The Failure of Welfare Reform 

How Bill Clinton’s signature legislative achievement tore America’s safety net.

President Bill Clinton signs the welfare reform bill in Washington on Aug. 22, 1996.

Stephen Jaffe/Reuters

If you want a sense of how thoroughly America’s welfare system has decayed thanks to the reforms Bill Clinton signed into law two decades ago, consider Arizona.

Despite being home to one of the nation’s most crushing child poverty rates, the state has all but stopped giving cash assistance to its needy. During 2014, for every 100 poor families with children in Arizona, just 8 families received aid. And even that tiny fraction is likely to shrink. Last year, while trying to chip away at a $1 billion budget deficit, lawmakers lowered the maximum amount of time Arizonans could receive welfare payments before being kicked off the rolls permanently—it’s now just 12 months.

This was a first. Most states enforce a five-year time limit. Some, including Arizona, have gone as low as two years. None had tried the one-year-and-you’re-out approach.

The move was expected to save up to $9 millionfor perspective, the state’s board of tourism spends about three times that much annually—while cutting off aid to some 2,700 children. Nonetheless, Republican state Sen. Kelli Ward, who is now challenging John McCain for his U.S. Senate seat, said that the 12-month cutoff would “encourage the able-bodied to treat welfare like a safety net rather than a hammock.” The line might have been more convincing if benefits for a family of three in Arizona didn’t already max out at a miniscule $278 per month. Some hammock.

The death of welfare in Arizona isn’t an exception; the program has shriveled just as badly in many other states, especially across the Deep South and West. Even in most of the states that are more generous than Arizona about giving benefits to the poor, welfare has still dwindled during the postrecession era. That’s because its funding was never designed to grow along with inflation—or with the U.S. population. For all intents and purposes, welfare is becoming a zombie system rather than the bridge from poverty to work that its reformers envisioned.

And, of course, it was the signature legislative achievement of Bill Clinton’s presidency.

* * *

The word welfare tends to get thrown around a bit loosely; conservatives have a habit of using it to describe any program for the poor, from food stamps to housing assistance. Traditionally, though, the term has referred to the federal system that sends cash to poor parents and their kids.

For most of the 20th century, that system was Aid to Families With Dependent Children, a program launched during the Great Depression to help widows and their families. Over time, it transformed largely into a source of financial support for single mothers who had never married or who were divorced or separated. By the time Clinton vowed to “end welfare as we know it” during his first presidential run, AFDC had been maligned for decades. Critics blamed it for incentivizing the dramatic rise of out-of-wedlock childbirths that started in the 1960s. (The hypothesis was that mothers were choosing not to marry so they could keep their free checks from the government.) These critics claimed that the program bred poverty and dependence by discouraging work.

Making matters worse, the program’s rolls unexpectedly surged by 27 percent in the early 1990s. (Experts disagreed how much of the increase was due solely to the weak economy and how much was because of the rise of all those single moms, among other social changes.) By the time the numbers peaked in 1994, more than 5 million families, including about 9.6 million children and 4.6 million adults, relied on AFDC. That year, three-quarters of voters told the American National Election Studies that they favored the idea of capping the time families could receive welfare benefits at just two years. More than half of respondents said the government should spend less on welfare.

Americans value work (often to a fault) and have always been philosophically uncomfortable with letting adults live on the public dole. None other than Franklin Roosevelt once said that to give out “relief” was “to administer a narcotic, a subtle destroyer of the human spirit”—an eloquent 1930s way of describing welfare as a hammock. But much of the resentment toward AFDC also boiled down to race. By the time Ronald Reagan was regaling audiences with tales of Cadillac-driving welfare queens scamming taxpayers, African Americans made up about 40 percent of the program’s caseload. Studies found that whites’ opinions about welfare recipients were strongly tied to their feelings about blacks in general and in particular, their work ethic. Few artifacts capture how race dominated the issue more arrestingly than the notorious 1996 New Republic cover urging President Clinton to sign the welfare reform bill then in Congress. It featured a picture of a black woman cradling an infant while smoking a cigarette, with the words “Day of Reckoning” splashed above. Inside, the magazine made its case:

Why, after all, do we care about welfare? Not for fiscal reasons; Aid to Families with Dependent Children (AFDC), the main welfare program, consumes only about 1 percent of the federal budget. We care because welfare is implicated in America’s gravest social problem, the existence of isolated, depressed neighborhoods, the vast majority either black or Hispanic, where intact families and working fathers are practically nonexistent. You can argue about welfare’s role in creating this underclass, but there is little doubt that welfare sustains it.

