Review of Sanction Policies and Research Studies
Final Literature Review
Department of Health and Human Services Office of the Assistant Secretary for Planning and Evaluation
Project Officer: Elizabeth Lower-Basch
Mathematica Policy Research, Inc. 600 Maryland Ave., SW, Suite 550 Washington, DC
Project Director: LaDonna Pavetti
March 10, 2003
This report is available on the Internet at:http://aspe.hhs.gov/hsp/TANF-sanctions03
- Review of state sanction policies
- Review of Research Findings
- Synthesis of Findings and next Steps
- Key Dimension of State Sanction Policies
- Other TANF Work-Related Policies Influencing the Role of Sanctions in Welfare Reform
- Reported Rates of Sanctioning
- Characteristics More Prevalent Among Sanctioned than Non-sanctioned TANF Recipients
- Barriers to Employment Experienced More Commonly by Sanctioned than Non-sanctioned TANF Recipients
- Employment Rates and Returns to TANF Assistance for Sanctioned and Non-sanctioned Clients
- Conditions Experienced More Commonly by Sanctioned than Non-sanctioned TANF Recipients
The 1996 Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) made unprecedented changes to the welfare system in the United States, eliminating the 60 year-old AFDC program and replacing it with a block grant to states to create the Temporary Assistance for Needy Families (TANF) program. A system that once emphasized the accurate delivery of cash benefits is now focused on encouraging families to make the transition from welfare to work. As a part of this shift, the range of circumstances in which families' welfare benefits can be reduced or canceled has dramatically increased. In particular, sanctions--financial penalties for noncompliance with program requirements--have become central features of most states' TANF programs. The primary goal of sanctions is to convince clients that there are immediate consequences associated with the decisions they make. Sanctions have long been used to enforce program requirements and, with the emergence of "full-family" sanctions that remove all of a family's cash grant, have taken on a much greater significance.
Although there is a general consensus that sanctions have been one of the most important policy changes implemented through state welfare reform efforts, they are among the least studied. In this paper, we summarize what is known about the role they play in welfare reform. The first section is a review of state TANF sanction policies. In this section, we use existing information to describe the structure and stringency of work-oriented sanctions, their cost, the context in which they are applied, and strategies to encourage compliance. The second section is a review of research findings on sanctions--including the incidence and duration of sanctions, characteristics and circumstances of sanctioned families, and the impacts and the implementation of sanctions. The final section concludes with a summary of the gaps in our knowledge of the role of sanctions in welfare reform.
Review of state sanction policies
The Structure and Stringency of Work-Oriented Sanctions
States have used the flexibility provided under TANF to develop different approaches to sanctioning. Table 1 describes four key dimensions that capture the structure and stringency of various state sanction policies: (1) the type of sanction, (2) its minimum duration, (3) the requirements to reverse it, and (4) approach to multiple instances of noncompliance. State policy choices on each of these dimensions are presented in Appendix A and B.
|Dimension||State Approaches||Number of States|
|Type of Sanction||Partial||15|
|Pay for Performance||1|
|Minimum Duration||No minimum, until compliance||28|
|Cure Requirements||Willingness to comply||9|
|Period of compliance||26|
|Repeated Noncompliance||More stringent sanction type||10|
|Longer minimum duration||32|
|Stricter cure requirements||24|
|Reapplication for benefits||24|
|Lifetime ban on assistance||7|
|Source: Welfare Rules Database, Urban Institute 2000; State Policy Documentation Project.|
Type of Sanction. States have implemented four different types of sanctions: (1) partial; (2) gradual full-family; (3) immediate full-family; and (4) pay-for-performance. Fourteen states and the District of Columbia have implemented a partial sanction. As the name implies, when a partial sanction is imposed, a family's cash assistance grant is reduced but they continue to receive some portion of their benefits. The most common approach to imposing a partial sanction is to eliminate the noncompliant adult(s) from the grant, as all states did prior to the implementation of welfare reform. Some states that impose a partial sanction have deviated from this structure and instead reduce the family's grant by a specified percentage.
Seventeen states have implemented an immediate full-family sanction. When such a policy is in place, a family loses all of their cash assistance soon after they are identified as being noncompliant. In some states, these cases become "zero-grant" cases and are counted as part of the TANF caseload. In most states, the case is closed with a sanction closure code, enabling reviewers to distinguish families exiting TANF because of a sanction from those who have left for other reasons.
Nineteen states have implemented either gradual full-family or pay-for-performance approaches to sanctions. These approaches include elements of both partial and full-family sanctions. Under a gradual full-family sanction policy, failure to comply with work requirements leads to a grant reduction for a period ranging from one to six months, depending on the state. If a family comes into compliance before the end of this period, their full grant is restored. If they remain noncompliant at the end of this period, the entire grant is eliminated. Pay for performance, implemented only in Wisconsin, can resemble either a partial or full-family sanction, depending on whether a family is fully or partially noncompliant. Under this model, a family receives assistance only for the hours they participate in required work activities. If they do not participate at all, they do not receive any assistance and the policy operates in the same manner as an immediate full-family sanction. However, if they participate, they receive payment for those hours and the policy functions like a partial sanction.
Sanction Duration. States have taken two different approaches to deciding how long a sanction must remain in place after it is imposed. These approaches represent different philosophies about how to encourage and reward program compliance. Twenty-eight states immediately lift a sanction once a family comes into compliance with program requirements, in order to maintain an immediate connection between a family's choices and the receipt of benefits. Twenty-three states impose a minimum sanction period ranging from one to three months for a first instance of noncompliance. This is often implemented for practical as well as philosophical reasons- it acknowledges that it usually takes some period of time to restore a grant once it has been reduced or eliminated, and has as a philosophical aim that reminds clients that there are consequences for noncompliance. A minimum sanction period was used under the pre-TANF Job Opportunities and Basic Skills (JOBS) training program in an effort to eliminate a "revolving door" effect, where families were believed to move in and out of sanction status frequently.
Cure Requirements. States must also decide what a family must do in order to come into compliance, usually referred to as "curing" the sanction. States have taken two different approaches to defining cure requirements. Nine states require that a family simply indicate their willingness to comply in order to have their grant restored, while 26 states require a family to show actual compliance. (Information is not readily available for the remaining 16 states.) What is needed to demonstrate compliance varies widely, ranging from as little as two days of participation in Arizona to 30 days of compliance in South Carolina and New York. In a few states, the cure requirement depends on the nature of the noncompliance.
Approaches to Repeated Incidence of Noncompliance. Ten states impose a more stringent type of sanction if a family moves in and out of sanction status--for example moving from a partial to a gradual full-family sanction or from a gradual to an immediate full-family sanction. Thirty-two states increase the stringency of the sanction by imposing a longer minimum duration, often up to six months, and stricter cure requirements. In almost half the states, families with multiple sanctions must reapply for benefits rather than having their case reinstated. In seven states, multiple sanctions can lead to a lifetime ban on benefits.
The "Cost" of Work-Oriented Sanctions
It is more complicated to examine the cost of a sanction to a family and what might affect their response to it than to examine states' policies. All else equal, a full-family sanction imposes a higher penalty for non-compliance than a partial sanction, and thus provides a greater incentive to comply. However, because TANF grants vary from state to state, the amount of the family's penalty depends not only on the type of sanction but also on the amount of the maximum benefit available for a particular family size. TANF sanctions can also affect the receipt of other public benefits, potentially increasing the cost of the sanction (see Appendix C-1). States have the option to terminate Medicaid coverage for non-pregnant adults, and 19 states have chosen to do so. If an individual is subject to both TANF and food stamp work requirements, a state must disqualify the individual from food stamps and has the option to disqualify the whole family if all children in the household are over the age of six; 16 states have elected this latter option. Under federal public housing and Section 8 certificate or voucher rules, families losing income due to work-related sanctions cannot qualify for the rent reductions that would otherwise occur. Finally, in some states, families who are sanctioned are not eligible for subsidized child care.
In five states, all with partial sanctions, the most stringent financial penalty under TANF associated with a work-related sanction is $100 or less (see Figure 1). One of the five states strengthens this sanction by disqualifying the whole family from receiving food stamps and one terminates the adult's Medicaid (see Appendix C-2). At the other extreme, in fifteen states, the financial cost of a work-oriented sanction is more than $400. All of these states impose a gradual or immediate full-family sanction. Almost one-half of the states that eliminate the full food stamp grant fall in this category, as do about one-third of the states that eliminate the adult's Medicaid. Data on what happens to child care for sanctioned families is not available for all states. In 13 of the 23 states where information is available, sanctioned families retain eligibility for child care; in the other ten states, sanctioned families either lose their eligibility or must have it redetermined.
Figure 1. Financial Cost of Most Stringent Work-Related Sanctions for a Family of Three
The Interaction with Other State TANF Work-Related Policies
The role sanctions play in welfare reform may be influenced not only by the structure and cost of sanctions, but also by the context in which they are applied. There are two factors that appear particularly important in understanding the reach and imposition of sanctions. The first is the presence of any pre-approval work-related requirements, and the second is the state's approach to exemptions (see Table 2).
Pre-approval work-related requirements are important to consider because their use may result in a state not opening cases that in another might be opened and eventually sanctioned. When benefits are denied because an applicant does not complete the pre-approval requirements, these requirements function almost as an immediate full-family sanction, although the family never receives TANF benefits. More families could end up being sanctioned in states that do not have pre-approval requirements simply because more families were successful in opening a TANF case. Most states have implemented at least one of these pre-approval options--including attending an orientation, signing an employability plan or a personal responsibility contract, and completing applicant job search requirements--and many have implemented more than one (see Appendix D).
