A few years ago, old-paradigm welfare experts still were saying that the mammoth welfare caseloads of those days sprang from economic and social injustice. These Jeremiahs claimed that the new-paradigm approach—workfare—could do little to reduce these caseloads, because while most welfare recipients wanted to work, there weren't enough jobs. And even if jobs were available, shortages of day care, transportation, and job training made employment unrealistic for them. Yet, early work-related reforms in Wisconsin and Oregon cut welfare rolls by as much as one-fourth in a single year. Then, after federal welfare reform kicked in during early 1997, caseloads in the Aid to Families with Dependent Children program—re-christened Temporary Assistance to Needy Families (TANF)—fell 30 percent nationwide.
As hostile data overran their initial position, the old guard regrouped. Sure, they admitted, welfare caseloads fell in some states, but robust state economies were responsible, not workfare. Moreover, these experts warned, the reductions represented the "easy" cases; declines would taper off once reforms reached the harder-to-employ. Workfare supporters stuck to their guns, crediting work requirements for the diminished welfare rolls.
Who was right? If we disaggregate the national data, which mask huge state variation—some states reduced caseloads by 75 percent or more, while others have barely budged them—we can answer the question empirically.
If the robust economy is the prime mover, then states with speedy caseload decline should have zesty economies, right? But rates of caseload decline don't correlate to such key economic variables as unemployment rates or job growth. Indeed, states with higher rates of caseload reduction have slightly higher rates of unemployment than states with smaller reductions.
In fact, rates of state welfare decline closely track the different ways in which different states implemented workfare. Of particular importance is the firmness of a state's sanctions. States with tougher workfare programs treat a welfare check much like a paycheck, which the recipient must earn through a closely supervised job search, community-service work, or participation in on-the-job training. If the recipient fails to do his part, the state can cut off his welfare check. At the other extreme are states with lenient sanctions: if a recipient in these states fails to perform a required activity, she risks losing only the so-called parent's portion of the TANF check, some 12 percent of the whole. The bulk of the check—supposedly the "child's portion"—continues, impervious to bad behavior.
Ranking the 50 states by magnitude of caseload reduction, we find that nearly all those with big drops have prompt, stiff sanctions for bad behavior. In contrast, nearly all the states with low caseload reduction let recipients off easy. States with moderate sanctions clump in the middle. The two biggest states, California and New York—with a full third of the national TANF caseload between them—have wimpy sanctions. Not surprisingly, California ranks 36th among the states in recent caseload decline, while New York ranks 41st.
Another key variable is the timing of work requirements. States that require welfare recipients to work immediately upon enrollment tend to have greater reductions in dependence. Taking the most recent 18-month period with full data available (January 1997 to June 1998), it turns out that states with immediate work requirements and fast, tough sanctions cut welfare rolls 50 percent, on average. States with soft sanctions and no immediate work requirement rolled back caseloads only 14 percent.
Do the commendable effects of work programs taper off once states winnow the "easy cases"? Look to Wisconsin, which has the nation's toughest work requirements. The state's initial AFDC population, including a knot of hard-core dependency in inner-city Milwaukee, closely resembled the national caseload. For years, the pace of caseload reduction accelerated in Wisconsin as welfare rolls shrank. In recent months, the decline has slowed—but only after the welfare rolls had plummeted 90 percent. The remaining adult recipients must perform community service to get aid. By Wisconsin's standards, most states have barely begun to reform.