The United States’ means-tested welfare system consists of over 80 programs that provide cash, food, housing, medical care, and social services to poor and lower-income Americans. Total annual spending on these programs reached $1 trillion in 2015.[1] More than 75 percent of this funding comes from the federal government.

The last substantial reform of welfare, enacted in 1996, transformed the Aid to Families with Dependent Children (AFDC) program into the Temporary Assistance for Needy Families (TANF) program. Mandatory federal work requirements for recipients were at the heart of the change, which led to significant decreases in the program’s rolls, increased work among former recipients, and historic reductions in child poverty.

Re-starting Welfare Reform

Although the welfare reform of the 1990s was popular and initially successful, it was actually quite limited. Of 80 welfare programs, only TANF was reformed, and even in TANF, the vigor of reform has nearly disappeared. Welfare reform should be rejuvenated and expanded by making the following changes.

1. Set the proper goals for welfare. The goal of welfare should not be to reduce poverty through an ever-larger welfare state. Rather, the goals should be to increase self-sufficiency (having an income above poverty level without relying on government welfare aid); enhance productive participation in society; and improve personal well-being and upward mobility.

On the surface, the simplest goal of welfare is to reduce poverty by raising incomes; however, this goal can be pursued by two contradictory approaches. On one hand, government can simply guarantee welfare benefits at higher and higher levels. On the other hand, government could promote self-sufficiency: the ability of individuals to support their families above the poverty level through their own efforts without the need for welfare aid. The principal mechanisms of prosperous self-sufficiency are work, earnings, and marriage. A conservative approach to welfare would seek to strengthen, not undermine, these mechanisms.

In launching the War on Poverty, President Lyndon Johnson sought to pursue this second course. Johnson promised that his war would reduce welfare dependence. He aimed not merely to reduce the “symptoms” of poverty through more welfare, but also—and more important—to end the “causes” of poverty by changing abilities and behaviors so that welfare became unnecessary.[2]

Welfare dependence pushes individuals to the margins of society; it impedes the upward mobility of children. Welfare also undermines personal psychological well-being. Well-being includes happiness, self-respect, competence, accomplishment, a sense your life is meaningful and valued by others, and strong interpersonal bonds.[3] A decrease in the capacity for self-support and an increased dependence on welfare would tend to significantly diminish well-being.

In particular, a healthy marriage is one of the two most important factors contributing to personal happiness.[4] Marriage is also a very strong factor in promoting the upward mobility of children.[5] But since the beginning of the War on Poverty, marriage has been nearly wiped out in many low-income communities.[6] Replacing husbands with welfare checks has degraded personal well-being for men, women, and children in low-income neighborhoods.

2. Clearly enumerate the total cost of the means-tested welfare in the Congressional Budget Resolution and set annual spending caps on total welfare spending. As noted, government spent over $1 trillion last year on means-tested welfare programs. (Social Security and Medicare are not included in this figure.) Means-tested cash, food, and housing spending alone totals nearly twice the amount needed to eliminate all poverty in the U.S.[7]

But the size of the welfare state is largely hidden; the real aggregate cost of welfare programs is almost never presented or discussed. Instead, welfare is debated one program at a time. This piecemeal approach distorts perceptions and enables the left to pretend that the welfare state is tiny and that the solution to problems is simply to spend more money. By contrast, confronted with actual total spending on the poor, the left has great difficulty explaining where all the money goes.

Rational debate about the welfare state must begin by highlighting the vast sums taxpayers already devote to the poor. To accomplish this, the annual Congressional Budget Resolution should clearly state the aggregate total spending on all means-tested programs (including state spending on federal programs). It should also set forth aggregate spending projections for the next 10 years.

Discussing reductions in individual welfare programs without providing the context of overall welfare spending is inherently misleading. Any proposed reduction in future means-tested spending should be presented initially as a percentage reduction in total welfare spending. For example, it is more appropriate to discuss a 1 percent reduction in $1 trillion of spending than it is to discuss the same change as a $10 billion reduction in one or two programs seen in isolation.

