Homelessness and policing: the criminalization debate

In the last fifteen years, the share of US cities with ordinances criminalizing public sleeping has roughly doubled. Anti-camping laws, anti-panhandling laws, anti-sitting laws, anti-loitering laws, and "civility" codes have proliferated. In June 2024, the US Supreme Court in Grants Pass v. Johnson ruled that such ordinances are constitutional even when no shelter is available. This is now the dominant US policy environment.

The question that matters for student research and policy work is whether enforcement reduces homelessness. The answer, based on roughly two decades of evaluation work, is unusually clear.

What enforcement does

There is consistent empirical evidence on several specific outcomes of criminalization-heavy policy:

• It does not reduce the homeless population. Aggregate counts in cities with intense enforcement do not differ meaningfully from comparable cities with less enforcement. People are displaced, not housed.
• It increases the cost of homelessness. Police time, court time, jail time, and ambulance time for the same population end up substantially higher in enforcement-heavy regimes.
• It increases mortality among the unsheltered. A 2023 University of Colorado study estimated that aggressive enforcement of camping bans was associated with elevated death rates, primarily through loss of medications, disruption of medical care, and exposure during forced relocations.
• It produces criminal records that make future housing harder. A misdemeanor or felony record (especially a felony) materially raises the threshold for getting an apartment, getting a job, and accessing many federal housing programs.
• It damages trust in outreach. Outreach workers report that after sweep waves, the unsheltered population becomes much harder to find and much less willing to engage with anyone wearing a uniform or perceived to be government-affiliated.

What enforcement doesn't do

Enforcement does not, by available measures:

• Reduce the underlying rate of homelessness in the affected city.
• Encourage shelter uptake, except in cases where the offered shelter exactly meets the person's needs (which is rare).
• Lower the visibility of homelessness for more than a few weeks at a time before encampments re-form, usually in a less visible location.

What does work

The cities that have substantially reduced visible homelessness — Houston, parts of Salt Lake City, Helsinki, Anchorage during a sustained period — share a different pattern. They:

1. Built a real housing-placement pipeline first (Houston added ~25,000 people to permanent housing between 2011 and 2022).
2. Did sustained voluntary outreach before any clearance.
3. Made specific, tailored housing offers — not a generic shelter referral — to each person at the encampment.
4. Cleared sites only after the residents had been housed, often with their voluntary participation.

In every case, the enforcement was not the cause of the reduction; the housing pipeline was.

Grants Pass v. Johnson: what changed in 2024

The 2024 Supreme Court ruling overturned the 2018 Martin v. Boise Ninth Circuit precedent that had held cities couldn't criminalize sleeping outside if no shelter was available. Grants Pass found such ordinances do not constitute "cruel and unusual punishment" under the Eighth Amendment.

What this changed:

• Legally: US cities are now free to enforce anti-camping ordinances regardless of shelter availability.
• Practically: Some cities responded by ramping up enforcement; others did not. Some explicitly committed to not changing their approach.
• Politically: The decision created cover for elected officials who wanted to authorize sweeps but had been blocked by Martin. Several state legislatures (Tennessee, Texas, Missouri) have since passed state-level criminalization statutes.

The ruling is about constitutional ceiling, not policy floor. A city can choose to use the legal latitude Grants Pass opened up, or it can decline to. Most of the question now is local political.

The policing reform side

Separate from the criminalization debate is the broader question of who responds to homelessness-related 911 calls. Several cities have adopted mental-health response teams that send mental-health clinicians and EMTs rather than police to calls involving unhoused people in non-violent crisis:

• CAHOOTS (Eugene, OR) — operating since 1989, often cited as the model. Responds to about 17% of all Eugene 911 calls without police involvement.
• STAR (Denver) — Support Team Assisted Response, launched 2020. Hundreds of thousands of dispatches since.
• B-HEARD (NYC) — Behavioral Health Emergency Assistance Response Division. Mixed results in early evaluations.

The evidence on these programs is mostly positive: they reduce arrests of mentally-ill homeless people, reduce ambulance and ER use, and free up police time for actual crime — at roughly one-tenth the cost of standard police response.

What students should know

If you're writing about homelessness and policing for a class:

• Distinguish "enforcement against homelessness as a policy lever" (e.g. anti-camping ordinances) from "policing reform as it intersects with homeless calls" (e.g. CAHOOTS-style alternative response). They are different debates with different evidence.
• Cite the National Coalition for the Homeless's Housing Not Handcuffs report series for criminalization data. Cite the Vera Institute and individual city audits for sweep-cost data.
• Don't assume that visibility = total homelessness. PIT counts capture single-night unsheltered totals; criminalization affects visibility much more than it affects the total.
• The Helsinki and Houston cases are useful precisely because they are not American exceptions — they are evidence that the alternative model works at scale when funded.

Related: Encampments and why sweeps don't work · What actually works: Housing First · Common myths.