What Did States Do?

Flexibility and devolution were hallmarks of the 1996 reforms. After enumerating four broad goals — support needy families, reduce welfare dependency and increase work, reduce out-of-wedlock childbearing, and promote the formation of two parent families — and establishing a set of rewards and penalties tied to those goals, the new act devolved primary responsibility for the actual design and implementation of welfare programs to the states. In state law and in practice, states overwhelmingly emphasized the first two goals while all but a few ignored the second two. Equally important, nearly every state added a new goal — to reward work and reduce poverty for welfare recipients who took jobs, at least until their months on welfare reached the state’s time limit on benefits.

Programmatically, what did states do with their new responsibilities and flexibility? Most did three things. First, they emphasized “work first” (and de-emphasized education and training) by requiring virtually all welfare recipients to begin searching for work immediately. These mandatory employment service programs also differed from past efforts in the frequency and intensity of the sanctions states imposed for failure to comply, including full-family sanctions that ended the entire family's welfare grant. Notably, only a handful of states and localities relied on “work for your benefits” work-experience programs or subsidized public employment to achieve these goals.

Second, in a little noticed but path breaking development, most states also helped to make work pay by allowing welfare recipients to keep more of their earnings without losing supplemental cash support. By not counting some portion of earnings when calculating welfare benefits, states allowed welfare recipients who took jobs to combine low-wage work with welfare benefits, in effect using welfare benefits as an “earnings supplement” to boost incomes.

Third, states placed limits on the number of months a family could receive welfare benefits, although the nature, enforcement, and, thus, the reality of those limits varied widely. While 17 states have established time limits shorter than the federal limit, several of the largest states — including California, Indiana, Michigan, and New York — either do not have a time limit or have substantially modified the federal limit, choosing instead to use state funds to pay benefits for those who exceed the federal lifetime limit.

Not surprisingly, the block grant framework — and, thus, the reality that TANF is a flexible funding source, not a program — spawned tremendous diversity among the states in the mix of mandates, incentives, and time limits employed, as well as in the emphasis placed on one or the other of these component parts. Some states — Iowa, Michigan, and Montana, for example — have dramatically increased participation in work activities by emphasizing mandates. Taking advantage of the caseload reduction credit, other states have placed less emphasis on mandates. Florida, Louisiana, Ohio, and Utah adopted time limits that are significantly shorter than the federal 60-month maximum and have enforced them strictly. Michigan and Vermont have no time limit at all. California, Connecticut, and Minnesota, among other states, use incentives in the form of generous income disregards to encourage work. These policy options are not mutually exclusive; on the contrary, most states are doing some or all of these things.

The direction a given state took also depended on local circumstance. States with big cities were preoccupied with making the transition from an education-first to a work-first orientation, and tended to focus first and foremost on mandates and the new message that welfare was a temporary source of support. Predominantly rural states had to focus on building the service network required to engage everyone, solving the transportation problems that make engagement difficult, and addressing the lack of employment opportunities that often characterize rural economies and tribal areas.

In addition to these programmatic strategies, states have availed themselves of TANF’s flexibility by transferring substantial sums to the Child Care Development Block Grant and the Social Services Block Grant. A handful of states also pushed the outer limits of TANF’s flexibility by counting state funds spent on other low-income programs against their TANF Maintenance of Effort spending requirements, in effect freeing up state dollars for other purposes.

Research Results: What Is Known

Work: A wide range of welfare reform strategies has increased employment and earnings among single mothers. Education and training played an important supporting role in the most effective programs.

• Nearly all of the approaches states have used — mandatory employment services, earnings supplements, time limits, and various combinations thereof — increased welfare recipients' employment and earnings.

• Mandatory employment services programs that tailored services to the needs of individual recipients — so-called “mixed strategy” programs that required some participants to begin by looking for work and others to start with education and training — proved more effective than more rigid approaches that simply assigned all participants to either a job-search-only program or an education and training-only program with little regard to their individual needs.

• About one in four welfare recipients did not find jobs even when followed for up to five years, in part because they were unable to surmount a range of significant employment barriers. Among those who did work, annual earnings were often low, and many remained poor or near poor, even when income from food stamps and the EITC were added. These findings underscore the need to develop new strategies to improve the employment and earnings prospects of the hard to employ whom welfare reform have left behind, to enhance job retention and advancement for the working poor, and to ensure that qualified working families receive Food Stamps and related benefits.

