urban institute nonprofit social and economic policy research

Extreme Poverty Rising, Existing Government Programs Could Do More

Read complete document: PDF

PrintPrint this page
Share on Facebook Share on Twitter Share on LinkedIn Share on Digg Share on Reddit
| Email this pageE-mail
Document date: April 01, 2002
Released online: April 01, 2002

Number B-45 in Series, "New Federalism: National Survey of America's Families"

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

Many have heralded welfare reform as an unqualified success. More single parents are working, and welfare caseloads have plummeted. The official poverty rate decreased from 13.7 percent in 1996 to 11.8 percent in 1999. While many commentators debate the causes of these improvements—citing the strong economy, federal and state policies designed to "make work pay," and welfare reform—few dispute the strong positive outcomes.1

Less well known, however, is the unfortunate story buried within these positive outcomes. When all types of income are taken into account, extreme poverty (income below 50 percent of the federal poverty level) has increased. Despite phenomenal growth in the U.S. economy in the late 1990s, more children lived in single-parent, extremely poor families in 1998 than in 1996. This surprisingly negative outcome largely reflects an increase in the number of low-income families that have either left, or chosen not to enroll in, government support programs (including cash welfare and food stamps).

The reasons for the decline in program participation are difficult to pinpoint. Changing attitudes about seeking help from the government, the increasing complexity of program rules, and states' new welfare programs have no doubt all played a role in reducing participation. The decline suggests that many families are not getting the assistance they need and may be facing greater hardship.

This brief sets out to understand how decreased program participation has affected poverty since welfare reform. It reviews changes in poverty from 1996 to 1998 and compares families' current economic status with a scenario that assumes full participation in government support programs.

The findings show that the government offers a stronger safety net than is delivered. In 1998, if all families with children participated in the post-reform government safety net programs for which they qualified, poverty would have declined by more than 20 percent and extreme poverty by 70 percent. This decline provides a strong rationale for changing existing programs to provide "family-friendly" delivery systems and more standardized eligibility requirements. By increasing participation in programs already in place, policymakers can take an important step toward improving the economic well-being of children.

Data and Methods

The analysis uses the 1997 and 1999 National Surveys of America's Families (NSAF) and the Urban Institute's Transfer Income Model, version 3 (TRIM3). The NSAF is a large, nationally representative survey of the civilian, noninstitutionalized population under age 65 that overrepresents low-income families. Like similar government surveys, each round of the NSAF asked families about income received during the prior year (1996 and 1998). The TRIM3 model corrects the NSAF's underreporting of government benefits (a condition that plagues all surveys in this area),2 adds the value of food stamps and the earned income tax credit (EITC), imputes federal payroll and income taxes, and estimates annual out-of-pocket child care expenses.

The TRIM3 model corrects for the underreporting of Supplemental Security Income (SSI), Temporary Assistance for Needy Families (TANF), and food stamp benefits by applying individual state rules to determine eligibility for each individual and family in the survey file.3 The model then compares benefit eligibility with income sources reported by NSAF families and adds benefits for families that failed to report them by matching detailed control totals developed from state and federal administrative data.

The augmented NSAF data provide estimates of families' total disposable income—defined as all sources of cash income plus the value of food stamps and the EITC, minus federal taxes and out-of-pocket child care expenses.4 We report poverty estimates using cash income alone (following the official Census Bureau estimate) and using an alternative measure based on disposable income.5

The TRIM3 program eligibility estimates allow us to calculate the change in income that would occur if all families took the benefits available to them. The "full-participation scenario" holds earnings and nonwelfare sources of income constant, but recalculates eligible families' SSI, TANF, and food stamp benefits. Owing to data limitations, the estimates do not take into account ineligibility because of immigrant status. Thus, the estimates of poverty reduction achieved by full program participation are somewhat overstated.6

Changes in the Government Safety Net

Federal and state welfare reforms enacted since 1996 have affected families' benefit eligibility and participation in a number of ways. Some of the changes positively affected participation. For example, states that raised the amount that individuals can earn and still receive TANF benefits increased the number of eligible families and encouraged participation. In contrast, state diversion strategies, such as requiring substantial proof of job search prior to enrollment, may have discouraged program participation.

The 1996 federal reforms to other parts of the safety net also affected benefit eligibility and participation. In addition to restricting SSI eligibility for children, the legislation eliminated SSI and food stamp benefit eligibility for legal, noncitizen immigrants. (Subsequent legislation enacted in 1997 and 1998 restored eligibility for some noncitizen groups.) The reforms also cut food stamp benefits by lowering the payment standard, freezing the standard deduction, increasing the amount of income counted against benefits, and limiting food stamp eligibility to three months for able-bodied adults without dependents working less than 20 hours per week. Finally, the legislation gave states more funding and consolidated programs for child care assistance.

