Effects on Economic Outcomes

The U.S. economy flourished during the 1990s, at the same time that the states implemented reforms to their welfare programs. By 2001, the welfare caseload had fallen to about 2.1 million families, less than half its all-time peak of 5 million families in 1994. The fraction of welfare recipients participating in welfare-to-work activities, or actually working, increased rapidly. Employment rates and earnings of single mothers rose substantially. Family income rose. Poverty declined. Even the prior upward trend in nonmarital fertility leveled off.

Because these improvements occurred as welfare reform took place, some observers point to welfare reform as the cause. However, other policies also changed during the 1990s, notably the increases in the Earned Income Tax Credit, the uncoupling of subsidized health insurance from welfare receipt, and the increases in the minimum wage. The long and robust economic expansion was an important change in itself. Therefore, some of the improvements probably resulted from the strong economy and from changes in other policies, rather than from welfare reform.

Ideally, in designing welfare programs, lawmakers would consider the separate effects of each of the reform policies. Toward that goal, our synthesis aims to answer the question: What is the effect of a given pol-icy (such as work requirements) on a given outcome (such as welfare use), holding all else equal? We also consider the effects of reform as a whole: the combined effects of the bundle of policies implemented together. We summarize our results in four tables.

Table 1 shows the effect of different welfare reforms and reform as a bundle on the five key economic outcomes: welfare use, employment, earnings, income, and poverty. An orange cell indicates that a reform (or bundle of reforms) has been found by most studies to contribute to an increase in an outcome, such as increased welfare use or increased employment. A brown cell indicates that a reform is associated with a decrease in an outcome. Note that either an increase or a decrease can be desirable, depending on the outcome. Gray cells indicate that roughly equal numbers of studies have identified increased and decreased results. Blank cells indicate that no studies exist at all.

Stars indicate the strength of the evidence provided by the existing literature. Four stars indicate the strongest evidence (several high-quality studies, most of which yield similar and significant estimates). Three stars indicate an intermediate level of evidence (two or three high-quality studies). Two stars indicate only weak evidence (a single high-quality study or two moderate-quality studies with significant estimates). One star indicates that the available studies offer insufficient evidence to assign a direction of effect with a minimum level of confidence.

Mandatory work-related activities (either employment or job training) have been studied more than any other reform. A substantial body of evidence shows that they generally reduce welfare use while increasing employment and earnings. However, when implemented in conjunction with the traditional welfare benefit structure, in which benefits fall nearly dollar for dollar as earnings rise, the work requirements have little effect on income. The earnings gained are not enough to offset the benefits lost.

Nonetheless, the evidence suggests that work requirements do decrease poverty. This finding may imply that the activities raise income primarily for those who live just below the poverty line. This interpretation is consistent with other evidence that such programs offer greater income benefits to relatively advantaged welfare recipients than to relatively disadvantaged ones.

Sanctions for noncompliance with work requirements may include either a partial or full reduction in the monthly welfare grant. Some evidence shows that sanctions and time limits both reduce welfare use, but we know relatively little about their effects on other economic outcomes.

Of all the policies shown in Table 1, only strong financial work incentives, either alone or in combination with work requirements, have simultaneously raised income and reduced poverty. (These effects were modest.) Strong financial work incentives allow welfare recipients either to keep more of their welfare checks as their earnings rise or to receive income supplements conditioned on full-time work. However, in contrast to the other reform policies, financial work incentives also increase welfare use, because families remain eligible for benefits even at higher levels of earnings.

The row in Table 1 that combines mandatory work-related activities with strong financial work incentives illustrates a central point. The row shows that it is possible to require work, to raise income, and to reduce poverty all at the same time. The key is to combine a work requirement with a relatively generous financial work incentive, so that earnings rise more rapidly than benefits fall. The trade-off for raising income, however, is higher welfare use.