A second contextual issue that is important in understanding the role of sanctions in welfare reform is the proportion of the caseload subject to work requirements. When considered for the caseload as a whole, a low incidence of sanctions could signal many things, including the possibility that relatively fewer people are subject to work requirements and therefore at risk of being sanctioned.
What work requirements a state imposes upon families with young children is a key determinant of the proportion of the caseload that is potentially at risk of being sanctioned. The majority of states (28) require all families with a child over the age of one to participate in work activities (see Table 2). Four states require all families to participate and 13 require families with children over the age of four months to participate. Only 6 states do not require families to participate until their youngest child is older than 12 months
A second factor that influences the proportion of the caseload that is required to participate in work activities and, therefore, subject to sanctions, is a state's treatment of families facing various personal and family challenges. In an effort to understand the choices states were making regarding participation in program activities for these families, several years ago, the Urban Institute classified states into three groups based on their treatment of people who had previously been exempt from program participation because of their own or a family member's disability or illness (Thompson et al. 2000). Using their classification, thirteen states have adopted universal work participation requirements, meaning that they expect nearly everyone to participate in program activities; in many cases the activities that count towards participation in these states are broader than in other states, including activities such as substance abuse and mental health treatment. In these states, case managers also often have considerable flexibility in defining the amount and types of required activities and in granting exemptions from the work requirements. The remaining states that could be categorized fell into two groups of equal size; the first group continued to use the pre-TANF JOBS program participation categories and the second adopted participation requirements that were broader than the JOBS program, but did not expect everyone to participate. In states that use the JOBS criteria for identifying who is required to participate in work activities, families are exempt from program participation if they are ill or capacitated or caring for an ill or capacitated household member. In these states, staff generally relies on local medical professional to document a recipient's inability to participate. States that adopted broader participation requirements than the JOBS program often grant fewer formal exemptions but still maintain a process for identifying recipients who may experience various personal or family challenges that might interfere with their ability to find or sustain employment.
|TANF Work-Related Policy||Number of States|
|Pre-employment Requirements (any)||28|
|Signed Employability Plan||6|
|Signed Personal Responsibility Contract||14|
|Applicant Job Search||16|
|Exemptions for Recipients Caring for Young Children|
|3 months or younger||13|
|12 months or younger||28|
|over 12 months||6|
|Approach to Exemptions for Disabled Recipients|
|JOBS Participation Requirements||17|
|Broader Participation Requirements||17|
|Universal Participation Requirements||13|
|Source: Pavetti et al. 1998; State Policy Documentation Project; Thompson et al. 1998; Welfare Rules Database, Urban Institute 2000|
Key Implementation Tasks
Putting sanction policies into practice is a complex endeavor. In order to impose sanctions in a meaningful way, states must create the foundation for a mandatory employment program. This includes determining who is required to participate in work activities and to which activities recipients will be assigned, ensuring that sufficient slots are available, developing systems for monitoring compliance and initiating the sanction process, and implementing strategies for encouraging compliance.
Within this framework, the implementation of TANF sanctions involves six key tasks. The first is clearly informing recipients about what is expected of them and the consequences for not following through. This typically begins during an initial orientation and may be repeated throughout the service delivery process. The second is identifying clients who may be unable to participate in program activities because they lack child care or transportation or face other personal and family challenges that may make participation difficult. TANF agencies use a variety of approaches to assess recipients' needs and identify those who may be unable to participate in regular work activities. While some use an up-front assessment to make this determination, others use the labor market as a test, granting an exemption only after a recipient has been unsuccessful in participating in employment activities or in finding employment. Due to the nature of the personal and family challenges faced by TANF recipients (e.g., domestic violence, substance abuse, mental health issues, learning disabilities), the assessment and exemption process often is an ongoing one.
The third task is monitoring participation in required activities. Because TANF recipients may be participating in a variety of activities provided by a broad range of providers, completion of this task often requires extensive coordination and communication between multiple agencies. Because many TANF recipients experience what are sometimes referred to as "hidden barriers" to employment, issues that might affect a recipient's ability to participate may not surface until after participation commences. In some cases, these circumstances may provide "good cause" for not meeting participation requirements. In others, they might be severe enough to result in an exemption from program requirements. The distinction between good cause and exemptions in the current TANF environment is an important one. When a TANF recipient is exempted from program requirements, she is no longer required to participate in program activities. In contrast, "good cause" often is used to grant an "excused absence" from participating in program activities, usually for a limited period of time. Some common reasons for granting "good cause" include attendance at medical, school or court appointments, presence of an unstable housing situation, and participation in mental health or substance abuse treatment.
The fourth task involves defining what will trigger the start of the sanctioning process. For example, a sanction may be imposed immediately upon non-compliance or after a specified number of weeks of non-compliance. What triggers a sanction may be different for initial and subsequent sanctions. The fifth task involves defining the actual process for imposing a sanction, including the timing and content of any notices that will be sent to alert recipients that the sanction will be imposed and when the sanction will actually take effect. The final task involves establishing a process for re-engaging sanctioned recipients in program activities. At a minimum, this would include the specific requirements for curing the sanction, but may also include procedures for appealing a sanction decision and development of outreach or other strategies to encourage families to come into compliance.
To date, no systematic data has been collected on how sanctions are being implemented, probably because many of the key decisions are left to local offices, making it to difficult to collect this information in a cost-effective manner. There may be considerable variation from office to office, and possibly from one worker to another within the same office. Factors that might influence the implementation of sanctions include the "culture" of the welfare office, especially the strategies used to communicate the importance of work and to encourage compliance, staff workloads and the complexity of the service delivery system. In an ideal world, no one would actually bear the cost of being sanctioned because the threat of a sanction itself would encourage program compliance, either through active participation in job search or other work-related activities or by more accurate reporting of earnings by recipients who are already employed. However, early evidence shows that sanctions do not work in this way for many TANF recipients, making implementation procedures a critical component of understanding the role sanctions play in welfare reform.
Review of Research Findings
This review of the literature covers five primary aspects of TANF sanctions: (1) the incidence and duration of sanctions, (2) the characteristics of sanctioned families, (3) the circumstances of sanctioned families; (4) the impact of sanctions on various outcomes, and (5) the implementation of sanctions. We gathered and reviewed all of the known studies that examine these aspects of sanctions. Many of these studies focus specifically on sanctions but in some cases, sanctions are examined as a part of a larger study. This review provides a context for assessing various dimensions of TANF sanctions, and identifying important gaps in the information available on the role sanctions play in encouraging TANF recipients to meet the program's employment goals.
The Incidence and Duration of TANF Sanctions
Information on the incidence and duration of sanctions can improve our understanding of the role they play in welfare reform in two different ways. It can provide some insight into the extent to which sanctions are used to encourage compliance with work and other program mandates. It can also provide information on the number of families who may be at higher risk of various hardships because their financial resources have been reduced.
State data suggest that, at least in some states, large numbers of families have been sanctioned for failure to comply with program requirements. For example, in Virginia, in State Fiscal Years (SFY) 1996 and 1997, 3,777 families were sanctioned for failure to sign an Agreement of Personal Responsibility or participate in the state's work program (Gordon and Agodini 1999). In Indiana, in SFY 1995, 7,810 sanctions were imposed on families, up from 917 the previous year (Holcomb and Ratcliffe 2000). During a six month window in 1994 and 1995, 4,200 families were assigned to Iowa's Limited Benefit Plan (the equivalent of a full-family sanction with a six-month minimum period), although slightly more than half of these assignments were canceled before the family lost cash assistance because they came into compliance with the state's work requirement (Fraker et al. 1997).
- Owing to differences in methodology, studies reported a wide range of estimates of the incidence of sanctions.
Rates of sanctioning, rather than raw numbers, are needed to compare the incidence of sanctions across states and across time. Studies that examined the incidence of sanctions reported rates ranging from 5 to 60 percent (see Table 3). The differences in these estimates appear to primarily reflect variations in the methodology used to calculate sanction rates. The studies we reviewed used three different groups as the base for estimating the prevalence of sanctions: (1) current TANF recipients, (2) a cohort of current or new recipients followed over time, and (3) closed TANF cases. Studies that used the current caseload as the base reported the lowest sanction rates; cohort studies reported the highest.
The U.S. General Accounting Office (GAO 2000a) and Koralek (2000) calculated the incidence of sanctions as the fraction of the current caseload in sanction status. Using this methodology, GAO concluded that the "proportion of TANF families who actually lose part or all of their TANF cash benefits as a result of sanctions is not large." During an average month in 1998, about 135,800 families, or five percent of the TANF caseload, received reduced or no TANF benefits as a result of sanctions for failure to comply with TANF and other work responsibilities. GAO's state estimates ranged from a low of 1 percent in Rhode Island to a high of 29 percent in North Carolina. In states that impose only a partial sanction, this methodology provides a meaningful measure of the prevalence of sanctions. However, for states that impose immediate or gradual full-family sanctions, it does not. In a follow-up letter to the study requesters, GAO (2000b) acknowledged that the estimates of full-family sanctions provided to them by the states counted the sanction during the first month that the sanction was imposed; however, these cases were not included on an ongoing basis. Therefore, the figures presented by GAO do not reflect the cumulative effect of sanctions. Most likely, this resulted in a substantial underestimation of the number of families affected and the extent to which sanctions are used to encourage compliance with work mandates. Similarly, GAO's conclusion that most sanctions are partial rather than full-family may not hold true upon examination of data on all sanctions over time.
