Poverty in the United States in 2021

December 27, 2022 R47354

Poverty in the United States in 2021
December 27, 2022
The federal government publishes poverty statistics using two measures: the official poverty
measure and the Supplemental Poverty Measure (SPM). These two measures tell different stories
Joseph Dalaker
about who is poor in the United States because they measure different things. Both measures
Analyst in Social Policy
compare the resources of a family or unrelated individual against a measure of need for that same

family or individual. If the resources (measured in dollars) are less than the measure of need (also
measured in dollars), the family or unrelated individual is considered to be in poverty; if

resources are greater than or equal to the measure of need, the family or individual is classified as
not being in poverty. The measures differ as to what is counted as resources or included in the measure of need.
 Under the official poverty measure, the measure of need was originally computed using family expenditure
data from 1955 and food costs in 1962. Using the cost of a tightly constrained food budget, and the average
share of family income that was spent on food, dollar amounts were computed to represent the overall
income levels (the poverty thresholds) at which a family whose basic needs overall might have been
similarly constrained. These official poverty thresholds have been updated annually for inflation. For the
resources necessary to meet that level of need, the official poverty measure counts income in the form of
cash only, before taxes—meaning that tax credits and the value of noncash benefits are not counted.
 Under the SPM, the measure of need is based on recent spending data from the Consumer Expenditure
Survey; namely, a percentage of median family spending on food, clothing, shelter, and utilities (plus an
extra 20% for miscellaneous expenses such as personal care products), as opposed to having been
computed once and indexed forward for inflation (as is done for the official measure). For the resources
necessary to meet that level of need, the SPM uses after-tax income (which includes tax credits), estimates
the value of certain noncash benefits (such as food assistance), and subtracts some expenditures (such as
work-related expenses, child care expenses, and medical expenses paid out-of-pocket) that families cannot
use toward the categories of basic needs that are used to define the SPM poverty level. This approach is
intended to better reflect the economic choices families currently face, and to better reflect the effects of
government programs on the low-income population, than does the official measure.
For the official measure in 2021, no statistically significant change from 2020 was detected in the overall poverty rate—the
percentage of people who are in poverty—or in the number of people in poverty in the United States: 11.6% of the
population, or 37.9 million people, lived in poverty. Poverty among most segments of the population held steady, with some
exceptions:
 the poverty rate fell among children (from 16.0% in 2020 to 15.3% in 2021) and rose among the aged (from
8.9% to 10.3%);
 as more people obtained full-time year-round work, the poverty rate fell among workers (from 5.0% to
4.7%) and rose for nonworkers (from 28.8% to 30.0%); and
 the poverty rate rose among people identifying racially as Asian (from 8.1% to 9.3%).
The SPM tells a different story—one that reflects the effects of taxes, refundable tax credits such as stimulus payments and
child tax credits, and in-kind assistance such as food assistance, which were expanded to address economic effects of the
COVID-19 pandemic. In 2021,
 the overall poverty rate under the SPM fell 1.4 percentage points (to 7.8% from 9.2% in 2020),
 nearly 4.5 million fewer people lived in poverty (25.6 million, down from 30.0 million in 2020) according
to the SPM,
 SPM poverty rate declines were widespread among demographic groups, and
 the SPM child poverty rate reached the lowest level ever recorded, having fallen to 5.2% from 9.7% in
2020; this reflects the effects of policies targeted toward children and families with children.

Congressional Research Service

Poverty in the United States in 2021

Introduction
The federal government publishes poverty statistics using two measures: the official poverty
measure and the Supplemental Poverty Measure (SPM). These two measures tell different stories
about who is poor in the United States because they measure different things. Both measures
compare the resources of a family or unrelated individual against a measure of need for that same
family or individual. If the resources (measured in dollars) are less than the measure of need (also
measured in dollars), the family or unrelated individual is considered to be in poverty; if
resources are greater than or equal to the measure of need, the family or individual is classified as
not being in poverty. The measures differ as to what is counted as resources or included in the
measure of need.
 Under the official poverty measure, the measure of need was originally computed
using family expenditure data from 1955 and food costs in 1962. Using the cost
of a tightly constrained food budget, and the average share of family income that
was spent on food, dollar amounts were computed to represent the overall
income levels (the poverty thresholds) at which a family whose basic needs
overall might have been similarly constrained. These official poverty thresholds
have been updated annually for inflation. For the resources necessary to meet that
level of need, the official poverty measure counts income in the form of cash
only, before taxes—meaning that tax credits and the value of noncash benefits
are not counted.
 Under the SPM, the measure of need is based on recent spending data from the
Consumer Expenditure Survey; namely, a percentage of median family spending
on food, clothing, shelter, and utilities (plus an extra 20% for miscellaneous
expenses such as personal care products), as opposed to being computed once
and indexed forward for inflation (as is done for the official measure). For the
resources necessary to meet that level of need, the SPM uses after-tax income
(which includes tax credits), estimates the value of certain noncash benefits (such
as food assistance), and subtracts some expenditures (such as work-related
expenses, child care expenses, and medical expenses paid out-of-pocket) that
families cannot use toward the categories of basic needs that are used to define
the SPM poverty level. This approach was intended to better reflect the economic
choices families currently face, and to better reflect the effects of government
programs on the low-income population, than does the official measure.
For the official measure in 2021, no statistically significant change from 2020 was detected in the
overall poverty rate—the percentage of people who are in poverty—or in the number of people in
poverty in the United States: 11.6% of the population, or 37.9 million people, lived in poverty. By
the official measure, poverty among most segments of the population held steady, with some
exceptions: the poverty rate fell among children and rose among the aged; it fell among workers
as more people obtained full-time year-round work, and it rose for nonworkers; and it rose among
people identifying racially as Asian.
The SPM tells a different story. The SPM differs from the official poverty measure in that the
SPM takes account of taxes and noncash resources in ways the official measure does not. Under
the SPM, the overall poverty rate fell 1.4 percentage points in 2021 (to 7.8% from 9.2% in 2020),
and nearly 4.5 million fewer people lived in poverty (25.6 million, down from 30.0 million in
2020). Under the SPM, poverty rate declines were widespread among demographic groups,
reflecting the noncash assistance and refundable tax credits provided in response to the COVID-
Congressional Research Service

1

Poverty in the United States in 2021

19 pandemic. In particular, the poverty rate for children under the SPM fell to 5.2% from 9.7% in
2020, which was the lowest rate ever recorded for that group under the SPM, and reflected the
effects of policies targeted toward children and families with children.
This report presents a general overview of poverty in the United States. It introduces the concepts
and data sources used in defining and measuring poverty. It then offers a historical perspective on
poverty at the national level by presenting trend data on the official poverty measure. Next, the
report focuses on poverty by demographic group, mainly by comparing 2021 estimates with
2020, along four characteristics:
 family structure, because poverty is defined according to the composition, needs,
and resources of families, and because antipoverty interventions have often been
targeted to families;
 age, because age groups vary in the types and sources of income available to
them, and because congressional policymaking has often focused on children and
the aged population;
 race and Hispanic origin, because poverty rates among these demographic groups
historically have had wide differences; and
 work status, because economic well-being is typically tied to the current or past
work of oneself or one’s family members.
State poverty rates are then presented to provide a geographical perspective on poverty
throughout the United States. Lastly, the report describes the SPM, a newer measure that
improves upon some of the official poverty measure’s limitations, and illustrates how the SPM
offers a different view of poverty than the official measure. This different view is particularly
relevant for examining the impact on poverty of the refundable tax credits and other measures
Congress undertook to counteract the recession related to the pandemic.
Poverty Data As Estimates: Survey Data Collection
and Poverty Measure Definitions
The numbers and percentages of those in poverty presented in this report are based on the Census
Bureau’s estimates.1 While this official measure has been regarded as a statistical yardstick rather
than a complete description of what people and families need to live,2 it does offer a measure of
economic hardship faced by the low-income population. The poverty measure compares family
income against a dollar amount called a poverty threshold, a level below which the family is
considered to be poor. The Census Bureau releases these poverty estimates every September for
the prior calendar year. Most of the comparisons discussed in this report are year-to-year. The
report only considers a number or percentage to have changed from the previous year, or to be

