Is Devolution Working? Federal and State Roles in Welfare

The decade of the 1990s was an ambivalent period for American federalism. While the national government strengthened control over economic policies to assure uniformity and help businesses compete nationally and globally, Republican leaders in Congress worked to decentralize many public services, moving control from Washington down to governments “closest to the people,” especially state governments, a clear majority of which were headed by Republican governors.

During 1995—96, then-Speaker Newt Gingrich and other GOP leaders of the 104th Congress made block grants a centerpiece of their “Contract with America.” They supported omnibus block grants for Medicaid, food stamps, and welfare to enhance opportunities for flexible action, especially by state governments. President Clinton consistently vetoed these block grants. Then, in August 1996, on the eve of the presidential election and against the advice of many of his liberal advisers, he signed one—the welfare reform bill—into law.

A Devolution Revolution

The Personal Responsibility and Work Opportunities Reconciliation Act of 1996, the welfare block grant, has fundamentally transformed the roles of national, state, and local governments, as well as thousands of nonprofit and for-profit groups in the field of human services.

The 1996 law is complicated—even, in a sense, schizophrenic. It sets strict federal requirements about work and limits on how long families can receive federally aided welfare benefits, now called Temporary Assistance for Needy Families (TANF). But it also encourages states to develop their own welfare reforms by adopting variants on the work-first approach, emphasizing quick engagement in work-related activities and moving people off public assistance.

States not only have accepted but have frequently embraced the goal of getting people into jobs, especially unsubsidized jobs, as quickly as possible. Moving beneficiaries into jobs and reducing caseloads have been critical operational goals going all the way down to local welfare offices. Work participation rates are at historic highs, particularly in states that had already been reforming welfare under federal waivers.

States have used their flexibility to signal the emphasis on work. They have organized new ways to deliver services, the most striking and widespread being linkages to workforce development systems. Many states have demanded that people who apply for cash assistance first search for jobs, cooperate with child support enforcement efforts, or engage in other activities before they qualify for benefits.

States have increased their services, especially those involving work supports, such as child care and transportation, while greatly reducing their cash assistance. In 1999, the median state spent only about 55 percent of its total “welfare” budget on cash assistance. Most states do not appear to have greatly reduced total spending on social services, and some have actually increased it. The “race to the bottom” that some critics anticipated never took place. One reason may be the design of the block grant—TANF offered states generous funding during a time of economic growth. But another reason may have been the greater political legitimacy accorded to welfare programs because family heads who are working are seen as more deserving of support.

Federal agencies have given states flexibility in using their welfare block grant funds. Rules are loose, permitting states to assist the heads of working families without making their families subject to federal time limits. The task of defining “work activities” has been turned over to the states. And Washington has not rigorously audited state spending.

State Variations

While all states share the focus on work first and the activist approach to reforming their human service systems, many variations exist in tone, structure, spending, and operations. States differ, for example, in their operational goals. Some try to meet TANF’s performance requirements by engaging a large and growing share of their cash assistance caseloads in work activities. Others reduce caseloads, either by moving beneficiaries into jobs or providing them with alternatives to continuing cash assistance. Although some states are beginning to incorporate other goals in their welfare systems, work participation rates and caseloads remain dominant criteria in day-to-day operations, if for no other reason than that failure to meet the required work participation rates, after factoring in a caseload reduction credit, can exact big financial penalties.

States differ even more in how they try to achieve these goals. Minnesota, Michigan, and Kansas, for example, all emphasize work engagement rather than caseload reduction, but each pursues a different approach. Minnesota emphasizes financial incentives to work. It gives cash assistance to families well over the poverty level by offering generous benefits and disregards. It also supplements earnings with a state earned income tax credit and provides substantial work supports, including a child care program available to many working families. The aim is to encourage poor families to mix earnings and assistance and increase total income (and to increase work participation rates among families on the rolls).

Michigan’s special “Project Zero” sites tried to reduce to zero the number of families on welfare and without earnings. The state essentially eliminated federal time limits by committing itself to use state funds to assist people beyond the five-year limit. But instead of financial incentives, Michigan uses a wide range of training, employment, and educational services through its workforce development system to move people into jobs and keep them there.

