Building an Employment Focused Welfare System:

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

This report was prepared for the U.S. Department of Health and Human Sevices, Office of the Assistant Secretary for Planning and Evaluation, under contract number HHS-100-95-0021, Delivery Order 10.

The full report is available in both HTML and PDF formats. Those wishing to print this document may wish to use the PDF version of the report.

Chapter 1

This report summarizes the results of an in-depth review of the policy choices and experiences of five states that are engaged in making the transition from the more traditional welfare focus on cash assistance to the new Work First focus on employment.

The Context for Reform

The Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996 made unprecedented changes to the nation's welfare system. In place of the former Aid to Families with Dependent Children (AFDC) program and the Job Opportunities and Basic Skills Training (JOBS) program, PRWORA created Temporary Assistance for Needy Families (TANF). TANF gave states far greater flexibility to design their own cash assistance and welfare-to-work programs with the intent of transforming welfare into a work-oriented, transitional assistance program for families in need. At the same time, PRWORA does include important restrictions on how block grant funds are to be spent and several mandates concerning the work-related aspects of TANF. In particular, TANF:

• Requires states to achieve higher work or work-related participation rates for recipients;

• Defines participation rates in ways likely to increase participation in work activities and limit participation in education or training;

• Imposes a work requirement on recipients after two years of benefit receipt; and,

• Replaces the open-ended entitlement to cash assistance which had been the core element of AFDC with a five-year maximum lifetime limit on a family's receipt of the federally-funded portion of TANF benefits.

The PRWORA work-related requirements and lifetime limit, coupled with increased flexibility and freedom from federal regulations, present states with a wide range of program design issues, policy choices, and tradeoffs that touch upon all dimensions of the current welfare system. Whether states are in the early stages of implementing changes in response to the federal welfare reform legislation or simply on the lookout for new ideas and practices, the experiences of other states can inform policy makers and administrators as they confront and reassess these issues in their own states.

Employment Focused Welfare Reform: Work First and Other Strategies

In order to encourage and stimulate the cross-fertilization of ideas across states, the U.S. Department of Health and Human Services asked the Urban Institute to document key practices and strategies states have used thus far to make their welfare systems more employment focused, particularly with respect to strategies emphasizing quick entry into the labor market. Six local sites in five states were selected for intensive examination:

• Indiana: Indianapolis (pop. 817,604) and Scottsburg (pop. 22,528)
• Massachusetts: Worcester (pop. 718,858)
• Oregon: Portland (pop. 614,104)
• Virginia: Culpeper (pop. 30,528)
• Wisconsin: Racine (pop. 182,982)

These states were chosen because they exemplify a range of strategies to achieve the common goal of increasing employment among welfare recipients. Although they vary in the generosity of their cash assistance, all five have experienced well above-average caseload declines and low unemployment (Table 1). Strong economies in each of the five states provided a supportive context for implementing strategies to replace welfare with work.

Work First is a concept at the core of most state efforts to shift to systems that emphasize work rather than cash assistance. All five states included in this report have implemented Work First programs. These programs share a common philosophy regarding work: any job is a good job and program efforts are geared toward helping recipients enter the paid labor force as quickly as possible. Thus, Work First programs emphasize job search programs rather than education and training as a first step to self-sufficiency.

As we delved into the details of these states' Work First programs, we discovered two distinct models. The first we call the Work First, Work Mandate model. The second we call the Work First, Participation Mandate model. As these names imply, the two models differ in how narrowly permissible work activity is defined.

The Work First, Work Mandate model—exemplified by Massachusetts and Virginia—mandates participation in employment (unsubsidized, subsidized, community work experience) as a requirement for continuing to receive cash assistance only a few months after initial receipt of benefits. The Work First, Participation Mandate model—exemplified by Oregon, Indiana, and Wisconsin (pre W-2)—includes job search, education, and training as allowable forms of participation throughout the period of cash receipt. Although our study states were relatively evenly divided between the two models, for the country as a whole, the predominant model by far is the Work First, Work Participation approach.

We also identified several key program dimensions that cut across the two Work First models and have important roles in shaping overall program design. Table 1.2 provides an overview of the study states in the two Work First groups, along with their key cross cutting characteristics. These include: fewer exemptions from participation, more stringent sanctions, diversion efforts, time limits, more generous earned income disregards, and organizational changes to support the increased focus on work.

The Study

The information in this report is drawn primarily from site visits to state and local welfare offices between January and March of 1997. It is important to note that this is too early to capture fully the implications and impacts of the new federal reform law. The work-based reforms that are the focus of this study were also implemented at different times between 1993 and 1996. Since passage of PRWORA and our site visits, Indiana and Wisconsin have implemented new work-based reforms. Wisconsin's new W-2 program includes changes which make it more closely resemble a Work First, Work Mandate model. The other three study states have made few changes to their programs since the time of our site visits. Thus, our review of these study states' experiences covers different stages in the evolution of their new employment focused strategies.

To learn more about Work First programs in the study states, discussions were held with (1) state welfare administrators and/or senior program staff; (2) local welfare administrators and front-line eligibility and welfare-to-work staff; (3) staff from employment and training service providers; and (4) employers participating in subsidized employment programs and/or community work experience programs. With the exception of Wisconsin, the information obtained through these discussions is supplemented with administrative data on mandatory recipient participation in welfare-to-work program activities, placement in unsubsidized employment, and receipt of cash assistance over a 12-month period.

