Mainstream reporting on SF’s homeless issue by and large indicates the majority of the city’s homeless are San Franciscans. For example, in May, The Standard wrote that “71 percent of [homeless people] surveyed reported living in San Francisco, 24 percent in other California counties, and four percent outside of California.” In June, Vox reported that “the vast majority of those homeless in California (nine out of 10) had been living in the state before losing their homes — bucking the idea that maybe people are flocking to the sunny West Coast to live outside in the nicer weather.” More recently, city supervisor Dean Preston told UnHerd that “the biggest driver of why folks are on the street [in San Francisco] is because they lost their jobs, income, or were evicted from their homes.” 

Mainstream reporting on where exactly SF’s — and California’s — homeless are from is based on two canonical sources: surveys from the Department of Homelessness and Supportive Housing (HSH), and a study of 3,200 homeless people in California, conducted by the Benioff Homelessness and Housing Initiative at UCSF. However, both suffer from flawed methodology that renders their data completely unreliable — though even this unreliable data still suggests most homeless in San Francisco aren’t from the city, and that most homeless in California are from out of state.

Every two years, over the course of a single night, HSH volunteers walk around the city, counting the number of people they see sleeping in cars and on the street. This is called a “Point in Time” (PIT) survey, and it’s a federally mandated requirement for receiving federal funding for homeless services. During the weeks following the PIT headcount, homeless people are surveyed “in order to profile their experiences and characteristics” (the most recent of which surveyed 768 of them). These two evaluations form the basis for the data the HSH publishes in its “comprehensive reports.”

The PIT and self-reported survey methodologies severely restrict the range of conclusions you can draw from their data. PIT doesn’t control for the fact that the number of homeless people sleeping in publicly visible spaces fluctuates nightly, nor does the HSH even try to contend with the fact that demographic and place-of-origin data are never independently verified or corroborated. Further, it ignores the unique incentives at play in San Francisco: if a homeless survey respondent says he’s not from SF, he may believe that would put him at risk of being denied the direct cash transfers, free housing, and other homelessness services the city offers. This is not even to mention the reports emerging about homeless people coached by activists to lie about where they’re from.

But even the HSH’s flimsy methodology is unable to hide the damning facts their data — however unreliable — implies. A quote from Pirate Wires’ August article, “San Francisco’s Homeless Ticking Time Bomb” (emphasis added):

In 2022 [HSH] reported that 71 percent of people on San Francisco’s streets were living in the city when they became homeless. According to HSH methodology, “place of residence” can include living with friends, family, partner, or even a motel. This means that if someone stays with friends or family in SF for a month (or even a night) before winding up homeless on the city’s streets, the HSH [can] count them as ‘from San Francisco.’ On top of this, the Department classifies these “San Franciscans” in only three ways — those living in the city for less than a year prior to becoming homeless (17 percent, according to the latest data), between one and 10 years (52 percent) and for 10+ years (31 percent). But a duration window as expansive as between one and 10 years makes it impossible to know whether this 52 percent figure includes more people closer to the one- or ten-year ends of the residency spectrum.

In other words, the HSH’s data directly states that almost a third (2,248) of SF’s homeless aren’t from the city, and of those it claims are from the city, the department says almost 20% (869) had arrived in SF less than a year before becoming homeless, and it cannot rule out the possibility that over half (2,661) of them could have been couch surfing with friends, staying in motels, etc., for a year after moving from somewhere else. Viewed from this perspective, the HSH’s most recent data, which counted 7,754 homeless, doesn’t rule out the possibility that the majority (5,778) of San Francisco’s homeless are not from San Francisco. Nevertheless, the SF Chronicle and SF Standard have used this data to claim that most of the city’s homeless are actually from the city.