3. Measure poverty, income, living standards, and inequality correctly. Developing sound welfare policy requires accurate information. Regrettably, current government surveys do not and cannot measure incomes, poverty, and receipt of welfare correctly. This cripples efforts to devise rational policies.

Each year, the Census Bureau claims that nearly 50 million Americans live in poverty, but “poverty” as measured by the government has little relationship to the ordinary sense of the word. The actual living standards of the poor differ greatly from conventional perceptions. The government’s own data show that the typical poor family in the U.S. has air conditioning, a car, and cable or satellite TV. As of 2009, half of the poor have computers; 43 percent have Internet, and 40 percent have a wide-screen plasma or LCD TV. The U.S. Department of Agriculture reports that only 4 percent of poor children were hungry for even a single day in the prior year due to lack of funds for food. Only 6 percent of poor households are overcrowded. The average poor American has more living space than the average non-poor individual living in Sweden, France, Germany, or the United Kingdom.[8]

The problem lies in how the government defines and measures income. The government defines a family as poor if its income falls below specified income thresholds. For example, the poverty income threshold for a family of four is $24,036.[9] But out of the $1 trillion in government welfare spending, only about 5 percent is counted as income for purposes of measuring poverty or inequality. According to the government, entire programs such as food stamps, the earned income tax credit (EITC), and public housing are not “income” and have no impact on poverty. Given this bizarre measurement, it should be no surprise that government reports show that poor people spend $2.40 for every $1.00 of income the Census claims they have.[10]

Moreover, even when government surveys do attempt to report welfare benefits, they severely undercount them. People responding to government surveys simply do not report all the welfare benefits they receive. To obtain accurate information, respondent reports of welfare use or non-use in annual Census surveys should be corroborated at the individual level by administrative records from programs such as food stamps, public housing, the EITC, Medicaid, and the State Children’s Health Insurance Program (SCHIP).

However, even corroborated income surveys would still underreport incomes because they would miss substantial off-the-books earnings on the part of low-income persons. This means that alternative surveys of consumption and material conditions give the best picture of actual living conditions at the low end of the income scale. Such surveys should be given greater prominence.

Finally, the Obama Administration has created an extremely misleading new poverty measure in which future poverty income thresholds will rise in direct proportion to any increase in median income in society as a whole. Under this measure, poverty is no longer linked to any material conditions; moreover, poverty can be reduced only if the incomes of the “poor” are rising faster than the incomes of everyone else.[11] Funding for the Obama poverty measure should be eliminated.

4. Require able-bodied non-elderly adults receiving welfare benefits to work, prepare for work, or at least look for a job under supervision as a condition for receiving aid. Welfare programs should assist those in need, but welfare should not be a one-way handout. Work requirements in welfare promote greater self-support and establish a reciprocal obligation between the beneficiary and the taxpayers who fund the benefits. Work requirements should be introduced or greatly strengthened for three groups: able-bodied, non-elderly adults without dependents receiving food stamps; work-capable TANF recipients; and able-bodied parents receiving food stamps.

Work Requirement for Able-Bodied Adults Without Dependents (ABAWDs) Receiving Food Stamps. The food stamp program is the second largest means-tested welfare program. In 2014, government spent $83.1 billion on the program.[12] In recent years, the most rapidly growing group of food stamp recipients has been able-bodied adults without dependents. ABAWDs are adults between the ages of 18 and 49 who are not disabled and who have no children to support. In 2014, nearly five million ABAWDs received food stamps each month;[ 13] few are employed.[14]

ABAWDs who receive food stamps should be required to work, prepare for work, or look for work in exchange for receiving benefits. In FY 2014, Maine implemented a work requirement for ABAWDs. After the implementation of the work requirement, Maine’s ABAWD caseload dropped substantially, by 80 percent within just a few months. If a federal work requirement for ABAWDs were enacted and achieved the same level of success as was achieved in Maine, the reform could save taxpayers up to $9.7 billion annually.[15]

Strengthen Work Requirements in the Temporary Assistance for Needy Families Program. The TANF program was created by the welfare reform law of 1996. A key premise of that law was that work-capable TANF recipients should be required to work or prepare for work. However, the current work programs in TANF are quite lax. In the average state, more than half of work-capable recipients are completely idle. Typically, state welfare bureaucracies engage less than a fifth of recipients in activities intended to increase employment and reduce dependence.[16] The federal TANF work rules should be greatly strengthened to require that two-thirds of non-employed TANF recipients engage in training, perform community service, or at least search for a job under supervision in exchange for their benefits.

