physical disabilities and/or health limitations may affect welfare recipients' ability to work in several ways. Some recipients face health limitations so severe that they cannot even seek employment, while others experience restrictions on the kind or amount of work they can do. Still other recipients may be able and willing to work, but may be kept from doing so because they are especially dependent on the Medicaid coverage they receive by virtue of being a welfare recipient to cover their ongoing and high medical expenses.
Housing instability highlights the important role environmental factors may play in a recipient's ability to make an effective transition from welfare to work. Recipients who are homeless may find it impossible to search for housing and sustain employment at the same time. Others may be hampered by not having a telephone number or stable address where an employer can reach them. Still others may be so demoralized by their current living situation that they cannot muster the personal resources needed to sustain an ongoing search for employment.
Given the increasing recognition of the role family violence plays in the lives of recipients, it is important to view the transition from welfare to work as a transition for families rather than just as a transition for recipients. When a parent goes to work, children must adjust to a new routine that often reduces their time with their parents. Similarly, male partners, friends and relatives must adjust to new time constraints as well as substantial shifts in existing roles, responsibilities and relationships. To succeed over the long-term, these families are likely to need far more ongoing family-centered support than traditional welfare-to-work programs generally provide.
According to case managers, low skills affect the transition to employment in obvious and not-so-obvious ways. Recipients with low skills generally can only qualify for jobs that offer low pay, irregular hours, and few benefits. As a result, welfare is often a better option financially. But for some recipients, low skills or the presence of a learning disability is synonymous with a life of failure that started long before they first received welfare. The end result is very low self-esteem and an overwhelming fear of change. It is extremely difficult for recipients who are certain they will fail at any attempts to make it on their own to take the first step towards self-sufficiency.
It is difficult to define and measure the prevalence of these circumstances and their impact on employment. Whether a particular issue prevents, limits or does not affect an individual's employment potential depends upon the interplay of this issue with her other personal characteristics and life circumstances. While across a population, one might expect the existence of such an issue to decrease the probability of working and/or the expected wage rate, it will not represent a barrier for all individuals. For example, a woman with a special-needs child who receives support from nearby extended family may able to work while a similar woman who has few friends or relatives on which to rely cannot work. Some potential barriers that carry a negative stigma, such as chemical dependency or mental health problems, are likely to be under reported, especially if the prevalence is based on self-reports. In contrast, the extent to which respondents report that health problems keep them out of the labor market may be over reported, especially if the respondent has not worked in some time and has only limited knowledge of the current labor market. Finally, a number of circumstances that may affect employment cannot be adequately measured by large surveys. Depression, low self-esteem and the existence of domestic violence and its psychological impacts are a few examples of issues that welfare recipients may not reveal in a two-hour survey but that anecdotal evidence suggests are highly prevalent and can discourage self-sufficiency.
II. How Prevalent Are These Potential Barriers to Employment?
Given methodological problems mentioned above and the minor role welfare-to-work programs have played in the lives of welfare families, only a modest amount of research is available on the special challenges welfare recipients and their families must overcome to make a successful transition from welfare to work. Most of the available information focuses on those challenges that are easiest to measure (health limitations) or are more controversial (chemical dependence). This section reviews the available evidence on the prevalence of the potential barriers to employment identified above, noting gaps or inconsistencies in the available studies where appropriate. A summary of the major studies included in this section can be found in Appendix A.
A. Physical Disabilities and/or Health Limitations
Reports on several programs and demonstration projects note that health limitations are a significant barrier to employment for some welfare recipients. The Family Development and Self-Sufficiency (FaDSS) program in Iowa states that approximately 12 percent of their caseload is chronically and physically ill (Iowa Department of Human Rights, 1996). In the Riverside County GAIN program in California, "health problems" were the reason 41 percent of deferred participants were exempted (Maynard, 1994). Program staff from New Chance, a welfare-to-work program for teenage dropouts, reported that mothers’ illnesses and that of their children constitute one of the most frequently given reasons for absenteeism, in part because the young women do not always know how to assess the seriousness of the illness and often stay home unnecessarily (Quint et al., 1991). A range of studies has been conducted using a variety of data sources to determine the rate of prevalence of disabilities among all women receiving AFDC, not just those participating in welfare-to-work programs. As Table 1 indicates, using self-reported measures of the existence of chronic illnesses or disabilities, there is little variation among estimates of the percentage of recipients who report they have a health condition or work disability that prevents them from working, but substantial variation among estimates of the prevalence of disabilities or health conditions that limit the amount or kind of work recipients can do.
Between 10.1 to 11.0 percent of recipients report they are unable to work because of a disability or serious medical condition. The percentage of recipients who report any limitation in the amount or kind of work they can do ranges from 16.6 percent to 30.8 percent, with no obvious reason for the difference in these estimates (Loprest and Acs, 1995; Acs and Loprest, 1995; Meyers et al., 1996). Approximately 8.4 percent to 9.5 percent of the women report that they need help with or are unable to perform specific functional tasks including dressing, eating, bathing, using the telephone or walking up stairs, while an additional 11.7 percent report some difficulty performing daily activities (Loprest and Acs, 1995).
Other measures of health status, that may be related to either chronic or temporary conditions, are also relevant to the ability of women to become employed. These measures include the number of overnight hospital stays, trips to the emergency room, and visits to the doctor. About 7 percent of women on AFDC report that they were confined to bed for health reasons for over 30 days in the past year. Almost 4 percent of women on AFDC were in the hospital for more than 5 days, and 8.7 percent had at least 16 or more visits to a doctor. Regardless of the nature of the underlying condition, such interaction with the health care system undoubtedly takes away from a woman’s ability to maintain steady employment, particularly in jobs that may not have benefits such as sick leave (Loprest and Acs, 1995).
AFDC recipients who have disabilities differ from their counterparts who do not have disabilities in several ways. Acs and Loprest (1995) find that recipients with disabilities are older on average than women on AFDC without disabilities. For example, 46.6 percent of women on welfare with disabilities are over age 35 compared to only 18.1 percent of those without disabilities. Furthermore, they are more likely to be married, less likely to have a high school education (40.6 vs 53.0 percent) or have young children. Acs and Loprest also found that a higher percentage of women on welfare for at least 12 out of 32 months have at least one functional limitation (23.4 percent) compared to the percentage of those on welfare for shorter periods of time (16.4 percent).
These and other studies have shown higher rates of prevalence of disabilities than studies have shown to be the case in the general population. Adler (1993) reports that approximately 19 percent of AFDC recipients between the ages of 15 and 45 report disabilities compared to only 10 percent of all women in that age range. Loprest and Acs (1995) also found that over one-fifth of the AFDC population feels that they are in poor (4.5 percent) or fair (17.5 percent) health. Similarly, Meyers et al. (1996) report that 12.6 percent of a sample of AFDC recipients from California stated that they are in "poor health". A much smaller proportion of all women ages 18 to 64, only 8 percent, report that they are in poor (2 percent) or fair (6 percent) health (Loprest and Acs, 1995).
B. Mental Health Problems
Very few studies in the literature examine the prevalence of mental illness among welfare recipients, possibly because mental health problems are much harder to identify than physical disabilities. Some of the previously discussed studies include measures of mental illness within their categories of illness and disability. However, they do not encompass all forms of mental disability and under reporting may occur.
A study of JOBS-eligible mothers with children between the ages of three and five in Fulton County, GA (primarily Atlanta and the surrounding suburbs) found that 42 percent of this group of mothers showed high level of depressive symptoms, more than twice the percentage found in community samples (Moore et al., 1996). FaDDS in Iowa indicates that 39 percent of its caseload statewide have mental health issues. The Riverside County GAIN program reports that 17 percent of those participants deferred from participation are deferred for emotional or mental health or substance use problems (Maynard, 1995). The New Chance Demonstration Project found that 28.4 percent of the young mothers it served were at high risk of clinical depression, and 22.5 percent were at some risk. Data from the 1989 National Health Interview Survey suggests that a substantially smaller percentage of recipients are likely to suffer from serious mental illness. Just 2 percent of all adult women and 4 percent of all adults with incomes below the poverty threshold have serious mental illness (Barker et al., 1992).
