Can Work Experience Programs Work for Welfare Recipients?

Harry J. Holzer
Harry Holzer
Harry J. Holzer Nonresident Senior Fellow - Economic Studies, Center for Economic Security and Opportunity, John LaFarge Jr. SJ Professor of Public Policy - Georgetown University' McCourt School of Public Policy

November 2, 2002

Work experience programs have played a limited role in welfare reform to date. While virtually all states and local areas have developed such programs, they are limited in both number and size. Only 40,000 Temporary Assistance to Needy Families (TANF) recipients are enrolled in work experience programs nationwide in any given month, and four states (Illinois, New Jersey, New York, and Ohio) account for about two-thirds of total enrollment. Most states have chosen to use TANF funds for other purposes, such as child care or other work supports, rather than for job creation.

However, state interest in work experience may rise over the next few years, especially if TANF is reauthorized with a higher set of work requirements on welfare recipients and states than has existed under current law. Specifically, TANF now requires 50 percent of recipients to be involved in work activities, at twenty hours per week for those with children under the age of six and thirty hours otherwise. Under the President’s proposal and the bill passed by the House of Representatives, the requirements would rise to 70 percent of recipients and forty hours per week for mothers, of which at least twenty-four hours would have to be in paid or unpaid work, while the remainder could be devoted to training or other work-related activities. Many states have met the current participation requirements through credits for caseload reductions, which were quite dramatic during the period from 1996 to 2000. But caseloads are very unlikely to keep falling dramatically over the next several years, and the credits will be of much less use in meeting participation requirements. Thus, states may feel substantial pressure to raise employment among welfare recipients, perhaps by instituting mandatory work programs.

Several additional factors might also lead to wider adoption of work experience programs in the near future. If slow economic growth continues, and results in slack labor markets for an extended period, both current and former recipients may face more difficulty finding employment than they did in the very tight labor markets of the late 1990s. As time limits for welfare become effective and are enforced in many states, cash assistance for those without work may be limited, which might in turn generate calls for public job creation. Widespread concern about the “hard-to-employ,” and limited eligibility for unemployment insurance among these former recipients, might reinforce public support for these calls.

This brief considers what we know about these programs: what forms they can take, the issues involved, the evidence on their effectiveness to date, and the implications of these findings for policymakers at the state and local levels.

The Many Forms of Work Experience Programs

Work experience programs have taken many forms over the years and been applied to many different groups and situations. These have included the large-scale Public Service Employment (PSE) program of the 1970s, the more recent Summer Youth Employment Program, and a variety of smaller efforts for disadvantaged adults (e.g., Supported Work in the 1970s) and youth (e.g., the Youth Incentive Entitlement Pilot Project [YIEPP] of the late 1970s and the more recent Youth Service Corps and Youth Build programs).

All of these examples entailed public funding of jobs, either in the public or private sectors, which paid an hourly wage to each worker and often included other supports and services. The most recent incarnations of this approach are usually called “transitional jobs.” These programs have been implemented in a variety of local areas, with prominent examples appearing in Chicago, Philadelphia, and the state of Washington. Transitional jobs programs provide a short-term (i.e., three to twelve months) paid employment experience plus supports for welfare recipients or other very disadvantaged workers, including ex-offenders. Workers receive hourly wages and are also eligible for other public benefits, including the Earned Income Tax Credit (and sometimes cash benefits under TANF).

A quite different approach involves mandatory work programs for welfare recipients, commonly known as “workfare.” Prominent examples have included the Community Work Experience Programs (CWEP) of the 1980s and New York City’s program in the 1990s. (Wisconsin’s welfare program also includes a subsidized employment component that is something of a hybrid of the two approaches.) Under “workfare,” welfare recipients are expected to work in return for their cash grants, but do not receive hourly wage payments and are not eligible for the Earned Income Tax Credit.

The primary goal of most of the paid work experience programs described above has been to improve future transitions to private sector employment, though less intensive variants of this approach have also been used to raise the current employment rates of those with poor job prospects. In contrast, the primary purpose of “workfare” has been to raise current work activity among welfare recipients. The desire to raise work activity in the latter case is motivated not only by the hope that it will improve their future employment prospects, but also by a belief in the work ethic and a desire to impose reciprocal obligations on those receiving public assistance. Other possible benefits of mandatory work programs for governments include deterring welfare applications and use by those who already have jobs or can readily find work on their own.

