In 2021, the Coronavirus pandemic interrupted homelessness data collection, specifically the Point-in-Time Count. Thus, the current version of the State of Homelessness reflects limited updates as compared to the previous year. Since 2020 was the last year for which full Point-in-Time data is available, the analysis and charts focused on that year continue to be highlighted here. Updated and new sections are flagged in the report’s subheadings.
In January 2020 , there were 580,466 people experiencing homelessness on our streets and in shelters in America.i Most were individuals (70 percent), and the rest were people in families with children. They lived in every state and territory, and they include people from every gender, racial, and ethnic group. However, some groups are far more likely than others to become homeless.
Special Populations. Historically, policymakers and practitioners at every level of government have focused special attention on specific populations and subpopulations.
For example, decision-makers are often concerned about children and young people due to their developmental needs and the potential life-long consequences of hardships in early in life. People in families with children make up 30 percent of the homeless population. Unaccompanied youth (under age 25) account for six percent of the larger group.
People experiencing “chronic homelessness” belong to another group that is often singled out for attention. These individuals have disabilities and have also: 1) been continuously homeless for at least a year; or 2) experienced homelessness at least four times in the last three years for a combined length of time of at least a year. Chronically homeless individuals are currently 19 percent of the homeless population.
Finally, due to their service to our country, veterans are often analyzed separately from the larger group. They represent only six percent of people experiencing homelessness.
Populations Most at Risk. Although the homeless population is diverse, inequalities are evident among subgroups. To identify meaningful differences among groups, it is necessary to look beyond overall population counts. Comparing rates of homelessness (or a group’s homeless counts within the context or its overall size) reveals which groups are more likely to experience homelessness (or which ones are more at-risk of being in these circumstances).
Risk is significantly tied to gender, race, and ethnicity.ii Males are far more likely to experience homelessness than their female counterparts. Out of every 10,000 males, 22 are homeless. For women and girls, that number is 13. Gender disparities are even more evident when the focus is solely on individual adults (the most significant subgroup within homelessness). The overwhelming majority (70 percent) are men.
Race is another significant predictor. As with so many other areas of American life, historically marginalized groups are more likely to be disadvantaged within housing and homelessness spheres. Higher unemployment rates, lower incomes, less access to healthcare, and higher incarceration rates are some of the factors likely contributing to higher rates of homelessness among people of color.
Numerically, white people are the largest racial group within homelessness, accounting for more than a quarter-million people. However, historically marginalized racial and ethnic groups are often far more likely to experience homelessness. The reasons for the disparities are many and varied but tend to fall under the umbrellas of racism and caste. Throughout American history, private actors have contributed to the status quo, but so has government via actions and inactions resulting in limited housing opportunities, suppressed wages, and other unhelpful outcomes.
Native Hawaiians and other Pacific Islanders have the highest rate of homelessness (109 out of every 10,000 people).iii Groups such as Native Americans (45 out of every 10,000) and Black or African Americans (52 out of every 10,000) also experience elevated rates. Importantly, these rates are much higher than the nation’s overall rate of homelessness (18 out of every 10,000).
Unsheltered Homelessness. The nation has a system of temporary shelters that reaches 354,000 people on a given night. However, some still sleep in locations not ordinarily designated for that purpose (for example, sidewalks, subway trains, vehicles, or parks). Unsheltered people are considered particularly vulnerable due to their exposure to the elements and lack of safety, among other things.
Homeless programs and systems provide shelter for most people experiencing homelessness (61 percent in 2020 ). However, significant variation exists among populations and subpopulations. For example, children are often a priority for homeless services systems. As a result, families with children are least likely to be unsheltered (only ten percent of unsheltered people were living in families with children). However, young people not living with their families do not enjoy the same access to services—50 percent of unaccompanied homeless youth in 2020 were unsheltered.
Individuals experiencing homelessness on their own are particularly at risk of being unsheltered. In 2020, for the first time since data collection began, the majority of those who were homeless as individual adults (51 percent) were unsheltered. These circumstances were most likely for those who are chronically homeless, with 66 percent living without any shelter at all.
Note on COVID-19 Impacts (*Updated for 2022). The most recently available nation-wide unsheltered data are from the 2020 Point-in-Time Count. Pandemic-related health concerns disrupted counts of unsheltered people in 2021. While some CoCs made such data available for that year, the nationwide count will not be fully updated until late 2022 or early 2023, leaving a significant hole in available knowledge on homelessness.
Trends in Homelessness
Between 2019 and 2020, homelessness nationwide increased by two percent. This change marked the fourth straight year of incremental population growth. Previously, homelessness had primarily been on the decline, decreasing in eight of the nine years before the current trend began.
