Welfare Reform, Phase Two

In 1996, welfare reform was rarely far from the headlines. across the country, states were overhauling their cash assistance programs for poor families. That summer, Congress passed and President Clinton signed a deeply controversial revamping of the federal-state system. The new law ended public assistance as a federal entitlement, in favor of a complex system of state block grants with work requirements and time limits.

Thanks to a fortuitous rendezvous with a full-employment economy, a great many people who had been receiving welfare found jobs. Many, but certainly not all, improved their livelihoods. As long as welfare caseloads were declining, states enjoyed a surplus of welfare funds and thus an expanding pool of resources to redirect to other benefits and services for low-income families.

Eight years later, the picture could hardly be more different. Caseloads are no longer falling in most of the country, and funding remains flat at 1996 levels. It seems virtually inevitable that there will be significant cuts in benefits and services funded through the welfare block grant in the coming years.

For the last two years, Congress and the White House have been unable to agree on terms for reauthorizing the 1996 welfare reform law, which expired in 2002. The discussion has been mired in disputes about required “work-participation rates” for welfare recipients. When the Senate briefly debated the issues in March, the press barely noticed. It now appears probable that the welfare program will receive yet another temporary extension -- its seventh so far -- and crucial decisions about its future will be postponed until the next Congress.

The share of the nation's poor children receiving welfare assistance is now at a historic low, even though employment among single-parent families has declined since 2000. This suggests a lot of hidden suffering.

In important ways, welfare reform was more successful than many liberals (and conservatives) expected in 1996. However, if the national goal was to reward work and reduce poverty, the 1996 law was profoundly incomplete. Unless Congress and the White House reconsider their approaches, a combination of stagnant funding and bureaucratic mandates will result in a welfare system that helps steadily fewer people with each passing year, while the set of supports for working families outside the welfare system remains an uneven patchwork.


During the late 1980s and early 1990s, there was a rapid increase in the nation's welfare caseload, with the number of families receiving assistance growing from 3.7 million in 1988 to 5 million in 1994. It's never been entirely clear how much of this caseload growth was due to economic downturn, demographic changes, or other factors, but it led to a strong perception that the system was “out of control” and that its incentives were perverse. In his 1992 campaign, Bill Clinton famously promised to “end welfare as we know it.” An emerging consensus among policy-makers held that some combination of compulsion and support was needed to link families receiving welfare with employment. There was far less consensus about the “right” mix of compulsion and help.

The new welfare law enacted in 1996 did four main things: It cut low-income assistance, most severely affecting benefits for legal immigrants; it made changes intended to improve child-support enforcement; it consolidated and expanded funding for child care; and it repealed the old Aid to Families with Dependent Children (AFDC) entitlement and replaced it with the Temporary Assistance for Needy Families (TANF) block grant. A previously sacrosanct federal guarantee of minimal cash assistance to needy families was ended. Instead, TANF provided each state with an annual lump sum. This roughly reflected what the state had been receiving from the federal government under the old welfare program in the mid-1990s.

With its block grant, each state was free to design its own cash assistance program, subject to two main federal rules: With limited exceptions, federal funds couldn't be used to provide assistance for more than five years, and states had to meet federal work-participation rates each year for families receiving assistance. States could impose other restrictions on assistance, and -- as turned out to be important -- states could use block grants for a wide range of activities besides operating cash assistance programs.

In implementing TANF, almost all states established programs of time-limited assistance. Most embraced a strong “work-first” philosophy, which discourages education and training and encourages recipients to take any available job as a condition of receiving benefits.

During this period, the nation coincidentally entered a period of economic growth, with demand for low-wage labor so strong that a two-decade decline in real wages for low-earning workers reversed and real wages began growing at the bottom. And, as a result of other deliberate policies, notably a large expansion of the federal Earned Income Tax Credit (EITC) coupled with minimum-wage increases enacted in 1996, low-wage workers realized income gains. In addition, health-care access was broadened for children outside the welfare system, and states began expanding child-care assistance for families not receiving welfare.

Academics are still debating the relative importance of each factor, but it's clear that in the first years of TANF, there was an unprecedented growth in employment among low-income single parents and an unprecedented drop in the welfare caseload (from 5 million families at its peak to a little more than 2 million). Not surprisingly, the jobs tended to be low-wage and the families were often still poor or near poor, but because welfare cash benefits are so low, even a family entering minimum-wage employment generally saw an increase in disposable income. The EITC helped, as did state policies of “disregarding” a portion of earned income in calculating the welfare check, which allowed working recipients to receive partial benefits for a limited time.

Even in the early years, however, there was a darker side. While most families leaving welfare got jobs, a significant minority didn't. Those families that either didn't or couldn't meet the requirements of the new system were more likely to be those with the least education, the shortest work history, and the most serious barriers to employment.

The fact that families could leave with jobs and still be poor, and that some families left without jobs, meant that the share of poor children receiving welfare assistance plummeted. Child poverty fell during the late 1990s, but welfare caseloads fell much faster, and the share of poor children receiving any assistance dropped from 62 percent in 1994 to 33 percent in 2002.


