FROM ISSUE NUMBER 14 ~ WINTER 2013 GO TO TABLE OF CONTENTS

TEXT SIZE A A A

The Unfinished Work of Welfare Reform

ROBERT RECTOR and JENNIFER A. MARSHALL

Among the public-policy achievements of the past two decades, welfare reform may simultaneously be the best known and least understood. It is now remembered as a bipartisan triumph that ended "welfare as we know it," to use President Clinton's phrase, transforming the character of federal anti-poverty policy. The true history, however, is less august: The struggle to enact welfare reform was in fact bitter and arduous, and the policy changes that finally became law were but a first step toward a real transformation of the system.

The lip service that welfare reform today receives from all sides of our politics obscures how sharply the issue divided the left and right. New York Democratic senator Daniel Patrick Moynihan called the 1996 reform law a "brutal act of social policy" and a "disgrace" that would dog proponents "to their graves." Marian Wright Edelman, president of the Children's Defense Fund (a liberal advocacy group), said welfare reform would "leave a moral blot on [Bill Clinton's] presidency and on our nation that will never be forgotten." Her husband, poverty-law specialist Peter Edelman, resigned his post as an assistant secretary at the Department of Health and Human Services in protest of the law.

Conservatives heralded the change with an equal and opposite fervor. They believed that the reform would help restrain welfare spending, but they also felt the law served a crucial social purpose — that its work requirement offered a hand up rather than a handout to people in need, and thus held the key to reducing welfare-state dependency.

The immediate results clearly vindicated the conservative hypothesis about "workfare," as droves of former (and potential future) welfare dependents became productive employees in the private economy. So successful was the policy overhaul, in fact, that many conservatives concluded that their work on welfare was finished. But the reactions of both sides were overwrought: Liberals' dire predictions that millions more Americans would fall into poverty and that social dysfunction would increase proved mistaken; conservative workfare, meanwhile, has become the victim of its own success.

Because the reform law was a breakthrough, many saw it as a terminus rather than as the mere beginning of a long-term effort to transform the welfare state. And a massive, decades-long overhaul is what the welfare state desperately needs. For all the hype about the 1996 law, it dramatically reformed only one of nearly 80 federal means-tested programs providing aid to the poor. The reform replaced the largest cash-welfare program, Aid to Families with Dependent Children (AFDC), with a new program, Temporary Assistance for Needy Families (TANF), which included work requirements and a time limit on aid — all funded through block grants to states. But to put that achievement in perspective, in fiscal year 2011, federal spending on TANF was $17.1 billion — a mere 2.4% of the $717 billion the federal government spent on all means-tested welfare programs.

And while conservatives believed the '96 law would bring welfare spending under control, it in fact barely paused the overall growth of that spending. TANF spending has remained relatively flat since 1996, but overall means-tested spending has nearly doubled. By 2008, such means-tested welfare spending had risen to 5% of gross domestic product, up from an average of 4.4% in the 1990s when TANF was created. Since the beginning of the War on Poverty in the mid-1960s, means-tested spending as a share of GDP has increased, on average, between one-half a percentage point and a full percentage point per decade. All evidence indicates that this trend will continue, with welfare spending as a share of GDP hovering around 6% in the decade ahead. These numbers demonstrate that the 1996 reform was far too narrow to slow the overall growth of the welfare system.

To make matters worse, last summer, the Obama administration issued a policy directive from the Department of Health and Human Services dismantling the core of the '96 reform: the TANF work requirements that have sharply reduced dependence and increased employment. HHS unlawfully claimed the authority to waive these requirements; in their place, the department attached alternative conditions to welfare benefits that would be useless at best, and deeply harmful at worst, to the goal of helping lower-income Americans avoid dependency. That there was even a dispute over whether the waiver in fact gutted welfare reform's work requirements — as there was through much of the summer and during the height of the presidential campaign — reveals just how little journalists, politicians, and voters understand about how and why the '96 law worked.

By considering what was most essential to the '96 reform, as well as what the administration has done to undermine it, we can gain a better understanding of precisely why the Obama directive must be reversed. We can also better apply the lessons of success to determine how to carry on the work of welfare reform where it left off, and how to truly transform our welfare system so that it offers a path out of poverty and dependency.

