WE CAN AFFORD A JAIL CELL. CAN WE AFFORD A FRONT DOOR?

VAULT · JUSTICE AGAINST THE EPSTEINERS

Published by | Langhorne P. Breedlove

MONDAY, NOVEMBER 9, 2026 · 2:22 AM ET

Tags: Incarceration Costs · Housing · Jail Diversion · Public Safety · Homelessness · Fiscal Accountability

Dear fellow American proletariat: Suppose a person cannot afford a room. The county may have no available housing assistance, a treatment appointment scheduled six weeks from now, and a helpful brochure about personal responsibility. But should that person be arrested, a building with a staffed bed, three meals and a taxpayer-funded booking process suddenly becomes available. Apparently the government can find a place for someone to sleep as long as the door locks from the outside.

The title asks an uncomfortable question, not an accounting conclusion. Housing and incarceration are not equivalent services, and the people occupying their beds are not interchangeable populations. Correctional budgets include security, medical care, personnel, facilities and other duties. Housing budgets include rent, construction, maintenance and sometimes substantial supportive services. To compare them fairly, we need matching years, locations, populations, scope and real marginal costs.

ROOTS · TWO INSTITUTIONS BUILT FOR DIFFERENT PURPOSES

American prisons and local jails are institutions of the criminal justice system; local jails often hold people awaiting court action or serving shorter sentences, while state prisons generally serve people convicted under state law. Public housing and homelessness assistance arise from different legal mandates, serving people based on housing need and program eligibility. The existence of a jail cell does not establish that housing rather than detention would be lawful for its occupant.

The modern Housing First model prioritizes access to stable housing without requiring prior abstinence or completion of treatment, while offering services intended to sustain tenancy. It is distinct from simply handing somebody an apartment key and walking away. The City and County of Denver launched a supportive-housing initiative in 2016 to test whether housing plus services could interrupt repeated cycles of homelessness, jail stays and emergency care. (Source 02)

A federal constitutional distinction matters as well. The Supreme Court's 2024 decision in City of Grants Pass v. Johnson held that the Eighth Amendment does not categorically bar enforcing generally applicable public-camping laws against people experiencing homelessness. It did not establish that homelessness must be criminalized, nor did it resolve every other statutory or constitutional challenge to a particular enforcement practice. (Source 07)

TODAY · WHAT THE PUBLIC ACTUALLY SPENDS

In a 2025 analysis of state corrections and prison population data for 2023, USAFacts reported approximately $63.6 billion in combined state corrections spending and a median expenditure of $60,989 per prisoner among states. The reported state figures ranged from under $20,000 in Mississippi to $284,976 in Massachusetts. Six states' reporting combined jail and prison populations, so the measure is not a standardized national price for one year in prison. (Source 01)

The Bureau of Justice Statistics reported 658,700 people in local jails at midyear 2024 and roughly 7.9 million admissions over July 2023–June 2024. Of those held at midyear, 68% had not been convicted of their current charge. Unconvicted is not synonymous with harmless, improperly detained or legally eligible for immediate release. (Source 03)

Housing costs face different measurement problems. HUD counted 745,652 people experiencing homelessness on one night in January 2025, including 266,320 in unsheltered settings. The one-night estimate does not represent every person who experienced homelessness during the year, and it cannot be matched directly against a jail census from a different month and year. (Source 04)

The comparison often circulated online is a prison's annual cost per person versus the monthly rent of a local apartment. That is like comparing the operating cost of a hospital with the price of a mattress. It tells us the hospital is expensive, but not what it costs to treat a particular patient or what fraction of the budget disappears when that patient goes home.

