Skip to main content

Safety Net for Poorest Weakened After Welfare Law But Regained Strength in Great Recession, at Least Temporarily

A Decade After Welfare Overhaul, More Children in Deep Poverty

A close examination of the effectiveness of U.S. safety net programs, using more complete income figures than are widely available, reveals important shifts in recent years.  In the decade after Congress altered the welfare system in the mid-1990s, the safety net grew more effective at assisting working-poor families with children but less effective at protecting Americans from deep poverty — that is, at lifting their incomes above half the poverty line — and children’s deep poverty increased.  During the Great Recession, safety net policies for both the poor and deeply poor grew much stronger, bolstered by temporary recovery initiatives, and prevented what likely would have been a large surge in deep poverty.  But the expiration of those temporary initiatives could push deep poverty upward again.

To paint a fuller picture of the public income support system for low-income families, this analysis uses an expanded poverty measure similar to the federal government’s Supplemental Poverty Measure (SPM) to address well-known weaknesses in the official poverty statistics.[1]  Like the SPM, this expanded poverty measure counts more types of income, subtracts work-related and medical expenses that the official poverty measure ignores, and uses a modernized poverty line, which we adjust each year for inflation.  The analysis also goes beyond a number of past studies of deep poverty by correcting for the underreporting of key government benefits in the Census data.[2]  These corrections expose a rise in deep poverty for children that is masked in the uncorrected data by changes in the degree of underreporting over the period.

We focus on children, the age group for whom evidence is strongest that deep poverty causes lasting harm.  The analysis spans 1995 through 2010, paying particular attention to the decade between 1995 and 2005, a period that provides an especially revealing look at the changing safety net:  1995 was the year before Congress enacted a major overhaul of the welfare system, while 2005 was a year with comparable economic conditions.  The unemployment rate in 2005 (5.1 percent) was similar to the rate in 1995 (5.6 percent), as was the percentage of adults employed, which stood at 62.7 percent in 2005 and 62.9 percent in 1995.

Between 1995 and 2005, while overall child poverty declined significantly, the safety net became less effective at protecting children from deep poverty.  The share of children living in deep poverty rose from 2.1 percent in 1995 to 3.0 percent in 2005.    

  • Among children whose families were below half the poverty line before counting government benefits, income from safety net programs lifted 88 percent of them above half the poverty line in 1995 (and thereby lowered the deep poverty rate among children by seven-eighths).  By comparison, in 2005, safety net benefits lifted just 78 percent of such children out of deep poverty.
  • If the safety net had remained as effective at keeping families out of deep poverty in 2005 as in 1995, 1.2 million children would have been below half the poverty line in 2005; instead, 2.2 million were.
  • In other words, an additional 1 million children lived in deep poverty in 2005 due to the weakening of the safety net over the previous decade. [3]

The safety net similarly became less effective at protecting job seekers from deep poverty.  Among very poor unemployed workers in families with children who spent part or all of the year looking for work, the safety net lifted 82 percent above half of the poverty line in 2005, down from 91 percent in 1995.

The safety net for the very poor regained part of its previous strength during the Great Recession.  This improvement in large part reflected the various temporary measures enacted under Presidents Bush and Obama in 2008-2009 to support the struggling economy, such as emergency federal unemployment benefits, expanded tax credits for working families (the Child Tax Credit and Earned Income Tax Credit, or EITC), and higher monthly benefit levels in the SNAP program (formerly known as food stamps).  It also reflected the growth in various safety net programs that expand automatically to meet rising need when the economy deteriorates.

The strengthened safety net prevented poverty and deep poverty from surging in the recession.   

  • Not counting government programs, the poverty rate would have risen under our measure from 25.3 percent in 2005 to 29.4 percent in 2010.  Counting these programs,the poverty rate did not rise (it was 15.0 percent in 2005 and 14.8 percent five years later, not a statistically significant difference), even though the unemployment rate was much higher in 2010 than in 2005. 
  • Notcounting government programs, the deep poverty rate would have risen substantially in the recession (from 15.0 percent in 2005 to 18.0 percent in 2010).  The deep poverty rate after counting benefits from these programs remained unchanged, at 3.3 percent.  For children, the deep poverty rate, notcounting the safety net, would have risen from 13.2 percent to 16.8 percent.  Counting the safety net programs, it declined from 3.0 percent to 2.6 percent.

The revival of the safety net may prove short-lived, however, as most of the temporary recovery initiatives have expired.  These expirations will likely have a significant impact on children’s poverty and deep poverty rates, given the anti-poverty effectiveness of the expansions during the years they were in effect.  

Two programs with important expansions that have not expired are the EITC and Child Tax Credit.  CBPP estimates that these expansions, now scheduled to expire at the end of 2017, kept about 1 million children above the poverty line in 2013.

The remainder of this analysis explores these findings in detail, including:  how our poverty measure affects poverty and deep poverty trends; the extent to which the safety net’s protection from deep poverty weakened from 1995 to 2005; why the safety net weakened for the poorest families; the effects on single-parent families; the resurgence of the safety net during the Great Recession; and why this resurgence will likely prove to be largely temporary.   Appendices provide further details regarding: how correcting for underreporting affects our results (Appendix A); year-by-year trends in poverty and deep poverty (Appendix B); and the methods used in this analysis and the effects of individual programs on poverty and deep poverty (Appendix C).

Expanded Poverty Measure Reveals Stronger Safety Net in Any Given Year But Rise in Deep Poverty

The poverty measure in this analysis differs from standard poverty measures in several respects.  In accordance with National Academy of Sciences (NAS) recommendations (and as many experts favor), it counts government noncash benefits such as food assistance and rent subsidies as income and subtracts from income various taxes paid, work expenses, and out-of-pocket medical expenditures.  It also adopts a somewhat modernized poverty line, calculated in accordance with NAS methods.  This poverty line is “anchored” at 2010 — that is, it starts with a poverty threshold for 2010 calculated by the Bureau of Labor Statistics based on the low end of what most Americans spent on food, clothing, shelter, and utilities, and adjusts this threshold backward in earlier years for price inflation.  Our poverty calculation also slightly broadens the family unit to include certain unmarried partners.  Finally — and of particular note — it corrects for the underreporting of certain government assistance income in the Census data. [4]  (See Appendix A for more on the importance of correcting for underreporting.)

Overall, our measure largely resembles the SPM,[5] with the addition of corrections for underreporting. 

Using this expanded measure reduces the percentage of children considered poor, and especially deeply poor.  The child poverty rate in 2010 drops from 22 percent under the official poverty measure to 15.3 percent.  The share of children below half the poverty line drops from 9.9 percent to 2.6 percent.  (See Figure 1.)

Over time, our measure, like the official measure, shows a decline in child poverty in the first decade after the welfare law:  from 22.6 percent in 1995 to 16.7 percent in 2005.  (See Figure 2.)

For deep poverty, however, our measure reveals a troubling trend not evident in the official figures.  The share of children below half the poverty line rose by more than one-third between 1995 and 2005 (from 2.1 percent to 3.0 percent), reflecting the weakening of a significant aspect of the safety net, as discussed in the next section.  

Safety Net Protection from Deep Poverty Weakened Over 1995-2005

Between 1995 and 2005, public benefits became less effective at shielding children from deep poverty.  In 1995, public benefits lifted above half the poverty line 88 percent of children who would otherwise have been in deep poverty; by 2005, this figure had dropped to 78 percent.  This decline largely reflects dramatic declines in the protective role of cash assistance programs financed by the Temporary Assistance for Needy Families (TANF) block grant in the late 1990s and early 2000s.  SNAP’s protective role also weakened in that period.

  • In 1995, Aid to Families with Dependent Children (AFDC), TANF’s predecessor, lifted out of deep poverty some 61 percent of the children who would otherwise have been below half of the poverty line.  By 2005, this figure for TANF was only 22 percent.
  • In 1995, SNAP protected from deep poverty 62 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 weakened, the number and percentage of children and parents in deep poverty grew significantly.  The number of children and parents living below half of the poverty line (after benefits are counted) rose from 2.5 million in 1995 to 2.8 million in 2000 — despite a booming economy and declines in both unemployment and regular poverty — and continued rising to 3.5 million by 2005, a ten-year increase of 1 million or 41 percent.

This rise in deep poverty did not result from declines in earnings or in other income from private, market-based sources.  Using non-safety-net income alone, children’s deep poverty rate fell from 17.2 percent in 1995 to 13.2 percent in 2005, reflecting labor market gains in the late 1990s and a movement from welfare to work for some of the poorest families.  The sharp decline in the safety net’s protection from deep poverty, however, more than offset that improvement, leaving a rising number of children below half the poverty line once benefits are counted.

The increase thus directly reflected the weakening of aid.  If public benefits had kept the same share of children out of deep poverty in 2005 as in 1995, 1 million fewer children would have lived below half of the poverty line in 2005 — 1.2 million, rather than the actual 2.2 million.