Welfare: Requiring participation in mandatory employment services and time limiting benefits decreased welfare receipt.

• Two of the approaches states used — mandatory employment services and time limits — reduced the amount of cash benefits paid out (usually because welfare receipt declined). Earnings supplement programs, by contrast, typically increased benefit payments and, thus, cost more than traditional welfare programs.

• The most effective mandatory programs returned to the government more than $2 in lower welfare costs and higher taxes paid for every $1 invested, a stunning achievement for any social program. Put another way, it would have cost these states more not to operate these programs.

Income and Hardship: Only programs that supplemented earnings increased income; while other program strategies did not increase income, they did not leave families financially or materially worse off.

• Only those programs that included provisions to supplement low earnings, usually by allowing recipients to keep some of their welfare benefits when they took jobs, increased income. The programs’ rules typically required parents to work full time in order to receive supplement payments.

• Programs that combined mandates, earnings supplements, and time limits — as most states currently do — increased income in the period before the time limit went into effect, but the income gains disappeared once the time limit was reached and welfare support was withdrawn.

• To date, there is little evidence that either mandates or time limits substantially increase material hardship. Conclusions regarding time limits must be considered tentative, however. They are based on only two studies, both of which offered substantial protections to vulnerable families and both of which were tested in an unusually strong economy.

Children: Whether or not children benefit depends on the program strategy and the age of the child.

• Welfare reform programs that led to increases in mother’s employment and income — specifically, those that included earnings supplements — consistently improved the school performance of elementary school-age children. The measured improvement is equivalent to raising their scores on a math or reading test from the 25^th to the 30^th percentile, and the positive effects persist throughout the five year follow-up period. By contrast, work mandates or time limits alone had few effects on young children; there was no consistent pattern of benefit or harm.

• The data on infants and toddlers are too limited to permit definitive conclusions. Evidence from two studies reveals little systematic harm or benefit to very young children’s later achievement or schooling when their mothers go to work.

• Regardless of program approach, policies that led to increases in mothers’ employment led, in turn, to small negative effects on adolescents’ school performance, although these policies did not lead to increases in more serious problems like school suspensions, dropout rates, or teenage childbearing. As mothers went to work, their teen children received less supervision, were more likely to work in excess of 20 hours per week, and or be responsible for the care of their younger siblings — activities that may have interfered with schooling.

Family and Marriage: Little is known about how to promote marriage or strengthen families through welfare.

Few programs increased the likelihood that a single parent would marry. Intriguingly, however, one of the earnings supplement programs did have a large and lasting effect on the likelihood that two-parent families would stay together.

• Several programs reduced the incidence of domestic violence, possibly because work meant less reliance on others, or because work meant less time at home, or because welfare systems are now offering more services for victims of domestic violence.

• Among low-income noncustodial fathers of welfare-dependent children, programs that combined employment and parenting services with more responsive child support rules increased child support payment rates. For the least employable and least involved fathers, they also increased employment and father involvement. While positive, these gains were small, suggesting that more intensive programs are needed.

Participation and Mandates: States have made large strides in increasing the percentage of welfare recipients who are working or participating in welfare-to-work activities, but they would have difficulty achieving TANF participation rates.

• Achieving high rates of participation in program services requires intensive staff outreach, ongoing monitoring and case management, as well as extensive resources to pay for the program activities and the support services needed to engage everyone. In any given week, a significant percentage of recipients are unable to participate — some are ill, some are between program assignments, some are awaiting child care, some do not meet the hours threshold, and so on. As a result, programs must work with almost all those targeted by a mandate in order to obtain the high participation rates required of them.

• Following TANF’s enactment, most states devoted additional resources to services, monitoring and case management activities, and they engaged a wider range of recipients than ever before. Nonetheless, if caseload reductions had not occurred, few states would have been able to meet both the hours requirement and the participation rate established in TANF.

• Programs that actively enforced mandates by reducing the welfare grants of those who did not participate produced higher participation rates than did low-enforcement programs. Beyond a threshold level, however, further increases in sanctioning rates were not associated with higher participation rates.

Implications for Reauthorization

Expanding the Role of Education and Training. The 1996 welfare reform’s “work first” emphasis was, in part, a reaction to the perceived shortcomings of the 1988 Family Support Act (FSA) reforms, which had strongly encouraged education and training in the hope that it would help people get better jobs. To some extent, this swinging pendulum of action and reaction in federal policy mimics the movement between a work-first and an education-first approach that has characterized policymaking in state after state. At its extreme, “work first” becomes “work only.” When administrators realize that not everyone can get a job, the pendulum swings back toward the point where everyone is assigned to education and training, few people are getting jobs, costs are high — and the pendulum again begins its return swing.