1 The national-level data in this report were obtained from the report by John Creamer, Emily A. Shrider, Kalee Burns,
and Frances Chen, Poverty in the United States: 2021, U.S. Census Bureau, Current Population Reports number P60-
277, September 13, 2022, at https://www.census.gov/library/publications/2022/demo/p60-277.html (hereinafter,
“Creamer et al., 2022”), and the detailed tabulations and the Current Population Survey Annual Social and Economic
Supplement (CPS ASEC) public use file that accompanied the release of that report.
2 Creamer et al., 2022, Appendix A. The characterization of the poverty measure as a statistical yardstick goes back
decades. See, for example, “U.S. Changes Yardstick on Who Is Poor,” Chicago Tribune, May 3, 1965, section 1B, p. 4.
Congressional Research Service

2

Poverty in the United States in 2021

different from another number or percentage, if the difference has been tested to be statistically
significant at the 90% confidence level.3
Effect of the Pandemic on Survey Collection and Survey Estimates
In addition to affecting people’s poverty statuses, the pandemic also affected the collection of the
data used to measure poverty, which in turn affects how the data may be interpreted. The official,
annual data on income and poverty are measured retrospectively by a Census Bureau survey
conducted in February, March, and April of the following year.4 While data for 2021 (collected in
2022) were obtained using both in-person and telephone interviewing, data collection for income
in 2020 (collected in 2021) was affected by social distancing measures that the Census Bureau
implemented from March 2020 to September 2021.5 As a result, any interviews during time that
previously would have been conducted in person, by a field representative visiting respondents at
their homes, were conducted by telephone only. That temporary switch to telephone-only
interviewing contributed to a lower response rate,6 which affected the estimates: those who did
not respond were more likely to have had low incomes than high incomes.7 Further, even though

3 Not every apparent difference in point estimates is a real difference. The official poverty measure uses information
from the CPS ASEC, which surveys about 95,000 addresses nationwide. All poverty data discussed here are therefore
estimates, which have margins of error. Error in this case refers to a difference from the true data that is caused by
using a sample instead of the entire population, not mistakes in computation or biases from imperfect data collection or
processing. Even if a survey were implemented perfectly and had collected complete and accurate information from all
respondents in the sample, surveying a different sample would likely yield slightly different estimates of the poverty
population or the poverty rate. Thus, even if the true poverty rate were exactly the same in two different years, it is
possible to get survey estimates that appear different. To report that a change has occurred in the poverty rate—that is,
that the difference between the estimates is likely not caused by sampling variability—the difference has to be large
enough that fewer than 10% of all possible survey samples would produce a difference that large (and, conversely, 90%
of the samples would not). Such a difference is said to be statistically significant at the 90% confidence level. Point
estimates whose differences are not statistically significant are described as such in this report.
4 The CPS, the source of most of the data in this report, asks an extra set of questions (the ASEC) in February, March,
and April, with most of the data collection taking place in March, in order to obtain detailed information about the
previous year’s income. According to the U.S. Census Bureau, “It was thought that since March is the month before the
deadline for filing federal income tax returns, respondents were likely to have recently prepared tax returns or be in the
midst of preparing such returns and could report income more accurately than at any other time of the year”; U.S.
Census Bureau, Current Population Survey: Design and Methodology, Technical Paper 66, October 2006, Chapter 11,
“Current Population Survey Supplemental Inquiries,” p. 11-5, https://www.census.gov/prod/2006pubs/tp-66.pdf.
5 The CPS is usually conducted using both telephone and in-person interviewing. For details, see U.S. Census Bureau,
Current Population Survey: Design and Methodology, Technical Paper 66, October 2006, https://www.census.gov/
prod/2006pubs/tp-66.pdf.
6 According to Jonathan Rothbaum and Charles Hokayem, “During the onset of the pandemic, response rates fell
sharply to 73.0% in March 2020 and 69.9% in April 2020. In March and April of this year [2021], response rates
climbed to 76.2% and 78.8%, respectively, but were still below the prepandemic trend.” See 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,” on the Census
Bureau’s “Research Matters” blog entry for September 14, 2021, at https://www.census.gov/newsroom/blogs/research-
matters/2021/09/pandemic-affect-survey-response.html. This decreased response rate could have many contributing
causes, such as a respondent’s lack of a telephone, difficulties in finding a telephone number associated with a
household selected for the survey, some respondents’ lower willingness to respond to a telephone call than to an in-
person visit, or other reasons. See Jonathan Rothbaum and Adam Bee, Coronavirus Infects Surveys Too: Nonresponse
Bias During the Pandemic in the CPS ASEC, U.S. Census Bureau, working paper number SEHSD WP2020-10,
September 2020, at https://www.census.gov/library/working-papers/2020/demo/SEHSD-WP2020-10.html.
7 Every year, when processing the CPS ASEC data, the Census Bureau applies weighting procedures that attempt to
correct for nonresponse, so that totals by age, sex, race, and Hispanic origin match independently computed totals
based on administrative data. These procedures reduced but did not eliminate nonresponse bias. In a research paper
(i.e., separate from the official reports), Census Bureau analysts estimated the likely effects of nonresponse bias on the
Congressional Research Service

3

Poverty in the United States in 2021

in-person interviewing had resumed when the 2021 data were collected, response rates did not
return to their pre-pandemic levels, and nonrespondents were still more likely to have low income
than high income.8 Thus, this drop in survey response tempers the conclusions that may be drawn
from the data, particularly when examining longer historical trends.9 The information in this
report necessarily uses the official estimates. Detailed experimental estimates that adjust for
nonresponse bias specific to the pandemic are not available for the characteristics discussed in the
report.
How the Official Poverty Measure is Computed
The Census Bureau determines a person’s poverty status by comparing his or her resources
against a measure of need. For the official measure, the term resources is defined as total family
income before taxes, and the measure of need is a dollar amount called a poverty threshold. There
are 48 poverty thresholds that vary by family size and composition. If a person lives with other
people to whom he or she is related by birth, marriage, or adoption, the money income from all
family members is used to determine his or her poverty status. If a person does not live with any
family members, his or her own income is used. Only money income before taxes is used in
calculating the official poverty measure, meaning this measure does not treat in-kind benefits
such as the Supplemental Nutrition Assistance Program (SNAP, formerly known as food stamps),
housing subsidies, or employer-provided benefits as income. Because the official measure uses
income before taxes, it also excludes refundable tax credits such as the Earned Income Tax Credit
and the Child Tax Credit, as well as stimulus payments that were made as refundable tax credits.
The poverty threshold dollar amounts vary by the size of the family (from one person not living
in a family, to nine or more family members living together) and the ages of the family members
(how many of the members are children under 18 and whether or not the family head is 65 or
older). Collectively, these poverty thresholds are often referred to as the poverty line. As a rough