C. Health or Behavioral Problems of Children
Parents whose children have chronic health problems, disabilities or behavioral problems must attend to a variety of needs such as multiple doctors appointments, meeting with school officials, providing care for children with special needs, or remaining home to care for sick children. Data on emergency visits among AFDC families show that 55 percent of families with a severely disabled child reported at least one trip to the emergency room in the last year compared to 32 percent of all AFDC families (Meyers et al., 1996).2 Particularly in the low-wage labor market, it is difficult to find jobs that allow enough flexibility to meet these increased demands on a parent’s time, especially when only one parent lives with the child. Research shows that ten percent of poor families in which parents are employed faced a "family illness burden" of two to three weeks, and an additional 27 percent faced a burden of three weeks or greater.3 However, 38 percent of working parents living in poverty and 45 percent of working single parents did not have any paid sick leave in a given year.4 Over eighty percent of mothers living in poverty had less than one week of sick leave available to them (Heymann et al., 1995).
The prevalence of disabilities and functional limitations among children on AFDC varies depending on the definitions and sources of data used by the researchers, as shown in Table 2 (Loprest and Acs, 1995; Meyers et al., 1996). Rates of mild forms of disability range from 2.4 to 4.0 percent. This level of disability includes limitations in activity other than major activity (such as school attendance), or chronic conditions that limit the amount of things a child can do without making them unable to partake in usual activities for a child his or her age. Measures of moderate limitations on activities range from 9.3 to 10.1 percent and include children with some limitations in major activities or the necessity to attend special school or classes. The proportion of children in AFDC families who have severe levels of disability, including serious health conditions, limited school attendance or inability to perform major activity, ranges from 1.7 to 9.3 percent. Overall, these studies estimate that between 11.1 percent and 21.1 percent of children receiving AFDC have some level of limitation or disability.5
Program and caseload data also indicate that a high proportion of AFDC families have children with special needs or disabilities. FaDDS reports that 25 percent of the families it serves have children with special needs. In case record reviews conducted in Colorado, Iowa and Utah, 15.2, 17.5 and 13.0 percent of the cases respectively included documentation in the case record noting a child behavioral or medical problem as a potential or actual barrier to JOBS participation or employment (Pavetti and Duke, 1995).
Families with disabled children tend to be worse off on a number of measures than families without disabled children. Women with special needs children were more likely to rate their health as poor and were more likely to report that they personally had a disability which limited or prevented work. Families with at least one disabled child were more likely to have had a child or an adult go hungry and to have had a child go without medical care over the past year (Meyers et al., 1996). The level of impairment among AFDC children is not necessarily higher than among the general population, depending on the specific measures used. Estimates of disabilities among the general population of children range from .9 percent of those under 15 (McNeil, 1993) to 2 percent of those under 18 for severe disabilities or chronic illnesses (Newacheck and Taylor, 1992) and 5.2 percent (McNeil, 1993) to 9 percent (Newacheck and Taylor, 1992) for more moderately severe chronic conditions. Approximately 15.2 percent of children have chronic illnesses which cause major limitations in activities or serious pain or discomfort (Aday, 1992) and up to 30 percent have chronic illnesses when relatively mild conditions including mild allergies are included (Newacheck and Taylor, 1992), although both estimates exclude cancer and mental health problems.
D. Alcohol and Drug Use
Using data from the National Household Survey on Drug Abuse (NHSDA), several recent studies have analyzed the prevalence of alcohol and drug use among AFDC and other social program participants at the national level. These studies also include comparisons of the prevalence of substance use problems among AFDC and non-AFDC households. The data presented in Table 3 show that estimates of the prevalence of substance use problems are extremely sensitive to the definition used and the population studied. Estimates range from a low of 6.6 to a high of 37 percent (U.S. Department of Health and Human Services (DHHS), 1995a, 1995b; Center on Addiction and Substance Abuse, 1995). The lowest percentage is obtained using a definition of daily alcohol use.
During the last year among all AFDC households.6 The definition and subgroup yielding the highest estimate is binge drinking two or more times or any use of drugs (including marijuana) during the last year estimated for AFDC recipients between the ages of 18 and 24. While daily drinking for the last year may be too restrictive a definition to identify the percentage of the caseload whose transition from welfare to work may be slowed due to substance use problems, the definition that includes any drug use is almost certainly too broad.
To be able to better examine the relationship between substance use problems and employment, one study conducted by DHHS (1995c) analyzed the prevalence of substance use problems using two different definitions of "functional impairment." This study, using data from both the 1991 and 1992 NHSDA studies, estimates that approximately 4.9 percent of female AFDC recipients have a significant functional impairment related to substance use problems and an additional 10.6 percent are somewhat impaired by substance use problems. Significant impairment includes persons identified as:
1) dependent on alcohol and drunk at least weekly; or
2) dependent on an illicit drug other than marijuana and used an illicit drug at least monthly or used heroin at least once in the past year.
Recipients defined as somewhat impaired by substance abuse problems are those individuals identified as:
1) not dependent on alcohol or any illicit drug but used an illicit drug at least weekly or was drunk at least weekly; or
2) dependent on an illicit drug other than marijuana and used an illicit drug less than monthly and did not use heroin; or
3) dependent on alcohol but was drunk less than weekly; or
4) dependent on marijuana and not otherwise meeting criteria for maximum impairment.
Data available on participants in welfare-to-work programs produce similar estimates. A 1992 comparison of male and female recipients of GA and AFDC found that 16 percent of male recipients and 5 percent of female recipients were alcohol dependent and 32 percent of males and 13 percent of females were drug dependent (Schmidt, 1992, cited in Sisco and Pearson, 1994). Sisco and Pearson (1994), examining the prevalence of abuse in female AFDC recipients enrolled in Project Independence, a welfare to work program in Montgomery County, Maryland, found higher rates of substance abuse: 11.4 percent of participants studied were identified as alcoholics and another 14.7 percent as possible alcoholics while 9.4 percent of the sample were identified as drug abusers and 16.7 percent as drug users.
7 When alcohol and drug abuse were considered simultaneously, 16.1 percent of the sample were alcoholics and/or drug abusers and an additional 20.8 percent of the sample were possible alcoholics and/or drug users. In New Chance, 12 percent of enrollees across all sites used alcohol and 15 percent used illegal drugs to such a degree that it interfered with their participation in the program. However, there was substantial variation across the programs studied.
Most studies that examine the prevalence of substance use problems look at current rather than past problems. However, research shows that past abuse may also affect a recipient's ability to make a successful transition into the labor market. In their research on the relationship between substance use problems and employment, Renwick and Krywonis (1992) note that abstinence does not necessarily eliminate the effects of abuse on employment, particularly the psychological barriers such as low self-esteem. The FaDDS program in Iowa reports that approximately 26 percent of the families it serves had substance use problems in the past.
In general, most measures of alcohol or drug use produce higher estimates among AFDC than non-AFDC households (See Table 3). While 7.8 percent of mothers in AFDC households are estimated to have had at least one drink daily or almost daily for the past year, the comparable estimate for all mothers is just 3.8 percent. Similarly, 9.1 percent of AFDC mothers but just 4.0 percent of all mothers have had five or more drinks in one sitting more than three times in the past month. Similar patterns emerge for estimates of drug use. An estimated 12.6 percent of mothers living in households receiving AFDC report past-month drug use compared to only 5.1 percent of all mothers (DHHS, 1994b).
