President Biden has introduced a new plan that would greatly increase the welfare state. The plan would provide $78 billion per year in new cash grants to families who owe no income tax. The $78 billion in new cash benefits would be added to the nearly half a trillion dollars that government currently spends on cash, food, housing, and medical care for lower-income families with children.REF

If enacted permanently, the Biden cash grant plan would constitute the second-largest expansion of means-tested welfare entitlements in U.S. history. In constant dollars, its annual cost would dwarf the initial costs of the Medicaid, food stamp, and Aid to Families with Dependent Children programs. Only Obamacare would be more expensive.

The Biden plan would increase child cash grants from the current level of $2,000 per child to $3,000 for each child from six to 17 years of age and $3,600 for children under six. Critically, the Biden plan would eliminate all existing work obligations from the current child credit program. By doing so, President Biden rejects the foundational principles of welfare reform established in the 1990s.

That reform was rooted in the concept that welfare should no longer be a one-way handout. Instead, welfare should be based on reciprocal obligation: Society should support those who need assistance, but able-bodied recipients of aid should in turn be required to work or at least prepare for work in exchange for the aid given.

The Biden plan abandons that principle. It would eliminate work obligations from one of the largest cash welfare programs and restore the principle of welfare as an unconditional entitlement.

By increasing cash benefits while eliminating work requirements, the Biden plan would increase dependence and reduce work. This would be harmful to the poor, to taxpayers, and to society in general.

The Means-tested Welfare System

Means-tested welfare programs provide benefits that are not available to the general population but are restricted to persons who are poor or have lower incomes. Means-tested programs provide cash, food, housing, medical care, and targeted social services. The federal government alone operates nearly 90 separate means-tested aid programs.REF

In 2019, before the start of the COVID-19 recession, federal and state governments spent $1.24 trillion on means-tested aid.REF (These figures do not include Social Security or Medicare, which are not means-tested.) Some 73 percent of this spending came from the federal government, and 27 percent came from state and local governments. A major component of the means-tested welfare system is the nearly half-trillion dollars spent by the government on cash, food, housing aid, and medical care for poor and lower-income families with children.

As noted, means-tested programs give aid in a variety of forms. However, irrespective of the form of aid given, means-tested programs will have five main features:

1. Income Limits. Means-tested programs restrict benefits to persons or families with incomes below certain limits. Families with incomes above the cutoff level do not receive aid. By contrast, a universal program will provide benefits to everyone without regard to income.
2. Categorical Eligibility Standards. In addition to having income limits, means-tested programs may have categorical eligibility limits or may be non-categorical/broad-based. A categorical aid program will restrict eligibility to certain types of families or individuals. For example a categorical program may provide benefits only to specified eligibility groups such as disabled persons or families with children. By contrast, a non-categorical or universal program will be available to all lawful residents below the income limits.
3. Maximum Benefit Levels. Typically, a means-tested aid program will vary the benefits provided within the eligible group according to non-welfare income received. The maximum benefit is usually but not always given to families or persons with no earnings.
4. Benefit Phasedown Rates. The maximum benefits of a program are normally phased down as earnings incrementally rise. For example, with a phasedown rate of 30 percent, benefits would be cut by 30 cents for each added dollar of earned income. Phasedown rates are sometimes called the marginal tax rates although they usually have nothing to do with actual taxes.
5. Conditional Requirements. Programs will be either conditional or unconditional. Conditional aid programs will have positive behavior requirements. For example, a program may have a work requirement stating that an able-bodied recipient must be employed or engage in activities that prepare for work in order to receive benefits. Or a program might have a drug test. Individuals who fail to fulfill the positive performance standards will become ineligible for aid or will have benefits cut. By contrast, unconditional aid programs have no requirements and are explicitly neutral with respect to recipient performance.

Maximum Benefit Levels and Phasedown Rates

The maximum benefit level is quite simply the maximum dollar value of benefits that an eligible person or family may receive under a program. Typically, the maximum benefit will be given to families with zero earnings. The maximum benefit is sometimes called the income guarantee level.

Nearly all means-tested benefit programs have benefit phasedown rates by which benefits are reduced gradually as earned income incrementally rises.REF Typically, benefits will be phased down as earnings rise until benefits reach zero. The earnings level at which benefits reach zero can be called the benefit end point. It can also be called the break-even point. In most cases, the earnings level at benefit end point is the same as the program income limit.