The remainder of this report of organized as follows. Chapter 2 discribes the central features of Work First programs in place in the five study states at the time of our visits. The next six chapters review the key cross cutting characteristics in turn: differing groups targeted for mandated participation (Chapter 3), differing sanction policies for noncompliance (Chapter 4), differing generosity in earned income disregards (Chapter 5), different strategies for diverting families from cash assistance in the first place (Chapter 6), and differing organizational and staffing changes to facilitate a work-oriented focus (Chapter 7). Chapter 8 concludes this summary report by reviewing actual recipient participation patterns and outcomes in light of the policies documented in the previous chapters.

Chapter 2

The philosophy underlying Work First programs is that job advancement and higher wages will come from the actual experience of working—rather than from the building of skills through education and training. Based on the assumption that most recipients are capable of finding work, employment is both the goal and the expectation under Work First. This is a major shift from the focus of the former JOBS program, which was on education and training as a precursor to efforts to find a job. More directly work-oriented activities were permitted but not emphasized under JOBS and, when used, typically came after education and training rather than before.

Work First has become the dominant approach under welfare reform. A major focus of this study was to obtain more information on how Work First actually works at the local level. We found that Work First does not represent a brand new set of welfare-to-work program services but rather that the intensity, mix, and format of services found under the former JOBS program are used differently to achieve more employment focused welfare reform. Typical practices now include: sequencing activities so that job search precedes other types of activities; making greater use of unpaid work experience and/or subsidized employment; and limiting education and training or requiring they are combined with more directly work-related activities.

WHAT RECIPIENTS ARE REQUIRED TO DO

PRWORA requires recipients to participate in "work activities" after two years of benefit receipt but allows states to impose participation requirements even sooner. All five study states require mandatory recipients to begin participating in program activities far sooner than the federally imposed two-year maximum. The five states have also defined what activities are required of recipients in a variety of different ways.

How Work First, Work Mandate States Define Participation

Both the Work First, Work Mandate study states move clients who do not find jobs within a short period into unpaid work experience or, to a far lesser extent, subsidized employment. Massachusetts requires mandatory recipients who have not found at least 20 hours a week of unsubsidized employment within 60 days after benefits are authorized to participate in 20 hours a week of unpaid work experience or subsidized employment. Before the 60-day work requirement takes effect, recipients are permitted but not required to participate in job search, education, or training. Virginia requires recipients to engage in job search immediately after benefits are authorized. If the recipient has not found at least part-time unsubsidized employment within 90 days after benefits are authorized, she must work at least 30 hours a week in unpaid work experience or subsidized employment.

How Work First, Participation Mandate States Define Participation

The Work First, Participation Mandate states provide recipients somewhat greater flexibility than Work First, Work Mandate states in terms of what activities meet participation requirements. Indiana requires a minimum of 20 hours of participation per week. If job ready recipients are not successful in an initial job search, the 20-hour participation requirement can be met by participating in unpaid work experience, continuing job search, or combining the two. Until June 1997, recipients determined not job ready could meet the 20-hour participation requirement by combining education or training with job search and/or unpaid work experience. (Since mid-1997, all recipients are considered job ready for participation requirement purposes; education and training activities cannot be used to meet the 20-hour participation target.)

Oregon requires recipients to engage in an up-front job search of 40 hours a week, part of which must occur before benefits are authorized. Those not finding a job within 30 days must engage in some combination of subsidized employment, education, training and continued job search for at least 20 hours a week. Among the five study states, Oregon's participation mandate is unique in that recipients with substance abuse or mental health problems may meet the participation requirement through engaging in treatment activities.

Wisconsin (pre W-2) required recipients to engage in 60 hours of job search over a 30-day period prior to benefit authorization. Those who did not find a job were required to participate for a minimum of 20 hours a week in activities that varied with age of youngest child. Those with a child between three and nine months were required to participate in some combination of parenting classes, job readiness activities, and basic education (job search was neither required nor discouraged). Parents of children between nine and 12 months were required to engage in at least 20 hours of job search/job readiness. Those with children over age one were required to participate a minimum of 20 hours, but generally closer to 30-35 hours a week, in program activities. For the vast majority, this meant engaging in job search/job readiness activities but recipients were also permitted to participate in short-term training, education and unpaid work experience.

JOB SEARCH AND JOB READINESS

Job search program components range on a continuum from little more than telling applicants to look for work and monitoring their progress to highly intensive and structured assistance. Although we observed variation in the design and implementation of job search components in the five study states, our discussions with staff did not produce information that would suggest that any one approach to job search was particularly more promising or effective than another. The prevailing view among staff was that local economies were so vibrant and low-wage jobs were so plentiful in the five study states that only those with serious problems (which could not be addressed through a job search component) had difficulty finding a job. In conjunction with a strong economy, the specifics of the job search seemed less important compared to the more general practices of (1) making job search the first activity, (2) emphasizing the importance of finding any job as quickly as possible, (3) imposing stricter participation and work requirements, and (4) enforcing stiffer penalties for noncompliance.

Applicant Job Search

In keeping with the general Work First premise that job search should precede any other type of employment-preparation activity, the five study states all require or encourage recipients first to engage in a job search. Some states have taken this notion to its ultimate conclusion by requiring applicants to conduct a job search before benefits will be authorized. Wisconsin (pre-W-2) and Oregon both had such a applicant job search component, although the up-front requirement was structured differently (see box for a summary description).

Types of Job Search Assistance

Common types of job search assistance included providing instruction on: (1) practical skills directly tied to finding a job such as resume preparation and (2) job readiness skills designed to better prepare recipients for the world of work, such as instruction on proper work habits and attire and life skills (e.g, parenting, time management, financial management and budgeting).