One reason that the rates for the AFDC population are higher than for the general population is that the AFDC population is younger on average, and younger persons are more likely to report illegal drug use. While less than fifty percent of all females in AFDC households are 30 years of age or younger, 71.1 percent who report past-month drug use are age 30 or younger. Furthermore, among females, those 20 to 25 years of age have the highest rates of self-reported drug use — 12.6 percent of all women in this age range report past month use and 25.8 percent report past-year use, and the comparable rates for AFDC recipients of this age are 16.8 percent and 31.6 percent respectively (DHHS, 1994a). Another study using data from the 1991 NHSDA reports that 24 percent of all parents age 20 to 25 report illicit drug use in the past year and 12 percent report past-month use while 34.6 percent of all parents receiving AFDC report past year use and 19.6 report past month use. Substance-abusing mothers are also more likely to have young children, at least in part due to the fact that they are younger themselves (DHHS, 1994b).
Even though substance abuse, especially illicit drug use, is often identified as a central city problem, research suggests that it is not limited to central cities or specific geographic locations. According to the 1991 NHSDA study completed by DHHS, program participants living in small metropolitan areas are more likely to report drug use than those living in either larger cities or rural areas — 13.5 percent of persons in AFDC households in small metropolitan areas report past month drug use compared to 9.5 percent of those in large metropolitan areas and 9.4 percent of those in non-metropolitan areas. AFDC and Medicaid participants who report past-month drug use generally are less likely to live in large metropolitan areas than are all program participants. While 49 percent of all female AFDC recipients live in large metropolitan areas, only 41 percent of those who have used drugs in the past month reside in such areas.
A majority of state and local welfare directors agree that substance and alcohol treatment services are a necessary component of efforts to move welfare recipients off welfare into work. According to a study conducted by the Legal Action Center (1995), 65 percent of state and local welfare directors say drug and alcohol treatment services are extremely important to help welfare recipients get off welfare and stay off and a total of 99 percent say such services have at least some role to play.
E. Domestic Violence
Domestic violence among welfare recipients has been increasingly noted as a major barrier for many women attempting to become self-sufficient (Raphael, 1995; Lloyd, 1995; Davis and Kraham, 1995). Women currently involved in abusive relationships may be unable to attend class or work regularly because of disruptive and threatening behaviors of their partners both at home and on the job. Also, both current and past victims may suffer from psychological and physical effects of the abuse. Because women in abusive relationships are often afraid or embarrassed to report they are involved with an abuser, abuse is often not reported as a problem until it is completely out of control. Consequently, estimates of prevalence of domestic violence rely heavily on reports by individual programs working with families and other anecdotal evidence which tends to vary widely. Stronger research on the link between welfare usage and domestic violence is clearly needed.
The Taylor Institute (1995) reports that at one employment training program in Chicago, 58 percent of participants who entered the program between July 1, 1993 and June 30, 1994 were current victims of domestic violence and an additional 26 percent were past domestic violence victims. Seventeen percent of all participants also reported being incest survivors or past victims of sexual assault. The Washington State Institute for Public Policy found that 60 percent of a representative sample of the state welfare population reported some type of abuse (physical and/or sexual) compared to 35 percent for a comparison group of women drawn randomly from neighborhoods likely to have high rates of public assistance. More than half (55 percent) reported being physically abused by a spouse or boyfriend compared to 28 percent for the at risk sample. Other community programs across the country, informally surveyed by the Taylor Institute, estimated that anywhere from 20 to 80 percent of participants are currently experiencing some form of domestic violence. FaDDS in Iowa found that 18 percent of the women in families it served were current victims of domestic violence; 47 percent were past victims and 19 percent were adult victims of rape or sexual violence/assault. For comparison, in the general population, a large-scale national survey conducted in 1985 found that 11.6 percent of married or cohabitating male partners carried out one or more violent acts and 3.4 percent severely assaulted their partners during the year of the survey (Straus and Gelles, 1990).
It is interesting to note that JOBS case record reviews undertaken in three states that have implemented welfare reform demonstration projects reveal much lower rates of domestic violence, possibly reflecting the reluctance of women on welfare who are living with abusive partners to acknowledge they are living with a boyfriend who may or may not be the father of their children. Researchers found that only 8.3 percent of cases in Utah had family violence documented as a potential or actual barrier to JOBS participation and/or employment. In Iowa and Colorado, only about 6 percent of the case records indicated evidence of family violence (Pavetti and Duke, 1995).
Women who live with or are involved with abusive partners, in addition to experiencing direct detrimental impacts on employment, also lack positive assistance that a supportive partner can provide. And, for many AFDC families, regardless of the existence of domestic violence per se, the lack of external support systems can make attempts to become self-sufficient much more difficult. In New Chance, young mothers with some source of emotional support attended program activities for significantly more hours than those without any support. It is important to note that boyfriends or male partners are not the only significant others who may discourage program participation. For example, nine percent of New Chance enrollees had mothers or close relatives who discouraged their participation and one program had as many as 45 percent of its enrollees experiencing negative responses to participation by partners or family members (Quint et al., 1991).
F. Child Welfare
Some AFDC families are involved with the child welfare system in ways that may affect their participation in employment or training programs and their transition to self-sufficiency. Involvement with child protective services (CPS) and the foster care system may signal that a family is struggling with a multitude of intense problems that may detract from efforts to achieve economic self-sufficiency. On the other hand, mothers whose children have been removed from the home or whose children are at risk of removal may gain access to intensive social services to assist the mother and the family in addressing personal and family problems. One of the early lessons of the family preservation movement is that the crisis of a child's removal, or threat of removal, provide the incentive and opportunity for a family to make significant changes (Barthel, 1992).
Involvement in the child welfare system can be measured in a variety of ways — prevalence of abuse and neglect, cases reported to child protective services (CPS), substantiated or indicated reports, and placement rates in foster care. One of the best sources of national data is the national incidence study of child abuse and neglect in 1986, which included estimates of the prevalence of child abuse and neglect among families of different income levels. An estimated 3.2 percent of children from low-income families were estimated to be abused or neglected, including cases that are not reported to the CPS system (Sedlak, 1988). The estimated prevalence was 5.4 percent under a slightly broader definition of abuse and neglect. These rates are about twice as high as the overall rates of 1.6 percent and 2.5 percent reported for children nationwide.8 About half of the low-income children at risk of harm involved cases of neglect; the other half involved physical, sexual, or emotional abuse (Sedlak, 1988).
The number of AFDC children experiencing child abuse and neglect is likely to be higher than the 5 percent rate for low-income families (defined as families with income less than $15,000). Child abuse and neglect appears to be more prevalent among the poorest of poor families, including families living in very poor housing conditions (Pelton, 1989; Besharov, 1994). Furthermore, child neglect, and to a lesser extent child abuse, is more common among single-parent families than other families (Polansky, Gaudin & Kilpatrick, 1992; Besharov, 1994).
Besharov (1994) estimates that "fewer than one in five welfare families are reported for suspected abuse or neglect" in any one year. He cites national data from the American Humane Association that suggest that approximately half of the families reported for any form of child abuse or neglect receive public assistance. Data from the National Child Abuse and Neglect Data System (NCANDS) suggests that national reports on child abuse and neglect totaled 1.8 million reports on approximately 2.7 million children in 1991 (National Center for Child Abuse and Neglect, 1993). As a rough estimate, therefore, 900,000 of the reports on 1,350,000 children may have involved AFDC families, suggesting an incidence of 21 percent in terms of reports per AFDC family or 16 percent in terms of reported children per AFDC children, or roughly one in five welfare families.9
Empirical data on the overlap between the welfare and child welfare populations has been analyzed by the Chapin Hall Center for Children, using four years of administrative data from the state of Illinois. Goerge et al. (1993a) report that among children under age 20 who were receiving AFDC, food stamps, and Medicaid in June 1990, 11 percent had been identified as indicated victims of child abuse or neglect at least once within the past four years. This rate is 2.5 times higher than the 4.4 percent reported for the overall child population for Illinois.10 The rate of indicated cases of abuse and neglect among families (rather than children) is similar, 12 percent (Goerge et al., 1993b). A third analysis showed that among Illinois first-born children who were born in 1988 and had received AFDC at least once in their first five years of life, 6 percent had indicated reports of abuse or neglect within that same five-year period (Gorge et al., 1993c). The comparable rate among the 1988 cohort of first-born children across the state was 4 percent.