The challenge for policymakers is to find ways to maintain the employment orientation that underlies reform’s success, while opening the door to additional education and training. Results from carefully designed tests of job search-first programs, education-first programs, and mixed-strategy programs provide strong support for the idea that education and training have an important, although probably subsidiary, role to play in the future of welfare reform. The evidence indicates that both job search-first and education first-strategies are effective, but neither is as effective as a strategy that combines the two, particularly a strategy that maintains a strong employment orientation while emphasizing job search first for some and education first for others, as individual needs dictate. There is little evidence to support the idea that states should be pushed to one or the other extreme.

Welfare reform’s success in reducing caseloads and increasing employment adds new urgency to this debate. These accomplishments have led states to begin experimenting with job retention and advancement strategies to help former recipients further secure their foothold in the labor market and reduce their long-term reliance on other government benefits such as food stamps and child care. Investments in customized training or community college coursework to increase skills — sometimes in concert with release time from work — are among the many strategies states are beginning to use TANF resources to support. 

Adding Services for the Hard to Employ. As caseloads have fallen and as the five-year time limit approaches, states increasingly find themselves working with people who have a range of persistent, multiple, and, sometimes, severe employment barriers, such as substance abuse and depression, that make it difficult to get and keep a job. Treatment programs play a small but important part in states’ efforts to ameliorate these problems and promote employment among the hard to employ. If engagement in these activities does not count towards meeting their participation requirements, state officials have less incentive to work with these populations. It also places a funding obstacle in the way of the newly emerging focus states are now compelled to place on rehabilitation as a way to address the problems of the hard to employ. Recognizing this need, the Administration’s plan would allow engagement in treatment programs to count towards the participation standard, but only for three months out of every twenty-four. Six months might be a more realistic maximum.

Enhancing State Flexibility to Reward Work and Benefit Children. Although poverty reduction was not a TANF goal in 1996, most states’ conforming legislation included provisions to reward work. New research evidence shows that earnings supplement programs increase employment and income and that when the supplements are generous children benefit. By tying cash payments to earnings, these programs have cut the Gordian knot that has baffled welfare policy since the English Poor Laws — no longer do payments to poor families inevitably mean less work effort. This development has enabled states to refocus on welfare’s original purpose — to help children — without reducing the self-sufficiency efforts of their parents. Thus, states can now choose between program strategies that emphasize caseload reductions and strategies that emphasize benefits for children, while retaining the program's focus on increasing parental employment.

Several aspects of the current law, however, make it difficult for states to craft strategies that benefit children.

At the heart of the problem is the inherent conflict between earnings supplement and time-limit policies. Time limits tell recipients to “get a job, leave welfare, and bank your remaining months of eligibility.” Earnings supplements tell recipients to “get a job, stay on welfare, and let welfare supplement your earnings.” Implementing the two policies together virtually guarantees that a substantial number of people who take jobs while on welfare will unwittingly exhaust their months of welfare eligibility.

To avoid this outcome, states have two choices. One is to use the federally required state maintenance-of-effort (MOE) dollars to create either a “separate” or “segregated” state program for the working poor. By relying on state funds instead of federal funds, the federal time-limit clock is not ticking. The second alternative is to classify earnings supplement benefits as “non-assistance,” a categorization that allows certain payments such as employer subsidies, job retention bonuses, work expense payments, and so forth, not to be considered “assistance” under TANF, and, thus, not to be counted against the time-limit clock. Unfortunately, both strategies have shortcomings. The first places the fiscal burden of paying for supplements entirely on the state. The second exposes states to federal audits and the risk that the federal government will not accept the states’ definition of “non-assistance.” Without assurance of federal TANF reimbursement for long-term earnings supplement payments, states have been reluctant to choose these options.

TANF reauthorization could end these risks by either allowing states the option of stopping the federal time-limit clock when recipients take full-time jobs or, alternatively, by expanding and clarifying the definition of “non-assistance" to include ongoing cash payments or earnings supplements made to full-time workers. The Administration’s proposal to clarify what counts as non-assistance presents such an opportunity. Either strategy would enable states to create separate programs with federal financial participation to pay earnings supplements to the working poor outside the welfare system, effectively resolving the inherent message conflict that now exists between time limits and incentives, without fear of losing federal reimbursement. In a fixed-block-grant environment, this change would have no federal fiscal implications.