It hardly seemed to matter that almost as many of AFDC’s beneficiaries were white as black. In the white public’s mind—even to the editors of a putatively center-left policy magazine—it was a handout to poor people with dark skin.

While racial resentment drove much of the welfare debate, the truth was that AFDC had flaws. Some families did end up chronically dependent on the program—which served nobody well, given how meager the benefits were. Almost a quarter of mothers on welfare spent 10 years or more on its rolls, and at any given moment, those women made up a majority of all active cases. Some parents faced personal difficulties that kept them from working, such as health problems or learning disabilities. Others were afraid to lose their benefits, which shrank quickly with each dollar they earned from a job; many women from this group did go back to work, but off the books. “Welfare was a terrible system,” says Johns Hopkins sociologist Kathryn Edin, who has spent much of her career studying welfare programs and their recipients. “It paid women too little to survive. And it made you feel like scum. These are not good things for women who are the primary caretakers of children.”

Starting with Lyndon Johnson, presidents had attempted to reform AFDC to make the program better at moving families from public assistance to employment. Ronald Reagan signed legislation that required states to create job-training and job-placement programs for welfare recipients and ensure a modest percentage of parents receiving aid were also working, though the rules were never strongly enforced. Both he and George H.W. Bush started giving states waivers that allowed them to experiment with their own reforms; neither saw much in the way of immediate results. When his turn came, Bill Clinton at first embraced the waiver approach, too—he ultimately gave 43 states permission to tinker with welfare’s structure. Many introduced policies such as limiting how long families could receive cash assistance or stricter work requirements that later became part of the federal reform. But when push came to shove, Clinton chose to dismantle AFDC entirely and start from scratch.

The president’s original reform bill, proposed in 1994, was both more incremental and more generous than what eventually passed into law. In keeping with his campaign pledges, it did require younger welfare recipients to go to work after two years, but in return it guaranteed them a (low-paid) public sector or government-subsidized job; there were lots of loopholes on those time limits as well. The original version of the bill also would have further expanded the earned income tax credit for low-wage workers—a credit that Clinton had already doubled in 1993—and beefed up under-resourced state child support systems. In short, it was the kind of bill one might expect from a liberal-ish president who wanted to reorient the welfare system toward helping parents get work.

It was also not the sort of thing that Republicans who’d seized control of Congress after the historic 1994 midterm elections were ever going to consider seriously. The party had campaigned on a more punitive vision of welfare reform contained in its Contract With America, which also called for a good deal of federal budget slashing. Conservatives proceeded to craft their own reform legislation over the next two years, amid liberals’ pushback.

The GOP’s plan included lots of sticks to prod mothers to work, such as time limits on how long they could receive benefits, but little in the way of carrots such as, say, guaranteed jobs. New York’s patrician Sen. Daniel Patrick Moynihan predicted that the GOP’s plan would leave hordes of homeless children “sleeping on grates.” Ted Kennedy of Massachusetts called it “legislative child abuse.” Clinton’s own Department of Health and Human Services produced a report suggesting it would send 1.1 million children into poverty.

Senator Daniel Patrick Moynihan talks to reporters November 9 about an administration study sent to Congress on a Senate walfare reform plan endorsed by President Clinton.
Sen. Daniel Patrick Moynihan in Washington in 1995.

Mike Theiler/Reuters

The president vetoed two versions passed by Congress. But he eventually made the election-year calculation to compromise and sign a somewhat more moderate third bill that attracted backing from Democrats including John Kerry of Massachusetts, Russell Feingold of Wisconsin, Tom Harkin of Iowa, and Joe Biden of Delaware. It was still divisive enough that three HHS officials resigned in protest. (One of them, Peter Edelman, went on to write a long article about welfare reform for the Atlantic, titled “The Worst Thing Bill Clinton Has Done.”) Amid the acrimony, the Personal Responsibility and Work Opportunity Reconciliation Act became law on Aug. 22, 1996. At the signing ceremony, the president brought along a former welfare mother from his home state of Arkansas, Lillie Harden, as a speaker and stage piece. “In a sweeping reversal of Federal policy, President Clinton today ended six decades of guaranteed help to the nation’s poorest,” the New York Times proclaimed.

And what, exactly, arose in its place? Old welfare had been a federal entitlement. The program was more generous in some parts of the country than others, because states set their own eligibility rules and cash benefit levels. But in the end, Washington helped cover the cost for each and every new enrollee, no matter how many signed up. If a family was poor enough, the support was guaranteed.

