Mental illness and homelessness: it's not what you think

When most people think of homelessness, they picture a person they passed on the street: visibly disorganized, talking to themselves, perhaps frightening. The natural inference is that the person is mentally ill and that mental illness is the reason they're on the street. The natural follow-up is that homelessness is mostly a mental-health problem and that more psychiatric beds would solve it.

This intuition contains some truth and a lot of error. Untangling it changes what kind of solutions make sense.

The actual numbers

Across multiple studies of US and Canadian homeless populations, the share with a serious mental illness (schizophrenia, bipolar disorder, severe depression, severe PTSD) is roughly 20-25% of all homeless adults — about three times the rate in the general population, but still a minority.

Among the chronically homeless — people who have been homeless a year or longer, or have had four episodes in three years and have a disabling condition — the share rises to about 30-50%. That's the population most visible to the public.

Among the transitionally homeless — people who experience homelessness once briefly and return to housing — the rate of serious mental illness is barely different from the general population.

So the popular image of "the homeless person" significantly overrepresents the visible chronic population and significantly underrepresents the much larger transitional population, which mostly looks like everyone else.

Cause or consequence

Even among people with both a mental illness and a housing situation, the relationship between the two is not what casual observation suggests.

For some people, mental illness comes first. A severe mental illness disrupts work, relationships, and self-care; the person loses housing as a consequence. This is the path that most people imagine.

For many others, homelessness comes first and mental illness follows or worsens because of it. Living unsheltered means sleep deprivation, chronic stress, sustained fear, violence (homeless people are far more likely to be assaulted than to assault), loss of social contact, exposure, and routine humiliation. These conditions reliably produce or worsen psychiatric symptoms in anyone — including people who were mentally healthy when they became homeless. PTSD rates among chronically homeless people are extreme.

This matters because it changes what works. If mental illness is the cause, the policy answer is more treatment. If mental illness is also a consequence of unstable housing, the policy answer is housing — which then makes treatment work.

What the research shows about housing first

Multiple controlled studies have tested the same question: do you have to treat mental illness before housing someone, or can housing come first and treatment follow voluntarily?

The largest study is Canada's At Home / Chez Soi trial, which enrolled 2,148 chronically homeless people with serious mental illness across five cities, randomly assigned half to Housing First (housing with no treatment preconditions) and the other half to treatment-as-usual. The Housing First group spent dramatically more time housed (about 75% of the follow-up period versus about 40%) and had better outcomes on most other measures — including mental-health symptoms.

That last finding is the important one. Putting people in housing first did not require their mental illness to get worse first. It improved.

The mechanism is straightforward: it is nearly impossible to do effective mental-health treatment with someone who is sleeping outside. The person cannot show up to appointments reliably. They cannot store medications safely. They cannot sleep enough to get well. They cannot access voluntary services because the prerequisites (an address, a phone, a place to be) don't exist.

Housing them creates the conditions under which treatment can work. The reverse — withholding housing until they're stable — assumes a kind of stability that homelessness systematically prevents.

The deinstitutionalization argument

A common claim is that homelessness rose in the US starting in the 1980s because of the closure of state psychiatric hospitals — that we used to house mentally ill people in institutions and now they sleep on the street.

There is some truth to this. The state hospital census did fall dramatically, from about 559,000 patients in 1955 to under 40,000 today. But the timing doesn't fit cleanly. Most of the deinstitutionalization happened in the 1960s and 1970s. Visible homelessness emerged as a mass phenomenon in the 1980s, alongside two other developments: a sharp reduction in federal housing subsidies under the Reagan administration, and a tightening of housing supply in many cities.

Most analyses now treat deinstitutionalization as a contributing factor but not the dominant one. The dominant factor is the same as everywhere else: housing costs rose faster than wages, and the cushion that absorbed precarious people disappeared.

If deinstitutionalization were the dominant cause, we'd expect homelessness rates to correlate with state-hospital closures. They don't. They correlate with housing costs.

The case for more psychiatric capacity

None of this is an argument against more mental-health services. The US in particular has badly inadequate psychiatric capacity: too few inpatient beds for acute crises, too few community mental-health centers for ongoing care, too few psychiatrists who take Medicaid, almost no integrated treatment for people with co-occurring substance use and severe mental illness. All of this should be expanded.

What it is not is a substitute for housing. A psychiatric bed is not a home. Sending someone with severe mental illness to a 72-hour psychiatric hold and then discharging them back to the same street where they were picked up does not produce sustained improvement — it produces a cycle of admissions, often called the "psychiatric revolving door."

The combination of more psychiatric capacity and more housing produces the outcomes everyone says they want. Either alone does not.

Practical implications

• If you are a voter: the most consequential thing you can do is support housing supply. That is the binding constraint. More psychiatric beds also help, but only if there is housing on the other end of the discharge.
• If you are a donor: organizations doing Housing First plus integrated behavioral health outperform organizations doing one or the other. See our effective charities list.
• If you have a family member experiencing this: in most US states, you cannot involuntarily commit an adult unless they are an imminent danger. This is a real legal constraint. The best you can usually do is connect them with a case manager, ensure they have access to medication and benefits, and keep the door open. Patience is sometimes the most loving option.
• If you are working in the field: the Housing First model is widely understood; the under-recognized challenge is integrating mental-health care into the housing rather than running it as a separate referral track.

Bottom line

The popular framing — that homelessness is mostly mental illness and that the answer is more treatment — is significantly wrong on the first claim and only partially right on the second. Most homelessness is not mental illness. Most mental illness on the street was made worse by homelessness, not the other way around. And the treatments that work depend on the person having a home to receive them in.

The right framing is the boring one: build housing, fund integrated services, and let people get well in the order that actually works.

Read more: What actually works: Housing First · Common myths · Homelessness by population.