The sanction rates reported in studies of closed cases range from 10 to 28 percent (see Table 3). These studies used the administrative code for case closure to calculate the fraction of cases that have closed due to sanctions, providing a meaningful measure of the number of families who may be at higher risk for various hardships because their cash grants have been eliminated. However, they provide an incomplete picture of the extent to which sanctions are being used to encourage compliance. Missing from these data are families who are currently subject to a partial sanction and those who were previously subject to a partial or full-family sanction but have since reversed it. Ovwigho et al. (2002) suggest that, at least in Maryland, the use of full-family sanctions has increased over time. Among early cohorts of leavers in Maryland (October 1996 to March 2001), only 10 percent of cases were closed because of a full-family sanction. Among later cohorts (April 2001 to March 2002), this rate almost doubled, increasing to 18 percent. South Carolina's 29 percent sanction rate may be higher than that of other states because the estimate is calculated only for families subject to the work requirement.
The highest estimates of the incidence of sanctioning are reported in studies that follow a cohort of recipients or new applicants over time. Fein and Lee (1999) found a sanction rate of 60 percent for all sanctions, and 52 percent for work-related sanctions for a random sample of TANF cases followed over an 18-month period. Holcomb and Ratcliffe (2000) reported a sanction rate of 45 percent for work-related sanctions for a random sample of new cases followed over a 12-month period. While they are not without problems, these estimates provide the most reliable picture of the extent to which sanctions are imposed, and the most complete accounting of the number of families who have experienced sanctions. Still, these estimates may not account for all recipients who are ever affected by a sanction. Sanction notices, intended to warn clients of a pending sanction, motivate some to participate who may not have otherwise. Recipients who respond to these notices would be affected by a sanction, but would not be counted as a sanctioned TANF recipient.
|Study||State||Sanction Type||Base||Period of Time||Sanction Rate (%)|
|GAO (2000a)||Nationwide||TANF caseload||Average month in 1998||5|
|Koralek (2000)||South Carolina||Immediate Full Family||TANF caseload||Average month during 1998-1999||16|
|Born et al. (1999)||Maryland||Immediate Full Family||Closed cases||Cases closed in Oct 1996-Mar 1998||7.3|
|Ovwigho et al. (2002)(1)||Maryland||Immediate Full Family||Closed cases||Cases closed in Apr 2001-Mar 2002||18.3|
|Edelhoch et al. (2000)||South Carolina||Immediate Full Family||Closed cases||Cases closed in Oct 1996-Mar 1997||28|
|Westra and Routley (2000)||Arizona||Gradual Full Family||Closed cases||Cases closed in Jan 1998-Mar 1998||20|
|Fein and Lee (1999)||Delaware||Gradual Full Family||Cohort||Longitudinal Dec 1996-June 1998||60|
|Holcomb and Ratcliffe (2000)||Indiana||Partial||Cohort of new cases||Longitudinal May 1996-Apr 1997||45|
Unfortunately, we no of no studies that have been conducted to examine the proportion of TANF families who receive a sanction notice and comply before the sanction is imposed.
- Sanctions tend to be imposed shortly after TANF clients begin receiving benefits.
TANF clients appear to be most likely to be sanctioned shortly after they begin receiving assistance, a pattern that is consistent with a work-first philosophy. Tracking the patterns of sanctions, researchers found that most occur within the first three months of program entry (Fein and Lee 1999; Koralek 2000; Holcomb and Ratcliffe 2000). Studying patterns of TANF sanctions in Indiana, Holcomb and Ratcliffe (2000) reported that 56 percent began within the first three months, and 81 percent by six months. Fein and Lee (1999) found that 43 percent of sanctions occurred within the first month of enrollment.
- Many TANF recipients cure their sanction and remain in sanction status for relatively short periods of time.
Reporting findings from data collected in seven states, the GAO (1997) concluded that of those sanctioned, 18 to 47 percent return to welfare. Another GAO study (2000a) estimated that in ten states, an average of about one-third of sanctioned clients came into compliance after receiving a full or partial sanction. This finding is consistent with other studies examining the percentage of clients who return to TANF after a sanction (Fein and Lee 1999; Fraker et al. 1997; Holcomb and Ratcliffe 2000). In Iowa, Fraker et al. (1997) found that 53 percent of sanctioned clients reversed their sanction. In Indiana, Holcomb and Ratcliffe (2000) documented that 55 percent come into compliance within a year after the sanction is imposed. Of those who reverse their sanction, 28 percent do so in the first month and about two-thirds reverse their sanction within three months after the sanction is imposed. Two-thirds of TANF recipients interviewed in the "Three City Study" conducted in Baltimore, Boston and San Antonio reported trying to get their benefits back, and about half of all sanctioned clients were successful in doing so (Cherlin et al. 2001).
- Most sanctioned clients are sanctioned only once, but a modest fraction are sanctioned, come into compliance and then are sanctioned again.
Most clients reverse their sanction and return to cash assistance, get a job, or remain off of assistance and seek alternative sources of support. However, a modest fraction of sanctioned clients cure their initial sanction, and then are sanctioned again after failing to comply with program activities a second or third time. Tracking the dynamics of welfare sanctions, Holcomb and Ratcliffe (2000) estimated that about one-fifth of sanctioned clients received more than one sanction in a year-long period. Nixon et al. (1999) found that in Iowa about one-quarter of sanctioned clients were sanctioned more than once. These were a more hard-to-employ group of TANF recipients than those who experienced only one sanction. According to clients who repeat sanctions, personal and family challenges, poor client-case manager communication, lack of transportation, and lack of child care contribute to their repeated problems with noncompliance.
Characteristics of Sanctioned Families
In an effort to better understand the use and implications of sanctions, several studies examined the characteristics of sanctioned recipients, and most compared them to non-sanctioned recipients. Only a few studies collected data explicitly for this purpose; most relied on the administrative data available on all TANF recipients or for a subset, such as closed cases. These studies--particularly when they compare the characteristics of sanctioned and non-sanctioned recipients--provide some insight into the families affected and whether certain characteristics make TANF recipients more likely to be sanctioned. Of particular interest is whether sanctioned recipients exhibit characteristics that may make it more difficult to comply with program requirements. While there was some variation, most studies found that sanctioned families are more likely than non-sanctioned families to exhibit one or more characteristics that make them harder-to-employ. In these studies, families may not be sanctioned at the time the study was conducted.
- Sanctioned families exhibit many of the characteristics that have traditionally been associated with longer welfare stays.
Studies of welfare dynamics conducted prior to reform found that recipients who were African American, young, never married and poorly educated were more likely to receive welfare for long periods of time (Pavetti 1995). Studies comparing sanctioned and non-sanctioned families found that the former exhibit many of these characteristics (see Table 4). With two exceptions, these studies found African-Americans over-represented among sanctioned families. For example, Kalil et al. (2002) found that more than two-thirds of sanctioned clients were African American, compared to about half of non-sanctioned clients. Several studies found that sanctioned families were more likely to be living in a large household, have never been married or not living with a partner, and be young. According to Westra and Routely (2000), 55 percent of sanctioned TANF clients have never married compared to half of non-sanctioned clients. Two studies, Born et al. (1999) and Koralek (2000) found that, on average, sanctioned clients were about two years younger than those not sanctioned. Hasenfeld et al. (2002) reported that TANF recipients under the age of 24 are somewhat more at risk for sanctions than older recipients. The one study that looked at the age at which a recipient had her first child found that 53 percent of sanctioned clients were 20 years old or younger when they had their first child, compared to 45 percent of non-sanctioned mothers (Born et al. 1999).
|Studies||Location of Study||African American||Never married or not living with a partner||Larger household size/ more children||Young adult||Began childbearing at a young age|
|*Born et al. (1999)||Maryland||ns||x||x||x|
|Cherlin et al. (2001)||Boston, Chicago and San Antonio||ns|
|*Edelhoch et al. (2000)||South Carolina||x||x||ns|
|Fein and Lee (1999)||Delaware||x||x|
|Kalil et al. (2002)||Michigan||x||x|
|Koralek (2000)||South Carolina||x||x||x|
|*Mancuso and Linder (2001)||California||x|
|*Westra and Routely (2000)||California||x||x|
|* Indicates studies that compare sanctioned and non-sanctioned leavers. ns - Variables were included but not significant.|
- Sanctioned TANF recipients are more likely than their non-sanctioned counterparts to be long-term welfare recipients and to experience human capital deficits such as limited education and lack of work history.
Studies of TANF recipients found that human capital barriers are strongly associated with unemployment (Kalil et al. 2002). Studies of sanctioned families consistently found that such barriers are even more common among these recipients (see Table 5). Between 30 and 45 percent of TANF recipients lack a high school diploma or GED (GAO 2000a). Among sanctioned recipients, this proportion is substantially higher--between 44 and 54 percent (Cherlin et al. 2001; Edelhoch et al. 2000; Fein and Lee 1999; Kalil et. al. 2002; Koralek 2000; Mancuso and Lindler 2001; Westra and Routely 2000). Other human capital deficits that affect sanctioned recipients more frequently include limited work experience and lack of job skills. For example, Hasenfeld et al. (2002) found that sanctioned recipients were twice as likely to have not worked in the past three years. Sanctioned recipients are also more likely to have received welfare for long periods of time. Edelhoch et al. (2000) found that sanctioned clients are twice as likely to have received cash assistance for 60 months or longer.
|Study||Human Capital Deficits||Logistical Barriers||Personal and Family Challenges|
|*Born et al. (1999)||x||n/a||n/a|
|Cherlin et al. (2001)||x||x||x|
|*Edelhoch et al. (2000)||x||n/a||n/a|
|Fein and Lee (1999)||x||x||n/a|
|Kalil et al. (2002)||x||x||x|
|*Mancuso and Linder (2001)||x||x||x|
|*Westra and Routely (2000)||x||n/a||n/a|
|* Indicates studies that compare sanctioned and non-sanctioned welfare leavers. n/a - Variables not included in the analysis.|
- Studies consistently found that lack of transportation is more common among sanctioned than non-sanctioned TANF clients. Few studies examined lack of child care as a barrier to employment; of those that did, findings varied.