The changes made the already complex safety net system even more difficult to navigate. Low-income citizen families can apply for cash assistance through SSI (if there is a severely disabled family member) or TANF (if they have not reached a benefit time limit). Low-income families can also apply for in-kind assistance through the Food Stamp program and for child care and housing subsidies. Each of these programs, however, has very different minimum income requirements, rules about work, and asset tests (see Zedlewski 2002). In addition, unlike many benefits, child care and housing assistance are not guaranteed and are awarded on a first-come, first-served basis. Medical assistance may be available to low-income families, but most states limit coverage for parents to those that have cash assistance or have recently left TANF. Coverage for their children, however, is more broadly available. Working families receive assistance through the refundable EITC.7 Legal immigrant families may or may not be eligible for some of these safety net benefits depending on the program, their state of residence, and (sometimes) their work history. As the next section shows, the drop in participation following the reforms led to higher extreme poverty among families with children.

The Poverty Story: 1996–1998

Our examination of participation's effect on family well-being looks at two types of poverty measures. The first is the official U.S. measure of poverty, which compares families' cash income to the poverty threshold but excludes two of the most important government benefits—food stamps and the EITC. Because of the importance of food stamps (a "near cash" benefit) in reducing poverty, we also extend the cash-income measure to include these benefits.

The second, more comprehensive measure includes all government assistance, but deducts federal payroll and income taxes as well as out-of-pocket child care expenses from income to adjust for the different needs of working and nonworking families.8 The official U.S. "cash-income poverty" measure yields higher poverty levels than the more comprehensive "disposable-income poverty" measure because it omits significant sources of income. The disposable income poverty measure better reflects families' economic status because it takes into account their buying power.

Changes in both the cash and disposable income measures of poverty, reported below, show a disturbing trend. While overall family poverty declined significantly, extreme poverty for persons in all families with children and those in single-parent families shrunk much less, and by the disposable-income measure it rose. The difference between having income below one-half of the poverty level and 50–100 percent of the poverty level makes a big difference in economic hardship. A family of three is poor when its income falls below $14,269; it is extremely poor when its income falls below $7,135 (in 2001 dollars).


Poverty is lower when more sources of government assistance are taken into account (figure 1). For example, 20 million persons were poor in 1996 according to the disposable-income measure, compared with 23.7 million persons by the official poverty measure. Food stamps accounted for a large share of this difference (2.3 million fewer poor persons). The net effect of the refundable EITC decreased poverty by another 1.4 million persons, despite deductions for other tax payments and out-of-pocket child care expenses.

Figure 1

Both cash-income and disposable-income poverty declined between 1996 and 1998 for individuals in families with children, including those living in single-parent families. Based on cash income, about 3.8 million fewer persons in families with children were poor in 1998 than in 1996. The decline in poverty for all persons is slightly less when food stamps are included in income (3.5 million persons) and when total disposable income is used to measure poverty (3.3 million).

The change in poverty for persons in single-parent families follows a similar pattern. The cash measure of poverty shows the largest decline between 1996 and 1998 (2 million persons). The measure that includes food stamps reveals a decline of 1.7 million persons. A reduction in the use of food stamps mitigated the improvement in cash-income poverty for persons in single-parent families.

Extreme Poverty

Factoring in government assistance has a dramatic effect on extreme poverty levels (figure 2). The number of persons in families with children counted as extremely poor is more than halved when food stamps are considered income (3.2 million compared with 7.5 million), and even fewer persons are classified as extremely poor based on disposable income. Food stamp benefits have an especially large effect on extreme poverty figures because benefits are significant at this income level and a large share of these families participate.

Figure 2

According to the cash-income measure, extreme poverty declined substantially between 1996 and 1998 among persons living in families with children. About 6.3 million persons lived in extreme poverty in 1998, 1.2 million fewer than in 1996. By this measure, fewer persons in single-parent families also lived in extreme poverty in 1998 than in 1996, although the reduction in extreme poverty was relatively small.

These declines in extreme poverty shrink as more sources of income are taken into account. For example, when food stamps are included as income, extreme poverty among families with children was unchanged between 1996 and 1998. The disposable income poverty measure shows a slight increase in extreme poverty for all families with children. Fewer families in extreme poverty participated in food stamps in 1998 than in 1996, and increased work-related expenses (taxes and child care) further reduced disposable incomes.