By contrast, the safety net became more effective over this period at protecting children from less severe poverty.  The share of otherwise poor children that public benefits lifted above the poverty line rose from 24 percent in 1995 to 35 percent in 2005.  This is because support for low-income working families with incomes modestly below (or modestly above) the poverty line grew considerably more robust.  The large EITC expansion enacted in 1993 completed its phase-in in 1996, and policymakers created the Child Tax Credit in 1997 and expanded it in 2001 and 2003 (and again in 2008 and 2009) to reach additional low-income working families and to increase tax-credit amounts for many families.  Further, SNAP participation among low-income working households has risen markedly since 2000 due to bipartisan efforts at federal and state levels such as streamlining paperwork requirements and changing eligibility rules that disqualified families that own a modest car for commuting to work.  The extension of public health insurance through Medicaid and the Children’s Health Insurance Program (CHIP) also helped many low-income families, particularly working families.[6]

As a result, the safety net for working-poor families grew more effective at reducing poverty between 1995 and 2005.  Among children with family members who worked all or part of the year and whose non-safety-net income was below the poverty line, public programs lifted 39 percent above the poverty line in 2005, up from 26 percent in 1995.

Why Did the Safety Net Weaken for the Poorest Families Through 2005?

During the 1990s, federal and state policy changes — including the 1996 welfare law, which created the TANF block grant to replace the AFDC cash assistance program — constricted access to basic cash assistance for the poorest families with children while expanding supports for low-income working families through enlargement of tax credits, child care subsidies, and health insurance coverage.  In the mid- to late 1990s, the combination of a booming economy and these policy changes pulled and pushed large numbers of low-skilled adults, especially single mothers, into the paid labor force and increased their income related to work.  Poverty among families with children fell.

Even as some families climbed out of poverty, warning signs suggested that others were slipping deeper into poverty.  Despite the strong economy, many families had periods of joblessness due both to the volatile nature of many low-wage jobs and to individual barriers to employment, and deep poverty among children edged up.  Then, the 1990s economic boom ended, and the economy experienced a shallow recession in 2001 followed by years of labor-market weakness before the Great Recession started after 2007.  Between 2000 and 2005, employment rates fell for single mothers, and deep poverty among children, as measured in this analysis, climbed further.

The largest single reason that the safety net protected fewer children against deep poverty in 2005 than in 1995 was the loss of cash assistance following the 1996 welfare overhaul.  The number of children that cash assistance (AFDC or TANF) kept above half the poverty line fell from 2.4 million (3.4 percent of all children) in 1995 to 609,000 (0.8 percent of all children) in 2005.  (See Figure 4.)

Eligible families’ participation in TANF cash assistance also fell steeply, from 84 percent of those eligible in 1995 to just 40 percent in 2005, Department of Health and Human Services data show.  The decline chiefly affected families at high risk of deep poverty, since most states’ TANF income eligibility limits are well below the poverty line.[7]  

The reasons for the sharp drop in TANF participation vary across states.  Some states adopted policies and administrative practices that discouraged many poor families from applying, made it more difficult for families to complete the application process, or led many families to leave the program even when they did not have a job.  

Many families lost assistance because of strict welfare-to-work rules and policies that terminated assistance to families that did not meet various program requirements such as participating in structured job search workshops for as many as 40 hours per week or looking for jobs on their own for 20 or 30 hours per week, often without child care assistance.  While such requirements may not sound onerous, some families that confront a serious crisis or have mental or other health issues or experience domestic abuse or substance abuse problems can experience difficulty complying.  Research has found that many families that lose assistance due to sanctions have significant mental health (or other health) problems or experience domestic abuse or other barriers to employment, which can make strict adherence to such requirements more difficult for them.  In addition, some families lost benefits if parents missed meetings due to transportation or child care problems, or other logistical barriers created or exacerbated by poverty.  In addition, in some states, time limits on assistance barred some very poor families from receiving assistance.[8]

On balance, the decline in cash assistance among out-of-work poor single parents outweighed the employment gains among such parents over TANF’s first decade.  The number of non-working single parents fell by 400,000 between fiscal years 1995 and 2005 (from 4.1 million to 3.7 million), according to CBPP analysis of average monthly Census data.  During the same period, the number of non-working adults receiving cash assistance fell by 3 million (from 4.1 million to 1.1 million), according to published AFDC and TANF administrative data.  These two sets of numbers — the former from survey data and the latter from case records — are not precisely comparable (for example, the case data include a small number of married parents).  Nonetheless, the fact that the number of jobless single parents who receive cash welfare assistance fell much more steeply than the overall number of jobless single parents indicates that the number of single-parent families that received neither earnings nor cash assistance — and hence were especially vulnerable to deep poverty — swelled. 

Rigorous evaluations of experimental state welfare-to-work programs of that era confirmed that many of these programs led to an increase in deep poverty even as they reduced overall poverty.  In one federal evaluation of 11 early 1990s pilot programs that included work requirements and other elements of the 1996 welfare law, deep poverty rates rose significantly among families in six of the programs (and declined in none) in comparison with “control group” families randomly assigned to a more traditional welfare program.  By contrast, overall poverty rates (that is, the share of participants below the full poverty line) fell significantly in five programs (and increased in none).[9]  

All 11 programs raised employment rates, but most also raised the share of families at any given time who had neither employment nor welfare.[10]  Similarly, in one of the few rigorous, random-assignment studies of income distribution in a time-limited welfare-to-work program — Connecticut’s Jobs First pilot study — researchers concluded that “there are definitely negative [income] effects on some women, and positive effects on others” and that losses “are concentrated at the lower end, with the positive ones concentrated in the upper half” of the participants’ income distribution.[11]

In addition to creating TANF, the 1996 welfare legislation made large cuts to SNAP and other programs, which further weakened the safety net for the poorest families. The SNAP cuts included both across-the-board benefit reductions and targeted eligibility changes that denied benefits to many legal immigrants, unemployed childless adults, and others.  While policymakers later restored some of the SNAP cuts, others remain in place.  Another reason why SNAP rolls dropped sharply in TANF’s initial years is that families traditionally had enrolled in SNAP and cash assistance jointly, and as cash assistance plummeted, many families also stopped receiving SNAP despite remaining eligible.  

As a result, SNAP lifted many fewer children from deep poverty in 2000 and 2005 than in 1995, although it, like the safety net as a whole, subsequently regained strength and, by 2010, had substantially surpassed earlier participation levels.

Rise in Deep Poverty Primarily Affected Single-Parent Families

The shrinking role of TANF helps to explain the stark differences seen in deep poverty trends by family type.  Unmarried families with children, who make up the majority of TANF participants, showed by far the largest increase in deep poverty rates from 1995 to 2005.  For people living in these families (generally single-parent families, though some also include an unmarried partner[12] and others consist of children being raised by a grandparent or other guardian), the likelihood of being below half the poverty line nearly doubled between 1995 and 2005, from 2.8 percent to 5.3 percent.

By contrast, the deep poverty rate for people in married-couple families with children — a group that seldom received cash assistance under AFDC — remained essentially unchanged, at 1.3 percent in 1995 and 1.2 percent in 2005.  (See Figure 5.)

Childless individuals saw a modest increase in deep poverty.  The deep poverty rate for individuals not living in families with children rose from 3.9 percent to 4.3 percent between 1995 and 2005.

One contributor to the rise in deep poverty was changes in safety net policies toward non-citizens.  The 1996 welfare law barred many lawfully present immigrants from receiving major forms of federal assistance such as SNAP, TANF, and Supplemental Security Income (SSI).  (Undocumented immigrants were already ineligible for these programs.)  Deep poverty rates rose for children in households that included non-citizens, from 4.3 percent in 1995 to 6.1 percent in 2005.  Poor children living with non-citizens experienced sharp losses of family income even though most of the children were themselves citizens.  Research by Harvard University economist George Borjas has also shown that the immigrant restrictions led to significantly higher food insecurity in affected households, with food insecurity rates rising 5 percent for every 10 percent decline in the population receiving public assistance.[13]

While non-citizen households were hit particularly hard, children in citizen-only households were also affected by the changes in the safety net.  Deep poverty rates for children in households where all members were citizens rose from 1.8 percent in 1995 to 2.3 percent a decade later. 

In Great Recession, Strengthened Safety Net Prevented Surge in Poverty and Deep Poverty

As the Great Recession took hold in 2008, Congress and President Bush took three temporary steps to support income, shore up consumer spending, and soften the recession’s impacts on struggling families:  a one-time “stimulus rebate” for most households, an expansion of the Child Tax Credit to reach more low-income working families, and a federal program of extended jobless benefits called Emergency Unemployment Compensation.

After the recession deepened in late 2008, Congress and the incoming Obama Administration adopted a stronger recovery package including several income-support initiatives.  The largest were a temporary Making Work Pay Tax Credit of up to $400 per worker ($800 per couple), further expansions of tax credits for working families (the Child Tax Credit and the EITC), additional unemployment benefits, temporary nutrition assistance in the form of higher SNAP benefits, and a one-time payment in 2009 chiefly for seniors and people with disabilities.

These initiatives, combined with growth in existing safety net programs such as SNAP and regular state unemployment benefits — which, by design, expand automatically to meet rising need — shored up family incomes against the effects of burgeoning layoffs, reductions in work hours, and a shrinking economy.  Between 2005 and 2010, government assistance became much more effective against poverty and deep poverty, as reflected under our expanded poverty measure.