Work Requirements for Able-Bodied Parents on Food Stamps. In an average month in 2014, there were 8 million to 9 million able-bodied parents receiving food stamp benefits; around half of these were not employed.[17] In general, families with an employed parent will have incomes substantially higher than those who are fully dependent on welfare. Moreover, a family with a parent who works full-time (even at low wages) will typically have a combined income from earnings and welfare that is well above poverty. Welfare programs should therefore seek to promote parental employment rather than long-term dependence. Able-bodied, non-elderly parents who have received food stamp benefits for over three months and who are not currently employed should be required to undertake training, community service, or at least supervised job search in exchange for benefits.

5. Promote healthy marriage in low-income communities. Marriage is the greatest protector against child poverty. Children born to a married mother and father are about 80 percent less likely to be poor compared to children in single-parent homes. Yet more than 40 percent of children are born outside of marriage in the United States annually.[18]

Ironically, nearly all means-tested welfare programs impose significant penalties against marriage. If low-income fathers and mothers marry, in most cases, their welfare benefits will be cut and their combined income will fall. The overall marriage penalties in the welfare state should be reduced by expanding the EITC for married couples; funds for this can be obtained from the other reforms presented in this paper.[19]

Three of the four legislative goals of the TANF program involve promoting healthy marriage. Regrettably, for two decades nearly all states have simply ignored the legislation. A substantial portion of any future TANF funding must be set aside explicitly to meet the original pro-marriage goals of the legislation; these funds should be used to reduce marriage penalties and to operate other promising pro-marriage programs. For example, PREP for Strong Bonds, a marriage-strengthening program for enlisted men in the Army, has been found in random assignment evaluation to reduce divorces by nearly 50 percent over a two-year period.[20]

6. Reject the ineffective pseudo-federalism of welfare block grants and begin to implement real federalism by transferring fiscal responsibility for low-income housing from the federal government to the states. More than 75 percent of total government spending on means-tested welfare comes from the federal government. Moreover, most of the state portion of welfare spending goes to a single program: Medicaid. Aside from Medicaid, welfare is nearly 90 percent federally funded. Making the federal government rather than state and local governments responsible for welfare makes no sense within our constitutional structure.

Block grants are a commonplace nostrum for restoring “federalism” in welfare, but welfare block grants are pseudo-federalism and bad public policy. Under a block grant welfare system, the federal government would collect hundreds of billions through taxes and then dump the money on state governments with the idea that they will spend it more wisely than the U.S. Congress can. In reality, the opposite is true: Block grants for welfare programs are a recipe for inefficiency and have rarely, if ever, led to conservative welfare innovation. By contrast, placing operational and fiscal responsibility for welfare programs at the same level of government yields greater accountability, efficiency, and fiscal prudence.

Restoring real federalism in welfare would mean that state governments would not only operate welfare programs, but also pay for those programs with state revenues. A first step toward real federalism would be to return control and fiscal responsibility of housing programs incrementally to the states. The federal government currently pays for over 90 percent of the cost of subsidized housing for poor and low-income persons at a cost of $52 billion per year.[21] Federal funding for means-tested housing programs should be phased down at a rate of 10 percent per year, reaching zero funding at the end of a decade. Each state should be allowed to determine how and to what extent it will replace federal housing programs with alternative programs designed and funded by state and local authorities.

7. Reform the earned income tax credit. The EITC is the nation’s largest means-tested cash welfare program. Although the EITC in some cases reduces federal income taxes owed, 85 percent of its expense goes to “refundable tax credits,” which are simply cash welfare grants to individuals who have no federal income tax liability. In 2015, refundable EITC payments cost $60 billion.[22]

According to the IRS, erroneous, often fraudulent over-claims account for 28 percent to 39 percent of all EITC payments.[23] The total over-claim amounts may be as high as $24 billion per year. Erroneous payments are generally the result of unverified income claims and payments to non-parents who do not reside with the child. Fraud can be greatly reduced by requiring income verification before payments are made and limiting eligibility to custodial parents and legal guardians.

