Community Crisis Response as Constitutional Alternative

“…services, programs, and activities [provided] in the most integrated setting appropriate to the needs of qualified individuals with disabilities.” — the integration mandate, as construed in Olmstead v. L.C. (1999)

Start with the provocation, because it is where the thinking begins even if it is not where it can safely end. The claim, in its strong form, is that funding the police can increase crime rather than reduce it: every dollar into enforcement is a dollar not into the shelter and the recovery bed that would have prevented the encounter, and an officer with time on their hands generates work — an arrest that manufactures a record that manufactures the next arrest. It is a genuinely counterintuitive principle, and this essay takes it seriously enough to test it, which means conceding at the outset that in its strong, general form it does not survive contact with the evidence. What survives is narrower, better supported, and enough: for one specific category of emergency — the person in psychiatric crisis — the police-first response does not reduce harm. It converts a medical event into a criminal one, and there is already a body of law that names the conversion for what it is.

I. The Wager

The wager is that the mental-health crisis is the case where the abolitionist provocation and the disability-rights statute meet, and that the statute is the stronger ground. The essay does not need the general claim that police cause crime; it needs the specific, defensible claim that dispatching an armed officer to a psychiatric emergency is both empirically worse and legally vulnerable — worse because it criminalizes and endangers where a trained civilian would de-escalate, and vulnerable because the Americans with Disabilities Act, as construed in Olmstead, requires public services in “the most integrated setting appropriate,” and a jail cell is the least integrated setting there is. The series this essay belongs to is named Bandi Chhorr — the liberator of prisoners — and the argument is a Bandi Chhorr argument in the exact sense: it is about the machinery that puts people in cells who should never have been near one, and the door that law already leaves open.

II. The Provocation, Tested

Take the strong claim honestly. Does funding the police increase crime? The mainstream empirical literature says no in the aggregate: the best-identified studies in economics generally find that more police reduce serious crime, particularly violent crime, through deterrence and incapacitation. An essay that asserted the opposite as established fact would be doing exactly what this institute accuses its opponents of — selecting the science that flatters the conclusion. So the strong form is set aside, with the reason stated plainly.

But the provocation contains two mechanisms that survive the aggregate, and they matter. The first is budget substitution: public money is finite, and dollars committed to enforcement are, at the margin, dollars not committed to housing, treatment, and crisis care — the upstream interventions that prevent the situations enforcement is then dispatched to manage. This is not a claim that police cause crime; it is a claim about opportunity cost, and it is simply true of any fixed budget. The second is net-widening: when the only tool dispatched to a category of human trouble is one that arrests, the trouble gets processed as crime, because that is what the tool does. A man shouting on a corner in the grip of a psychotic break is, to a social worker, a medical emergency; to a patrol car, he is a potential arrest, and the patrol car is what was sent. The crime rate in that category is partly an artifact of what we chose to send. Neither mechanism rescues the strong claim. Both are enough to build the narrow one.

III. The Category That Breaks the Aggregate

The aggregate hides the category. Somewhere between a fifth and a quarter of American police killings, by widely cited estimates, involve a person with mental illness, and a large share of officer time is spent on calls that are medical and social rather than criminal. This is the category where the police-first default is not merely suboptimal but actively counterproductive: the armed, uniformed, time-pressured responder is close to the worst possible first contact for a person whose crisis is fear, disorganization, or delusion, and the encounters end, with grim regularity, in an arrest or a body. The Sheehan case is the emblem: officers called to help transport a woman in psychiatric crisis from her group home broke down her door, pepper-sprayed her, and shot her. The tool that was sent did what the tool does.

Against that default sits a documented alternative. CAHOOTS — Crisis Assistance Helping Out On The Streets — has run in Eugene, Oregon since 1989, pairing a medic with a crisis worker and dispatching them, unarmed, to exactly these calls. In 2019 CAHOOTS teams answered on the order of 24,000 calls and requested police backup roughly 250 times — a rate that tells you how few of these situations ever needed a badge in the first place. Denver’s STAR program, launched in 2020 on the CAHOOTS model, rerouted non-criminal 911 calls to unarmed responders, and its evaluation found reductions in low-level crime in the served area — the opposite of what the “less policing means more crime” intuition predicts, in precisely the category where the intuition fails. The alternative is not theoretical. It exists, it works, and where it has been tried it has freed the police to do the thing only police can do, while doing better the thing police were never suited for.