While the means appear arcane, the end is eminently clear. Message confusion between time limits and earnings disregards undermine both program strategies. By giving states greater latitude and the promise of federal financial participation when they choose to run separate programs for the working poor, time limits would continue to apply to welfare recipients who were not working, while earnings supplement policies could reward those who do the right thing and take jobs. And if the resulting state programs are sufficiently generous, available evidence suggests that better school performance among elementary school children would result.

The income of low earners would also be bolstered by ensuring that qualified workers receive the Food Stamp, health insurance, and child care benefits for which they remain eligible. Congress has built a safety net around work, but studies that have followed welfare leavers have found that fewer than half of the low-wage workers who qualify for these benefits receive them. Some job takers, unaware that they remain eligible for a range of benefits, do not stay in contact with the welfare office after they get jobs; other recipients are inadvertently cut off by the stringent quality-control system’s penchant for unnecessarily penalizing cases with earnings. The Administration’s proposed changes to the Food Stamp program would make it substantially more worker friendly; similar changes could be beneficial in the health insurance and child care areas.

Modifying Participation Standards and the Caseload Reduction Credit. The 1996 law established what many thought was an unachievable participation standard: In order to maintain their full TANF block grant, states would have to have 50 percent of the single-parent caseload and 90 percent of the two-parent caseload working or participating in approved activities for 30 or more hours per week. Most knowledgeable observers thought that no state would be able to meet these targets; yet all states did. Because a state’s participation standard is reduced by an amount equal to the percentage point reduction in its caseload since 1995, and because caseloads fell by 50 percent or more, most states’ effective participation standards are now at or near zero.

The states’ dramatic success in reducing caseloads has made the question of how to set participation standards in welfare reform’s next phase potentially one of the most contentious reauthorization issues. Some observers would like to end the caseload reduction credit because it sends the message that caseload reductions are the main goal of TANF. The Administration wants to end the credit to keep the pressure on states' performance. Not surprisingly, states would like to remain free of the participation standard and, thus, prefer to keep the caseload reduction credit in force.

As has already been noted, how Congress defines “participation” — the rate, how it is calculated, what activities count, and the number of hours of activity required — is one of several signaling mechanisms it can use to communicate to states what it wants. In an attempt to strike a balance between reinforcing the act’s original focus on work and the need to broaden the range of allowable activities to include more education and training as well as other services, the Administration’s proposal would make several important changes to the framework established in 1996. Notably, it would:

• increase the required participation rate to 70 percent and gradually eliminate the caseload reduction credit;

• increase the number of hours of required participation to 40 per week; after three months, 24 hours per week must be work; and

• broaden the rules to allow education and training activities to count but only towards the remaining 16 hours of required activity each week.

In assessing these proposed changes, it is important to address these questions:

Are the new standards achievable? Determining whether a particular participation standard is “feasible” depends on what counts as participation (the numerator) and who gets counted (the denominator) when determining the rate. None of the welfare-to-work programs that MDRC has evaluated to date — including the most effective programs — would have achieved either the participation rates currently in place (ignoring the caseload reduction provision) or the new rates being considered by the Administration, primarily because few of them could have met the weekly hours requirement.

For example, in a just-completed study that began in the 1990s, MDRC collected uniquely detailed participation data from several successful mandatory welfare-to-work programs to determine what the participation rate would have been had these programs been required to meet a 20-hour per week participation standard. We found that even though all of these programs vigorously enforced the participation mandate, increased employment, and reduced welfare, their monthly participation rates did not exceed 10 percent. Rates might rise to about 15 percent if reasonable assumptions are applied to take account of changes in the law that allow people with earnings to continue collecting welfare, to remove those who are sanctioned from the calculation, and to provide an employment credit (for three subsequent months) for people who left welfare for work. Only if criteria are relaxed substantially to count any activity in the month, regardless of the number of hours, could these same sites have reached participation rates of roughly 50 percent. 

To use a current example, MDRC is studying welfare reform in four large urban areas. Using the more generous definition of participation — ever participated in a month — the cities in the Urban Change study achieved rates ranging from about 30 percent to 50 percent of all adult recipients. These rates count all types of activities (including people assigned to education or substance abuse programs) and do not take into account the actual number of hours that people participated.

