Homelessness is about housing

The solution is to legalize more and more kinds of it

Something that I took for granted in one section of “Some Thoughts on Faculty Lounge Politics” that I realize not everyone knows or agrees with is that homelessness is fundamentally continuous with general housing policy.

This is at odds with how it is normally constructed in American politics, in which you have a discrete population of people who do not have a home and then a question of what to do about them. As a last resort, this tends to become a law enforcement question, but in essentially every city, there is also a social service provider community that is trying to help the homeless in non-punitive ways. So then you have a discourse about the funding of social services and what boundaries law enforcement should enforce. There are then micro-niche issues about how shelters for people experiencing homelessness should be run. And, there is a perennial conversation about root causes — the people sleeping on the streets are often suffering from addiction problems or mental health difficulties, and we maybe need to straighten those issues out somehow or other.

These are all, I think, valid questions to ask in the sense that people need to take the world as they find it and deal with situations that exist.

But broadly speaking, I think it misconstrues the problem as a “homelessness” problem with primary adjacencies to mental health, addiction services, and law enforcement. I think it’s a housing problem with primary adjacencies to questions like “why was New York City’s population falling even before the pandemic?”

Nobody goes there anymore, it’s too crowded

The Covid-19 pandemic was unfortunate for many, many, many reasons, but for housing policy junkies, one downside is we wound up paying no attention to the 2019 NYC population decline since the following decline in 2020 was so clearly about the pandemic. In 2021, there is sure to be a post-pandemic bounceback and a lot of stories about New York’s struggle to come back.

What happens in New York is always important both because it is by far America’s largest city, but also because the overwhelming concentration of media there tends to mean issues are read through a New York City lens.

So the question of why people were fleeing NYC in 2019 — before the pandemic, before the school closures, before the rise in the murder rate — should have been an analytically significant issue. People can leave cities for all kinds of reasons, after all, but most of them would come down to “it became a worse place to live.”1 But what you expect to see if somewhere becomes a worse place to live is that the price of living there falls thanks to low demand.

But what we actually observed was a strong increase in rents over the decade.

Rather than New York becoming a worse place to live, it became a more expensive place to live. Construction of new units did not keep pace with the increase in demand. So the population fell as families who might have had three or four people living in a two-bedroom apartment were replaced with childless professional couples, and wealthy people combined units to make larger dwellings for themselves.

Demand for housing is rising but the supply does not rise at the same rate, so people get squeezed out and the city’s population falls as people head for cheaper pastures in the Sunbelt. But what if you’re squeezed out of housing and you don’t leave for somewhere else? Well, then you’ve got nowhere to live.

Homelessness is high where housing is expensive

Leftists often raise the true-but-misleading factoid that the number of people experiencing homelessness on any given day is generally much smaller than the number of homes that are unoccupied on any given day.

But as we covered in “Homelessness and Vacant Houses,” it’s not the case that homelessness is high where vacancy rates are high. Indeed, it’s the opposite — the vacancy rate is lower in places with more homelessness.

Here’s an analysis looking at states:

And here’s one looking at large metro areas:

That’s because unusually high levels of homelessness and unusually low levels of vacancy are both caused by the same thing — scarcity of dwellings.

According to the United States Interagency Council on Homelessness’ 2019 count, there were roughly as many unhoused in New York state on a given day as in Texas, Florida, Pennsylvania, Georgia, and Illinois combined, despite New York having a much smaller population. That’s because New York City circa 2019 was a huge engine of displacement, featuring net population losses and also people who, for whatever reason, got squeezed out without going anywhere.

We think of the “homeless person” and the “moved to North Carolina to be closer to family and also to afford more space for the kids” as two entirely different types of people. But it’s a single underlying phenomenon. And who ends up in which category will come down to a mix of luck, whether or not you do in fact have family in North Carolina, and whether or not the forces pushing you out of the high-cost area come up on you slowly so you have time to plan.

The top six states for homelessness rates are New York, Hawaii, California, Oregon, Washington, and Massachusetts, which is basically just a list of states where housing is expensive. The highest homelessness rate in Texas is in Austin, which is the city with the most expensive housing.

More homes, less homelessness

I know it’s a little bit counterintuitive to people, but in a broad sense, the biggest thing we could do to take a bite out of homelessness in the medium term is build more luxury condos.