One reason benefits are phased down gradually is to avoid benefit “cliffs.” For example, suppose a program provided $10,000 in benefits to every family with an annual income below $30,000. In that situation, a family would lose all $10,000 if its income rose by one dollar from $30,000 to $30,001. The benefit “cliff” would cause the family to lose $10,000 for one extra dollar in earnings. To avoid such irrational outcomes, most programs phase down benefits incrementally as earning rise.

Phasedown rates are usually described in percentage terms that represent the amount that benefits are cut for each $1.00 in added earnings. This incremental benefit reduction decreases the effective wage rate of the recipient. For example, with a phasedown rate of 100 percent, benefits are cut by $1.00 for each added dollar of earnings; the effective wage rate of each added hour of work would be zero. With a benefit phasedown rate of 50 percent, benefits will be cut by 50 cents for each dollar of extra earnings; in this case, a worker who earns $10 per hour will lose $5 in benefits for each added hour of work. His effective wage equals $5.00 per hour (the hourly wage minus the benefit reduction).

Chart 1 shows the benefits that would be provided in a hypothetical program with a maximum benefit of $10,000 at zero earnings with a 25 percent phasedown rate. The benefit levels appear on the Y-axis, and earnings levels appear on the X-axis. In this program, benefits are cut by $250 for each additional $1,000 in earnings. This program would give $7,000 in benefits at $12,000 in annual earnings. (At that earnings level, the maximum benefits would be cut by $3,000 or 25 percent times $12,000 in earnings.) At $20,000 in earnings, the benefits would be $5,000. At $40,000 in earnings, the benefits would be zero. Thus, $40,000 represents the benefit end point or break-even point for the program.REF

Examples of Benefits Levels and Phasedown Rates

Means-tested benefits are particularly important to families with children. For simplicity in presentation, the rest of this paper will focus on those families, although the same principles will apply to other recipient groups. Families with children generally receive benefits from multiple means-tested programs simultaneously. It is therefore useful to discuss the aggregate maximum benefit level and the aggregate phasedown rate of combined programs. The maximum value of combined benefits will always be higher than that of a single program, and the aggregate phasedown rates are also likely to be higher.

Chart 2 shows a set of hypothetical welfare programs with a traditional benefit structure. The Y-axis represents the aggregate benefits received, and the x-axis represents the earnings of the recipient family.REF Line A on the chart shows maximum aggregate benefits of $20,000 with the benefits phased down at a 100 percent rate: For each dollar of additional earnings, the benefits are cut by one dollar. The benefit end point for line A is $20,000 in earnings. Line B shows the same $20,000 benefit package subject with a phasedown rate of 50 percent; it has a benefit end point of $40,000 in earnings. Line C represents the same package with a phasedown rate of 25 percent: the benefit end point is $80,000 in earnings.

Varying Phasedown Rates

Chart 2 demonstrates a simple mathematical relationship between the maximum benefit, phasedown rate, and benefit end point in traditional welfare programs. The benefit end point will always equal the maximum benefit times the inverse of the phasedown rate. (The inverse of a fraction is that fraction turned upside down, so the inverse of ¼ is 4.) Therefore, Chart 3 shows that if the maximum benefit is $20,000 and the phasedown rate is 25 percent (or ¼), then the benefit end point will be $80,000 ($20,000 x 4 = $80,000).

For a given maximum benefit level, reducing the phasedown rate makes the line on the chart less steep and proportionately increases the benefit end point.REF As the benefit end point increases, the number of families eligible for benefits rises in roughly the same proportion.REF As the number of eligible families rises, the cost of the program will also rise in roughly the same proportion.

Consequently, reducing phasedown rates will be extremely expensive to taxpayers. Overall, for a program with a maximum benefit at zero earnings, reducing the phasedown rate from 100 percent to 25 percent will quadruple the benefit end point and roughly quadruple program costs.

The Impact of Welfare Benefits on Work and Employment

The maximum benefit level and the phasedown rates of programs affect labor supply and earnings through two mechanisms: an income effect and a net wage or phasedown effect.REF Individuals make choices between labor and “leisure.” Labor in this sense means work for pay, and leisure means all non-labor activities including household chores, child care, recreation, and socializing. Leisure is assumed to be a normal good, meaning that as income rises, individuals will seek to obtain more leisure along with other goods and services. The income effect occurs because as income rises, the individual, on average, will seek to “consume” more leisure as well as more purchased goods and services; this generates an increase in time spent on leisure and a decrease in time spent on labor.

