Skip to main content

Safety Net Effective at Fighting Poverty But Has Weakened for the Very Poorest

As mounting job losses threaten to push more Americans into poverty and make poor families still poorer, a new examination of the public benefits system finds that it is more effective in reducing poverty than previously known but has become less effective over the past decade in protecting Americans from deep poverty.

To paint a fuller picture of the effect of the public benefits system (sometimes referred to as the “safety net”) in reducing hardship, this analysis adopts changes to the Census Bureau’s official poverty measure recommended by the National Academy of Sciences (NAS), as explained in the methodological appendix. It also uses data from the U.S. Department of Health and Human Services on the receipt of public benefits that are more complete than Census data.

The good news is that the safety net reduces poverty substantially and is more effective at reducing poverty than has generally been recognized. When both broad social insurance benefits such as Social Security and programs targeted on low-income people such as food stamps are considered, the safety net lifts tens of millions of people out of poverty. More specifically, in 2005 (the latest year for which comprehensive data are available), the safety net as a whole:

  • Cut the number of Americans living in poverty by nearly half (44 percent), lifting 31 million people above the poverty line.[1]
  • Reduced the severity of poverty for those who remain poor, increasing their average disposable income from 29 percent of the poverty line to 64 percent.
  • Helped protect Americans from the deepest extremes of poverty, cutting by 7.3 million — or more than three-quarters — the number of children living below half the poverty line. It also lifted 8.0 million children above three-quarters of the poverty line. (This analysis uses a poverty line equal to about $21,400 in 2005 for a couple with two children in a community with average housing costs, consistent with NAS recommendations.)
  • Was more effective at lifting children in less-deeply-poor families from just below the poverty line to above the poverty line than it had been a decade earlier. Among children whose non-benefit income was between 75 percent and 99 percent of the poverty line, public programs lifted 65 percent above the poverty line in 2005, up from 51 percent in 1995.

The bad news is that the safety net has weakened over the last decade for families with children that have the lowest incomes and are in greatest need of help due to joblessness or other crises. In 2005, the safety net as a whole:

  • Protected a smaller share of children from deep poverty than it used to. In 1995, the safety net lifted above half the poverty line 88 percent of children whose family incomes were lower than that before counting safety net benefits. By 2005, this percentage had declined to 76 percent. If the safety net had been as effective at keeping children out of deep poverty in 2005 as it was in 1995, there would have been 1.1 million very poor children in 2005; instead, there were 2.4 million.
  • Protected fewer jobless workers from deep poverty than it used to. Among very poor unemployed workers looking for work in any given week, the safety net lifted 60 percent above half of the poverty line in 2005, down from 70 percent of very poor unemployed workers in 1995. [2]

Since these data were collected, the economy has entered a major recession, and Congress enacted the American Recovery and Reinvestment Act, designed to boost economic growth and ameliorate the harshest impacts of the recession on struggling families. The recovery package included many provisions that strengthen the safety net, though in most cases the improvements are designed to be temporary. These include a temporary boost in food stamp benefits, temporary expansions in the Earned Income Tax Credit and the Child Tax Credit, new incentives for states to make their unemployment insurance systems more accessible to jobless workers, and new funding for states that see an increase in the number of families receiving basic cash assistance through TANF programs and states that expand short-term help and subsidized employment programs for poor families.

These provisions will soften the impact of the recession on the extent and depth of poverty. (A previous Center analysis projects that the expansions in the EITC, Child Tax Credit, and the new Making Work Pay tax credit will stop 1 million children from falling below the poverty line. [3]) When the recession abates, it will be important to measure precisely the impact of these temporary measures and consider what longer-lasting improvements should be made in the safety net.

Programs Lift Millions of Americans Out of Poverty and Deep Poverty

There are two principal categories of income-support programs: those that provide benefits more or less regardless of income and those that limit assistance to people with low or modest incomes.