The Work First, Work Mandate states—Massachusetts and Virginia—offered little in the way of structured assistance in practical job search skills or job readiness skills. In Massachusetts, recipients were encouraged but not required to attend a short session that combined general program orientation with tips on how to conduct a job search. In addition, on their own initiative, recipients could request and receive job search assistance from Department of Employment Services staff. In some sites (e.g., Worcester), these staff were co-located in the local welfare offices and in other sites, they were co-located in one-stop employment and training centers. In Culpeper (VA), the majority of a recipients' job search was carried out on an independent basis with oversight and monitoring provided by welfare-to-work program staff. Recipients' job search activities were supplemented with a week-long, half-day workshop on job search and job readiness.

The Work First, Participation Mandate states placed more emphasis on developing recipients' job readiness skills. Portland (OR) and Indianapolis/Scottsburg (IN) provided well-developed job clubs that included instruction on job search and job readiness skills combined with a guided and monitored job search. Wisconsin (Racine) also combined job search and job readiness assistance but not in the form of a structured job club targeted to welfare recipients.

As part of Wisconsin's effort to centralize all employment and training services within a single Workforce Development Center, recipients in Racine had access to the same centralized resources and attended the same job search/job readiness workshops as all other job seekers. Local staff reported that, although there was some concern that the least motivated or confident recipients found the new system intimidating, most recipients liked "being treated like everyone else" and adapted easily to the less structured format.

All of the study states except Massachusetts (because job search was not mandatory there) monitored job search activities. Monitoring practices included various combinations of requiring verification of employer contacts, requiring submission of self-reports of employer contacts that were not monitored, and tracking recipient attendance at job search classes and time spent in other job search or job readiness activities.

UNPAID WORK EXPERIENCE AND SUBSIDIZED EMPLOYMENT

Unpaid work experience and subsidized employment are both acceptable means of meeting the new federal work participation requirements. As such, they are of special interest to states looking for ways to meet their increasingly stringent work participation targets. Unpaid work experience, also known as community work experience or community service, typically provides recipients the opportunity to work for public and nonprofit employers in exchange for welfare benefits. Subsidized employment provides recipients the opportunity to work for private employers who receive a subsidy for hiring recipients that is financed out of diverted welfare benefits.

Since these strategies played a very small role in the former JOBS program, many states had little experience with unpaid work experience programs, even less with subsidized employment programs, and none with large-scale versions of either. Our study states were chosen, in part, because some did have more experience with these types of programs than many other states, due in part to waivers granted by the Department of Health and Human Services since 1993. Our two Work First, Work Mandate states—Massachusetts and Virginia—have well- developed unpaid work experience programs but have had little success in implementing subsidized employment. One of our Work First, Work Participation states—Oregon—has a well-developed subsidized employment program but makes only minimal use of unpaid work experience.

Unpaid Work Experience

Staff generally viewed unpaid work experience, at least in principle, as a valuable way to prepare clients who are not quite ready for the world of work. Staff also told us that informing recipients that they would be placed in such a position proved to be a very effective incentive for recipients to seek and find unsubsidized employment. Other factors have also contributed to remove the need for any large-scale unpaid work experience program in the study states. These include the vigorous economy, which has led many recipients to find jobs or simply close their cases prior to imposition of the work requirement or soon after, enforcement of strict sanction policies (discussed further in Chapter 4), and generous earned income disregards (discussed further in Chapter 5) that allow recipients to combine welfare with unsubsidized employment.

So far, the study states have not been unduly burdened by the administrative and operational challenges commonly associated with operating such programs (developing slots, keeping slots filled, monitoring participation, and serving as a liaison between recipients and site providers). Both Massachusetts and Wisconsin found it helpful to dedicate certain staff to such tasks. In Massachusetts, recipients were also encouraged to develop their own "community service" positions, thereby generating additional slots and more flexibility for recipients to develop assignments that fit their interests and schedules.

The handful of unpaid work experience providers across the five states who shared their experiences with us reported that they did not find involvement in such a program overly burdensome and were generally pleased with their participants. Several noted that participants rarely remained in a slot more than a month or two, making involvement in unpaid work experience programs particularly well-suited to employers with jobs that demanded little in-house training and were easily transferable from one person to another.

These experiences suggest that even states that opt to define work activities primarily as unpaid work experience may not encounter as much need for recipients to be placed in this work activity or as heavy an administrative burden as might be anticipated—at least when the economy is strong and unsubsidized entry-level jobs are plentiful.

Subsidized Employment

Of the study states, both Work First, Work Mandate states—Massachusetts and Virginia—originally envisioned implementing large subsidized employment programs but encountered significant barriers to implementation that stymied their efforts. In contrast, Oregon's experience with subsidized employment has been much more positive.

The implementation barriers encountered in Virginia and Massachusetts are typical of efforts to implement subsidized employment programs on any scale. On the employer side, staff reported that employers seemed generally unaware of the possibility of subsidies for hiring welfare recipients; even when they were aware of the program, they were uninterested or reluctant to participate. The financial incentives did little to persuade most employers to participate because of the potential administrative burden it entailed and the stigma attached to hiring welfare recipients. On the program side, insufficient staff resources were invested in combating employer reluctance and reducing administrative burdens associated with their participation. On the recipient side, those who were good candidates for subsidized employment generally could find (and preferred) unsubsidized jobs.