Logistical barriers, such as transportation and child are common among TANF recipients (Pavetti 2002). Transportation appears to be an even greater barrier for sanctioned recipients (see Table 5). For example, Cherlin et al. (2001) reported that 19 percent of sanctioned clients said someone in their household owned a car, compared to 35 percent of non-sanctioned clients. A study of welfare mothers in Michigan reported that 41 percent of non-sanctioned TANF recipients lacked access to transportation, compared to 59 percent of those sanctioned (Kalil et al. 2002). In California, it was estimated that 79 percent of CalWORKs clients in conciliation cited lack of transportation as a barrier to employment (California Department of Social Services 2001). The two studies that compared access to child care among sanctioned and non-sanctioned families reached different conclusions, possibly because they examined different groups of recipients. In examining sanctioned and non-sanctioned leavers, Mancuso and Linder (2001) found that sanctioned clients were more likely to indicate that child care is a barrier to employment. However, Hasenfeld et al. (2002), examining current recipients, found no difference.
- Sanctioned families are more likely than non-sanctioned families to experience some but not all personal and family challenges.
Personal and family challenges, especially mental health and domestic violence, are common among TANF recipients, especially those who remain on the TANF caseload (Pavetti 2002). Few studies compared the presence of these issues in the lives of sanctioned and non-sanctioned recipients; those that did often found that the results varied depending on the issue examined. Studies consistently found that alcohol and drug problems are greater among sanctioned families (Cherlin et al. 2001; Hasenfeld et al. 2002; Mancuso and Lindler 2001). Mancuso and Lindler (2001) found that almost one-fifth of sanctioned clients may have a drug or alcohol addiction compared to less than 10 percent of non-sanctioned families. Studies that examined the presence of mental health problems found no difference among sanctioned and non-sanctioned recipients. Some studies looking at differences in domestic violence and physical health problems found differences between sanctioned and non-sanctioned clients, while others did not. For example, Kalil et al. (2002) found that 25 percent of sanctioned clients had experienced severe domestic abuse within the last year, almost twice the rate for non-sanctioned clients. Hasenfeld et al. (2002) found substantially lower rates of domestic abuse among sanctioned clients (14 percent) compared to Kalil et al. (2002) and no differences between sanctioned and non-sanctioned TANF recipients. In her study of the health of poor urban women, Polit et al. (2001) found that women with multiple barriers--including physical abuse, risk of depression, a chronically ill or disabled child--were more likely than other recipients to have been sanctioned in the prior year.
Involvement with the child welfare system is quite common among TANF families (Courtney et al. 2001; Needell et al. 1999; and Shook 1999). For example, Courtney et al. (2001) estimate that over half of TANF recipients in Milwaukee County, Wisconsin have been investigated by Child Protective Services. Needell et al. (1999) tracked child welfare involvement of families receiving cash assistance in ten counties in California. They found that 27 percent of children who received aid in 1990 experienced a child maltreatment report within 5 years. While some studies suggest that child welfare involvement may be more common among sanctioned families (Colville et al. 1997; Shook 1999), other studies do not find this relationship. In a pre-TANF study of case closures due to sanctions, Colville et al. (1997) found that sanctioned families were about 50 percent more likely to have contact with protective services, even prior to being subject to work requirements. However, using administrative data and case reviews with 400 TANF recipients in Utah, Derr and Cooley (2002) found that sanctioned and non-sanctioned families both had high rates of child welfare involvement, but that the former were no more likely to have an open child welfare case three years after TANF case closure.(2)
- When accounting for the interaction and influence of various demographic characteristics and potential employment barriers simultaneously, few factors appear to significantly predict whether a family will be sanctioned or not.
Two studies used econometric models to identify the predictors of TANF sanctions. Hasenfeld et al. (2002), controlling for county and ethnicity, identified the following variables as significant risk factors for sanctions: limited work history, lack of transportation, large numbers of children, younger than 24, a non-native English speaker, and self-reported substance abuse. In a similar study of Michigan welfare recipients, Kalil et al. (2002) found fewer predictors of sanction status. Including a range of variables, they found only four that significantly predict sanction status: being African American, not cohabiting, being under the age of 24 or over the age of 35 and having less than a high school education. In both studies, a variety of characteristics and barriers to employment--including access to child care, physical or mental health problems, and domestic violence--were not found to be significant predictors of sanction status.
Circumstances of Sanctioned Families
Sanctions are intended to encourage TANF recipients to comply with program requirements. Some families may respond to the sanction by finding employment. Other families may fail to comply, and have their benefits reduced or eliminated. Some of these sanctioned families may have other sources of support, such as unreported earnings or income from family members, while others may lose their only source of income support. Depending on their circumstances prior to and actions after a sanction, a family may fare worse, the same, or better after being sanctioned. Studies attempting to assess the well-being of sanctioned families examined: (1) employment, (2) their return to the welfare systems, and (3) the presence of hardships. The research available in this area consistently found that sanctioned families are less likely than their non-sanctioned counterparts to be employed, and more likely to return to the welfare system. Fewer studies examined the presence of hardships, but those that did found that sanctioned families experience them at a higher rate.
- Employment rates and earnings are lower for sanctioned than non-sanctioned families.
Studies that examined the employment status of sanctioned families found that between 36 and 55 percent are employed at some point after case closure (see Table 6). Employment status is measured from as little as four months to as long as two years after case closure. However, there does not appear to be any relationship to the level of employment and the period of time over which it is measured. While these rates indicate that a modest fraction of sanctioned recipients may find employment after their TANF case closes or may have been working prior to case closure, their employment rates are substantially lower than those for non-sanctioned recipients. Using administrative data, Born et al. (1999) and Edelhoch et al. (2000) found a 20 percent gap between the employment rates of these groups; Westra and Routely (2002) placed the gap at 10 percentage points. When Westra and Routely (2000) examined information provided by recipients through a telephone survey, the gap widened to 21 percentage points. Born et al. (1999) reported that sanctioned clients in Maryland who were working earned approximately $600 less per quarter than non-sanctioned clients, which may be a result of limited educational attainment and other human capital deficits. Studies that looked at the types of jobs sanctioned recipients hold found that they work in sectors similar to other welfare leavers--food service, clerical, and sales (Edelhoch et al. 2000). According to Fraker et al. (1997), sanctioned TANF recipients in Iowa who were employed after case closure worked an average of 31 hours per week and earned about $170 per week.(3) About one third (36 percent) of those employed reported that health insurance was available through their job; however, due primarily to the cost, only about 11 percent actually received it. Approximately 30 percent of the jobs offered paid sick leave.
- Studies reported a wide range in the estimated number of sanctioned TANF recipients who return to the rolls; most studies found higher rates of return for sanctioned than non-sanctioned families.
If sanctioned recipients are unable to find employment on their own or lack other sources of support, they may return to the TANF rolls in order to meet their basic needs. Estimates on the fraction of sanctioned recipients who return to the TANF rolls range from 19 to 50 percent (see Table 6). The 19 percent estimate comes from a California study (Mancuso and Linder 2001). Because California only imposes a partial sanction, it is possible that families who leave assistance have other reliable sources of income. The 50 percent estimate comes from self-reports from participants in the Welfare, Children & Families Three City Study. About two-thirds said they had tried to get their benefits back and half said they had been able to get them back (Cherlin et al. 2001). Because these are self-reports, they might include people who reversed a sanction before losing benefits as well as those who left the rolls and then returned.
|Study||Period measured||Percentage of clients who are employed||Percentage of clients who return to TANF|
|Bloom and Winstead (2002)||6 months after case closure||55||n/a||45||n/a|
|Born et al. 1999||6 months after case closure||38||58||38||22|
|Cherlin et al. 2001||Varies by site||36||n/a||50||n/a|
|Edelhoch et al. (2000)||2 years after case closure||46||66||36||34|
|Fein and Lee (1999)||4 months after case closure||39||n/a||32||n/a|
|Fraker et al. (1997)||Within 6 months of case closure||53||n/a||20(4)||n/a|
|Mancuso and Lindler (2001)||1 year after case closure||n/a||n/a||19||n/a|
|Westra and Routely (2002)||1 year after case closure||42||52||40||33|
- Sanctioned recipients are more likely to experience material hardships than their non-sanctioned counterparts.