The net wage or phasedown effect occurs because traditional welfare benefits are phased down as earnings rise. REF As noted, a program with a benefit phasedown rate of 33 percent will cut benefits by 33 cents for each extra $1.00 of earnings. A worker who on paper earns $10.00 per hour will get only $6.67 of net income for each hour worked. As the net financial gain for each hour of work is reduced, the incentive to work is also reduced. The net wage effect or phasedown effect measures the reduction in work due to the reduced incentive.

Welfare and Work: Evidence from the Negative Income Tax Experiments

The effect of welfare on work was tested in a series of large-scale random-assignment, controlled experiments in the 1970s called the negative income tax (NIT) experiments.REF These experiments tested the behavioral effects of experimental cash assistance programs that varied the maximum benefits provided and the benefit phasedown rates. Lower-income families and individuals were randomly assigned to experimental groups that received the new benefits and control groups that did not. The effect of the benefits on work effort was evaluated by the differences between experimental and control groups. NIT experiments were run in Seattle, Denver, Gary, Indiana, New Jersey, Pennsylvania, and rural areas in North Carolina and Iowa.

The NIT experiments clearly showed that welfare benefit programs are very inefficient in raising family incomes. More generous benefits reduced the work and earnings of husbands, wives, and single parents who received the aid.">REF Although the goal of the experiments was to raise family income, the income gained through added welfare was offset to a considerable degree by the income lost through reduced employment. Each $1,000 in added benefits was offset by a $660 reduction in earnings.REF This means that $3,000 in government benefits was required to cause a net increase of $1,000 in family income.REF This may help to explain why traditional welfare programs have had such difficulty in reducing poverty.

The NIT experiments clearly showed that welfare benefit programs are very inefficient in raising family incomes. Each $1,000 in added benefits was offset by a $660 reduction in earnings. This means that $3,000 in government benefits was required to cause a net increase of $1,000 in family income.

Even worse, although the NIT experiments lasted only three to five years, they had a negative effect on earnings of participants that persisted long after the programs ended. Each $1.00 of higher benefits provided by the experimental programs led to a $5.00 drop in the lifetime earnings of recipients.REF

Why Did the NIT Programs Reduce Work?

The simple view is that welfare reduces work because of the phasedown of benefits and the net wage effect. According to this view, high benefit phasedown rates (which make benefits drop more quickly) will reduce the effective return to work and thereby should reduce labor effort. By contrast, decreasing phasedown rates (making the benefits fall more slowly) should increase work among recipients.REF But reality is more complex.

In fact, the NIT experiments showed that welfare clearly reduced hours of work, but this decrease was due predominantly to the income effect, not the net wage effect. Raising the maximum benefit given to recipients clearly reduced work, but altering the phasedown rates of those benefits did not have a clear employment effect. Higher maximum benefits clearly led to reduced earning, but slowing the benefit phasedowns in programs did not mitigate the earnings loss.

For example, the report on the largest experiment, the Seattle/Denver Income Maintenance Experiment, stated: “We find the major portion of the labor [reduction] response to the NIT treatments is due to the income effect.”REF By contrast, the experiment showed “no conspicuous pattern with respect the wage rate [i.e., phasedown rates]” on work. In fact, contrary to common expectations, slower phasedowns in some program models were linked to lower, not higher work levels.REF

The same pattern was found in the Gary, Indiana, NIT experiment, which examined the impact of various welfare benefit designs on black families, both married couples and single mothers. The experiment showed significant labor supply reductions for husbands and female heads but not for wives.REF Female heads reduced hours worked by up to 30 percent. The NIT report stated that “these [labor] responses seem primarily responses to the guarantee level rather than the tax rate.”REF (The guarantee level is the maximum benefit, and the tax rate means the benefit phasedown rate.) Researchers found that variations in the benefit phasedown rates had no “perceptible effect on labor supply.” By contrast, the income effect generated by increases in maximum benefits had statistically significant negative impacts on labor.REF

In the New Jersey NIT experiment, most of the work reduction occurred among wives. Again, the decreases in work were driven by income effect due to higher benefits. Variation in the phasedown rates did not have an impact.REF

Explaining the Benefit Phasedown Puzzle

To many, the NIT results seem counterintuitive. The factors behind the perplexing results can be illustrated in Charts 3 and 4. Chart 3 shows a simple increase in maximum benefits, which are raised from $20,000 to $25,000. The benefit phasedown rate is unchanged; the lines are parallel. Both plans phase down benefits by 50 cents for each dollar of earnings. Thus, in both plans, the benefit reduction cuts the effective wage of workers in half.