In the block-grant structure, a state got the same amount of money no matter how much its welfare caseload fell (or rose). With the welfare caseload decline, spending for cash assistance fell from $23 billion in 1994 to $11 billion in 2000. As a result, states now had billions of dollars in federal block-grant money each year to redirect to other purposes. Where did the money go? Broadly, it went in three directions: expanding child care, expanding other low-income benefits and assistance, and reducing other state spending.

By 2000, states were using nearly $4 billion in TANF funds for child-care assistance to low-income families, an amount larger than the entire federal child-care block grant. Child-care spending across the nation tripled between 1996 and 2000 (from $3 billion to $9 billion), and the number of children receiving child-care subsidies doubled. Admittedly, this was from a very low starting point: With the expansions, an estimated 2.25 million children were receiving subsidies (among about 15.7 million eligible children). Nevertheless, the progress was real. There were large increases in child-care assistance to low-income working families outside the welfare system. Moreover, states used the new funds to expand after-school programs, establish or expand child-care–Head Start collaborations, and increase spending for initiatives to improve child-care quality.

Second, states began using block-grant funds for benefits and services for low-income families that may not be receiving, or might never have received, welfare under the old AFDC program -- state earned-income tax credits, expanded transportation assistance, expanded services for victims of domestic violence, teen-pregnancy-prevention initiatives, demonstration projects to provide employment retention and advancement services for low-wage workers, even college tuition aid. States also increasingly began to use the funds to meet escalating costs in their child welfare systems. Many state administrators and advocates now look back, fondly, at a several-year period in which states were literally looking for good ideas for new initiatives to fund with TANF dollars.

Finally, the law let states cut their own spending for low-income assistance to 75 percent or 80 percent of 1994 spending levels, and most states did so. Moreover, states could use TANF funds to substitute for existing state spending for low-income families, so long as they satisfied their “maintenance of effort” level for state spending. While there are no precise figures, it seems clear that states were able to use federal TANF funds to substitute for billions of dollars of state spending.

Overall, then, the picture in the first years of implementation was mixed, but certainly more positive than many had feared. Employment rose, poverty fell, and states redirected welfare dollars to expand supports for families outside the welfare system. TANF was a “leaky bucket,” but it was still redirecting funds from the old welfare system into new supports for a much broader population of low-income working families.


If most things seemed to be moving in a positive direction in the first years of TANF, the opposite has occurred since then. First, when the nation's economic downturn began, the growth in employment among low-income single parents stopped (employment rates of single parents have fallen since 2000). And, at about the same time, the national caseload decline also stopped. Welfare caseloads are now growing in some states and falling in others, but, overall, the national caseload in 2003 was about the same, or only slightly lower, than it was in 2001.

People disagree about why the caseload hasn't gone up -- how much is due to time limits, more restrictive application procedures, more extensive requirements for families receiving assistance, less willingness to turn to welfare, restrictions on legal immigrants, etc. What's clear, though, is that the number of families receiving welfare assistance remained flat at a time when the number of families in poverty grew by half a million.

At the same time, states entered into the worst fiscal crisis since World War II, with aggregate state budget shortfalls approaching $80 billion in 2003. Ideally, the federal government would play a strong countercyclical role by providing fiscal relief during economic downturns, as almost all states have annual balanced budget requirements. But the block-grant structure was designed not to be countercyclical, i.e., it would provide states with the same amount of funding regardless of changes in the numbers in need. The law's small “contingency fund” had numerous design flaws so serious that not a single state was able to access it during the downturn. And a federal fiscal-relief bill was enacted to provide general help to all states, but the amount of funding -- $20 billion -- could, at most, address a limited share of a single year's shortfall.

The fact that TANF caseloads stopped declining meant that TANF was no longer an expanding source of funding for low-income benefits and services. But, under the pressure of state budget shortfalls, states drew on TANF funds to save programs that would otherwise be cut and to fill budget holes. States not only began spending their full block grants but also began dipping into reserves. For the last three years, states have essentially spent $2 billion more than their annual block grants each year. At this pace, most reserves will be gone within another year or two.

The impacts of the fiscal squeeze have resulted in cuts in basic assistance, employment services, and social services across the county. West Virginia is cutting cash welfare grants by 25 percent. Montana cut basic grants to avoid cutting child care. Indiana cut benefits for working families. Wisconsin cut job retention and training services for working families as well as literacy grants. Pennsylvania cut low-income housing assistance, family savings accounts, and violence-prevention programs that had been funded with TANF dollars. The District of Columbia cut adult basic education, eviction mediation, and a tuition-assistance program previously funded with TANF dollars.

