Sarah B. Hunter is a senior behavioral scientist at RAND, where she worked on the LA LEADS research. Until January 2026, she co-directed the RAND Center on Housing and Homelessness.
The release in late July of the Los Angeles Homeless Services Authority’s 2026 point-in-time count has reignited public debate over the trajectory of the city’s homelessness crisis. These official figures show a 7.9% increase in unsheltered people — those living in tents, cars or on the street — and a 3.4% rise within the city limits overall, which includes those in shelters and temporary housing.
For Angelenos trying to make sense of the problem, the data will be contradictory to what they have been seeing and reading in the news. Initiatives such as Inside Safe have visibly cleared large tent encampments from sidewalks and parks, and local leaders have promoted these successes. How, then, are the numbers going up?
Golden State is a reader-supported publication. No billionaires tell us what to do. Help us to stay independent with a tax-deductible donation.
The answer lies in the way street homelessness has changed in the last few years, which has left the methodology used by the Homeless Services Authority struggling to keep pace. The LAHSA point-in-time census has been undercounting for a few years, and there are reasons to think that even with this uptick, the 2026 result is still not reflecting the full scale or the true distribution of the homelessness crisis.
The changing shape of the crisis
At RAND, where I’m a researcher, we conducted a five-year study of the three neighborhoods that have been epicenters of the homelessness crisis: Skid Row, Hollywood and Venice. Called the Los Angeles Longitudinal Enumeration and Demographic Survey, or LA LEADS, it used only professional canvassers. These trained field staff systematically walked and drove every block of these neighborhoods every two months to count every person living unsheltered.
We also used a strict quality control process: Two people covered each block independently. We then compared their totals, and if their findings differed by more than 10%, the block got recounted. This rigorous approach captured a level of detail that is not possible with an annual, volunteer-based point-in-time count.
The RAND data revealed, for example, that clearing tent encampments does not automatically produce a net reduction in homelessness. Instead, it often alters how people survive on the street.
Mariel Garza
In the three neighborhoods we tracked, tent-dwellings fell by half between 2021 and the end of last year. But those reductions were offset by more people living in cars, vans and RVs — as well as many more people living literally unsheltered on sidewalks, in alleys or in alcoves, commonly referred to as “sleeping rough.” Tents and makeshift structures are the only form of homelessness that is shrinking.
For every four tents removed, our May 2026 update reported, approximately three rough sleepers or vehicle-dwellers were added to the surrounding area. Rough sleeping has now become the most common form of homelessness in Hollywood, Skid Row and Venice.
Why the official count underestimates the problem
This shift presents a challenge for the LAHSA point-in-time count, which uses thousands of volunteers who visually survey census tracts over a few nights. This methodology relies on easy-to-see indicators of homelessness, such as tents. An individual wrapped in a dark blanket, tucked away for safety in a bush, under a bench or behind a bus shelter, is much more difficult to spot. Likewise, volunteers frequently struggle to accurately distinguish vehicles that someone is living in from those simply parked in a neighborhood.
By removing the easiest-to-count subpopulation — tent dwellers — while missing the rise in hidden rough sleepers and vehicles, the LAHSA count metrics have obscured the true scale of the crisis. Yes, the numbers rose this year, which is dispiriting. Worse is the fact that they are still likely a profound undercount.
Those of us working on LA LEADS noticed the undercount two years ago. The point-in-time count had closely matched our numbers in 2022 and 2023, but a discrepancy of about 26% emerged in 2024. By 2025 and 2026, the gap was one-third. The undercount is most severe in Skid Row, where it missed 39% of the true homeless population.
Karin Klein
This has tangible consequences. Public funding — including Measure A sales tax revenue— is distributed geographically based on the LAHSA count findings. Resources are being diverted away from the very neighborhoods that require the most intensive healthcare, safe parking and housing infrastructure.
An intergovernmental blame game
The worsening numbers have also strained the intergovernmental relationships required to address the crisis. LAHSA is a joint city-county authority. The mayor and the Los Angeles City Council appoint half of its governing board, and the five members of the Los Angeles County Board of Supervisors appoint one member each. It’s impossible for either city or the county to escape blame for the agency’s structural deficiencies as captured in a highly critical audit earlier this year.
Assigning accountability for the rise in homelessness is also messy because of city-county overlapping jurisdictions. The city manages land use, local law enforcement and programs such as Inside Safe. However, it’s the county of Los Angeles that administers the clinical treatments (including for physical health, mental health and substance use) and the social services that are vital to getting people to move indoors and remain there.
The absence of a tent on a sidewalk does not equate to a person in permanent or transitional housing. To build a sustainable strategy for reducing homelessness, LAHSA must fix its baseline metrics. Using professional field teams to independently cross-check volunteer counts would significantly improve the accuracy of homelessness data. That would help Los Angeles target efforts and funding on the complex health needs of rough sleepers, the specific ways to help those living in vehicles, and the neighborhoods where both are on the rise.
