Government’s Pandemic Response Turned a Would-Be Poverty Surge Into a Record Poverty Decline
Bolstered by pandemic relief, economic security programs kept 53 million people above the poverty line in both 2020 and 2021, far surpassing the previous 2009 record of 40 million people protected from poverty. Public programs transformed what, without the programs, would have been a near-record surge in poverty in the COVID-19 recession into a record one-year overall poverty decline in 2020 and a record one-year decline in children’s poverty in 2021. The strong policy response brought the U.S. poverty rate in 2020 to its lowest level on record in data back to 1967, and then to a new record low of 7.8 percent in 2021, an analysis of Census and other historical data shows. Government aid helped prevent a sharp rise in poverty for all major racial and ethnic groups and narrowed differences in poverty rates between groups, yet wide inequities remain.
In 2021, the American Rescue Plan’s Child Tax Credit expansion and the third round of stimulus payments, in conjunction with other relief measures, spurred the largest one-year drop in child poverty on record, driving child poverty down to an all-time low of 5.2 percent. The Rescue Plan by itself lifted more people, including more children, above the poverty line with government assistance in a single year than any other piece of legislation enacted in more than 50 years. Children are particularly vulnerable to poverty’s long-term effects on health, education, and earnings.
The strong government response prevented the pandemic economy from adding further to the already disproportionate risk of poverty borne by Black and Latino individuals, including children. Government policies helped keep 30 percent of all Black children and as many as 24 percent of Latino children above the poverty line in 2021, as well as 10 percent of Asian and white children.
With the expiration of temporary pandemic assistance, poverty likely rose in 2022. The Census Bureau will release 2022 poverty data in September 2023.
These findings emerge from our analysis of the Census Bureau’s Current Population Survey (CPS) data and historical Supplemental Poverty Measure (SPM) data produced by Columbia University’s Center on Poverty and Social Policy. We use the Census Bureau’s SPM data when available (for 2009 through 2021) and the Columbia SPM data for prior years. We create an “anchored” series by using the 2021 SPM poverty line adjusted for inflation. Unlike the official poverty measure, the SPM counts income from non-cash programs such as the Supplemental Nutrition Assistance Program (SNAP, formerly known as the Food Stamp Program) and rental assistance, and from refundable tax credits for families with low or moderate incomes — the Earned Income Tax Credit (EITC) and Child Tax Credit — as well as cash benefits such as Social Security and unemployment insurance.
For many families these annual poverty figures likely obscure periods of turbulence and uncertainty within the year. In 2020, in particular, millions of jobless families spent weeks waiting for delayed unemployment insurance and other benefits, some households did not qualify for aid, and many of the benefits that Census figures count as received in 2020 arrived late in the year or not until the following year. But the net reduction in annual poverty during a time of deep economic crisis is remarkable, nonetheless.
COVID Relief Achieved Historic Gains Against Poverty Overall and for Children
Government policies played a historic role in reducing poverty in 2020 and 2021, building on a longer-term trend of government assistance becoming more effective at reducing poverty. The growing effectiveness of economic security programs helped bring down poverty rates over the last half-century.
Economic security programs lowered the number of people in poverty by just 1 percent in 1967. By 2019, that figure was 46 percent. That is, in 2019, before accounting for government benefits, about 74 million people had pre-tax incomes below the poverty line. Counting benefits (including tax credits) and subtracting taxes lowers that number by 34 million people (or 46 percent). In 2020 and 2021, bolstered by pandemic relief, economic security programs reduced poverty by 64 percent and 67 percent respectively — all-time highs in poverty-reduction effectiveness with data back to 1967.
Changes in the Census Bureau’s survey methods make data after 2017 not strictly comparable to earlier years. But the Census Bureau provides enough data about this survey transition to make clear that the anti-poverty effectiveness of economic security programs reached all-time highs in 2020 and 2021, and the SPM poverty rate reached record lows in 2020 and 2021 by our measure (the SPM with an inflation-adjusted poverty line).
The SPM poverty rate in 2021 was 7.8 percent, a record low, and well below the rate of 12.2 percent in 2019 just before the pandemic and much lower than the 27.6 percent rate in 1967 by our measure. Not considering government assistance or taxes, poverty has improved only modestly over the past five decades, falling from 28.0 percent to 23.8 percent between 1967 and 2021. (See Figure 1.)
The impact of government poverty-reduction efforts has been just as striking among children, a particularly important success given child poverty’s long-lasting consequences and the health and economic benefits of reducing child poverty.
The one-year drop in child poverty in 2021 was the largest ever, in data back to 1967. (See Figure 2.) The year-over-year decline of 4.4 percentage points brought the child poverty rate in 2021 to an all-time low of 5.2 percent.
In 2019, government policies cut child poverty by 40 percent. Bolstered by pandemic relief legislation, economic security programs cut child poverty by 60 percent in 2020 and 76 percent in 2021, both all-time highs with data back to 1967. In the late 1960s, in contrast, counting government assistance and taxes made child poverty modestly higher because the federal tax code at the time taxed a substantial number of families with children into poverty (or deeper into poverty).
Since that time, federal policymakers have established and expanded the EITC and the Child Tax Credit. Nationwide implementation of SNAP in the 1970s and its increased effectiveness in reaching more eligible people in more recent decades have lifted millions of additional children out of poverty as compared to the late 1960s. In 2021, counting government assistance and taxes drops the child poverty rate from 21.8 percent to 5.2 percent. (See Figure 3.)
By lowering poverty, economic security programs can make a lasting difference, particularly for children. A congressionally chartered report from the National Academy of Sciences noted the difference that anti-poverty programs can make: “The weight of the causal evidence indicates that income poverty itself causes negative child outcomes . . . . Many programs that alleviate poverty either directly, by providing income transfers, or indirectly, by providing food, housing, or medical care, have been shown to improve child well-being.”
Recent studies of nutrition assistance and the Earned Income Tax Credit, for example, have found evidence that assistance early in childhood leads to children earning more as young adults. Other studies have linked various income assistance programs such as food assistance and housing subsidies with healthier birthweights, lower stress for mothers (measured by reduced stress hormone levels in the women’s bloodstream), better nutrition in childhood, higher scores on reading and math tests and other measures of school performance, more high school completion, higher rates of college entry, and better health in adulthood.
Economic Recovery Legislation Drove Historic Poverty Reductions in Both 2020 and 2021
The historic poverty reduction in 2020 and 2021 was driven by four major legislative packages enacted in 2020 and 2021 that provided stimulus payments, expanded unemployment insurance, increased nutrition and housing assistance, improved tax credits, and made health coverage more accessible.
In both 2020 and 2021, economic security programs bolstered by economic recovery legislation lowered the number of people in poverty by 53 million, far exceeding the prior record of 40 million set in 2009. The percentage-point reduction in the poverty rate due to government assistance net of taxes (16.2 percentage points in 2020 and 16.0 percentage points in 2021) likewise exceeded the previous record (13.2 percent in 2009).
The three policies that reduced poverty the most as measured by the Supplemental Poverty Measure were stimulus payments (formally known as Economic Impact Payments), expansions in unemployment insurance, and the expansion of the Child Tax Credit. All three of these policies were temporary and were no longer in effect by 2022.