Oregon's "JOBS Plus" subsidized employment program was notably different from typical state efforts because it was designed by the private sector and, thus, supported by employers and tailored to their needs. Immediately upon implementation, thousands of letters to employers were sent on the letterhead of the private business whose president created JOBS Plus. Since its initial implementation, the program continues to be aggressively marketed by designated welfare-to-work staff and members of local welfare-to-work program councils.

In addition, participation in Oregon's subsidized employment program has been made easy for the employer. The only requirements are to sign an agreement to abide by a minimal number of program rules and provide a mentor. There is no obligation to hire the participant at the end of that time but firms are typically prepared to do so. Participants also receive an array of additional supports that includes any special clothing needs; child care and transportation assistance, continued Medicaid eligibility; and after 30 days of participation, a dollar-per-hour-worked contribution by the employer to a fund the recipient can later draw on to pay for education and training.

ESTABLISHING STRONGER LINKAGES WITH EMPLOYERS

Changing the welfare system to be more employment focused has accentuated the need to expand linkages with employers in order to support efforts to place recipients in unsubsidized employment and work activities. In the past, little program attention was paid to developing stronger ties with employers, despite their importance in the welfare-to-work transition. The focus instead was on helping clients attain longer term education and training, helping clients learn good job search skills, and identifying promising job leads. Employers, for their part, did not typically see welfare recipients as promising labor prospects.

In our study states, we observed three types of efforts to bridge the gap between employers and welfare-to-work programs: involving employers in program design and implementation, working to develop job slots (job development) and helping recipients keep them (job retention), and having employers speak at job search sessions.

In general, bringing employers to the table was viewed by many welfare staff as key to making employers feel vested in the process. Neither job development nor job retention were widely used in the study states. The strong economies reduced the perceived need to develop additional job opportunities but many staff saw a strong need to implement job retention strategies that would reduce the high propensity of recipients to lose jobs. Finally, staff reported that the "real life" dimension of hearing from employers themselves at job search sessions about what they valued in an employee seemed to give the material added salience for recipients.

EDUCATION AND TRAINING IN A WORK FIRST ENVIRONMENT

With its emphasis on quick entry into the workforce, there is no predefined role for education and training in a Work First environment. The challenge is to tailor these activities to fit the new environment rather than to jettison them altogether. States have approached this challenge by: (1) allowing education and training only after completion of an unsuccessful job search, (2) permitting them only when combined with more directly work-oriented activities, and (3) shifting the focus toward short-term programs that recipients can begin immediately.

Implementing these changes is challenging. They involve restructuring linkages with education and training providers. Some of these may be unprepared, resistant, or outright opposed to the whole Work First approach. Coupling education and training with more work-oriented activities can also be a deterrent to participation in these skills preparation activities. Many recipients in a Work First environment reportedly do not engage in education and training because they cannot balance the competing demands of work and education and training and family. Supporters of coupling education and training with more work-oriented activities emphasized that non-welfare working people have to make the same choices.

THE HARD TO SERVE: A CRITICAL YET LARGELY UNADDRESSED CHALLENGE

The vigor of the job market is clearly pivotal to the ease with which welfare recipients find work. But some recipients face barriers to employment that cannot be relieved simply by the availability of jobs. These barriers include little to no previous work experience and low basic skills, as well as mental health and drug abuse problems, domestic violence, child behavior problems, and legal problems. Most staff interviewed in the study states favored Work First's strong focus on immediate steps to find employment. At the same time they expressed concern about the harder to employ and the need for supports other than job search for those recipients.

Subsidized employment or unpaid work experience may help recipients whose only barriers are lack of work experience and low basic skills. But people with severe personal and/or family problems are likely to need more. They may be able to handle community service positions, but these may not lead to unsubsidized employment unless other supports are available. Officials and staff often pointed out that the need for such additional supports is increasing under Work First, precisely because program success in moving recipients into jobs has left behind a recipient population that is increasingly hard to serve. Approaching time limits and stricter participation requirements are intensifying this sense of urgency.

Of the study states, Oregon (Portland) had devoted the most emphasis and resources to dealing with the harder-to-serve population. Case managers are assigned specifically to identify and resolve barriers for those who do not find jobs. Recipients are required to attend an addictions awareness session, where they undergo drug screening. Mental health counselors are located on-site for recipient counseling and case manager consultation. Recipients with identified mental health or substance abuse problems are required to participate in treatment or counseling, which counts toward meeting the participation requirement.

Staff generally perceived that these types of investments in harder-to-serve recipients have paid off. They also noted that, although the average cost of serving welfare-to-work recipients has gone up, these costs are not as high as expected, because a substantial share of the cost of the more expensive specialized services is covered by other agencies and/or spending streams.

A key initial challenge when implementing Work First was to change the group mindset about recipient employability, according to administrators and staff. The basic premise is that most recipients can and should work. However, administrative reports a growing recognition that Work First is not the final answer for all recipients, and that some will require additional supports in order to achieve self-sufficiency.

Chapter 3

Expanding participation by narrowing exemptions is one of the most significant changes a state can make in reshaping its welfare-to-work programs to emphasize employment. The former JOBS program's primary exemption criterion was to exclude adult primary caretakers with children under age three (or age one, at state option). In addition, recipients caring for an ill or disabled household member, working at least 30 hours a week, living in a remote area, or with a medical or psychological problem that affected their ability to work were also exempt. Under PRWORA, states are free to design their own exemption policies, although the law's escalating work participation rates argue for narrowing exemption criteria, especially regarding the age of the youngest child. With the exception of Massachusetts, all states in this study moved toward narrowing their policies regarding who should be exempt from work or work-related activities. Within this general trend, the extent to which exemptions were narrowed varied.