2019 income and poverty estimates; an updated analysis was published online for the 2020 income and poverty
estimates. For a detailed analysis of the lower response to the CPS ASEC conducted in 2020, see Jonathan Rothbaum
and Adam Bee, Coronavirus Infects Surveys Too: Nonresponse Bias During the Pandemic in the CPS ASEC, U.S.
Census Bureau, working paper number SEHSD WP2020-10, September 2020, at https://www.census.gov/library/
working-papers/2020/demo/SEHSD-WP2020-10.html. A succinct and less technical explanation by Jonathan
Rothbaum is available on the Census Bureau’s “Research Matters” blog at https://www.census.gov/newsroom/blogs/
research-matters/2020/09/pandemic-affect-survey-response.html. The updated analysis for the 2021 CPS ASEC by
Rothbaum and Hokayem is available on the “Research Matters” blog at https://www.census.gov/newsroom/blogs/
research-matters/2021/09/pandemic-affect-survey-response.html.
8 For a comparison of income between respondents and nonrespondents, and between data with weights adjusted for
income and without income adjustments, using the 2017 through 2022 CPS ASEC linked to administrative and other
survey data, see the Census Bureau’s website; Jonathan Rothbaum and Adam Bee, “How Has the Pandemic Continued
to Affect Survey Response? Using Administrative Data to Evaluate Nonresponse in the 2022 Current Population
Survey Annual Social and Economic Supplement,” U.S. Census Bureau’s “Research Matters” blog, September 13,
2022, at https://www.census.gov/newsroom/blogs/research-matters/2022/09/how-did-the-pandemic-affect-survey-
response.html.
9 For an analysis of nonresponse in the 2020 estimates, see 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,” on the Census Bureau’s “Research Matters” blog entry
for September 14, 2021, at https://www.census.gov/newsroom/blogs/research-matters/2021/09/pandemic-affect-survey-
response.html. In an analysis of the 2019 estimates that attempted to correct for response bias, the U.S. poverty rate was
estimated to have been 11.1% (tying the previous low of 11.1% in 1973), higher than the official 2019 estimate of
10.5%. See Jonathan Rothbaum and Adam Bee, Coronavirus Infects Surveys Too: Nonresponse Bias During the
Pandemic in the CPS ASEC, U.S. Census Bureau, working paper number SEHSD WP2020-10, September 2020, at
https://www.census.gov/library/working-papers/2020/demo/SEHSD-WP2020-10.html.
Congressional Research Service

4

Poverty in the United States in 2021

guide, the poverty line in 2021 can be thought of as $27,740 for a family of four, $21,559 for a
family of three, $17,529 for a family of two, or $13,788 for an individual not living in a family,
though the official measure is actually much more detailed.10
The threshold dollar amounts are updated annually for inflation using the Consumer Price Index.
Notably, the same thresholds are applied throughout the country: no adjustment is made for
geographic variations in living expenses.11
The official poverty measure used in this report is the federal government’s definition of poverty
for statistical purposes, such as comparing the number or percentage in poverty over time. A
related definition of poverty, the poverty guidelines published by the Department of Health and
Human Services (HHS), is used for administrative purposes such as eligibility criteria for
assistance programs and will not be discussed in this report.12
The Supplemental Poverty Measure: Its Relevance in Relation to
the Official Measure
Over the past several decades, criticisms of the official poverty measure have led to the
development of an alternative research measure called the SPM, which the Census Bureau also
computes and releases. Statistics comparing the official measure with the SPM are provided at the
conclusion of this report.
The SPM includes adjustments to reflect geographic variations in housing costs, and the
estimated effects of taxes and in-kind benefits (such as housing, energy, and food assistance) on
poverty, while the official measure does not. The SPM also takes a more expansive approach than
the official measure in recognizing relationships among household members for the purpose of
identifying how those members share costs and pool resources. Furthermore, while one-time
payments such as economic stimuli are not considered as part of the official definition of income,
these payments are considered as resources in the SPM. Because some types of tax credits and
noncash benefits provide financial help to families and individuals in poverty, the SPM may be of
interest to policymakers, particularly in light of the economic stimulus payments and tax credits
provided in 2021 in response to the COVID-19 pandemic.
The official measure provides a comparison of the population below poverty over a longer period
than does the SPM, including some years before many current antipoverty assistance programs

10 To provide a general sense of the poverty line, the Census Bureau computes weighted averages of the thresholds
within each family size. For example, a family of three may consist of any of the following combinations: three adults,
two adults and one child, or one adult and two children. Each combination has its own distinct threshold. The $21,559
figure cited represents an average of those family combinations, adjusted to reflect that some types of three-person
families are more common than others. The averages are a convenience for the reader, but are not actually used to
compute poverty status for statistical reports. In actual computations, 48 thresholds are used in the official measure.
11 Unlike the poverty thresholds that are used to compute official poverty statistics, the Health and Human Services
(HHS) poverty guidelines used for administrative purposes do include separate amounts for Alaska and Hawaii.
12 The official poverty measure described in this report was established in the Office of Management and Budget’s
Statistical Policy Directive 14, May 1978, reproduced on the Census Bureau’s website at https://www.census.gov/
topics/income-poverty/poverty/about/history-of-the-poverty-measure/omb-stat-policy-14.html. It states that the official
measure is to be used for statistical purposes, but should not be construed as required for administrative purposes.
Though the poverty guidelines published by HHS use the official thresholds as part of their computation, the HHS
poverty guidelines are collectively a distinct poverty definition and are often used as a criterion in federal assistance
programs. The HHS poverty guidelines are often referred to as the federal poverty level or FPL. See CRS Report
R44780, An Introduction to Poverty Measurement, for further discussion.
Congressional Research Service

5

link to page 10 link to page 4 link to page 6 Poverty in the United States in 2021

had been developed.13 To the extent that policymakers may be interested in historical trends, the
official poverty measure affords a longer consistent view.
Historical Perspective Under the Official Poverty
Measure
Figure 1 shows a historical perspective of the number and percentage of the population below the
poverty line. The number in poverty and the poverty rates are shown from the earliest year
available (1959) through the most recent year available (2021). Because the total U.S. population
has grown over time, poverty rates are useful for historical comparisons because they control for
population growth.
Poverty rates fell through the 1960s. Since then, they have generally risen and fallen according to
the economic cycle, though during the two most recent expansions poverty rates did not fall
measurably until four to six years into the expansion. Historically notable lows occurred in 1973
(11.1%), 2000 (11.3%), and 2019 (10.5%).14 Poverty rate peaks occurred in 1983 (15.2%), 1993
(15.1%), and 2010 (15.1%).15
Poverty rates tend to rise during and after recessions, as opposed to leading economic indicators
such as new housing construction, whose changes often precede changes in the performance of
the overall economy. The poverty rate’s lag is explainable in part by the way it is measured: it
uses income from the entire calendar year.16

13 While their methodology is not discussed in this report, researchers at Columbia University have developed an
anchored SPM, which estimates what the SPM would have been in previous years before the data necessary for
computing the SPM according to current methods were available. See Christopher Wimer, Liana Fox, Irv Garfinkel,
Neeraj Kaushal, and Jane Waldfogel, “Trends in Poverty with an Anchored Supplemental Poverty Measure,” Institute
for Research on Poverty Discussion Paper No. 1416-13, December 11, 2013, at https://www.irp.wisc.edu/resource/
trends-in-poverty-with-an-anchored-supplemental-poverty-measure/.
14 The rate in 2019 is the lowest numerically, but suffered from nonresponse bias, as was described in the
“Introduction” section. Before 2019, the poverty rates in 1973 and 2000 had been considered to be tied for the lowest
measured poverty rate because they are not statistically different from each other.
15 These poverty rates may not necessarily be distinguishable from the poverty rates in their adjacent years. See
footnote 3 for an explanation of statistical significance.
16 For further historical information about poverty and recessions, see CRS Report R45854, Trends in the U.S. Poverty
Rate after Recessions, by Joseph Dalaker; and CRS Report R46939, Underemployment, Recessions, and Poverty, by
Joseph Dalaker.
Congressional Research Service