HOW TO FIGHT POVERTY

America's welfare system has served neither the poor nor the taxpayer well. When Lyndon Johnson launched the War on Poverty in the mid-1960s, he intended it to strike "at the causes, not just the consequences of poverty." The aim of that effort, he explained, was "not only to relieve the symptom of poverty, but to cure it and, above all, to prevent it."

President Johnson's goal was not to create a massive system of ever-increasing welfare benefits for an ever-larger number of beneficiaries. Instead, he sought to increase self-sufficiency, enabling recipients to lift themselves up beyond the need for public assistance. "[M]aking taxpayers out of taxeaters" was Johnson's stated mission; "[w]e want to offer the forgotten fifth of our people opportunity and not doles," he declared.

Since Johnson uttered those words, the United States has spent nearly $20 trillion on means-tested welfare programs. And while the material living conditions of the poor have improved in that time, dependence on public assistance has only grown.

While America's poor do of course experience real material hardship, it is typically limited in severity and scope. The average poor person is far from affluent, but his lifestyle is hardly marked by the Dickensian privation suggested by advocacy groups and the media. Today, the typical poor American lives in a house or apartment that is in good repair and not overcrowded. In fact, the average home of a poor American is larger than the home of the average citizen of France, the United Kingdom, or Germany. Ninety-six percent of poor parents state that their children were never hungry at any time during the past year because the family could not afford food. This does not mean that the poor are living high, but they are living significantly better than past generations of poor Americans.

Still, many Americans are very poor in another sense. Material poverty has been replaced by a far deeper "behavioral poverty" — a vicious cycle of unwed childbearing, social dysfunction, and welfare dependency in poor communities. Even as the welfare state has improved the material comfort of low-income Americans by transferring enormous financial resources to them, it has exacerbated these behavioral problems. The result has been the disintegration of the work ethic, family structure, and social fabric of large segments of the American population, which has in turn created a new dependency class.

That dependency is measurable. According the U.S. Census Bureau, some 46 million Americans, including 16 million children, were "poor" in 2010. But it is important to note what this figure represents. To measure poverty, the Census Bureau looks at the number of Americans living below certain income thresholds. In evaluating income, however, the bureau omits more than $800 billion in means-tested government cash, food, housing, and medical benefits; it takes into account only what Americans earn on their own, without government assistance. The Census Bureau figure is therefore not a very accurate measure of "poverty" — since, as discussed above, many Americans enjoy substantially higher levels of material well-being than their incomes would suggest, precisely because of these government benefits. It is, however, a good measure of the number of Americans who are self-sufficient.

And what does this measure tell us? That there has been essentially no improvement in self-sufficiency since the War on Poverty began more than four decades ago. In 1966, the share of the population living below the poverty (self-sufficiency) threshold was 14.7%; by 2011, it had actually risen — to 15.0%. While the material living conditions of less affluent Americans may have improved over time, when it comes to President Johnson's original goal — reducing the "causes" rather than the mere "consequences" of poverty — our welfare policies have failed. Many parts of the population are in fact now less capable of self-sufficiency than they were when the War on Poverty began.

Among families with children, one major driver of this dependence is the collapse of marriage. In a staggering 71% of all officially poor families with children, the parents are unmarried. Indeed, the rate of poverty among single-parent and married-parent families presents a stark contrast: More than a third of single-parent families with children are poor, compared to only 7% of families with married parents. Overall, the children of married parents are 82% less likely to be poor than are the children of single mothers. This relationship between marriage and poverty persists across lines of race and educational attainment.

Data from the Fragile Families and Child Wellbeing Survey (conducted jointly by Princeton and Columbia) show that, if unmarried mothers remain single, more than half — 56% — will be poor. By contrast, if single mothers married the biological fathers of their children, only 17% would remain poor. These results are based on the actual earnings of the biological fathers of the children in the representative survey population, not on assumed or hypothetical earnings. Marriage would thus be likely to reduce the expected poverty rate of the children of non-married mothers by two-thirds.

