Richard Vogel/AP Photo
A homeless man walks with his dog along a street in the Skid Row section of Los Angeles, January 11, 2024.
Despite enjoying decades of bipartisan support, Housing First programs have become a target in the culture wars. Housing First provides housing and wraparound services that are not predicated on sobriety or psychiatric treatment. Conservative critics challenge this approach, arguing instead that people should be required to work and participate in abstinence-based addiction treatment or faith-based programs to be eligible for housing.
These opponents of Housing First begin from the flawed premise that human pathology is the root of homelessness. By contrast, research clearly shows that housing shortages, wage inequality, and high rents—not mental illness, substance use, or even poverty—are its real causes. These structural factors explain why homelessness is more common in Seattle and San Francisco than in Detroit or Oklahoma City, where there is still plenty of addiction and mental illness, but where people with these conditions spend their time indoors.
Still, even if homelessness is primarily a housing market and financial insecurity problem, does it automatically follow that Housing First is the solution? As sociologists who study urban inequality and homelessness, we have seen Housing First’s successes and failures firsthand. We believe it is an essential component of homelessness policy.
But Housing First faces legitimate questions that should come as much from the left as the right. Is it successful and cost-effective? Under what conditions? Can it be improved? And what about people whose behavioral issues are so severe that they cannot thrive, even in programs based on Housing First principles?
The question of whether Housing First is effective is complex. The answer depends on how success is defined. Unfortunately, proponents and opponents usually elide this complexity, highlighting the outcomes that bolster their arguments while ignoring those that don’t.
Housing First faces legitimate questions that should come as much from the left as the right.
Supporters point to decades of research showing that about 85 percent of the people who enter Housing First programs remain housed years later—a huge success compared to older, “treatment as usual” models. Critics focus instead on substance use and mental illness among Housing First residents. From their perspective, Housing First programs do little to improve behavioral health and merely move “social pathologies into physical housing complexes.”
In fact, the research is mixed. According to the most recent meta-analysis, research does show improvements in health and well-being among Housing First participants. But studies also suggest a high degree of variability in these outcomes. Disturbingly, it appears that the mental and physical health of participants in some Housing First programs isn’t better than for those who remain on the streets.
In our work, we’ve seen how programs with well-staffed community case management teams that conduct daily visits, take residents to appointments, and build close relationships can significantly improve residents’ health and well-being. We’ve also seen how understaffed and underpaid providers struggle to prevent evictions, keep staff safe, and keep clients out of jail. We’ve spoken with people who feel safer sleeping outside than in notoriously decrepit and dangerous housing. This may be obvious, but it bears repeating: Good care cannot be provided on the cheap. It is imperative that programs receive robust funding and that providers be properly trained and compensated.
Consider Roxanne (not her real name), a Black woman in her fifties, whom one of us, Neil Gong, got to know while studying supportive housing in Los Angeles. A longtime street drinker with serious mental illness, Roxanne had just been housed by a local nonprofit and L.A. County’s Department of Mental Health. Her case manager Carlos checked in with her, kept her fridge stocked, and reminded her what the doctor said: More drinking could be fatal.
Carlos hoped stable housing and regular visits would enable him to build trust and slowly coax Roxanne into reducing her alcohol consumption. But the clinic was underresourced and had lost staff; remaining case managers were overwhelmed and would have to focus only on people in immediate crisis. Carlos worried he’d find Roxanne dead, and indeed, within the year he did.
This tragedy shows how underresourced Housing First descends into mere “tolerant containment”—keeping people out of the way while accepting problem behaviors. Sadly, Roxanne’s death and other tragedies are routinely manipulated for ideological purposes. In San Fransicko: Why Progressives Ruin Cities, provocateur Michael Shellenberger misleadingly uses her story and Gong’s research to critique Housing First. Omitting the discussion of staffing shortages, Shellenberger concludes that “Gong found that leniency toward drug and alcohol abuse could result in death by overdose.” A critique from the left—that poorly funded programs are set up to fail—becomes moralizing about liberal “leniency” from the right.
The next point of contention is cost-effectiveness. Critics say Housing First is simply too expensive; proponents insist that it saves money. Who is right? It is true that by keeping people in homes and out of emergency rooms and jails, Housing First programs can reduce costs in overall municipal expenditures. Yet many of those who experience homelessness do not cycle through ERs or have costly health issues. Housing this part of the population might be the right thing to do, but it won’t necessarily save money.
We have more than enough evidence to know that homelessness is not simply a matter of human failure.
Cost is also tied to city-level housing markets and regulations. Houston has worked miracles with Housing First, but Los Angeles has floundered. The difference lies in the fact that Houston has abundant housing supply and few restrictive zoning laws. In Los Angeles, efforts to build new units have been thwarted by NIMBYism, high costs, and the uncertainty that surrounds all construction projects in California. To make progress in addressing homelessness where housing is in short supply and construction delays are inevitable, we need to invest in short-term shelter options while taking steps to expand the housing supply in the long run.
Finally, critics of Housing First note that many people with serious mental health issues remain on or return to the streets. It’s true that about 15 percent of people who are placed in Housing First programs either voluntarily leave, get evicted, or are arrested and jailed. Our observations strongly suggest that with the right investments, we can meaningfully reduce the size of that group. But it is also true that a very small percentage might require alternatives like long-term psychiatric hospitalization.
So where should policymakers go from here? Our observations lead us to believe that variability in the health and well-being of Housing First participants has everything to do with the quality of the care and services people receive while housed. Conservative prescriptions for tough mandated services as a prerequisite for housing are not backed by evidence, but expanded and often improved services are needed. Hiding people away in underresourced, dangerous, and depressing housing facilities should not be equated with success. And to deny the fact that some people require in-patient hospitalization serves no one. Instead, in the rare cases where coercion may be necessary, efforts must be made to protect the rights of people who are committed and ensure quality of care in those facilities.
We have more than enough evidence to know that homelessness is not simply a matter of human failure. But Housing First initiatives do face challenges, and not all criticism coming from the right is inaccurate. Progressives might worry that acknowledging any merit in critiques of Housing First will simply add fuel to the fire. Surely this is a risk. But sweeping the challenges under the rug will not serve the unhoused in the long run.