DIFFERING PROGRAM COVERAGE

The broad range of coverage options exercised by the study states translates into mandatory participation for anywhere from 20 percent and upwards of 90 percent of the state's TANF caseload. These exemption choices bear no relationship to whether a state has adopted a Work First, Work Mandate versus Work First, Participation Mandate approach.

Of the Work First, Work Mandate states, Massachusetts extended its work mandate for able-bodied recipients to cover only those whose youngest child was over age six. Whether a recipient was exempted because of a medical or psychological problem depended on a very structured medical review. Virginia, in contrast, though it preserved most of the JOBS exemptions, extended its work mandate to cover all able-bodied recipients whose youngest child was over 18 months.

All the Work First, Participation Mandate states imposed narrower exemption criteria than Massachusetts. Indiana eliminated exemptions for recipients living in remote areas but retained the JOBS exemption for recipients with children under age three and the disability exemption. By the end of 1998, however, the exemption based on the age of the youngest child will have been lowered to 12 weeks. Oregon eliminated virtually all the JOBS exemption criteria—leaving only VISTA volunteers, women in their last pregnancy trimester, and recipients caring for a child under 12 weeks exempt from the participation requirement. Wisconsin initially lowered the able-bodied exemption to recipients whose youngest child was under age one. At the beginning of 1997, however, almost all exemptions were eliminated with the exception of those with children under 12 weeks of age.

BASIC EXEMPTION TRADE-OFFS

Program coverage is an important influence on the degree to which welfare can be transformed from a cash-based to a work-based system. The principle that all recipients should be required, as a condition of cash benefit receipt, to engage in activities leading them to work depends on universal coverage.

Exemptions run counter to this underlying rationale but also serve two major purposes of their own. First, up-front exemption criteria enable states to screen out high-cost cases and thus control their expenditures. Second, exemption criteria can be used to narrow the mandatory participant pool to a more uniformly job-ready group and expose them to the full intensity of job search and related activities. A larger and more heterogeneous mandatory pool requires increased expenditures or the spreading of the same resources more thinly over a group with more diverse needs.

These tradeoffs play out in different ways across states, reflecting differences in priorities, financial considerations, program philosophies, and responses to new federal participation rate requirements. Concerns over child care and services for the harder-to-serve populations weigh particularly heavily as these tradeoffs are made.

Exemptions and Child Care Issues

The two states with the narrowest age of youngest child exemption criteria for recipients—only those caring for children under 12 weeks—have both significantly increased their financing of child care and made it a near-guarantee for the working poor as well. The local sites we visited in Portland (OR) and Racine (WI) reported that waiting lists for low-income care assistance had been eliminated and that there was an adequate supply of child care providers in these areas.

Indiana and Massachusetts both cited child care costs as a reason for choosing relatively broad criteria for the able-bodied exemption. Indiana chose to focus its resources initially on efforts to increase actual participation among recipients with no children under age three than to spread limited resources over a broader pool that would need more child care. Having made great strides towards achieving this broad participation goal, the state has since begun to phase in its lower youngest child exemption age to under 12 weeks. Massachusetts chose to restrict its mandate to recipients with only school-age children, providing child care only in the summer months. The state has recently increased its funding for child care, but is targeting this assistance to families seeking jobs on their own.

Exemptions and the Harder-to-Serve

Of our five study states, Oregon and Wisconsin have gone the furthest in including harder-to-serve recipients in their participation mandates. For example, since 1993, Oregon expanded its definition of allowable participation activities to include substance abuse treatment and mental health counseling, in order to require participation of these recipients but also provide them appropriate supportive services to enable them to do so. As of January 1997, recipients in Wisconsin who would previously have received a disability exemption had started being referred to the Division of Vocational Rehabilitation for assessment and assistance in developing an activity plan and additional strategies were being developed in preparation for the implementation of W-2.

Complexities Added to Exemption Choices by PRWORA

The PRWORA's two-year work requirement, lifetime benefit limit, and escalating work participation requirements are all powerful incentives pushing states to narrow their exemption policies. In response to these incentives, states are lowering the youngest child's age exemption. Only seven states, including our study states of Virginia and Massachusetts, now have youngest child exemption policies above one year. If states do continue to reduce their caseloads by expanding participation to a greater share of the TANF caseload, there will be pressure to include increasing numbers of recipients who need more assistance to make the transition to work. How fast they do so, and how far they take such an expansion is likely to depend on whether states believe they have the knowledge and resources to succeed with the welfare-to-work transition of formerly exempt recipients (e.g., those with young children or physical or mental disabilities). In general, states have narrowed exemptions as part of the larger effort to move toward a more work-oriented system but most have stopped short of committing to the development of full-scale programs that require universal participation.

Chapter 4

Sanctions (fiscal penalties assessed on recipients' grants) and time limits raise the stakes for welfare recipients by imposing financial penalties on those who do not follow participation requirements (sanctions) or fail to become self-sufficient (time limits). States have long had the ability to impose sanctions which are imposed in response to specific actions. Time limits represent a new and significant departure from traditional welfare in that they set an arbitrary deadline on benefit receipt.

SANCTIONS: RAISING THE STAKES OF NONCOMPLIANCE

Many states have been moving in recent years from invoking sanctions as a warning to actually imposing them when recipients fail to comply with program requirements. Stricter enforcement has often been accompanied by stiffer penalties in response to the common perception that sanctions under JOBS were too weak, even when enforced, to be an effective deterrent to noncompliance.