A second major cause of child poverty is a lack of parental work. Even in good economic times, parents in the average poor family with children perform a total of only 800 hours of work per year — the equivalent of one adult working 16 hours per week. If this amount were increased to the equivalent of one adult working full time throughout the year, the poverty rate among these families would drop by two-thirds.

But while work and marriage offer reliable paths out of poverty and toward self-sufficiency, the welfare system perversely remains hostile to both. Major programs such as food stamps, public housing, and Medicaid continue to reward idleness and penalize marriage. If welfare could be turned around to encourage work and marriage, persistent poverty would quickly decline. This was precisely the insight behind welfare reform.

HOW AND WHY REFORM HAS WORKED

In the four decades preceding the 1996 welfare reform, the number of Americans on welfare had never significantly decreased. By 1995, nearly one in seven children was on AFDC. The typical AFDC beneficiary was on welfare for an estimated average of 13 years. And such welfare dependency gets passed from one generation to the next: For example, in a 2002 study, researcher Marianne Page of the University of California, Davis, found that if a woman's family had received welfare during that woman's childhood — be it AFDC, General Assistance, food stamps, or Supplemental Security Income — that woman was roughly three times as likely to receive welfare as an adult. The '96 reform sought to disrupt this cycle of inter-generational dependence by moving families with children off the welfare rolls through increased work and marriage.

The success of that reform tells an extraordinary story. Within five years of the enactment of TANF, caseloads dropped by approximately 50%. As caseloads plummeted, employment and earnings among low-income single parents surged upward. Employment of never-married mothers increased by 50%, employment of single mothers with less than a high-school education increased by two-thirds, and employment of young single mothers between the ages of 18 and 24 approximately doubled. Child poverty among the demographic groups most affected by the new policy also fell dramatically. For a quarter-century before the reform, poverty among single mothers and among black children had remained stubbornly high. Immediately after the reform, poverty among both groups experienced a dramatic and unprecedented decline, reaching all-time lows.

Against conventional wisdom, the effects of welfare reform have been the greatest among the most disadvantaged single parents facing the greatest barriers to self-sufficiency. Both decreases in dependency and increases in employment were largest among those most inclined to long-term dependence — that is, among younger, never-married mothers with little education.

How did this happen? The welfare reform enacted in 1996 had three main elements, of varying degrees of effectiveness. First, the reform placed a five-year time limit on the receipt of TANF benefits. But the time-limit policy contained large loopholes; in the 15 years since welfare reform's enactment, only about 2% of TANF recipients have been removed for reaching the five-year limit. Overall, the requirement has had little or no impact on caseload reduction.

The second main component of the reform, which changed the funding structure of cash welfare, had a greater effect. The old AFDC program was an open-ended entitlement: If caseloads went up, state governments received more federal funds. The new TANF program, by contrast, used a fixed, block-granted funding system. If caseloads increased, state governments were forced to bear the extra cost. If caseloads fell, however, state governments had to receive their fixed amounts of federal funding and could use the surplus for other state projects. This policy created enormous financial incentives for states to reduce lifelong welfare dependency among their populations.

But it was the third element — mandatory work requirements — that had the greatest impact. The 1996 law required that roughly a third of people on the TANF rolls in each state work or prepare for work as a condition of receiving aid. The TANF work requirements were the real motor behind welfare reform — the main reason that caseloads fell rapidly during the five years following the law's enactment.

These core "workfare" requirements were spelled out in Section 407 of the 1996 law. Section 407 established a three-part mandatory work system, defining who was required to work, how much they were required to work, and what constituted qualifying work-related activities. Under Section 407, around 30% to 40% of the "work eligible" adult TANF case-load (defined as non-disabled adults receiving TANF benefits) is required to engage in "work activities." (The law nominally requires that 50% of this caseload be engaged, but provides a variety of exemptions and deductions that substantially reduce the effective participation rates.)