Kansas, like Michigan, uses a service strategy. But rather than focusing on employment needs and services, case managers draw from a mix of services and assessment tools to deal with a wide range of family needs, including learning disabilities, mental health, and substance abuse.

Likewise, states that stress caseload reduction also use varying structures and services. Like Minnesota, Wisconsin offers incentives to which families are expected to respond. But Wisconsin built its reforms around a vision of welfare strongly rooted in the “real world” of work. It stresses independence from cash assistance, while Minnesota mixes cash assistance and earnings to maximize income. Wisconsin gives cash only for actual hours “worked” in approved activities—and without regard to family size. Access to publicly subsidized work is restricted (most families are eligible only for services and in-kind benefits), and services (such as child care and medical benefits) require co-payments. Texas also signals independence from public assistance, in part by emphasizing thecosts and inconvenience of cash assistance through itsstrong anti-fraud efforts and extensive “up-front” requirements.

Some states offer varied services to minimize cash assistance caseloads. West Virginia has tried to keep families from using up their cash assistance by diverting them to food stamps and other sources of support and by offering in-kind benefits (such as dental benefits and clothing vouchers) to help families get by without getting or staying on the welfare rolls. Finally, some states have devolved considerable control and responsibility for developing programs down to localities, in most cases county governments, though sometimes private contractors. In these “compound systems,” like the one in Ohio, several different strategies or variants develop within a single state.

Some of this state-to-state variation seems to grow out of earlier differences among states under the old Aid to Families with Dependent Children (AFDC) program. States that emphasize caseload reduction tend, for example, to be states that offered (and still offer) relatively low cash benefits, while states that focus on work participation tend to be those that provided (and still provide) relatively generous cash assistance. States with decentralized administrative structures under AFDC generally have decentralized structures under TANF. But there are important exceptions. First, goals are not always predictable based on spending and benefit levels. Tennessee, for example, is a low-benefit state that stresses work participation. To motivate clients to engage in work and educational activities, it offers cash bonuses for achieving certain thresholds, such as completing a training or educational program. Second, some states with centralized systems under AFDC, such as Florida and Washington, now give broad discretion to local districts, administrators, or case managers. Third, as states develop distinct service strategies, differences are arising that were not at all evident under AFDC.

How Is Devolution Working?

The editors of the Brookings Review posed the question, “Is Devolution Working for Welfare?” The answer, we believe, is: “It depends on the criteria by which reform is judged.” The five perspectives that follow are merely suggestive of the different criteria, not definitive. Many people viewing the “new world of welfare” have still different, and often shaded, views.

Let a Thousand Flowers Bloom

To the extent that the block grant was expected to be a federalism initiative, an instrument for permitting different welfare reform strokes for different state and local governments, YES, welfare reform is working. According to this perspective, state and local preferences and conditions should really matter, not only to take into account economic circumstances and resources that can influence programeffectiveness, but also to fashion welfare programs that command support and ownership within a state’s or locality’s political culture. As noted, states and localities already vary widely in how they implement welfare reforms and appear to be moving in ever more distinct directions.

Encourage Work and Reduce the Rolls

In a second view, the controversial 60-year history of AFDC finally pushed political leaders (Republicans and Democrats alike) to act boldly in 1996 “to end welfare as we know it!” If ending cash support without work requirements is the critical goal, then, YES, the new signaling and sanctions of the 1996 welfare reform act are working. The rolls are down more than 50 percent nationally since 1995 (before most major AFDC-waiver reforms were implemented). Employment rates for single mothers with young children grew from 50 percent to 62 percent between 1995 and 1999. Average state work participation rates for adult caregivers on cash assistance reached nearly 40 percent in 1999, much higher than under previous welfare reforms. True, much of this growth in employment and decline in dependence is due to the strong national economy. And only a handful of states (such as Oregon and Wisconsin) have been able to engage nearly all their cases in some sort of work. But even the typical states have seen major and rapid changes in employment both off and on the welfare rolls, and there are good reasons for attributing a sizable share of these changes to the new systems of welfare.