But, at each level of earnings, recipients will receive $5,000 more in benefits in Plan B compared to Plan A. The increase in benefits will produce an “income effect” that will decrease work. The phasedown rates are unchanged, so there would be no wage effect. Increasing the maximum benefit creates an income effect that reduces work; therefore, overall, shifting from Plan A to Plan B unambiguously reduces work.

Raising the Maximum Benefit

Chart 4 shows a different hypothetical scenario. The maximum benefit remains fixed at $20,000, but the phasedown rate is shifted from 50 cents for each added dollar of earnings in the plan on line A to 25 cents for the plan on line B. With the phasedown rate reduced, the effective wage rate for workers is much higher on line B than on line A; this increases the incentive to work.

However, there is a second, less obvious change shown on the chart. Note that even though the maximum benefit was not changed, the benefits received will be higher in line B than in line A at each earnings level along the X-axis. For example, a family earning $20,000 on line A will receive $10,000 in benefits. The same family will receive $15,000 in benefits in the plan on line B. This $5,000 increase in benefits will generate an “income effect” that will reduce work and earnings. The increased benefits and work-reducing income effect will occur automatically whenever phasedown rates are decreased (and the lines on the chart become less steep).

Thus, on Chart 4, there is a net wage effect that will increase work and an income effect that will reduce work. The two effects offset and neutralize each other. As a consequence, in most cases, cutting benefit phasedown rates (i.e., slowing the phasedown of benefits) will not increase work.

These offsetting effects explain why altering the phasedown rates in the NIT experiments had little impact on labor. In reality, reducing phasedown rates with a constant maximum benefit will not increase work and may reduce it.

The report on the NIT experiments in Seattle and Denver clearly describes the countervailing effects shown in Chart 4. The report explains that if maximum benefits are held constant, reducing the benefit phasedown rate (making the line on the chart less steep) will not increase employment because this generates “both an income effect and a substitution [or phasedown] effect, which are offsetting.”REF In fact, if the maximum benefit of a program is held constant, lowering the benefit phasedown rate (slowing the decline in benefits and making the line on the chart less steep) appeared counterintuitively to reduce work, suggesting that the “the income effect dominates the substitution effect.”REF

The Effect of Expanding Eligibility

There is a final effect of reducing benefit phasedown rates that is strong and unambiguously anti-work. Note that in Chart 4, reducing the benefit phasedown rate from 50 percent in line A to 25 percent in line B raises the benefit end point from $40,000 to $80,000. A whole group of families with incomes between $40,000 and $80,000 who were not eligible for benefits under the original Plan A become eligible under Plan B. Roughly speaking, the shift from Plan A to Plan B doubles the recipient population.

The impact of extending new benefits to the $40,000 to $80,000 income group is unambiguously anti-work. This group will receive a new benefit, which raises income and reduces work through the income effect. The group is also subject to an entirely new 25 percent benefit phasedown; this reduces the effective hourly wage of workers in this income class by 25 percent and thereby reduces work through the net wage effect.

Phasedown Rates and Program Costs

Some may reject the argument presented above concerning the offsetting effects when benefit phasedown rates are lowered. They may continue to believe that the main anti-work effects in welfare are due to high benefit phasedown rates and that the only way to promote work is to reduce the benefit reduction rates (making the lines on the charts less steep).REF

This approach runs into difficulty because, on average, the comprehensive marginal tax rates (CMTR) for lower-income households are not greater than those of the rest of society. (The comprehensive marginal tax rate represents the joint effect of the actual taxes paid by households plus the phasedown rates of any benefits they receive.) The median marginal tax rate for a given year for the lowest income quintile appears to be 37.8 percent compared to 36 percent for the middle quintile.REF

It is true that there is a greater variation in CMTR among low-income households because those households may participate in a greater variety of government aid programs. Those with the highest CMTR are likely to receive aid from a number of high-cost, high-benefit programs such as Section 8 housing. But most low income households do not face greater CMTR than the rest of society.REF

In addition, those who continue to believe that work is reduced among the poor primarily due to the rapid phasedown of benefits and that work can be incentivized only by a broad slowing in those phasedowns face a fundamental dilemma. Average phasedown rates can be reduced by only two strategies. The first is to significantly reduce the maximum benefits given to most families who do not work or who work little. The second is to significantly reduce the phasedown rates across a broad array of benefit programs (making the lines on the charts less steep). As this paper has demonstrated, the latter approach will greatly increase taxpayer costs while substantially enlarging the number of those who are dependent on government aid.