  • The number of people that the safety net kept above the poverty line rose from 30 million in 2005 to 45 million in 2010.  The number of children kept out of poverty rose from 7 million in 2005 to 12 million in 2010.
  • The number of people that the safety net kept above half the poverty line rose from 35 million in 2005 to 45 million in 2010. 
  • The strengthened safety net prevented a surge in poverty.  Not counting government programs, the poverty rate would have risen under our measure from 25.3 percent in 2005 to 29.4 percent in 2010.  Counting these programs, the poverty rate was essentially unchanged (15.0 percent in 2005 and 14.8 percent in 2010, not a statistically significant difference). 
  • It also prevented a surge in deep poverty.  Not counting government programs, the deep poverty rate would have risen from 15.0 percent in 2005 to 17.9 percent in 2010.  Counting these programs, it remained unchanged at 3.3 percent.  For children, the deep poverty rate, notcounting the safety net, would have risen from 13.2 percent to 16.8 percent.  But counting these programs, the rate declined from 3.0 percent in 2005 to 2.6 percent in 2010.

This does not mean that Americans avoided economic hardship in the recession.  Millions of families suffered wrenching layoffs, foreclosures, evictions, reduced job hours, and loss of savings.  Median pre-tax household income dropped nearly $3,400 from 2007 to 2010 (in 2010 dollars), by official Census figures.  Still, on balance, the number and percentage of children in families with annual income below the poverty line, and the number and percentage in deep poverty, did not increase, a striking accomplishment in the worst recession since the 1930s.

These figures contrast with the official poverty statistics, which show a bleaker trend, with both poverty and deep poverty rising in the recession and the percentage of people in deep poverty setting a record high in 2010 (with data going back to 1975).  But the official statistics don’t reflect most of the safety net’s added anti-poverty impact because that impact came in the form of noncash benefits (such as SNAP) and tax-based assistance (such as the expanded Child Tax Credit), which the official poverty measure does not count.

Much of Safety Net’s Recent Strength Was Temporary

The majority of temporary safety net initiatives implemented in the Great Recession have now expired.  The Making Work Pay credit ended in tax year 2010, while most of the expansions in unemployment insurance (UI) benefits and SNAP expired by the end of 2013. 

These expirations will likely have a significant impact on children’s poverty and deep poverty, given the expansions’ anti-poverty effectiveness.  The temporary UI expansions kept about 800,000 children out of poverty in 2010, according to a previous CBPP analysis.[14]  The temporary SNAP expansion kept 450,000 children out of poverty that year, while the Making Work Pay credit kept nearly 440,000 children out of poverty.  (The figures in that analysis, unlike other figures in this report, do not correct for underreported income.  With corrections, the estimated effect of the SNAP expansion in particular would have been larger.)

Two programs with important expansions that have not expired are the EITC and Child Tax Credit.  In both cases, the expansions are scheduled to expire at the end of 2017.  CBPP estimates that these two expansions together kept about 1 million children above the poverty line in 2013.[15]

The expirations have come at a time of progress in the labor market.  But although the number of officially unemployed workers fell by 3.4 million between 2010 and the end of 2013, this is somewhat misleading as it ignores workers who have given up seeking work and people able to find only part-time employment.  More revealing is the employment rate (the share of the total adult population with work), which dropped from 62.7 percent in 2005 to 58.5 percent in 2010 and has since recovered only a fraction of the lost ground; it stood at 59.3 percent in April 2015, the latest available figure.  Some of this decline reflects the aging of the baby boom into their retirement years, which has increased the non-working adult population.  But the employment rate of people in their prime working years, ages 25-54, also remains well below its level at the start of the recession (77.2 percent in April 2015, versus 79.7 percent in December 2007).

Of particular note are the trends in unemployment benefits.  As the economic recovery has progressed, the overall amount of UI benefits paid has declined markedly.  Some of this reduction reflects the improving labor market and is a positive sign.  But there are also signs that unemployment insurance assistance declined faster than need, with the contraction placing some upward pressure on the poverty rate. 

Between 2010 and 2012, both the number of workers receiving UI and the number lifted out of poverty by UI fell much faster than the number of unemployed.  If the number of people kept out of poverty by UI had fallen at the same rate as the number of unemployed, 1 million fewer people would have been poor in 2012 under the official poverty definition, and the official poverty rate would have declined between 2010 and 2012 rather than remaining statistically unchanged.[16]

In addition, while long-term unemployment has receded somewhat, it remains unusually high.  Partly as a result, the number of unemployed workers not receiving unemployment benefits is actually higher today than at any point during the Great Recession.  Similarly, the percentage of unemployed workers who receive unemployment benefits — 27 percent — is now at the lowest level on record, with data back to 1971.

Conclusion

In the decade following a major overhaul of the nation’s welfare system in 1996, the safety net shifted focus, becoming more effective at helping low-income working families climb above the poverty line but less effective at protecting children from deep poverty.

In response to the Great Recession, Congress and two Presidents acted to protect families and the economy by bolstering the safety net with tax credits, stimulus rebate payments, and other anti-recession initiatives.  The strengthened safety net staved off a surge in poverty and deep poverty when the economy slumped.  The initiatives were temporary, however, and joblessness remains elevated.  Even after the employment rate eventually returns to its pre-recession level, the question of how to better protect the nation’s poorest families and children will remain.

Appendix A:
The Importance of Correcting for Underreported Income

Correcting for the underreporting of government benefit income has an important impact on our understanding of poverty and deep poverty, separate from the impact of counting more types of income and adopting other features of the NAS-based poverty measure.  Adopting an NAS-based measure cuts the children’s deep poverty rate in 2010 nearly in half, to 5.5 percent (compared with 9.9 percent under the official definition).  But correcting for underreporting of SNAP, SSI, and TANF (using the TRIM model developed by the Urban Institute) lowers the rate by half again, to 2.6 percent.

Correcting for underreporting has a particularly pronounced effect on our findings for trends in deep poverty between 1995 and 2005.  Before applying the corrections, the analysis shows that the children’s deep poverty rate fell or held steady during this decade.  After applying the TRIM corrections in combination with the NAS-based approach, the more complete measure shows a weakening of the cash welfare assistance part of the safety net and a resulting rise in the deep poverty rate for children (from 2.1 percent to 3.0 percent) in the decade after welfare reform.  (See Figure 6.)

The corrections are important; uncorrected figures miss billions of dollars per year of income from government benefit programs.  In the case of cash assistance, the data source we use, the annual March Current Population Survey (CPS), appears to have missed about $9.3 billion in total AFDC benefit payments in 1995 (in 2010 dollars) compared with administrative program records.  By 2005, after TANF replaced AFDC, the CPS appears to have missed a smaller amount — $6.1 billion — largely as a result of TANF’s dwindling size.  Put another way, there was less actual assistance to undercount in 2005 than in 1995.[17]

While TRIM’s corrections are not perfect, they greatly improve the accuracy of the survey data with respect to the total size of the safety net.  According to the uncorrected survey data, for example, the total value of cash welfare benefits from AFDC was $15 billion in 1995, while our corrected data show about $20 billion (not adjusting for inflation).  Compared with administrative records, the corrected figures capture over 90 percent of the true total benefits that year, while the uncorrected figures capture about 70 percent of true benefits.

An alternative method of correcting for underreporting, not based on TRIM, also shows a rise in children’s deep poverty between 1995 and 2005.  That method — which uses the percentage of total benefit payments missed by the Census data to adjust programs’ anti-poverty effectiveness — confirms our finding that the children’s deep poverty rate rose by 0.9 percentage points from 1995 to 2005.[18]

The underreporting of welfare benefits may help explain conflicting findings from previous studies on deep poverty.  In groundbreaking work, Christopher Wimer and his colleagues estimate poverty and deep poverty rates back to 1967 using a version of the SPM.  They find little change in the children’s deep poverty rate since 1996.  By contrast, Yonatan Ben-Shalom, Robert Moffitt, and Karl Scholz show a rise in all-age deep poverty rates from 4.5 percent in 1993 to 6.6 percent in 2004, and Luke Shaefer and Kathryn Edin show a doubling of children’s extreme poverty (the number of U.S. children in families with monthly cash incomes below the equivalent of $2 per person per day) from 1996 to 2011. [19]  However, Wimer and colleagues used uncorrected data from the CPS (the same data source we use here with corrections), which has a relatively high level of underreporting of TANF and other benefits, whereas the other two studies used data from the Survey of Income and Program Participation, which generally suffers somewhat less from underreporting.[20]

Appendix B:
Year-by-Year Detail Shows Role of Business Cycle, Policy

This analysis focuses on the years 1995, 2005, and 2010.  Year-by-year data tell a more nuanced but consistent story.  The deep poverty rate for children rose sharply in the first years after enactment of the 1996 national welfare legislation, peaking at 3.1 percent in 1998.  (See Figure 7.) 

As the economic boom continued, the rate declined to 2.4 percent in 2000, but rose again after the recession of 2001 took hold and remained high as weak labor market conditions lingered for several years.  The rate dipped slightly in the years before the Great Recession, which began after December 2007.

As noted, the share of children in deep poverty barely increased in the Great Recession once noncash benefits and tax credits are counted and underreporting is corrected.  This was surprising, given the depth of the economic crisis and the resulting sharp rise in the percent of children with little non-safety-net income.  The percentage of children below half the poverty line not counting benefits ticked up from 13.0 percent in 2007 to 13.7 percent in 2008 (the first, and mildest, year of the recession), then rose to 16.4 percent in 2009 and 16.8 percent in 2010.  (See Figure 8.)  But after counting benefits, the rate stayed below 3.0 percent. 