When I first started making the case for land use reform, we had two legs to our argument. One was basic economic theory — more supply equals less scarcity equals lower prices. The other was cross-sectional analysis — metro areas with laxer land use regulation see more construction and lower prices. But we can now add a third leg to that analysis in terms of detailed empirical work. Here’s a good roundup of five studies from the UCLA Lewis Center that they did in February, and in April, we have “Local Effects of Large New Apartment Buildings in Low-Income Areas” and it says the same thing:

We study the local effects of new market-rate housing in low-income areas using microdata on large apartment buildings, rents, and migration. New buildings decrease rents in nearby units by about 6 percent relative to units slightly farther away or near sites developed later, and they increase in-migration from low-income areas. We show that new buildings absorb many high-income households and increase the local housing stock substantially. If buildings improve nearby amenities, the effect is not large enough to increase rents. Amenity improvements could be limited because most buildings go into already-changing neighborhoods, or buildings could create disamenities such as congestion.

The fact that the new buildings are not “affordable housing” in the regulatory sense doesn’t mean that they have no impact on affordability. Nor does the fact that a person existing on the margins of homelessness would be unable to live in such a building change the fact that allowing their construction reduces homelessness by increasing the affordability of other units.

With this new empirical work, I’m increasingly confident saying that the key thing really is to legalize as much market-rate housing as possible. Not because there’s anything wrong with subsidized housing. But because market-rate housing generates tax revenue by bringing in new affluent residents, and that revenue can be used to finance an affordable housing trust fund or new public housing or whatever else you want. Cities looking to tackle homelessness should upzone for more market-rate construction and enhance their existing social service spending in line with the increased revenue.

But they should also take a look at the range of building types that they allow.

A bad house is better than no house

Paytung Chung wrote a post about the history of housing in the District of Columbia that has stuck with me ever since I first read it eight years ago:

The 1951 sci-fi classic “The Day the Earth Stood Still” inadvertently shows us how. Klaatu, a level-headed extra-terrestrial emissary, escapes captivity at Walter Reed Army Medical Center. He wanders down Georgia Avenue, away from not-yet-nuclear-weapon-free Takoma Park, and attempts to disappear into everyday DC.

To do so, Klaatu checks into a boarding house at 14th and Harvard in Columbia Heights. Each room houses one or two people, and as such there’s scant privacy to be had: everyone overhears everything.

This is convenient for Klaatu, who knows little of Earthlings’ simple ways, but probably annoying for the Earthlings. Conditions like these were common in DC homes at the time.

And, indeed, this was the typical mode of accommodation for a non-wealthy, non-married person in the United States who wanted to get out of mom and dad’s house and work in the big city.

These Library of Congress photos show life in a D.C. boarding house:

You can see two beds in a tiny room here. In a boarding house, you would share a bathroom with others on the hall and use a shared dining room.

In America today, this persists as a mode of accommodation for college students, but it’s largely vanished.

That’s mostly for good reasons. Thanks to widespread car ownership, there’s more space to build homes. And because the United States is a lot richer than it was in Klaatu’s time, we can afford to build larger structures. But obviously not everyone can afford to live in a modern one-bedroom apartment or there wouldn’t be a homelessness problem to blog about. The issue, as is common in American planning, is that we took something that’s normal (people want a place to park their car, people want their own bathroom in their dwelling) and we made it mandatory without regard to its implications for affordability on the margin.

I used to think of this as largely a case of unintended consequences or good intentions gone awry. But then I found the American Society of Planning Officials’ 1957 guidance on rooming houses and they are actually pretty clear that the regulation is pretextual.

They open by quoting an October 18, 1957 editorial in the St. Louis Post-Dispatch which says:

If rooming houses are permitted to spread to the city's one- and two-family neighborhoods, there is not much use in talking brave words about fighting blight. Rooming houses are not compatible with one- and two-family districts. When the rooming houses come in, the families move out — and the whole area starts down hill. If St. Louis is to retain its many fine family neighborhoods, the rooming houses will have to be kept where they belong.

And the report agrees that “even under the strictest codes, rooming houses are out of place in some neighborhoods.” But they go on to further argue that in addition to using zoning to keep boarding houses out of single-family districts, cities can use minimum quality regulations to keep out undesirable types even where they are allowed:

Zoning is not the only tool available to control the blighting effects of rooming houses. Housing codes in an increasing number of cities require that decent — though often minimal — standards be maintained in them. Besides protecting the roomers, enforcement of these codes can do a great deal to assure that rooming houses do not harm districts in which they are properly located.

