dissenting:
I respectfully dissent.
Cash asserts a claim for failure to supervise. The analytical framework for such a claim is set out in Walker v. City of New York, 974 F.2d 293, 297-98 (2d Cir.1992). Cash concedes that Walker controls, arguing that “the present case meets the three requirements for liability this Court delineated in Walker.” Cash Br. at 37. Under Walker, the plaintiff must show: first, a policymaker knew to a “moral certainty” that an employee would confront a given situation; second, the situation presents the employee with a difficult choice or there was “a history of employees mishandling the situation”; third, the wrong choice by the employee would frequently “deprive citizens of constitutional rights.” Walker, 974 F.2d at 297-98.
Perhaps recognizing that there was no “history of employees mishandling the situation” at the Erie County Holding Center (“ECHC”), the majority sidesteps Walker by framing the issue as “the adequacy of defendants’ own actions to prevent sexual contact between guards and prisoners consistent with their affirmative duty to protect prisoners in their custody.” Op. at 336. However, that is a claim of failure to supervise — one that fails the stringent test set out in Walker.
In finding a basis for a jury verdict adverse to Erie County and the sheriff (whom the jury found was not even negligent), the majority opinion relies on notice that existing measures were insufficient, and the availability of a measure that would be more effective. I disagree on both scores. Taking its holdings together, the opinion can be read (and will be read) to impose strict liability on municipalities and policymakers for any incidents that arise in a prison.
*345The female plaintiff was raped by a male guard. The jury considered whether the County and the sheriff were deliberately indifferent to the risk that prisoners would be sexually assaulted by guards. But the opinion re-casts the relevant risk in general and expansive terms: the “risk of sexual exploitation posed by male deputies guarding female prisoners.”1 Op. at 335.
Nothing supports even this generalized risk other than the complaint of inmate Allen, three years earlier — a complaint that was investigated, but that yielded ambiguous conclusions establishing no more than that the inmate exhibited herself sexually to guards who did not report her, and that one or more guards gave her commissary items. One guard was found to have violated policy and was given a three-day suspension without pay. Thus the Alen complaint was not ignored: It provoked an investigation; and the investigation resulted in discipline. This is not deliberate indifference to sexual exploitation, and far less is it deliberate indifference to the risk of rape. In faulting the sheriff and County for handling the Alen complaint in a way that could evidence deliberate indifference to rape, the opinion elides critical particulars of that complaint:
Alen conceded that she knowingly lied when she alleged that: (1) she had sexual intercourse with the guard; (2) a female guard was complicit; (3) she had physical evidence — a condom — that could be linked to the guard. Joint Appendix at 126-27.
Alen had a history of threatening guards and making false allegations to “get even” with them for “not getting her what she wanted.” Id. at 115, 120. She racked up 28 violations of prison rules in the 30 days around the incident. Id. at 115.
The investigator found that Alen’s allegation was made at least partly out of self-interest: “Ms. Alen stated she had the [condom] and she wanted to know what was ‘in it’ for her”; and her statement to investigators was in exchange for the potential of a statement by the prison superintendent to the judge on her behalf. Id. at 118,128.
She conceded that she seduced the guard: “[H]e didn’t force hisself on me or notin’. An, I totally seduced him. I mean, I totally went out my way to get him.” Id. at 142. “Um, when I found out he was interested in me, I made advances at him.” Id. at 130.
That was her standard practice: “I do put on shows for the officers.... [I]f I see dey interested ... dats when I go [i]n for the kill. I be like, ‘Oh, you like what you see?’ An, I be like, ‘Well give me some cigarettes.’ ” Id. at 141.
There is no evidence of sexual misconduct by guards at the ECHO prior to Cash’s assault besides these questionable allegations from one inmate, three years earlier — at a facility through which 17,000 inmates pass every year, Trial Tr. at 674.
The majority opinion implicitly concedes that the Alen incident may have been insufficient to put the sheriff and the County on notice that Cash might be raped; so the opinion relies as well on evidence that the sheriff was aware of incidents at other New York correctional facilities. Op. at 336-37.
*346That is adding nothing to nothing. If the evidence in this case amounts to sufficient warning of a criminal sexual assault, then a supervisor or government is always on notice of the risk of sexual abuses in prisons, and will always be liable when, sooner or later, something bad happens. The majority opinion is thus unbounded: It combines an ever-present risk with an inferred “proactive responsibility,” Op. at 335, in a way that constitutes strict (and vicarious) liability. And nothing limits the opinion to conduct by guards, or to sexual conduct. Did a warden or sheriff, a guard or a County know that sometime in past years one inmate hit another? Or that a guard observed or tolerated sexual misconduct by an inmate and received insufficient discipline for failing to report it (or for a gift of candy)? Or that something like that happened someplace else in the state? If so, they could be held liable as well for every act of prisoner-on-prisoner violence or sexual misconduct (even rape). To hold a municipality and its policymaker liable in this way eviscerates the Supreme Court’s limitations on municipal and policymaker liability.
The measures taken by defendants to prevent sexual exploitation of inmates were stringent: a no-excuses policy that is integral to training, that is enforced by supervision, that is reinforce’d by threat of discipline in a written notice, and that was implemented by an investigation and discipline following the only relevant inmate grievance in prior years. Yet the majority opinion holds that the jury could find that the defendants knew that the measures taken to protect Cash were insufficient. Op. at 336-37. The majority opinion “must assume” (and I agree) that the only ground on which the jury could have found deliberate indifference is failure to implement a policy (urged by plaintiffs expert witness) that the ECHC should have had a policy altogether preventing “unmonitored one-on-one contact between male guards and female prisoners.” Op. at 338. The expert offered as good practice (in the opinion’s words) “to pair a female officer with a male officer whenever direct interaction with a female prisoner is required.” Op. at 331.
Among the absurdities here is that no guard can know when direct contact may become required; in prison, interventions are not always by appointment. And at the risk of being obvious, this policy would either impose enormous incremental costs or would halve the personnel available for supervision of the facility (and thereby increase the risk of prisoner-on-prisoner violence and abuse).
In any event, the risk associated with having men and women interact in a closed environment is bred in the bone; it means nothing to say that the prison authorities should anticipate it. Abating that risk is another matter. If the majority opinion is sound, the only effective solution would be to have no guards of the opposite sex in women’s or men’s prisons. The majority opinion does not take account of the considerable ramifications. Because male inmates greatly outnumber female inmates, the resulting curtailment of opportunity for female guards would likely trigger valid Title VII suits. People with known same-sex preferences may not be able to serve as guards in any prison. And in another sphere, since military officers are responsible for their subordinates, we could not have mixing of the sexes in the military, unless (I suppose) the officers are paired off.
* * *
Finally, the majority casually discards— in a footnote, without explanation, Op. at 341 n. 8 — the district court’s own conclusion that the jury charge was deficient and that a new trial was warranted. (The *347magistrate judge would have sua sponte ordered a new trial had he not awarded judgment as a matter of law to the County and the sheriff. See Decision & Order, Cash v. Cnty. of Erie, No. 04 Civ. 182, at 6 n. 5 (W.D.N.Y. Mar. 10, 2009).)
. This framing of the risk is incompatible with the concession in the majority opinion that '‘unmonitored one-on-one interactions between a guard and a prisoner of [a] different sex[ ] [are] not [themselves] unconstitutional,” and with the citation (with approval) of a case that ‘‘reject[s] [the] conclusion that every male guard is a risk to the bodily integrity of a female inmate whenever the two are left alone.” Op. at 336 (internal quotation marks omitted). No effort is made to resolve this inconsistency.