Recovery measures started bringing income assistance to low- and middle-income families in 2008, with Congress and President Bush enacting a one-time stimulus rebate credit of $1,200 or more for a low-earning couple with two children and earnings of at least $3,000, as well as federal Emergency Unemployment Compensation and an expansion of the Child Tax Credit to working parents earning as little as $8,500.

Policymakers enacted stronger temporary recovery measures in 2009 as the depth of the recession became more evident.  As previously discussed, these measures helped bring deep poverty rates for children down to 2.6 percent by 2010, although much of this assistance has since expired.

The patterns suggest that both business cycles and changes in the safety net have strong effects on children’s deep poverty.

Appendix C:
Methods and Detailed Tables

This Appendix describes how we performed these calculations, including how we determined individuals’ poverty status and the numbers of people lifted out of poverty and deep poverty.

Determining Poverty Status and Deep Poverty

This analysis builds on a 2009 CBPP report that examined the effect of the safety net on poverty and deep poverty.[21]  Like the earlier report, this analysis starts with person-level data from the Census Bureau’s Current Population Survey (CPS) Annual Social and Economic Supplement.  Our approach seeks to address well-known criticisms of the official Census Bureau poverty measure, including:

  • In the official measure, income does not include noncash and tax-based benefits, which have grown as a means of support for poor families in the last three decades.
  • The CPS significantly undercounts certain benefits.  (Census’s counts of program participants and aggregate benefit payments received typically fall well short of the actual totals shown in administrative records.)
  • The Census measure does not take into account resources that households must devote to taxes, work expenses, and health care and that are thus unavailable to meet basic needs such as food, clothing, and shelter.
  • The poverty line does not vary by geographic area to reflect the significant differences in the cost of living across the country.
  • Annual adjustments in the poverty line do not effectively track changes in the cost of basic needs.

To address these concerns, this analysis takes the following steps, based largely on recommendations of the National Academy of Sciences’ 1995 expert panel on poverty measurement:[22]

  • Accounting for noncash benefits and taxes.  The official measure counts only cash income, which includes sources such as earnings from work, interest, dividends, child support, and government cash payments such as Social Security, UI benefits, and monthly payments from TANF and the SSI program for very-low-income seniors and people with disabilities.  Our measure includes these programs but, following NAS recommendations, also includes the value of government noncash benefits (other than medical benefits) and the net impact of federal and state income and payroll taxes.  Specifically, we include the value of the Supplemental Nutrition Assistance Program (SNAP, formerly food stamps) and housing assistance,[23] school lunch subsidies, and low-income home energy assistance.  We also count tax credits for working families such as the EITC and Child Tax Credit, while subtracting federal and state income taxes and payroll taxes paid.  We do not subtract state sales tax because the data needed to do so are not available.  We use payroll tax estimates provided by Census on the CPS file.  We use federal and state income tax estimates based on a model created by the National Bureau of Economic Research (and known as TAXSIM); these are more consistent over time than the Census Bureau’s own income tax estimates, which Census is continually improving. [24]
  • Correcting for underreporting of benefits.  We go beyond the NAS recommendations by using income data on TANF, SSI, and food stamps from the Transfer Income Model (TRIM) developed for the Department of Health and Human Services.[25]  The TRIM model starts with information from the CPS and adds to it by assigning benefits to some individuals in the CPS who did not answer all of the CPS questions.  For example, someone who skipped a question about whether he or she received food stamps might be assigned food stamp income.  Unlike in the CPS, the total number of recipients in the TRIM data is designed to approximate true totals from each program’s administrative records.[26]
  • Subtracting medical out-of-pocket (MOOP) expenses and work expenses (including child care).  Following NAS recommendations, we subtract these expenses from income. Because these expenses are not included in the survey data for most years of our study period, we estimate them by adapting parameters and formulas used previously by the Census Bureau and the Bureau of Labor Statistics (BLS) for their published NAS series and made available by Census and BLS staff.[27]
  • Counting more people as family members.  Poverty status is determined family by family (that is, you are poor if your family is poor).  In the official poverty measure, the family unit includes people related by birth, marriage, or adoption who reside together.  The NAS panel did not recommend a specific family unit but suggested exploring inclusion of other household members, particularly unmarried partners.  In certain cases, therefore, we add unmarried partners to the family.  We do this only if one of the partners is the head of household, because the relationship data needed for including other unmarried partners are not available for the early years of our study.  Also, we only include longer-term partners, since temporary partners appear to contribute less of their income to the family than long-term partners do.[28]  We consider partners long-term if they were present in the household both at the time of the survey and 12 months earlier.  We also assign certain young children who appear to be in foster-care-like situations to the family of the head of household.[29]
  • Adjusting the poverty line itself.  We use an updated poverty line based on NAS methods and reflecting living standards as of 2010.  Specifically, we start with an NAS-based poverty line of $27,499 for a two-parent, two-child family in 2010, calculated by BLS.[30]  This level reflects what a relatively frugal two-parent, two-child family at the 33rd percentile of the spending distribution, living in an average-cost community, spent that year on very basic needs:  food, clothing, shelter, utilities, and medical care, plus 20 percent for miscellaneous items.[31]  (The 33rd percentile means that one-third of families spent less than this amount and two-thirds spent more.)[32] 

Setting aside medical needs — which the NAS proposed treating as a deduction from income rather than a part of the poverty threshold — this poverty line represents $25,510 a year for food, clothing, shelter, utilities, and other miscellaneous expenses (see box).[33]  We reduce this level with the consumer price index (CPI) to reflect the lower price levels in earlier years.[34]  Thus, unlike in many NAS-based poverty measures that evolve with changing living standards, our threshold is “anchored” to living standards in 2010 adjusted only for inflation.  We calculate variations for different family sizes and geographic locations based on NAS-based procedures.[35]

Poverty Lines for a Two-Adult, Two-Child Family: Official Threshold and Threshold Following NAS Recommendations, 2010

Official Poverty Definition $22,113
NAS with Medical Needs 27,499
Medical Share (7.233%) 1,989
Non-Medical Share 25,510

Source: U.S. Census Bureau, Bureau of Labor Statistics, and CBPP calculations.  This version of the NAS threshold is sometimes referred to by Census as the MIT-CE-GA threshold.  The abbreviation indicates that Medical out-of-pocket needs are In the Threshold; the threshold is Geographically Adjusted based on local housing costs; and the threshold was updated from previous years using the change in median Consumer Expenditure on basic needs (though, as noted in the text, in this report we deflate this 2010 threshold for earlier years using consumer prices in order to maintain comparability over time).

Census staff and academic experts have experimented with a variety of ways to implement the NAS poverty measurement recommendations that vary in technical details.  Recent Census Bureau publications have discussed and presented eight variations of the NAS poverty measure, as well as a newer refinement of the NAS measures known as the Supplemental Poverty Measure (SPM).  Data for measuring the SPM are not available before 2009.  For the purposes of this analysis, therefore, we used an NAS-based measure that can be measured as consistently as possible over time.

We use the same income measure and poverty lines to determine deep poverty as we use for poverty.  We define deep poverty as income below 50 percent of the NAS poverty line.  (We also provide an alternative definition:  income below 75 percent of the poverty line.[36])

One final adjustment to our data addresses the concern that certain people who appear to be in deep poverty are self-employed business owners who may possess substantial assets but are experiencing temporary business losses.  To reduce this concern, this report omits families with negative cash income from our counts of families in deep poverty.

Calculating Who Is Lifted Out of Poverty and Deep Poverty

This analysis examines people whose family income is lifted above the poverty line — or above half (or three-quarters) of the poverty line — by one or more forms of government assistance.

In this analysis, a person is considered to be lifted out of poverty by a given program or group of programs if his or her family’s total income from all sources is above the poverty line but would be below the poverty line if the income type in question were excluded.

In 2005, for example, 47.5 million people lived in families with income below the poverty line by our measure.  But if income from all government safety net programs (including both means-tested and non-means-tested programs) were excluded, nearly 72.2 million people would have below-poverty income.  The difference — about 24.6 million — is the number lifted above the poverty line by the safety net.

Under this method, a family can be lifted above the poverty line by more than one program.  For example, consider a family with $15,000 in earnings, $2,000 in SNAP food assistance, and $4,000 in EITC; assume that the poverty line for this family is $20,000.  The family’s non-EITC income (earnings plus SNAP) totals $17,000, which is below the poverty line.  Adding the EITC brings the family to $21,000, or above the poverty line.  Therefore, the family is considered to be lifted out of poverty by the EITC.

Our method also counts this family as lifted out of poverty by SNAP: the family’s non-SNAP income (earnings plus EITC) totals only $19,000, but adding SNAP pushes the family to $21,000. Many people are lifted out of poverty by a combination of benefits, so more than one of the benefits in our analysis can be credited with raising them above the poverty line.  Our figures on the safety net’s overall anti-poverty impact, however, count each family only once.

Table 2 below shows, for several programs, the steps in the calculation.  The second and third rows of the table show the actual number and percentage of people below the poverty line.  The next eight rows show how many people would bepoor if a particular benefit were excluded.

For example, while 47.5 million Americans were actually in poverty in 1995 by our measure, 51.4 million would have been in poverty without the income they received from the SNAP program.[37]   The difference between these two figures, roughly 3.9 million, is the number of people lifted above the poverty line by SNAP that year.