New welfare was something else entirely: a diminishing pool of money that guaranteed the poor precisely nothing. The reform bill replaced AFDC with a program called Temporary Assistance for Needy Families. On paper, it required more welfare enrollees to work. It also limited how long families could receive federally funded cash benefits to five years or less over a lifetime—in theory, single parents would be forced to find jobs sooner or later (in practice, there were exceptions that would let some families keep benefits longer than five years). But most importantly, it turned welfare into what’s known as a block grant system: Each state would receive an annual lump sum of cash from the federal government that lawmakers were then expected to spend on their own customized welfare programs. (States were also required to put up some of their own matching funds.) While this may just sound like tedious budget talk, the change had two enormous consequences.

First, the block grant structure capped federal spending on welfare. States would get their yearly payments and no more. In practice, this put a ceiling on the number of families who could enroll. And that ceiling was set to drop. As a budget savings measure, Congress chose not to index the grants to grow with inflation. They would amount to the same $16.5 billion in 2007 as in 1997, meaning the program’s purchasing power was rigged to shrink with time.  

Second, block grants gave states virtually unchecked freedom to reshape welfare as they saw fit. Lawmakers were allowed to spend their federal funding on anything, so long as it was connected loosely to one of TANF’s four official goals: providing cash aid to the poor; promoting work, job training, and marriage; reducing out-of-wedlock pregnancies; and increasing two-parent families. These were extremely broad ambitions that left plenty of room for interpretation.

Welfare reform’s architects assumed that, in addition to cash benefits, states would devote much of the grant money to helping poor parents find employment by providing services such as job training and child care. Ultimately, though, the idea was to let states experiment with minimal interference from Washington and figure out what worked best for them. Unlike the federal waivers Clinton had granted, which required governors to produce a thorough evaluation of each policy reform—often using randomized control trials—states under TANF didn’t even have to prove they were getting results in moving families from welfare to work. They could simply use their money to help single mothers track down jobs, or not. They could keep sending checks to poor families, or not. It was up to them.

At the signing ceremony, Clinton stated that the bill “should represent not simply the ending of a system that too often hurts those it is supposed to help, but the beginning of a new era in which welfare will become what it was meant to be: a second chance, not a way of life.”

* * *

A decade in, the 1996 welfare reforms were considered enough of a triumph for Bill Clinton to take a victory lap on the New York Times op-ed page. TANF “has proved a great success,” the former president wrote in 2006, one that had created “a new beginning for millions of Americans.” Many on the left still disagreed, but James Traub seemed to sum up the mainstream attitude when he wrote in the New York Times Magazine, “In fact, welfare reform proved so successful at both shrinking caseloads and reducing poverty levels that the issue lost its urgency.” In the years since, conservatives have become even more enthusiastic about the bill. Republican House Speaker Paul Ryan, for instance, has called welfare reform a “crown jewel” of U.S. social policymaking.

The argument that reform succeeded typically goes something like this: In the years after it passed, the welfare rolls plummeted by about half. Single, never-married mothers went to work—their employment rate shot from 63 percent in 1996 to an unprecedented 76 percent, while their poverty rate tumbled by one-fifth. With its work requirements and time limits, this is exactly what TANF was intended to do.

Of course, much of this progress was due to the then-roaring economy. Clinton’s expanded EITC, which made low-paying jobs more lucrative, also lured women into the workforce. But economists have generally concluded that welfare reform played a positive role as well. In a survey of the research literature, Johns Hopkins University economist Robert Moffitt suggests that by 2001, reforms probably dropped the welfare caseload by 20 percent and increased employment among single mothers by about 4 percent. University of Wisconsin Chancellor Rebecca Blank and the University of Michigan’s Robert Schoeni find that the creation of TANF was associated with a roughly 2 percentage point fall in the poverty rate for less educated women.

But plenty of other families only suffered. The Urban Institute’s Pamela Loprest and Sheila Zedlewski found that during the early postreform era, about one-third of single parents were jobless soon after leaving welfare. Those who did find work often earned no more than what they lost in benefits; studies have concluded that anywhere from 42 to 74 percent of those who exited the program remained poor. Meanwhile, states began enrolling fewer new families in welfare. As the rolls shrank, a new generation of so-called disconnected mothers emerged: single parents who weren’t working, in school, or receiving welfare to support themselves or their children. According to Loprest, the number of these women rose from 800,000 in 1996 to 1.2 million in 2008.   