Stimulus payments kept the annual incomes of 11.6 million people in 2020 and 9.0 million people in 2021 above the poverty line. Unemployment insurance benefits kept 5.6 million people out of poverty in 2020 and 2.3 million in 2021. (This compares to 0.5 million in 2019.) (See Figure 4.)
The temporary pandemic unemployment programs significantly increased the coverage, duration, and adequacy of unemployment benefits compared to regular unemployment insurance, substantially reducing hardship and providing important stabilization and impetus for recovery for a sharply declining economy. These steps were taken because the permanent unemployment insurance system does not cover many unemployed workers and often provides inadequate benefits.
The full poverty-reducing effect of unemployment insurance is likely even higher than the above figures suggest, moreover, because the survey data do not capture a substantial amount of unemployment benefits. According to one estimate, unemployment benefits’ effect on poverty in the pandemic was twice as high when one corrects for this underreporting. (See methodological appendix for more discussion of underreported income.)
The effectiveness of stimulus payments in reducing poverty partly reflected design and implementation improvements compared to similar stimulus payments in 2008, during the Great Recession. In the earlier crisis, payments went only to individuals who had filed tax returns, and only individuals with sufficient tax liability received the full amount. The 2020 and 2021 stimulus payments were the first time the IRS provided direct cash payments to households with no minimum earnings threshold or tax filing requirement, so people with the lowest incomes were eligible for the full rebate amount.
And, unlike in 2008, the Treasury Department was able to deliver benefits automatically to recipients of Social Security, Supplemental Security Income, railroad retirement, and certain veterans’ benefits, rather than forcing them to file tax returns that were otherwise unnecessary. The IRS created simpler portals for filing for the payments for those who don’t usually file a tax return, boosting participation, though significant numbers of people who were eligible for payments still missed out because they didn’t apply for them.
2020 Relief Led to Record Drop in Poverty
Government assistance in 2020 turned what would have been a near-record surge in poverty that year into a record one-year poverty decline. Not considering government assistance or taxes, the share of people in poverty would have risen in 2020 by the second-largest amount on record: 2.5 percentage points (8 million people). But when government assistance and tax policies are included, the share of people with annual income below the poverty line fell in 2020 by 3.1 percentage points (10 million people), the largest amount on record in data back to 1967. (See Figure 5.) (Poverty fell again in 2021 to a new record low, as described below.)
In 2019, 34 million fewer people had incomes below the poverty line when government assistance and taxes are considered (40 million compared with 74 million when they are not counted). In 2020, counting government assistance and taxes lowered the number of people in poverty by 53 million (from 82.3 million to 29.5 million).
Much of that poverty reduction came from the Coronavirus Aid, Relief, and Economic Security (CARES) Act, which included the first round of stimulus payments and a sizable expansion of unemployment benefits. These benefits lifted more people above the poverty line with government assistance — approximately 11 million people — than any other piece of legislation enacted since at least 1967, when our SPM poverty data begin.
Other legislation in 2020 — including temporary SNAP expansions enacted in March and a second round of stimulus payments enacted in December — brought annual income above the poverty line for several million additional people in that year.
American Rescue Plan Reduced Poverty Even More
The following year, the American Rescue Plan lifted even more people — approximately 16 million — above the poverty line through its stimulus payments and temporary expansions of the Child Tax Credit, unemployment insurance, and the Earned Income Tax Credit.
The Rescue Plan was particularly important for children. Two provisions of the law alone — the Child Tax Credit expansion and the third pandemic stimulus payment (which, unlike the CARES stimulus payment was as large for children as for adults) — lifted far more children above the poverty line with government assistance in a single year — more than 5 million children — than any other piece of legislation on record in data back to at least 1967.
Bolstered by the Rescue Plan, economic security programs overall (including long-standing programs) lifted 12.2 million children above the poverty line in 2021, also an all-time high. (The previous records were 10.7 million children in 2020, and before that, 9.0 million children in 2009.) The child poverty rate in 2021 was 16.7 percentage points lower when counting government assistance and taxes than not counting them, also a record reduction. (The previous records were 14.4 percentage points in 2020 and 12.0 percentage points in 2009.)
The Rescue Plan’s one-year Child Tax Credit expansion played a central role in these reductions. It raised the maximum credit amount from $2,000 to $3,600 for children under 6 and $3,000 for children aged 6-17 (the first time 17-year-olds were included). Most importantly, for the first time it made the Child Tax Credit “fully refundable,” meaning the full credit was available to children in families with low or no earnings. When the expansion expired at the end of 2021, the credit reverted to the flawed design left in place by the 2017 Trump tax law. As a result, an estimated 19 million children — or more than 1 in 4 children under age 17 — will get less than the full Child Tax Credit or no credit at all this year because their families earn too little, while families with much higher incomes (up to $400,000 for married couples) will receive the full $2,000 credit for each child.
Without the American Rescue Plan’s Child Tax Credit expansion (but with the stimulus payments and other relief measures in place), the child poverty rate would have dropped from 9.6 percent in 2020 to 8.1 percent in 2021. With the Child Tax Credit expansion also in place, the rate fell to 5.2 percent. (See Figure 6.)
Without the Child Tax Credit expansion, some 2.1 million more children would have been in poverty in 2021, based on Census estimates of families’ taxes.
The Rescue Plan for the first time delivered up to half of the Child Tax Credit as advance monthly payments from July to December 2021 rather than all at tax time. Note that Census’ poverty calculations count the entire Child Tax Credit as 2021 income, following Census’ long-standing practice of counting tax-based benefits in the relevant tax year. In reality, most families received half of the credit in early 2022 when they filed their 2021 taxes.
Other relief measures, such as expansions in SNAP benefits, assistance for those at risk of eviction, and expansions of the EITC, helped further reduce poverty and hardship.
Relief measures that provided uninterrupted health insurance coverage for Medicaid enrollees and lowered or eliminated premiums for Affordable Care Act (ACA) marketplace enrollees boosted health coverage. These health coverage measures helped to keep the uninsured rate from rising in 2020 and 2021, which is highly unusual during a major economic downturn. In fact, according to the Census Bureau’s American Community Survey, the uninsured rate fell to 8.6 percent in 2021, to match 2016’s record low. Administrative data show that Medicaid enrollment rose by over 15 million people between February 2020 and December 2021, while ACA marketplace enrollment also reached a record high. The federal response to the pandemic also included increased support for child care, paid leave, and fiscal aid to states, cities, counties, and tribal governments.
Annual poverty figures for 2020 put the best face on a turbulent year. The annual figures average together periods when families received substantial relief — including stimulus payments and boosts to jobless benefits, food assistance, and more — with periods when their incomes were down, obscuring the weeks and months of hardship when families were waiting for aid to arrive or wondering if Congress would renew expiring benefits or pass additional stimulus payments.
Moreover, more than $163 billion in the second round of stimulus payments wasn’t enacted until December 27, 2020, yet most of these dollars are counted in Census’ 2020 annual income data because, as noted above, Census counts tax credits (including the stimulus payments, which were delivered through the tax system) based on the tax year for which they are received, not the year in which they were paid.
Even if Census had counted this income in 2021 rather than 2020, however, the SPM poverty rate would still have declined in 2020 by the largest amount since 1968 and reached its lowest level since 1967, we estimate. Despite delays and other issues that arose as new relief measures were implemented, poverty in both 2020 and 2021 would have been vastly worse without the unprecedented relief.
