Housing is one of our best tools for ending mass incarceration. It does more than put a roof over people’s heads; housing gives people the space and stability necessary to receive care, escape crises, and improve their quality of life. For this reason, giving people housing can help interrupt a major pathway to prison created by the criminalization of mental illness, substance use disorder, and homelessness.

For this briefing, we examined over 50 studies and reports, covering decades of research on housing, health, and incarceration, to pull together the best evidence that ending housing insecurity is foundational to reducing jail and prison populations. Building on our work detailing how jails are (mis)used to manage medical and economic problems and homelessness among formerly incarcerated people, we show that taking care of this most basic need can have significant positive downstream effects for public health and safety.

A Venn diagram showing some of the ways in which homelessness, mental illness, substance use disorder, and criminalization and incarceration overlap.

Using housing to interrupt cycles of incarceration

Homelessness, substance use disorder, mental illness, and incarceration are deeply intertwined experiences. Around 45% of adults in the United States who have been diagnosed with serious mental illness¹ also have a co-occurring substance use disorder. People with such dual diagnoses are 12 times more likely to be arrested than people with neither diagnosis. This is borne out in prison populations. A study of Iowa’s state prisons, for example, found nearly 54% of people with serious mental illness also had a substance use disorder.

Drug and alcohol disorders can be both a cause and a consequence of being unhoused. Many people use drugs to self-medicate and cope with mental illness and the constant stressors of homelessness. One point-in-time count of homelessness across the United States in 2022 found that roughly 21% of unhoused people had a “severe” mental illness, and 16% engaged in “chronic substance abuse.” As frequent targets of aggressive policing, unhoused people face constant threats of criminalization. A recent survey of 441 unhoused people in Colorado found 36% had been arrested for a crime of homelessness, 70% had been ticketed, and 90% had been harassed by police, while recent research shows that 1 in 8 city jail bookings in Atlanta involved a person experiencing homelessness. In the end, policing homelessness creates a vicious cycle of poverty and confinement, where basic needs are never met: Formerly incarcerated people are almost ten times more likely to be unhoused than the general population, and 52,000 people who left correctional facilities in 2017 directly entered shelters.

For many years, housing policy was dominated by moralistic views of homelessness, which held that people just needed to take matters into their own hands and pull themselves up by their bootstraps. Drug use and mental illness were deemed character flaws and personal weaknesses, incompatible with housing or employment, and total sobriety and participation in treatment were required to receive what few services were available. Thankfully, decades of advocacy have begun to supplant these ideas with more effective supportive housing models like Housing First.

Housing First programs provide low-barrier permanent supportive housing with wraparound voluntary mental health services and case management. These programs have demonstrated success in ending chronic homelessness and improving quality of life, especially among people experiencing both substance use disorder and mental illness.² Though advocates and scholars have for years urged Housing First facilitators to better target services toward the needs of people with criminal legal system involvement, research has shown these programs are effective in reducing arrests and incarceration even when they aren’t tailored specifically to criminalized populations.

Take, for example, this report which examined results from a 10-year follow-up with participants in the New York City Frequent Users System Engagement program (NYC FUSE) – a supportive housing program working with housing providers in the city, including Housing First practitioners. Compared to a closely matched comparison group, the researchers found that participants spent an average of 95 fewer days in jail, and 256 fewer days in shelters, over the 10-year period.

Other programs focused specifically on arrest, incarceration, and reentry have shown equally impressive results:

Summary of findings from four studies of supportive housing programs serving people with a history of criminal legal system involvement.