Vox’s claim that most of California’s homeless are Californians relies on research conducted by the Benioff Homelessness and Housing Initiative at UCSF, established in 2019 with a $30 million donation from Salesforce CEO (and Prop C supporter) Marc Benioff and his wife Lynne. The UCSF survey, published this June, describes itself as “the largest representative study of homelessness in the United States since the mid-1990s, [providing] a comprehensive look at the causes and consequences of homelessness in California.” Its contributors include dozens of statisticians, epidemiologists, and anthropologists.

Unfortunately, the study was subject to methodological flaws similar to the HSH’s. UCSF’s research relies on uncorroborated, self-reported data collected through questionnaires issued to around 3,200 homeless people throughout California. Critically, UCSF’s survey doesn’t ask respondents where they’re originally from, but rather “where they were last housed,” which the paper describes as (emphasis added) “the last non-institutional setting where they stayed for one month or more, or the last institutional setting where they stayed for three months.” To this end, researchers found that the majority (68%) of survey respondents became homeless after being “housed” somewhere they didn’t pay rent: 

• The majority1 of those who were “housed” in “non-institutional settings” — 49% of total — were not on a lease when they were “housed”

• The majority2 of those who were “housed” in “institutional settings” — 19% of total — were “directly” from jail or prison, most3 of whom were homeless before being in jail or prison

These figures indicate the majority of previously “housed” “Californians” had come from out of state, couch surfed for at least four weeks, then became homeless, or came directly from jail or prison (and were homeless before that). The most generous possible interpretation here is that UCSF simply didn’t ascertain where the majority of the state’s homeless are from. Regardless, the study’s authors conclude, somewhat triumphantly, “90 percent of participants became homeless in California, having been last housed in the state. People who experience homelessness in California are Californians.”

UCSF’s “landmark study,” as Vox called it, does not only do nothing to disprove the claim that most of California’s homeless are from out of state, it has a clear ideological bent. Note the phrase “people who experience homelessness in California are Californians” from above. The researchers bely their dogmatism with this and other statements rife in the report. For example, in their discussion of reasons for homelessness, the authors incoherently blame it on homophobia, among a variety of identity-based factors:

Homelessness does not happen in a vacuum. It occurs in conjunction with structural conditions that produce and reproduce inequalities. These conditions include… the ongoing effects of classism, racism, sexism, homophobia, and transphobia.

In their discussion of “people who experience homelessness,” the authors claim “colonization” is not only a cause of homelessness but that it somehow prevents people from exiting homelessness:

People who are members of populations marginalized by racism and colonization face economic and structural disadvantages. These structural disadvantages lower the threshold for people who face them to become homeless and prolong episodes by creating barriers to exiting homelessness.

And earlier in that chapter, the researchers write that “exposure” to structural racism causes homelessness:

[When] we ask who is homeless, we find that those with certain individual vulnerabilities to homelessness—either because of a health condition or exposure to structural racism—are at increased risk of homelessness.

Combined with its poor methodology that still fails to mask the suggestion in the data that most of the study’s respondents are from out of state, the researchers’ clear ideological bias makes their work impossible to rely on for a true account of where the state’s homeless are from. Nevertheless, the most authoritative outlets in mainstream media, such as Fortune, The New York Times, The Guardian , The Atlantic, and many others, have uncritically repeated the researchers’ claims.

“For over two decades, I’ve interviewed the homeless, usually about their dogs,” veteran investigative journalist and San Franciscan Susan Reynolds told us when we asked her where she thinks most of the city’s homeless are from. “At first, one said he was from San Francisco, but as we talked more over lunch, he told me he was from Ukiah.” She went on: “If you go deeper with these folks, they will usually be honest. And in my experience, 99% of them are not from San Francisco. When I would ask what elementary school they went to or what neighborhood they grew up in, the truth would come out.”

But for both San Franciscans and Californians, the truth about homelessness in their communities has always been plain as day. The real question is: how long can politicians and the media can keep lying about it before losing control of the narrative?

-Brandon Gorrell and Sanjana Friedman