COVID Relief Helped Prevent a Sharp Rise in Poverty for All Major Racial and Ethnic Groups and Narrowed Differences in Poverty Between Groups
Black and Latino people entered the pandemic with lower income and fewer assets due to discrimination and other structural barriers to opportunity. This meant that many elements of the pandemic response that targeted those with the greatest need had particularly large, positive impacts on Black and Latino people.
Not counting government assistance, poverty would have risen by 2.6 percentage points for Black, 2.2 percentage points for Latino, 3.8 percentage points for Asian, and 2.3 percentage points for white individuals between 2019 and 2020. Once that assistance is counted, poverty fell by 4.8 percentage points for Black, 5.7 percentage points for Latino, 3.1 percentage points for Asian, and 2.0 percentage points for white individuals.
Economic security programs are particularly important in reducing racial inequities in child poverty. Expanded government assistance prevented a sharp rise in child poverty across racial and ethnic groups in 2020 and drove down child poverty rates in 2021. (See Figure 7.) Government assistance and taxes helped keep 30 percent of all Black children and as many as 24 percent of Latino children above the poverty line in 2021, as well as 10 percent of Asian and white children, though poverty after accounting for government programs remains three times higher among Black and Latino children than white children.
Economic security programs reduced differences in child poverty rates by race and ethnicity by over half in 2020 and by over two-thirds in 2021. Before counting government assistance and taxes, the poverty rate for Black children in 2021 was 26 percentage points higher than for white children. For Latino children, it was 20 percentage points higher. Once government assistance and taxes are accounted for, the poverty rate for Black children was 5 percentage points higher than for white children. For Latino children, it was 6 percentage points higher.
These differences were the narrowest they’ve ever been, with data back to 1967. Prior to the pandemic, the poverty rates for both Black and Latino children were always at least 13 percentage points above the corresponding rates for white children and, since 1967, averaged about 28 percentage points higher.
The Rescue Plan’s one-year Child Tax Credit expansion in 2021 played a key role in this historic narrowing of racial and ethnic differences in child poverty. Making the full Child Tax Credit available to children in families with low incomes lowered child poverty among all races and ethnicities, and especially helped Black, Latino, and American Indian or Alaska Native (AIAN) children, whose families are overrepresented in low-paid work due to past and present hiring discrimination, inequities in housing opportunities, access to good schools, and other sources of inequality.
By lifting larger shares of Black, Latino, and AIAN children above the poverty line, the expansion narrowed differences in child poverty rates by race and ethnicity. The Child Tax Credit expansion helped keep 5.1 percent of all Black children, 4.1 percent of AIAN children, and as many as 4.0 percent of Latino children above the poverty line in 2021, as well as 1.8 percent of white and 1.5 percent of Asian children. (See Figure 8.)
After 2021, Congress allowed the Child Tax Credit to return to an earlier, smaller, and less equitable design that denies full participation in the credit to families with low or no earnings. Because children in families with low incomes currently cannot receive the full Child Tax Credit, roughly 46 percent of Black children, 39 percent of AIAN children, 37 percent of Latino children, 17 percent of white children, and 15 percent of Asian children get less than the full credit or no credit at all.
Making the credit more available to these children would lower their poverty, push back against long-standing inequities, and ensure all families can offer their children a stronger foundation of income security and opportunity in difficult times.
COVID Relief Built on Previous Success of Economic Security Programs in Reducing Poverty and Racial and Ethnic Inequities
Over the last half-century, economic security programs such as Social Security, food assistance, tax credits, and housing assistance have reduced poverty for millions of people and reduced inequities in poverty by race and ethnicity. At the same time, barriers to opportunity, including discrimination and other obstacles to employment, education, and health care, remain large and keep poverty rates much higher for some racial and ethnic groups than others.
Between 1970 and 2019 the poverty rate fell for all groups, but it fell even more for Black and Latino people: by 30 and 26 percentage points, respectively, compared with 10 percentage points for white non-Latino people, we calculate. Even with this reduction, poverty rates for Black and Latino people remained far above the white poverty rate. Our analysis by race and ethnicity begins in 1970, the first year that data on Latino ethnicity are available.
Economic security programs have become more effective at reducing poverty and narrowing the differences in poverty rates across racial and ethnic groups over the last five decades. In 1970, families’ government benefits and the taxes they paid lowered the white poverty rate and the Black poverty rate by 2 percentage points and increased the Latino poverty rate by 2 percentage points. In contrast, in 2019, before the pandemic, accounting for government benefits and taxes lowered the white poverty rate by 11 percentage points, Black poverty by 15 percentage points, and Latino poverty by 10 percentage points.
In 2021, economic security programs, enhanced by COVID relief legislation, lowered the white poverty rate by 14 percentage points, Black poverty by 23 percentage points, and Latino poverty by 18 percentage points. (See Figure 9.)
For children, government programs played an even larger role in lowering poverty and narrowing racial and ethnic inequities. Between 1970 and 2019, the Black child poverty rate fell by 36 percentage points, the Latino child poverty rate by 34 percentage points, and the white child poverty rate by 12 percentage points, though the white child poverty rate remained far lower than the poverty rate for Black and Latino children. More than half of the declines in Black, Latino, and white child poverty were driven by the increasing effectiveness of government assistance at lifting families’ incomes above the poverty line. (See Figure 10.)
COVID relief measures accelerated this progress in 2020 and even more so in 2021. For example, for Black children, the poverty rate fell to 8.2 percent in 2021 from 16.7 percent in 2020 and 21.3 percent in 2019. This is stunning progress — in 2019 more than 1 in 5 Black children lived in families with incomes below the poverty line. In 2021, fewer than 1 in 10 did.
Despite this progress, past and present discrimination in both the private sector and public policies kept overall poverty rates in 2021 nearly twice as high among Black and Latino people (both 11.2 percent) than among white people (5.7 percent). Child poverty reflected the same dynamic, with Black and Latino child poverty rates at 8.2 and 8.3 percent, respectively, compared to 2.7 percent among white children.
Stronger Longer-Term Policies Would Avoid Coverage Gaps in Future Recessions and Enable Further Progress
Large-scale, emergency action in the pandemic was remarkably successful at protecting households from serious hardship, but difficult delays and gaps in aid were common, in part because our underlying economic security programs do too little even in normal economic times to help support families, children, and workers who struggle to afford the basics. Many jobless individuals waited weeks for unemployment insurance due to slow and outdated unemployment insurance benefit systems. Others weren’t covered by stimulus payments or other assistance because of their own or a family member’s immigration status, or for other reasons. For example, early stimulus payments omitted dependents aged 17 and older. Some short-term indicators of economic strain or uncertainty such as food pantry utilization increased.
Many of these problems could have been averted if stronger economic security policies had been in place before the pandemic. For example, making permanent reforms to the unemployment insurance system would better support workers in the next recession.
The expiration of these emergency actions since 2021 almost certainly raised poverty in 2022 when measured using the Supplemental Poverty Measure. The Census Bureau will release 2022 poverty data in September 2023. The end of stimulus payments, the EITC and Child Tax Credit expansions, and expanded unemployment benefits almost certainly outweighed the still important job growth and wage gains for lower-paid workers that helped boost incomes in 2022. To recapture and build on the progress against poverty made during the pandemic, policymakers will need to strengthen economic supports for families who face difficulties affording the basics.