Material hardships TANF recipients face include borrowing money to pay bills or falling behind on payments, not having enough food, problems paying for medical care, and experiencing a utility shut-off, among others (Cherlin et al. 2001; Edelhoch et al. 2000; Fraker et al. 1997; Kalil et al. 2002; Mancuso and Lindler 2001). Cherlin et al. (2001) reported that sanctioned families were twice as likely as non-sanctioned families to say they lack adequate food, and five times as likely to borrow money to pay bills. A quarter of sanctioned TANF clients said they used a food pantry (compared to 19 percent of non-sanctioned clients), and about one-quarter said that they received emergency clothing (compared to 15 percent of non-sanctioned recipients). Based on telephone interviews with TANF recipients, Kalil et al. (2002) discovered that sanctioned clients were twice as likely to experience a utility shut-off (21 percent, compared to 9 percent for non-sanctioned clients). In addition, about one-third of sanctioned clients engaged in hardship activities within the last six months, compared to 14 percent of non-sanctioned clients.(5) About half of sanctioned clients indicated that they expect to experience hardship within the next year, while about a quarter of non-sanctioned clients had the same response. It is important to note that none of these studies establish a causal relationship between sanctions and hardship. In some cases, the family may not have been sanctioned at the time the study was conducted. While it is possible that sanctions may increase the likelihood that a family experiences various hardships, it is also possible that the same characteristics that lead families to be sanctioned may result in greater experience of hardship.
Sanctioned clients access a variety of resources and supports after TANF case closure to address their basic needs. A few studies found that sanctioned recipients often rely on emergency services such as food banks and homeless shelters after TANF case closure (Cherlin et al. 2001; Kalil et al. 2002). Other sources of support include friends and family, or government assistance programs such as food stamps, Medicaid, and Supplemental Security Income. Most sanctioned welfare recipients access Medicaid and food stamps. Estimates of the extent to which they rely on other sources of support vary widely. Edelhoch et al. (2000) reported that about one-quarter received income from someone outside the home. Fraker et al. (1997) found that 65 percent of sanctioned clients received support from their parents. Between 16 and 35 percent of sanctioned clients received regular child support (Edlehoch et al. 2000; Mancuso and Lindler 2001).
Two studies found that the circumstances of sanctioned TANF recipients improve over time. Edelhoch et al. (2000) found that almost half of sanctioned clients were working two years after case closure, compared to about one-fifth at case closure. Interviewing sanctioned clients at 6 and 12 months, Mancuso and Lindler (2001) documented that the resources and general family stability of sanctioned families increased over time.
- One study suggests that sanctioned families are more likely to have an infant or toddler who is hospitalized.
In a study of 2,718 families who received welfare, Cook et al. (2002) examined the potential influence of welfare sanctions on the health and food security of young children. They found that infants and toddlers in sanctioned families have a 30 percent higher risk of having past hospitalizations and a 90 percent higher risk of being hospitalized at the time of an emergency room visit than children in non-sanctioned families. In the same study researchers found that sanctioned families have a 50 percent higher risk of being food insecure than non-sanctioned families.(6) Although the study authors suggest that sanctions may cause these adverse outcomes, it is also possible that the characteristics that make it difficult for a family to comply with welfare requirements may also lead to greater hospitalizations or emergency room visits. For example, a child with a chronic illness such as asthma may make it hard for a family to hold a job or comply with work requirements.
The Impact of TANF Sanctions
Sanctions are intended to change TANF recipients' behavior; the hope is that they will encourage recipients who would not otherwise participate in work activities to do so, leading to higher levels of program participation, increased exits for work and lower TANF caseloads. In the absence of an experiment where families are randomly placed into groups, one that is subject to a sanction and one that is not (or one subject to a more stringent sanctioning policy and one to a more lenient policy), it is difficult to determine what impact sanctions have and whether stricter sanctions produce greater behavioral changes. While we can observe the number of families who have been sanctioned and describe their characteristics, we cannot capture the number and characteristics of families who may have changed their behavior to avoid being sanctioned. In the absence of this information, some studies exploit the variation in state sanction policies to examine whether stricter sanction policies lead to greater TANF caseload declines. The evidence suggests this is the case--stricter sanctions may increase TANF exits--but more research is needed.
- A few studies suggest that more stringent sanctions lead to greater welfare exits and caseload declines, although most offer little insight into how these changes occur.
A study that examined the impact of waiver policies on welfare exits found that more stringent sanction policies are associated with increased employment exits (Hofferth, Stanhope, and Harris 2000). A second study that examined the relationship between a state's sanction policy and the change in its TANF caseload estimated that the presence of an initial full-family sanction is associated with a 25-percent higher caseload reduction rate than that found in states with weak sanctions (Rector and Youssef 1999). A study examining the relationship between welfare reform policies, governmental quality, and caseload changes found qualitatively similar results (Mead 2000).
In an earlier study, Mead (1997) concluded that well-performing welfare offices make program expectations clear and threaten sanctions for non-participation, but rarely need to impose sanctions on recipients. Conversely, welfare offices that do a poor job of clearly stating recipient expectations and perform poorly in job placement and other performance measures frequently sanction recipients. Accordingly, high rates of caseload declines may be due to different office performance--either high performance that encourages recipients to find work, or poor performance wherein expectations are unclear and noncompliance is likely, leading to exits without work.
Findings from the 11 programs in the National Evaluation of Welfare-to-Work Strategies (NEWWS) suggest that programs need to enforce work-related mandates in order to achieve high rates of participation in employment activities. Programs with high levels of enforcement of a participation mandate tended to have higher participation rates than programs with low levels of enforcement. However, within high enforcement programs, researchers found no association between the frequency of sanctions or the length of sanctions and program outcomes (Hamilton and Scrivener 1999). It is important to note that these findings are based on data gathered prior to the passage of PRWORA and included programs that implemented partial rather than full-family sanctions.
Implementation/Local Office Practices
In many states and localities, sanctions--particularly those affecting the full family--are relatively new to case managers and recipients. Work requirements and partial sanctions existed under the JOBS program in many states prior to reform, but few families were subject to them and they were rarely enforced. Enforcement is an important factor in the way sanctions create a work-oriented assistance system, and who is adversely affected (Pavetti and Bloom 2001). Lax enforcement can nullify their effect. Efforts to promote compliance, including providing clear information on sanctions, may encourage families to take appropriate steps to achieve self-sufficiency, and reduce the number who lose benefits due to sanctions.
State and local office procedures for implementing sanctions vary considerably, and may influence the role of sanctions in welfare reform. Thus, these procedures and the decisions that program administrators, intermediaries and case managers make provide an important context in understanding the role of sanctions. To date, there has been scant systematic review of how sanction policies are being implemented in local welfare offices. The research that has been done suggests that recipients often aren't clear about sanction policies, and that implementation of these policies varies considerably.
- Recipients often are not clear about participation requirements, sanction policies, and processes to "cure" a TANF sanction.
Recipients are typically informed about program requirements and sanctions for noncompliance during orientation. In a report that synthesizes findings from existing studies (pre-TANF) about client participation, Hamilton and Scrivener (1999) reported that during any given month 37 percent of the welfare caseload had not completed an orientation, and as a result, may not be informed about sanctions. In addition, the quality of the orientation and the way it is conducted varies considerably within local welfare offices. Some recipients receive a clear message about what is expected of them and the consequences for noncompliance, while others may be less informed.
Even when informed, some TANF recipients do not fully understand what is expected of them. A study of second assignments to Iowa's Limited Benefit Plan found that one-quarter of parents who were sanctioned did not understand the program rules clearly (Nixon et al. 1999). An Inspector General's report of sanctions found that even though local offices explained sanctions to clients repeatedly and in a logical format, many TANF clients did not fully understand them (U.S. Department of Health and Human Services 1999). Clients often knew that they might lose their benefits if they didn't do what was expected of them, but rarely understood what benefits they would lose and for how long.
- State and local practices vary considerably and can influence the number of families who are sanctioned.
The enforcement of sanctions may be influenced by case managers' ability to identify barriers to employment, time to monitor and track participation in program activities, approach to case management, and comfort initiating a sanction. Case managers have primary responsibility for encouraging and monitoring program participation and considerable discretion in initiating a sanction. Within local offices, some case managers initiate sanctions more frequently than others. Some states take steps to promote fairness in the implementation of sanctions. For example, welfare case managers in South Carolina receive training on sanctions and their implementation (Koralek 2000).
Variation in local practices may contribute to differences in sanction rates even when the policies are the same. In a study of Virginia's VIEW program, researchers looked at the implementation of TANF sanctions in five different communities (Pavetti et al. 1998). They found that differences in the philosophy and approaches to sanctions contributed to differences in the rates of sanctions between the study sites (sanction rates range from 11 and 35 percent). During the first 18 months of welfare reform implementation sanction rates varied across Maryland's counties from a low of 2.6 percent of all case closures to a high of 21.2 percent (Born et al. 1999). Koralek (2000) documented variation in the rates of sanctions and in the use of conciliation reviews in South Carolina. Sanction rates between the five communities studied range from 17 to 25 percent. The differences in the use of conciliation reviews were even greater, between 18 and 36 percent. During in-depth site visits Koralek (2000) found substantial differences in the approach and use of sanctions between these communities. Overall, local offices influence the amount of training, the "message" sent to clients about sanctions, and the process for initiating and implementing a sanction.
Implementation of strategies to encourage program compliance may also influence sanction rates in various communities. The Cuyahoga County Safety Net program in Ohio re-engages sanctioned families in work activities through phone calls and home visits by participating community agencies (Goldberg and Schott 2000). In the District of Columbia, contracted service providers conduct outreach home visits to determine the client's service needs and re-engage them in employment activities. In Minnesota, two intervention programs use legal advocacy aimed at reversing imposed sanctions. The Legal Aid Society of Minneapolis helps participants cure their sanctions by either proving their eligibility for exemptions or documenting their compliance. Some of these programs report considerable success in reversing inappropriate sanctions or helping recipients to cure them. For example, the Saint Cloud Area Legal Services in Stearns and Benton counties successfully resolved 88 percent of the cases referred to them between May 1998 and November 2000 (Collins and Obrecht-Como 2001).