Table 2 further highlights the trade-offs involved as policymakers seek to design welfare programs. For each of the three major policy reforms—work requirements, financial work incentives, and time limits—a check mark indicates whether that policy is likely to achieve one of three possible objectives of reform: reducing welfare dependency, increasing work, or raising income. From our synthesis, we conclude that work mandates reduce welfare use and increase employment, but they do not necessarily raise income. In contrast, strong financial work incentives raise employment and income but do not decrease welfare use. Time limits have similar effects as work mandates.

Depending on the decisionmaker's objectives, different combinations of policies may be preferred over others. Policymakers who place a greater priority on reducing dependency and increasing employment will favor work requirements and time limits. Policymakers who place more weight on raising income and reducing poverty will favor financial work incentives.

Returning to Table 1, we conclude that the combined effects of the reforms implemented during the 1990s have been a reduction in welfare use; an increase in employment, earnings, and income; and a reduction in poverty. However, the favorable effects on income and poverty represent what happened before most welfare recipients could have reached their time limits. The effects may become less benign as more recipients exhaust their benefits.

 
Effects on Child Well-Being

Our knowledge of the effects of welfare reform on child well-being is quite limited, but there appear to be both positive and negative effects on children, and the effects can vary by age (see Table 3). For example, work requirements seem to have mixed effects on preschool children, a positive effect on the health of grade-school children, and a negative effect on school achievement among adolescents. We cannot explain these discrepancies.

The most positive effects on children are associated with financial work incentives, most likely because family income increases as welfare recipients combine work and welfare. Strong work incentives alone or in combination with work requirements generally appear to decrease behavior problems and school achievement problems among children in grade school. However, the same reform likely increases both kinds of problems among adolescents. Depending on how financial work incentives are combined with work mandates, they may either improve health among grade-school children or contribute to impaired health among the same age group.

Because of the very limited evidence available, we have even less clarity about the overall effects of reform on child well-being than on economic outcomes. The bottom row of Table 3 shows a mixed effect on the behavior problems of grade-school children, a salutary effect on the health status of the same children, a mixed effect on the behavior problems of adolescents, and a detrimental effect on the achievement problems of adolescents. Yet these results should be interpreted cautiously. They are all based on two studies of bundles of reforms that are not very representative of those implemented in other states.

Furthermore, we know very little about the effects on children of parental responsibility requirements, such as providing child immunizations and other preventative health care. We also know virtually nothing about the effects on children of sanctions, time limits, and family caps.

Yet the early evidence suggests that welfare reforms affect children of different ages differently. There seem to be countervailing forces that both promote and impair behavioral, social, cognitive, and physical development. The prevailing effects are likely to depend on the strength of the opposing forces, on the child's stage of development, and on other circumstances.

Clouding our understanding further, a nearly universal limitation of our conclusions is that they apply mostly to the short run. Understanding the longer-term effects of reform is especially pertinent for child well-being. Some aspects of child well-being, such as cognitive skills, are likely to take much longer to change than parental work habits. Even the shorter-term aspects of child well-being, such as behavior problems, could vary as children are exposed to cumulatively lower levels of welfare use and cumulatively higher levels of parental employment.

Whether the effects identified in this research will continue in the longer run—or whether they will be attenuated or exacerbated—remains to be seen. Long-run information on the effects of current welfare reforms on family structure and child well-being is crucial.

 
Effects on Other Outcomes

We are almost certain that mandatory work-related activities have neither a positive nor a negative effect on marriage or fertility (see the black cells in Table 4, which denote that a policy contributes to no change in an outcome). This conclusion is based on five years of follow-up data for seven programs and two years of follow-up data for five additional programs.

The rest of Table 4 reveals another series of trade-offs. The evidence indicates that work requirements (and other policies that leave income unchanged) can undermine other measures of well-being, such as food security and children's health insurance coverage. Meanwhile, strong financial work incentives (and other policies that raise income) can enhance food security, children's health insurance coverage, and savings.

Only the most limited evidence suggests that financial work incentives may also promote marriage—a principal goal of the 1996 legislation. This finding comes from a single random assignment experiment that is not corroborated by any others.