6

link to page 4 link to page 16

Poverty in the United States in 2021

Figure 1. Poverty Rate and Number of Persons in Poverty: 1959 to 2021
(poverty rates in percentages, number of persons in millions; shaded bars indicate recessions)

Source: Congressional Research Service (CRS), based on poverty data from Table A-4 of John Creamer, Emily
A. Shrider, Kalee Burns, and Frances Chen, Poverty in the United States: 2021, U.S. Census Bureau, Current
Population Reports number P60-277, September 13, 2022, at https://www.census.gov/library/publications/2022/
demo/p60-277.html. Recession dates were obtained from the National Bureau of Economic Research at
https://www.nber.org/research/data/us-business-cycle-expansions-and-contractions.
Notes: The 2019 and 2020 estimates were biased downward because of increased nonresponse associated with
telephone-only interviewing during the pandemic; response rates for 2021 did not return to their pre-pandemic
levels (for details, see the “Introduction” section). A summary of methodological changes to the Current
Population Survey Annual Social and Economic Supplement (CPS ASEC) and the poverty measure in other years,
with references to technical descriptions of the changes, is available in the Annual Statistical Supplement to the
Social Security Bulletin, Appendix C, at https://www.ssa.gov/policy/docs/statcomps/supplement/2022/apnc.pdf.
Poverty by Demographic Group
Just as the overall poverty rate for the United States experienced no statistically detectable shift
between 2020 and 2021, poverty rates by demographic group largely held steady. Notable
exceptions were the decline in the poverty rate among children and rise among the aged
population, the rise among those identifying as Asian, and the decline in the rate among workers
overall—which was accompanied by a rise in the number of full-time year-round workers.
Family Structure
Because poverty status is determined at the family level by comparing resources against a
measure of need, vulnerability to poverty may differ among families of different compositions. In
this section, poverty data by family structure are presented using the official poverty measure,
with families defined as persons related by birth, marriage, or adoption to the householder (the
person in whose name the home is owned or rented). In the “Supplemental Poverty Measure”
section of this report, a different definition will be used.
In general, women have higher poverty rates than men: 12.6% compared with 10.5% in 2021
(neither changed significantly from 2020). Historically, families with a female householder and
no spouse present (female-householder families) have had higher poverty rates than both married-
couple families and families with a male householder and no spouse present (male-householder
Congressional Research Service

7

link to page 11 link to page 13

Poverty in the United States in 2021

families). This remained true in 2021: the poverty rate among female-householder families was
23.0% (not different from 2020), compared with 12.0% for male-householder families and 4.8%
for married-couple families (Figure 2).
Even though survey nonresponse was more common among low-income families than high-
income families, and female-householder families are more likely to be low income than the other
family types, the 2021 female-householder poverty rate is the latest in a series of lower poverty
rates for this group compared with previous decades.17
No discernible changes were detected in the poverty rates for married-couple families (4.8% in
2020), male-householder families (12.0%), or persons living alone or with non-relatives only
(19.5%).
Figure 2. Poverty Rates of Families by Family Structure: 2021
(poverty rates in percentages)

Source: Congressional Research Service (CRS), based on poverty data from Table A-1 of John Creamer, Emily
A. Shrider, Kalee Burns, and Frances Chen, Poverty in the United States: 2021, U.S. Census Bureau, Current
Population Reports number P60-277, September 13, 2022, at https://www.census.gov/library/publications/2022/
demo/p60-277.html.
Notes: The poverty rates above include only families with a householder (the survey’s reference person for the
household; typically the person in whose name the home is owned or rented). The Census Bureau defines a
family as those living together related by birth, marriage, or adoption.
Age
When examining poverty by age, the three main groups (under 18, 18 to 64, and 65 and older) are
noteworthy for distinct reasons. People under age 18 are typically dependent on other family
members for income, particularly young children below their state’s legal working age. People
aged 18 to 64 are generally thought of as the working-age population and typically have wages
and salaries as their greatest source of income. People aged 65 and older, referred to as the aged
population, are often eligible for retirement, and those who do retire typically experience a
change in their primary source of income, such as from earnings to Social Security.
As shown in Figure 3, poverty fell among children and rose among the aged population in 2021.
Among children, 11.1 million (15.3%) lived in poverty, down from 11.8 million (16.0%) in 2020.

17 Poverty rates for female-householder families are available from 1959 onward. Until 1964, the rates for this group
were estimated to be above 40%. From 1964 through 1997, poverty rates for female-householder families were
between 30% and 40%, and from 1998 to 2014, they hovered close to or below 30% except during the years following
the Great Recession, when they peaked above 30%. From 2015 to 2021, the poverty rates for this group remained
below 30%.
Congressional Research Service

8

link to page 16 link to page 13 link to page 10 Poverty in the United States in 2021

Official poverty rate estimates use family income before taxes and thus do not count tax credits or
stimulus payments—using post-tax income produces a steeper drop in the measured poverty rate
for children in 2021 (discussed in the “Supplemental Poverty Measure” section). Among the aged
population, 5.8 million (10.3%) lived in poverty in 2021, up from 4.9 million (8.9%) in 2020.
Because Social Security constitutes the majority of family income among the aged below or near
poverty, the rise in their poverty rate can be partially explained by comparing the cost-of-living
adjustment (COLA) for Social Security with the inflation-adjustment of the poverty thresholds.18
The COLA for Social Security paid in January 2021 was 1.3%, which was lower than the increase
in the poverty thresholds from 2020 to 2021 (4.7%); this suggests that the rise in poverty
thresholds outpaced the rise in income for some seniors.19 The working age population
experienced no significant change in their number in poverty or their poverty rate: 21.0 million
(10.5%) were below poverty in 2021.
Figure 3 illustrates poverty rates historically by age because the overall poverty rate (seen in
Figure 1) masks the historical decline in poverty among the aged population. Before 1974, the
poverty rate for those aged 65 and over was the highest of the three age groups. In 1966, people
aged 65 and over had a poverty rate of 28.5%, compared with 17.6% for those under 18 and
10.5% for working-age adults. By 1974, the poverty rate for people aged 65 and over had fallen
to 14.6%, compared with 15.4% for people under 18 and 8.3% for working-age adults. Since
then, people under 18 have had the highest poverty rate of the three groups. The poverty rate
among the 65-and-older population eventually fell below the poverty rate of the working-age
population, and until 2021 had trended below that group since the early 2000s.

