The Austerity Myth · Investigation · 2014–2021
It costs more to jail someone than to house them.
Supportive housing keeps people housed and, for the expensive high-need cases, costs less than the jail-and-ER status quo it replaces. The evidence is randomized, not anecdotal.
FACT
§2 · Thesis
A prison cell costs more than a housing unit — and the randomized evidence says housing people cuts the jail and emergency bill.
The expensive option is the one that fails: prison refills itself, while housing tends to end.
The number
$33,274
the average cost of one year in state prison (Vera), from $14,780 in Alabama to $69,355 in New York — more than supportive housing, which runs about $10,000 to $25,000.
Vera Institute, The Price of Prisons
§5 · Graded Claim
Housing First keeps people housed, and the proof is randomized trials — the $110M At Home/Chez Soi study across five cities and 2,000+ people, plus the 2018 National Academies consensus that the housing-stability evidence is strong.
FACT
The Mental Health Commission of Canada's At Home/Chez Soi RCT and the National Academies' 2018 review.
§5 · Graded Claim
For high-need people, supportive housing cuts jail stays, police contacts, and emergency care — the Denver five-year randomized trial of 361 chronically homeless people found roughly 34 to 40 percent fewer jail stays and police contacts.
FACT
The Urban Institute's 2021 randomized evaluation of the Denver Supportive Housing Social Impact Bond. It is the direct housing-versus-jail test.
§5 · Graded Claim
Prison is a compounding cost: about 68 percent of released state prisoners are rearrested within three years and 77 percent within five — partly because a record locks people out of jobs and apartments, so the system refills the cells it pays for.
FACT
Bureau of Justice Statistics recidivism data (2005–2014 cohort) and the recidivism literature on employment and housing barriers.
§5 · Graded Claim
The honest limit, kept on purpose: the savings are strongest for high-need, high-cost people, thinner at the population level, and for the most disabled the win is stability and avoided jail and ER costs, not a paycheck.
FACT
The 2018 National Academies review, which finds the cost-offset evidence strongest for high-utilizer groups; the taxpayer flip applies mainly to the non-chronic majority.
§6 · Record vs Narrative
Two ways to spend on the same person.
Jail them
- $33,274 a year on average, up to $69,355; city jails far higher.
- About 68% rearrested within three years.
- A multi-year cost that refills itself.
House them
- About $10,000 to $25,000 a year in supportive housing.
- Randomized-trial gains in housing stability, fewer jail and ER contacts.
- Usually time-limited; can end as people re-enter work.
▦ Still unknown
The Utah '91% reduction in chronic homelessness' headline is oversold.
This line is still blank
It leaned on counting-definition changes and later rebounded; the real case rests on the randomized trials, not that number.
Help us fill it →§7 · Why it matters now
Why it matters now — and where it connects.
The country spends more than $33,000 a year to hold one person in prison, and most are rearrested within five years, while housing costs less and tends to end. 'We can't afford to house people' chooses the more expensive path. That is the double standard the Austerity Myth hub keeps across every program.
▸ The Austerity Myth →
▦ Ledger gaps
Help us fill these lines.
- OpenThe per-person, city-by-city cost gap between a jail bed and a housing unit.
- OpenHow much of the housing cost is recovered as non-chronic residents re-enter work.
Help fill these →