In keeping with that trend, researchers have also found a gradual uptick in what economists call deep or extreme poverty. Johns Hopkins’ Edin and Luke Shaefer, now of the University of Michigan, reported that the number of American households with children living on less than $2 in cash per person each day grew 159 percent, from about 636,000 in 1996 to 1.65 million in 2011. Even if you treat the value of food stamps as cash, the number rose some 80 percent, to 857,000. In their book $2.00 a Day, Edin and Shaefer describe women and children living on the fringes of society, relying on homeless shelters and selling their own plasma to get by. “Some of those people are ending up in very frightening conditions that don’t even look like America,” Edin tells me.

Some have questioned the data underlying Edin and Shaefer’s findings, in part because poor families have a habit of underreporting both their incomes and the government benefits they receive. But while social scientists can quibble about issues such as survey reliability, it’s undeniably disturbing that, over time, a rising number of families have started claiming to be utterly impoverished—it suggests that something has gone wrong.

It also comports with other research findings. Using a technique that accounts for both the value of government benefits and underreporting, the progressive Center on Budget and Policy Priorities found that the proportion of children living beneath half the U.S. poverty line rose from 2.1 percent in 1995 to 3 percent in 2005—an increase of about 700,000 kids. It concluded: “The largest single reason that the safety net protected fewer children against deep poverty was the loss of cash assistance following the 1996 welfare overhaul.”

* * *

Under TANF, the U.S. no longer has a single welfare system. Instead, it has 50 different ones, some far less helpful to the poor than others. But if you want a single snapshot that sums up welfare’s collapsed role in the safety net, you could look at the CBPP’s TANF-to-poverty ratio. In 1996, for every 100 poor families with children, 68 families received welfare cash benefits. By 2014, it had declined to just 23 for every 100, or about 1.6 million families total. In about a dozen states, mostly in the West and Deep South, fewer than 10 out of every 100 poor families received aid. In Georgia, the number was about 6 for every 100; in Texas, about 5 in 100; and in Louisiana, it’s about 4 in 100. This is all especially dispiriting when you realize that in most cases, parents who don’t get cash benefits also don’t get the job training or child care benefits that were supposed to make welfare a steppingstone to employment. “TANF is not providing cash or helping people get connected to work,” says Liz Schott, a senior fellow at the CBPP.

These minuscule enrollment rates aren’t just the result of tight eligibility rules. Even though families have to be living below roughly half the poverty line before they can get cash benefits in most states, there are still lots of qualified families across the country that never enroll. State governments have become adept at finding ways of discouraging the poor from signing up: Some of the moves, such as drug testing, make headlines; others are less attention-grabbing but perhaps even more onerous, such as requirements that force applicants to spend hours and hours on job searches before they start receiving benefits.

One classic example is Georgia, where the TANF rolls dropped by nearly half from 2006 to 2014. As Neil deMause recounted in Slate, the state’s welfare commissioner pushed local offices to deny benefits and prevent people from signing up by virtually “any and all means necessary.” They handed out flyers telling applicants “TANF isn’t good enough for any family” and “We believe welfare is not the best option for your family.” They made applicants “go through 60 job searches a week, or come to 8 orientations”—the sort of thing that would be all but impossible for, say, a high school dropout with children in a small town without a lot of employers. And all it took was a missed appointment here or a dropped deadline there for a poor family to lose its monthly support check. The commissioner even handed out Zero candy bars to state workers to drive home her message about whittling down the welfare rolls.

States have worked hard to box poor families out of welfare because TANF gives them strong incentives to do so. Under the program, states are required to make sure that half of their welfare recipients are either working or involved in some sort of “work-related activity” such as vocational training. If it fails, a state can theoretically lose federal funding. Of course, governors can hit these work targets by helping welfare families find jobs. But another easier way is simply to enroll fewer recipients in the first place—especially ones who might have trouble getting employed. Perversely, Washington has made this approach even more appealing by lowering states’ work targets as a reward for successfully reducing their caseloads. From a lawmaker’s perspective, kicking someone off welfare—or keeping him or her from ever signing up—is a sweet two-for-one deal.