We do not know how financial work incentives may affect marriage when they are conditioned on full-time work or combined with work requirements. We also know little, if anything, about the effects of sanctions, time limits, family caps, or parental responsibility requirements on marriage, fertility, or other measures of well-being.

 
Unknown Effects

The tables show that our knowledge is strongest regarding the effects of welfare reform policies on welfare use, employment, earnings, and income. Our knowledge is weakest regarding the effects on children, especially preschool children, and on broader measures of well-being.

Among the various policies, we know a lot about the consequences of mandatory work-related activities. We know nearly as much about financial work incentives (either alone, when conditioned on full-time work, or when combined with mandatory work-related activities). We know much less about the consequences of other reforms. Overall, just under half the cells in our tables (120 out of 242 cells) remain empty, indicating that no research has been done to assess the consequences of the policies. Another 37 cells (those with a single asterisk) are nearly empty.

Some of the gaps in knowledge are highly relevant for policy. There have been very few studies of the effects of welfare reform on the participation of poor families in the Medicaid program or on the health care coverage of children more generally. This omission is particularly troublesome in light of the apparent decreases in Medicaid enrollment following the implementation of TANF—despite 15 years of policy initiatives designed to increase the coverage of poor children.

With an increasing emphasis on work for mothers of children as young as age one, it is also unfortunate that so little is known about the effects of welfare reform on child development prior to school entry. This issue is especially relevant for policies aimed at improving early care and education.

Despite continued interest among many policymakers in promoting two-parent families and reducing out-of-wedlock births, little is known about how reform policies other than work mandates affect marriage and fertility. The short-run nature of most of the evidence limits our understanding of whether reform has accomplished its goals in these areas. Since marriage and fertility involve substantially more inertia than other aspects of behavior, we would expect the effects of welfare reform on such outcomes to become apparent only over a longer horizon. It should come as little surprise that most of the evidence from high-quality studies to date is both mixed and statistically insignificant.

A more general omission is any understanding of how reform has affected family decisions to enter welfare. Random assignment experiments reveal how policy reforms affect those who are either applying for or already on welfare. The experiments give no information about how families decide to apply for aid in the first place. There have been only a few econometric studies that focus specifically on welfare entry. This omission is significant, because recent findings indicate that as much as half of the recent decline in the welfare caseload is attributable to declining rates of entry.

The federal reauthorization proposals now on the table include modifications to current reform policies, such as work mandates. We have no direct evidence to weigh the merits of the proposed changes, other than extrapolating from experience with current policies. Moreover, our evidence on the effects of the reforms is limited to a time in which the strong economy played a key role in reducing welfare use and increasing self-sufficiency. We can only speculate on how welfare recipients may fare in a weakened economy.

In sum, the short-term results have been mostly positive, but the outcomes were assessed in a period when the economy was booming and before time limits were reached. Even under these conditions, there have clearly been trade-offs among the policies. We also suspect that progress will become harder to sustain as welfare recipients begin to reach their time limits in significant numbers. For these reasons, it is important that the federal and state policymakers who design welfare reform policies review and revise them with realistic expectations based on what we already know about the likely consequences.

 
Related Reading

Consequences of Welfare Reform: A Research Synthesis, Jeffrey Grogger, Lynn A. Karoly, Jacob Alex Klerman, RAND/DRU-2676-DHHS, 2002, 331 pp., no charge.

A Decade of Welfare Reform: What We've Learned About Child Well-Being, RAND/RB-5068-DHHS, 2002, 2 pp., no charge.

A Decade of Welfare Reform: What We've Learned About Welfare Usage and Economic Outcomes, RAND/RB-5067-DHHS, 2002, 2 pp., no charge.

Estimating the Effect of Work Requirements on Welfare Recipients: A Synthesis of the National Literature, Lynn A. Karoly, RAND/CT-185, 2001, 18 pp., $5.00.

A Stock-Flow Analysis of the Welfare Caseload: Insights from California Economic Conditions, Jacob A. Klerman, Steven J. Haider, RAND/DRU-2463-DHHS, 2001, 42 pp., no charge.