Long-term progress has been modest. In 2020, the number of unhoused people was only 10 percent lower than in 2007 (the first year of nationwide data collection).
Uneven Progress. While overall progress on ending homelessness has been modest, there is significant variation among subgroups. Some have experienced striking reductions in their counts.
Veterans are a good example. Currently, 83 communities and 3 states have announced that they ended veteran homelessness (meaning that systems can ensure that homelessness is rare, brief, and one-time). Nationally, veteran homelessness decreased 47 percent since the point at which it peaked in 2009.
Homeless families with children are another group that decreased in size — 27 percent between 2007 and 2020. And, before homelessness began increasing again in 2016, chronic individual homelessness dropped by 35 percent since 2007.
Multiple causes could explain why these populations have experienced periods of greater reductions than the overall homeless population. Some subpopulations have benefitted from greater attention and/or resources after being prioritized by national-, state-, and local-level stakeholders. However, factors external to homeless services systems also contribute to outcomes. Regardless, these subgroups provide proof that significant reductions in homelessness are possible and have occurred.
Other populations have been left behind from this progress — primarily individual adults. The number of individuals experiencing homelessness has remained static over time, decreasing by a mere 1 percent between 2007 and 2020.
Although most veterans and chronically homeless people fall under the umbrella category of “individuals,” the majority of individuals do not belong to one of these subgroups. As a result, individual homeless adults who are not veterans or chronically homeless have typically not been the focus of special attention or resources.
Even more troubling, in recent years, previous and significant gains made to reduce the number of individuals experiencing chronically homelessness has been quickly eroding. As noted above, the size of this group had decreased significantly in the period before 2016. However, between 2016 and 2020, their numbers have surged by 43 percent.
Unsheltered Homelessness on the Rise. Since data on homelessness has been collected, unsheltered homelessness has largely trended downward. By 2015, it had dropped by nearly a third.
However, between 2015 and 2020, there was a reversal of that trend. The unsheltered population surged by 30 percent , almost wiping out nearly a decade of reductions. The number of people currently living unsheltered was virtually as high as it was in 2007.
The rising trend of unsheltered homelessness impacts nearly every major subgroup—including people of every race, ethnicity, gender, and most age groups. Only children (people under 18) have realized an overall decrease in unsheltered homelessness during the surge that was evident at least through 2020.
Sheltered Homelessness on the Decline (*New for 2022). While unsheltered homelessness was on the rise during the period leading up to the pandemic, fewer and fewer people were staying in shelters. The downward trend in shelter usage continued into 2021.
An easy assumption would be that systems are failing—that they’re providing fewer people with shelter, leaving more and more people to sleep outside. However, the reality is much more complicated.
Overall, homeless services systems have actually increased their capacity to serve people. As illustrated in the below visualization (Permanent vs Temporary Bed Inventory Trends, 2007-2021), systems have been steadily growing their available bed numbers. However, they have been increasingly focusing their resources on permanent housing rather than temporary shelter. Thus, more and more people may be benefitting from housing and services, but an increasing share is living in permanent housing as opposed to languishing in temporary shelters. Further, growth in overall bed numbers is likely failing to keep pace with the number of new people entering homelessness, and specifically unsheltered homelessness.
This overall trend predated the pandemic. However, HUD’s AHAR Part 1: Point-in-Time Estimates of the Sheltered Homelessness (February 2022) highlights some COVID-19-related factors that likely contributed to decreases the size of the sheltered population between 2020 and 2021. These include:
- reduced shelter capacity due to social distancing requirements, and
- reduced inflow into shelters tied to eviction moratoria and stepped-up investments in homelessness prevention and diversion.
Ending homeless is an ongoing challenge throughout America. However, the severity of the challenge varies by state and community. Locating the areas experiencing the most significant challenges and directing additional attention – including new resources – towards them could result in meaningful reductions in homelessness. There are two ways to evaluate geographic variations—counts and rates.
Counts. Examining the jurisdictions with the largest homeless populations is informative. Many also have the highest populations, overall. For example, California is the most populous state in the union and also has the largest number of people experiencing homelessness. Similarly, the Continuums of Care (CoC) with the largest homeless populations include highly populous major cities (e.g., New York City, Los Angeles, and Seattle) and Balance of State CoCs encompassing numerous towns and cities.
Fifty-seven percent of people experiencing homelessness in 2020 were in five states (California, New York, Florida, Texas, and Washington ). Half were in the top twenty-five CoCs identified in the State & CoC Ranking 2020 chart. Thus, a significant share of this national challenge is in a small number of places with large homeless counts. Meanwhile, most communities have relatively small homeless populations to serve. This should impact how the problem is addressed.