Whatever states save by giving less cash to the poor, they can spend elsewhere. Thanks to its efforts to keep people off welfare, Georgia today only spends about 8 percent of its more than $500 million welfare budget on basic cash assistance. Where does the rest of the money go? A good chunk of it—at least $211 million—is spent not on supporting poor families and helping parents find work but on propping up the state’s underfunded child protection and adoption systems. In this, Georgia is not alone. In recent years, Arizona has spent more than 60 percent of its TANF budget on child welfare services (meaning things such as funds for child protective services and adoption assistance); Texas has devoted more than 50 percent.*

Which brings us to one of the most fundamental failures of TANF. As former Reagan administration official Peter Germanis has written, in much of the country, the program has degenerated into a “slush fund.” Early on, governors and legislators really did increase their spending on things such as work programs and child care to ease women off cash assistance. But with time, they’ve diverted the money to patch up their budgets. Today, the CBPP reports that about 19 percent of all TANF spending goes to a grab-bag category of “nonassistance”—more than double its share in 1997, and more than what’s now spent on programs such as job training (8 percent), child care (16 percent), or refundable tax credits (8 percent) and almost as much as basic cash aid (26 percent).

Much of that “nonassistance” nationwide is made up of child welfare services (meaning things such as funds for child protective services and adoption assistance), though not all of it. Michigan, for instance, has spent TANF dollars on college scholarships. Louisiana, which makes an unusually detailed breakdown of its TANF budget public, spends more on prekindergarten than cash assistance, while also dividing the money between domestic abuse prevention, child welfare, drug treatment, drug court, “abortion alternatives” such as crisis pregnancy centers, and other assorted programs.

Louisiana, Georgia, Arizona, and other states have been able to turn TANF into a slush fund for the simple reason that the law lets them. Anything that can be lightly tied to the program’s extremely broad ambitions—help the poor find work, save the two-parent family, etc—is considered an acceptable use of money. Meanwhile, a small amount of child welfare spending had been authorized under AFDC, so the category was considered fair game.

“The vision of TANF was that states would be spending the money on cash assistance and on work and work supports. And it’s turned out not to be the case” in some states, says Brookings Institution senior fellow Ron Haskins, who played a key role in drafting welfare reform as a Republican House aide and who is still generally a defender of the legislation. “We didn’t anticipate it.”

Others have been more unsparing about design failures that have motivated states to spend less helping the poor so they can spend more elsewhere. Germanis is now a civil servant in the Office of Family Assistance, the office that administers TANF as well as other federal aid programs. He writes that “when it comes to the TANF legislation, Congress got virtually every technical detail wrong.” TANF’s acronym, he says, should stand for “Truly a National Failure.”

Impoverished families can still rely on other safeguards against destitution. As welfare has withered, food stamps have become far more important—as of 2012, a three-person family in the median state could get more in benefits under the Supplemental Nutrition Assistance Program than it could under TANF. (In the 1980s and ’90s, welfare would have provided more.) Some women may be turning to the disability rolls for support in lieu of traditional welfare. The working poor, meanwhile, are in many ways better off than they were before the Clinton era, thanks largely to his expansion of the EITC. Though it often flies under the radar, the tax credit has become a far bigger source of support for poor and lower-middle-class Americans than welfare itself ever was—it sends around $70 billion to about 29 million working families annually, ultimately lifting at least 6.2 million people out of impoverishment.

Update, June 1, 2016: This article has been updated to clarify that Arizona’s TANF spending on child care services includes things such as funds for child protective services and adoption assistance. (Return.)

It’s clear that not everyone has suffered in the new world of welfare. But the biggest losers have tended to be society’s most vulnerable. Without cash, poor families can’t cover their basic needs; they can’t pay their phone bills, buy clothes for their children, or put gas in their car tanks. They can’t live normal, predictable lives that help them find and keep the jobs we’ve told them they should have. 

One of the most frustrating parts of welfare reform’s legacy is how easy it is to imagine the way things might have turned out better. Whether or not putting single mothers to work was a worthy policy goal, it wasn’t a goal that required turning the entire welfare program into an easily raided cookie jar. Arguably, it wasn’t even a goal that required any new legislation—Clinton could have continued letting states implement waivers to experiment with new policy reforms. That might have led to innovations that eased poor parents into the workforce while still ensuring that others could receive cash help as needed. Instead, he agreed to give states an evaporating pot of money and carte blanche to spend it how they pleased, setting up the program for slow self-destruction.

Unless something significant changes, we should only expect the welfare system’s dysfunctions to grow more severe. Inflation has already eaten about a third of the block grants’ value. Time will continue to consume them, and it will make it hard for high-cost states such as New York and California—which account for roughly half of all cash assistance cases in the country—to maintain benefits for the poor. TANF tore a hole in the safety net. That hole will just keep growing.