Three Types of Sanctions

Three major types of sanctions, as exemplified in the study states, are currently used by states: adult-only, full-family, and pay-for-performance. There was no relationship between the Work First model a state adopted and the severity of its sanction policy.

Adult-only sanctions deduct that portion of the grant covering the adult who is not complying. As used by Indiana, they amount to about one-third of the average grant and continue the assistance to the children. The standard JOBS sanction was an adult-only sanction.

Full-family sanctions eliminate the entire grant. Virginia relied solely on full-family sanctions. Massachusetts and Oregon used them only when less severe sanctions had failed. Massachusetts imposed adult-only sanctions for the first three months of noncompliance. Oregon imposed a $50 a month sanction for the first two months, an adult-only sanction for the next two months, and a full-family sanction only after four months of noncompliance.

Pay-for-performance sanctions are an innovative variation developed in Wisconsin. Designed to replicate regular work experience, they tie the penalty directly to the hours an adult fails to participate in the required activities—just as a regular employee receives no pay for hours not worked. The maximum penalty is the most severe of the five study states (and possibly the nation), because it includes the family's Food Stamp allotment as well as its full-family cash grant.

Balancing the Dual Intents of Sanctions

Sanctions have two purposes. The first is to motivate recipients to participate. The second is to hold them accountable for their actions. If a state's primary aim is to motivate, it may choose to focus staff resources on encouragement and on addressing barriers to participation. If the main goal is accountability, states will be more likely to favor swift enforcement and leave responsibility for coming back into compliance completely up to the recipient.

All the study states sent written notification to noncompliant recipients that their benefits were going to be cut if they did not contact their case manager to establish good cause or come back into compliance. Beyond this standard practice, how the dual intents of sanctions were translated into practice varied widely.

The only study state making a major effort to encourage recipients to come back into compliance before the sanction was imposed was Oregon. Before initiating any sanction, case managers were required to contact the recipient to determine whether there was either good cause or removable barriers to compliance. Before progressing to a full-family sanction, case managers were also required to conduct an intensive one-on-one session to identify any barriers to participation, make a home visit and develop a plan that relied on other community agencies to address the children's safety.

In the other four study states, staff were not required to make any special efforts—other than sending warning letters—before the sanctions went into effect. Indiana and Massachusetts provided a grace period to enable the recipient to come into compliance. Virginia and Wisconsin simply put the sanction into effect the month following the written notification.

Minimizing the "Revolving Door"

All five states relied on either minimum sanction periods or some demonstration of good faith on the part of the noncompliant recipient to minimize situations in which sanctions were lifted only to be reimposed again. This "revolving door" associated with the implementation of sanctions was reportedly a common problem under JOBS that diluted the seriousness of the message to recipients and placed a heavy administrative burden on staff.

Three of the states imposed minimum sanction periods. Wisconsin's minimum was a straight one-month period—as long as the recipient was in full compliance in subsequent months—because the pay-for-performance design allows recalculation of the basic grant (with or without a sanction deduction) every month. Virginia and Indiana imposed progressively longer minimum sanction periods. In order to get their benefits restored, recipients must comply with program rules even though they still continue to be penalized until the minimum sanction period elapses. The intent is to reinforce the lesson that there are, indeed, consequences for noncompliance but the practical outcome is also to reduce the incentive to come back into compliance. In Virginia, on the second occasion a recipient received a full-family sanction, the benefit could not be restored for three months. In Indiana, the corresponding minimum sanction period after the second incidence of noncompliance was a year.

The good faith approach was used in Massachusetts and Oregon. In Massachusetts, a recipient who was not employed within 60 days was given a 10-day grace period to begin community service. If the recipient was still not in compliance after another 30 days, a full-family sanction was applied and the case closed. To be reinstated after that point, recipients had to comply with participation requirements for two weeks. In Oregon, the activity required to restore benefits was decided jointly by recipient and case worker and had to be completed within a 24-hour period.

Assessing the implications and impacts of the tradeoffs states make in setting sanction policies is difficult because it depends on the reasons why recipients are not participating in required activities. What proportion have other sources of income or do not think compliance is "worth the hassle?" What proportion would participate with greater encouragement and more service support? Greater understanding of the dynamics of noncompliance would help states as they try to balance the dual aims of sanctions within the constraints of administrative efficiency.

TIME LIMITS: RAISING THE STAKES OF CONTINUED DEPENDENCY

Although evidence is trickling in regarding time limits, critical questions remain that this study cannot answer about their impact on patterns of work and welfare receipt, and how families who hit the time limits fare in the short- and long-run. Here we review the characteristics of the time limit policies of the study states and how they can interact with other key welfare-to-work strategies.

Key Time Limit Characteristics

All five states imposed a 24-month limit on assistance receipt before the time limit took effect. They differed, however, in the time period over which this limit was applied: 84 months in Oregon, 60 months in Massachusetts and Virginia, 48 months in Wisconsin, and 24 months in Indiana. In all the study states except one, the total family benefit was terminated when recipients' reached their time limit. In Indiana, reaching the time limit resulted in a reduction of the grant amount equivalent to only the adult portion of the grant.

Policies for back-end extensions of, and exceptions from, the time limit were more varied. These differences are not correlated with the type of Work First model used by the study states. Indiana and Virginia both allowed extensions for up to one year—the former if the recipient had been substantially in compliance but still failed to find a job; the latter if the local unemployment rate was above 10 percent or the extension would allow the recipient to complete an employment-related course of training or education.