Despite tremendous progress in reducing poverty and advancing equity over the last 50 years, large gaps in U.S. anti-poverty policies have continued to leave U.S. children more exposed to poverty than other nations’ children. Before the pandemic, 20 percent of U.S. children lived in families with incomes below half of the national median, the poverty measure most commonly used for international comparisons. This is a much higher share than in any of the world’s 18 other similarly wealthy nations, where between 3 and 15 percent of children are poor. Poverty in the United States also stands out when using an absolute poverty measure to compare nations.
The sheer number of U.S. families who sometimes struggle to meet basic needs is not well understood because data on such hardships seldom span more than a year of a family’s life. Many more families face hardship — that is, an inability to afford basic needs — and poverty at some point over multiple years than in a single year.
CBPP analysis of data from the Census Bureau’s Survey of Income and Program Participation (SIPP), newly redesigned to allow tracking of people’s hardships over multiple consecutive years, reveals how widespread economic insecurity was even before the pandemic. More than 1 in 4 households, including more than 1 in 3 households with children, experienced a major form of hardship — specifically, an inability to afford adequate food, shelter, or utilities — in one or more of the years 2014, 2015, and 2016. Among Black and Latino households with children, roughly 1 in 2 reported one of these hardships, as did more than 1 in 4 white households with children.
Even many households who are currently in the middle of the income scale may encounter hardship over time. Among the middle third of households with children (ranked by their current annual income), nearly 1 in 3 reported one of these hardships over that three-year span, either because their annual incomes were lower in one of the two other years or because they experienced hardship despite having middle incomes.
Some temporary policies put in place during the pandemic proved effective at combatting problems that long predated the pandemic and point the way to policy advances the nation should adopt permanently. The most notable examples are policies to better support children in families with low incomes: an expanded Child Tax Credit that provides the full credit to children in the lowest-income families, increased support for child care and housing, and summer food benefits to prevent an increase in food insecurity when school is out. Congress took a modest step in December 2022 when it created a permanent summer food assistance program in the year-end funding bill; but action in these other areas is needed to make a sizable dent in child poverty.
Building on the strong research base for effective policies in reducing poverty and hardship, supporting healthy child development, and broadening opportunity, policymakers should make the investments needed to address economic and health insecurity and glaring disparities in hardship and opportunity across lines of race and ethnicity. These investments would help families meet everyday challenges, with all the long-term payoffs that economic security for children provides. They would also ensure that the nation’s economic security infrastructure is prepared to respond more quickly and reliably to meet the needs of families and the economy in the next recession or economic crisis.
Appendix I: Methodological Notes
This appendix provides some more detailed explanation of our methods and discusses some data quality issues of the Current Population Survey (CPS), which is the main source for the poverty estimates that Census produces and those we present in this report.
Defining Economic Security Programs
In this report we use the terms “economic security programs,” “government assistance,” and “government benefits” interchangeably. In our calculations these benefits include: Social Security, unemployment insurance, workers’ compensation, veterans’ benefits, Temporary Assistance for Needy Families (TANF), state General Assistance, Supplemental Security Income (SSI), SNAP, the National School Lunch Program, the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), rental assistance (such as Section 8 and public housing), home energy assistance, the EITC, and the Child Tax Credit. Benefit figures for 2008-2010 also reflect a number of temporary federal benefits enacted in response to the Great Recession: the 2008 stimulus payment, 2009 economic recovery payment, and 2009-2010 Making Work Pay Tax Credit. Benefit figures for 2020-2021 include economic impact payments and figures for 2021 include broadband assistance. Taxes, subtracted from family resources, are family members’ federal and state income and payroll taxes.
Poverty Hit Record Low in 2021 Even After Accounting for Recent Data Issues
Precise year-to-year poverty comparisons in the CPS are hampered in recent years by three challenges: the pandemic’s effect on data collection, recent methodological changes in the CPS, and the timing of certain government assistance payments. Even after accounting for all three of these issues, however, the data show that poverty reached a record low in 2021.
The pandemic’s effect on CPS data collection made figures less reliable, particularly in the case of figures for 2019, collected in 2020. A Census assessment noted that fewer people completed the survey than usual in the pandemic, especially those with low incomes; it concluded that Census’ own median household income figures were likely about 2.9 percent too high in 2019 and 2.0 percent too high in 2020 as a result. Supplemental poverty rates were likely 0.3 percentage points too low in 2019 and 0.2 percentage points too low in 2020. Adjusted data are not available for 2021, but even if we were to adjust the 2021 poverty rate upwards by 0.3 percentage points, 2021 would still be a record low poverty year. The SPM poverty rate in 2021 was 7.8 percent compared to 9.1 percent in 2020, the previous record low.
A second challenge in the CPS was a set of recent changes the Census Bureau made to its survey processing system, which make data for 2018 and later not strictly comparable to earlier years. But Census provides enough data on this survey transition to make clear that poverty in 2021 reached a record low even compared to years before 2018. Using data comparable over the period 1967-2017, the poverty rate in 2017 stood at a then record low of 13.9 percent. Using data comparable to 2018 and later, Census recomputed the 2017 poverty rate to be 13.0 percent. The 2021 poverty rate of 7.8 percent is much lower than that recomputed 13.0 poverty rate, and therefore lower than the years before.
A third data issue is how to count certain government payments made at the cusp between 2020 and 2021. Most U.S. residents received $600-per-person stimulus payments in a second round of COVID Economic Impact Payments (EIP2s). Counting the approximately $141 billion in EIP2 payments when they were made — overwhelmingly in the first two months of 2021 — would lower the SPM poverty rate in 2021 from 7.8 percent to 6.9 percent. Census counted these payments as 2020 income, following long-standing Census practice to count tax-based benefits in the relevant tax year. In addition, we calculate that if Census had counted EIP2s in 2021 rather than 2020, SPM poverty would have been 3.8 million people (about 1.2 percentage points) higher in 2020. From 2020 to 2021, the number of individuals in poverty would have declined by some 6.6 million more than under Census’ approach to counting the payments.
Appendix Table 1 shows poverty rates using different methodological adjustments discussed in this section. Taking these into account is particularly important when making year-to-year comparisons. For example, to accurately compare 2017 and 2018 one should use data based on the updated processing system for both years. Using those data, poverty fell by 0.3 percentage points between 2017 and 2018. If instead one used the old processing system data for 2017 (13.9 percent) and the new processing data for 2018 (12.7 percent), one would erroneously conclude that poverty fell by over 1 percentage point. Similarly, when comparing 2018 to 2019, it’s best to use the Census weights that adjust for non-response bias. Using those adjusted weights, one finds that poverty fell by 0.6 percentage points between 2018 and 2019. Without using those adjusted weights, one would erroneously conclude that poverty fell by a whole percentage point in 2019.
|APPENDIX TABLE 1|
|Poverty Rates Using Different Methodological Adjustments|
|Old processing system, original weights||New processing system, original weights||New processing system, non-response-adjusted weights||New processing system, original weights, count EIP2s in year payments made|
Note: Figures use the federal government’s Supplemental Poverty Measure (SPM) with the original thresholds produced by the U.S. Bureau of Labor Statistics, which are not anchored to a specific year.