"Work activities" are defined very broadly to include unsubsidized employment, government-subsidized employment, on-the-job training, up to 12 months of vocational education, community-service work, job searches (for up to six weeks) and job-readiness training, high-school or GED education for recipients under age 20, and high-school or GED education for those 20 or over if combined with other listed work activities. Eligible individuals are required to engage in such activities for 20 hours per week if they have children under age six in the home and for 30 hours per week if they do not. Subject to the overall participation rate, states are free to determine which recipients will be required to engage in work activities.

This TANF workfare system is simple and flexible, allowing states a wide range of choices in meeting their participation standards. And the "workfare" approach also yields a number of important benefits for welfare recipients and society at large. The public overwhelmingly believes that able-bodied adult welfare recipients should be required to work, prepare for work, or at least seriously look for work as a condition of receiving aid. In requiring some TANF recipients to engage in constructive activity in exchange for benefits, workfare thus transforms welfare from a one-way handout into a system of reciprocal obligation. Moreover, demanding work as a condition of receiving benefits reduces the relative economic utility and attractiveness of remaining idle on welfare: If a recipient has to work anyway, he might as well hold a job that likely provides more compensation than welfare and allows for self-sufficiency. These improved incentives have resulted in fewer welfare enrollments, shorter spells of welfare dependency, and smaller caseloads across the country.

The implementation of workfare had the added effect of helping people avoid slipping into welfare dependency in the first place. By definition, an able-bodied applicant for welfare claims that he cannot find employment and therefore needs aid from the taxpayers. In many cases these claims are true, but it is also true that large numbers of people will take free handouts if the government is offering them, regardless of whether they really need them. A work test applied at the point of entry into the welfare system helps separate these two groups. If workfare requires recipients to begin serious efforts toward achieving self-reliance at the outset, many people who do not really need the aid will simply choose not to join the welfare rolls. The sharp drop in the TANF welfare caseload after 1996 was caused as much by a decline in new enrollments as it was by the increase in departures from welfare.

By deterring unnecessary entries into the welfare system, and by steering recipients toward paying jobs to help them get out of that system, workfare has also increased long-term earning potential among the affected population. Time spent on welfare never looks good on a résumé, and welfare dependence erodes important work habits and job skills and reduces the sorts of contacts with other employed people that can lead to future job opportunities. To the extent that welfare enrollment can be avoided, then, it boosts earning prospects. And for those already in the system, workfare programs provide skills training, job-readiness preparation, and employment-search services that help connect recipients to jobs, speeding the transition from welfare to work.

By reducing unnecessary welfare enrollments and shortening the time spent on welfare, the '96 workfare requirement significantly reduced caseloads, generating substantial savings for taxpayers. It also offered those who did require government assistance the support and training they needed to achieve self-sufficiency — and to become productive contributors to the broader economy. This was the heart and soul of welfare reform, and this is why it has been deemed a success.

OBAMA UNDOES WORKFARE

Among the left, however, resistance to workfare has a long history. Liberals opposing work requirements blocked welfare-reform efforts during the Nixon administration. Throughout the 1980s, congressional Democrats obstructed President Ronald Reagan's efforts to require AFDC recipients to work. Bill Clinton may have campaigned on requiring work in welfare, but his actual legislative proposal offered only token work requirements: Introduced in the summer of 1994, Clinton's "Work and Responsibility Act" contained no time limits and required only 7% of the adult AFDC caseload to participate in a work program.

In 1996, a Republican Congress crafted a new welfare-reform law with workfare as its centerpiece. Half of the Democrats in the House and Senate voted against it. The liberal opposition in Washington was so strong that Clinton vetoed welfare reform twice before finally signing it. But workfare was extremely popular, with more than 80% of the public supporting it in polls. Warned by his political advisor Dick Morris that opposition to welfare reform could cost him re-election, Clinton eventually bowed to pressure and signed the bill. Even then, however, he could not convince many within his own administration to get on board. Most of his cabinet opposed the bill.

When TANF faced reauthorization in 2002, liberals aggressively (though ultimately unsuccessfully) sought to repeal the federal work standards. A few years later, during the next reauthorization battle, freshly elected senator Barack Obama opposed efforts to renew the TANF program, denouncing the law's work requirements on the Senate floor. When Obama was elected president, liberal opponents of workfare judged that they finally had the power they needed to undermine TANF's work requirements — even if they did not have the legal authority to do it without an act of Congress.