While these results are discouraging, they do not mean that higher participation rates could not be met. To do so, MDRC’s research suggests, the weekly hours requirement would have to be relaxed, and the rules would need to take specific account of several practical realities involving the changing status of people, the slots and services required, and the administrative difficulty of monitoring participation. Even in a tightly managed program, for example, a substantial number of recipients will be between activities at any time — they will have recently finished one activity and will be waiting for another to begin. Some recipients will be in the midst of a noncompliance review process that may lead to sanctions. Others may not be able to participate fully because they are temporarily ill or disabled, or caring for a disabled family member, or awaiting the outcome of an application to the Supplemental Security Income program.

States will have to confront a number of administrative challenges as well. The unsubsidized jobs open to recipients often do not provide 40 hours of work each week, and it is often impractical to try to add 5 or 10 hours of activities to a nearly full-time workweek. In addition, most program services are not designed to last for 40 hours a week, thus participants would have to be enrolled in multiple activities. To satisfy a “work only” participation standard for 24 hours per week would probably require states to develop large numbers of work experience or community service slots, a potentially expensive undertaking. And satisfying 40-hour participation standards would require major increases in childcare funding. All of these challenges are magnified in rural areas.

Finally, few states will be able to meet the reporting requirements. It is extraordinarily difficult and expensive to monitor hours of attendance for large numbers of welfare recipients being served by multiple providers. In most large cities that MDRC has studied, program records can supply information on assignments to activities and whether or not clients show up, but not the hours they actually attend.

In short, these considerations suggest that to achieve very high participation rates, there will need to be a very broad range of countable activities, flexibility in the number of hours required, and a measurement system that accounts for inevitable periods of down time and incomplete attendance.

Are the standards likely to generate more effective state TANF programs? It is hard to argue that the current system is ineffective. Even without effective participation standards to meet, welfare time limits have driven states to communicate the message that welfare is temporary, engage much of the welfare population in at least some activities, and emphasize work first in its programming. As a result, unprecedented numbers of people have left welfare for work. If states were to restructure their welfare-to-work programs to try to achieve the new vision, it is not at all clear that the result would be stronger programs. Indeed, the most successful programs MDRC has studied would not have met these participation standards and would have had to change substantially in order to do so.

While the ends the Administration’s plan attempts to accomplish are laudable — that is, seeking a balance between allowing additional services and retaining a focus on work — the means entail what appear to be unnecessary risks. Essentially, the Administration’s proposal would force states to increase the use of work-experience programs, possibly at the expense of the successful job search programs that have been most state programs’ first line of action. Instead of focusing on getting people off of welfare, states may become preoccupied with keeping everyone busy while they are on welfare.

Can the balance the Administration is seeking be struck without risking unintended effects on state programs? Those who would end the caseload reduction credit and reestablish effective participation standards have a point — an effective rate of zero doesn’t send the right message. And while states did spend more money on services when state budgets were flush, now that they are lean there is increased pressure to shift state dollars into other areas, suggesting that TANF service levels could suffer. Participation standards could help keep state attention focused on the need and the requirement to maintain their level of effort by matching resources to the goals of welfare reform.

How might the Congress go about establishing new participation standards? It might begin by borrowing a page from the Administration’s playbook. The Administration proposes to build an outcomes-based performance management system, whereby states will have to develop goals, and then measure and report their performance against those goals. Participation should be made an explicit part of this effort.

To give impetus and consequence to this effort the Committee might want to consider the following actions: First, invest in helping states establish the measurement and reporting systems that would be required to provide meaningful information on actual participation. Second, use the next three-year period to benchmark state performance, and then use these actual participation rates to establish individual state participation standards. Third, gradually phase out the caseload reduction credit. Fourth, while awaiting the new benchmark levels, leave the current participation standard in place, but allow states more freedom than they have now to count participation in education and training and substance abuse and mental health treatment programs. States should also be allowed to count participation in job search for up to four months in a year, a change that would facilitate state efforts to continue their work-first emphasis.

Exercising Caution on Time Limits. The Administration’s bill recommends few changes in the law’s time-limit provisions. Apart from the treatment of working families (discussed earlier), this approach seems consistent with the research findings available to date, which have not found evidence that time limits have caused significant harm.