The report notes that “many roomers are real down-and-outers, and the atmosphere of a rooming house in which they predominate is likely to be bleak” and complains that “hundreds of zoning ordinances have loopholes that permit group living arrangements.”

It took a while, but over the generations, the planners have been very successful at mostly eliminating the accommodations for down-and-outers2 with the consequence that if you are down and out in a city where real estate is expensive, you end up on the street.

Up from homeless shelters

Right now the state-of-the-art goal in well-intentioned policymaking is to try to build more homeless shelters so that people who don’t have homes aren’t sleeping rough.

A federal judge in Los Angeles is even attempting to require the city to provide shelter space for all of the thousands of homeless Angelenos living in Skid Row. The city, of course, has no capacity to actually do this and is appealing the decision. The argument made by the judge is that homelessness in Los Angeles is racist, due to what you can see here is an overrepresentation of African Americans among the unhoused.

This seems a little far-fetched as a legal theory, especially given the conservative majority on the Supreme Court. In general, a slight majority of those experiencing homelessness are white, and the ethnic group that most disproportionately suffers homelessness is Native Hawaiians/Pacific Islanders, which is pretty clearly because they overwhelmingly live in Hawaii, a state with very expensive housing.

Regardless of that legal theory, the irony of the homeless shelter is it completely replicates all the issues with the boarding house — the living conditions are not good and the neighbors regard it as an undesirable land use — except at the public’s expense and without the residents having the dignity of a place of their own where they can store their things.

It’s likely that even if we fixed all the land-use problems, there would still be some need for social capacity to provide this kind of emergency shelter. But there’s no good reason to have such a gap between the crappiest apartments for lease and the shelter. We ought to allow the full spectrum of market phenomena — including plenty of shiny new buildings for rich people and the subdivision of older buildings into boarding houses for down-and-outers and the occasional visiting alien.

One key point is that most people who find themselves down-and-out due to job loss, a breakup with a partner, a fight with family members, a dispute with a landlord, an emergency expense, or whatever other problem will in fact get their feet under them and bounce back. But it’s much harder to bounce back if you sink into a situation where you’re exposed to the health and safety problems of being on the street, don’t have easy access to a shower so you can be presentable when looking for work, and don’t have any place to keep your stuff.

More of everything

Now, what’s true is there are people out on the streets this week and they need help now. What I’m talking about here are mostly mid-range solutions, not crisis measures.

That said, we really ought to fix this problem rather than being stuck perennially in crisis mode. These are locally focused issues, so the precise nuances of the relevant zoning codes and planning terrain will vary greatly. But here’s a great account of the rise of Single-Room Occupancy (SRO) dwellings in New York — that’s the higher density equivalent of D.C.’s rooming houses — and their deliberate destruction, which ends up a call to lift the rules that prevent the construction of new ones. That’s something both Kathryn Garcia and Andrew Yang have talked about, and they’re right.

But I do want to emphasize that far and away the best use-case for SROs, rooming houses, and other subprime rental housing types is in a context when a lot of fancy new buildings are being built.

Under the current economics, it doesn’t make a ton of sense to invest a large sum of capital in constructing a brand new boarding house. And if you do, it would be an upscale boarding house targeting frugal recent college grads, not down-and-outers. There’s nothing wrong with that idea — housing is housing, as I’ve emphasized, and the markets are all connected — but it wouldn’t “bring back” the cheap housing of yore.

What brings back that very cheap housing is that in the context of lots of new high-end development, there’s less demand for the older housing stock. Some of it will get renovated to be shiny and new again. But historically, the very cheap housing came from subdividing old units that nobody really wanted in order to serve an even-more-downscale market.

Now don’t get me wrong! I’m not saying that living in some flophouse is the housing utopia or the future of housing for everyone. The main reason we no longer have lots of people living two-to-a-room in units with no bathroom is that we’re not living through the Great Depression and World War II anymore. Large, modern American houses are great. And we should try to have full employment and a generous welfare state and all the other good things that make it possible for people to enjoy high living standards. But in the United States today, even as most people have housing that is much better than Klaathu’s boarding house, some people are too poor for their housing and our current program is to offer them something much worse instead.


Historically we see two main things lead to community population decline in America. One is technology shocks, as when the automobile made it suddenly much less valuable to live within walking distance from a train station or a commercial strip. The other is economic decline, as jobs become less available in a particular place.


San Francisco does retain a non-trivial number of Single Room Occupancy (SRO) dwellings, and I’m sure there are a few scattered hither and yon in the rest of the country.