Source: CBPP analysis of U.S. Census Bureau's Current Population Survey
Main Findings Are Similar When Using Relative Rather Than Anchored SPM Thresholds
Unless otherwise noted, figures in this report use an SPM threshold “anchored” to 2021. That is, poverty is defined as having family resources below the Supplemental Poverty Measure’s poverty threshold established by the Bureau of Labor Statistics for 2021, adjusted in earlier years for inflation using the Labor Department’s consumer price index retroactive series. (The threshold is also adjusted for family size and composition, home ownership status, and local housing costs. In 2021, this threshold was $31,453 for a two-adult two-child family renting in an average-cost community.) Figures may differ from figures in previous CBPP reports that were anchored to an earlier year or published Census SPM figures, which are not anchored.
Another version of the SPM — the unanchored or “relative” SPM — takes into account evolving societal standards of poverty over long periods of time by updating the poverty thresholds each year based on gradual changes in what most families spend on basic necessities.
Poverty declines less over the long term when a relative SPM poverty measure is used; that’s because under a relative measure, the poverty threshold in earlier years is lower in inflation-adjusted terms than the threshold in 2021. As the amount that families spend on necessities rises in real terms, the relative measure’s poverty threshold rises in inflation-adjusted terms.
But under both relative and anchored versions of the poverty measure, poverty rates reached record lows in 2020 and 2021, overall and for children. Similarly, under both measures, the decline in the number and percentage of people poverty in 2020 were the largest on record, and the decline in the number and percentage of children in poverty in 2021 were the largest on record, in data back to 1967.
This analysis uses an anchored series to ensure that trends are purely due to changes in families’ resources, not changes in the poverty thresholds, in accordance with this paper’s focus on the evolving role of government assistance. (Census’ official poverty measure also uses thresholds that are adjusted each year only for inflation.)
The supplemental tables in Appendix II of this report include trends using both the anchored and relative SPM thresholds.
Correcting for Underreporting Shows Even Larger Impact of Economic Security Programs
Census’ counts of program participants typically fall well short of the totals shown in actual administrative records. Such underreporting is common in household surveys and can affect estimates of poverty. The extent of underreporting can vary by program and by year. In 2020, less than one-half of unemployment benefits and SNAP benefits were captured by the CPS. In 2021, less than one-third were. This implies that the anti-poverty effectiveness of those programs could be at least twice as high as what the uncorrected data suggests in those years. Programs like SSI that provide regular monthly payments that people receive over longer periods tend to be underreported less. Roughly 90 percent of SSI benefits were captured by the CPS in 2020 and 2021. (See Appendix Table 2.)
|APPENDIX TABLE 2|
|Comparison of Benefit Totals in Administrative and Survey Data |
In billions of dollars (not adjusted for inflation)
|Administrative data (NIPA)||28||28||537||339|
|Survey data (CPS)||15||18||218||90|
|Percent captured by CPS survey||56%||64%||41%||26%|
|Administrative data (NIPA)||58||55||95||150|
|Survey data (CPS)||29||25||39||46|
|Percent captured by CPS survey||50%||47%||41%||31%|
|Administrative data (NIPA)||57||58||58||57|
|Survey data (CPS)||48||48||50||52|
|Percent captured by CPS survey||85%||84%||87%||92%|
Sources: CBPP analysis of U.S. Census Bureau Current Population Survey (CPS) public use files for 2019-2022 and U.S. Bureau of Economic Analysis National Income and Product Account (NIPA) tables 2.1 and 3.12.
The figures in this report do not correct for underreporting of benefits. In previous analyses, we have corrected for the underreporting of SNAP, SSI, and TANF benefits using data from the Transfer Income Model (TRIM) produced by the Urban Institute for the U.S. Department of Human Services Assistant Secretary for Planning and Evaluation. We do not use TRIM data in this report since they are not yet available for 2020 or 2021. TRIM data start with the same survey data that the Census Bureau uses but TRIM uses an alternative method of filling in respondents’ skipped survey questions that more closely approximates actual total program participation and payment levels for key benefit programs.
Making these corrections increases the number of people lifted above the SPM poverty line by all government assistance in 2018 from 37 million to 41 million and lowers the poverty rate from 12.7 percent to 11.5 percent. See Appendix Table 3 for program-by-program figures.
|APPENDIX TABLE 3|
|Number of people lifted above SPM poverty line in 2018 by selected programs |
|Correcting for underreported benefits||6.0||4.0||0.7|
|Administrative data (NIPA)||1.4||0.5||0.2|
|Survey data (CPS)||2.8||1.0||0.4|
Note: Figures use the federal government’s Supplemental Poverty Measure (SPM).
Sources: CBPP analysis of U.S. Census Bureau's March 2019 Current Population Survey. Corrections for underreported benefits from Department of Health and Human Services/Urban Institute Transfer Income Model (TRIM).
When analyzing poverty rates for all ages over time, we find that these corrections for underreporting make only a modest difference in most years. However, they matter more for trends in child poverty and in years when government assistance has expanded, such as during the Great Recession.
Correcting for underreporting is also particularly important when measuring “deep” child poverty, often defined as income below half of the poverty line, because underreported benefits make up a larger share of the total income of those with the lowest reported incomes.
For example, correcting for underreported TANF, SSI, and SNAP benefits reveals that the share of children in single-mother families living in poverty fell but the share living in deep poverty rose during the first decade after policymakers made major changes in the public assistance system in the mid-1990s that made SNAP and TANF less available and accessible to poor families, as we reported in 2020. That paper focuses on children in single-mother families (the population most affected by the cash assistance changes under the 1996 law that created TANF) between the years 1995 and 2005. This period, the first decade under the new law, is suitable for assessing the law because the years represent comparable points in the business cycle; 2005 also predates benefit expansions during the Obama Administration including the ACA, which lowered deep poverty by lowering families’ out-of-pocket medical expenses. The paper finds that deep poverty for children in single-mother families rose from 5.4 percent in 1995 to 7.4 percent in 2005.
Another way to correct for the underreporting of government benefits is to augment survey data by linking it with actual payment data from survey respondents’ confidential program records. A recent Census Bureau working paper uses multiple linked administrative sources to correct underreporting of earnings and a variety of other cash income sources. This lowers the official child poverty rate by 0.57 percentage points, not a statistically significant change. So far, no analysis using linked data has examined deep poverty in single-mother families with a Supplemental Poverty Measure that corrects for the underreporting of TANF, SSI, and SNAP in the key years of 1995 and 2005.
Analysis Draws on Census Estimates of Federal and State Taxes and Credits
The CPS does not ask families directly about taxes or tax credits; instead, Census estimates them for families, which avoids adding needlessly to an already long survey. Census bases its estimates on respondents’ survey income. One limitation of the approach is that in practice the credit is not received until the family files its taxes the following year. Families that anticipate receiving a tax credit, however, may reduce their tax withholding accordingly and see higher take-home pay in the current year.
