tents
© unknownHomeless Camp • Los Angeles, California
In recent years, the rate of homelessness in California has grown substantially. The Golden State now ranks first among the 50 states at 4.4 per 1,000 residents, with LA and San Francisco often hosting vast homeless encampments (or tent cities, as they're also known). I say "often" because these encampments are frequently cleared away by police, only to reappear somewhere else.

Many have blamed California's homelessness problem on the cost of housing there. One prominent dissenting voice is the activist Michael Shellenberger, who has argued - most famously in his book San Fransicko: Why Progressives Ruin Cities - that it's largely the result of untreated mental illness and addiction.

In June, Margot Kushel and colleagues at the Benioff Homelessness and Housing Initiative at UC San Francisco published findings from "the largest representative study of homelessness in the United States since the mid-1990s". They interviewed 3,200 homeless people across eight locations, with effort being made to achieve representativeness.

So, what were the findings?

Remarkably, 82% of the homeless people in their sample said they'd had a mental health condition, and almost 1 in 4 said they'd had hallucinations (which are indicative of serious conditions like schizophrenia and borderline personality disorder). By contrast, the lifetime prevalence of any mental condition in the general population is only 46%.
graph report
© unknown
The researchers also asked participants whether they currently had a mental health condition, and 66% said they did.

As for addiction, 65% of homeless people in the sample said they'd abused drugs or alcohol at some point in their lives. And among that sub-sample, two thirds said they'd done so before their first episode of homelessness. Which means that homelessness cannot have caused them to start abusing drugs or alcohol (more plausibly, it was the other way around).
chart
The lifetime prevalence of drug abuse in the U.S. is around 8%, while for alcohol abuse its around 18%. Assuming that drug and alcohol abuse are completely independent (which they probably aren't) the life prevalence of either is 26%.

Overall, the survey confirms that people with mental health conditions and a history of substance abuse are massively overrepresented among the homeless in California ­- consistent with Shellenberger's arguments.

However, there doesn't seem to be much correlation between the rate of homelessness and the prevalence of mental illness or addiction across U.S. states - which is wholly inconsistent with Shellenberger's arguments. California is neither the most mentally ill nor the most addicted state.

What's more, there does seem to be an association between the cost of housing and the rate of homelessness (which is hardly surprising). Indeed, homelessness is most common in liberal states that tend to restrict housing development, notably California and New York.

Here's what appears to be going on:
Mental illness and addiction can help to explain which individuals become homeless, but they can't explain why some states have higher rates of homelessness than others. Why so much homelessness in California? As the conventional wisdom says, it's the cost of housing.