On July 12, 2012, the Obama administration's Department of Health and Human Services issued a policy directive rewriting the successful welfare-reform law of 1996. The HHS directive strikes directly at Section 407, allowing states to waive the TANF work requirement — thereby gutting the program of its most critical reform element and savaging both the letter and the intent of the '96 law. This move was not just counterproductive, but illegal. In establishing welfare reform, Congress had made the core work requirements of the TANF program mandatory and non-waiveable. It had explicitly protected those requirements from any future administration that might seek to weaken them.

In attempting to explain its unlawful changes, the Obama administration claimed the authority to waive the TANF work requirements through a legal device called Section 1115 waiver authority. Section 1115 of the Social Security Act states that "the [HHS] Secretary may waive compliance with any of the requirements" of specified parts of various laws; the purpose was largely to allow states the flexibility to launch test projects and other policy experiments. With this in mind, laws enacted through the Social Security Act since 1935 — including the '96 welfare reform — specifically enumerate which parts of the statute may be waived by the HHS secretary using Section 1115 authority. Indeed, in order for a provision of a law to be waived under Section 1115 authority, that provision must be listed in Section 1115 of the Social Security Act itself. The idea is to avoid granting open-ended authority to the HHS secretary to waive important provisions of welfare and health-care laws at will.

As noted above, the work provisions of the TANF program are contained in Section 407 of the welfare-reform law. And Section 407 of the welfare-reform law was deliberately not listed in Section 1115 of the Social Security Act. Section 407 was titled "Mandatory Work Requirements" for a reason; those requirements cannot legally be waived. The only section of the TANF law that can be waived is Section 402, which describes "state plans": reports that state governments must file to HHS describing the actions they will undertake to comply with the many requirements established in the other sections of the TANF law. The authority to waive Section 402 provides the option to waive state reporting requirements only, not to overturn the core requirements of the TANF program contained in the other sections of the '96 law. The Obama administration asserts that because the work requirements established in Section 407 are mentioned as an item that state governments must report about in Section 402, all the work requirements can be waived.

Contrary to press reports, the Obama administration did not merely "tweak" the welfare-reform law's workfare architecture. Nor were the changes merely temporary responses to the recession. They constitute a permanent change in the very character of the TANF program. HHS asserted the authority to "waive compliance" with every provision of Section 407 — which means all of the work rules in the TANF law are now subject to the secretary's whim.

In its guidance memorandum and related documents, HHS outlined the types of changes it was seeking in the TANF program under this newly (and lawlessly) asserted authority. The department stated that it would significantly lower the already lenient required work-participation rates in TANF; broaden the definition of "work activities"; replace the requirement that recipients engage in work activities for 20 to 30 hours per week with looser standards, perhaps as little as one hour per week; and in fact replace the TANF work-participation requirements entirely with alternative standards based on "employment exits." At the same time, HHS promised to reduce the verification procedures that monitor state workfare programs — allowing states to run pseudo work programs that exploit newly loosened and vague federal rules, without public scrutiny. All of these changes are likely to dramatically increase the number of TANF recipients who receive welfare checks without working.

The guidance memorandum further asserted that HHS will use waivers to promote state policies that use a "comprehensive universal engagement system in lieu of certain participation rate requirements" (emphasis added). The term "universal engagement" generally means a policy of seeking to have all adult TANF recipients engage in "federally countable activities" for at least one hour per week. As Jacqueline Kauff and Michelle Derr noted in a 2008 study of TANF for the research firm Mathematica, such "federally countable activities" are defined very broadly — including pursuits such as visiting a doctor or looking for day care. "Universal engagement" can be a helpful policy if used in conjunction with existing work standards. What HHS proposes, however, is not to add "universal engagement" to existing work requirements, but rather to replace those requirements with activities that have nothing to do with paid work.