Synthesis of Findings and next Steps
Research on the role of sanctions in welfare reform is in its infancy. The studies that examine sanctions do so in many different ways, making it difficult to draw strong conclusions about various aspects of sanction policy. Below, we assess the current state of knowledge in each of the various aspects of TANF sanctions and suggest ways in which it could be expanded.
The Incidence and Duration of Sanctions
Owing to methodological differences, the research on the incidence of sanctions can be extremely confusing to interpret. GAO (2000a), which used currently sanctioned cases to assess the incidence of sanctions, almost certainly presented an underestimate in all but the 14 states that impose partial sanctions. Studies examining closed TANF cases can--using closure codes--provide an accurate representation of the number of families whose cases have been closed due to sanctions, and who may be at greater risk for hardships associated with lower income. But they fail to capture the extent to which partial sanctions are used, either as a final or first stage of sanctioning, and thus also underestimate the extent to which sanctions are used, although to a lesser degree than the GAO 2000a and similar studies.
Studies that look at sanctioning over a period of time for a cohort of recipients or new applicants provide the most accurate estimate of the extent to which sanctions are imposed to encourage compliance. While not completely comparable, the two studies that used this methodology found similar rates of sanctioning for work-related requirements (45 and 52 percent) even though one was conducted in a state with a gradual full-family sanction and the other in a state with a partial sanction. If families were more likely to respond to harsher penalties, and everything else were the same, one would have expected to see a lower incidence of sanctions in the state with a gradual full-family sanction.
Considerably less information is available on the duration of sanctions. Studies that have looked at duration found that sanctions last for a short period of time for at least a modest fraction of recipients. For example, Holcomb and Ratcliffe (2000) found that 28 percent reversed their sanction within a month of the sanction. Two-thirds cured their sanction within three months. Only twenty percent of sanctioned clients remained in sanction status for six or more months. This study and others that look at the duration of sanctions do not examine who eventually comes into compliance and who does not. Also missing is an explicit discussion of the number or fraction of families who never respond to sanctions.
To further our knowledge of the incidence and duration of sanctions, it would be useful to conduct cohort studies in multiple states using the same methodology, including the same sampling frame, selection period, and follow-up period. Ideally, the sample would include cases from the full TANF caseload from a certain period who would be tracked over time. The sample would then include families who remained on TANF without getting sanctioned and those who left for reasons other than a sanction. These groups could then serve as comparison groups for the sanctioned groups. Gathering data from a large enough number of states implementing various sanction policies will enable better understanding of the role of sanctions generally play, and whether there are notable differences in the incidence of sanctions by the type of sanctioned imposed.
The Characteristics and Circumstances of Sanctioned Families
Studies conducted on the characteristics and circumstances of sanctioned families presented relatively consistent findings (see Table 7). These studies found that sanctioned recipients are more likely than non-sanctioned recipients to have personal characteristics, human capital deficits, transportation barriers or personal and family challenges that make them harder to employ. They are more likely to return to TANF and less likely to be employed. These families are also more likely to experience various hardships and, on average, have lower household income. While these findings are generally consistent, the measures used are not, making it impossible to compare the magnitude of the differences across studies. The area in which these findings are most incomplete is in the presence of personal and family challenges such as substance abuse, mental health and domestic violence. Few studies captured this information; those that did didn't always reach the same conclusions. The information on the greater presence of material hardships is also difficult to interpret because it may be unrelated to the imposition of the sanction (i.e., sanctioned families may have experienced greater hardships prior to the sanction).
Few data necessary to examine families' characteristics and circumstances are readily available from routine administrative data files. Thus, to obtain more detailed information, survey data are needed. Surveying a cohort of recipients at multiple times would provide the most comprehensive information. Attaching a longitudinal survey component to the multi-state study recommended earlier would enable a simultaneous analysis of the characteristics and circumstances of sanctioned and non-sanctioned recipients, and of incidence and duration. Key topics to include in the survey would be the presence of various barriers to employment and the experience of material hardship. Standardized measures would make it possible to compare these results with those of other studies examining the TANF population as a whole. One possible way to better examine whether sanctioned families are more likely than non-sanctioned families to experience hardships would be to examine the frequency of these hardships using a multivariate analysis that controls for individual differences. This would address, at least partially, the issue that sanctioned and non-sanctioned families may be different even before the sanction is imposed.
In the absence of a new multi-site study, cross-state analysis of the recent ASPE-funded state surveys of current TANF recipients could substantially increase our understanding of the employment barriers among sanctioned and non-sanctioned families, and whether there are similar patterns observed across states. Six states fielded the same survey (with a few state-specific questions added), producing comparable data on barriers to employment among states that use a variety of sanction policies. The states are linking administrative data with survey responses, making it possible to identify families who were sanctioned at the time the survey sample was selected. With additional administrative data, it would be possible to identify families sanctioned prior to or after the selection month. For example, if you followed these families for a year or two after the time they were selected into the sample you could track them over time using administrative data to see who gets sanctioned.
|Study||More likely to be Hard-to-Employ||More likely to return to TANF||Less likely to be employed||Have less household income||More likely to experience material hardships|
|*Born et al. (1999)||x||x||x||x||n/a|
|Cherlin et al. (2001)||x||n/a||n/a||x||x|
|*Edelhoch et al. (2000)||x||n/a||x||x||x|
|Fein and Lee (1999)||x||x||x||x||n/a|
|Kalil et al. (2002)||x||n/a||n/a||n/a||x|
|*Mancuso and Linder (2001)||x||n/a||n/a||x||N/a|
|*Westra and Routely (2000)||x||x||x||x||x|
|* Indicates studies that compare sanctioned welfare recipients to other welfare leavers. n/a - Variables not included in the analysis.|
Then, you could use the combined administrative and survey data to conduct a detailed analysis of who is at risk of a sanction. One possible limitation is that the survey samples were selected from the entire TANF caseload, and may not yield sufficient cases for analysis of sanctioned and non-sanctioned cases. The capture of sufficient cases depends on several factors, including the size of the original survey sample, the state sanction policy, the rate at which sanctions are imposed and the length of follow-up to identify cases that are eventually sanctioned.
The Impact of Sanctions
Studies that attempt to examine the impact of sanctions focus primarily on whether full-family sanctions have a greater impact than partial sanctions. The few studies that examined this difference concluded that full-family sanctions increase the likelihood that a recipient will leave welfare for work and result in greater caseload declines. The caseload decline result is predictable: if, in two states with identical TANF caseloads, one imposed partial and the other imposed full-family sanctions in equal numbers, the caseload in the latter would decrease by a greater amount simply because of the mechanics of the sanction. Judging whether one type of sanction is more effective than the other requires looking beyond caseload declines to measure self-sufficiency over time through employment, earnings, other sources of income and receipt of other public benefits such as food stamps and Medicaid. We would also want to look at broader measures of family and individual functioning to examine any impact the practice might have on child or family well-being. The one study that examined the impact of sanctions on employment exits did find that work exits were more common when families were subject to more stringent sanctions, however, this study was conducted at the start of welfare reform and the results may be different now that states have fully implemented their TANF programs. Additional research in this area could help to better assess how much full-family sanctions contribute to greater employment rates among welfare recipients and how these results compare to other policies such as work incentives.
Even with more and better data, studies that exploit the variation in state sanction policies will not be able to definitely prove whether full family sanctions produce better outcomes than partial sanctions. Because states that have implemented full family sanctions in conjunction with many other policy and programmatic changes, some of which are difficult to measure, there is always some worry that a state's sanction policy may be capturing many other elements of its welfare reform policies. A well-designed random assignment demonstration project could help to isolate the impact of full family sanctions. Under such a demonstration, some families would continue to be subject to a partial sanction, others to either a gradual or immediate full-family sanction. Recipients would be randomly assigned to one of the two groups. With the exception of a different sanction policy, the two groups would receive the same treatment. They would have access to the same services and would be subject to the same earned income disregard, time limit and other policies. Researchers would follow participants for an extended period and collect and analyze data on welfare receipt, employment and earnings, other sources of income, and receipt of other benefits such as Medicaid and food stamps. In the absence of such a demonstration project, our understanding of the role of sanctions in welfare reform will always be incomplete.
The Implementation of Sanctions
There is scant literature on the implementation of sanctions. Although there is some evidence to suggest that sanction rates vary from one local office to the next, there has been very little research to assess what might contribute to these differences. Much of what we know about the implementation of sanctions comes from studies done by advocacy groups who are aiming to ensure access to public benefits.
The ideal implementation study would be a multi-state study that examines variation in local communities working under the same sanction policy framework and across states with different types of sanction policies. The information collected and analyzed for the study would include qualitative information from individual and group interviews with line staff and local and state program administrators, observation of key program activities such as orientation and sanction reviews, case record reviews, and reviews of program documents such as sanction notices. This information would be combined with administrative records data on the number of sanctions imposed and the number cured. Analysis of the data would be structured to identify the key dimensions of the implementation of sanctions, how they vary across sites, and how key implementation issues relate to the number of sanctions imposed and cured in a particular locality.
Bloom, Dan and Don Winstead (2002). "Sanctions and Welfare Reform: Policy Brief No.12." Washington, DC: The Brookings Institute.
Born, Catherine, Pamela Caudill, and Melinda Cordero (November 1999). "Life After Welfare: A Look at Sanctioned Families." Baltimore, MD: University of Maryland, School of Social Work.
Brown, June Gibbs (1999). "Temporary Assistance for Needy Families: Educating Clients About Sanctions." Washington, DC: Department of Health and Human Services, Office of Inspector General.