Disposable-income measures show extreme poverty among persons living in single-parent families—those most at risk of needing welfare—increasing more substantially. About 300,000 more persons in single-parent families lived in extreme poverty in 1998 than in 1996. The rise reflects shifts in benefits and expenses. Many single parents who moved into the labor market during this period left welfare and the Food Stamp program. Their earnings, however, were not sufficient to offset the income loss resulting from the drop in food stamp benefits. In addition, while many of these parents became eligible for the EITC, they also began to pay payroll taxes and child care expenses. In sum, disposable income declined sufficiently to move more parents and their children into extreme poverty.

One other change in program policy may have also affected the well-being of extremely poor families. While our analysis does not deduct out-of-pocket health care costs, some evidence suggests that these costs (or at least the risk of higher costs) increased. In particular, the NSAF showed that Medicaid coverage declined significantly among extremely poor single parents, from 48 percent in 1997 to 41 percent in 1999. While the uninsured rate of these parents did not change significantly (because of a slight gain in private health coverage), the decline is notable because Medicaid provides the fullest package of coverage with limited, or no, out-of-pocket costs.

Full Participation Would Brighten the Poverty Story

The full-participation analysis shows that poverty and extreme poverty would have been substantially lower if families took full advantage of the government safety net (figure 3). For example, based on disposable income, if all families with children participated fully in the government programs they qualified for, 3.8 million fewer persons would have been poor and 2.0 million fewer persons would have been extremely poor compared with the actual 1998 data. The potential decline in poverty is substantially smaller when using the less-comprehensive cash-income measure (1.6 million).

Figure 3

For persons living in single-parent families, the decline in disposable-income poverty resulting from full participation would have been even more substantial. About 1.8 million fewer persons in single-parent families had incomes below poverty under the full-participation scenario than in the actual 1998 data. This difference could have reduced the number of poor individuals in single-parent families from 9.9 million (shown in figure 1) to 7.1 million. And 1.4 million fewer persons would have been in extreme poverty—more than enough to offset the 300,000-person increase in extreme poverty actually experienced by single-parent families between 1996 and 1998.

The potential poverty decreases are particularly notable for those in extreme poverty (figure 4). The full-participation scenario shows extreme poverty declining about 70 percent and poverty dropping 20 percent in 1998. Government transfer programs are much more likely to move families above 50 percent of poverty than to lift families out of poverty. TANF, for example, pays a maximum cash benefit equal to about 36 percent of poverty for a family of three, on average. Families that combine substantial earnings with food stamps and the EITC, however, can move above poverty, as long as their child care costs do not eat away a substantial portion of their income.

Figure 4

Of course, full participation in government programs would increase government program costs. The TRIM3 model estimates that food stamps, TANF, and SSI would have cost $23 billion more in 1998 (in 2001 dollars) if all low-income families with children took advantage of these programs' benefits. Estimates based on full participation (even for families eligible for low benefits who might find the "cost" of participating particularly high) may somewhat overstate these costs. In addition, potential increases in cash TANF benefits ($10 billion) arising from the greater demand may have restricted states from using TANF dollars to provide some new benefits for working poor families (such as child care). Participation in TANF would also count against benefit time limits, potentially deterring participation by working families that could combine a small welfare check with earnings. Still, the rise in government costs would be significant and must be weighed against the high costs of poverty to individuals and to society.

Hardships Facing Families in Poverty

Families with children living in extreme poverty find it difficult to provide basic food and shelter for their families. Indeed, about 4 in 10 extremely poor families report two or more problems providing food for their families, and 3 in 10 report difficulties paying rent or utilities in the prior 12 months (table 1). Almost a quarter report living in crowded housing (more than 2 persons per bedroom), and 1 in 10 have lived without a telephone for one month or longer.

Table 1: Type and Number of Economic Hardships Reported by Families with Children, by Disposable Income Level (percent), 1999
  Extremely Poor Poor Near Poor
Economic Hardship (Below 50% of FPL) (50%-100% of FPL) (101-150% of FPL)
   Two or more food problems 42 41 32*
   Crowded housing 24 24 18
   Housing costs exceed 50% of income 46 17* 7*
   Unable to pay rent or utilities 30 32 29
   Moved in with others 5 6 3
   Without a phone for month or more 10 6 6
   Without car 40 41 20
Number of Hardships:
   One 25 35* 33*
   Two 31 23* 20*
   Three or more 30 24 13
Source: The 1999 National Survey of America's Families.
Note: Measures of poverty are based on disposable income, defined as all cash plus food stamps and the EITC, less taxes and child care expenses.
FPL = federal poverty level
*Significantly different from results for extremely poor at the 90 percent confidence level.

