Changes in Participation in Means-Tested
Programs

In the early 1990s, the number of people receiving benefits from Aid to Families with Dependent Children (AFDC), a major cash assistance program for nonelderly U.S. residents, increased rapidly. Concerned about the nation's welfare system, lawmakers in 1996 replaced AFDC with a block-grant program, called Temporary Assistance for Needy Families (TANF), and made other significant changes in federal welfare policies. Since then, the number of AFDC/TANF recipients has dropped by more than 60 percent--from about 13 million to 5 million. Analysts disagree about the extent to which the new program, the robust economy of the late 1990s, or other factors played a role in that decline.

The substantial drop in the number of TANF recipients raises questions about whether participation in other means-tested programs has followed similar patterns. This brief looks at trends in participation in some of the major means-tested programs for low-income nonelderly people--AFDC/TANF, the earned income tax credit (EITC), Food Stamps, Medicaid, and Supplemental Security Income (SSI)--from fiscal year 1977 to 2003 (the most recent year for which data about participation are available).

Unlike the pattern with AFDC/TANF, those other four programs have generally experienced increases in participation in recent decades. Moreover, as of 2003, each of those programs served more low-income nonelderly people than did TANF, the program commonly thought of as welfare. The federal government also spends more on those programs than on TANF. In 2004, federal outlays for TANF totaled $18 billion, compared with $24 billion for Food Stamp benefits, $30 billion for SSI benefits for nonelderly recipients, $33 billion for the EITC, and $114 billion for the federal share of Medicaid spending on benefits for nonelderly people.

Future participation in, and federal costs of, those five programs will depend on many factors, such as the level of benefits, demographic trends, labor market conditions, and health care costs. Under current law, changes in the total federal cost of those programs are likely to be dominated by health care costs, since Medicaid is by far the largest and fastest growing of the programs.
 

Major Federal Means-Tested Programs

The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA) created the Temporary Assistance for Needy Families block-grant program and eliminated the Aid to Families with Dependent Children program that had been in place. In general, under AFDC, single parents (usually mothers) who met certain income and resource restrictions--the means test--and who had a child under age 18 were eligible for cash aid. Under TANF, such parents may still be eligible for cash aid and work-related activities, including job training and child care. But typically, they must meet certain work or training requirements to receive benefits, and they can only receive aid for 60 months during their lifetime.⁽¹⁾ Although AFDC had similar restrictions on income and assets, it did not have strict work or training requirements or a lifetime limit on benefits.

Various other federal or federal/state programs provide financial support to low-income nonelderly people. The earned income tax credit is a refundable tax credit available to workers with low income. If the amount of the credit exceeds the amount of federal income tax that a person owes, the excess is payable directly to the taxpayer. The Food Stamp program provides a monthly allotment of benefits to qualifying households. Recipients normally receive an electronic benefit card that they can use to buy approved food products at approved stores. Medicaid provides federal matching funds so that states can offer medical assistance to poor children and their families and to some disabled and elderly people. A more recent program, the State Children's Health Insurance Program (SCHIP), created in 1997, provides federal matching funds so that states can extend medical coverage to uninsured children (and sometimes parents) in low-income families. The federal Supplemental Security Income program provides monthly cash payments to low-income elderly, blind, and disabled individuals.
 

Factors That Affect Participation

The number of participants in federal means-tested programs depends, at least partially, on the number of people who are eligible and thus on population growth.⁽²⁾ Since 1977, the number of U.S. residents under age 65 has increased by more than 30 percent--from 196 million to 257 million people--which by itself would suggest that the number of participants in the programs discussed above should have risen as well. Other demographic factors, such as long-term trends in child birth, marriage, and disability, also play a role. For instance, if the number of single mothers in the population grows over time, the number of participants in programs for which a single mother might be eligible, such as TANF, is also likely to increase.

Another factor that influences participation is the value of a program's benefits relative to the value of work, activities such as schooling, and alternative forms of support. Participation is likely to increase as the relative value of benefits increases. Benefit levels are generally set by policy decisions, but earnings--an important alternative to program participation--change with the state of the economy. For instance, a rise in real (inflation-adjusted) wages for low-skilled workers would boost the payoff from work for many actual or potential program participants.