Issues and Concerns Over Work Experience Programs

In either form, work experience programs lead to higher employment rates, while they are in effect, for those who are either unable or unwilling to find employment on their own. For instance, employment prospects in the private sector appear to be quite limited for a subset of current and former welfare recipients who have very little recent work experience and suffer from a number of “barriers” to work—such as poor skills, physical or emotional disabilities, substance abuse problems, and so forth. Indeed, Sheila Zedlewski and Pamela Loprest of the Urban Institute estimate that 80 percent of recent welfare recipients suffer from at least one of these barriers, and that 30 to 40 percent suffer from more than one.

Other individuals, such as ex-offenders, suffer similar problems and additional ones as well (such as legal barriers to their employment in many sectors and great employer distrust). These difficulties are compounded during economic downturns, or in localities where few jobs exist. Some individuals on public assistance may also be reluctant to accept low-wage jobs, especially if they pay little better than what they receive on welfare, and will not work unless required to do so.

Work experience programs should raise employment rates among most of these individuals. When combined with training, supports such as childcare or transportation, and treatment of the other barriers they face, such experience might also improve job-readiness or other skills and provide a transition to unsubsidized work in the private sector. In addition to increasing employment rates, public jobs and work experience programs may also produce goods and services that are socially valued, particularly in poor communities where public amenities (such as buildings and parks) are often unkempt or in great disrepair.

What, then, are the potential limitations of these approaches? First, they can be quite costly; second, they may generate little net new employment if they substitute for employment that is already available in the private or public sectors; third, they may generate little in the way of new goods or services, especially of a kind that is socially valued; and fourth, they may have little effect on the future employment and earnings of participants.

What has the evidence shown on these issues? Costs, of course, depend on the nature of the program and on the support services provided. Most of the transitional jobs programs reviewed in a recent Mathematica report cost $1,000 to $2,000 per participant per month, with administrative and support costs often exceeding those of wages. In contrast, mandatory work experience or “workfare” programs have zero wage costs (by definition), but administrative costs that have run anywhere from $1,000 to $8,000 per participant on an annual basis.

In their recent survey of public employment programs, David Ellwood and Elisabeth Welty note that estimates of displacement of existing employment, as percentages of the new employment generated by these programs, range from relatively low (about 20 percent) to quite high (about 70 percent). But displacement can be reduced in a variety of ways—for example, by targeting workers (and/or areas) with poor employment prospects in the absence of these programs; by designing jobs that do not overlap significantly with existing private or public sector jobs; and by limiting the duration of the programs. Smaller programs are more likely to achieve these goals than are larger-scale efforts, though even some of the latter (such as the YIEPP demonstration, which guaranteed jobs for youth in entire communities) have generated significant net employment expansion when carefully targeted towards the right workers and jobs.

As far as the value of the goods and services produced is concerned, Ellwood and Welty note that estimates of the ratio of value of output relative to wage and administrative costs in work experience programs have ranged from .25 to over 1.00. Again, careful design of programs of moderate scale are most likely to generate socially valued outputs, but even in larger-scale programs (such as the one currently operating in New York City) there has been evidence of improved local amenities (such as park cleanliness) that are valued by residents.

Perhaps the most difficult goal for work experience programs to achieve is the improvement of future employment and earnings prospects. A few programs have clearly succeeded in this regard—most notably, the Supported Work program for welfare recipients. Public Service Employment also generated improved post-program earnings for disadvantaged women, though less so for adult men and youth.

The evaluations also suggest that combining work experience with training and other supports, sometimes in an intensive fashion, is more likely to generate post-program improvements in employment and earnings. For instance, most of the Community Work Experience programs in the 1980s evaluated by the Manpower Demonstration Research Corporation did not significantly improve the future earnings of participants. The largest positive impact was achieved by the San Diego Saturation Work Initiative Model (SWIM) program, which combined mandatory work with job search assistance and limited training. However, little positive impact was achieved in West Virginia, which ran a relatively large program that relied almost exclusively on work requirements. The impacts of work experience on subsequent employment and earnings have not been estimated in the more recent evaluations of work experience and transitional jobs programs, though the Mathematica report on transitional jobs strongly suggests a positive impact.

Of course, the more intensive the approach, the higher the program costs. For example, the Supported Work program improved the subsequent earnings of welfare recipients by as much as a few thousand dollars. It was also a very expensive program to operate—although one that in the end generated greater social benefits than costs. Programs that emphasize mandatory work, but provide few other important supports and services, are less likely to improve post-program earnings, although they may well increase current employment and produce valued goods and services, as noted.