THE DENVER EXPERIMENT · ACTUAL EVIDENCE, NOT WISHFUL ARITHMETIC

Denver's supportive-housing evaluation randomly assigned 724 eligible people with chronic homelessness and frequent justice-system interactions to an offer of supportive housing or usual community services. Across three years, people referred to housing had an average of 38 fewer jail days, nearly two fewer jail stays, four fewer arrests and 560 additional days of housing assistance than those receiving usual care. (Source 02)

The investigators also reported that about half of the program's total annual per-person cost was offset by reduced use of other public services, notably jail, ambulance and emergency-department services. That does not mean the entire initiative became free. It means the net public cost was materially lower than its gross cost under that local program's assumptions and participating population. (Source 02)

A Los Angeles County evaluation of the Just in Reach supportive-housing program found significantly reduced jail use after housing placement for eligible justice-involved people with complex health needs. Researchers estimated a first-year net program cost of $6,202 per participant, but with a wide 95% confidence interval from a net saving of $3,543 to a net cost of $15,947. In other words, a precise national savings promise would exceed what even that local study could establish confidently. (Source 05)

WHY THE COST COMPARISON CAN MISLEAD

01 · AVERAGE IS NOT MARGINAL

A prison's average annual expenditure divided by its population includes salaries, buildings, contracts and costs that may remain even when one person leaves. Avoiding one jail day may reduce some meals, medical services or overtime, but may not permit closing a housing unit or reducing fixed staffing. Large sustained declines could change the cost structure. A credible budget should model the actual unit, shift or facility affected.

02 · JAIL IS NOT PRISON

A local jail detention is not the same as a state prison sentence. They have different populations, costs, jurisdictions and lengths of stay. Using the USAFacts state corrections median to value Denver jail days would combine unlike systems and produce a false estimate. (Sources 01, 02, 03)

03 · HOUSING IS NOT JUST RENT

For some people, a safe apartment and modest rental assistance may be enough. Others need ongoing clinical care, case management, disability accommodations, crisis response or intensive support. Construction, acquisition, utilities and upkeep also matter. Subtracting a market rent quote from an annual prison cost ignores both service delivery and which government actually pays the bill.

04 · SAVINGS TO WHOM?

A city might pay for supportive services while a federal voucher pays rent and the county captures reduced jail costs. Medicaid, emergency departments, state corrections and hospitals may experience different effects. A program can reduce overall public-system costs while increasing the particular city department's appropriation. Fiscal claims must identify the payer and recipient of each offset.

PUBLIC SAFETY, RIGHTS & QUALIFYING DIVERSION

Housing can stabilize an individual without eliminating criminal accountability. Eligibility for diversion or community supervision must consider the alleged offense, statutory authority, credible danger, victim protections, court orders and the person's treatment needs. A person requiring secure custody is not automatically a housing-program candidate. Nor should poverty or disability become a substitute for evidence of criminal wrongdoing.

A well-designed diversion program should preserve defense counsel, notice and appeals, judicial review where required, and individualized risk assessments that can be independently evaluated. Treatment should be clinically appropriate, with safeguards against coercive or unnecessary conditions. Housing should remain an option for people who were never arrested; waiting for a booking number is a spectacularly inefficient eligibility rule.

THE TEN-YEAR LEDGER · WHO SAVES WHAT?

The baseline begins with actual correctional staffing and facility budgets, expected jail days, state prison populations, community-supervision costs, homelessness assistance, housing subsidy and construction expenditures, and relevant Medicaid and crisis-treatment spending. For an honest comparison, select one jurisdiction and fiscal year, define a legally eligible cohort, then calculate both gross and marginal costs.

In year one, measure housing acquisition or rent assistance, assessment and outreach, clinical support, staff, evaluation, security, administration and court coordination. Over ten years, model housing retention, expected jail days averted, facility and staffing changes that can actually occur, medical and shelter utilization, inflation, provider shortages and possible public-safety consequences. Count savings only once and distinguish taxpayer outlays from benefits to participants.

Illustrative arithmetic can clarify but cannot substitute for local cost records: if a cohort of 100 eligible people avoids 30 jail days each, that is 3,000 jail days avoided. At a hypothetical avoidable cost of $40 per day, the directly avoidable jail expense would be $120,000—not 100 times the average annual state-prison spending. Both the 30-day change and $40 marginal rate in this example are assumptions, not results from a particular jurisdiction.