Goerge et al. (1993c) also estimated foster care placement rates among AFDC families. They report that 2 percent of Illinois first-born children receiving AFDC at some point in their first five years also experienced at least one foster care placement. The placement rate was 0.8 percent for children who did not receive AFDC. Analyses of foster care placement rates among different census tracts in New York city provide additional evidence that foster care placements are highest among neighborhoods with large proportions of poor families, female-headed families, and among children in sub-families (Wolczyn, personal communication, March 22, 1996). Another analyst of trends in New York city data estimated that "perhaps 15 percent" of AFDC cases in New York city also enter the child welfare system (Berlin, 1992). Finally, a rough measure of the national foster care placement rate among welfare families can be calculated by dividing the Federal foster care caseload into the AFDC caseload. In the average month of fiscal year 1995, 260,700 children received Federal foster care payments and there were 9,245,000 children on AFDC, resulting in a placement rate of 2.8 percent.
G. Housing Instability
Lack of stable, affordable housing can also be a serious barrier to stable employment for some low-income families. Homelessness may be the result of other family problems, such as domestic violence. Or, persistent homelessness may be the cause of additional family problems such as child behavioral problems or substance abuse. In either case, not having stable or safe housing, makes finding and keeping steady employment more difficult. Angela Browne (1993) cites studies of homeless families, in which the mothers were described as having more fragmented or nonexistent support networks, less contact—supportive or otherwise—with family members and more contacts with victimizing and substance-abusing men than their housed counterparts. Homeless women/mothers also manifest higher rates of chemical dependence, depression, and other medical and psychological symptomatology than do their housed counterparts (Buckner et al., 1993).
In the New Chance Demonstration project, almost half (48 percent) of all enrollees faced a housing problem that interfered with their ability to attend the program. In five of the sites, more than 60 percent of the young women had a housing problem of this kind, with the proportion ranging from 11 to 90 percent across all the programs (Quint et al., 1991). In Utah, homelessness was reported as a barrier to employment in 9.3 percent of the case records reviewed (Pavetti, 1995).
H. Low Basic Skills and Learning Disabilities
A substantial proportion of welfare recipients have extremely low levels of schooling, and even more have extremely low reading levels despite what might be a higher number of years in school. Some recipients with serious deficits will be qualified to compete for only a limited pool of jobs, making it more difficult for them to find employment. Others may have such serious skill deficits that they are unable to perform even the most rudimentary tasks, making it difficult for them to find employment without some additional education or basic skills training. Zill et al. (1991) report that at least 10 percent of AFDC mothers have only a grade-school education. Pavetti (1993) shows that low skills is one of the strongest predictors of long-term welfare receipt. Nightingale et al. (1991) estimated that between 25 to 40 percent of AFDC recipients may have learning disabilities.
Only 26.5 percent of the New Chance program participants had a 10th grade reading level or above, and nearly 30 percent were at the 6th grade reading level or below. Out of 16 programs, 14 had at least 25 percent of their participants reading at the 6th grade level or below, and 6 programs had at least 40 percent of their participants at these levels. Many vocational training programs that are linked with jobs require reading levels at the 8th grade level or above, resulting in limited training options for recipients with the greatest skill deficits.
I. Summing Up
Table 4 provides an overview of the range of estimates included in the literature for each of the personal or family challenges highlighted above. The wide range of estimates for each of the potential barriers to employment is the most striking pattern observed in this table, reflecting in part the limited attention this type of analysis has received. With the exception of the presence of serious health problems, there is at least a ten percentage point difference between the highest and lowest estimate of the prevalence of each of these potential barriers to employment.
These differences result largely from different approaches to defining and measuring potential barriers to employment. Some of the figures are based on nationally representative samples of welfare recipients, while others indicate prevalence rates among specific segments of the AFDC population such as teen mothers. The methods used to collect the prevalence rates vary from brief surveys of recipients themselves to interviews of program staff who worked with recipients for over a year. More than anything, the range of estimates reflects the absence of accepted criteria for how to define and measure the prevalence of the challenges recipients and their families may face as they make the transition from welfare to work. These figures also capture very different levels of severity associated with the particular barrier. For example, housing instability in one study is defined as the existence of housing problems that interfere with participation, while the definition in another study is actual homelessness.
III. What Fraction of the Caseload is Likely to Experience One or More Barrier to Employment?
The studies reviewed above provide important insights into the type and extent of barriers faced by welfare recipients. It is certainly useful for programmatic purposes to understand the prevalence of individual barriers in order to determine the need for particular kinds of services. However, because families may face more than one potential barrier to employment, it is impossible to use this information to construct an overall estimate of the percentage of the caseload that may need more assistance than traditional welfare-to-work programs generally provide to successfully make the transition from welfare to work. Very different implications derive from whether the majority of welfare recipients each face one barrier or a much smaller proportion face several. In this section, we briefly review the existing literature and analyze data from the National Longitudinal Survey of Youth (NLSY) to examine the prevalence of multiple challenges which may impede employment.
Because estimating the prevalence of multiple barriers to employment depends upon the ability to document a wide range of potential barriers, only a limited number of studies provide such estimates. The most comprehensive analysis is a survey of New Chance program staff that documented the extent to which the following issues interfered with program participation: discouragement of program participation by mother, other close relative, or partner; participant's or close family member's use of alcohol or illegal drugs; housing instability; current physical abuse; and physical or sexual abuse as a child. Staff reported that 74 percent of the young women experienced at least one of the above problems (broken into 10 categories). Of those who experienced any problem, more than half (54 percent) registered more than one. This survey involved fourteen varied sites and covered 617 enrollees whom staff knew well enough to indicate whether such a problem existed (Quint et al., 1991). Unlike other studies, the responses are not self-reported but are based on established relationships built over time.
A case record review in Utah that simultaneously examined the presence of children's behavioral or medical problems, family violence, homelessness, temporary and chronic medical problems or disabilities, chemical dependency, mental health and recent work history found that 34 percent of the total sample experienced more than one of these potential barriers. However, the presence of these barriers was not equally distributed across the sample. Just five percent of recipients receiving assistance for six months or less out of a 24-month period reported more than one barrier to employment compared to 36 percent of recipients who received assistance for between 7 and 18 months and 52 percent who received assistance for 19 to 24 months (Pavetti, 1995).
Because only limited information is available on the extent to which families experience multiple challenges that may impede their transition from welfare to work, we examine data from the NLSY to provide additional insight. The NLSY is a national panel survey of 6,403 male and 6,283 female youth who were between the ages of 14 and 22 in 1979, the first year the survey was conducted. Those interviewed include: (1) a nationally representative sample of young people; (2) additional subsamples of African-American, Hispanic and low-income white youth and (3) a subsample of youth enlisted in the four branches of the military as of September 30, 1978. Most of the military subsample was dropped after the 1984 interviews and the low-income white subsample was dropped after the 1991 interviews. The tabulations presented here are based on data for the 4,535 women who were interviewed in 1992. All estimates are weighted to adjust for the presence of the over-sampled groups and sample attrition.
Each year the NLSY collects data on wide range of topics including employment, AFDC, Food Stamp, unemployment compensation and Supplemental Security Income (SSI) receipt, training and schooling, fertility and marital status. In addition, over the time period covered by this analysis, the survey collected data that makes it possible to analyze the prevalence of the following potential barriers to employment: (1) medical problems of the household head; (2) chronic medical problems of children in the household; (3) alcohol and drug use; (4) mental health problems (depression) and (5) low basic skills.11 Since this list of barriers is far from exhaustive, the estimates of the overall prevalence of barriers are likely to be underestimated and are best interpreted as lower bounds.