Even Worse Consequences in Universal Programs Such as Child Allowances

The fiscal consequences of universal programs such as child allowances are even worse. These programs provide uniform benefits to nearly the whole population irrespective of income. As a broad policy for reducing poverty, the cost of such programs is prohibitive.

For example, a means-tested program that offered every family with children cash benefits equal to the federal poverty level and phased the benefits down at 40 percent would cost about $167 billion per annum.REF As Chart 5 shows, the same program operated as a universal program with a zero phasedown rate would cost $956 billion, almost six times as much.

A Simple Solution: Work Requirements

There is one simple solution to increase work and reduce dependence without slashing benefits or greatly expanding the cost and scope of programs. That solution is work requirements.

When a work requirement is established in a welfare program, recipients are expected to find employment. If a job is not immediately available, the recipient will be required to engage in various supervised activities that prepare for and lead to employment. These may include supervised job search, job prep programs that help recipients prepare resumes or learn interview and job search skills, education, formal job training, or community service.REF The fundamental idea is that recipients are no longer permitted to sit at home and simply collect welfare checks; instead, they are required to go to the welfare office or similar agency and consistently engage in activities aimed at increasing self-support.

Attaching a work obligation to welfare benefits fundamentally changes the nature of welfare. A recipient no longer faces a choice between labor and leisure (broadly defined). Instead, the recipient faces a choice between two types of labor: employment or work-like activities mandated by the welfare agency. By removing the ability of recipients to receive recurring welfare payments without effort, a work requirement greatly reduces the attractiveness of welfare dependence relative to employment. The work requirement will reduce dependence and increase self-support through employment or greater reliance on relatives and parental partners.

The principle of work requirements in welfare is based on common sense and has nearly universal public support. The public supports aid to those who need it, but they do not want welfare to be a one-way handout. The key is reciprocity: If aid is given to able-bodied individuals, they should be required to take positive steps toward self-support in exchange for the assistance given. Polls show that close to 90 percent of Americans agree that “able-bodied adults that receive cash, food, housing, and medical assistance should be required to work or prepare for work as a condition of receiving those government benefits.”REF

The History of Work Requirements

Work requirements played a central role in the welfare reform enacted in the 1990s. In 1992, Bill Clinton was elected President on the pledge to “end welfare as we know it.” At that time, the welfare state was clearly failing: One of every seven children in the U.S. was dependent on the Aid to Families with Dependent Children (AFDC) program. Work among recipients was very low, and the typical family received AFDC benefits for 14 years. Unwed childbearing had been rising for decades.

Within a few years, welfare reform was enacted. For the first time, recipients of cash aid were required to work or prepare for work as a condition of receiving benefits. In response, the welfare caseload experienced its first significant decline in a half-century. Within a few years, dependence had plummeted by 60 percent.REF At the same time, employment of less-skilled single parents surged.REF Child poverty, which had been static for decades, fell at an unprecedented rate, especially among black children.REF

While welfare reform was passed on a broad bipartisan basis, it was never accepted by the progressive left, which strongly opposed work requirements. For two and a half decades, the left has consistently sought to block, weaken, or remove work requirements across all welfare programs. The Biden child allowance plan, if enacted, would represent a decisive victory for the left in those efforts.

The Biden Child Allowance Plan

The Biden child allowance plan would increase the refundable credits/cash grants from $2,000 per child to $3,000 for each child six to 17 years of age and $3,600 for children under six. Two-thirds of the new benefits provided (or $78 billion per year) would be cash grants to families who owe no income tax. The $78 billion in new benefits would be added to the nearly half a trillion dollars government currently spends on cash, food, housing, and medical care for lower-income families with children.

The Administration suggests that these changes would be limited to a single year to help families suffering under the COVID-19 recession. In fact, the Biden plan is based on legislation that would create permanent new entitlements.REF Reports indicate that the real goal is the permanent expansion of the welfare state.REF

If enacted permanently, the Biden cash grant increase would constitute the second-largest expansion of means-tested welfare entitlements in U.S. history. In constant dollars, its annual cost would dwarf the initial costs of the Medicaid, food stamps, and AFDC. Only Obamacare would be more expensive.