The first category, sometimes called “universal programs,” includes the major social insurance programs, such as Social Security and unemployment insurance. Social Security alone lifts nearly 22 million Americans above the poverty line, including nearly 14 million seniors, who rely heavily on Social Security income in their retirement years. Unemployment insurance lifted about another 1.2 million Americans above the poverty line in 2005. [4] The universal category also includes other widely available benefits such as the Child Tax Credit, which lifted 2 million people above the poverty line in 2005, including 1.2 million children.[5]

Programs in the second category are often referred to as “means-tested” programs. They include programs like Supplemental Security Income (SSI) for low-income seniors and people with disabilities; cash assistance programs funded by the Temporary Assistance for Needy Families (TANF) block grant; programs that provide non-cash benefits like food stamps (now called the Supplemental Nutrition Assistance Program), housing assistance, and the Earned Income Tax Credit (EITC). They, too, play a large role in reducing the extent and severity of poverty. In 2005, means-tested benefits lifted 14 million low-income Americans above the poverty line.

Figure 1 summarizes the poverty-reducing effects of major means-tested benefits. [6] For example, the EITC lifted 5.1 million Americans above the poverty line in 2005. This includes 2.6 million children — more than any other single program.

It should be noted that housing programs reach relatively few households (due to limited funding) but have strong antipoverty effects for those they do reach. In 2005, about 10 million Americans in more than 4 million low-income households received housing assistance, well below the number that received food stamps, SSI, EITC, or any of the universal programs we examined. [7] Yet, as a share of those assisted, housing assistance reduced poverty more effectively than any other program: 44 percent of those who received housing assistance would be considered poor without it but were above the poverty line when their housing benefit was counted.

The means-tested safety net also plays a crucial role in easing deep poverty, particularly for families with children. In 2005, means-tested benefits cut by 56 percent the share of the nation’s children who were below three-quarters of the poverty line; they reduced the share of the nation’s children who were below half the poverty line by 72 percent.[8]

The appendix tables provide program-by-program details on the numbers of people lifted above half and three-quarters of the poverty line. Among other things, it shows that the Food Stamp Program does more than any other single program to protect children from deep poverty; in 2005, food stamps lifted 1.7 million children younger than 18 above half the poverty line and 2.5 million children above three-quarters of the poverty line. It also shows that after Social Security (which is not means-tested), SSI lifts the most Americans of all ages above half the poverty line — 3.8 million in 2005.

Effectiveness Against Deep Poverty Has Weakened

Unfortunately, over the past decade, means-tested benefits have grown less effective at shielding children from deep poverty. This largely reflects dramatic declines in the antipoverty effects of cash assistance programs financed by the TANF block grant, as well as reductions in the effectiveness of food stamps that were quite pronounced in the late 1990s and early 2000s (though food stamps' effectiveness has improved since then).

  • In 1995, all means-tested benefits together lifted 87 percent of children who would otherwise have been below half of the poverty line out of deep poverty. By 2005 this figure had dropped to 72 percent.
  • In 1995, AFDC (which preceded TANF) lifted 62 percent of children who would otherwise have been below half of the poverty line out of deep poverty. By 2005 this figure for the TANF program was just 21 percent.
  • In 1995, the Food Stamp Program lifted out of deep poverty 61 percent of children who would otherwise have been below half of the poverty line. By 2005 this figure had dropped to 42 percent.

As the means-tested safety net for the poorest families has weakened, the number of children in deep poverty has grown significantly. The number of children living below half of the poverty line rose from 1.4 million in 1995 to 1.7 million in 2000 and to 2.4 million by 2005, nearly a 75 percent increase. The percentage of all children in deep poverty increased from 1.9 percent to 3.2 percent over this decade.


In fact, if means-tested benefits were as effective at keeping children out of deep poverty in 2005 as they had been in 1995, some 1.2 million fewer children would have lived below half of the poverty line in 2005 — a reduction of more than half.

