part of EQUORA Institute
HUEN

Mental Fast Food™ · Data Visualisation · Standing summary

Why is there
homelessness in every city?

Ask anyone and you'll get a confident answer: drugs, mental illness, bad choices, warm weather. Each feels obvious. Each is mostly wrong — at least about why one city has more than another. The research here is unusually clear, it just hasn't reached the street.

Source
HUD PIT 2024 · US Census · Apartment List · Pew Charitable Trusts · "Homelessness Is a Housing Problem" (Colburn & Aldern, 2022)
Note
City rates are per 10,000 residents; figures rounded. Individual risk factors explain who, not how many.

Interference topic · What the individual story explains, and what it doesn't

What if the individual story is true, yet still doesn't explain the citywide pattern?

I.
The obvious answers

Four answers
almost everyone gives.

When a city's homelessness rises, the explanations come fast. They all share one flaw: they describe traits that exist everywhere — yet homelessness is concentrated in specific places.

"It's the drugs."
Arkansas and West Virginia have some of the worst opioid burdens in the US — and some of the lowest homelessness rates. Cheap housing.
"It's mental illness."
Serious mental illness is roughly as common across US cities. Homelessness is not. It tracks rents, not diagnoses.
"It's poverty / bad choices."
Detroit is the poorest big US city — highest unemployment, ~33% adult poverty — yet its homelessness rate is about a fifth of San Francisco's.
"It's the warm weather."
Then why are cold, expensive Boston, New York and Portland near the top, and warm, cheap Houston near the bottom? Climate barely moves the rate.

None of these is nonsense at the level of one person — addiction, illness or a job loss really can tip someone onto the street. But that's about who becomes homeless. It says almost nothing about how many a city has. For that, one number does most of the work.

II.
The switch

Same places,
two different stories.

Each dot is a major US city. The vertical axis is always the same — how many people are homeless per 10,000 residents. Only the horizontal axis changes. First: poverty. If the obvious answer were right, the dots should line up. Watch what they do — then switch the axis to housing cost. (Hover or tap any dot for its name and numbers.)

Switch the axis

On the poverty axis the cloud is shapeless — Detroit sits poor-but-low, San Francisco sits less-poor-but-high. The intuition breaks. Flip to rent and the same dots snap into a line: the more expensive the housing, the more homelessness. Nothing about the cities changed — only what we measured them against.

The third view zooms out to whole countries (OECD harmonised data, so they're actually comparable). The housing-cost pull is still there — but it's noisier, because a country is not a market: national policy can override it. Finland carries a heavy rent burden yet very little homelessness, because "Housing First" deliberately breaks the link. The lever is real; politics decides whether anyone pulls it.

III.
Who vs. how many

Two questions
we keep mixing up.

The confusion at the heart of this topic is treating one question as if it were the other.

Individual factors
Who ends up homeless?
Addiction, mental illness, a lost job, a fled relationship, leaving prison or care. These decide which vulnerable person falls — like who catches a cold in a draughty room.
Structural factor
How many beds are missing?
How much affordable housing exists relative to people who need it. This decides how many fall in total — it's the temperature of the room. Lower the rent and fewer people hit the street, whatever their personal story.

Both are real. But policy aimed only at individuals — treatment, policing, "personal responsibility" — fights the cold one sniffle at a time while the room stays freezing. Raise the temperature and far fewer people get sick in the first place.

IV.
The exceptions

Cities that
broke the pattern.

If housing is the lever, then places that pulled it should look different. They do — in three very different ways.

Helsinki, Finland
Gave housing first — no conditions.
Finland's "Housing First" hands people a permanent flat before sobriety or treatment. Rough sleeping in Helsinki is nearly gone; shelter beds fell from thousands to ~50. The only EU country where homelessness fell for over a decade.
Proof it's solvable — but fragile: in 2025 the number rose again as living costs climbed and benefits were cut.
Houston, USA
Just kept housing cheap — and built.
No European welfare state, no warm-weather magnet effect reversed. Houston has the lowest rents of the big US cities and builds freely — and a homelessness rate roughly 15–19× lower than San Francisco's, despite faster population growth.
You don't need a perfect system. You need enough housing people can afford.
Tokyo, Japan
Almost no one on the streets — really?
Tokyo reports a few hundred rough sleepers in a city of 14 million — partly real (flexible welfare, easy building, cheap rooms), partly definition. The number behind the number
A low count can mean a solved problem — or a hidden one. Always ask what's being counted.
V.
So why everywhere?

So why is there
homelessness everywhere?

Because almost every big, desirable city has let housing get scarcer and pricier than the people at the bottom can survive — and the few that didn't are the few with little or no street homelessness.

That's the quietly radical part. Homelessness looks like a story about damaged individuals, so we reach for individual fixes and individual blame. But the amount of it in a city is mostly a story about a market and a policy choice — how many affordable homes a place allows to exist. The people are the ones who fall through; the size of the gap is on us.

What you can actually do with this isn't pity or panic — it's a sharper question.

Next time your city debates homelessness, notice which question is being answered. If every proposal is about who — more policing, more treatment mandates, moving camps along — and none is about how many homes exist, then the room is still freezing and everyone's arguing about coats. The honest fix is unglamorous: let more affordable housing be built, and keep people in the homes they have. That fixes the temperature of the room. It doesn't guarantee everyone stays warm — individual crises still happen, and individual support still matters — but it makes the scale of the problem tractable rather than permanent.

How a place measures its own housing gap honestly — beyond a one-night street count — is exactly the kind of question worth getting right rather than guessing.

EQUORA Institute →
The number behind the number

Why Tokyo shows so little
visible homelessness

Japan reports one of the lowest street-homelessness rates on Earth — a few hundred in Tokyo, under ~3,000 nationwide. Part of that is genuine: flexible cash welfare ("livelihood protection"), looser building rules that keep housing cheaper, and dense low-cost rooms.

But part is definition. Japan officially counts only people sleeping in parks, riverbeds and stations, surveyed by daylight visual patrol in winter. It misses the "net café refugees" — tens of thousands in Tokyo alone, with national estimates of 100,000–300,000 — living in 24-hour internet cafés, plus the doya districts of welfare-dependent former day labourers. Under most Western definitions they'd be counted as homeless.

So the lesson cuts both ways: Japan really did shrink street homelessness — and it also moved part of it out of view. A number is only as good as the definition behind it.

How Japan counts homelessness →

TakeawayMeasurement is itself a decision. What an official definition leaves out disappears from the statistics, not from reality — in Japan the tens of thousands sleeping in net cafés are absent from the homelessness count for exactly this reason. When you see an improving indicator, it is worth asking whether the definition changed along the way.

Scientific background
Homelessness — structural determinants and Housing First evidence
Research provenance
This page comes out of research at the EQUORA Institute and captures one state of that work rather than a settled institutional position. That state rests on the findings available at the time of publication; later findings appear here only where the page has been updated, which the date shows. AI takes part throughout the research process as a thinking partner; responsibility for interpretation and publication remains human.
Published: 27 June 2026 · Last content update: 17 August 2026
Papp László · EQUORA InstituteHow We Research →

The research is free — and will stay free. If it matters to you, there's a way to say so.

Support the work →