The value of program benefits is also affected by any social stigma associated with participation and by the stringency of any requirements that have to be met to participate. As noted above, many TANF recipients must work in order to receive benefits. The more stringent the work requirement, the less valuable the TANF benefit is likely to be to a recipient.

Another factor that could restrict participation in TANF is the 60-month lifetime limit on eligibility. Some states have promised to use state funds to continue aiding people who reach that limit, and each state can exempt a portion of its caseload from the limit. Information currently available about the effects of time limits suggests that they have not often been binding.
 

Trends in Participation

Participation by nonelderly people in the earned income tax credit, SSI, and Medicaid has increased almost every year since 1977 (see Figure 1). By contrast, participation in Food Stamps and AFDC/TANF remained fairly constant or trended upward between 1977 and 1994 and then declined significantly. But whereas AFDC/TANF participation has fallen steadily since 1994, the number of Food Stamp participants began rising rapidly again after 2000 as the labor market weakened. (Participation is measured slightly differently for those various programs; see Box 1 for details.)

Figure 1.

Source: Congressional Budget Office.

Notes: AFDC = Aid to Families with Dependent Children; TANF = Temporary Assistance for Needy Families; SCHIP = State Children's Health Insurance Program.

The data series shown here and their sources are described in Box 1. Data for AFDC/TANF and the earned income tax credit cover participants of all ages, including the elderly, and data for the tax credit represent families, not individuals.

Some of the changes in participation in federal means-tested programs are associated with changes in federal policies, although distinguishing the effects of legislative initiatives from the effects of economic and demographic trends can be difficult. In the case of the EITC, the number of families claiming the credit frequently rose in the 1980s and 1990s following major policy changes that increased benefits, expanded eligibility, or both. The Tax Reform Act of 1986 raised EITC benefits significantly, and participation grew by about 75 percent over the next four years. In 1990, lawmakers expanded the credit further, with the expansion phased in over three years. And in 1993, lawmakers raised EITC benefits sharply for families with two or more children and expanded eligibility to include childless workers. Between 1993 and 2003, participation in the EITC program grew by nearly 30 percent.

For the federal SSI program, annual increases in participation accelerated in the early 1990s after a 1990 Supreme Court decision (in Sullivan v. Zebley) that revised the definition of disability for children and led to more children being eligible for SSI benefits. Lawmakers enacted a more restrictive definition of disability for children in 1996 in PRWORA. However, that change was not followed by a significant decline in the total number of nonelderly recipients of SSI benefits.⁽³⁾ In the late 1990s and early 2000s, restrictions on the eligibility of immigrants and substance abusers dampened growth in SSI participation.

For Medicaid, increases in the number of nonelderly beneficiaries coincided with policy changes that loosened eligibility and with substantial growth in the cost of health care (and thus, potentially, in the value of health care services).⁽⁴⁾ In the 1980s and early 1990s, eligibility for Medicaid was expanded to include more children and pregnant women from low-income families. Those changes were followed by significant growth in the number of nonelderly beneficiaries. In 1996, PRWORA broke the linkage that had existed between participation in Medicaid and participation in the AFDC and SSI programs. Finally, the Balanced Budget Act of 1997 created SCHIP to make health care more accessible to low-income children. That program has grown rapidly in the past several years.

The AFDC/TANF program did not experience the same consistent increases in participation that the EITC, SSI, and Medicaid/SCHIP programs did. Participation in AFDC remained relatively stable from 1977 to 1989, but in 1990 it began to rise rapidly, peaking in 1994. Participation in AFDC then began to decline, and participation in TANF continued to fall through 2003. Analysts believe that the drop since 1994 resulted, in some combination, from a strong economy during most of that period, the shift in requirements between the AFDC and TANF programs, and other policy changes that occurred at the same time, including expansions of the EITC.