Another limitation of Census’ tax estimates is that they do not account for restrictions on EITC and Child Tax Credit eligibility for filers or dependents who lack a Social Security number, which bars certain families that include immigrants from receiving these credits. Accounting for this restrictive policy would reduce the anti-poverty effects of these tax credits and would increase the estimated poverty rates for certain groups. In particular, it could increase estimated poverty rates perhaps by nearly 3 percentage points for Latino children and nearly 1 percentage point for Asian children.
A final limitation of Census’ tax estimates is that they assume 100 percent participation in tax credits among eligible families, and no participation among ineligible families. It is challenging to assess whether Census’ tax estimates overstate or understate the tax system’s effect on poverty as a result. One study matched Census’ tax estimates with each survey member’s confidential IRS records and found many people receiving no EITC despite appearing eligible in the survey data while other filers received payments despite appearing possibly ineligible (though they usually had low incomes).
On balance, actual EITC payments lower poverty rates by 1.2 percentage points in the linked data for 2014, rather than the 1.5 percentage points Census’ approach indicates. On the other hand, in dollar terms, the EITC and the Child Tax Credit are not overcounted in the CPS. The CPS estimates show a total of $39 billion in EITC payments and $113 billion in Child Tax Credit payments in 2020, less than the actual $62 billion in EITC payments and $119 billion in Child Tax Credit payments shown in IRS records.
Accounting for immigration-related restrictions in the EITC and Child Tax Credit and less than 100 percent participation in the credits, however, would be unlikely to change the conclusions that poverty rates reached record lows in 2021 both overall and for children, and the one-year declines in poverty in 2020 and children’s poverty in 2021 were the largest on record.
As impressive as the poverty reductions of 2020 and 2021 are, it is also important to keep in mind that the poverty figures here reflect annual income as calculated by the Census Bureau. Households in need in the pandemic did not always receive benefits as quickly as needed. One lesson of the pandemic is that many households’ needs could have been met faster if unemployment insurance eligibility rules and administrative systems had been modernized before the pandemic and if the expansion of the Child Tax Credit in 2021, including full refundability and monthly payments, had already been in place.
Appendix II: Supplemental Tables
Supplemental tables can be accessed here:
|Tables using 2021 SPM poverty line adjusted for inflation|
|Tables using "relative" SPM thresholds (that is, adjusted annually for changes in basic needs spending)|
 Chris Wimer et al., Historical Supplemental Poverty Measure Data, Columbia Population Research Center, 2017, https://www.povertycenter.columbia.edu/historical-spm-data-reg. The CPS data we accessed partly via IPUMS-CPS, University of Minnesota, www.ipums.org.
 Using a recent year’s SPM threshold and adjusting it back for inflation (using the Labor Department’s consumer price index retroactive series) creates an SPM series “anchored” to recent years’ poverty standards. Some analysts prefer the anchored SPM to the standard or “relative” SPM, which allows poverty thresholds to grow slightly faster than inflation as living standards rise across decades. This analysis uses an anchored series to ensure that trends are purely due to changes in families’ resources, not changes in the poverty thresholds. Chris Wimer et al., “Trends in Poverty with an Anchored Supplemental Poverty Measure,” Columbia Population Research Center Working Paper No. 13-01, December 2013, https://academiccommons.columbia.edu/doi/10.7916/D8RN3853. We anchor all poverty thresholds to 2021, the latest SPM threshold available. Figures in this report may differ from figures in previous CBPP reports that were anchored to an earlier year or from published Census figures, which are not anchored. Although poverty declines less over the long term when a relative SPM poverty measure is used, under both relative and anchored versions of the poverty measures, poverty rates reached record lows in 2020 and 2021, overall and for children.
 Figures account for all public benefits (including permanent programs such as Social Security, food assistance, rental vouchers, regular state unemployment insurance, Temporary Assistance for Needy Families (TANF), the EITC, and the Child Tax Credit, as well as pandemic programs such as stimulus payments and supplemental unemployment benefits and food assistance. The figures subtract Census estimates of families’ federal and state income taxes and payroll taxes. Consistent with similar comparisons by the Census Bureau and other analysts, this comparison looks at the effect of assistance programs on family resources and presumes no change in behavior.
 In 2018 Census released data based on an updated processing system. To facilitate comparisons across time, Census released two versions of 2016 and 2017 data. One version is comparable to 2018, and the other version is comparable to earlier years. When making comparisons in this report we use the appropriate version depending on what years we’re comparing. For details on the changes in the Census survey, see Liana Fox, “The Supplemental Poverty Measure: 2018,” U.S. Census Bureau, October 2019, https://www.census.gov/content/dam/Census/library/publications/2019/demo/p60-268.pdf.
 The differences in poverty rates and anti-poverty effectiveness of economic security programs between 2020/2021 and earlier years are larger than can be attributable to changes in survey methods. The 2017 poverty rate using thresholds anchored to 2021 was 14.2 percent when using the new methods and 15.2 percent when using the previous methods. Government assistance and taxes reduced poverty by 44 percent when using new methods and 41 percent when using the previous methods in 2017.
 Samantha Waxman, Arloc Sherman, and Kris Cox, “Income Support Associated With Improved Health Outcomes for Children, Many Studies Show,” CBPP, May 27, 2021, https://www.cbpp.org/research/federal-tax/income-support-associated-with-improved-health-outcomes-for-children-many.
 SNAP was previously known as food stamps. In the early 1960s, the Food Stamp Program was a pilot program operating only in some counties. It became a nationwide program in 1974 after successive expansions to additional counties during the 1960s and 1970s.
 George Fenton, “New Study Finds Income Support in Childhood Increases Future Earnings,” CBPP, June 21, 2022, https://www.cbpp.org/blog/new-study-finds-income-support-in-childhood-increases-future-earnings; Martha J. Bailey et al., “Is the Social Safety Net a Long-Term Investment? Large-Scale Evidence From the Food Stamps Program,” National Bureau of Economic Research Working Paper, April 2020, http://www.nber.org/papers/w26942.pdf.
 Arloc Sherman and Tazra Mitchell, “Economic Security Programs Help Low-Income Children Succeed Over Long Term, Many Studies Find,” CBPP, July 17, 2017, https://www.cbpp.org/research/poverty-and-inequality/economic-security-programs-help-low-income-children-succeed-over.
 Nick Gwyn, “Historic Unemployment Programs Provided Vital Support to Workers and the Economy During Pandemic, Offer Roadmap for Future Reform,” CBPP, March 24, 2022, https://www.cbpp.org/research/economy/historic-unemployment-programs-provided-vital-support-to-workers-and-the-economy.
 Jeff Larrimore, “Poverty in 2021 Corrected for Unemployment Insurance Underreporting,” SSRN, September 13, 2022, https://ssrn.com/abstract=4217956. Figures published by Census show that unemployment insurance reduced SPM poverty by 1.7 percentage points in 2020 and 0.7 percentage points in 2021. Larrimore finds that correcting for underreporting of unemployment insurance benefits would reduce SPM poverty in 2020 and 2021 by an additional 1.8 and 1.2 percentage points respectively.
 Kris Cox, Samantha Jacoby, and Chuck Marr, “Stimulus Payments, Child Tax Credit Expansion Were Critical Parts of Successful COVID-19 Policy Response,” CBPP, June 22, 2022, https://www.cbpp.org/research/federal-tax/stimulus-payments-child-tax-credit-expansion-were-critical-parts-of-successful.