The bottom section of Table 2 shows the percentage by which the particular benefit reduced the number of poor people.  For instance, SNAP (then called food stamps) reduced the number of poor people by 8 percent in 1995.  Put another way, the 3.9 million people lifted above the poverty line by SNAP in 1995 made up 8 percent of those whose income not counting SNAP was below the poverty line.

Tables 3 and 4 repeat this information for people in families below 75 percent and 50 percent of the poverty line, respectively.  For example, Table 3 shows that SNAP reduced the number of Americans below 75 percent of the poverty line by about 24 percent in 1995, relative to the number who would be this poor without that income.  Table 4 shows that SNAP reduced the number of Americans below 50 percent of the poverty line by 38 percent in 1995.

Table 5 focuses on people in unmarried families with children — the group most likely to receive TANF — and shows the total number of such people at all income levels and the percentage in deep poverty (below 50 percent of the poverty line) with and without counting government benefits.  Counting government benefits, the deep poverty rate for this group nearly doubled from 2.8 percent in 1995 to 5.3 percent in 2005.  Not counting government assistance, their deep poverty rate would have declined from 35.6 percent to 27.3 percent during the same period, indicating that a weakening of assistance explains the rise in deep poverty. 

Table 5 also splits people in unmarried families with children into four groups:  female-headed families where the head of household has no partner present, male-headed families where the head of household has no partner present, families where the head of household lives with a long-term partner (one present 12 months earlier), and those where the head of household lives with a short-term partner (one not present 12 months earlier).  As previously noted, we count long-term (but not short-term) unmarried partners as part of the householder’s family when determining the family’s poverty status.  The inclusion of long-term partners’ income helps explain why the deep poverty rate is much lower for families with long-term partners than for families with short-term partners (whom we do not include in the family for the purpose of determining poverty status).

Table 1: Persons in Poverty (Below 100% of Poverty Line) and
Deep Poverty (Below 75% and 50% of Poverty Line), By Age: 1995-2010

Poverty line is NAS (MIT-CE-GA) poverty line for 2010 deflated for earlier years using CPI-U.

Income includes cash after taxes, plus noncash benefits (SNAP, school lunch, housing and energy assistance) minus out-of-pocket medical and work expenses. 

Income also includes one-time stimulus payments in 2008 and Economic Recovery Payments in 2009.Income from TANF, SSI, and SNAP is corrected for underreporting using baseline estimates from the HHS/Urban Institute Transfer Income Model (TRIM).

Federal and state income taxes are from NBER’s TAXSIM9 model.  Payroll tax is from Census tax model.

Housing subsidies are valued at Fair Market Rent in excess of 30 percent of household income.

Persons under 75% and 50% of poverty exclude those in families with negative cash income.Unmarried long-term partners are treated as part of the head of household’s family, as are persons under age 15 not living with family members.

Number (in thousands)
  All Ages   Children Under 18
  Poor Under 75% Under 50%   Poor Under 75% Under 50%
1995 47,512 20,477 6,967   16,112 6,408 1,523
1996 47,326 21,134 7,019   15,791 6,513 1,571
1997 46,690 21,326 7,674   15,683 6,778 2,025
1998 43,738 20,860 8,177   14,603 6,661 2,198
1999 41,721 19,113 7,674   13,241 5,574 1,952
2000 40,423 18,620 7,024   12,754 5,509 1,746
2001 41,414 20,171 8,394   12,531 5,776 2,162
2002 42,197 20,259 8,950   12,331 5,398 2,117
2003 42,898 20,978 9,358   12,397 5,587 2,233
2004 43,216 21,295 9,511   11,996 5,422 2,069
2005 44,132 21,644 9,554   12,374 5,646 2,195
2006 43,517 21,533 9,172   12,190 5,617 1,920
2007 44,624 21,797 9,257   12,613 5,558 2,012
2008 44,114 22,418 10,022   11,921 5,511 2,141
2009 42,881 21,266 10,046   11,008 5,067 2,093
2010 45,215 22,779 10,177   11,476 5,305 1,974
Percent              
1995 18.0 7.7 2.6   22.6 9.0 2.1
1996 17.7 7.9 2.6   22.2 9.1 2.2
1997 17.4 7.9 2.9   21.9 9.5 2.8
1998 16.1 7.7 3.0   20.3 9.2 3.1
1999 15.2 7.0 2.8   18.3 7.7 2.7
2000 14.6 6.7 2.5   17.6 7.6 2.4
2001 14.7 7.2 3.0   17.3 8.0 3.0
2002 14.8 7.1 3.1   16.8 7.4 2.9
2003 14.9 7.3 3.2   16.8 7.6 3.0
2004 14.8 7.3 3.3   16.2 7.3 2.8
2005 15.0 7.4 3.3   16.7 7.6 3.0
2006 14.7 7.3 3.1   16.5 7.6 2.6
2007 14.9 7.3 3.1   17.0 7.5 2.7
2008 14.6 7.4 3.3   16.0 7.4 2.9
2009 14.1 7.0 3.3   14.7 6.8 2.8
2010 14.8 7.4 3.3   15.3 7.1 2.6

 

Table 2: Effects on Poverty (Below 100% of Poverty Line) of Specified Income Sources
Poverty measure based on NAS methods and 2010 NAS poverty line adjusted for inflation. TANF, SSI, and food stamp income is corrected for underreporting using data from HHS’s TRIM model. See Table 1 for additional notes.
Number (in thousands)
  All Ages   Under 18
  1995 2000 2005 2010   1995 2000 2005 2010
Population 264,314 276,540 293,834 306,110   71,148 72,553 73,985 74,916
Poor 47,512 40,423 44,132 45,215   16,112 12,754 12,374 11,476
As a percentage of population 18.0% 14.6% 15.0% 14.8%   22.6% 17.6% 16.7% 15.3%
 
  Would be poor if the following family income were not counted:
  All public benefits* 72,166 65,075 74,412 90,251   21,246 17,526 19,053 23,360
  Social Security 66,817 59,647 65,006 68,973   17,545 13,868 13,614 12,972
  Unemployment Insurance 48,917 41,385 45,269 50,023   16,680 13,155 12,749 12,761
  Means-tested benefits* 60,822 52,111 59,722 69,016   21,947 17,642 19,132 22,181
  TANF 50,670 41,733 45,408 46,524   17,926 13,479 13,076 12,228
  SSI 50,622 43,796 47,552 49,265   17,132 13,742 13,359 12,634
  SNAP (formerly food stamps) 51,426 42,739 47,773 53,656   18,408 14,144 14,467 15,713
  Housing assistance 51,696 43,945 48,298 49,900   17,949 14,065 13,958 13,188
  EITC and Child Tax Credit** 52,339 46,577 52,660 55,782   18,585 16,067 16,798 16,984
  (Example: In 1995, 51.4 million Americans lived in families whose non-food stamp income was below the NAS poverty line.)
 
  Lifted above the poverty line by the following family income:***
  All public benefits* 24,654 24,652 30,280 45,036   5,134 4,772 6,679 11,884
  Social Security 19,305 19,224 20,874 23,758   1,433 1,114 1,240 1,496
  Unemployment Insurance 1,405 962 1,137 4,808   568 401 375 1,285
  Means-tested benefits* 13,310 11,688 15,590 23,801   5,835 4,888 6,758 10,705
  TANF 3,158 1,310 1,276 1,309   1,814 725 702 752
  SSI 3,110 3,373 3,420 4,050   1,020 988 985 1,158
  SNAP (formerly food stamps) 3,914 2,316 3,641 8,441   2,296 1,390 2,093 4,237
  Housing assistance 4,184 3,522 4,166 4,685   1,837 1,311 1,584 1,712
  EITC and Child Tax Credit** 4,827 6,154 8,528 10,567   2,473 3,313 4,424 5,508
  (Example: In 1995, 3.9 million Americans lived in families whose income was lifted above the poverty line by food stamps.)
 
  Lifted above, as a percentage of those who would be poor without the specified income source:
  All public benefits* 34 38 41 50   24 27 35 51
  Social Security 29 32 32 34   8 8 9 12
  Unemployment Insurance 3 2 3 10   3 3 3 10
  Means-tested benefits* 22 22 26 34   27 28 35 48
  TANF 6 3 3 3   10 5 5 6
  SSI 6 8 7 8   6 7 7 9
  SNAP (formerly food stamps) 8 5 8 16   12 10 14 27
  Housing assistance 8 8 9 9   10 9 11 13
  EITC and Child Tax Credit** 9 13 16 19   13 21 26 32
  (Example: In 1995, food stamps lifted above the poverty line 8 percent of Americans who would otherwise have been poor.)

* “All public benefits” includes the listed programs plus workers’ compensation, veterans’ benefits, home energy assistance, the 2010 Making Work Pay Tax Credit, and the net effect of taxes paid. “Means-tested” benefits are TANF, SSI, SNAP, housing and energy assistance, means-tested veterans’ benefits, the EITC, and the refundable portion of the Child Tax Credit.