18 Among the aged in poverty, Social Security constituted 56.0% of family income, on average (median: 79.7%).
Among the aged between 100% and 149% of poverty, Social Security constituted 79.8% of family income, on average
(median: 99.8%). This percentage was lower but still sizeable among higher-income seniors (an average of 40.2% and
median of 34.6% for seniors at or above 150% of poverty); author’s computations, using the 2022 CPS ASEC public-
use file.
19 The COLA paid in January 2022 was 5.8%, which was greater than the 1.3% paid in January 2021. Poverty statistics
for the aged population for 2022 will not be available until September 2023. For a discussion of how COLA for Social
Security is computed and historical COLA amounts, see CRS Report 94-803, Social Security: Cost-of-Living
Adjustments, by Paul S. Davies and Tamar B. Breslauer.
Congressional Research Service

9

link to page 4 link to page 14

Poverty in the United States in 2021

Figure 3. Poverty Rates by Age: 1959 to 2021
(poverty rates in percentages; shaded bars indicate recessions)

Source: Congressional Research Service (CRS), based on poverty data from Table A-5 of John Creamer, Emily
A. Shrider, Kalee Burns, and Frances Chen, Poverty in the United States: 2021, U.S. Census Bureau, Current
Population Reports number P60-277, issued September 13, 2022. Recession dates were obtained from the
National Bureau of Economic Research at https://www.nber.org/research/data/us-business-cycle-expansions-and-
contractions.
Notes: The 2019 and 2020 estimates were biased downward because of increased nonresponse associated with
telephone-only interviewing during the pandemic; response rates for 2021 did not return to their pre-pandemic
levels (for details, see the “Introduction” section). A summary of methodological changes to the Current
Population Survey Annual Social and Economic Supplement (CPS ASEC) and the poverty measure in other years,
with references to technical descriptions of the changes, is available in the Annual Statistical Supplement to the
Social Security Bulletin, Appendix C, at https://www.ssa.gov/policy/docs/statcomps/supplement/2022/apnc.pdf.
Race and Hispanic Origin20
Poverty rates vary by race and Hispanic origin, as shown in Figure 4. In surveys, Hispanic origin
is asked about separately from race; accordingly, people identifying as Hispanic may be of any
race.21 Comparing poverty rates in 2020 and 2021, among the Asian population the poverty rate
rose from 8.1% to 9.3%. No statistical differences were observed in the poverty rates for the
White non-Hispanic, Black, American Indian and Alaska Native, two or more races, or Hispanic
populations.22

20 Since 2002, federal surveys have asked respondents to identify with one or more races; previously, they could choose
only one. The groups in this section represent those who identified with one race alone. Another approach is to include
those who selected each race group either alone or in combination with one or more other races. Those data are also
available on the Census Bureau’s website at https://www.census.gov/library/publications/2022/demo/p60-277.html,
where they are published in Creamer et al., 2022; and in accompanying historical data tables.
21 Except for the two or more races population and the Hispanic population, the racial categories listed in this section
include those identifying with one race only.
22 The Asian, Black, American Indian and Alaska Native, and two or more races populations all include Hispanics.
Congressional Research Service

10

Poverty in the United States in 2021

Figure 4. Poverty Rates by Race and Hispanic Origin: 2021
(poverty rates in percentages)

Source: Congressional Research Service (CRS), based on poverty data from Table A-1 of John Creamer, Emily
A. Shrider, Kalee Burns, and Frances Chen, Poverty in the United States: 2021, U.S. Census Bureau, Current
Population Reports number P60-277, September 13, 2022, at https://www.census.gov/library/publications/2022/
demo/p60-277.html.
Notes: People of Hispanic origin may be of any race. Additionally, respondents may identify with one or more
racial groups. Except for “All persons,” “Two or more races,” and “Hispanic,” the remaining groups shown
include those who identified with one race only. The “non-Hispanic White alone” group includes only the White
non-Hispanic population, while the “Black alone,” “Asian alone,” and “American Indian and Alaska Native alone”
groups include persons who identify as Hispanic. Data for Native Hawaiians and Other Pacific Islanders are not
shown separately.
Work Status
While having a job reduced the likelihood of being in poverty, it did not guarantee that a person
or his or her family would avoid poverty. Among the population aged 18 to 64 living in poverty,
34.3% had jobs in 2021. Poverty rates among workers in this age group were 4.7% for all workers
(down from 5.0% in 2020), 1.8% for full-time year-round workers (up from 1.6% in 2020), and
12.2% for part-time or part-year workers (up from 11.3% in 2020). Because many persons who
were working part-time or part-year in 2020 obtained full-time year-round work in 2021, poverty
rates for each of those separate groups rose while the poverty rate among workers as a whole
fell.23 Among those who did not work at least one week in 2021, 30.0% were poor (up from
28.8% in 2020).

23 The seeming inconsistency—that the component groups experienced poverty rate increases while the total of workers
aged 18 to 64 experienced a poverty rate decrease—is not actually inconsistent. All numbers below are rounded; details
may not sum to totals.
Full-time year-round: 2020: 1,600,000 poor out of 100,000,000 (1.6%); 2021: 2,000,000 poor out of 111,000,000
(1.8%) ↑
Part-time or part-year: 2020: 6,000,000 poor out of 53,000,000 (11.3%); 2021: 5,100,000 poor out of 42,000,000
(12.2%) ↑
Total (after rounding): 2020: 7,600,000 poor out of 154,000,000 (5.0%); 2021: 7,200,000 poor out of 153,000,000
(4.7%) ↓
In 2021, the total of full-time year-round workers rose by 11 million and the total of part-time or part-year workers fell
by 11 million. This is a net change; the CPS ASEC sample is not the same in consecutive years, so it is not possible to
identify who exactly moved into full-time year-round work status, or directly compute their poverty rate. However, the
net gain of full-time year-round workers in 2021 was more likely to be poor than full-time year-round workers overall
Congressional Research Service

11

link to page 16 link to page 6 link to page 16 Poverty in the United States in 2021

Because poverty is a family-based measure, a change in one member’s work status can affect the
poverty status of his or her entire family. Among all 18- to 64-year-olds who did not have jobs in
2021, 57.8% lived in families in which someone else did have a job.24 Among 18- to 64-year-olds
with income below the poverty line and without jobs, 21.7% lived in families where someone else
worked.25
Poverty Rates by State26
Poverty is not equally prevalent in all parts of the country. Figure 5 shows states with relatively
high poverty rates across parts of the Appalachians, the Southwest, and the Deep South.
Mississippi’s poverty rate (18.1%), seemingly the highest, was in a statistical tie with Louisiana’s
(17.3%) and New Mexico’s (16.7%—see footnote 3 for a discussion of statistical significance).
The poverty rate in New Hampshire (5.6%), seemingly the lowest, was not statistically
distinguishable from that in Utah (7.5%). When comparing poverty rates geographically, the
official poverty thresholds are not adjusted for geographic variations in the cost of living—the
same thresholds are used nationwide. As such, an area with a lower cost of living accompanied by
lower wages will appear to have a higher poverty rate than an area with a higher cost of living
and higher wages, even if individuals’ purchasing power were exactly the same in both areas.

in 2020, thereby driving up that group’s poverty rate. Meanwhile, the net loss of part-time or part-year workers in 2021
was less likely to be poor than part-time or part-year workers overall in 2020, meaning that those left behind in 2021
were more likely to be poor, also driving up that group’s poverty rate. Taken together, however, 400,000 fewer workers
were poor in 2021 than in 2020—a faster rate of decline than the population of workers ages 18-64, meaning the
overall poverty rate of workers fell.
24 Author’s tabulation using 2022 CPS ASEC public use file.
25 Ibid.
26 These state estimates are based on multi-year averages using 2020-2022 CPS ASEC data (which collected income
data for calendar years 2019-2021), rather than from the 2022 CPS ASEC alone. Because the CPS ASEC surveys
90,000 to 100,000 addresses nationwide, it is sometimes difficult to obtain reliable estimates for small populations or
small geographic areas—the sample may not have selected enough people from that group or area to provide a
meaningful estimate. Using multi-year averages therefore affords greater statistical precision for comparing state areas.
Three-year averages for 2019-2021 are used in Figure 5 and in comparing poverty rates across states. Alternatively, the
American Community Survey (ACS) is typically recommended by the Census Bureau for estimates at the state level
and smaller areas.
Congressional Research Service