Housing Program	Arrest & jailing outcomes	Other positive outcomes
NYC FUSE	95 fewer days in jail	256 fewer days in shelter
Denver Social Impact Bond (SIB)	40% reduction in arrests	40% reduction in shelter stays
	30% reduction in jail admissions	65% reduction in use of emergency detox services
Returning Home-Ohio	40% less likely to be arrested again	Remained in community for longer before rearrest
	61% less likely to be incarcerated again
Solid Start (Missouri)	Not measured	Compared to the comparison group, participants felt more:
		Independent and integrated into the community
		Capable of stabilizing life
		Confident and secure in their housing placement
		Distanced from unsafe and criminogenic environments
		Aided in release transition
		Able to build a support network
		Clear in describing pragmatic future plans
		More personally responsible and capable of taking action

The Denver Social Impact Bond (Denver SIB)

The city of Denver, Colorado launched a housing initiative in 2016 for people experiencing long-term homelessness who had frequent interactions with police and emergency health services. The initiative, which is no longer active, provided housing subsidies with limited requirements, voluntary intensive clinical treatment and case management services, and assistance navigating the criminal legal system. In a study comparing people in the Denver SIB program to those receiving “services as usual” in the community, researchers found program participants spent significantly more time in housing: 77% percent stayed in their housing after 3 years, and they used shelters 40% fewer times than the comparison group. They also experienced 34% fewer police interactions and 40% fewer arrests than their peers. Denver SIB participants spent 27% fewer days in jail, and were booked into jail 30% less often. Finally, participants used emergency detoxification services 65% less often than the control group while using preventative and community-based care more often.

The Returning Home-Ohio Pilot Project

The Returning Home-Ohio Pilot Project, funded largely by the Ohio Department of Rehabilitation and Correction, linked disabled incarcerated people who had a history or risk of housing instability to supportive housing upon their release. The pilot, which became a permanent program in 2012, was implemented in 2007 and reached 13 prisons. It provided coordinated prerelease reentry planning, housing, and supportive services in five Ohio cities. Comparing participants to similarly situated formerly incarcerated people, researchers found participants were 40% less likely to be rearrested and 61% less likely to be reincarcerated.³

St. Louis, Missouri’s Solid Start Program

One study out of Missouri provides strong evidence of Housing First programs’ potential to encourage positive shifts in attitude and self-perception, which are important for successful reentry and desistance from crime. The Solid Start program provided housing for one year to about 30 men on parole at a time, who entered either directly from prison or after a short stay in the community. Participants were eligible if they had experienced over 10 years of incarceration, little community support, substantial child support or other financial obligations, no consistent work history, a maxed-out sentence, or a mild-to-moderate mental health disorder. The program provided housing subsidies as well as coordinated services and case management, and required participation in weekly group therapy sessions. According to data from 2010, Solid Start participants reported fewer problems and greater satisfaction with their accommodations compared to a group of similar men on “traditional parole.” They also felt more self-sufficient, and like they could overcome financial obstacles to independent living, viewing the program’s support as temporary. Solid Start participants also felt better integrated into the community and capable of stabilizing their lives thanks to their independent home placement. They were less likely to report that they were living in undesirable or criminogenic environments and were able to describe future plans with more clarity than the comparison group.

Housing First works, but it doesn’t solve everything

Housing can help people dramatically improve their lives, but these programs are not a panacea. They depend on affordable housing units and access to funding to operate, and those resources are extremely limited. Simply giving someone a place to live does not guarantee that they are being properly cared for, either. They may have particular safety and service needs that are not guaranteed or readily available through these programs, or the kinds of housing available may not be conducive to their social, spiritual, or cultural needs and values. Even successful models like Housing First struggle to help everyone given these constraints.

Supportive housing programs, like all housing programs, are filling gaps caused in large part by insufficient and discriminatory housing policies. They provide subsidies for housing, but must compete for funding and open housing units. Fewer open apartments, higher rents, long waiting lists, and the struggle of cobbling together funding scattered between different government agencies to cover everything from rent and down payments to physical and mental health care all make it exceedingly difficult to house someone. Add to this discrimination against tenants by landlords and neighbors, persistent policing in areas where housing is provided, and the struggle of supporting people amid rising costs of living, and placing and sustaining participants becomes even more difficult.

Some Housing First program workers have noted that many of these factors, and the overall intense urgency of clients’ housing and health needs, means they are constantly stuck in crisis mode and rarely able to plan or work through issues with their clients. While the model is successful at producing housing stability, many providers have felt that people should be staying in the programs longer. The goal of graduating people out of the program can actually be counterproductive to their work in some cases, introducing pressure and stress, or encouraging people to avoid graduation out of fear.