Oregon exempted those who are participating in required activities when they reach the time limit for as long as they continue to comply with program rules. Oregon also allowed extensions for up to three months for the care of family members who are seriously ill. Massachusetts and Wisconsin (pre W-2) allowed extensions on a case-by-case basis—the former with no specific criteria, the latter to recipients who had made diligent but unsuccessful efforts to find a job or could not find work because of low ability, emotional or family problems, or poor local labor market conditions.

Interactions with Other Employment Focused Strategies

Time limits interact with sanctions, work requirements, and earned income disregards in ways that may not have been anticipated or fully appreciated when initially designed.

With respect to sanctions, there are several points to be made. Due to full-family sanctions in Virginia and Massachusetts, most noncompliant individuals will have been completely cut off cash assistance long before reaching the time limit, leaving only recipients who met the requirements to be penalized by a time limit. In Oregon, in contrast, the time limit policy exempts individuals who comply with requirements, protecting "worthy" recipients from it and cutting drastically the proportion of the caseload likely to be affected by a time limit. In addition, full-family sanctions that do not automatically close cases have more negative implications in a time-limited environment. In Virginia, for example, unless the sanctioned family requests to be taken off the rolls, the time limit clock runs even though no benefit is being received.

Unless work requirements and time limit exemptions are synchronized, they can result in groups being dropped from the rolls due to the time limit without ever having been subject to any work requirement. In Massachusetts, for example, this happens because almost a third of the caseload is subject to the time limit but not the work requirement—putting this group at a disadvantage in a state that has specifically chosen a Work First, Work Mandate approach in order to maximize recipients' chances of gaining unsubsidized employment.

Finally, time limits have important implications for earned income disregard policies. Although earned income disregards are intended to encourage recipients to combine welfare with work, months on assistance all count towards the time limit, even if a recipient cuts her grant amount by working during those months. Most states have increased their earnings disregards, but very few exempt working recipients from the time limit. Therefore, recipients who continue to receive benefits while working in unsubsidized jobs — due to earned income policies in some states — are also using up the finite months of assistance available to them.

Chapter 5

Earned income disregards provide a means to increase recipients' family income and ability to combine work and welfare. The former AFDC program's treatment of recipient earnings was commonly criticized for not giving recipients sufficient incentive to seek or keep a job. Recipients were given a $90 work expense disregard for as long as they worked and remained on the rolls, an additional $30 a month for the first 12 months of employment, and an additional one-third of any remaining earnings for the first 4 months of employment. A major problem with this formula was that it loaded the incentive into the first 4 months. After that, any income a recipient earned above $120 a month led to a dollar-for-dollar reduction in the family cash grant—removing any financial incentive to earn.

Prior to TANF, thirty-two states obtained waivers to change their earned income disregards. The most common change was to remove the time limit and increase the fraction of income that was disregarded. Since this change increased costs, several states offset the cost increase by reducing the amount of income that could be disregarded, restricting the disregard to only part of the caseload, or concentrating it on the lowest-earning recipients. PRWORA gives states total flexibility to determine how to treat earnings when calculating eligibility and benefit levels for applicants and recipients.

EARNED INCOME DISREGARD POLICIES IN THE STUDY STATES

Among the five study states, the earned income disregard changes in the Work First, Work Mandate states—Virginia and Massachusetts—were the most generous. While the Work First program approach used by these states calls for recipients that have not found jobs to be engaged in work programs soon after benefits are authorized, these states also give recipients much more financial incentive to obtain unsubsidized employment and provide continued financial support for those that do work. The following provides an overview of key earned income disregard changes in the five study states.

Work First, Work Mandate States

The earned income disregards in the Work First, Work Mandate states were designed to reward recipients' efforts to enter the paid labor force and stay employed to a far greater degree than those in the Work First, Work Participation states.

Massachusetts raised the proportion of earned income disregarded from $30 and one-third to $30 and one-half, removed the time limit on the disregard, and extended the disregard to all recipients subject to the time limit on benefit receipt. As an effort at cost containment, these changes were accompanied by a 2.75 percent reduction in benefits for all families subject to the time limit on benefit receipt. Virginia continued the time-limited $30 and one-third earned income disregard. However, instead of retaining the dollar-for-dollar reduction in benefits for all income earned over $120 a month, the state chose to continue to provide recipients with a full grant as long as the combined earnings and cash assistance did not exceed the federal poverty line.

The difference between these two states also illustrates the new flexibility states have to design financial work incentives. Under the old AFDC rules, the amount of income which families who combined welfare and work had to spend depended entirely on the generosity of a state's welfare benefit. Now it depends on a whole set of interactions.

Massachusetts, for example, is a relatively high benefit state. Under the old AFDC rules a working recipient in a three-person case could have continued to receive some cash assistance until earnings reached $670 a month. Virginia, in contrast, is a relatively low benefit state. Under the old AFDC rules a working recipient in Virginia would have lost all assistance when earnings reached $380 a month. Under the new rules, recipients working part-time (20 hours a week) at the minimum wage receive more assistance in Massachusetts than Virginia—$387 versus $291 a month. But recipients working full-time at the minimum wage now receive more cash assistance in Virginia than in Massachusetts—$291 versus $149 a month—a reversal of the old system that is due to Virginia's policy decision to preserve the total grant amount as long as the recipient's total income is below the poverty level.

Work First, Participation Mandate States

In sharp contrast to the Work First, Work Mandate states, the only recipients who benefit from their states' earned income disregards in the Work First, Participation Mandate states are those who work part-time at very low wages; none receive as much cash support as similar working recipients in the Work First, Work Mandate study states (see Table 5.1).

