RAND research suggests Los Angeles’ homelessness problem is changing form as tents decline, rough sleeping rises and more of the remaining tent population becomes concentrated in Skid Row.
Los Angeles spent the past two years reporting declines in its official homeless count. But new research suggests that the visible improvement may tell only part of the story.
In three Los Angeles neighborhoods tracked by RAND, Hollywood, Skid Row and Venice, the number of people experiencing unsheltered homelessness remained statistically flat through January 2026 even as the way people survived on the street changed dramatically.
Tents declined. Rough sleeping and vehicle dwelling increased. And by January 2026, nearly 90% of the tents remaining across the three study areas were concentrated in Skid Row, compared with 60% four years earlier.
The findings raise a practical question for Los Angeles:
What happens when homelessness becomes harder to see, harder to count and potentially harder to reach?
RAND’s five-year Los Angeles Longitudinal Enumeration and Demographic Survey, or LA LEADS, followed unsheltered homelessness in Hollywood, Skid Row and Venice using professional field staff who systematically counted people every two months.
The study found that tent dwelling across the three neighborhoods fell by roughly half since 2021, including a 23% drop in 2025.
But the remaining tents became increasingly concentrated in Skid Row.
By January 2026, nearly 90% of the tents remaining in the three study neighborhoods were in Skid Row, up from 60% four years earlier. Skid Row was also the only one of the three neighborhoods to experience continuous growth in its homeless population during the study period.
That does not establish that people whose tents disappeared from Hollywood or Venice physically moved to Skid Row. It does show that as tent dwelling declined elsewhere, the remaining tent population became increasingly concentrated there.
The concentration matters because Skid Row remained the neighborhood with the greatest concentration of need among the three areas RAND studied.
Parriva has previously examined the legal and policy pressures surrounding the neighborhood in Court in Session on Skid Row, but the RAND findings add another dimension: the changing form of homelessness may be making the population itself harder to measure.
The most significant change may not be where people are. It is how they are living.
RAND found that rough sleeping, sleeping without a tent, vehicle or other shelter, increased 20% in 2025, reaching its highest level in four years.
Vehicle dwelling increased 11%.
By January 2026, rough sleeping accounted for about 44% of the unsheltered population in the three study neighborhoods, compared with roughly 30% in 2021 and 2022.
That creates a fundamental measurement problem.
A tent is visible.
A person sleeping under a bench, behind a bus shelter, in an alley or wrapped in a dark blanket can be much harder to identify.
A vehicle presents a different problem: someone sleeping inside a car can be difficult to distinguish from someone who simply parked there.
RAND’s research found that the population has increasingly shifted toward precisely those less visible forms of homelessness.
What happens when a tent disappears?
RAND’s survey data also suggest that encampment clearings may be contributing to the shift from tent dwelling to rough sleeping.
Nearly half of rough sleepers surveyed said they had lost a dwelling during the previous year. Of those, 46% said the dwelling had been confiscated or towed by government officials or service providers.
RAND says those findings suggest some people may move from tent dwelling to rough sleeping following encampment clearings.
That does not prove that every encampment clearing produces a rough sleeper, or that sweeps are responsible for the overall increase.
But it points to a consequence that can be missed if success is measured primarily by what disappears from sidewalks and parks.
The absence of a tent is not the same thing as the end of homelessness.
RAND researchers say rough sleepers can also be harder to engage and often have greater clinical needs, making the changing composition of street homelessness important for outreach and health services.
The official count is telling a different story
Los Angeles’ official numbers have not actually continued declining in 2026.
The Los Angeles Homeless Services Authority’s 2026 results reported a 1.2% increase in the countywide homeless count and a 3.4% increase in the City of Los Angeles.
LAHSA said the overall increases were not statistically significant.
For unsheltered homelessness, LAHSA reported increases of 3.3% countywide and 7.9% in the city.
That matters because the RAND research is not showing that the official count has simply missed an otherwise dramatic decline.
Instead, the two systems are producing different pictures of a population whose composition is changing.
RAND’s professional enumeration closely matched LAHSA’s official numbers in 2022 and 2023. The gap then widened, reaching roughly one-third in RAND’s 2025–26 comparison.
In Skid Row, RAND researcher Sarah Hunter estimated that the LAHSA count missed 39% of the population RAND’s professional enumeration identified.
That figure should be understood as a comparison between two methodologies; not as proof that LAHSA’s number is simply “wrong.”
Why the methods produce different results
LAHSA’s Point-in-Time Count uses teams surveying assigned areas during the annual count. LAHSA has also made methodological changes, including digital data collection, improved mapping and additional staffing and support for difficult-to-access areas. LAHSA’s homelessness-count information
RAND’s LA LEADS study used professional field staff to conduct repeated, systematic enumeration of the three study neighborhoods throughout the year. It concluded in January 2026 and represented the largest ongoing professional count of unsheltered people in Los Angeles outside LAHSA’s annual Point-in-Time Count. RAND’s methodology and analysis
The issue RAND identifies is not simply that one method counts tents and the other does not.
It is that the composition of homelessness has changed.
When more people are sleeping without tents or living in vehicles, the visual indicators that make homelessness easier to identify become less reliable.
RAND argues that this changing form of homelessness may help explain why its professional enumeration has increasingly diverged from the official count.
The accuracy of homelessness counts can matter beyond public perception.
Los Angeles County’s Measure A Local Solutions Fund uses a formula based on the Point-in-Time Count and similar measures of people experiencing homelessness.
For fiscal year 2026–27, the Los Angeles County Measure A funding framework says the formula is 90% based on a three-year rolling average of the Point-in-Time Count and 10% based on Census data. The fund is projected to distribute approximately $96.97 million annually.
That creates an important policy question.
If the population becomes harder to see and a systematic undercount develops, could that eventually affect how resources are distributed geographically?
The available evidence does not establish a specific dollar amount that Skid Row or another neighborhood has lost because of RAND’s estimated undercount.
What is documented is the connection between homelessness-count data and the County’s funding formula.
That makes the accuracy of the underlying data a practical issue, not simply a statistical argument.
The RAND research leaves several questions that matter for the next stage of the homelessness response:
How many people moved from tents to rough sleeping?
How many moved into vehicles?
How many were successfully housed?
How many moved outside the three study areas?
How many were offered shelter or housing?
How many remain in places that traditional counting methods have difficulty finding?
Those questions are more useful than simply asking whether the number of tents went up or down.
They also point to a broader issue for Los Angeles: whether homelessness policy is keeping pace with the changing population it is supposed to serve.
RAND recommends tailoring homelessness interventions to neighborhood conditions and to the needs of people sleeping in different ways, rather than relying on one encampment-focused approach for every population.
The measurement problem is becoming a policy problem
Los Angeles does not have to choose between believing the official count and believing RAND.
The two systems are measuring homelessness differently.
What the RAND research shows is that the difference between them has grown—and that the gap is particularly significant in Skid Row, where homelessness is heavily concentrated.
That deserves attention because a count is not merely a number.
It helps government understand the size and location of a problem, plan outreach and services, and make funding decisions.
And when the form of homelessness changes, the measurement system has to change with it.
For Los Angeles residents trying to understand what is happening on the streets, the central lesson is straightforward:
A tent disappearing from a sidewalk is an observable event. The disappearance of homelessness is much harder to measure.
And in Skid Row, the people who remain may increasingly be the people the traditional system has the hardest time seeing.