Compare continued standard services, targeted diversion and supportive housing, and expanded permanent housing with integrated care. Publish first-year gross cost, true offset, net fiscal effect and separate outcomes for safety and wellbeing. The proposed AP program has no defensible national ten-year savings total until eligible population, per-person services, jurisdictional costs and implementation schedule are specified.

WHAT COULD GO WRONG?

Housing shortages can prevent referrals from becoming placements. Insufficient support workers can undermine stability. Poor screening may expose residents or victims to risks, while excessively restrictive conditions may exclude people who could benefit safely. Expanding supervision instead of replacing confinement can increase the number of people entangled in the justice system.

Critics of Housing First question whether a low-barrier approach can sustain recovery and public order without stronger treatment expectations. Supporters point to housing retention and reduced justice-system use in trials such as Denver's. HUD's 2026 public interpretation of national homelessness trends criticized Housing First, while independent local evaluations identify positive outcomes in specific populations. National trend data alone cannot isolate the effect of one policy across changing rents, migration, housing supply and economic conditions. (Sources 02, 04)

The responsible approach is to evaluate comparative safety, housing stability, participant choice and real costs. It must permit secure custody when legally necessary and professional care when housing alone is not enough. The point of evidence is to find where an intervention works, not to declare that one slogan wins every case.

WHY IT MATTERS

The question is not whether an apartment is cheaper than a prison in some abstract universe. It is whether investing in safe housing and appropriate services can prevent a particular person's unnecessary detention, repeated crisis use and worsening health at a lower social and fiscal cost than the available alternatives.

The Denver randomized trial offers measurable grounds for further investigation; its local sample and net costs place limits on what can be claimed nationally. The better standard is simple: protect the public, preserve rights, measure results, and stop assuming that a locked door is always the least expensive way to handle somebody else's hardship. (Source 02)

SOURCES & FURTHER READING

01 · USAFacts — How Much Do States Spend on Housing Prisoners?, updated October 31, 2025; state corrections spending for 2023. https://usafacts.org/articles/how-much-do-states-spend-on-prisons/

02 · Urban Institute — What We Learned from the Denver Supportive Housing Social Impact Bond Evaluation, updated November 22, 2023. https://www.urban.org/policy-centers/metropolitan-housing-and-communities-policy-center/projects/denver-supportive-housing-social-impact-bond-initiative/what-we-learned-evaluation

03 · Bureau of Justice Statistics — Jail Inmates in 2024: Statistical Tables, September 8, 2026. https://bjs.ojp.gov/library/publications/jail-inmates-2024-statistical-tables/web-report

04 · U.S. Department of Housing and Urban Development — 2025 Annual Homelessness Assessment Report, May 29, 2026. https://www.hud.gov/news/hud-no-26-037

05 · RAND-affiliated study — Just in Reach Pay for Success: Impact Evaluation and Cost Analysis of a Permanent Supportive Housing Program, 2023. https://pmc.ncbi.nlm.nih.gov/articles/PMC10187556/

06 · Urban Institute — Costs and Offsets of Providing Supportive Housing to Break the Homelessness–Jail Cycle, July 15, 2021. https://www.urban.org/research/publication/costs-and-offsets-providing-supportive-housing-break-homelessness-jail-cycle

07 · U.S. Supreme Court / Cornell Legal Information Institute — City of Grants Pass v. Johnson, June 28, 2024. https://www.law.cornell.edu/supremecourt/text/23-175

10 RECOMMENDED ARTICLES, ESSAYS & POLICIES

→ Explore Housing & Rehabilitation Before Incarceration Where Appropriate

→ Explore Safe, Dignified Public Housing

→ Explore Abolishing For-Profit Prisons & Rehabilitating Inmates

→ Explore Ferries, Locks & Public Waterways

→ Explore Redesign the Faces of American Currency

→ Explore Blue Dogs

→ Explore Libertarians

→ Explore Reformers, Abolitionists & Civil Rights

→ Explore Elite & Industrial Power

→ Explore Public Banking & Postal Financial Services

WE ARE ONE.  ·  E PLURIBUS UNUM.