Who is homeless in the United States? A 2025 update

This is the third installment of our annual look at what the data say about who is homeless in America. Our 2024 analysis is here and our 2023 analysis is here.

Homelessness data in the United States come from annual point-in-time (PIT) counts, where local volunteers spread out to count the sheltered and unsheltered population experiencing homelessness on a single night in January. In 2024, there were approximately 771,400 homeless individuals in the country, an increase of 118,300. The PIT data provide a window to an understanding of who is homeless in America, even though they are not able to capture everyone who lacks adequate shelter. (See the end of this article for the definitions the Department of Housing and Urban Development uses in the PIT counts.)

Some groups are overrepresented in the homeless population compared with their share of the total U.S. population. Our analysis finds that Americans who identify as Hispanic, Black, Native Hawaiian or Pacific Islander, Native American, or male have a greater presence in the homeless population than in the overall population. Americans who identify as Asian, White, or female and those under 18 have shares in the homeless population that are lower than their share in the overall population (Figure 1).

This is our third installment looking at national PIT counts, and Figure 2 shows how homelessness for different groups has changed in those three years. In all, homelessness has increased from 1.75 per 1,000 people in 2022 to 2.3 per 1,000 people in 2024, a 30 percent increase. There are many potential contributing factors. One possibility, research suggests, is a connection between rising rent prices and rising homelessness. Of course, many goods and services experienced inflation above 2 percent starting in 2021, rental prices included. But inflation in the housing sector has proved particularly sticky, coming in at an annual rate of 4.8 percent in December. Another notable trend is the decline in rental vacancies, which have been falling since 2009. The HUD report identified two other factors as well, pointing out that many of the municipalities with the largest increases in homelessness “reported that they experienced an increase from people who were displaced from natural disasters or due to immigration.”

Figure 2 includes two bars for 2024 because the Department of Housing and Urban Development (HUD), which tabulates the PIT counts, changed its reporting in 2024 to break down race categories by Hispanic and non-Hispanic ethnicity. Because the U.S. population counts also break down race categories by Hispanic and non-Hispanic ethnicity, the new method likely offers a more faithful representation of the racial and ethnic self-identification of the homeless population. We include the data for 2024 using the previous methodology to provide a comparison with the data from 2022 and 2023. In particular, the old method shows a decline in the homeless population among Hispanic Americans, while the new method shows a jump. This is a result of the change to methodology.

For demographic characteristics other than race and ethnicity, the methodology does not change the estimates. Figure 2 shows that the share of children under the age of 18 who are experiencing homelessness increased 33 percent between 2023 and 2024, the largest jump of any age group. The PIT data counted approximately 150,000 children experiencing homelessness. A similar number of Americans aged 55 or more were also experiencing homelessness.

Having a roof over your head depends on where you live

The experience of being homeless varies by place. In Los Angeles, most homeless people live outside—a quite different experience than in New York City, where a legal “right to shelter” has almost every homeless person under a roof. The rest of the country is in between (see Figure 3).

The share of the homeless population that is unsheltered declined between 2023 and 2024 in each of the four place categories examined in Figure 3. The biggest decline occurred in major cities other than NYC and LA, where the unsheltered share fell by 5 percentage points.

This change occurred against a backdrop of considerable debate over how municipalities could and should treat homeless populations. In 2024, a city in Oregon sought the right to treat sleeping outside as “unsanctioned camping” by issuing tickets and making arrests. In June, the Supreme Court decided that such penalties do not constitute “cruel and unusual punishment” even when shelter beds aren’t available.

The causes of homelessness are complex, and so are the remedies. The PIT data describe who is homeless in a basic sense but lack important elements that could help explain why some people end up homeless. This is harder information to know, but it matters for how different solutions might help different groups of people. A 2024 Institute working paper by Institute advisor Ayşe İmrohoroğlu and Kai Zhao looks to address that gap using a sophisticated theoretical model that combines income, health, and housing. In the model, homelessness can happen when someone experiences a negative event affecting their income or health that leaves them unable to afford the cheapest rental unit.

The economists use the model to compare the effectiveness of different policies in reducing homelessness in the long run. They find that existing voucher programs do help reduce homelessness. To reduce homelessness further, their results suggest it is more effective to increase the number of people receiving vouchers than to increase rent subsidies or offer cash transfers.

Continuing to invest in efforts to take stock of homelessness is essential to understand how policies affect the housing security of different groups of Americans.