For purposes of this analysis, we divide potential barriers to employment into two different categories denoting the level of severity of the potential barrier. Conditions potentially posing the most serious barriers to employment include:
(1) not seeking work due to a medical condition;
(2) experiencing depression between five and seven days a week;
(3) extensive alcohol or drug involvement defined as being concerned about being an alcoholic, having had problems at work or school because of drinking or using cocaine or crack repeatedly (more than 100 times in one's lifetime); or
(4) extremely low basic skills defined as scoring in the bottom decile of the women's distribution of the Armed Forces Qualifying Test (AFQT), a measure of basic skills shown to be highly correlated with employment and earnings.
Moderate barriers are defined as those barriers that may restrict the amount or kind of work that a woman can do or may interfere with her ability to sustain continuous employment. These potential barriers include:
(1) presence of a medical problem that limits the amount or kind of work a person can do;
(2) presence of a child with a chronic medical condition;12
(3) some physical indication of problem drinking including the shakes, loss of memory or drinking first thing in the morning;
(4) experiencing depression between three and five times per week;
(5) repeated use of marijuana (more than 100 times in one's lifetime); or
(6) very low basic skills defined as scoring between the 10th and 25th percentile of the AFQT distribution for women.
As noted above, since there has been relatively little work done on examining potential barriers to employment, there is no generally accepted definition for what constitutes a potential barrier to employment. Because the definitions used in the literature differ from study to study, it is impossible to choose definitions to correspond with existing research. Thus, the definitions used in this analysis reflect the best use of the data available in the NLSY and do not necessarily correspond to the data that is used in the studies reviewed above.
The data presented in Table 5 show that just over 30 percent of the AFDC caseload report a serious barrier to employment and two-thirds report either a serious or modest barrier, excluding low basic skills.13 When one takes into account extremely low or very low skills, the percentage of the caseload with a serious barrier to employment increases to 54 percent and the percentage of the caseload with a moderate or severe barrier increases to 89 percent.
Extremely low basic skills are much more common that other potentially serious barriers to employment. Fully 33 percent of the AFDC caseload scored in the bottom decile on the AFQT test. The next most commonly reported serious barrier to employment is almost constant depression (five to seven days per week) reported by 13 percent of recipients, followed by a medical problem that prevents work reported by 10 percent of all recipients. Nine percent of recipients report extensive cocaine or crack use and five percent report they are concerned about being an alcoholic.
The most commonly reported moderate barriers to employment are very low skills reported by 31 percent of recipients and the presence of a child with a chronic medical condition and some physical indication of problem drinking such as having the shakes, loss of memory or drinking first thing in the morning, both reported by about twenty percent of the caseload. About 16 percent of the caseload report extensive marijuana use, 11 percent report being depressed between 3 and 5 days a week and seven percent report a medical condition that limits the amount or kind of work they can do.
The data presented in Table 5 also show that the incidence of these potential barriers to employment is higher among respondents who are currently receiving welfare than those who are not. In most, but not all cases, the difference in prevalence is greater for the more serious barriers to employment. Welfare recipients are about five times as likely as non-recipients to report they are not seeking work due to a medical problem and about 40 percent more likely to report a medical condition that limits the amount or kind of work they can do. Recipients are nearly three times as likely as all women to report being depressed five to seven days out of a week and 1.55 times more likely to report being depressed between three and five days per week. While there is virtually no difference in the prevalence of serious alcohol or drug use, recipients are more than twice as likely to report some physical indication of problem drinking. Low basic skills are much more common among the AFDC recipients than among non-recipients. Current welfare recipients are 4.35 times as likely to score in the bottom decile and 2.15 times as likely to score between the 10th and 25th percentile on the AFQT than non-recipients.
The data presented in Table 6 show that the majority of the caseload who experience any of these barriers to employment experience only one rather than multiple barriers. Excluding low skills, almost 90 percent of recipients who report a serious barrier to employment, report just one serious barrier. Even including extremely low skills, 77 percent of the caseload with serious barriers experience just one serious barrier to employment. About sixty percent of those with either serious or moderate barriers report just one barrier. When very low skills are included the percentage experiencing just one barrier declines to 44 percent. However, the prevalence of multiple barriers to employment could change quite dramatically if additional barriers such as housing instability or domestic violence were included.
IV. The Relationship Between Potential Barriers to Employment, Time on Welfare and Actual Work Experience
As noted earlier, the mere existence of many of these personal and family challenges does not necessarily mean that all AFDC recipients who experience these challenges will be unable to make the transition from welfare to work. Nonetheless, one would expect that the existence of any of the above-mentioned issues would impact some recipients' ability to work as evidenced by a lower probability of leaving welfare given the existence of a particular condition, longer stays on welfare and/or higher rates of unemployment.
This section begins by examining the available evidence on the relationship between specific personal and family challenges and employment. Then, data from the NLSY are used to examine the relationship between the existence of several potential barriers to employment and the length of time recipients spend on the welfare rolls. This section concludes with an analysis of NLSY data on the actual work experiences of recipients and non-recipients experiencing and not experiencing potential barriers to employment.
A. The Relationship Between Potential Barriers and the Likelihood of Employment
There are relatively few studies that examine the relationship between personal and family challenges and the likelihood of employment. Acs and Loprest (1995) note that most studies that examine the relationship between disabilities and exits from welfare find little evidence that the existence of disability affects the likelihood of exiting AFDC. However, their research shows that there does appear to be a relationship between the presence of a severe disability and exits from welfare. Women with functional limitations had a 3.8 percent chance of exiting AFDC for work in a given four-month period compared to 7.6 percent for other AFDC recipients. Women with severe limitations (needs help or unable to perform certain daily activities) had only a 2.5 percent probability of leaving AFDC for employment. When compared to women without any disabilities, women with less severe disabilities but some level of limitation do not experience significantly different probabilities of leaving AFDC for work.
Recent research on the relationship between expected medical expenditures and employment suggests that the lower exits for work may not be entirely due to the presence of a serious medical condition or physical disability but may instead be related to the need for health insurance to cover the expenses associated with these conditions. Moffitt and Wolfe (1992), using data from the 1984 and 1986 SIPP to develop a family-specific variable based on expected medical expenditures and health conditions, estimate that increases in expected Medicaid benefits exerted strong negative effects on employment and positive effects on AFDC participation among welfare recipients. However, these effects appear to be concentrated among those recipients with the highest medical expenditures. Their simulations also indicate that extending private health coverage to all working female heads would lower the AFDC caseload by 10.7 percent, and an increase in benefits of private plans up to Medicaid levels extended to all female workers would decrease AFDC participation by nearly 25 percent and increase employment levels by 18 percent.
The research available on the extent to which having a child with a disability hinders the ability of the parent to work is not entirely consistent. Acs and Loprest (1995) find little evidence that a mother’s chance of exiting welfare for work declines because of the existence of a child’s disability. However, Meyers et al. (1996) find in their sample of California AFDC recipients that the presence of two or more disabled children does decrease the probability of mothers' employment.14 After controlling for education, family characteristics and SSI receipt, mothers of more than one child with a disability were significantly less likely to work—by approximately 70 percent. Also, the presence of at least one severely impaired child reduced the odds of working by about half (Meyers et al., 1996). Similarly, Wolfe and Hill (1995), using a 1984 panel of the SIPP, found that the presence of a disabled child significantly reduced the probability of employment among a sample of single mothers.
As in the case of families with a disabled adult, families with children most likely to experience a particularly acute illness may have an especially high tendency to remain on welfare in order to maintain Medicaid eligibility for their families. Wolfe and Hill (1995) projected that extending health insurance coverage to all children would increase the probability that single mothers work by nearly 12 percent and increase the probability that single mothers with a disabled child would work by 22.6 percent. Providing coverage only to children with disabilities still increases the probability of their mothers working, but by only 8.1 percent because such a plan leaves other children in the family uncovered.