12

Poverty in the United States in 2021

Figure 5. Three-Year Average State Poverty Rates: 2019-2021
(poverty rates in percentages)

Source: Congressional Research Service, using data from the U.S. Census Bureau, Current Population Survey,
2020-2022 Annual Social and Economic Supplements, “Interrelationships of Three Year Average State Poverty
Rates: 2019–2021,” supplementary table to accompany John Creamer, Emily A. Shrider, Kalee Burns, and Frances
Chen, Poverty in the United States: 2021, U.S. Census Bureau, Current Population Reports number P60-277,
September 13, 2022, at https://www.census.gov/library/publications/2022/demo/p60-277.html; also see
https://www2.census.gov/programs-surveys/demo/tables/p60/277/
interrelationships_poverty_state_grid_table.xlsx.
Notes: Data by state are based on three-year averages in order to reduce sampling error.
Supplemental Poverty Measure
Criticisms of the official poverty measure led to the development of the SPM. Described below
are the development of the official measure, its limitations, attempts to remedy those limitations,
the research efforts that eventually led to the SPM’s first release in November 2011, and a
comparison of poverty rates in 2021 based on the SPM and the official measure.27
How the Official Poverty Measure Was Developed
The poverty thresholds were originally developed in the early 1960s by Mollie Orshansky of the
Social Security Administration. Rather than attempt to compute a family budget by using prices
for all essential items that low-income families need to live, Orshansky focused on food costs.28

27 For a more thorough discussion of the SPM’s development and methodology, see CRS Report R45031, The
Supplemental Poverty Measure: Its Core Concepts, Development, and Use.
28 While Orshansky did not attempt to compute a complete basket of goods and services, her focus on food costs was
Congressional Research Service

13

Poverty in the United States in 2021

Unlike other goods and services such as housing or transportation, which did not have a generally
agreed-upon level of adequacy, minimum standards for nutrition were known and widely
accepted. According to a 1955 U.S. Department of Agriculture (USDA) food consumption survey,
the average amount of their income that families spent on food was roughly one-third. Therefore,
using the cost of a minimum food budget29 and multiplying that figure by three yielded a figure
for total family income. That computation was possible because USDA had already published
recommended food budgets as a way to address the nutritional needs of families experiencing
economic stress. Some additional adjustments were made to derive poverty thresholds for two-
person families and individuals not living in families to reflect the relatively higher fixed costs of
smaller households.
Motivation for a Supplemental Poverty Measure
While the official poverty measure has been used for nearly 60 years as the source of official
statistics on poverty in the United States, it has received criticism over the years for several
reasons. First, it does not take into account benefits from most of the largest programs that aid the
low-income population. For instance, it uses money income before taxes—meaning that it does
not necessarily measure the income available for individuals to spend, which for most people is
after-tax income. Therefore, any effects of tax credits designed to assist persons with low income
are not captured by the official measure. The focus on money income also does not account for
in-kind benefit programs designed to help the poor, such as SNAP or housing assistance.
The official measure has also been criticized for the way it characterizes families’ and individuals’
needs in the poverty thresholds. That is, the method used to compute the dollar amounts used in
the thresholds, which were originally based on food expenditures in the 1950s and food costs in
the 1960s, does not accurately reflect current needs and available goods and services.30 The
official measure also does not take account of the sharing of expenses and income among
household members not related by birth, marriage, or adoption. And, as mentioned earlier, the
official thresholds do not take account of geographic variations in the cost of living.

already a more detailed empirical approach to poverty measurement than were the dollar amounts used in the 1964
Economic Report of the President, issued by the Council of Economic Advisers (chapter 2, “The Problem of Poverty in
America”). In that report, a flat figure of $3,000 was used for all families and $1,500 for unrelated individuals. See also
Economic Report of the President (1964), https://fraser.stlouisfed.org/title/45#8135. For a thorough history of the
official poverty measure, see Gordon Fisher, The Development of the Orshansky Thresholds and Their Subsequent
History as the Official U.S. Poverty Measure, 1992, rev. 1997, reproduced on the Census Bureau’s website at
https://www.census.gov/library/working-papers/1997/demo/fisher-02.html.
29 The stringency of this food budget, called the Economy Food Plan, was characterized by Betty Peterkin and Faith
Clark, “Money Value and Adequacy of Diets Compared with the USDA Food Plans,” Family Economics Review,
September 1969, p. 8: “Diets were considered good if they provided the recommended allowances (1963) for all
nutrients, and fair or better if they provided at least two-thirds of the allowances.” They presented results of a 1965
survey of urban families indicating that less than 50% of families on the Economy Food Plan had a fair or better diet
(implying at least 50% did not), while less than 10% of the families on the plan had a good diet; see https://archive.org/
details/familyeconomicsr6251inst_48.
30 Criticisms have been discussed in the mainstream press as well as academia. A 1988 article (Spencer Rich, “Drawing
the Line Between Rich, Poor,” The Washington Post, September 23, 1988, https://www.washingtonpost.com/archive/
politics/1988/09/23/drawing-the-line-between-rich-poor/60f5dbeb-dab3-4a42-819a-2dea34e7854e/) documented
dissatisfaction about the official measure. This came from both those claiming it was too high, citing its failure to
capture the effects of in-kind benefits for the poor and its overstatement of inflation, and those claiming it was too low,
based on the fact that if the thresholds were derived using more recent household consumption data, they would be
based on roughly five times the cost of food, not three times as Orshansky had computed in the early 1960s.
Congressional Research Service

14

link to page 18 link to page 18 Poverty in the United States in 2021

In 1995, a panel from the National Academy of Sciences issued a report, Measuring Poverty: A
New Approach, which recommended improvements to the poverty measure.31 Among the
suggested improvements were to have the poverty thresholds reflect the costs of food, clothing,
shelter, utilities (FCSU), and a little bit extra to allow for miscellaneous needs;32 to broaden the
definition of family; to include geographic adjustments as part of the measure’s computation; to
include the out-of-pocket costs of medical expenses in the measure’s computation; and to subtract
work-related expenses from income. An overarching goal of the recommendations was to make
the poverty measure more closely aligned with the real-life needs and available resources of the
low-income population, as well as the changes that have taken place over time in their
circumstances, owing to changes in the nation’s economy, society, and public policies (see Table
1).
After over a decade-and-a-half of research to implement and refine the methodology suggested by
the panel, conducted both from within the Census Bureau as well as by other federal agencies and
the academic community, the Census Bureau issued the first report using the SPM in November
2011.33
Table 1. Differences Between the Official and Supplemental Poverty Measures

Official Poverty Measure

Supplemental Poverty Measure
Resource units
People related by birth, marriage, or adoption
People related by birth, marriage, adoption, plus
(families)
(official Census Bureau definition of family).
unrelated and foster children, and cohabiting
People aged 15 and older not related to anyone
partners and their children or other relatives (if
else in the household are considered as their
any) are considered as “SPM resource units”
own economic units.
(sharing resources and expenses together).