Changes in government programs that increase program participation could substantially ease economic hardship. The types of hardships measured by the NSAF were about the same for individuals living in poor and extremely poor families. The number of hardships reported by the less poor, however, was significantly lower (47 percent of families with incomes between 50 and 100 percent of the poverty level reported two or more hardships, compared with 61 percent of extremely poor families). In addition, near-poor families (income of 100–150 percent of the poverty level) reported significantly fewer hardships than those living in poverty. These findings suggest that full participation in government assistance programs, which would lift many families out of poverty into the near-poor category, would ease the economic hardships currently facing poor families with children.


The 1996 federal reforms shifted responsibility for cash assistance to the states, cut food stamps and SSI benefits, and expanded child care assistance. At least in part because of the changes in the safety net, participation in government assistance programs has declined. At the same time, extreme poverty based on disposable income increased, especially among single-parent families. Potential reductions in poverty that could have been achieved by combining increased earnings with government assistance were diminished.

As noted, it is impossible to isolate the reasons for low and declining program participation. Some of the drop likely reflects changing family attitudes about using government assistance. Studies also show that the complexity of program rules and fear of stigma have played a role in discouraging participation (see, for example, McConnell, Ponza, and Cohen 1999). A recent report by the U.S. General Accounting Office (2001), which documents the wide array of government assistance programs and income eligibility policies, concludes that the federal government should support state and local demonstration programs that test alternative benefit delivery systems.

The recent weakening in the economy will likely put more families at risk of poverty, making program changes that welcome, rather than discourage, families from enrolling in appropriate assistance programs even more essential. While the demonstration programs recommended by the GAO may ultimately identify the most effective benefit-delivery strategies, incremental improvements could be implemented immediately. Standardizing income eligibility, work rules, asset tests, and treatment of different family members' incomes for TANF families as well as for working poor families could raise participation levels. In addition, information sharing across the agencies that administer these programs—including human services offices, housing authorities, the public health service, and the social security office—would likely streamline the eligibility maze confronting low-income families. States must ensure that TANF rules designed to encourage work do not discourage families from using interim government assistance. Public and private sector groups also have a role to play—effective outreach that educates families about public assistance and facilitates their participation would help. At the agency level, efforts to develop family-friendly offices that can accommodate the schedules of working parents are needed.

Our results demonstrate that private and public sector efforts to increase program participation could reduce poverty and economic hardship.

Alternative ways of increasing the disposable incomes of families—such as minimum wage increases and investments in education and training—should not be forgotten. In these tough economic times, however, these strategies seem less saleable than steps that improve the safety net already offered by the government.


1. Recent positive outcomes since welfare reform have led some to argue that studies predicting higher levels of child poverty as a result of welfare reform were wrong. However, these studies all assumed that federal benefit time limits were fully implemented, that the economy performed as it did in 1995, and that proposed benefit cuts were permanent. We now know that assumptions about the economy proved to be too pessimistic (at least through 2000) and that some of the immigrant benefit cuts have been restored. Further, federal benefit time limits will only gradually roll out across the states between 2001 and 2003.

2. Wheaton and Giannarelli (2000) show a deterioration in reports of government assistance in the March Current Population Survey, the basis for the federal government's poverty estimates. The use of corrected NSAF data for this analysis eliminates the chance that a change in income is due to higher levels of underreporting of government assistance in one year relative to the other.

3. The model uses the same procedures as a caseworker to determine eligibility. These procedures are explained fully in Giannarelli, Morton, and Wheaton (forthcoming).

4. The NSAF asked families how much out-of-pocket expenses they incurred for child care in the past month. We use these reports to extrapolate annual child care spending for employed families. Note that the use of out-of-pocket child care spending indirectly accounts for increased government spending on child care subsidies because these subsidies reduce families' out-of-pocket costs.

5. Note that our estimates omit the value of some of the nutrition programs for low-income families (school lunch and school breakfast) and the value of housing assistance (because of incorrect reports of assistance in round one of the NSAF).

6. For example, based on Current Population Survey estimates of the number of legal immigrants ineligible for food stamps in 1998, immigrant benefits are overstated by 7 percent.

7. Congress recently increased the child tax credit and made it refundable. This credit will provide additional assistance to families with children, but is outside the time period analyzed here.