Participation in the Food Stamp program followed a similar pattern as participation in AFDC/TANF during the 1980s and 1990s--declining slightly throughout most of the 1980s, increasing sharply between 1988 and 1994, and then falling rapidly. One probable reason is that in 1996, PRWORA restricted eligibility for some groups of recipients, including able-bodied adults without dependents and legal residents who are not U.S. citizens. The decline in Food Stamp participation is also partially explained by the decline in AFDC/TANF participation during the 1990s. Some people who left AFDC/TANF also stopped receiving food stamps, although they may have still been eligible for that program. Unlike participation in TANF, however, Food Stamp participation has risen since 2000. That increase is partly explained by outreach efforts and by the restoration of benefits for many noncitizens in subsequent legislation. The increase also coincides with the 2001 recession and its lingering effects on the labor market.
 

Trends in Benefit Levels

In most of the programs discussed here, the average monthly benefit received by participants stayed the same or rose over the 1977-2003 period, after adjustment for inflation (see Figure 2). Food Stamp benefits remained relatively constant, averaging about $80 per person per month (in 2003 dollars). The average SSI benefit grew slightly, from $400 per month in 1977 to $440 per month in 2003. Real increases were much more substantial in the EITC and Medicaid programs. The average EITC benefit tripled during that period, to almost $150 per family per month, and Medicaid benefits doubled, to nearly $300 per person per month. Only in the AFDC/TANF program did real cash benefits fall, from about $240 per month in 1977 to $140 per month in 2003.

Figure 2.

Source: Congressional Budget Office.

Notes: AFDC = Aid to Families with Dependent Children; TANF = Temporary Assistance for Needy Families.

The data series shown here and their sources are described in Box 1. Data for AFDC/TANF and the earned income tax credit cover participants of all ages, including the elderly. Data are not available for some programs for certain years (such as TANF in 1999).

The spike in Supplemental Security Income benefits in 1993 was probably caused in part by retroactive payments that were made to disabled children as a result of the Supreme Court's decision in Sullivan v. Zebley.

Would Participation and Spending Rise If the Economy Weakened?

The Congressional Budget Office (CBO) projects that if current laws do not change, spending on many of the means-tested programs discussed here will grow modestly over the next five years.⁽⁵⁾ The exceptions are TANF and SCHIP, for which annual spending is capped by law, and Medicaid, which is expected to grow rapidly. Under current law, nominal spending on Medicaid benefits for nonelderly recipients is projected to rise from $114 billion in 2004 to $184 billion in 2010 because of rising prices for health care, greater consumption of services, and (to a lesser extent) increased enrollment. CBO's projections assume a fairly healthy economy: an unemployment rate averaging about 5.2 percent per year and gross domestic product growing at a real rate of between 3 percent and 4 percent annually.

If the economy weakened, participation in means-tested programs might be expected to increase. However, during the most recent recession (in 2001), only the Food Stamp program showed a significant shift from earlier trends in participation, with the number of recipients beginning to grow after years of decline. For the other programs discussed here, the changes in participation that occurred during the recession could have reasonably been expected from previous trends or legislative changes.

Participation in SSI is unlikely to vary significantly with shifts in the economy. To receive benefits from the program, nonelderly people must document that they are blind or disabled--conditions that are unlikely to be affected by the state of the economy.

Medicaid also serves nonelderly disabled people, although the majority of participants under age 65 are children or nondisabled adults. As with SSI, the number of disabled Medicaid participants is not likely to be extremely responsive to changes in the economy. However, the number of children and nondisabled adults might increase during a recession--as it did in the 2001 recession--as people lose jobs and thus their health care coverage.

EITC participation remained fairly constant during the 2001 recession, perhaps because of two offsetting components of the eligibility requirements. On the one hand, people must have some earnings to receive the EITC, and the number of earners in the economy usually declines during a recession. On the other hand, for workers to receive the credit, their earnings must be below a threshold level, and a downturn could depress earnings for some people.

The number of participants in TANF declined during the 2001 recession, although the decrease was smaller than in previous years. That outcome suggests that TANF participation may not be very sensitive to cyclical variations.

The various changes in participation that occurred in 2001 might not be indicative of the changes that would take place during a more prolonged or deeper economic downturn. Over the long term, participation in the means-tested programs discussed here will largely be determined by trends in demography, the distribution of income, and the state of the economy. Federal spending on those programs will likewise be affected by such trends, although the effects could be dwarfed by the future growth of Medicaid spending.