31 Constance F. Citro and Robert T. Michael, eds., Measuring Poverty: A New Approach, Panel on Poverty and Family
Assistance: Concepts, Information Needs, and Measurement Methods, Committee on National Statistics, National
Research Council (Washington, DC: National Academies Press, 1995), available at https://www.nap.edu/read/4759/
chapter/1.
32 The portion of the SPM threshold that represents FCSU is set to 85% of the median FCSU expenditures among
families with children, according to the Consumer Expenditure Survey, with families in this case defined as the
consumer units measured within that survey. That amount is meant to represent a basic, modest level of FCSU. An
extra 20% of that amount is then added to represent other basic needs, such as personal care products, cleaning
supplies, and non-work-related transportation.
33 The effort to consolidate the previous research and create the SPM was done under the auspices of an Interagency
Technical Working Group (ITWG) led by the Office of Management and Budget (OMB) and received public
commentary via a Federal Register notice (Federal Register, vol. 75 no. 101, Wednesday, May 26, 2010, pp. 29513-
29514, https://www.federalregister.gov/documents/2010/05/26/2010-12628/developing-a-supplemental-poverty-
measure). The Federal Register notice referenced a report by the ITWG ("Observations from the Interagency Technical
Working Group on Developing a Supplemental Poverty Measure”), which has since been moved to a new URL at
https://www.census.gov/content/dam/Census/topics/income/supplemental-poverty-measure/spm-twgobservations.pdf.
The comments that the Census Bureau received on that report are available on the Census Bureau’s website at
https://www.census.gov/content/dam/Census/topics/income/supplemental-poverty-measure/redactedcomments.pdf.
These and additional methodological documents on the SPM are available at https://www.census.gov/topics/income-
poverty/supplemental-poverty-measure/guidance/methodology.html.
Congressional Research Service

15

link to page 16 Poverty in the United States in 2021

Official Poverty Measure

Supplemental Poverty Measure
Needs
 Vary according to family size and ages of
 Vary according to the size and composition of
(thresholds)
family members.
the resource unit (see above).
 Dol ar amounts based on the cost of a food
 Dol ar amounts based on consumer expenditure
plan for families in economic stress in the
data for food, clothing, shelter, and utilities
early 1960s, times three (with adjustments
(FCSU), with adjustments by homeownership
for two-person families and individuals).
and mortgage or rental status.
 Updated for inflation using the Consumer
 Based on five years of consumer expenditure
Price Index for All Urban Consumers (CPI-
data (not fixed at one point and trended
U).
forward), lagged one year from the most recent

for consistency with the CPS ASEC data available
No geographic cost adjustments.
for computing in-kind benefit amounts for the
SPM thresholds.
 Housing costs geographically adjusted for
individual metropolitan areas and the entire
nonmetropolitan area within states.
Resources
Money income before taxes (includes 18 private
Money income (both private and government
and government sources of income, including
sources) after taxes
Social Security, cash assistance, and other

minus: work expenses, child care expenses,
sources of cash income).
child support paid, out-of-pocket medical
expenses.

plus: tax credits (such as the Child Tax
Credit and the Earned Income Tax Credit)
and the value of in-kind benefits (such as food
and housing subsidies).
Source: Congressional Research Service (CRS) summary of methodological discussion from Appendix B in John
Creamer, Emily A. Shrider, Kalee Burns, and Frances Chen, Poverty in the United States: 2021, U.S. Census Bureau,
Current Population Reports number P60-277, September 13, 2022, at https://www.census.gov/library/
publications/2022/demo/p60-277.html.
Notes: For caveats, see the “Supplemental Poverty Measure” section of this report.
Official and Supplemental Poverty Findings for 202134
Compared with the official measure, the SPM takes into account greater detail of individuals’ and
families’ living arrangements and provides a more up-to-date accounting of the costs and
resources available to them. Because the SPM recognizes greater detail in relationships among
household members and geographically adjusts housing costs, it provides an updated rendering,
compared with the official measure, of the circumstances in which the poor live. In that context,
some point out that the SPM’s measurement of taxes, transfers, and expenses may offer
policymakers a clearer view of how government policies affect the population living in poverty
today. However, the SPM was developed as a research measure, and the Office of Management
and Budget set the expectation that it would be revised periodically to incorporate improved
measurement methods and newer sources of data as they became available; it was not developed
for administrative purposes.35 The fact that tax liabilities and credits are modeled, or that in-kind

34 Data in this section are available in Appendix B of Creamer et al., 2022.
35 The Health and Human Services Poverty Guidelines were developed for administrative purposes—they are a
simplification of the official poverty measure. For details, see CRS Report R44780, An Introduction to Poverty
Congressional Research Service

16

link to page 22 link to page 6 Poverty in the United States in 2021

benefits are estimated using limited data, can be useful to bear in mind when comparing SPM
estimates with official poverty estimates, or when any changes to the SPM methodology become
implemented in the future.36 Conversely, the official measure’s consistency over a longer time
span makes it easier for policymakers and researchers to make historical comparisons.
Under the SPM, the profile of the poverty population is different than under the official measure.
The SPM poverty rate was 3.8 percentage points lower in 2021 than the official poverty rate
(7.8%, compared with 11.6%; see Figure 6),37 and 1.4 percentage points lower than the
corresponding SPM poverty rate in 2020 (9.2%).
This is the second year in which the SPM poverty rate has been lower than the official rate: in
2019 and before, SPM poverty rates had been higher than official poverty rates. Two factors help
to explain this reversal, and both highlight policy responses to the pandemic in 2020 and 2021.
First, the SPM counts some resources that the official measure does not, such as refundable tax
credits (which include stimulus payments). The 2021 SPM poverty rate would have been 2.7
percentage points higher if the third round of stimulus payments had not been included in that
measure; the 2020 SPM poverty rate would have been 3.6 percentage points higher if the first two
rounds of 2020 stimulus payments had not been included. These payments did not exist in 2019.
Second, the SPM defines families differently from the official measure, and this can make cash
payments like unemployment insurance—which are included in both measures—appear to have
larger antipoverty effects in the SPM because more people in a family are counted as having
benefited from their receipt.38 The sharp increase in unemployment insurance (UI) payments in
2020, in response to layoffs related to the COVID-19 pandemic, highlighted how the SPM’s
family definition makes a difference in who gets classified as poor. Under the SPM,
approximately 0.7 million more persons in 2020 were counted as living in families that received

Measurement, by Joseph Dalaker.
36 For instance, work expenses such as commuting costs can be difficult to pin down precisely for every person or
family, because they often influence and are influenced by a person’s or family’s decision about where to live. Rather
than attempting to estimate the relevant work expenses for every family, in the SPM a flat amount is assigned to
workers, multiplied by the number of weeks they worked. Some researchers have also found that the tax model used in
the SPM underestimates refundable tax credits, in comparison with administrative data, which particularly affects
families with children. Therefore, refinements to the SPM methodology based on the ongoing SPM research may not
be trivial. Working papers that present results of research into SPM methodology may be found on the Census Bureau’s
website at https://www.census.gov/topics/income-poverty/supplemental-poverty-measure/library/working-papers.html.
37 To establish a more accurate comparison with the SPM, a set of poverty estimates using the official measure was
recomputed to include unrelated individuals under age 15 (such as foster children) who are not normally included in the
official measure. Additionally, both the SPM and the official poverty measure were affected by increased rates of
nonresponse in the CPS ASEC in 2021 and 2020 (see footnote 6 for detailed references). After adjusting for the
increase in nonresponse bias, the SPM poverty rate in 2019 (based on the 2020 CPS ASEC) was 11.9% compared with
11.1% for the official poverty rate. As with the estimates based on the official measure, the SPM estimates discussed in
this report are the estimates as published, because adjusted estimates are not available for the characteristics examined
in the report.
38 The official measure does not treat cohabiting partners as being in the same family unit, while the SPM does. As a
result, the antipoverty effects of unemployment insurance payments are greater as measured under the SPM. Among
members of cohabiting partner families, the SPM counted approximately 5.2 million fewer persons in 2021 compared
with the official measure (Creamer et al., 2022, Appendix Table B-4). In 2020, the corresponding figure was 4.2
million (see Liana Fox and Kaylee Burns, The Supplemental Poverty Measure: 2020, U.S. Census Bureau, Current
Population Reports number P60-275, September 2021, Appendix Table 1, at https://www.census.gov/library/
publications/2021/demo/p60-275.html; hereinafter, “Fox and Burns, 2021”). In 2019, the corresponding figure was 2.6
million (see Liana Fox, The Supplemental Poverty Measure: 2019, U.S. Census Bureau, September 2020, Appendix
Table 1, at https://www.census.gov/content/dam/Census/library/publications/2020/demo/p60-272.pdf).
Congressional Research Service