Critically, the Biden plan eliminates all existing work obligations from the child credit program. Under current law, a family must have at least $2,500 in annual earnings to be eligible for any cash grant. Moreover, the credit is designed to encourage work; it starts at a low value that increases as a family’s work and earnings increase.

Removes the Need for a Parent on Welfare to Work to Support the Family. The Biden plan would completely eliminate the existing work requirement and pro-work phase-ins. For the first time, non-working families would receive a full cash grant. The primary beneficiaries of this change would be non-working single-parent families. A non-working single parent with two children would receive a net windfall of $6,000 to $7,200 each year.

Advocates of the Biden proposals may argue that $6,000 per year for a single parent with two children is not enough to remove the need for the parent to work to support the family. The facts, however, suggest otherwise. In 2018, approximately 2.7 million families with children earned less than $2,500 during the year and were therefore ineligible for the current child cash grants.REF Of this group, two million were single-parent families. While these families did not receive the cash grants, they did receive extensive other benefits.

For example, a non-working single parent with two children would be eligible for approximately $11,000 in Medicaid benefits and would receive around $7,000 in food stamps and other nutrition benefits.REF About half of these non-working families would also receive either subsidized housing (average value $11,000 per year) or TANF without an active work obligation ($6,000 per year).REF

The Biden plan would destroy the foundations of welfare reform. The principle of reciprocal obligation, by which the taxpayer helps the recipient but the recipient in exchange is expected to take steps to help himself, is abandoned. The idea of mutual obligation is replaced by unconditional entitlement. In the future, remaining obligations would be unilateral: Taxpayers would be required to pay large sums to support welfare recipients, but recipients would have no obligations in return.

Finally, nearly three million single parents with little or no employment indirectly receive Earned Income Tax Credit (EITC) payments by fraudulently transferring the EITC work obligation to absent fathers or relatives.REF Altogether, non-working families will typically receive between $20,000 and $30,000 per year in multiple benefits; the Biden plan would add another $6,000 to $7,200 on top of that. All of the aid would be given unconditionally; no recipient would be required to take even a single step toward self-support.

The Biden plan would destroy the foundations of welfare reform. The principle of reciprocal obligation, by which the taxpayer helps the recipient but the recipient in exchange is expected to take steps to help himself, is abandoned. The idea of mutual obligation is replaced by unconditional entitlement. In the future, remaining obligations would be unilateral: Taxpayers would be required to pay large sums to support welfare recipients, but recipients would have no obligations in return.

Conclusion

Controlled experiments show that higher welfare benefits substantially reduce work and earnings among families with children. This occurs because welfare serves as a substitute for earnings and reduces the necessity for consistent employment and self-support.

The view that the phasedown rates of welfare benefits are the primary cause of welfare’s anti-work effects is inaccurate. In fact, if the maximum benefits in a welfare program are held constant, altering the benefit phasedown rates appears to have little impact on work or earnings.

Moreover, lowering benefit phasedown rates would greatly increase program costs and caseloads. This is particularly true of programs such as a universal child allowance that offer a uniform benefit to families irrespective of income. Such programs can be six times more expensive than an equivalent means-tested program and therefore cannot serve as a fiscally viable basis for a broad-based anti-poverty system.

By contrast, work requirements in welfare are a simple, cost-effective means for reducing dependence and increasing self-support. Work requirements based on the principle of reciprocal obligations between taxpayers and aid recipients have almost universal public support.

Government already spends nearly half a trillion dollars per year on cash, food, housing, and medical care for low-income families with children. The Biden plan would greatly increase cash grants while eliminating all existing work obligations in the current child credit program. It abandons the link between work and welfare established by welfare reform in the 1990s and re-establishes the principle of unconditional entitlement to taxpayer-funded benefits.

A better policy would be to strengthen work obligations for able-bodied recipients in such programs as Temporary Assistance to Needy Families, the EITC, subsidized housing, and food stamps. The welfare system should also be reformed to eliminate or reduce the stark marriage penalties embedded in most welfare programs. This policy would be a more effective means of reducing child poverty and improving the long-term well-being of adults, children, and society.

Robert Rector is a Senior Research Fellow in Domestic Policy Studies, of the Institute for Family, Community, and Opportunity, at The Heritage Foundation.