In sum, the legal requirement that 30% to 40% of each state's TANF caseload engage in clearly defined work-related activities for 20 to 30 hours per week will be replaced by a new standard under which the entire caseload can engage in vaguely defined activities for one hour per week. This gambit follows the habitual pattern of liberal opponents of welfare reform: using bold-sounding but vague language (like "universal engagement") to camouflage efforts to weaken workfare.

HHS has clearly explained that its aim is to develop policies and welfare rules "other than those set forth in Section 407," replacing the successful '96 requirements with alternative performance standards that are, at best, highly problematic. In a letter from HHS secretary Kathleen Sebelius to Utah senator Orrin Hatch — who had written the secretary to complain that the department's policy-guidance document would undermine workfare — the secretary stated that the department would rescind a state's exemption from federal work-participation standards if the state did not "demonstrate clear progress" toward the goal of "mov[ing] at least 20% more people from welfare to work compared to the state's past performance" within a year of receiving the exemption.

At first glance, this requirement sounds laudable. But a closer look shows that HHS would not in fact demand proof that a state had made progress toward a 20% increase in meeting work requirements. Rather, the state would have to demonstrate only a 20% increase in the number of so-called "employment exits" from its welfare program — that is, people who report leaving the program as a result of having found work. The problem is that "employment exits" are a weak and counterproductive measure of success in welfare reform, for four major reasons.

First, in the weak economy of the past few years, the number of employment exits from TANF has dropped by about one-fourth. Historical data show that the number of exits will almost certainly rebound on its own by a similar amount as the economy revives and hiring picks up. Virtually every state in the nation should therefore experience a 20% increase in employment exits over the coming years compared with its past performance. This means that every state would be entitled to an exemption from TANF's work requirements.

Second, about 1.5% of the monthly TANF caseload leaves the program because of increased employment each month, but an even larger portion of the caseload leaves the program for unknown or unspecified reasons. To be exempt from the TANF work requirements under the new HHS rules, the average state would need to raise its monthly employment exits from 1.5% to 1.8% of the caseload (since that would amount to a 20% increase). Many states could meet this target simply by tightening data collection, labeling some of these unknown or unspecified exits as "employment exits" instead. States could thus be exempted from TANF work requirements without experiencing any increase in the number of people leaving the system because they have actually found jobs.

Third, a 20% increase in employment exits is actually a very small change in terms of the number of individuals affected. Each month, the average state has a TANF caseload of around 40,000 families and sees around 600 employment exits from the program (the 1.5% figure noted above). Under the Obama administration's new welfare rules, that state can be fully exempted from the work standards in the '96 TANF law if it raises its employment exits from 600 per month to 720. A policy that exempts 39,000 welfare households from workfare participation just because an extra 120 have left the rolls clearly misses the point of welfare reform.

Finally, and crucially, an increase in employment exits is almost always a reverse indicator of reduced welfare dependence. Increases in employment exits are positively correlated with increases in caseloads and negatively correlated with caseload reduction. In other words, the number of employment exits usually rises as the welfare caseload rises, and it falls when the caseload falls. Why? Because welfare caseloads always have a certain amount of regular turnover. The larger the caseload, the greater the number of exits — simply because there are more people in the system to begin with.

Precisely because employment exits are linked to rising dependency, Congress deliberately excluded them as a performance measure when crafting the 1996 welfare-reform law. Indeed, by the employment-exit yardstick, the old AFDC program was a stunning success: Employment exits nearly doubled in the decade before the '96 reform as caseloads increased by a third. This same flawed measure would identify TANF as a dismal failure: Because total caseloads have fallen dramatically, employment exits have, too. Obviously, employment exits are a deeply problematic measure — and yet the Obama administration now seeks to make them the central performance standard of a radically revised TANF program. The Obama administration is not merely gutting welfare reform — it is perverting it.

BUILDING ON WELFARE REFORM

There is simply no question that the administration's lawless overturning of TANF's work requirements must be undone. Every avenue to this end must be pursued, since the effectiveness of the TANF program depends upon it. Rather than dismantling workfare, policymakers seeking to improve our welfare system — and to learn from both the failures and successes of the past several decades — must look for opportunities to expand its reach.