** Includes both the refundable (low-income) and non-refundable parts of the Child Tax Credit.

*** Equals the number that would be poor without this income minus the actual number in poverty.

 

Table 3: Effects on Deep Poverty (Below 75% of Poverty Line) of Specified Income Sources
Poverty measure based on NAS methods and 2010 NAS poverty line adjusted for inflation. TANF, SSI, and food stamp income is corrected for underreporting using data from HHS’s TRIM model. See Table 1 for additional notes.
Number (in thousands)
    All Ages   Under 18
    1995 2000 2005 2010   1995 2000 2005 2010
Population 264,314 276,540 293,834 306,110   71,148 72,553 73,985 74,916
Below 75% of Poverty Line 20,477 18,620 21,644 22,779   6,408 5,509 5,646 5,305
  As a percentage of population 7.7% 6.7% 7.4% 7.4%   9.0% 7.6% 7.6% 7.1%
 
  Would be below 75% of poverty if the following family income were not counted:
  All public benefits* 57,540 50,217 58,696 72,548   16,668 12,647 14,254 17,995
  Social Security 40,333 38,104 43,456 46,515   7,747 6,576 6,931 6,670
  Unemployment Insurance 21,576 19,252 22,349 26,553   6,823 5,731 5,845 6,272
  Means-tested benefits* 40,751 33,559 39,479 47,068   15,984 12,072 13,491 15,810
  TANF 25,900 20,340 23,353 24,052   9,698 6,510 6,663 6,002
  SSI 24,780 22,427 25,877 27,562   7,632 6,437 6,706 6,437
  SNAP (formerly food stamps) 26,834 22,159 26,702 31,113   9,931 7,400 8,257 9,128
  Housing assistance 24,904 21,967 25,502 25,864   8,689 7,098 7,295 6,629
  EITC and Child Tax Credit** 25,000 23,171 26,787 28,138   8,752 7,864 8,282 7,987
  (Example: In 1995, 26.8 million Americans lived in families whose non-food stamp income was below 75% of the NAS poverty line.)
 
  Lifted above 75% of the poverty line by the following family income:***
  All public benefits* 37,063 31,597 37,052 49,769   10,260 7,138 8,608 12,690
  Social Security 19,856 19,484 21,812 23,736   1,339 1,067 1,285 1,365
  Unemployment Insurance 1,099 632 705 3,774   415 222 199 967
  Means-tested benefits* 20,274 14,939 17,835 24,289   9,576 6,563 7,845 10,505
  TANF 5,423 1,720 1,709 1,273   3,290 1,001 1,017 697
  SSI 4,303 3,807 4,233 4,783   1,224 928 1,060 1,132
  SNAP (formerly food stamps) 6,357 3,539 5,058 8,334   3,523 1,891 2,611 3,823
  Housing assistance 4,427 3,347 3,858 3,085   2,281 1,589 1,649 1,324
 
EITC and Child Tax Credit**
4,523 4,551 5,143 5,359   2,344 2,355 2,636 2,682
  (Example: In 1995, nearly 6.4 million Americans lived in families whose income was lifted above 75% of the poverty line by food stamps.)
 
  Lifted above, as a percentage of those who would be below 75% without the specified income source:
  All public benefits* 64 63 63 69   62 56 60 71
  Social Security 49 51 50 51   17 16 19 20
  Unemployment Insurance 5 3 3 14   6 4 3 15
  Means-tested benefits* 50 45 45 52   60 54 58 66
  TANF 21 8 7 5   34 15 15 12
  SSI 17 17 16 17   16 14 16 18
  SNAP (formerly food stamps) 24 16 19 27   35 26 32 42
  Housing assistance 18 15 15 12   26 22 23 20
  EITC and Child Tax Credit** 18 20 19 19   27 30 32 34
 
(Example: In 1995, food stamps lifted above 75% of the poverty line 24 percent of Americans who would otherwise have been that deeply poor.)

* “All public benefits” includes the listed programs plus workers’ compensation, veterans’ benefits, home energy assistance, the 2010 Making Work Pay Tax Credit, and the net effect of taxes paid. “Means-tested” benefits are TANF, SSI, SNAP, housing and energy assistance, means-tested veterans’ benefits, the EITC, and the refundable portion of the Child Tax Credit.

** Includes both the refundable (low-income) and non-refundable parts of the Child Tax Credit.

*** Equals the number that would be below 75 percent of the poverty line without this income minus the actual number below 75 percent of poverty.

 

Table 4: Effects on Deep Poverty (Below 50% of Poverty Line) of Specified Income Sources
Poverty measure based on NAS methods and 2010 NAS poverty line adjusted for inflation. TANF, SSI, and food stamp income is corrected for underreporting using data from HHS’s TRIM model. See Table 1 for additional notes.
Number (in thousands)
    All Ages   Under 18
    1995 2000 2005 2010   1995 2000 2005 2010
Population 264,314 276,540 293,834 306,110   71,148 72,553 73,985 74,916
Below 50% of Poverty Line 6,967 7,024 9,554 10,177   1,523 1,746 2,195 1,974
  As a percentage of population 2.6% 2.5% 3.3% 3.3%   2.1% 2.4% 3.0% 2.6%
 
  Would be below 50% of poverty if the following family income were not counted:
  All public benefits* 43,412 36,643 44,151 55,142   12,226 8,258 9,780 12,602
  Social Security 23,659 23,445 28,474 30,686   2,475 2,432 3,170 3,021
  Unemployment Insurance 7,354 7,238 9,883 12,691   1,629 1,816 2,271 2,603
  Means-tested benefits* 25,664 19,358 23,765 28,988   10,823 7,204 8,319 9,943
  TANF 10,782 8,099 10,524 11,221   3,943 2,411 2,804 2,609
  SSI 10,290 10,294 13,245 14,345   2,264 2,349 2,913 2,877
  SNAP (formerly food stamps) 11,253 9,217 12,668 15,953   4,037 2,916 3,789 4,635
  Housing assistance 7,990 8,093 10,799 11,171   1,994 2,257 2,749 2,353
  EITC and Child Tax Credit** 8,195 8,715 11,122 11,717   2,122 2,590 2,927 2,677
  (Example: In 1995, 11.3 million Americans lived in families whose non-food stamp income was below the 50% of the NAS poverty line.)
 
  Lifted above 50% of the poverty line by the following family income:***
  All public benefits* 36,445 29,619 34,597 44,965   10,703 6,512 7,585 10,628
  Social Security 16,692 16,421 18,920 20,509   952 686 975 1,047
  Unemployment Insurance 387 214 329 2,514   106 70 76 629
  Means-tested benefits*
18,697
12,334 14,211 18,811   9,300 5,458 6,124 7,969
  TANF 3,815 1,075 970 1,044   2,420 665 609 635
  SSI 3,323 3,270 3,691 4,168   741 603 718 903
  SNAP (formerly food stamps) 4,286 2,193 3,114 5,776   2,514 1,170 1,594 2,661
  Housing assistance 1,023 1,069 1,245 994   471 511 554 379
  EITC and Child Tax Credit** 1,228 1,691 1,568 1,540   599 844 732 703
  (Example: In 1995, 4.3 million Americans lived in families whose income was lifted above 50% of the poverty line by food stamps.)
 
  All public benefits* 84 81 78 82   88 79 78 84
  Social Security 71 70 66 67   38 28 31 35
  Unemployment Insurance 5 3 3 20   7 4 3  
  Means-tested benefits* 73 64 60 65   86 76 74 80
  TANF 35 13 9 9   61 28 22 24
  SSI 32 32 28 29   33 26 25 31
  SNAP (formerly food stamps) 38 24 25 36   62 40 42 57
  Housing assistance 13 13 12 9   24 23 20 16
  EITC and Child Tax Credit** 15 19 14 13   28 33 25 26
  (Example: In 1995, food stamps lifted above 50% of the poverty line 38 percent of Americans who would otherwise have been that deeply poor.)

* “All public benefits” includes the listed programs plus workers’ compensation, veterans’ benefits, home energy assistance, the 2010 Making Work Pay Tax Credit, and the net effect of taxes paid. “Means-tested” benefits are TANF, SSI, SNAP, housing and energy assistance, means-tested veterans benefits, the EITC, and the refundable portion of the Child Tax Credit.

** Includes both the refundable (low-income) and non-refundable parts of the Child Tax Credit.

*** Equals the number that would be below 50 percent of the poverty line without this income minus the actual number below 50 percent of poverty.

 

Table 5: Persons in Unmarried Families with Children:
Percent in Deep Poverty Counting and Not Counting Public Benefits and Taxes

Poverty measure based on NAS methods and 2010 NAS poverty line adjusted for inflation.TANF, SSI, and food stamp income is corrected for underreporting using data from HHS’s TRIM model.

“Long-term” unmarried partners (those present in survey month and 12 months earlier) of the head of household are counted as part of the head of household’s family; i.e., their incomes are combined when determining poverty status, as are their needs. “Short-term” partners (all others) are treated as a separate family unit.

In multi-family or multi-generational households, marital status in this table reflects the immediate family, also known as the subfamily.

See Table 1 for additional notes.