 The CARES Act’s stimulus payments and unemployment insurance (UI) benefits combined to lift about 11 million people, including about nearly 3 million children, above the poverty line in 2020, as compared with the number who would have been considered poor in 2020 not counting that assistance, our analysis finds. The CARES Act authorized the first of two stimulus payments that Census counted toward 2020 income, which by itself lifted than 6 million people above the poverty line. The estimate assumes the CARES Act was responsible for approximately 73 percent (or more than 4 million) of the people lifted above the poverty line by UI benefits in 2020, which equals the percentage of total 2020 federal and state UI payments attributed to the Act in data from the Bureau of Economic Analysis (BEA). The analysis accounts for the overlap in poverty reduction among programs and uses a 2021 poverty threshold adjusted backwards for inflation.
The previous law that comes closest to the CARES Act in poverty reduction is likely to be the 2009 American Recovery and Reinvestment Act (ARRA), enacted in the Great Recession. We previously estimated that ARRA lifted nearly 7 million people above the poverty line in 2010, using an SPM-like measure. (The estimate predated the SPM and, using the SPM, would likely have been just under 8 million, we calculate.) See Arloc Sherman, “Poverty and Financial Distress Would Have Been Substantially Worse in 2010 Without Government Action, New Census Data Show,” CBPP, November 7, 2011, https://www.cbpp.org/research/poverty-and-financial-distress-would-have-been-substantially-worse-in-2010-without.
In BEA data back to 1929, no earlier year comes close to 2020’s year-over-year increase in inflation-adjusted per-person spending on government income assistance from programs — such as Social Security, nutrition assistance, or refundable tax credits — that are counted as family resources in the SPM. Social Security, created by the 1935 Social Security Act and expanded gradually over several decades, keeps more people above the poverty line than any other program (26 million people, including 1 million children, in 2021).
 The estimate assumes the American Rescue Plan is responsible for 47 percent of the number of people lifted above the poverty line by UI benefits in 2021; quarterly Bureau of Economic Analysis data indicate that the Rescue Plan was responsible for at least 47 percent of total UI payments in 2021.
 The Rescue Plan’s Child Tax Credit expansion and third round of stimulus payments together lifted 5.4 million children above the SPM poverty line in 2021, our analysis of Census Bureau data shows. Accounting for the Rescue Plan’s UI expansions would increase that figure further. It is difficult to find another piece of legislation that could come close to the Rescue Plan in its effect on annual child poverty rates. We previously estimated that ARRA lifted roughly 2.5 million children above the poverty line. (See Sherman, 2011). SPM data are unavailable before 1967 but no other year since at least 1950 had an increase in real per capita spending for overall government income assistance that would suggest a single-year decline in child poverty comparable to those observed in 2009, 2020, or 2021.
Kris Cox et al., “Top Tax Priority: Expanding the Child Tax Credit in Upcoming Economic Legislation,” CBPP, updated June 12, 2023, https://www.cbpp.org/research/federal-tax/top-tax-priority-expanding-the-child-tax-credit-in-upcoming-economic.
 Census’ published figures show a 9.7 percent child poverty rate in 2020. Using the 2021 thresholds adjusted back for inflation (rather than using the SPM's expenditure-based year-to-year adjustment) yields the 9.6 percent estimate for 2020 shown in Figure 6.
 Kalee Burns, Liana Fox, and Danielle Wilson, “Expansions to Child Tax Credit Contributed to 46% Decline in Child Poverty Since 2020,” U.S. Census Bureau, September 13, 2022, https://www.census.gov/library/stories/2022/09/record-drop-in-child-poverty.html. The poverty reduction estimate is sensitive to the order in which programs are considered. Not counting stimulus payments, the Child Tax Credit expansion would have lifted more than 3 million children above the poverty line in 2021. Counting the stimulus payments (and all other income), the credit expansion lifts 2.1 million children above the poverty line while additional children, already raised across the poverty threshold by the stimulus payments, are lifted slightly further above poverty. Estimates in this report calculate a program’s poverty impact after all other income is counted.
 Gideon Lukens, “Census: Uninsured Rate Matches a Record Low in 2021, Driven by Relief Provisions,” CBPP, September 15, 2022, https://www.cbpp.org/blog/census-uninsured-rate-matches-a-record-low-in-2021-driven-by-relief-provisions.
 Sharon Parrott, “Robust COVID Relief Achieved Historic Gains Against Poverty and Hardship, Bolstered Economy,” CBPP, June 14, 2022, https://www.cbpp.org/research/poverty-and-inequality/robust-covid-relief-achieved-historic-gains-against-poverty-and-0.
 Congressional Budget Office, “Estimate for Division N—Additional Coronavirus Response and Relief H.R. 133, Consolidated Appropriations Act, 2021 Public Law 116-260, Enacted on December 27, 2020,” January 14, 2021, https://www.cbo.gov/system/files/2021-01/PL_116-260_div_N.pdf
 This report uses the term “Latino” to refer to people of any race who identify as Hispanic or Latino in Census surveys. This language does not necessarily reflect how everyone who is part of this community would describe themselves. For example, gender-inclusive terms like Latinx and Latine can also be used to refer to this population. These terms have emerged in recent years to represent the diversity of gender identities and expressions that are present in the community. Unless otherwise specified, the racial and ethnic categories used in this report and supplemental tables in Appendix II are mutually exclusive and exhaustive. Individuals can be classified as white only, not Latino; Black only, not Latino; Latino (any race); Asian only, not Latino; or another race or multi-racial, not Latino.
The analysis uses Census Bureau estimates of income and payroll taxes. One notable limitation of Census’ tax estimates is that they do not account for restrictions on EITC and Child Tax Credit eligibility for filers or dependents who lack a Social Security number, which bars certain families that include immigrants from receiving these credits. Accounting for this restrictive policy would slightly increase estimated poverty rates for Latino people and consequently somewhat decrease the estimated reductions in poverty shown here.
 American Indian and Alaska Native (AIAN) children may be AIAN alone or in combination with other races and ethnicities. Poverty rates in Figure 8 and elsewhere in this report are calculated using data made public by the Census Bureau, which — to help protect households’ confidentiality — may differ slightly from figures Census publishes based on its internal files. For example, Figure 8 shows an SPM poverty rate of 8.3 percent for Latino children, whereas in Census publications this figure is 8.4 percent.
 Kris Cox et al., “Top Tax Priority: Expanding the Child Tax Credit in Upcoming Economic Legislation,” CBPP, updated June 12, 2023, https://www.cbpp.org/research/federal-tax/top-tax-priority-expanding-the-child-tax-credit-in-upcoming-economic.
 Danilo Trisi and Matt Saenz, “Economic Security Programs Reduce Overall Poverty, Racial and Ethnic Inequities,” CBPP, updated July 1, 2021, https://www.cbpp.org/research/poverty-and-inequality/economic-security-programs-reduce-overall-poverty-racial-and-ethnic.
 We don’t discuss Asian poverty trends in this section focused on long-term trends since poverty data for Asian people only became available starting in 2002. See our Excel appendix for these data.