8. See Citro and Michel (1995) for a full description of the alternative poverty measure recommended by the National Academy of Sciences. They also recommended deducting out-of-pocket health expenses from income.


Citro, Constance, and Robert Michel. 1995. Measuring Poverty, A New Approach. Washington, D.C.: National Academy Press.

Giannarelli, Linda, Joyce Morton, and Laura Wheaton. Forthcoming. "Using the NSAF in the TRIM3 Model." Washington, D.C.: The Urban Institute. Assessing the New Federalism Methodology Report.

McConnell, Sheena, Michael Ponza, and Rhoda Cohen. 1999. "Report on the Pretest of the Reaching the Working Poor and Poor Elderly Survey," vol. 1. Washington, D.C.: Mathematica Policy Research.

U.S. General Accounting Office. 2001. Means-Tested Programs: Determining Financial Eligibility Is Cumbersome and Can Be Simplified. Washington, D.C.: General Accounting Office Report 02-58.

Wheaton, Laura, and Linda Giannarelli. 2000. "Underreporting of Means-Tested Transfer Programs in the March CPS." American Statistical Association 2000 Proceedings of the Sections on Government Statistics and Social Statistics: 236–241.

Zedlewski, Sheila. 2002. "Family Incomes: Rising, Falling, or Holding Steady?" In Welfare Reform: The Next Act, edited by Alan Weil and Kenneth Finegold. Washington, D.C.: Urban Institute Press.

About the Authors

Sheila R. Zedlewski is director of the Urban Institute's Income and Benefits Policy Center. Her research deals with welfare reform, low-income program participation, and poverty. Her recent articles deal with the linkage between welfare reform and food stamp participation, and the changing characteristics of families participating in cash assistance programs.

Linda Giannarelli is a senior research associate in the Urban Institute's Income and Benefits Policy Center. Her research focuses on government programs for low-income families. She directs the Urban Institute's Welfare Rules Database project and is the principal investigator for the development and maintenance of the TRIM3 microsimulation model. She recently led the effort to add a child care component to TRIM3, which was used to estimate the number of children and families eligible for federally funded child care subsidies.

Joyce Morton is a senior programmer/analyst and one of the developers of the current version of the TRIM3 microsimulation model. Ms. Morton has extensive knowledge of the core programs modeled by TRIM3. She also has in-depth knowledge of large data sets including the Census Bureau's Current Population Survey (CPS), the Survey of Income and Program Participation (SIPP), and the National Survey of America's Families (NSAF).

Laura Wheaton is a research associate in the Income and Benefits Policy Center at the Urban Institute.

About the Series

This series presents findings from the 1997 and 1999 rounds of the National Survey of America's Families (NSAF). Information on more than 100,000 people was gathered in each round from more than 42,000 households with and without telephones that are representative of the nation as a whole and of 13 selected states (Alabama, California, Colorado, Florida, Massachusetts, Michigan, Minnesota, Mississippi, New Jersey, New York, Texas, Washington, and Wisconsin). As in all surveys, the data are subject to sampling variability and other sources of error. Additional information on the NSAF can be obtained at http://newfederalism.urban.org.

The NSAF is part of Assessing the New Federalism, a multiyear project to monitor and assess the devolution of social programs from the federal to the state and local levels. Alan Weil is the project director. The project analyzes changes in income support, social services, and health programs. In collaboration with Child Trends, the project studies child and family well-being.

This analysis and paper were funded by The David and Lucile Packard Foundation.

The ANF project has also received funding from The Annie E. Casey Foundation, the W.K. Kellogg Foundation, The Robert Wood Johnson Foundation, The Henry J. Kaiser Family Foundation, The Ford Foundation, The John D. and Catherine T. MacArthur Foundation, the Charles Stewart Mott Foundation, The McKnight Foundation, The Commonwealth Fund, the Stuart Foundation, the Weingart Foundation, The Fund for New Jersey, The Lynde and Harry Bradley Foundation, the Joyce Foundation, and The Rockefeller Foundation.

Topics/Tags: | Employment | Families and Parenting | Poverty, Assets and Safety Net

Usage and reprints: Most publications may be downloaded free of charge from the web site and may be used and copies made for research, academic, policy or other non-commercial purposes. Proper attribution is required. Posting UI research papers on other websites is permitted subject to prior approval from the Urban Institute—contact publicaffairs@urban.org.

If you are unable to access or print the PDF document please contact us or call the Publications Office at (202) 261-5687.

Disclaimer: The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders. Copyright of the written materials contained within the Urban Institute website is owned or controlled by the Urban Institute.

Email this Page