IV. The Statute Already Written

The constitutional argument is the essay’s spine, and it does not depend on the criminology at all. Olmstead v. L.C. (1999) held that Title II of the Americans with Disabilities Act carries an “integration mandate”: public entities must provide services in the most integrated setting appropriate to a person’s needs, and unjustified segregation of people with disabilities is itself a form of discrimination. The case was about institutionalization — warehousing in a hospital what could be served in the community — but the principle is general, and it has a sharp edge when turned on crisis response. If a person in psychiatric crisis can be safely and appropriately served by a community mobile-crisis team, then defaulting to the response that ends in a cell or a hospital hold is the provision of a service in the least integrated setting available. Where the integrated alternative exists and is not used, Olmstead’s logic says that is not merely bad policy. It is discrimination.

The doctrine is not yet nailed down, and honesty requires saying where it is open. Whether Title II reaches the arrest itself — the street encounter, not just the placement — is a question the Supreme Court took up in Sheehan and then declined to answer, resolving the case on qualified immunity and leaving the ADA question expressly undecided; the Ninth Circuit has held that Title II can apply to arrests, and other circuits have divided. So the argument here is not that the case law already commands community crisis response. It is that the integration mandate points unmistakably in that direction, that the doctrinal materials to get there exist, and that the failure to use a functioning, safer, more integrated alternative is exactly the kind of unjustified default Olmstead was written to reach. The law is not finished. It is aimed.

V. Why the Alternative Is Not Merely Kinder

The point that ties the provocation to the statute is that community crisis response is not offered here as compassion. It is offered as the intervention that avoids manufacturing the crime. The net-widening mechanism from Section II is the hinge: send the tool that arrests, and you generate the arrest; send the tool that de-escalates, and the same human event resolves without a criminal record attached to a person whose only offense was a medical emergency. The CAHOOTS ratio — hundreds of calls handled for every one that needed police — is the empirical measure of how much criminalization the police-first default was creating out of nothing. This is where VK’s provocation is right, precisely scoped: in this category, the enforcement response does create work for itself, not through malice but through the structural fact that its only verb is to arrest. Fund only that response, and you have funded the manufacture of exactly the records the Bandi Chhorr series exists to trace. The alternative breaks the loop not by being nicer but by removing the tool whose use was the harm.

Counter-case

The strongest objection comes from both flanks at once. From the empirical side: the aggregate deterrence literature is robust, communities that have cut police have sometimes seen violence rise, and generalizing from CAHOOTS — a mid-sized, relatively homogeneous city — to dense, violent urban cores is exactly the overreach the essay accuses others of. From the legal side: Olmstead is an institutionalization case, its extension to street encounters is contested and was pointedly left open in Sheehan, and building policy on a doctrine the Supreme Court declined to establish is advocacy dressed as analysis.

Both objections are conceded in part, and the concession is the argument’s discipline. The essay does not claim police reduce nothing or that the strong “funding police causes crime” thesis holds; it explicitly sets that aside. It does not claim Olmstead already commands crisis response; it says the mandate points there and the doctrine is unfinished. What the objections cannot touch is the scoped claim: that for the psychiatric-crisis category specifically, a functioning unarmed alternative exists, handles the overwhelming majority of calls without police, reduces rather than raises low-level crime where measured, and is the more integrated setting the ADA’s mandate prefers. The generalization the essay refuses to make — from this category to all policing — is precisely the generalization it accuses the provocation of over-making. The narrow claim is what is left when the honest deletions are done, and it is enough to carry the constitutional argument.

Stakes

The Bandi Chhorr series reads the American carceral machine the way the tradition reads a prison: as a set of doors, some of them openable by an argument the law already contains. This essay finds one such door. It does not need the sweeping claim that the police cause the crime they are paid to fight; it needs the exact claim that in the crisis category, the police-first default manufactures criminalization a community response avoids, and that Olmstead’s integration mandate gives that failure a legal name. The provocation was the way in — counterintuitive, overstated, and productive, the kind of claim that is wrong in general and load-bearing in particular. Scoped to where it is true, it becomes a constitutional argument for building the alternative and a diagnosis of why the default persists: not because it works, but because it is the only tool we funded, and a system with one tool will find, in every kind of human trouble, another nail.

---

Sources

• Olmstead v. L.C., 527 U.S. 581 (1999) — the ADA Title II integration mandate (“most integrated setting appropriate”).
• City and County of San Francisco v. Sheehan, 575 U.S. 600 (2015) — decided on qualified immunity; the question whether Title II applies to arrests was left undecided. Ninth Circuit precedent applying Title II to arrests.
• CAHOOTS (White Bird Clinic, Eugene, Oregon, since 1989); 2019 call-volume and police-backup figures. Denver STAR (2020), modeled on CAHOOTS, and its published evaluation finding reduced low-level crime.
• On the deterrence literature (more police reduce serious crime in the aggregate): standard findings in the economics of crime — cited as the counter-case the essay concedes.
• Companion entries in the series: the institute’s Bandi Chhorr research agenda (Legal Theory).