Number (in thousands)
    1995 2000 2005 2010
Total Persons (All Incomes) in Unmarried Families
  Total 38,662 37,959 42,788 47,168
  In female-headed families, no partner present 29,299 26,578 29,388 31,410
  In male-headed families, no partner present 4,766 4,996 5,649 5,284
  Head of household lives with long-term partner 3,043 4,593 5,887 8,100
  Head of household lives with short-term partner 1,554 1,792 1,865 2,373
 
Percent Below Half of Poverty Line
  Total 2.8% 4.1% 5.3% 4.6%
  In female-headed families, no partner present 2.7% 4.1% 5.7% 4.6%
  In male-headed families, no partner present 2.6% 3.3% 3.9% 2.6%
  Head of household lives with long-term partner 0.6% 1.4% 2.5% 2.4%
  Head of household lives with short-term partner 9.5% 12.7% 11.9% 16.9%
 
Percent Below Half of Poverty Line Not Counting Public Benefits and Taxes
  Total 35.6% 24.2% 27.3% 31.6%
  In female-headed families, no partner present 40.2% 27.9% 32.4% 35.9%
  In male-headed families, no partner present 18.0% 16.0% 15.3% 20.8%
  Head of household lives with long-term partner 15.8% 8.5% 9.7% 17.9%
  Head of household lives with short-term partner 42.6% 32.1% 39.2% 46.1%

End Notes

[1] Because the SPM is available only since 2009, it cannot be used for this analysis of trends since 1995, but a modestly modified version of the SPM can be used.  As explained in footnote 5 and in Appendix C, the poverty measure used here differs from the federal SPM in several technical ways, including using a 2010 poverty line adjusted only for inflation rather than for changes in living standards (measured by what Americans spend on basic needs).  (For an analysis that combines 2012 SPM data with corrections for underreported benefits similar to those used here, see Arloc Sherman and Danilo Trisi, “Safety Net More Effective Against Poverty Than Previously Thought,” Center on Budget and Policy Priorities, May 6, 2015, https://www.cbpp.org/research/poverty-and-inequality/safety-net-more-effective-against-poverty-than-previously-thought.)

[4] We correct the tendency of Census Bureau data to underreport income from three government assistance programs: Temporary Assistance for Needy Families, Supplemental Security Income, and SNAP (formerly food stamps).  The corrections come from the Transfer Income Model (TRIM) policy micro-simulation model developed by the Urban Institute.  TRIM starts with Census survey data but adjusts those data to more closely match actual numbers and characteristics of benefit recipients shown in program records.

[5] Differences between the poverty measure used in this analysis and the federal SPM largely reflect data limitations in the early years of our analysis.  Among these differences: (1) we adjust the poverty threshold over time only for inflation, while the threshold in the SPM grows from year to year with expenditures for basic needs; (2) our income measure does not count WIC or subtract child support paid; (3) our poverty thresholds do not vary by homeownership status; (4) we must use approximations for the value of out-of-pocket medical and work expenses (which both our measure and the SPM subtract from income) as well as rent levels (used to adjust the poverty thresholds locally) and rent subsidies; and (5) although our measure (like the SPM) expands who is counted in the family unit when determining poverty status, we do not include everyone who is included in the SPM.  For details, see Appendix C.

[6] Health insurance is not counted as income in the NAS-based measure we use, but having health insurance can reduce poverty and deep poverty rates under our poverty measure by reducing out-of-pocket medical expenditures.

[7] TANF’s low eligibility ceilings mean that many poor families aren’t eligible for TANF, but the ratio of families receiving TANF cash benefits to all poor families with children also fell dramatically during this period.  In 1995, for every 100 poor families with children (using the official poverty statistics), 76 families received cash assistance from AFDC, TANF’s predecessor.  By contrast, in 2005, for every 100 poor families with children, only 35 received cash assistance through TANF, and that figure is even lower today.

[8] The HHS estimates of declining participation rates in TANF do not include people affected by TANF time limits. Time limits — including the 60-month lifetime federal limit on receiving TANF assistance enacted in 1996 and, in many states, shorter state time limits — are an additional factor that contributed to the weakening of the cash welfare assistance component of the safety net.

[9] Stephen Freedman et al., National Evaluation of Welfare-to-Work Strategies Evaluating Alternative Welfare-to-Work Approaches: Two-Year Impacts for Eleven Programs, U.S. Department of Health and Human Services, June 2000, page ES-35, Exhibit ES-10, www.mdrc.org/publication/evaluating-alternative-welfare-work-approaches/file-full.  Ten of the 11 programs showed some increase in the share of participants below half the poverty line (six of them statistically significant), while nine showed declines in the overall poverty rate (five of them significant).  Taking a simple average across all 11 sites, deep poverty rose by 2.7 percentage points while regular poverty declined by 2.1 percentage points.  The study measured poverty using participant earnings, welfare, and food stamp income from administrative sources. 

[10] Freedman et al., Exhibit ES-9, page ES-34.

[11] Marianne Bitler, Jonah Gelbach, and Hilary Hoynes, “What Mean Impacts Miss: Distributional Effects of Welfare Reform Experiments,” American Economic Review, vol. 96, no. 4 (September 2006), www.ssc.wisc.edu/~scholz/Teaching_742/Bitler_Gelbach_Hoynes.pdf.

[12] Our poverty measure treats unmarried partners of the household head as part of the head’s family unit if they were consistently present, that is, living in the household both at the time of the survey and 12 months earlier.  See Appendix C for details.

[13] George J. Borjas, “Food Insecurity and Public Assistance,” National Bureau of Economic Research, NBER Working Paper 9236, October 2002, www.nber.org/papers/w9236.pdf.

[14] Arloc Sherman, “Poverty and Financial Distress Would Have Been Substantially Worse in 2010 Without Government Action, New Census Data Show,” Center on Budget and Policy Priorities, November 7, 2011, https://www.cbpp.org/research/poverty-and-financial-distress-would-have-been-substantially-worse-in-2010-without.

[15] Chuck Marr, Bryann DaSilva, and Arloc Sherman, “Letting Key Provisions of Working-Family Tax Credits Expire Would Push 16 Million People Into or Deeper Into Poverty,” Center on Budget and Policy Priorities, Updated February 20, 2015, https://www.cbpp.org/research/letting-key-provisions-of-working-family-tax-credits-expire-would-push-16-million-people.

[16] Arloc Sherman, “Why Isn’t Poverty Falling? Weakening of Unemployment Insurance Is a Pivotal Factor,” Center on Budget and Policy Priorities, October 7, 2013, https://www.cbpp.org/research/why-isnt-poverty-falling-weakening-of-unemployment-insurance-is-a-pivotal-factor.

[17] A more comprehensive examination of the missing benefits yields similar results.  When one includes two additional benefits, SNAP and SSI, in addition to TANF, and adjusts for the effects of the growing national population by expressing the underreported amounts in dollars per capita, the inflation-adjusted value of benefits missed by the CPS fell from $123 per person in 1995 to $105 per person in 2005.

[18] This alternative method, which serves to check our TRIM findings, differs from our main method in three ways:  it does not use TRIM adjustments; it uses official poverty guidelines rather than an NAS-based threshold; and, unlike the NAS method, it uses family disposable income (after taxes and noncash benefits) without netting out work expenses or medical out-of-pocket expenses or altering the family unit for which poverty status is determined.  To adjust for underreported income, the method starts with the share of children below half the poverty line not including SNAP, AFDC/TANF, or SSI (11.0 percent in 1995 and 7.5 percent in 2005).  It finds the share of children lifted out of deep poverty by each of the three programs in the uncorrected Census data for those years (cumulatively, 6.8 percent of children in 1995 and 3.2 percent in 2005).  For each program, it adjusts those uncorrected shares by dividing them by the ratio of Census-reported benefit payments to true aggregate benefit payments (which we find to be 65 percent in 1995 and 55 percent in 2005 for SNAP; 70 and 48 percent for AFDC/TANF; and 71 and 78 percent for SSI).  Finally, it applies those adjusted anti-poverty effects to the pre-transfer deep poverty rate to yield an adjusted deep poverty rate (1.0 percent in 1995 and 1.9 percent in 2005).  Although these levels are lower than in our TRIM/NAS-based calculations (in part because this method does not subtract medical and work expenses and uses the lower, official poverty line), the increases in both methods are of the same magnitude.

[19] Christopher Wimer et al., “Trends in Poverty with an Anchored Supplemental Poverty Measure,” Columbia Population Research Center, December 2013; Yonatan Ben-Shalom, Robert A. Moffitt, and John Karl Scholz, “An Assessment of the Effectiveness of Anti-Poverty Programs in the United States,” prepared for the 2012 Oxford Handbook of the Economics of Poverty, chapter 22; H. Luke Shaefer and Kathryn Edin, “Rising Extreme Poverty in the United States and the Response of Federal Means-Tested Transfer Programs,” National Poverty Center Working Paper 13-06, May 2013.

[20] Bruce D. Meyer, Wallace K. C. Mok, and James X. Sullivan, “The Under-Reporting of Transfers in Household Surveys: Its Nature And Consequences,” National Bureau of Economic Research, NBER Working Paper 15181, July 2009, http://www.nber.org/papers/w15181.

[21] Arloc Sherman, “Safety Net Effective at Fighting Poverty But Has Weakened for the Very Poorest,” Center on Budget and Policy Priorities, July 6, 2009, https://www.cbpp.org/research/safety-net-effective-at-fighting-poverty-but-has-weakened-for-the-very-poorest.

[22] National Research Council, Measuring Poverty: A New Approach(National Academy Press: 1995).