Indiana basically retained the AFDC earned income disregard rules. Therefore, recipients who earn the minimum wage and work only part-time are eligible for a $50 a month grant for four months. Recipients earning higher wages or working longer hours lose all their benefits immediately.

Oregon replaced the $30 and one-third rule with an earned income disregard of 50 percent. However, the pool of recipients who might benefit from this change is limited primarily to part-time working recipients with low wages because Oregon uses a gross income limit that is about one-third lower than the gross income limit used by most states. Under these rules a recipient earning the minimum wage and working 20 hours a week receives a grant of $222 a month. If the same recipient either worked full-time at the minimum wage or 20 hours a week at $7.50 an hour, she would no longer be eligible to receive any assistance.

WHAT HAPPENED TO RECIPIENTS OVER THE COURSE
OF THE YEAR FOLLOWING SYSTEM ENTRY?

Work First programs aim not only to engage a large fraction of recipients in program activities, but also to sustain participation and keep recipients from falling through cracks in the system. The only site for which we obtained data (not shown) that enabled us to see how well this goal was achieved is Worcester (MA). In that site, month-by-month movement of recipients through the system is quite dramatic.

By the second month after initial benefit receipt, 37 percent of mandatory recipients had already left the rolls and an additional 27 percent were either working in unsubsidized employment (17 percent) or participating in a work activity (10 percent). Most of the remaining mandatory recipients were participating in job search (29 percent). Only four percent of recipients were not engaged in any activity and it is likely that the majority of them were waiting for an unpaid work experience slot to be assigned to them. By the sixth month, the proportion not off the rolls and not in some form of approved participation had dropped to only 5 percent and continued to decrease thereafter.

This month-by-month information provides further support for a point made earlier in the report. Even though Massachusetts requires participation in unpaid work experience if mandatory recipients cannot find an unsubsidized job, the need for such slots in any given month remained modest, with most recipients either having left the rolls or combining welfare with unsubsidized jobs. The fraction of mandatory recipients who participated in unpaid work experience or subsidized employment in any given month never exceeded 14 percent.

It must be emphasized that the picture presented of recipients' experiences in Worcester (MA) reflects only a small proportion of the entire caseload. Only able-bodied recipients with no children under six years old are required to participate—the group with the fewest barriers to finding and keeping a job. This broad exemption policy contrasts sharply with most other states.

HOW MANY RECIPIENTS ARE ON WELFARE ONE YEAR AFTER INITIAL ENTRY?

While the primary goal of Work First is to move recipients into the paid labor force as quickly as possible, this approach does not focus on moving recipients off welfare permanently. As reflected in the employment rates achieved by the study states, the local sites examined here have been successful in implementing the goals of a Work First program approach.

At the same time, a significant fraction of the samples' mandatory recipients were receiving assistance at the end of a year—31 percent of recipients in Worcester (MA), 42 percent in Portland (OR) and 44 percent in Indianapolis (IN). These percentages reflect recipients who received assistance for all 12 months as well as those who left welfare but later returned. Recipients in either of these two groups may have combined work with welfare during the one year period. They serve as an important reminder that "replacing welfare with work" is an ongoing process and that reaching financial self-sufficiency, even in a Work First environment, requires a wide variety of policy and program responses.

About the Study

The Urban Institute
The Urban Institute is a nonprofit policy research organization established in Washington, D.C., in 1968. Its objectives are to sharpen thinking about society's problems and efforts to solve them, improve government decisions and their implementation, and increase citizens' awareness about important public choices. Institute researchers identify and measure the extent of social problems, assess developing trends and solutions to those problems, evaluate existing social and economic programs and policy options, and offer conceptual clarification and technical assistance in the development of new strategies. In pursuit of broader research and educational goals, Institute staff present their findings to representatives of federal, state, local governments, the media and other interested groups.

Study Authors
Pamela A. Holcomb, a senior research associate at the Urban Institute was project director. Other contributing researchers were Caroline Ratcliffe, a research associate at the Urban Institute and LaDonna Pavetti and Susan Riedinger, both formerly of the Urban Institute.

Study Funders
This study was funded by the U.S. Department of Health and Human Services (DHHS), Office of the Assistant Secretary for Planning and Evaluation (ASPE).

Acknowledgements
This study could not have been completed without the cooperation and assistance of many individuals. State and local agency staff, service providers and community work experience providers were extremely generous with their time and instrumental in helping us understand their policies, programs, practices and views on a wide range of issues. We are also indebted to Steve Trost and Stacy Poulos for their valuable assistance on the administrative data analysis and to Karen Foley and Rick Scully for their contributions throughout throughout the publication process. Great Appreciation is also owed to Felicity Skidmore for ler lead role in preparing the Executive Summary report. Finally, the report benefited from the careful review and suggestions of Karin Martinson, Sharon Parrot, Demetra Nightingale, Bob Lerman, Kathy Brennan and Carolyn O'Brien.

Ordering Information
Copies of both full the report and the executive summary can be obtained from: Ms. Ethel Norris (202-690-6809), Department of Health and Human Services, Assistant Secretary for Planning and Evaluation (ASPE), Human Services Policy, Room 404E, 200 Independence Avenue, S.W., Washington, D.C. 20201. The publications also may also be accessed through the websites of the Assistant Secretary of Planning and Evaluation (http://aspe.os.hhs.gov) or the Urban Institute http://www.urban.org). for information about other related U.I. reports, call UI Public Affairs at (202) 857-8709 (paffiars@ui.urban.org).