New Issues and Evidence

A primary motivation for the use of work experience programs is to impose a reciprocal obligation on those receiving welfare, rather than to increase future employment and earnings. Large-scale work programs can certainly help to achieve this limited goal. However, since minimum wage requirements (and other provisions of the Fair Labor Standards Act) apply to any mandatory work program, the amount of work experience covered by the TANF grant plus food stamps may be insufficient to fund twenty-four hours per week of work, especially in low-benefit states. Thus, states may need to supplement their mandatory programs with paid work experience in some cases, unless the new law allows them to receive credit towards meeting participation requirements in other ways.

The costs to states of supervision and other administrative activities may also be significant, as noted earlier. In fact, the Congressional Budget Office has estimated that the additional costs of child care and employment programs under the new House proposal would range from $8 to $11 billion. If sufficient additional resources are not provided in the federal legislation, most states would need to reallocate resources from expenditures (on items such as education, training, and childcare for former recipients) to meet these new work requirements. Indeed, the fact that most states have not chosen to spend many of their resources in this way before now suggests they do not view these work programs as the best use of their limited funds.

Interestingly, advocates of transitional jobs programs (such as Clifford Johnson of the National League of Cities) are often highly critical of large-scale mandatory work programs for welfare recipients. They argue that several features of transitional job programs that are necessary for achieving positive impacts on subsequent earnings are lacking in mandatory work programs—namely, the direct payment of an hourly wage, supplementation through the EITC and other benefits, flexibility in participation and job assignment, careful supervision, and provision of needed supports, such as counseling.

Some new evidence on these mandatory work programs for welfare recipients paints a more positive picture. For instance, a new report by the Manpower Demonstration Research Corporation shows that both welfare recipients and employers were very pleased with the mandatory Community Service Employment program undertaken in Vermont during the mid-1990s. Also, a recent report from the Urban Institute in Washington, D.C. shows that the New York City work experience program evolved from a “workfare-only” model initially to one that provided additional treatments and services for most participating recipients. But the Vermont program was extremely small, covering only about 2 percent of those required to work. The authors of the report on New York City also stress that the flexibility provided by the previous TANF law was critical to the successful implementation and evolution of the program there.

Thus, whether the gaps in impacts between these two program models are truly as great as some advocates claim, and whether implementation of the new law will force states to forego services and programs that they currently provide, will only be known for certain with time and further research.


The studies reviewed above suggest that carefully structured work experience programs can have a positive impact on the employment rates of disadvantaged groups and can generate socially valued goods and services. When poorly designed, they can be wasteful and have small net impacts on employment and output. Improving future (as opposed to current) employment outcomes is harder, and seems to require a greater expenditure of resources on a wider range of supports and services. As usual, a fairly clear tradeoff exists between more ambitious goals and the resources needed to achieve them.

Under any circumstances, these programs should be limited to those with very weak ability to find jobs on their own—though it is often difficult to make this determination on an individual basis. The case for work experience programs is also stronger during economic downturns, and in depressed local areas where few jobs are available in the private sector. Even in these situations, the length of time which participants can spend in publicly provided jobs should be limited, so as to generate incentives to seek other employment; and the jobs themselves should be carefully chosen to limit displacement of other workers.

Beyond these considerations, sensible decisions by state policymakers require that they be clear about what goal they are trying to achieve. Thus, if a policymaker’s only goal is to raise employment among current welfare recipients, this can be achieved in many cases without wage payments or the provision of additional services, though the administrative and child care expenses involved in doing so will not be trivial.

A stronger case might be made for providing jobs to those not currently on the rolls—for example, to those who have reached time limits or who are not eligible for any other “safety net” program (such as unemployment insurance). But during a sustained downturn, or in areas where local economies are depressed, the numbers of individuals who could benefit from such programs (and the length of time over which they might need support) could be substantial, and program costs could rise proportionately. For states that are financially strapped (as most currently are), any such undertaking without additional federal support is hardly feasible.

On the other hand, if the policymaker’s primary goal is to raise future earnings among those who face serious labor market problems, then the choice might be to create smaller and more carefully crafted “transitional jobs” programs with substantial supports and services provided at greater cost per participant. More research and evaluation are needed to precisely measure their impacts, but past evaluations already indicate some successful models. And, given the growing presence in the labor market of some very hard-to-place workers—such as ex-offenders—the case for moderate expansions of “transitional jobs” programs beyond those that exist for welfare recipients is a fairly strong one.