Unemployment and vocational instability do appear to be more prevalent among persons who abuse drugs or alcohol than among those who do not. One study (DHHS, 1994a) calculated that, while 21 percent of all females in AFDC households were unemployed, a higher percentage of drug users were unemployed — approximately 30 percent of female past-month drug users and 29 percent of female past-year drug users were unemployed. This study also shows that drug use is more prevalent among recipients who are unemployed. Nearly 16 percent of unemployed females in AFDC households report past-month drug use and nearly 30 percent report past-year use compared with 8.9 percent and 17.8 percent respectively for females who are employed full time. A second study also found that 17 percent of unemployed parents report past-month drug use compared with under 5 percent of full or part time employed parents (DHHS, 1994b).
Lloyd (1995) surveyed a random sample of women in a low-income Chicago neighborhood in order to uncover the effects of domestic violence on their labor market participation. More than one-quarter (27.7 percent) reported experiencing severe violence perpetrated by their male partners during their lifetimes and 39 percent reported having ever experienced any male aggression directed at them. Thirty-six percent reported experiencing direct verbal harassments or threats, 13.1 percent accounted physical aggression, and 8.6 percent reported severe aggression such as being beaten or raped. Surprisingly, Lloyd found no significant differences in current employment status among those who reported having experienced male aggression (in last twelve months or ever) and those who did not. However, she did find that those women who had been victimized were more likely to report having ever been unemployed (56.6 percent for never victimized, 69.4 percent for those reporting physical aggression and 67.2 percent of those who experienced severe aggression), to suffer from a range of physical or mental health problems, to earn less, and to be more likely to receive public assistance. Furthermore, regardless of the level of aggression, current and past victims reported higher rates of depression, anxiety and anger (in some cases 17 to 18 percentage points higher).
B. The Relationship Between Potential Barriers and Time on Welfare
Data from the NLSY (Table 7) shows that long-term recipients are more likely than either short or intermediate-term recipients to have very low skills, report they are not seeking work because of their own medical problems or report they are depressed between three and five days a week. They are also more likely to report some physical indication of problem drinking while intermediate-term recipients report the highest rates of extensive marijuana use. Intermediate and long-term recipients are both more likely than short-term recipients to have a child with a chronic medical condition.15
When the presence of any barrier is taken into account, according to the data presented in Table 7, long-term recipients are only slightly more likely than short-term recipients and no more likely than intermediate-term recipients to experience at least one serious barrier to employment, other than extremely low skills. About 26 percent of short-term, 30 percent of intermediate-term and 32 percent of long-term recipients report a potentially serious barrier to employment. When one includes moderate or severe barriers other than basic skills, the percentages rise to 58, 66 and 64 percent respectively.
When extremely low skills are included, serious barriers to employment are most prevalent among long-term recipients (54 percent), followed by intermediate-term (48 percent) and short-term recipients (40 percent). When moderate or severe barriers including very low basic skills are measured, 89 percent of long-term, 86 percent of intermediate-term, and 76 percent of short-term recipients experience at least one potential barrier.
C. The Relationship Between Potential Barriers to Employment and Actual Work Experience
Data from the NLSY (Table 8) show that even though recipients who experience barriers to employment may work less than recipients who do not experience such barriers, a substantial fraction do work, albeit intermittently. More than half of the recipients experiencing a serious barrier to employment worked in the current or previous year, as did 60 percent of recipients with a moderate or severe barrier and 60 percent of the overall caseload. The one exception to this result is the significant role of basic skills. Only 55 percent of recipients who scored in the bottom quartile of the women's AFQT distribution worked compared to 70 percent in the upper three quartiles. Only 44 percent of recipients in the bottom decile worked.
While employment was common among welfare recipients with and without barriers, continuous employment was not. Only seven percent of employed recipients with a serious barrier to employment other than low skills worked for at least 50 weeks out of the current or prior year in contrast to about one-quarter of recipients without a serious barrier and 19 percent of the caseload in general. When extremely low skills are included, the percentage of workers with a serious barrier to employment working for a full year increases to 11 percent.
In general, when barriers to employment are considered in the broad categories presented in Table 8, there is far more variation in the percentage of employed recipients working a full year than in the percentage of recipients working at all. Only recipients who report moderate or severe alcohol or drug problems have higher rates of employment (63 percent) than recipients with other barriers (55 to 56 percent). However, while only seven percent of workers with a medical problem worked full year, 20 percent of workers with a child with a chronic medical condition did so and between 12 to 15 percent of recipients with other potential barriers to employment worked full year.
The data presented in Table 8 further suggest that while intermittent work is common among welfare recipients, it is not the norm for non-recipients, including those with serious barriers to employment. Fully 89 percent of non-recipients worked at least part-year and 65 percent of those who worked did so for a full year. Similarly, 85 percent of non-recipients with a potentially serious barrier to employment (including low skills) worked and 67 percent of those who worked did so for a full-year; the comparable percentages for those with a moderate or serious barrier to employment are 88 and 65 percent, respectively.
These results are consistent with a study completed by the Manpower Demonstration Research Corporation on the employment potential of welfare recipients in California's GAIN Program. This study, which focused on a subset of recipients who were on AFDC for more than 24 of the 36 months following their entry into the program, found that most of this group of recipients could potentially have worked during the study period and over half did work. However, serious health and situational problems would have made continuous employment unrealistic for a substantial minority. More than half of this group were temporarily deferred from participation in GAIN activities for reasons other than part-time work, such as illnesses and severe family crises (Riccio et al. 1995).
V. Program Implications
In many ways this analysis raises as many questions as it answers. Given the data currently available on the prevalence of various potential barriers to employment, it is impossible to estimate precisely the fraction of the caseload that will need additional supports to find and/or keep employment. Using a conservative measure of barriers to employment presented in this analysis (the presence of a serious barrier to employment including low skills) and lack of participation in the labor force, we estimate that at least 25 percent of the caseload is likely to experience difficulties entering the labor force. When one accounts for the fact that the vast majority of those recipients who work do so only intermittently, the percentage of the caseload likely to need some additional assistance to fare well on their own over the long-term, especially within a time-limited welfare system, increases dramatically — to 51 percent.16
This analysis also highlights how difficult it will be to identify, especially in a new environment of time-limits and stringent work requirements, who will find employment on her own, who can benefit sufficiently from existing job training services to become employed, and who will need substantially more assistance to succeed in the paid labor market. A substantial fraction of the caseload that experience potential barriers do work, making the presence of such a barrier a relatively weak indicator of whether a family will be able to find employment with limited assistance. Thus, in an era of limited resources it will be important for programs to develop effective strategies to distinguish those families most in need of additional support.
It is, however, important to note that extremely low skills is the one factor that most distinguishes whether a recipient will enter the labor market or not. Only 44 percent of recipients scoring in the bottom decile on the women's distribution of the AFQT worked in the current or prior year. Recipients with other barriers to employment were substantially more likely to work. Given that one-third of all current recipients fall into this extremely low skill category, identifying appropriate labor market and alternative education and training strategies for this group of recipients is likely to be a challenge that will face most or all welfare departments as they attempt to transform the welfare system into a transitional, more employment-focused system.
The need to devote resources to help recipients keep as well as find jobs could not be more striking. Only 11 percent of those with a severe barrier to employment (including low skills) worked continuously over the past year. This analysis suggests that understanding the factors that lead to job loss and program interventions to support recipients beyond job placement will be necessary components of any strategy to transform the AFDC program into a transitional assistance program. As more and more recipients are expected to support themselves through paid employment, it is important to learn from existing programs currently grappling with how best to help recipients retain the jobs they find.
