Homelessness Fell in 2025 for the First Time Since 2016. Here's the Real Data.

The first real decline in nearly a decade, after a sharp recent surge

HUD's official 2025 Annual Homelessness Assessment Report found 745,652 people were experiencing homelessness on a single night in January 2025, a 3.3% decrease from 2024's record count of 771,480, and the first annual reduction since 2016. This followed two consecutive years of sharp increases: homelessness rose 18% from 2023 to 2024 alone, according to National Alliance to End Homelessness analysis of the same HUD data.

Independent preliminary tracking from Community Solutions, drawing on 170 local communities that released their own 2025 counts ahead of the official federal report, found a similar pattern: roughly a 2% national decline, with unsheltered homelessness specifically down about 3%, following back-to-back increases of approximately 15% (2022-2023) and 19% (2023-2024) in those same communities.

The longer-term trend remains genuinely alarming, even with this recent dip

Despite the 2025 decline, homelessness has risen 27% since 2013 nationally. More specific categories show even sharper long-term increases: unsheltered homelessness specifically rose 36% over that same period, and chronic homelessness, generally defined as long-term or repeated homelessness combined with a disabling condition, rose a striking 81%. This matters for interpreting the 2025 numbers correctly: a single year's modest decline doesn't offset a decade-plus trend of substantial growth, particularly in the most severe, hardest-to-address categories.

System capacity remains a fundamental, measurable constraint

According to National Alliance to End Homelessness analysis, no community in the country had enough permanent housing to serve everyone experiencing homelessness as of 2024, and the homeless response system nationally only had enough units to house an estimated 16% of households currently staying in shelter over the course of that year. Separately, an estimated 17,500 people per week entered homeless systems for the first time during 2024, illustrating that the flow of new people becoming homeless remains substantial even as total counts modestly declined the following year.

What's driving both the improvement and the ongoing debate over methodology

• HUD Secretary Scott Turner specifically attributed part of the decline to reduced numbers in "Sanctuary Cities," directly linking the homelessness decrease to immigration enforcement policy, a characterization not universally accepted by homelessness researchers or advocacy organizations
• California and New York, two of the states that had seen among the largest increases in prior years, posted meaningful year-over-year reductions in 2025 specifically, -2.8% and -7.9% respectively
• Point-in-Time count methodology itself remains a genuine, ongoing source of measurement uncertainty; Allegheny County, Pennsylvania specifically found its initial January 2025 unsheltered count (up 66% year-over-year) didn't match real-time street outreach data, prompting the county to conduct and publish a second, more carefully documented count in March that still showed a substantial, if smaller, increase
• HUD Secretary Turner has framed the data as evidence that "Housing First" policy, the dominant national approach prioritizing getting people into permanent housing quickly without prerequisite conditions like sobriety, has "failed to meaningfully reduce homelessness," a characterization housing policy researchers broadly dispute, pointing to the specific 2025 decline itself, and the fact it followed continued Housing First-oriented programming, as evidence against that framing

The core disagreement

One side of this debate generally argues Housing First and expanded federal housing assistance are still the right approach, and interprets the modest 2025 decline as early validation that continued investment works, arguing more resources, not a policy pivot, are what's actually needed given system capacity still serves only a fraction of people in shelter. The other side, reflected in current HUD leadership's framing, argues the decade-long 27% increase despite substantial Housing First-oriented spending shows the current approach has fundamentally failed, and favors conditioning assistance more heavily on factors like sobriety, work requirements, or immigration status, with the recent decline cited as evidence a policy shift, not more of the same approach, is producing results. Given genuine, unresolved measurement challenges at the local level (as Allegheny County's methodology review illustrates), both interpretations are working from data that even researchers acknowledge carries real uncertainty.