California Department of Social Services (2001). "Good Cause Establishment, Compliance and Curing of Sanctions Under the CalWORKs Program." Report to the Legislature. California Department of Social Services.
Cherlin, Andrew, Linda Burton, Judith Francis, Jane Henrici, Laura Lein, James Quane, and Karen Bogen (February 2001). "Sanctions and Case Closings for Noncompliance: Who Is Affected and Why." Policy Brief 01-1. Baltimore, MD: Johns Hopkins University.
Colville, Laura, Gerry Moore, Laura Smith, and Steve Smucker (1997). "A Study of AFDC Case Closures Due to JOBS Sanctions: April 1996 AFDC Case Closures." Michigan Family Independence Agency, Administration for Legislation, Budget & Analysis.
Cook, JT, Frank, DA, Berkowitz C, Black, MM, Casey, PH, Cutts, DB, Meyers, AF, Zaldiver, N, Skalicky, A, Levenson, S. Heeren, T. (2002). "Welfare Reform and the Health of Young Children." Archives of Pediatrics and Adolescent Medicine, vol. 156, pp. 678-684.
Courtney, Mark, Irv Paliavin, and P. Power (2001). "Involvement of TANF Applicants with Child Protective Services." Paper presented at the Association for Public Policy Analysis and Management Research Meeting. Seattle, WA, November 2, 2000.
Derr, Michelle K. and Vicky Cooley (2002). "Does Welfare Grant Sanctioning Increase the Prevalence of Child Welfare Involvement with TANF Families in Utah?" (submitted for publication).
Edelhoch, Marilyn, Qiduan Liu, and Linda Martin (1999). "The Post-Welfare Progress of Sanctioned Clients: A Study Using Administrative and Survey Data to Answer Three of Four Important Questions." South Carolina Department of Social Services, Office of Planning and Research.
Fein, David J. and Wang S. Lee (May 1999). "The ABC Evaluation: Carrying and Using the Stick: Financial Sanctions in Delaware's A Better Chance Program." Prepared for Delaware Health and Social Services. Cambridge, MA: Abt Associates Inc.
Fraker, Thomas M., Lucia A. Nixon, Jan L. Losby, Carol S. Prindle, John F. Else (May 1997). "Iowa's Limited Benefit Plan." Washington, DC: Mathematica Policy Research, Inc.
Gordon, Anne and Roberto Agodini (1999). "Early Impacts of the Virginia Independence Program: Final Report." Princeton, NJ: Mathematica Policy Research.
Hamilton, Gayle and Susan Scrivener (1999). "Promoting Participation: How to Increase Involvement in Welfare-to-Work Activities." New York, NY: Manpower Demonstration Research Corporation.
Hasenfeld, Yeheskel 'Zeke', T.J. Ghose, and Kandyce Hillesland-Larson (2002). "At Risk of Being Sanctioned: Comparing Sanctioned and Non-Sanctioned Welfare Recipients in California's CalWORKS." Paper presented at the Association for Public Policy Analysis and Management's 24th Annual Research Conference, November 7-9, 2002, Dallas, TX. Los Angeles, CA: Department of Social Welfare, UCLA School of Public Policy and Social Research.
Hofferth, Sandra, Stephen Stanhope, and Kathleen Mullan Harris. (2000). "Exiting Welfare in the 1990's: Did Public Policy Influence Recipients' Behavior?" Ann Arbor, MI: Population Studies Center at the Institute For Social Research, University of Michigan.
Holcomb, Pamela and Caroline Ratcliffe (Spring/Summer 2000). "When Welfare Recipients Fail to Comply with Work Requirements: Indiana's Experience with Partial Benefit Sanctions." Journal of Applied Social Sciences, volume 24, no. 1.
Kalil, Ariel, Kristin S. Seefeldt, and Hui-chen Wang (December 2002). "Sanctions and Material Hardship Under TANF." Social Service Review, vol. 76, no. 4.
Koralek, Robin (2000). "South Carolina Family Independence Program Process Evaluation." Prepared for South Carolina Department of Social Services. Washington, DC: The Urban Institute.
Mancuso, David, and Vanessa L. Lindler. (2001). "Examining the Circumstances of Welfare Leavers and Sanctioned Families in Sonoma County [CA], Final Report." Burlingame, CA: The SPHERE Institute.
Mead, Lawrence M. (1997). "Optimizing JOBS: Evaluation versus Administration." Public Administration Review, vol. 57, no. 2, pp. 113-123.
Mead, Lawrence M. (2000a). "Governmental Quality and Welfare Reform." Paper presented at the 2000 Annual Meeting of the American Political Science Association. Washington, DC, September 1, 2000.
Meckstroth, Alicia, Michelle Derr, Michael Ponza, Vinita Jethwani, and Jennifer Faerber. (2002). "Preparing Nebraska's Welfare Clients for Work and Addressing their Obstacles." Washington, DC: Mathematica Policy Research, Inc.
Needell, Barbara, Stephanie Cuccaro-Alamin, Alan Brookhart, and Seon Lee (1999). "Transitions from AFDC to Child Welfare in California." Children and Youth Services Review, vol. 21 nos. 9/10.
Nixon, Lucia A., Jacqueline F. Kauff, and Jan L. Losby (August 1999). "Second Assignments to Iowa's Limited Benefit Plan." Washington, DC: Mathematica Policy Research, Inc.
Ovwigho, Pamela Caudill, Catherine Born, Dorothy Ruck, Shafali Srivastava, and Carolyn Owens (October 2002). "Life After Welfare: Seventh Report." Baltimore, MD: University of Maryland, School of Social Work.
Pavetti, LaDonna (1995). "Who Is Affected by Time Limits?," Chapter 7, pp.31-34. In Welfare Reform: An Analysis of the Issues, ed. Isabel V. Sawhill. Washington, DC: The Urban Institute.
Pavetti, LaDonna (2002). "Helping the Hard-to-Employ," Chapter 15, pp. 135-142. In Welfare Reform and Beyond: The Future of the Safety Net, eds. Isabel V. Sawhill, R. Kent Weaver, Ron Haskins, and Andrea Kane. Washington, DC: Brookings Institution Press.
Pavetti, LaDonna A., Kathleen A. Maloy, Peter Shin, Julie Darnell, and Lea Scarpulla-Nolan (August 1998). "A Description and Assessment of State Approaches to Diversion Programs and Activities Under Welfare Reform." Report submitted to the U.S. Department of Health and Human Services, Administration for Children and Families and the Office of the Assistant Secretary for Planning and Evaluation.
Pavetti, LaDonna and Dan Bloom (2001). "Sanctions and Time Limits: State Policies, Their Implementation, and Outcomes for Families," Chapter 9, pp. 245-269. In The New World of Welfare, eds. Rebecca Blank and Ron Haskins. Washington, DC: Brookings Institution Press.
Pavetti, LaDonna, Nancy Wemmerus, and Amy Johnson. (1999). "Implementation of Welfare Reform in Virginia: A Work in Progress." Washington, DC: Mathematica Policy Research, Inc.
Polit, Denise F., Andrew S. London, and John M. Martinez (May 2001). "The Health of Poor Urban Women: Findings from the Project on Devolution and Urban Change." New York: Manpower Demonstration Research Corporation.
Rector, Robert E. and Sarah E. Youssef (1999). "The Determinants of Welfare Caseload Decline." Washington, DC: The Heritage Foundation.
Shook, Kristen (1999). "Does the Loss of Welfare Income Increase the Risk of Involvement with the Child Welfare System?" Children and Youth Services Review, vol. 21, nos. 9-10, pp. 781-814.
State Policy Documentation Project, http://www.spdp.org, a joint project of the Center for Law and Social Policy and the Center on Budget and Policy Priorities.
TANF Typologies Database (2000). Washington, DC: The Urban Institute.
Thompson, Terri, Pamela A. Holcomb, Pamela J. Loprest, and Kathleen Brennan (August 2000). "State Welfare-to-Work Policies for People with Disabilities." Washington, DC: The Urban Institute.
United State General Accounting Office (May 1997). "Welfare Reform: States' Early Experiences With Benefit Termination." Washington, DC: United States General Accounting Office.
United States General Accounting Office (June 2000b). "Letter in Response to GAO Report on State Sanction Policy: Additional Information about the Scope and Limits of Sanction Data Provided in Recent GAO Report on Temporary Assistance for Needy Families (TANF)." Washington, DC: United States General Accounting Office.
United States General Accounting Office (March 2000a). "Welfare Reform: State Sanction Policies and Number of Families Affected." Washington, DC: United States General Accounting Office.
Welfare Rules Database (1999). Washington, DC: The Urban Institute.
Welfare Rules Database (2000). Washington, DC: The Urban Institute.
Westra, Karen and John Routely (January 2000). "Arizona Cash Assistance Exit Study, First Quarter 1998 Cohort." Arizona Department of Economic Security.