17

Poverty in the United States in 2021

UI compared with the official measure (2.7 million compared in 2.0 million).39 In 2021, UI
receipt declined slightly compared with 2020, but it remained elevated compared with receipt
before the pandemic. Under the SPM, 0.4 million more persons were counted as living in families
that received UI than were counted under the official measure (nearly 2.0 million compared with
over 1.5 million). Before the pandemic, in 2019, approximately 9.3 million persons were poor
under the official measure but not the SPM, with 0.3 million in SPM-defined families and 0.2
million in official measure-defined families counted as having received UI.40
Fewer people aged 18 to 64 were estimated as being in poverty under the SPM (7.9%, compared
with 10.5% under the official measure), as were people under age 18 (5.2%, compared with
15.3%, with foster children included). The poverty rate for people age 65 and older was higher
under the SPM (10.7%) than under the official measure (10.3%).41
As with the overall poverty rate, the results for working-age adults in 2021 run counter to the pre-
pandemic comparisons between the SPM and the official measure, again, likely due in large part
to refundable tax credits, including stimulus payments.42 Working-age adults usually have a
higher SPM poverty rate than official poverty rate because the SPM deducts work expenses and
child care expenses from income, but in both 2020 and 2021 the increase in refundable tax credits
(including stimulus payments) made the SPM poverty rate lower for that group than under the
official measure.
While children under 18 have consistently already had SPM poverty rates below the official rate
(because refundable tax credits are often targeted to families with children), in 2020 and 2021 the
stimulus payments seem to have made this difference even more striking.43 Notably, the 2021
SPM poverty rate for children was 5.2%, the lowest SPM poverty rate on record for this group;
this was approximately 10.1 percentage points lower than their corresponding official poverty
rate, and 4.5 percentage points lower than their 2020 SPM poverty rate (9.7%).
While geographically adjusted thresholds usually result in regional differences in how the SPM
and official poverty rates compare, reflecting the differences in housing costs by region, that was
not the case in 2021. The poverty rate in 2021 was lower under the SPM than under the official

39 Author’s computations, using data from the 2021 CPS ASEC.
40 Author’s computations, using data from the 2020 CPS ASEC. The impact on the SPM of the increase in
unemployment insurance payments can also be seen by recomputing SPM poverty rates without including these
payments. If unemployment insurance payments were not counted in the SPM, the 2021 SPM poverty rate would have
been 0.7 percentage points higher and the 2020 rate would have been 1.7 percentage points higher; in contrast, the 2019
rate would have been approximately 0.2 percentage points higher (see Creamer et al., 2022, Appendix Table B-7, and
Fox and Burns 2021, Appendix Table 6).
41 The SPM deducts medical out-of-pocket expenses from income, while the official measure does not. Adults 65 and
older are more likely to incur those expenses than are other age groups; hence, SPM poverty rates for those 65 and
older are typically higher under the SPM than the official measure.
42 The SPM counts refundable tax credits, including the two stimulus payments in 2020 and the third round of stimulus
payments in 2021, but the official measure does not. If the SPM had not included stimulus payments, the 2021 SPM
poverty rate among 18- to 64-year-olds would have been approximately 2.5 percentage points higher, and
approximately 3.2 percentage points higher in 2020 (Creamer et al., 2022, Table B-7). These payments did not exist in
2019. If other types of refundable tax credits were not included, the 2020 SPM poverty rate would have been 1.3
percentage points higher and the 2021 rate would have been 2.3 percentage points higher. In contrast, the
corresponding 2019 SPM poverty rate would have been approximately 1.8 percentage points higher (Fox and Burns
2021, Appendix Table 6).
43 The 2021 SPM poverty rate among children would have been 3.1 percentage points higher if stimulus payments were
not included and 6.7 percentage points higher if other refundable tax credits were not included; the corresponding
figures for 2020 were 4.5 and 3.7 percentage points, respectively (Creamer et al., 2022, Table B-7). The 2019 SPM
poverty rate for children would have been approximately 5.7 percentage points higher if refundable tax credits were not
included (there were no stimulus payments that year). See Fox and Burns 2021, Appendix Table 6.
Congressional Research Service

18

Poverty in the United States in 2021

measure for all four major geographic regions of the country. In the Northeast, the SPM poverty
rate was 7.6% compared with 10.1% under the official measure; for the Midwest, 5.6% compared
with 10.3%; for the South, 8.4% compared with 13.2%; and in the West, 8.9% compared with
11.0%.
Figure 6. Poverty Rates Under Official Measure and Supplemental Poverty Measure
for the United States, by Age and by Region: 2021
(poverty rates in percentages)

Source: Congressional Research Service, based on data from Table B-3 in John Creamer, Emily A. Shrider,
Kalee Burns, and Frances Chen, Poverty in the United States: 2021, U.S. Census Bureau, Current Population
Reports number P60-277, September 13, 2022, at https://www.census.gov/library/publications/2022/demo/p60-
277.html.
Notes: Figures include unrelated individuals under age 15 (such as foster children), who are not usually included
in official poverty estimates.

Author Information

Joseph Dalaker

Analyst in Social Policy

Acknowledgments
Amber Hope Wilhelm, CRS Visual Information Specialist, assisted with the figures, and Calvin DeSouza,
CRS Geospatial Information Systems Analyst, assisted with mapping the data.
Congressional Research Service

19

Poverty in the United States in 2021

Disclaimer
This document was prepared by the Congressional Research Service (CRS). CRS serves as nonpartisan
shared staff to congressional committees and Members of Congress. It operates solely at the behest of and
under the direction of Congress. Information in a CRS Report should not be relied upon for purposes other
than public understanding of information that has been provided by CRS to Members of Congress in
connection with CRS’s institutional role. CRS Reports, as a work of the United States Government, are not
subject to copyright protection in the United States. Any CRS Report may be reproduced and distributed in
its entirety without permission from CRS. However, as a CRS Report may include copyrighted images or
material from a third party, you may need to obtain the permission of the copyright holder if you wish to
copy or otherwise use copyrighted material.

Congressional Research Service
R47354 · VERSION 1 · NEW
20

Download PDF

Download EPUB

Revision History

Dec. 27, 2022

HTML · PDF

Metadata

Topic areas

Domestic Social Policy

Economic Policy

Report Type: CRS Report

Source: CRSReports.Congress.gov

Raw Metadata: JSON