 Although some broad annual indicators of hardship, including food insecurity as measured by USDA, were statistically unchanged in 2020 and 2021, a variety of other evidence points to an increase during the pandemic in the number of households experiencing at least sporadic trouble making ends meet. This includes year-over-year increases in requests for help with food (up 97 percent), shelter (42 percent), and utilities (31 percent), according to an analysis of calls to the United Way’s “211” help line; a 55 percent rise in the number of people using food banks since before the pandemic, according to the food bank network Feeding America; and a near doubling (from $8 billion to $15.8 billion) in customers’ total estimated arrearages on electricity bills between December 31, 2019 and December 31, 2020, according to state energy assistance program directors.
See Alisha Coleman-Jensen et al., “Household Food Security in the United States in 2021,” ERR-309, U.S. Department of Agriculture, Economic Research Service, September 2022, https://www.ers.usda.gov/webdocs/publications/104656/err-309.pdf; Rachel Garg, Balaji Golla, and Matthew Kreuter, “Year 1 of COVID-19: Needs rise 59%,” Washington University in St. Louis Health Communication Research Laboratory, March 29, 2021, https://hcrl.wustl.edu/year-1-of-covid-19-needs-rise-59/; Paul Morello, “The food bank response to COVID, by the numbers,” Feeding America, March 12, 2021, https://www.feedingamerica.org/hunger-blog/food-bank-response-covid-numbers; Letter from National Energy Assistance Directors Association to Xavier Becerra, Secretary, U.S. Department of Health and Human Services, March 29, 2021, https://neada.org/wp-content/uploads/2021/04/LetterHHSReleaseofFundsARP.pdf.
 Nick Gwyn and Jenna Gerry, “Unemployment Insurance System Unprepared for Another Recession,” CBPP and National Employment Law Project, April 18, 2023, https://www.cbpp.org/research/economy/unemployment-insurance-system-unprepared-for-another-recession.
Arloc Sherman et al., “Widespread Economic Insecurity Pre-Pandemic Shows Need for Strong Recovery Package,” CBPP, July 14, 2021, https://www.cbpp.org/research/poverty-and-inequality/widespread-economic-insecurity-pre-pandemic-shows-need-for-strong.
 Janet C. Gornick and Marcus Jäntti, “Poverty,” Pathways, Stanford Center on Poverty and Inequality, 2016, https://inequality.stanford.edu/sites/default/files/Pathways-SOTU-2016-2.pdf.
 Sherman et al., 2021.
 Sharon Parrott, “The Nation Has Made Progress Against Poverty But Policy Advances Are Needed to Reduce Still-High Hardship,” CBPP, July 28, 2022, https://www.cbpp.org/research/poverty-and-inequality/the-nation-has-made-progress-against-poverty-but-policy-advances.
 CPS income and poverty data for 2019 were gathered in February, March, and April 2020. The Census data quality assessment was informed in part by employer pay records concerning CPS households that did and did not complete the survey in each of the last five years. Jonathan Rothbaum and Charles Hokayem, “How Did the Pandemic Affect Survey Response: Using Administrative Data to Evaluate Nonresponse in the 2021 Current Population Survey Annual Social and Economic Supplement,” U.S. Census Bureau, September 14, 2021, https://www.census.gov/newsroom/blogs/research-matters/2021/09/pandemic-affect-survey-response.html.
 These differences in poverty rates are calculated by CBPP adjusting the data for non-response during the pandemic using alternate weights released by Census in February of 2022 here: https://www.census.gov/data/datasets/2020/demo/cps/2020-pub-use-adj-nonresp-corono-pandemic.html.
 EIP2s were enacted on December 27, 2020; at least 99 percent of EIP2 dollars were paid in 2021, monthly Treasury figures show.
 The $600 stimulus payments were technically advance payments of a 2020 tax credit and were administered by IRS. For background and funding levels for all three stimulus payments, see https://www.irs.gov/statistics/soi-tax-stats-coronavirus-aid-relief-and-economic-security-act-cares-act-statistics.
 Researchers who prefer the relative SPM note that it aligns more closely with the recommendations of the 1995 National Academy of Sciences committee whose work led to the SPM. The committee found that, over decades, the view of the poverty line — that is, the minimum amount of income needed to obtain basic needs — held by both the public and poverty experts tends to rise slightly faster than inflation but more slowly than typical income levels, and this evolution is best captured by changes in the amount typical families spend on food, clothing, shelter, and utilities. The committee recommended updating the poverty threshold annually based on changes in basic needs spending, as the relative SPM poverty measure does. For more information on the development of SPM thresholds, see the SPM website of the U.S. Bureau of Labor Statistics at https://www.bls.gov/pir/spmhome.htm.
 Danilo Trisi and Matt Saenz, “Economic Security Programs Cut Poverty Nearly in Half Over Last 50 Years,” CBPP, updated November 26, 2019, https://www.cbpp.org/research/poverty-and-inequality/economic-security-programs-cut-poverty-nearly-in-half-over-last-50.
 Danilo Trisi and Matt Saenz, “Deep Poverty Among Children Rose in TANF’s First Decade, Then Fell as Other Programs Strengthened,” CBPP, February 27, 2020, https://www.cbpp.org/research/poverty-and-inequality/deep-poverty-among-children-rose-in-tanfs-first-decade-then-fell-as. Another study corrects for underreporting of SSI and a number of other income sources and concludes that deep poverty did not rise between 1995 and 2016 for people in single-parent families. That study, however, does not correct for underreporting of TANF income, corrects underreporting of SNAP in 2016 but not 1995, and relies heavily on an unusual and weakly justified inflation adjustment. Kevin Corinth, Bruce D. Meyer, and Derek Wu, “The Change in Poverty from 1995 to 2016 Among Single Parent Families,” Becker Friedman Institute, University of Chicago, March 2022, https://bfi.uchicago.edu/wp-content/uploads/2022/03/BFI_WP_2022-45.pdf.
 Adam Bee et al., “National Experimental Wellbeing Statistics,” U.S. Census Bureau, February 2023, https://www.census.gov/content/dam/Census/library/working-papers/2023/demo/sehsd-wp2023-02.pdf. This analysis only examines 2018 and doesn’t account for SNAP and the impact of the tax system. For the official (cash) measure of poverty, Bee et al. find administrative records make a much smaller difference on overall poverty rates than does work by Bruce Meyer and his colleagues, possibly in part because of different assumptions about when administrative records are themselves incorrect.
 For more information on the Census tax model, see Daniel Lin, “Methods and Assumptions of the CPS ASEC Tax Model,” SEHSD working paper 2022-18, U.S. Census Bureau, November 30, 2022, https://www.census.gov/content/dam/Census/library/working-papers/2022/demo/sehsd-wp2022-18.pdf.
 Unpublished CBPP analysis of 2017-2019 American Community Survey data.
 See Table 8 in Maggie R. Jones and James P. Ziliak, “The Antipoverty Impact of the EITC: New Estimates from Survey and Administrative Tax Records,” Center for Economic Studies (CES) research paper, U.S. Census Bureau, April 2019, https://www2.census.gov/ces/wp/2019/CES-WP-19-14.pdf.
 See Table 5 in Daniel Lin, “Methods and Assumptions of the CPS ASEC Tax Model,” SEHSD working paper 2022-18, U.S. Census Bureau, November 30, 2022, https://www.census.gov/content/dam/Census/library/working-papers/2022/demo/sehsd-wp2022-18.pdf.