The erosion of the safety net’s antipoverty effectiveness was concentrated at the very bottom of the income scale. For example, among children whose non-safety-net income was below 100 percent of the poverty line, the share lifted above the poverty line by all public benefits held steady at 37 percent from 1995 to 2005. Among children whose non-safety-net income was below 75 percent of the poverty line, the share lifted above three-quarters of the poverty line dropped from 67 percent to 59 percent, a decline of 8 percentage points. And for children whose non-safety-net income was below 50 percent of the poverty line, the share lifted above half of the poverty line dropped from 88 percent to 76 percent, a decline of 12 percentage points.

While the safety net for the poorest families has weakened, support for low-income working families with incomes just below or modestly above the poverty line has grown more robust. The main reasons are the large expansions in the EITC enacted in 1990 and 1993, expansions in the refundable portion of the Child Tax Credit in 2001, and the extension of public health insurance through Medicaid and the Children’s Health Insurance Program (CHIP) for low-income working families.

As a result, the effectiveness of the safety net in lifting less-deeply-poor children above the poverty line increased over this period. Among children whose non-benefit income was between 75 percent and 99 percent of the poverty line, public programs lifted 65 percent above the poverty line in 2005, up from 51 percent in 1995.

What's Behind the Weakening Safety Net for the Poorest Families?

During the 1990s, federal and state policy changes made basic cash assistance programs for the poorest families with children less accessible while increasing supports for low-income working families. During the mid- to late 1990s, a strong economy coupled with these policy changes resulted in large increases in employment among low-skilled workers in general and single mothers in particular. Poverty among families with children fell.

Even as some families climbed above the poverty line, however, there were warning signs that others were falling deeper into poverty. Most of these families had periods of joblessness, due both to the nature of low-wage work and to individual crises and barriers to employment. Making matters worse, the economic boom halted after 2000 and the brief recession of 2001 was followed by years of labor market weakness. From 2000 to 2005, employment rates fell for single parents, and both poverty and deep poverty among children rose.

The largest single reason why the safety net protected fewer children against deep poverty was the loss of TANF. Over the 1996-2005 period, TANF cash assistance programs served a shrinking share of very poor families with children. The number of children shielded from deep poverty by TANF cash assistance dropped by 1.6 million — from 2.2 million in 1995 to 645,000 in 2005.

Department of Health and Human Services data show a similarly steep decline in participation in TANF cash assistance as a percentage of families that are poor enough to qualify for it (and that meet other TANF eligibility requirements). In the 1980s and early 1990s, about 80 percent of families eligible for cash assistance through AFDC received assistance. By 2005, the last year for which data are available, this figure (for TANF cash assistance) had dropped in half, to 40 percent. The decline chiefly affected families at high risk of deep poverty; in most states, a family seeking to qualify for TANF must have cash income well below the poverty line (below half the poverty line in the majority of states).

The reasons for this sharp drop in receipt of TANF cash assistance vary across states. They include a range of policies and administrative practices that led many poor families to leave the program even when they did not have a job, to be discouraged from applying at all, or to fail to successfully complete the application process. Many families lost assistance because of strict welfare-to-work rules and policies that terminated assistance to families that could not meet program requirements. While requirements to attend orientation sessions or seek work before applying for aid may seem reasonable, some families in the midst of a serious crisis and those with mental health or other health issues may be unable to comply with them. Research has consistently shown that families that lose assistance due to sanctions often have significant mental health or other health issues and other barriers to employment that may inhibit their ability to meet various requirements.

Similarly, many states placed new requirements on families applying for assistance. In addition, in some states, time limits on assistance have barred some very poor families from receiving assistance.[9]

In addition to creating TANF, the 1996 welfare legislation made large cuts to food stamps and other programs that weakened the safety net for the poorest families. These included across-the-board food stamp cuts that affected millions of low-income Americans — including children, the elderly, and people with disabilities — as well as targeted eligibility changes that denied benefits to many legal immigrants, unemployed childless adults, and others. While some of the food stamp cuts were later restored, the majority remained in place in 2005. (Some of the remaining food stamp cuts were reversed in the 2007 farm bill, but others remain in place. The recovery package temporarily increased food stamp benefits.)