[23] Housing assistance includes federal, state, and local housing vouchers and public housing.  We calculate the value of rental assistance using a method that the Census Bureau developed for use with its NAS-based poverty measures. Specifically, for each assisted household, the annual value of housing assistance is 12 times the local monthly fair market rent reduced by the household’s required contribution, approximated as 30 percent of its annual cash income.  (We approximate fair market rents using a weighted average of HUD local Fair Market Rent levels for each family's state, broken down further by whether the family lives in a metropolitan or non-metropolitan area, from data available at www.huduser.org/portal/datasets/fmr.html.  We estimate the number of bedrooms in each housing unit using HUD occupancy rules.)  When more than one family or unrelated individual shares the same apartment, we assign each its per capita share of the household’s housing subsidy.  Since housing assistance cannot be used to meet other needs such as food or clothing, we, like the Census Bureau, place a cap on the value of housing assistance equal to housing's share of a poverty-level budget (that is, 44 percent of the poverty line for a family of that size and composition).

[24] Using TAXSIM means that we give up some features of the Census model, such as imputation of itemized deductions, but in return we are assured of a consistent model over time.  For details on the NBER TAXSIM version9 model, see www.nber.org/~taxsim.  Using TAXSIM with the CPS requires determining which individuals file taxes together.  In general, we treat each nuclear family as a tax filing unit; however, we assume that foster children are dependents of the head of the housing unit, as are certain clusters of relatives (called related subfamilies, typically a child of the head of household raising children) if they are living at home but have no earnings.

[25] TRIM is developed and maintained by the Urban Institute under contract with the Office of the Assistant Secretary for Planning and Evaluation at the Department of Health and Human Services.  Documentation of the TRIM model is at http://trim3.urban.org/T3Technical.php.  While the model was developed chiefly to allow users to compare current policies with proposed policies, we use data only for TRIM’s “baseline” (or current-policy) scenario.

[26] In producing the CPS files, the Census Bureau, like TRIM, also assigns benefits for some people with missing data. Unlike TRIM, however, Census does not use this process to try to match the actual number of recipients shown in program records.

[27] The CPS recently added questions on MOOP and child care expenses, but for the early years of our analysis, these data are not available.  To ensure consistency, we approximate them for each year using formulas adapted from the Census Bureau and BLS.  For out-of-pocket medical spending, we use the medical portion of the poverty line in a version of the NAS thresholds (described in the next bullet in the text) that includes out-of-pocket medical spending as a basic need.  We start with the medical share of the poverty line in 2010 — that is, 7.23 percent of the NAS-based poverty line for a family of two adults and two children (www.bls.gov/pir/spm/spm_threshold_200910.xls), or $1,989 — and adjust that amount up or down using a set of ratios provided by the Census Bureau that depend on the family's size, health insurance status, and age, thus assigning to each family a level of medical out-of-pocket (MOOP) expenditures in 2010 typical for its family type.  We deflate these values for other years using the overall CPI-U, as Census and BLS do when calculating NAS-based measures that use a fixed CPI-based threshold and include MOOP in the threshold.
Our estimates of work expenses also rely on formulas provided by the Census Bureau.  The formulas are based on data on median weekly out-of-pocket child care expenses and other work-related expenditures from the Survey of Income and Program Participation (SIPP). Values vary depending on year, weeks worked, number and ages of children (in the case of the child care formula), and other family characteristics.  (We use weekly values provided by the Census Bureau for 1996 through 2006; for other years, we extrapolate values using the CPI-U.)  Whether the family used paid child care in a given year is estimated with a probabilistic model.  The estimated value of work expenses is capped at the value of the worker’s earnings.  A couple’s child care expenses are further capped based on the earnings of the lower-earning spouse.

[28] Short-term partners may contribute less to family finances than long-term partners, either because they feel less invested in the family or simply because couples in which the partner for whatever reason contributes less tend to break up sooner.  We found evidence for this when we examined data on family hardships.  In an unpublished analysis, we examined 402 single parents living below the poverty line (according to the official annual cash poverty definition) with an unmarried partner, using SIPP hardship questionnaires from 1998, 2003, 2005, and 2011 and income data for the preceding 12 months.  Among these parents, having a partner with enough income to lift them and their children above the poverty line for the year was associated with significantly fewer material hardships, relative to having a poorer partner — but only ifthe partner was stable, that is, living with the family consistently for the previous 12 months.  For other, non-stable partners, bringing in enough income (while cohabiting) to lift the family’s annual income above the official poverty line for the year did not reduce the number of reported hardships.  This difference between stable and non-stable partners was statistically significant and suggests that the case for including long-term partners in the family unit for the purposes of poverty determination is more compelling than the case for including short-term partners.  (The analysis examined four hardships: food insecurity, falling behind on the rent or mortgage, having phone service cut off for failure to pay, and falling behind on gas or electric bills.  Results were similar after controlling for differences in race, ethnicity, gender, number of children, parent education, age, disability, and whether the partner was a biological parent of the children.)

[29] Following the approach used in the SPM and TRIM data, we treat household members younger than 15 living with no relatives as family members of the head of household.  This affects fewer than 200,000 poor children in our data in both 1995 and 2005.

[30] See “Poverty Thresholds for Two-Adult-Two-Child Family Following NAS Recommendations: 1999-2010,” downloaded from http://www.census.gov/hhes/povmeas/data/nas/tables/2010/web_tab5_povertythres2010.xls  (column 3).  Note that although the NAS poverty line was calculated by BLS and is found on the Census website, it may be too low.  While it is designed to reflect a minimal spending level for food, clothing, shelter, utilities, and medical needs, it leaves out the portion of shelter expenses that homeowners pay for mortgage principal. NAS poverty lines have excluded any amount for mortgage principal payments since the original 1995 NAS report, which noted that these payments were being left out for reasons of “processing convenience.”  Preferably, the report said, an amount for these payments would be included in the NAS poverty line.  Recently, BLS staff have developed a method for doing so.  This would raise the poverty line by about $2,000, which would raise the estimated number of people in poverty and deep poverty.

[31] See Thesia I. Garner and Kathleen S. Short, “Creating a Consistent Poverty Measure Over Time Using NAS Procedures: 1996-2005,” NAS Poverty Measurement Working paper, April 3, 2008, available at www.census.gov/hhes/povmeas/publications/povthres/experimental_measures_96_05v7.pdf.

[32] All else being equal, using the NAS-based poverty threshold instead of the official poverty line raises the poverty rate in 2010, as previously shown in Figure 2.  It also raises the percentage of people who would be poor when government taxes and benefits are not counted (from 25.3 percent to 31.2 percent).

[33] For the sake of consistency with the original NAS recommendations, we subtract our estimate of MOOP, described earlier, from both sides of the poverty equation — thresholds and income.  This has no effect on families’ poverty status but lowers the poverty threshold and, conceptually, treats medical expenditures as a deduction rather than a basic need.

[34] The NAS panel recommended updating poverty lines each year using a method that tends to rise slightly faster than the CPI.  (Under that approach, the poverty line is adjusted each year by the percentage change in median family expenditures on basic needs for U.S. two-child two-adult families.)  This alternative adjustment method is not comparable across recent years, however, due to improvements introduced in 2007 in the Consumer Expenditure Survey, the survey used to track spending on basic needs.  Therefore, while our analysis uses the NAS threshold for 2010 as a starting place, we adjust it over time using the CPI.  If we had used the NAS method of adjusting our thresholds over time (as we did in our original 2009 analysis of deep poverty), our poverty line would have risen somewhat faster, our estimated number of people in poverty would have fallen somewhat less, and the number in deep poverty would have risen somewhat more.

[35] Adjustments for family size use the “three-parameter” scale described by Garner and Short (see footnote 31 above) applied to the non-medical portion of the NAS poverty threshold.  Geographic adjustments are based on the ratio between the local and national HUD fair market rent (described in footnote 23), applied to the housing portion of the threshold.

[36] Income below 75 percent of the poverty line, as measured by our approach, is a more stringent definition of deep poverty than it may appear.  Traditionally, analysts have considered those below 50 percent of the poverty line to be in “deep” poverty.  Typically, however, the 50 percent threshold has been used in conjunction with a measure of cash income.  That is, a family with cash income below 50 percent of the official poverty line was considered to be in “deep poverty” even if the family received food stamps or housing assistance that lifted its total purchasing power above this threshold.  The 50 percent threshold represents a much stricter definition of deep poverty when one adopts, as we do, a broader poverty measure that includes noncash and tax-based benefits as income and corrects for underreporting of benefits.  The percentage of the population with cash below half the poverty line in 2010 using the official measure (6.7 percent) was closer to the percentage below 75 percent of the poverty line by our measure (7.4 percent) than to the percentage below half the poverty under our measure (3.3 percent).  In that sense, a 75-percent-of poverty standard under an NAS-based TRIM-corrected measure may be more comparable in severity to the traditional 50-percent-of-poverty standard under the official cash measure. 

[37] These calculations assume that a family’s other income would remain unchanged.  Although people’s behavior might change in the absence of government assistance, a recent comprehensive review of the literature by some leading economists in the field concludes that this would have a “negligible” effect on overall poverty rates.  Yonatan Ben-Shalom, Robert A. Moffitt, and John Karl Scholz, “An Assessment of the Effectiveness of Anti-Poverty Programs in the United States,” prepared for the 2012 Oxford Handbook of the Economics of Poverty, chapter 22.