Of the nearly 80 federal means-tested aid programs, only AFDC was significantly reformed in the '90s; none has been touched since. And, as noted above, even the reform that replaced AFDC with TANF was weaker and more limited than is generally understood. Half of able-bodied TANF recipients are typically idle on the rolls, collecting a check without performing any work activity. And the leniency and limited scope of the 1996 welfare reform has meant that total means-tested spending and overall dependence on government have continued to rise steadily in the post-reform period. Even worse, marriage in low-income communities has continued to erode, boosting dependence, poverty, and other social ills. Welfare reform should thus be extended by pursuing three goals: extending workfare, controlling welfare spending, and strengthening marriage.

First and foremost, work requirements should be established throughout the welfare system. There is no reason why only TANF should have a work requirement while other major welfare programs are treated like old-style entitlements. The food-stamp program and federal housing programs, for instance, could easily be reformed along the same lines by setting work as a condition of receiving benefits.

Workfare is particularly needed in the food-stamp program, the second most costly means-tested aid program. Food-stamp spending has exploded in recent years, from $19.8 billion in 2000 to $84.6 billion in 2011. Part of that growth was caused by the recession, but there is no evidence that spending discipline will return as the economy recovers: Under President Obama's most recent budget proposal, food-stamp spending will remain well above historic norms for the foreseeable future. The food-stamp program today discourages work, rewards idleness, and promotes long-term dependency in the way that AFDC once did; able-bodied food-stamp recipients should thus be required to work, prepare for work, or at least look for a job as a condition of receiving taxpayer aid. These work-activation requirements should be phased in gradually as the economy improves.

Second, the future growth of means-tested welfare spending should be limited. In fiscal year 2011, total federal and state spending on the roughly 80 means-tested federal welfare programs reached $927 billion. More than 100 million people, or a third of the U.S. population, received aid from at least one of these programs (Social Security and Medicare are not included in these figures). Average benefits amounted to roughly $9,000 per recipient. If converted to cash, means-tested welfare spending would be more than sufficient to bring the income of every American in the least affluent third of the population up to 200% of the federal poverty level — roughly $44,000 per year for a family of four. Since the beginning of the War on Poverty in the mid-1960s, government has spent $19.8 trillion (in inflation-adjusted 2011 dollars) on means-tested welfare. By comparison, the combined cost of all the wars in American history — from the Revolutionary War through the current war in Afghanistan — has been $6.98 trillion (in 2011 dollars). The War on Poverty has thus cost three times as much as all of our real wars combined.

These numbers are staggering, and without serious reform, they will only continue to grow. The president is certainly comfortable with rising welfare spending: After adjusting for inflation, Obama's increase in federal means-tested welfare spending during his first two years in office was two and a half times greater than any previous increase in federal welfare spending in American history. In fiscal year 2007, total government spending on means-tested welfare or aid to the poor was a record-high $657 billion. By fiscal year 2011, total government spending on means-tested aid had, as noted above, risen to $927 billion — a 40% increase. And Obama's proposed budgets would have total annual means-tested welfare spending remain permanently at its present bloated level of 6% of GDP over the next decade. Combined annual federal and state welfare spending will reach $1.56 trillion in 2022; overall, President Obama plans to spend $12.7 trillion on means-tested welfare in the next ten years.

To get control of this runaway spending, Congress should establish a cap on future aggregate welfare expenditures. It is critical that Americans understand the full price tag of the welfare state, and that policymakers treat some 80 means-tested programs as one system in need of massive reform. Liberals dominate the debate over welfare spending when each program is discussed in isolation without reference to the entire welfare state's size and scope — a trick that allows for vast understatement of the broader problem.

When the economy rebounds (or by 2017 at the latest), total spending on all federal means-tested welfare programs should be returned to the pre-recession levels of 2007, adjusted for inflation. Thereafter, aggregate annual federal welfare spending should grow no faster than inflation. Such a spending cap — which was the centerpiece of the Welfare Reform Act of 2011, introduced by Congressman Jim Jordan of Ohio and then-senator Jim DeMint of South Carolina — would save the taxpayers $2.7 trillion during its first decade.

