In their deliberations about reauthorizing the 1996 welfare reform law, lawmakers have stumbled over an issue that was not expected to stir great controversy: What proportion of recipients can realistically be expected to work or participate in work-related activities each month, what activities should count, and how many hours of participation per week should be required?

To qualify for their federal welfare block grants under the 1996 law, states must engage at least 50 percent of single-parent welfare recipients (90 percent of two-parent family heads) in work or work-related activities for at least 30 hours per week each month. But the reauthorization bills introduced in both houses of Congress would substantially ratchet up these standards. To meet them, most states would have to revamp current programs and make major new investments. Research conducted by MDRC during the 1990s sheds light on what more stringent participation standards could mean for states and for welfare recipients.

So why tinker with what’s already working? One reason is technical. Anticipating the difficulties states would have meeting the ambitious new targets, the 1996 law lowered the required rate of work participation by one percentage point for every percentage point the state’s welfare caseload dropped below its 1995 level. With a variety of forces pushing caseloads down, the “standard” most states have had to meet has therefore dropped to zero, masking the fact that only a handful of states have anywhere near 50 percent of single-parent recipients meeting the 30-hour requirement each month. As a result, there is widespread agreement that the caseload reduction credit should be modified — whether phased out gradually or replaced by or combined with an employment credit — to keep states focused on raising employment.

The caseload reduction credit aside, how much should participation standards rise under the new law, if at all? One proposal is to increase to 70 percent the proportion of a state’s welfare recipients who must work or engage in work-related activities each month. Under this plan, a recipient would be required to log at least 40 hours per week to qualify as a participant in a given month. After three months on the rolls, no more than 16 hours of the recipient’s weekly participation in skill-building activities such as education, training, or even treatment for illnesses like depression could count toward the state’s participation requirement.

Are the proposed new standards achievable? As described in MDRC’s guide to TANF reauthorization, What Works in Welfare Reform, most states would find it very difficult to meet them, and nearly all would have to modify their current programs substantially. The 40-hour requirement is particularly onerous, because individual circumstances and the availability of program slots are in constant flux. In any given week, a significant proportion of recipients are ill or caring for a family member, awaiting assessment and assignment to the next activity, or waiting for a child care slot or a program slot to open up or have had their work hours temporarily reduced by their employer. States would not only have to develop and pay for the necessary slots and manage the ever-changing status of program participants but also establish elaborate monitoring systems to track hourly participation. These demands would be especially challenging for low-grant states — most of them in the South — where the block grant allocation per participant leaves few resources to develop the necessary activities.

Would the new standards improve program performance? In states with low engagement and participation rates, they might. But states that already engage a substantial share of the caseload could face several unintended consequences. First, the evidence suggests that once a welfare recipient reaches a threshold level of work participation — say, 20 hours a week — more is not better. Furthermore, as explained in Moving People from Welfare to Work, the most successful programs that MDRC has evaluated did not achieve participation rates anywhere near the proposed levels. To do so, they would have had to replace the very services that made them effective with public jobs programs that require recipients to work for their benefits. Such work experience programs, MDRC has found, are much more expensive and no more effective in boosting employment or reducing welfare dependency than is job search alone. Indeed, evidence from the National Evaluation of Welfare-to-Work Strategies suggests that a mix of job search, education, and training works best, but proposed restrictions on education and training could limit states’ ability to continue offering a range of services. Finally, combined with the welfare time limits already on the books, the new participation standards would oblige states to engage everyone in work activities — even the hard-to-employ, many of whom have health-related problems requiring more than three months of treatment. A new MDRC project focusing on this diverse group, Enhanced Services for the Hard-to-Employ, will test interventions aimed at the most common barriers to their employment.

In short, while the goal welfare reformers hope to achieve through higher participation standards is laudable, the evidence suggests that changing the proportion of recipients who must participate, the minimum number of hours, what activities count, and the caseload reduction credit all at the same time may be asking more than states can handle, particularly in the current fiscal environment. A more modest change might be the best next step.