But while this approach is consistent with current findings, the final time-limit story has not yet been told. Relatively few welfare recipients have reached the 60-month maximum; indeed, in more than 10 states the federal time limit will not kick in until later this year. Moreover, we do not yet know how former recipients will fare over long periods without welfare. A recipient with preschool-age children who reaches a lifetime limit at age 25 would have to survive without welfare for many years. More definitive data on the longer-term consequences of a loss of welfare eligibility will become available only over the next two or three years. Given the current uncertainty, it would seem prudent to build in a review mechanism that would provide Congress with an opportunity to revisit the 20 percent exemption provision before the end of the next authorization period.

Investing in Learning and Sustaining Innovation.  Congressional support for research has built a remarkable body of knowledge about what works for families and children — as well as government budgets and taxpayers — with respect to welfare-to-work strategies, earnings supplements, and, to a lesser extent, time limits. But the AFDC waiver structure that nourished that effort no longer exists, even though the need to build our knowledge base is greater than ever. We face a large new agenda to develop and refine our understanding about job advancement and retention; the role of public employment in rural areas and tribal lands where unemployment is perennially high; what works for the hard to employ who have severe, persistent, and multiple employment barriers; how to best engage low-income adolescents as their mothers go to work; strategies to promote and sustain healthy marriages and work with noncustodial fathers who owe child support and are unemployed; and the role of faith-based institutions in service delivery. Initially, the block-grant structure and the surpluses states enjoyed as a welcome by-product of the remarkable economic expansion of the late 1990s fueled a new round of state-led innovation. But the economic slowdown coupled with states’ reluctance to commit state funding to potentially ineffective new endeavors that may be hard to roll back have limited the amount of experimentation in a number of critical areas. Creating the wherewithal for states and localities to engage in bold experimentation and in rigorous evaluation is paramount. An annual set-aside of program dollars that states could apply for to pay for pilot tests of new ideas — tests that include requirements for rigorous independent evaluations — is needed.

The Administration proposes two such resource pools to spur the development and testing of new approaches in the marriage field, but only one appears to include research requirements. An additional source of program dollars is needed to cover the other areas mentioned above. In addition, the Administration proposes broad waiver authority to enable states to consolidate and integrate programs. Past efforts to couple waiver authority with rigorous learning have been essential to building knowledge about what works. But here again, no learning agenda is described.

Conclusion

Welfare reauthorization will likely extend the law for at least five more years and possibly for as long as ten. Thus, the revised law needs to be sturdy enough, flexible enough, and prescient enough to meet the safety-net needs of the nation’s poor, not just through the current period of economic uncertainty but also into the next economic recovery and beyond. States’ efforts to respond to the changing nature of the low-income caseload and redefine the mission and structure of welfare and related social services agencies, accordingly, underscore the daunting challenges to flexibility the new act must accommodate. As welfare caseloads have fallen and employment has risen, the needs of the working poor and the hard to employ have come into sharper focus. And now new evidence showing that reform can also benefit younger school-age children without sacrificing the employment gains of their parents has opened up a range of new options for states. As state agencies sort out these developments, as well as what priority to place on each, their choices will hinge largely on how a reauthorized act responds to the new landscape they face.

The Administration’s proposed plan provides a constructive framework for addressing many of these issues, one that builds on the work focus that research shows has driven much of TANF’s success. Notably, it adds the improvement of child well being as a purpose of TANF, and it allows a greater use of education and training and other services. From a child-outcome perspective, as the Administration’s plan notes and as the research evidence confirms, the strategies states employ can play a vital role in improving child well being. Nearly every state now has in place policies that would increase both employment and income and, thus, have the potential to benefit young children’s school performance. Without additional efforts to resolve the inherent conflict between state incentive policies and time limits, that potential may not be realized. The reauthorization process could give states the tools to resolve this issue.

With respect to education and training, the trade-off the Administration proposes would toughen participation standards and eliminate of the caseload reduction credit in return for giving states increased flexibility to count education and training activities. But available evidence urges caution. To meet the standard being proposed, the most successful state welfare programs we have evaluated would have to radically restructure their programs. This restructuring could have the unintended effect of distorting priorities, diverting resources, and driving up costs for child care and work experience slots, with the potential consequence of undermining the very success we are now celebrating. Building better information systems and establishing benchmarks on actual participation might be the best next step.

Most important of all, we should not let the remarkable accomplishments of the last five years get lost in the details or the politics of reauthorization. It isn’t broken; it could be improved; the challenge is to adapt TANF to the changing environment while building on its success.