While all families are likely to experience some challenges in order to work, poor families living in adverse environments will bring a broader array of issues to the programs which serve them. Some recipients may need help locating resources to care for a special needs child, while others may need alcohol, drug or mental health treatment. Furthermore, the severity of the barriers facing these families will vary greatly. While many of these families will have developed strengths to cope with their life circumstances, some will still need significant additional support. To address the diverse needs of families, programs will need to be flexible and have access to a broad range of services not generally available within the welfare bureaucracy. With limited resources available to help recipients enter the labor market, one of the major challenges confronting state and local welfare offices will be to determine how to maximize the use of available resources and to restructure the design and delivery of other services to better meet the needs of families who have not been served or have not fared well in most welfare-to-work programs.
REFERENCES
Acs, Gregory and Pamela Loprest. 1995. "The Effect of Disabilities on Exits from AFDC." Paper presented at the 17th Annual Research Conference of The Association for Public Policy Analysis and Management, November 2-4.
Aday, L.A. 1992. "Health Insurance and Utilization of Medical Care for Chronically Ill Children with Special Needs: Health of Our Nation's Children, United States, 1988." Advance Data from Vital and Health Statistics No. 215. Hyattsville, MD: National Center for Health Statistics.
Adler, Michele. 1993. Disability Among Women on AFDC: An Issue Revisited. Report prepared for the Office of the Assistant Secretary for Planning and Evaluation. U.S. Department of Health and Human Services. Washington, D.C.: ASPE/DHHS.
Barker, Peggy, Ronald Manderscheid, Gerry Hendershot, Susan Jack, Charlotte Schoenborn, and Ingrid Goldstrom. 1992. "Serious Mental Illness and Disability in the Adult Household Population: United States, 1989," in Mental Health, United States, 1992. Center for Mental Health Services, National Institute of Mental Health, Rockville, MD.
Barthel, Joan. 1992. For Children's Sake. The Promise of Family Preservation. New York: Edna McConnell Clark Foundation.
Berlin, Gordon. 1992. Choosing and Measuring Interventions. Children and Youth Services Review, 14: 99-118.
Besharov, Doug, with Lisa Laumann. 1994. Don't Call it Child Abuse If It's Really Poverty. Paper presented at the "Social Policies for Children" conference, Princeton, New Jersey, May 25-27.
Browne, Angela. 1993. "Family Violence and Homelessness: The Relevance of Trauma Histories in the Lives of Homeless Women." American Journal of Orthopsychiatry 63(3): 370-384.
Buckner, John C., Ellen L. Bassuk, and Bonnie T. Zima. 1993. "Mental Health Issues Affecting Homeless Women: Implications for Intervention." American Journal of Orthopsychiatry 63(3): 385-399.
Center on Addiction and Substance Abuse. 1994. Substance Abuse and Women on Welfare. New York: Center on Addiction and Substance Abuse at Columbia University.
Davis, Martha F. and Susan J. Kraham. 1995. "Protecting Women's Welfare in the Face of Violence." Fordham Urban Law Journal 22: 1141-1157.
Dore, Martha Morrison. 1991. "Family Preservation and Poor Families: When >Homebuilding' Is Not Enough." Families in Society: The Journal of Contemporary Human Services
Goerge, Robert, Bong Joo Lee, John Van Voorhis, Teresa Sommer, Allen Hardin, Stephen Grant, Thomas Pavkov, Eboni Howard, Maria Alvarado, Paul Whiteneir and Diane Durante. 1993a. The Point in Time Multiple Service Use of Children and Adolescents in Illinois. Chicago, IL: Chapin Hall Center for Children.
Goerge, Robert, Bong Joo Lee, Teresa Sommer, John Van Voorhis, Allen Hardin, Stephen Grant, Thomas Pavkov, Eboni Howard, Maria Alvarado, Paul Whiteneir and Diane Durante. 1993b. The Point in Time Multiple Service Use of Illinois Families with Children. Chicago, IL: Chapin Hall Center for Children.
Goerge, Robert, Bong Joo Lee, Teresa Sommer, John Van Voorhis, Lucy Mackey, and Eboni Howard. 1993c. A Longitudinal Analysis of Public Aid and Child Welfare Experiences of Illinois Families with Children. Chicago, IL: Chapin Hall Center for Children.
Heymann, S. Jody, Alison Earle, and Brian Egleston. 1995. Parental Availability for the Care of Sick Children. Boston: Harvard Medical School, Department of Health Care Policy.
Iowa Department of Human Rights, Division of Community Action Agencies, Bureau of Community Services. 1996. FaDDS: Family Development and Self-Sufficiency Grant Program Fiscal Year 1995 Annual Report, Executive Report. Des Moines, IA: Iowa Department of Human Rights.
Legal Action Center. 1995. State, Local Welfare Officials See Important Role for Drug and Alcohol Treatment in Welfare Reform. Washington, D.C.: Legal Action Center, August 10.
Lloyd, Susan. 1995. The Effects of Domestic Violence on Female Labor Force Participation. Paper presented at the 17th Annual APPAM Research Conference. Washington, D.C: November 3.
Loprest, Pamela and Gregory Acs. 1996. Profile of Disability Among Families on AFDC. Washington, D.C.: Urban Institute Report to the Henry J. Kaiser Family Foundation.
Maynard, Rebecca A. 1995. "Subsidized Employment and Non-Labor Market Alternatives for Welfare Recipients." In The Work Alternative: Welfare Reform And The Realities Of The Labor Market, edited by Demetra Smith Nightingale and Robert H. Haveman. Washington, D.C.: The Urban Institute Press.
McNeil, J.M. 1993. "Americans with Disabilities: 1991-92." Current Population Reports P70-33. Washington, D.C.: U.S. Bureau of the Census.
Meyers, Marcia K., Anna Lukemeyer, and Timothy M. Smeeding. 1996. "Work, Welfare, and the Burden of Disability: Caring for Special Needs Children in Poor Families." Syracuse, NY: Center for Policy Research, Maxwell School of Citizenship and Public Affairs, Syracuse University, Income Security Policy Series, Paper No. 12.
Moffitt, Robert and Barbara Wolfe. 1992. "The Effect of the Medicaid Program on Welfare Participation and Labor Supply." The Review of Economics and Statistics: 615-626.
Moore, Kristin A., Martha J. Zaslow, Mary Jo Coiro, and Suzanne M. Miller. 1996. How Well Are They Faring? AFDC Families with Preschool-Aged Children in Atlanta at the Outset of the JOBS Evaluation. Washington, D.C.: U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation.
National Center on Child Abuse and Neglect. 1993. National Child Abuse and Neglect Data System. Working Paper 2. 1991 Summary Data Component. Washington, D. C.: DHHS/ACYF/NCAAN.
Newacheck, P. and W.R. Taylor. 1992. "Childhood Chronic Illness: Prevalence, Severity, and Impact." American Journal of Public Health 82: 364-371.
Nightingale, Demetra S., Regina Yudd, Stacey Anderson, and Burt Barnow. 1991. The Learning Disabled in Employment and Training Programs. Washington, D.C.: U.S. Department of Labor.
Pavetti, LaDonna Ann. 1993. The Dynamics of Welfare and Work: Exploring the Process By Which Women Work Their Way Off Welfare. Cambridge, MA: Harvard University. Dissertation.
Pavetti, LaDonna A. and Amy-Ellen Duke. 1995. Increasing Participation in Work and Work-Related Activities: Lessons From Five State Welfare Reform Demonstration Projects. Report prepared for the Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services. Washington, D.C.: The Urban Institute.
Pavetti, LaDonna, A. 1995. A...And Employment for All? Lessons from Utah's Single Parent Employment Demonstration Project." Paper presented at the 17th Annual Research Conference of The Association for Public Policy Analysis and Management. Washington, D.C.: November 2-4.
Pelton, Leroy. 1989. For Reasons of Poverty. New York: Praeger Press.
Polansky, Norman, James Gaudin and Allie Kilpatrick. 1992. Family Radicals. Children and Youth Services Review, 14: 19-26.
Quint, Janet C., Barbara L. Fink, and Sharon L. Rowser. 1991. New Chance: Implementing a Comprehensive Program for Disadvantaged Young Mothers and Their Children. New York: Manpower Demonstration Research Corporation.