Finally, welfare reformers must seek ways to strengthen marriage among lower-income Americans. This is the most difficult of the goals that must guide the continuing work of welfare reform, but it is also the most important. Married parents are easily the most effective insurance against child poverty, but increasing the rates of married childbearing has never seriously been considered as a tactic in the War on Poverty.

The connection between unwed childbearing on one hand and poverty and welfare dependence on the other is dramatic. A child born and raised outside of marriage is six times more likely to experience poverty than one who is born and raised by married parents in an intact family. Roughly three-quarters of welfare assistance to low-income families with children goes to single-parent families. The medical expenses associated with most non-marital births are currently paid for by taxpayers through the Medicaid system, and a wide variety of welfare assistance will continue to be given to both the mother and the child for the first two decades of the child's life.

In 2008, 1.72 million children were born outside of marriage in the United States. Most of these births were to mothers in the circumstances most poorly suited to raising a child alone: young women with educational attainment of a high-school diploma or less. More than two-thirds of births to women who were high-school dropouts occurred outside of marriage. Among women who had only a high-school diploma, slightly more than half of all births were out of wedlock. By contrast, among women with at least a college degree, only 8% of births were to unwed mothers. Overall, the unwed birthrate is above 40%; the Hispanic unwed birthrate is more than 50%; for blacks, it is higher than 70%.

Although most non-married parents aspire to remain together and to eventually marry, they generally lack the skills needed to build enduring relationships. Nor have they seen many models of healthy marriage in their own lives or heard about the socioeconomic significance of marriage. Tragically, on the issue of non-marital childbearing, our society's silence has been deafening for almost half a century. Many low-income young people have never been told that marriage is beneficial; they have never been told that having children outside of marriage is likely to have harmful consequences for their futures.

To combat poverty and dependence, it is vital to strengthen marriage. There is no easy public-policy path toward this goal, only modest steps to make a modest difference. But one plausible proposal would be to reduce the penalties against marriage in the welfare system. Currently, because benefits are reduced as a family's income rises, our welfare programs create disincentives to marriage. A mother will receive far more from welfare if she is single than if she has an employed husband in the home. For many low-income couples, marriage means a reduction in government assistance and an overall decline in the couple's joint income. The welfare system should be overhauled to reduce such counterproductive incentives.

A public-information campaign about the benefits of marriage could also help. To strengthen marriage, it is vital that at-risk populations gain a clear, factual understanding of the advantages marriage confers and the costs and consequences of non-marital childbearing. To develop this understanding, government and key institutions of civil society should establish a broad education campaign in low-income areas. This campaign should be similar to current efforts to convince young Americans of the importance of staying in school or to inform the public about the health risks of smoking. While the costs of such an effort would be small, its impact could be considerable.

THE UNFINISHED WORK

The '96 reform was only a first step toward reducing the needless dependency fostered by the structure of our welfare state. But it was a successful step, and the gains it has produced should be consolidated, not undone.

In the coming years, reformers have their work cut out for them. They must begin by resolutely opposing the Obama administration's efforts to dismantle the work requirements in TANF, which are the heart of the program and the key to the effectiveness of the 1996 reform. The administration's changes to the program are not only counterproductive, but unlawful. They must not be allowed to stand.

And to build on the successes of welfare reform and regain the initiative, reformers should pursue a three-part agenda. First, workfare must be expanded to other means-tested programs like food stamps and public housing. Second, the overall costs of all our ballooning welfare programs must be controlled through a cap on aggregate spending. Third, serious attention must be paid to the crisis of unwed parenting that drives so much of today's poverty.

The 1996 welfare reform was enacted even under a Democratic president because the public was clearly behind its basic goals. It was clear that America's welfare system was terribly dysfunctional and that addressing poverty and dependency would require making welfare beneficiaries work. These same truths are just as evident today; we are just as much in need of a robust reform effort with the strong support of the public. It is time to turn ending "welfare as we know it" from an old slogan into a reality.

Robert Rector is a senior research fellow in domestic policy at the Heritage Foundation. Jennifer A. Marshall is director of the DeVos Center for Religion and Civil Society at the Heritage Foundation.