Rates. Homeless counts are just one approach to understanding the nature of homelessness. Putting them into context adds nuance to the story. For example, suppose that 100,000 people were to experience homelessness in California (a state with more than 39 million people). Those would be far less challenging circumstances than 100,000 people being homeless in Wyoming (a state with roughly 575,000 people). Thus, it is helpful to consider the homeless population in relation to the general population.
Rates of homelessness vary widely across the country. For example, in 2020, the northeast Oklahoma CoC had the lowest rate in the country, reporting one person experiencing homelessness out of every 10,000 people. Meanwhile, the Humboldt County CoC in California had the highest rate of 126 people being homeless out of every 10,000.
Many of the states and CoCs with the highest rates of homelessness have the highest housing costs. For example, San Francisco had the fourth highest rate of homelessness in the country; and it has the nation’s highest housing wage (i.e., the hourly wage a full-time worker must earn to afford a modest home at HUD’s fair market rent ). Low-income people in such jurisdictions find it difficult to secure and keep housing they can afford, which directly.
Other jurisdictions with high rates of homelessness have high rates of poverty. For example, CoCs like Humboldt and Imperial City in California are listed above among the ten CoCs with the highest rates of homelessness in the country. They also have high poverty rates, exceeding 20 percent of their overall populations. Such jurisdictions have relatively low housing costs but have many people experiencing economic hardships, some resulting in homelessness.
The dashboard at the top of this page and the above rankings chart are helpful in making in-depth comparisons among states and CoCs. This allows jurisdictions to evaluate the severity of their challenges.
Homeless Assistance in America (*Updated for 2022)
The nation’s homeless services systems do not have enough resources to fully meet the needs of everyone experiencing homelessness. Thus, it is helpful to examine the difficult decisions they must make, including how much of their limited funds should be spent on temporary versus permanent housing.
Temporary Housing. Following two years of decreases, there was an uptick of nearly 7,000 temporary shelter beds between 2019 and 2021 . Currently, the overall number of temporary beds is 8 percent lower than the all-time high count which occurred in 2011.
Historically, America has not had enough shelter beds for everyone experiencing homelessness. Individual community circumstances vary. However, in examining national-level bed and population counts for 2020, systems only had enough year-round beds for 50 percent of individuals on the night of the PiT Count. Availability for families is far different. Collectively, the nation’s communities had enough shelter beds for nearly 100 percent of families experiencing homelessness throughout America (with a surplus of nearly 18,000 beds).iv
During the winter months, some communities temporarily supplement these year-round beds with seasonal ones. Thus, they may be able to serve more people during that time of the year. But, unfortunately, many people are unsheltered, sleeping on sidewalks, in abandoned buildings, or in other locations not meant for human habitation. This typically impacts individual adults, but some families with children are also in these situations.
Permanent Housing. CoCs have had recent years in which temporary housing offerings were on the decline while investments in permanent housing beds (Permanent Supportive Housing, Rapid Re-Housing, and Other) have been consistently increasing. Over just the last five years, these types of beds grew by 25 percent . Forty-five states and the District of Columbia have contributed to this trend.
These numbers reflect a shift in policy and funding priorities. The current era reflects a renewed emphasis on Housing First, or connecting people with permanent housing as quickly as possible. Currently, 59 percent of all homeless system beds are designated for permanent housing.
Common Forms of Assistance. Nationally, the most common forms of homeless assistance are Permanent Supportive Housing (39 percent of system beds) and emergency shelter (32 percent of system beds).
Over the last five years, the fastest growing forms of assistance have been Rapid Re-Housing (which is still relatively new, emerging as an eligible use of CoC program funds in 2012) and “Other Permanent Housing” (permanent housing other than Rapid Re-Housing or Permanent Supportive Housing). During that time period, the former expanded by 81 percent and the latter by 43 percent.
Only one type of intervention has been on the decline—transitional housing. There are 59 percent fewer beds in this category than there were in 2007. This shift is responsible for decreases in the overall availability of temporary housing in recent years. It further reflects the policy goal of moving more people into permanent housing as quickly as possible.
Indicators of Risk (*Updated for 2022)
In the lead up to the pandemic, the nationwide poverty rate had decreased for five consecutive years. In 2020, that streak ended and the number of people living in poverty spiked by approximately 3.3 million people. Overall, nearly 37.2 million people or 11.4 percent of the U.S. population were pushed into this group. Certain racial groups have even higher rates of poverty, including Black people (19.5 percent) and Hispanics/Latinos (17 percent). People living in poverty struggle to afford necessities such as housing.