Raphael, Jody 1995. Domestic Violence: Telling the Untold Welfare-to-Work Story. Chicago, IL: Taylor Institute.
Renwick, Rebecca and Marta Krywonis. 1992. "Personal and Environmental Factors Related to Employment: Implications for Substance Abuse Intervention." Journal of Rehabilitation: 23-28.
Riccio, James; Stephen Freedman and Kristen S. Harknett. 1995. Can They All Work? A Study of the Employment Potential of Welfare Recipients in a Welfare-to-Work Program. New York, New York: Manpower Demonstration Research Corporation.
Sedlak, Andrea (1988). Study Findings: Study of National Incidence and Prevalence of Child Abuse and Neglect: 1988. Washington, D.C.:DHHS/ACYF/NCAAN.
Sedlak, Andrea (1990). Technical Amendment to the Study Findings — National Incidence and Prevalence of Child Abuse and Neglect: 1988. Washington, DC: DHHS/ACYF/NCAAN.
Sisco, Carol B. and Carol L. Pearson. 1994. "Prevalence of Alcoholism and Drug Abuse among Female AFDC Recipients." Health and Social Work 19:75-77.
Schmidt, L. 1992. Shifting moral categories of the poor: Addiction and mothers on welfare. Paper presented at the Alcohol Epidemiology Symposium (Toronto, 1992) cited in Sisco and Pearson, 1994.
Straus, Murray A. And Richard J. Gelles. "How Violent are American Families? Estimates from the National Family Violence Resurvey and Other Studies." in Physical Violence in American Families. Edited by Murray A. Straus and Richard J. Gelles. New Brunswick, NJ: Transaction Publishers.
U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation and the National Institute on Drug Abuse. 1994a. Patterns of Substance Use and Program Participation. Washington, D.C.: DHHS/ASPE.
U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation and the National Institute on Drug Abuse. 1994b. Patterns of Substance Abuse Among Women and Parents. Washington, D.C.: DHHS/ASPE.
U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation and the National Institute on Drug Abuse. 1994c. Patterns of Substance Use and Substance-Related Impairment Among Participants in the Aid to Families with Dependent Children Program (AFDC). Washington, D.C.: DHHS/ASPE.
Wolfe, Barbara L. and Steven C. Hill. 1995. "The Effect of Health on the Work Effort of Single Mothers." The Journal of Human Resources 30: 42-62.
Zill, Nicholas, Kristin A. Moore, Christine Winquist Nord, and Thomas Stief. 1991. Welfare Mothers as Potential Employees: A Statistical Profile Based on National Survey Data. Washington, D.C.: Child Trends, Inc.
NOTES
1 Recipients with physical or mental health problems or recipients caring for a disabled child are currently exempted from participation in the JOBS program. Poorly educated, never-married mothers who first begin receiving welfare when they are teens are at the highest risk of long-term welfare receipt. African-American and larger families also have longer than average stays on the welfare rolls.
2 This is based on one randomly selected child in each family, not necessarily the child with the disability.
3 Family illness burden includes bed-days and days where activities were limited for children 0 - 4 (multiplied by 5/7ths to estimate effect on five-day work week), school loss days for older children, and work loss days for parents. 47 percent of poor families faced a family illness burden of less than one week, and 16 percent faced a burden of one to two weeks.
4 A working parent is defined as one who reported working 20 or more hours per week and was not self-employed. Poor families were defined as any family whose income was less than 125 percent of the poverty line.
5 Many of these measures do reflect emotional or mental as well as physical limitations on a child's activities. However, particularly for young children, it is difficult to diagnose disorders, for example learning disabilities, or to translate conditions into functional limitations for a child not yet in school.
6 Daily substance abuse may or may not constitute abuse depending on exactly how much alcohol is consumed. However, it is certainly one indicator of frequent use and possible dependency.
7 Alcoholics were identified by the SMAST test and possible alcoholics were those who had experienced (at least one) alcohol-related social problem such as work sanction, an arrest for drunk driving, relationship problems, neglect of family responsibilities, and hospitalization or counseling because of drinking. Drug abusers were identified by DAST scores and drug users were those who had either used illicit drugs or had experienced a drug-related social problem such as a work sanction, an arrest for illegal activities, a physical fight, relationship or family problems, medical illness, or withdrawal symptoms.
8 Estimates of prevalence rates were reduced by about 10 percent from these levels in a technical amendment to the study findings (Sedlak, 1990). The technical amendment did not report revised figures for low-income families.
9 There were 4,375,000 families and 8,515,000 children receiving AFDC in an average month in 1991 (Ways and Means, 1993).
10 One would expect the rate of indicated cases to be lower than the rate of overall child abuse and neglect, because indicated cases are limited to those cases that are reported, investigated and found worthy of referral to child welfare for follow-up evaluation and services. The higher overall rate in the Illinois study (4.4 percent) compared to the national incidence study (1.6 to 2.5 percent) is largely a result of the four-year period of analysis.
11 Data on all of these issues are not collected every year. When questions are asked in more than one year, data is used for the year closest to 1991. In 1980, all recipients were given the Armed Services Vocational Aptitude Battery (ASVAB) from which the AFQT (Armed Forces Qualifying Test), a common measure of achievement and basic skills, is derived. Data on medical problems of the household head are taken from the 1991 survey. Data on medical problems among children are taken from the 1992 Child Supplement. Data on alcohol use are taken from the 1989 survey while data on drug use come from the 1992 survey and data on depression come from the 1993 survey. Estimates of the prevalence of each of these issues among the welfare population in the year in which they were asked are almost identical to the numbers presented in the tables that follow.
12 A child is identified as having a chronic medical condition if he or she has any of the following: a learning disability, minimal brain dysfunction, hyperkinesis, hyperactivity; asthma; respiratory disorder/sinus infection; speech impairment; serious seeing difficulty; serious emotional disturbance; allergic conditions; crippled, handicapped, or orthopedic problem; mental retardation; heart trouble; chronic nervous disorder; chronic ear infections/problems; blood disorder/immune deficiency; and epilepsy or seizures.
13 This data is for the AFDC caseload at a point in time. Because the NLSY follows a cohort of respondents over time, the data presented here only represent the experiences of the portion of the AFDC caseload that was between the ages of 27 and 35 in 1991. Women in this group account for about 40 percent of the overall AFDC caseload. While not representative of the total AFDC caseload, this group represents a substantial fraction of those recipients who will be most affected by the proposed policy changes. Twenty-three percent of this group of recipients received welfare for 24 months or less; 29 percent for 25 to 60 months and 49 percent for more than 60 months.
14 Acs and Loprest also estimate models including only single-parent families to test the hypothesis that demands on time would be greater for these families but results were insignificant.
15 In contrast to the data presented in Table 5, this data is for all recipients who ever used welfare rather than for the caseload at a point in time. Using data for this group of women provides a larger sample size on which to base the comparisons that are presented here. Forty-five percent of women in the NLSY sample who ever received welfare did so for 24 months or less; 27 percent for 25 to 60 months and 27 percent for more than 60 months.
16 These estimates are derived using the information provided in Tables 5 and 9. For example, the 25 percent estimate is based on two pieces of information: 53.70 percent of the caseload has a serious barrier including low skills and 46.58 percent of this group did not work at all in the current or prior year (53.70*.4658=25.01). The higher estimates that take into account the fact that most recipients work only intermittently use the information on full-year employment at the bottom of Table 9. The 51 percent estimate, for example, is based on the fact that 53.70 percent of the caseload has a serious barrier including extremely low skills and that only 5.87 percent (53.42*.1099) worked for a full-year (53.70*(1-.0587)=50.55).
TABLES
Table 1
Table 2
Table 3
Table 4
Table 5
Table 6
Table 7
Table 8
APPENDIX
Table 9
Latest Research
Print this page