Los Angeles has been celebrating small signs of progress on homelessness, but new research suggests its numbers are flawed. For the past two years, the city’s official Point-in-Time Count has shown a modest decline in the number of people living outdoors, a reversal after a decade of relentless increases. But data from RAND suggests that the apparent turnaround may be partly an illusion. The research group warns that the city’s annual homelessness census is increasingly missing people who are still living on the street, giving the impression that the crisis is easing faster than it really is and hiding growing need in the neighborhoods that have long borne the brunt of the problem.
According to the latest findings from RAND’s Los Angeles Longitudinal Enumeration and Demographic Survey, or LA LEADS, Los Angeles Homeless Services Authority’s Point-in-Time Count is now underestimating unsheltered homelessness by nearly one-third in the areas most affected. RAND’s professional field teams have been conducting independent counts in Hollywood, Skid Row, and Venice since 2021, visiting the same streets every few months using trained enumerators who check and cross-verify their results. For the first two years their data matched LAHSA’s volunteer-based numbers almost exactly. By 2024 the gap had widened to 26 percent. By 2025 the discrepancy had grown to 32 percent.
If that level of undercount holds citywide, RAND estimates that roughly 7,900 people and dwellings were missing from this year’s official total. That is more than the entire unsheltered population of Orange County. The stakes are not just statistical. The Point-in-Time Count determines how more than 300 million dollars in federal and county funding is distributed each year through the Continuum of Care and the new Measure A program. Undercounting in high-need neighborhoods like Venice and Skid Row could mean millions of dollars in lost funding for outreach, shelter, and healthcare.
RAND’s latest analysis breaks new ground because it connects the counting problem to a deeper structural shift in how homelessness looks on the street. In the past two years, the number of people living in tents or vehicles has fallen, while the number of people sleeping directly on sidewalks or in doorways has sharply increased. The 2024 LA LEADS annual report, released in July, found that this most vulnerable group, known as “rough sleepers,” now represents about 40 percent of all unsheltered residents in the areas studied. In Hollywood and Venice the number of rough sleepers more than doubled even as the total number of people outdoors declined.
That change, RAND now says, helps explain why LAHSA’s official counts have become less reliable. The volunteers who conduct the annual Point-in-Time survey rely on visual tallies of tents, vehicles, and visible individuals. That system worked when most unhoused people were clustered in large encampments. As more people live without tents or vehicles, it becomes much harder to see and count them. RAND’s data show a clear pattern. Every year that the proportion of rough sleepers rises, the accuracy of the Point-in-Time Count falls.
The undercount is not even across neighborhoods. In 2025 the official total for Hollywood represented about 81 percent of RAND’s number, Venice about 76 percent, and Skid Row just 61 percent. The deeper the poverty and the more visible the crisis, the less accurate the city’s census. RAND warns that this unevenness could distort the flow of federal and local dollars, with some cities receiving up to 30 percent more funding per unhoused person than others.
Ironically, RAND notes, part of this problem may stem from the very policies that have driven recent progress. The city’s Inside Safe initiative, which relocates people from encampments to motel rooms, has been credited with reducing tents across the city. But by removing the easiest-to-count portion of the population, Inside Safe may also be making it harder to measure who remains outside. The more visible the success, the less visible the remaining crisis becomes.
RAND’s earlier reports foreshadowed this warning. The 2024 annual LA LEADS report showed that while unsheltered homelessness fell by 15 percent across Hollywood, Skid Row, and Venice, the overall well-being of those still on the street declined. People were moving more frequently, reporting worse health, and losing contact with outreach teams. The report found that 91 percent of unsheltered residents still wanted to be housed, but only 38 percent were on a housing waitlist and only 13 percent had ever been offered supportive housing.
In Venice, RAND’s enumerators recorded 554 people living outdoors in 2024, while LAHSA’s volunteers counted only 173 in the same area. At the time, researchers described the gap as concerning but isolated. The new data confirm that it was part of a broader trend.
The findings mirror what outreach workers have been describing for months. People pushed from RVs and encampments by enforcement or sanitation sweeps do not disappear. They scatter. Some move to less visible parts of the neighborhood, sleeping behind warehouses or near freeway embankments. Others drift between temporary motel placements and short-term safe parking sites. The visible tent decline that city officials cite as progress often corresponds to a population that is simply harder to find. For a volunteer counting cars and tents on a single night each January, those missing people are easy to overlook.
The result is a widening gap between perception and reality. To the public, smaller encampments and cleaner sidewalks look like success. To service providers, the work is becoming harder. Outreach teams must spend more time locating clients who no longer stay in the same place, while those clients often face worsening physical and mental health after losing the minimal protection of a tent or vehicle. RAND warns that if policymakers believe homelessness is falling faster than it really is, they may slow the expansion of housing programs just as the need for them is growing.
RAND also makes a rare, concrete recommendation: supplement the volunteer Point-in-Time Count with professional enumerators who can cross-check results in key neighborhoods during the same week. The change would cost little, could be funded through Measure A, and would sharply improve accuracy and public confidence. RAND calls it a “straightforward fix” in a field where simple solutions are rare.
The timing of RAND’s critique is sensitive. Mayor Karen Bass and other city leaders have pointed to consecutive declines in the Point-in-Time Count as proof that new strategies are working. Those results have helped justify Inside Safe’s expansion and strengthen arguments for continued emergency authority over city housing policy. RAND’s new report complicates that story. The data suggest that progress on paper is outpacing progress on the ground, and that the places most in need of resources are at risk of being shortchanged because they are also the hardest to count.
In Venice, the pattern is particularly striking. Between 2023 and 2024, RAND recorded a 22 percent drop in the number of people living outdoors, but almost none of that reduction came from housing placements. The decline was confined to vehicle dwellers. Tent numbers held steady and rough sleeping doubled. Venice also saw almost no Inside Safe operations during that period, suggesting that enforcement and displacement were the main forces at work. The city’s official count has consistently failed to capture the full population of people living outdoors in Venice. RAND’s new analysis shows that the undercount continued in 2025, with LAHSA counting roughly three-quarters of RAND’s total. If that gap persists, funding formulas based on those numbers will continue to penalize communities that already have the fewest services and the most need.
RAND’s broader conclusion is that Los Angeles has made progress, but it is fragile and uneven. Some people have been moved indoors, but many others have been rendered invisible. The people who remain outside are sicker, older, and harder to reach. They no longer live in the large encampments that draw attention from housed neighbors, reporters or politicians. They are scattered, anonymous, and missing from the data that drives public understanding.