In 2020, approximately 6 million Americans households experienced severe housing cost burden, which means they spent more than 50 percent of their income on housing. The overall size of this group had been gradually decreasing since 2014. However, the number of severely cost-burdened American households is still 6 percent higher than it was in 2007, the year the nation began monitoring homelessness data. And, more troublesome patterns may exist for notable subpopulations, including the lowest income people and female-headed households.
“Doubling up” (or sharing the housing of others for economic reasons) is another measure of housing hardship. An estimated 3.7 million people were in these situations. Some doubled-up individuals and families have fragile relationships with their hosts or face other challenges in the home, putting them at risk of homelessness. In 2020, the size of this group experienced an uptick after six years of decline. Currently, the doubled-up population size is 5 percent larger than it was in 2007.
Over a period lasting more than a decade, the nation has not made any real progress in reducing the number of Americans at risk of homelessness. In fact, these challenges are slightly worse. The trend lines in the above chart point to severe house cost and doubled-up numbers that are higher in 2020 than they were in 2007. Even more troubling, available data was generated amid the constantly evolving crises tied to the pandemic and other factors. For instance, the ultimate impacts of expiring eviction moratoria, fading Emergency Rental Assistance dollars, and 2022 rental cost inflation are unclear.
National-level data, which has been discouraging, can mask even more dire challenges in specific areas of the country. For example, since 2007, severe-housing-cost burdened households grew by 45 percent in Wyoming and 34 percent in Connecticut (numbers that are even higher than national-level population growth). Similarly, over that same time period, the number of people doubled up expanded by 136 percent in Nevada and 98 percent in Hawaii.
Sources and Methodology
Data on homelessness are based on annual point-in-time (PIT) counts conducted by Continuums of Care (CoCs) to estimate the number of people experiencing homelessness on a given night. The latest full counts (sheltered and unsheltered) are from January 2020. National-level sheltered-only data is available for 2021 (along with unsheltered data for about 40 percent of CoCs). Point-in-time data from 2007 to 2021 are available on HUD Exchange.
Rates of homelessness compare point-in-time counts to state, county, and city population data from the Census Bureau’s Population Estimates Program (Population and Housing Unit Estimates data tables, 2020 version). Rates for racial, ethnic, and gender demographic groups are drawn from the Census Bureau’s American Community Survey 5-year Data (2020 version).
Data on homeless assistance, or bed capacity of homeless services programs on a given night, are reported annually by CoCs along with point-in-time counts. These data are compiled in the Housing Inventory Count (HIC), which is also available on HUD Exchange for 2007 through 2021.
Data on at-risk populations are from analyses by the National Alliance to End Homelessness of the Census Bureau’s 2020 American Community Survey 1-year Estimates. Poor renter households with a severe housing cost burden are households whose total income falls under the applicable poverty threshold and who are paying 50 percent or more of total household income to housing rent. For people living doubled up, poverty is based on the composition and income of the entire household as compared to the poverty thresholds. A person is considered living doubled up based on his or her relationship to the head of household and includes: an adult child (18 years old or older) who is not in school, is married, and/or has children; a sibling; a parent or parent-in-law; an adult grandchild who is not in school; a grandchild who is a member of a subfamily; a son- or daughter-in-law; another relative; or any non-relative.
i Much of the data in this report is derived from the Point-in-Time Count published by U.S. Department of Housing and Urban Development. The agency publishes data on people served in shelter and other forms of homeless assistance housing over the course of a year—that information is not reflected in this report.
ii This report includes data on various racial and gender groupings that are a part of the Point-in-Time data collection process. HUD does not require data on other marginalized groups such as people with disabilities, older adults, or members of the LGBTQ community (other than people who identify as transgender or gender non-conforming).
iii The Pacific Islander and Native American groups are relatively small when compared to populations such as whites and Hispanics/Latinxs. This is one of the factors that makes them more difficult for homeless services systems and the Census to count them. There is a need to ensure that data collection efforts focused on these groups becomes more precise. However, available data suggests significant disparities and causes of concern that are worthy of discussion. See USICH, Expert Panel on Homelessness among American Indians, Alaska Natives, and Native Hawaiians (2012) and Oversight Hearing on Reaching Hard-to-Count Communities in the 2020 Census, 116th Congress (2020)(testimony of Kevin J. Allis).
iv Surpluses in family beds are partially tied to family housing being organized in units. For instance, each family may be assigned to a unit with four beds. If only two people are in the family, two beds will go unused. Some underutilization can’t be avoided but some systems may need to revisit how they organize family units and plan for the amount of space families will need.