The pressures have also led to sharp cuts in child care. Since 2001, at least 23 states have limited child-care assistance. For example, Nebraska, Ohio, and Oregon each cut back eligibility from 185 percent of poverty to 120 percent of poverty. In more than half of the states there are now either waiting lists for child-care assistance or no open slots for working families that aren't welfare recipients (or that have recently left welfare). States also significantly increased family co-payments. In Louisiana, for example, families that had been paying 15 percent of the cost of care had their co-payments increased to 70 percent of the cost. Other states cut quality initiatives. Kansas has cut back funding used to help teachers earn credentials and improve their teaching. In Maryland, funding for resource and referral services to help families find child care has been cut back by more than 30 percent. A key element of state efforts in the late 1990s -- the effort to ensure that no family needed to turn to welfare to get child-care help -- is thus increasingly being lost due to the combination of state budget squeezes and no new federal support for child care.

What's striking about these cuts is that they're each seen as individual state policy choices, rather than as flowing from decisions -- and lack of decisions -- at the federal level. If there had been a federal proposal to simply cut basic assistance, or cut child-care funding, there'd have been a political controversy and a national debate. But by establishing a “flexible” block-grant structure with flat funding, Congress virtually guaranteed that a steadily less responsive structure is simply perceived as the policy choices of each state, and that there is no national debate about the underlying federal policy that virtually compels state cuts.

TANF was scheduled to be reauthorized in 2002. Reauthorization should have focused on the next steps in anti-poverty policy: helping low-income workers get better jobs, ending restrictions on benefits for legal immigrants, and expanding work supports, most notably child care, for low-earning families. The administration proposed freezing TANF and child-care block-grant funding levels for the next five years, leaving TANF restrictions on benefits for immigrants unchanged and increasing the work participation requirements for TANF recipients.

The administration proposed to discourage TANF assistance for full-time participation in education and training, to encourage the use of unpaid work programs, and to evaluate states by whether welfare recipients were engaged in 40 hours of activity each week (regardless of whether they were employed or getting jobs). Understandably, there was little enthusiasm for this approach from states, but it became the centerpiece of the reauthorization debate, and for the last two and a half years the debate has been largely paralyzed by disputes about what the participation-rate rules should be for families receiving welfare assistance.

One hopeful area of policy debate has been child-care funding. The pending reauthorization bill in the Senate would increase child-care funding by $7 billion over five years, an amount that would meet the cost of new Senate work requirements for welfare recipients and also preserve current slots for other working families. The House, however, proposes only $1 billion, which would accomplish neither objective. Despite the best efforts of advocates, too much of the debate has focused on how much child care is needed just to meet welfare participation rates, and child-care funding generally is now being held hostage to the TANF reauthorization logjam.


The best result would be for Congress to adopt a reasonable reauthorization bill this year. Most likely, though, congressional action on TANF and child-care reauthorizations will be put off until after the 2004 election. If Congress is beginning anew next year, there are four crucial things to keep in mind.

First, the 1996 reforms involved two components: changing welfare and building work supports outside welfare -- the EITC, minimum-wage increase, improved child-support enforcement, broadened health care, expanded child care. Welfare families are now only one-third of families in poverty, and an even smaller share of the overall low-income population. There can't be continued progress if the key supports for working families don't improve. The next Congress needs to focus not just on welfare rules but on broadening the other supports that make it impossible for low-income families to get by in relatively low-wage work.

Second, the nation depends on a structure of supports for low-wage work because the nation has an extensive low-wage labor market, and much of the work available to families with limited skills is low wage. While states have increasingly come to recognize the limits of an extreme work-first approach, there remains a serious risk that federal law will impose such requirements on all states. It's essential that states are given the flexibility and the resources to focus on education and training, and to prioritize efforts to help workers get better jobs. In the late 1990s, there was much discussion of the need to build institutional support for efforts to help low-wage workers advance, but little was put in place, and much of that is now at risk from budget cuts.

Third, about one in four poor children in the United States now lives in immigrant families. Current law bars states from using TANF funds or providing federally funded health care to most legal immigrants who have lived in the United States for less than five years. In the long run, any efforts to address child poverty needs to include extending assistance and employment services to immigrant families, and it's essential to end the continuing discrimination against these families.

Finally, it may not be much of a surprise to say that money matters, but that is surely the TANF experience. When funding was (effectively) going up, states innovated and expanded supports; when funding stopped going up, states cut and retrenched. A block-grant structure turns out to have important pros and serious cons. On the one hand, it can provide for flexibility and innovation; on the other hand, block-grant money gets diverted to other uses, it's tremendously difficult to build in accountability, and funds that are adequate in boom times are hopelessly insufficient during recessions. What's more, once initial funding levels are set, it's hard to make a convincing case for increasing funding. At minimum, block grants need to be indexed for inflation and adjusted both to provide countercyclical funds and resources commensurate with their general mission.

There was a moment when TANF seemed as if it could be part of a new broad strategy not just to require work but to reward it. That effort is now in jeopardy. In 1996, America ended welfare as we knew it. Now we need to support and reward work -- and end poverty as we still know it.

Mark Greenberg is the director of policy at the Center for Law and Social Policy in Washington, D.C.