States have also largely eliminated cash assistance for unemployed childless individuals who do not qualify for unemployment insurance. Most states eliminated their general assistance programs — formerly the safety net of last resort for this group of people — in the late 1980s and early 1990s, except for programs for people with disabilities. Also, with some exceptions, the 1996 welfare law barred out-of-work childless adults age 18-50 from receiving food stamps for more than three months at a time. (The recent recovery act has temporarily relaxed this rule.)


The safety net protects millions of Americans from poverty and, especially, from deep poverty. In 2005 it protected 31 million people from poverty and kept 34 million from slipping below half of the poverty line. Nonetheless, this protection became weaker for children in the poorest families from 1995 to 2005.

Since 2005, much has changed. The unemployment rate has doubled as a result of the recession. The economic recovery legislation enacted early in 2009 provided a strong response, although the response is temporary, and joblessness and poverty tend to remain elevated for years after a recession formally ends. Even after joblessness eventually returns to its pre-recession level, the question of how to better protect the nation’s poorest families and children will remain.

Click here to read the methodological appendix (9pp.)

End Notes

[1] As explained in the appendix, this analysis considers a person to have been lifted out of poverty by a particular type of income if his or her family’s income from all sources — including the specified type of income — is above the poverty line but would be below the poverty line if that type of income were excluded.

[2] These figures are for workers who spent part of the year unemployed and looking for work. To estimate the average weekly number of unemployed workers in a given year, we weighted the data for each unemployed adult by the percentage of weeks during the year for which he or she reported having been unemployed.

[3] See Arloc Sherman, “Recovery Agreement Temporarily Expands Child Tax Credit for Large Numbers of Children in Every State,” Center on Budget and Policy Priorities, February 12, 2009.

[4] The antipoverty effects of unemployment insurance can rise substantially in times of high unemployment, particularly in years when Congress provides additional weeks of benefits through emergency legislation. For example, in 2002, a year in which Congress provided extended unemployment insurance benefits, unemployment benefits lifted 1.9 million Americans above the poverty line, more than twice as many as in 2000 (819,000) and well above the number in 2005 (1.2 million).

[5] The Child Tax Credit provides a partially refundable federal income tax credit of up to $1,000 per child. In 2005, the credit was available to families that earned over $11,000. (It was expanded temporarily in the American Recovery and Reinvestment Act of 2009 so that it is now available to families that earn over $3,000.) A married family with two children loses eligibility for the credit when its earnings rise above $150,000. (The cutoff for a single-parent family with two children is $115,000.)

[6] Numbers in Figure 1 do not sum to the total for all means-tested benefits because we count individuals who need help from a combination of programs to reach the poverty line as “lifted above the poverty line” by each program but count them only once in the total. In addition, in this analysis, “all means-tested benefits” includes the estimated value of energy-assistance subsidies, which Figure 1 does not show separately.

[7] In these data, about 36 million people in 2005 lived in families with food stamps. About 17 million lived in families with SSI. About 56 million lived in families with EITC income.

[8] In this analysis, figures for people in deep poverty (below 75 percent or 50 percent of the poverty line) exclude people in families with negative incomes (i.e., incomes below zero). Some of these families are likely to have substantial resources but to report temporary business losses that take their income for the year into negative territory. The number of people with negative cash income is very small (fewer than 300,000 in 2005) and has virtually no effect on our estimates of people protected from poverty and deep poverty.

[9] The HHS estimates of declining participation rates in TANF do not include those affected by TANF time limits. Time limits are an additional factor contributing to the weakening of the safety net.