Though somewhat more pro-work at low-income ranges than other welfare programs, the work incentives of the EITC are hampered because the value of the credit is not linked to actual hours worked. A parent who works all year at $10 per hour gets the same credit as a parent who works for half a year at $20 per hour. EITC benefit scales should be adjusted so that their value increases as the number of hours worked increases while not increasing overall spending on the program.

8. Reduce welfare fraud. The most common type of fraud in welfare involves “off-the-books” employment. In most welfare programs, benefits go down as earnings rise, but off-the-books employment is rarely reported to the welfare office; hiding earnings enables recipients to “double dip,” getting full welfare benefits that they are ineligible to receive while at the same time also receiving earnings from unreported jobs. An individual may also obtain an on-the-books job and fail to report it; this may permit the individual to double dip for up to six months before the welfare office discovers the unreported work. Work requirements such as those recommended in this paper substantially reduce welfare fraud because requiring a recipient to be in the welfare office periodically interferes with holding a hidden or unreported job. Recipients cannot be two places at once. Faced with a work requirement, many recipients with hidden jobs simply leave the rolls.

9. Reform social service and training programs by funding them on a pay-for-outcome basis. As noted, government spends more than $1 trillion per year on means-tested welfare. Some 90 percent of this spending goes to cash, food, housing, and medical benefits designed to prop up material living standards. The other 10 percent goes to programs aimed at improving human capabilities and changing behaviors in a positive direction. Examples of this second category are programs dealing with employment and training, child development, educational improvement, prisoner re-entry, drop-out prevention, and drug rehabilitation.

In this second category of program, the government generally pays for services provided rather than outcomes achieved. However, many desired outcomes for these programs can be clearly defined and measured. For example, programs may seek to reduce welfare dependence, increase employment, raise wages, increase high school completion, improve math and reading skills, reduce criminal activity and arrests, and decrease recidivism. Sadly, scientific evaluation shows that these social service programs rarely achieve strong positive outcomes.

Programs designed to build capability and enhance behavior can be improved by shifting from a funding-for-service model to a payment-for-outcome model. A portion of the payments made to state governments and individual grantees should be contingent on achievement of specified outcomes, and those outcomes should be verified by random assignment evaluation. For example, 20 percent of current funding in specified programs could be set aside as pay-for-outcome bonuses. States and grantees would attempt to win those bonuses by achieving specified performance outcomes.

10. Create greater employment opportunities for hard-to-employ individuals in low-income neighborhoods. Some low-skill individuals, such as former prison inmates, may have a particularly difficult time obtaining employment. Employment opportunities for these hard-to-employ workers may be expanded by reforming the Work Opportunity Tax Credit (WOTC).

The WOTC provides hiring incentives but without targeting truly disadvantaged job seekers ; in addition, the wage subsidies it provides are modest. A portion of WOTC funds should be retargeted at providing job opportunities to former inmates and similar hard-to-employ individuals in very low-income communities. Additional funds could be obtained from the EITC reforms outlined above. Job placement centers in poor communities, operating on a payment-for-outcome basis, could help participating individuals find jobs. The entire program should be evaluated with controlled random assignment experiments.

Such a targeted job creation program would differ greatly from proposed expansions of the EITC to non-parents.[24] In contrast to the proposals contained in this paper, such an EITC expansion for childless persons would simply increase welfare benefits. It would not be targeted to disadvantaged workers and would not increase job openings; it would, however, create significant new marriage penalties within the welfare state.

Conclusion

The foregoing reforms will increase efficiency in the use of taxpayer funds. They will reduce welfare dependence and decrease poverty through increased self-support. By promoting healthy marriage, they will enhance positive participation in society, increase upward mobility and improve personal well-being, thereby reducing the likelihood of future dependence and welfare expansion.

—Robert Rector is a Senior Research Fellow in the Institute for Family, Community, and Opportunity at The Heritage Foundation. Rachel Sheffield is a Policy Analyst in the Institute for Family, Community, and Opportunity.