- Appendix A: The Structure and Stringency of Work-Oriented Sanctions
- Appendix B: Approach to Repeat Acts of Non-Complianc
- Appendix C-1: The Cost of Most Stringent Work-Oriented Sanctions for a Family of Three
- Appendix C-2. The Cost of Most Stringent Work-Related Sanctions for a Family of Three (Number of States)
- Appendix D: Pre-Approval Requirements and Approach to Exemptions
|State||Type of Sanction1||Time to Full Family||Minimum Duration (months)||Cure Requirements|
|Alabama||Gradual Full Family||3 months||3||Participation in activities for 2 weeks|
|Alaska||Partial||None||Immediate compliance cures sanction|
|Arizona||Gradual Full Family||2 months||1||Participation for 3 days|
|Arkansas||Partial||None||Participation for up to 2 weeks|
|California||Partial||None||Varies by county|
|Colorado||Gradual Full Family||County discretion||1||County discretion|
|Connecticut||Gradual Full Family||6 months||3||Not available|
|Delaware||Gradual Full Family||4 months||2||Participation for 2 weeks|
|District of Columbia||Partial||None||Participation for 5 days|
|Florida||Immediate Full Family||None||Not available|
|Georgia||Gradual Full Family||3 months||None||Participant must initiate compliance during 3 month sanction and participate for 2 weeks|
|Hawaii||Immediate Full Family||None||Participation for 10 days|
|Idaho||Immediate Full Family||1||Participation for up to 2 weeks|
|Illinois||Gradual Full Family||3 months||None||Not available|
|Indiana||Partial||2||Willingness to comply|
|Iowa||Immediate Full Family||None||Willingness to comply|
|Kansas||Immediate Full Family||None||Willingness to comply|
|Kentucky 2||Immediate Full Family||None||Participation for 2 weeks|
|Louisiana||Gradual Full Family||3 months||3||Participation for 2 weeks|
|Maryland||Immediate Full Family||None||Not available|
|Massachusetts||Gradual Full Family||30 days||None||Participation for 2 weeks|
|Michigan 3||Immediate Full Family||1||Willingness to comply|
|Mississippi||Immediate Full Family||2||Participation required depends on activity missed|
|Missouri||Partial||None||Participation for 2 weeks|
|Montana||Partial||1||Willingness to comply|
|Nebraska||Immediate Full Family||1||Participation for 1 week|
|Nevada||Gradual Full Family||2 months||1||District office discretion|
|New Hampshire||Partial||1||Participation for 2 weeks|
|New Jersey||Gradual Full Family||3 months||1||Participation for up to 2 weeks|
|New Mexico||Gradual Full Family||6 months||None||Participation for 30 days|
|New York||Partial||None||Willingness to comply|
|North Carolina||Gradual Full Family||3 months||3||Not available|
|North Dakota||Gradual Full Family||6 months||1||Participation for 10 days|
|Ohio||Immediate Full Family||1||Willingness to comply at county's discretion|
|Oklahoma||Immediate Full Family||None||Participation for 2 weeks|
|Oregon||Gradual Full Family||4 months||None||Willingness to comply|
|Pennsylvania 4||Immediate Full Family||1||Not available|
|Rhode Island||Partial||None||Participation for 2 weeks|
|South Carolina||Immediate Full Family||None||Participation for 30 days|
|South Dakota||Gradual Full Family||1 month||None||Must begin participation prior to sanction period ending|
|Tennessee||Immediate Full Family||None||Participation for 2 weeks|
|Texas||Partial||1||Willingness to participate|
|Utah||Gradual Full Family||2 months||None||District office discretion|
|Vermont||Partial||None||Participation for 2 weeks|
|Virginia||Immediate Full Family||1||Demonstrate compliance after sanction period ends; the length of time depends on nature on non-compliance|
|Washington||Partial||None||Participation for 2 weeks|
|West Virginia||Gradual Full Family||6 months||3||Participation for an unspecified period of time|
|Wisconsin||Pay for Performance||None||Any participation|
|Wyoming||Immediate Full Family||None||Participation for 1 month|
|Source: Welfare Rules Database, Urban Institute 2000; State Policy Documentation Project.|
1 Some states impose different sanction policies for different groups of recipients. If a state has more than one sanction policy, we list the most stringent policy and identify the alternative sanction policy and to whom it applies in a footnote.
2 In Kentucky, an immediate full family sanction is imposed on families who do not complete an assessment. Families who complete an assessment and then fail to comply are subject to a partial, progressive sanction with benefits eventually being assigned to a payee.
3 In Michigan, an immediate full-family sanction is imposed if noncompliance occurs within the first two months of receiving assistance. If noncompliance occurs after the first two months, families are subject to a gradual full family sanction, with the full family sanction taking place after four months of noncompliance.
4 In Pennsylvania, during the first 24 months of benefit receipt, families are subject to a partial sanction. After the first 24 months, families are subject to an immediate full family sanction.
|State||Change in sanction type||Longer Minimum Duration||Lifetime Ban on Receipt of Assistance||Stricter Cure Requirements||Reapply for Benefits|
|District of Columbia||x|
|Source: Welfare Rules Database, Urban Institute 2000; State Policy Documentation Project.|
|State||Monthly Grant for Family of 3||Financial Penalty (Amount)||Financial Penalty|
(Percent of Cash Grant)
|Effect on Food Stamp Benefit||Adult Loses Medicaid Eligibility||Effect on Child Care|
|Arizona||347||347||100||Partial reduction initially||No||No effect|
|Connecticut||543||543||100||Partial reduction||No||No effect|
|Delaware||338||338||100||Reduced 100%||No||No effect|
|District of Columbia||379||103||27||Partial reduction||No||Ineligible|
|Georgia||280||280||100||Reduced 100%||No||Eligibility redetermined|
|Hawaii||570||570||100||Partial reduction||No||No effect|
|Illinois||377||377||100||Partial reduction||No||No effect|
|Iowa||426||426||100||Partial reduction initially||No||Eligibility redetermined|
|Maryland||399||399||100||Partial reduction||No||No effect|
|Minnesota||532||181||34||Partial reduction||No||Eligibility redetermined|
|Mississippi||170||170||100||Partial reduction initially||Yes|
|Montana||469||119||25||Partial reduction||No||No effect|
|Nebraska||535||535||100||Partial reduction initially||Yes||No effect|
|Nevada||348||348||100||Partial reduction||Yes||No effect|
|New Hampshire||550||363||66||Partial reduction||No|
|New Jersey||424||424||100||Reduced 100%||No|
|New Mexico||439||439||100||Partial reduction||Yes|
|New York||577||190||33||Partial reduction||No|
|North Carolina||272||272||100||Partial reduction||No|
|North Dakota||457||457||100||Reduced 100%||No|
|Oklahoma||292||292||100||Reduced 100%||No||Eligibility redetermined|
|Pennsylvania||403||403||100||Partial reduction||No||No effect|
|Rhode Island||554||142||26||Partial reduction||No||Ineligible|
|South Carolina||201||201||100||Partial reduction||Yes||Ineligible|
|South Dakota||430||430||100||Reduced 100%||No|
|Utah||451||451||100||Reduced 100%||No||No effect|
|Virginia||291||291||100||Reduced 100%||No||No effect|
|West Virginia||303||303||100||Partial reduction||No||No effect|
|Wisconsin||628||628||100||Partial reduction||No||No effect|
|Source: Welfare Rules Database, Urban Institute 2000; Welfare Rules Database, Urban Institute 1999; Typologies Project, Urban Institute 2000; GAO 2000; State Policy Documentation Project.|
|Amount of Financial Penalty||Total States||States with Full-Family Food Stamp Sanction||States that Terminate Adult Medicaid|
|Source: Welfare Rules Database, Urban Institute 2000; Welfare Rules Database, Urban Institute 1999; Typologies Project, Urban Institute 2000; GAO 2000; State Policy Documentation Project.|
|Pre-Approval work-related requirements||Approach to Exemptions (Universal Participation, Same as JOBS, Broader Participation)|
|State||Signed Employability Plan||Signed Personal Responsibility Contract||Orientation||Job Search|
|Alaska||Same as JOBS|
|California||x||Same as JOBS|
|Connecticut||Same as JOBS|
|Delaware||Same as JOBS|
|District of Columbia||Not Categorized|
|Kansas||x||Same as JOBS|
|Louisiana||Same as JOBS|
|Maryland||x||x||x||Same as JOBS|
|Massachusetts||Same as JOBS|
|Minnesota||Same as JOBS|
|Mississippi||x||Same as JOBS|
|Nebraska||Same as JOBS|
|New Hampshire||Broader Participation|
|New Jersey||x||Broader Participation|
|New Mexico||Same as JOBS|
|New York||x||Broader Participation|
|North Carolina||x||x||Not Categorized|
|North Dakota||x||Universal Participation|
|Pennsylvania||Same as JOBS|
|Rhode Island||Broader Participation|
|South Carolina||x||Broader Participation|
|South Dakota||Universal Participation|
|Tennessee||Same as JOBS|
|Texas||x||x||Same as JOBS|
|Vermont||Same as JOBS|
|Virginia||x||Same as JOBS|
|West Virginia||x||x||Broader Participation|
|Source: State Policy Documentation Project; Pavetti et al 1998; Thompson et al. 2000.|
1. This study is a continuation of the Maryland leavers study. The data used in Ovwigho et al. (2002) is comparable to the data used in the Born et al. (1999) study, but covers a different time period.
2. Approximately one-fifth of both sanctioned and non-sanctioned TANF recipients in Utah had an open child welfare case within three years after cash assistance case closure.
3. Employment information was gathered on TANF recipients' most recent jobs.
4. Measured 3 months after TANF case closure.
5. Hardship activities included: (1) pawning or selling personal possessions, (2) taking food or items from stores without paying for them, (3) searching in trash cans or begging, (4) engaging in any illegal activity, and (5) selling or trading food stamps.
6. Researchers used the USDA definition of food security, which is defined as, "the availabilityand access to nutritionally adequate and safe foods in socially acceptable ways."
"full-report.pdf" (pdf, 596.55Kb)