Welfare as a “New Social Bargain”

A third view is that welfare reform was meant to be a new social contract. Poor family heads must work; in exchange, governments will provide services to help them do that. The focus here is on services to help families be self-sufficient. In this view, the answer to whether devolution is working is, YES AND NO. Most states have increased spending on work supports, especially child care, child support enforcement, job placement services, and transportation. A few states are providing services involving skills development, short-term benefits for poverty prevention, and assistance for hard-to-employ clients burdened with learning disabilities, mental health problems, and substance abuse.

But there are still important limits to this social bargain. The first is resources. Even states that spend a lot on services, such as Minnesota, have waiting lists for child care benefits. Second, states are still struggling to connect welfare agencies to other service bureaucracies. Big differences in information systems, bureaucratic missions, and eligibility requirements divide states’ social service agencies and workforce development systems. These issues become ever more daunting as state welfare offices try to draw on agencies involved in mental health, substance abuse, child welfare, and other services to address obstacles to employment. Many states and localities have a long way to go to make services accessible to all who need them.

Reform and the Working Poor

In a fourth view, welfare hasn’t changed enough. Many poor and working-poor families aren’t getting benefits to which they are entitled—Medicaid, food stamps, child support from absent parents, the earned income tax credit, and child care. The concern is that working-poor families are not connected to social agencies. While many states are working on the problem, critics see their efforts as inadequate: too many people who need aid are not getting it. These critics believe that the new signals of welfare avoidance and work create problems for poor families that are not counterbalanced by providing decent jobs and related work-support services. Also welfare procedures remain complex and burdensome, and busy working families cannot easily comply with them. Other observers complain about a lack of data on what is happening to poor families, especially children and families who aren’t getting cash assistance.

Family Formation and Reproductive Practices

Even a cursory reading of the 1996 welfare law produces a jolt, for though the law stresses work, its strongest admonitory purpose is promoting marriage and reducing out-of-wedlock and teen births. And although efforts are being made to change personal behavior, states have not invested much in them. Part of the problem is organizational. Human service and employment agencies have little experience working with family planning clinics and public health agencies. Yet part is political. While state legislative debates over welfare reform showed considerable consensus over the work-related goals, they usually sparked strong disagreement over how to accomplish the family formation objectives. From this perspective on what welfare reform is supposed to accomplish, the answer is NO, the 1996 law isn’t working.

Looking Ahead

In the new world of welfare, two groups need particular attention. One is multi-problem families. The “silos” of governmental bureaucracies make it hard to look across these families’ needs and connect work-related human services, such as child care, job aid, transportation aid, and treatment of drug and alcohol problems, with other services, such as dealing with threats of domestic violence and child abuse. Some efforts to integrate services for these families are impressive, but they are few and far between.

The second group needing attention is the working poor—families whose head is working and whose goal is to stay in, and move up in, the labor force. The 1996 welfare reform law allows the use of federal funds to aid adults and children in working-poor families. New efforts under way in many states include informing people about and helping them receive “safety net” benefits (Medicaid and food stamps), as well as providing work-support services like child care, the earned income tax credit, transportation, and job services in a way that avoids stigma and preserves dignity. But, again, progress varies widely by state.

Three things are critical in making the changes required. The first is time. States and localities with the most impressive human service systems—Wisconsin, Michigan, Oregon, Riverside County in California—have been working on welfare reform since the 1980s or early 1990s. Even so, they still have many implementation issues to address. Other states may take even longer because of larger populations, more internal diversity, more autonomous counties or cities, or less professional bureaucracies.

Second, devolution demands stability in rules and funding. States and localities are making extremely complex institutional connections through contracts and interagency agreements, creating expensive information systems, and implementing vast training and retraining programs—all things that can be upended by unexpected changes in rules and funding. Rapid innovation and thorough planning at the periphery demands continuity at the center.

Third, devolution requires visibility for the whole range of families that the new welfare is expected to serve and at the level where service systems and policies are being formulated. If states and localities are to be accountable not just to the relatively few people on the cash assistance rolls but to the many working-poor and low-income families, more and timely information is needed about the economic status of all such families, even down to the county level. Local information of this sort does not now exist, and thus a major piece of an adaptable and accountable devolution initiative is still largely missing. Diversity may be the name of the game for shaping and implementing social policies in American federalism, but effective diversity needs stability at the top and information at the bottom.