There are steps Housing First programs can take to improve on their own, regardless of the housing or funding situation. Advocates for unhoused women and indigenous people have argued that these programs should be far more inclusive. The general approach of Housing First programs means they tend to engage a predominantly male street-dwelling population. Women and femmes, for example, tend to have distinct traumatic and gendered pathways to homelessness, and often avoid shelters for fear of violence – meaning they are often out of the recruitment range of Housing First programs. They also are more likely to be caring for children and require specific services and assistance that they may not get through typical Housing First programs. Although some research has suggested Housing First placements spread across a community’s existing residential buildings (known as “scattered site” housing) have better outcomes, these accommodations may not work best for women and femmes, who may benefit from shared (or “congregate”) settings due to a higher level of security and more communal spaces.

What constitutes a safe and stable “home” is also not universal. In Canada, indigenous participants in Housing First programs felt their accommodations left them disconnected from their community. Prohibitions and restrictions on having guests, the inability to participate in smudge or sweat ceremonies in provided housing units, and the way in which housing placement eligibility conflicted with the customary mobility of some indigenous people, all expose how simply giving people apartments may be preferable to housing precarity, but falls short of meeting everyone’s needs.

Studies of other models reinforce housing’s role in promoting public safety

While Housing First models have some of the most robust bodies of evidence to back them, research into other housing models reinforces the core elements of the model, and underscores how housing increases safety and stability, holding promise for challenging mass incarceration.

Supportive housing: While Housing First models are a kind of supportive housing, not all supportive housing programs follow the Housing First model. Some supportive housing programs do require treatment and abstinence. Others, such as the Housing and Urban Development Veterans Affairs Supported Housing (HUD-VASH) program provide housing subsidies, health care referrals, and case management but, unlike typical Housing First programs, do not offer substance use treatment as a core part of the program. One study of the HUD-VASH program found that while participants spent more time in housing and reported increased functioning and reduced substance use, veterans who had substance use disorders still needed more services than the program provided.

Transitional housing: Results from studies of transitional housing models, which provide housing and services to people for shorter periods of time on their way to more permanent housing, are also in harmony with the research on Housing First. One 2010 report from the Department of Housing and Urban Development found that providing people with housing quickly improves their stability and likelihood of remaining housed. It also found that longer periods in transitional housing were associated with better outcomes, confirming the benefits of long-term or permanent arrangements like those provided under Housing First. Programs like A New Way Of Life have demonstrated success under this model: In their 2022 annual report, they note that 41 women in their program were able to access permanent housing that year and 99% of women served were not reincarcerated. That being said, some analysis of research on halfway houses – a form of transitional housing under correctional control – is more ambiguous, and suggests that such a punitive model may be associated with higher rates of rearrests.

Recovery housing: In general, recovery housing or “sober living housing” is specifically for people with substance use disorders. These programs typically mandate treatment and/or abstinence to some degree. Some research indicates recovery housing leads to reductions in substance use, improvements in employment, and desistance from criminal activity. But it’s difficult to generalize because the level of abstinence required and definitions of “recovery” vary between programs. Even though research on Housing First tends to indicate that sobriety and treatment are not necessary to house people stably, it is probably best practice for Housing First and sober living houses to be developed in parallel. Emerging models like Housing Choice, which offer people choices between housing programs with various rules and requirements around abstinence and treatment, are experimenting with this conclusion.

Conclusion

People caught in cycles of incarceration and homelessness are not all alike; they have different pathways to those experiences as well as a range of needs. But housing is one special factor that can stabilize multiple aspects of a person’s life at once.

Available research strongly suggests that for most people, providing housing quickly, for as long as possible, with few conditions and as much choice and support as possible, is a practical way to improve people’s conditions, making it easier for them to manage other parts of their lives. The impact housing has on quality of life and a person’s relationships, attitudes, and sense of control are also key to reducing a person’s likelihood of arrest and incarceration, use of emergency services, and experience of other life crises.

Housing is not a universal remedy and existing housing models can be better supported and improved. But housing has the potential to be one of the most impactful investments to reduce incarceration without investing more in the criminal legal system itself.