dissenting.
This case raises the question whether the City of Milwaukee was required under Title II of the Americans with Disabilities Act (ADA), 42 U.S.C. § 12131 et seq., and the Rehabilitation Act, 29 U.S.C. § 794 et seq., to issue a special-use permit for the new outpatient mental health clinic that the plaintiffs (WCS) were trying to establish. The Milwaukee zoning code did not flatly prohibit this use, but it did require a special-use permit for properties located in the area where WCS wanted to locate. Concluding that the City was required as a matter of law to make reasonable accommodations to WCS’s disabled clients (and hence to WCS), and that WCS had shown that the special-use permit was such a *649reasonable accommodation, the district court granted summary judgment in WCS’s favor and ordered the City to issue the permit. The majority has vacated that order and remanded the case for further consideration of the question whether WCS can show that the City’s refusal to grant the permit stemmed from actual discrimination against mentally disabled users of WCS’s facility. I have no quarrel with the proposition that such a showing would entitle WCS to relief. In my view, however, further proceedings are unnecessary, because WCS has already successfully shown that it is entitled to the permit under the “reasonable accommodation” theory recognized in Title II of the ADA and the Rehabilitation Act. I would therefore affirm.
As the majority acknowledges, Title II of the ADA and the Rehabilitation Act prohibit intentional discrimination against qualified persons with disabilities, as well as measures that have a disparate impact on the disabled. (Henceforth I refer only to the ADA, because it is identical for all material purposes in this case to the Rehabilitation Act.) Where we disagree is over the extent to which Title II of the ADA also independently obliges covered persons to make reasonable accommodations for the disabled. The agency responsible for administering the ADA, the U.S. Department of Justice, has issued a regulation that unambiguously imposes such an obligation:
A public entity shall make reasonable modifications in policies, practices or procedures when the modifications are necessary to avoid discrimination on the basis of disability, unless the public entity can demonstrate that making the modifications would fundamentally alter the nature of the service, program or activity.
28 C.F.R. § 35.130(b)(7)(ADA). See also 28 C.F.R. § 41.53 (containing a similar provision implementing the Rehabilitation Act). That regulation says nothing about an antecedent need to prove pre-existing intentional discrimination or disparate impact. Instead, it authorizes prophylactic measures that are “necessary to avoid discrimination on the basis of disability,” unless the entity can satisfy the regulatory justification.
The most straightforward way to view this is as an affirmative obligation to take the steps that are necessary to bring the services available to the disabled up to the level that the nondisabled enjoy, which in one sense will require giving some benefits to the disabled that are unnecessary for their more fortunate fellow citizens. A person with no mobility problem would never miss a wheelchair ramp; a person with 20/20 vision has no need for an audible signal in an elevator that the desired floor has been reached. But those accommodations are essential for the disabled person to enjoy equal access to public services. The Supreme Court recognized this concept when, in US Airways, Inc. v. Barnett, 535 U.S. 391, 122 S.Ct. 1516, 152 L.Ed.2d 589 (2002), it acknowledged that reasonable accommodations “will sometimes require affirmative conduct to promote entry of disabled people into the work force.” Id. at 401, 122 S.Ct. 1516. Although Barnett held that in the normal run of cases an employer would not be required to override a bona fide seniority system in order to accommodate a disabled employee’s needs, it also held that the employee would be entitled to show that “special circumstances warrant a finding that, despite the presence of a seniority system ..., the requested ‘accommodation’ is ‘reasonable’ on the particular facts.” Id. at 405, 122 S.Ct. 1516. At least for purposes of Title I of the ADA, therefore, the accommodation requirement imposes spe*650cial responsibilities on employers that could not be derived from either the intentional discrimination theory or the disparate impact theory.
True, the majority says, but we are dealing with Title II of the statute, and it does not contain the same accommodation language as Title I. Literally speaking, that is true, but I do not understand the majority to be holding that the Department of Justice exceeded its authority when it issued the implementing regulations interpreting the requirements of Title II. To the contrary, it states that it is not taking issue with the regulation, 28 C.F.R. § 35.130(b)(7), in this case. If not, then Barnett’s approach to reasonable accommodation is binding on us here: the language of § 35.130(b)(7) is substantively identical to the language the Court was construing in Title I, 42 U.S.C. § 12112(b)(5)(A).
The majority avoids this outcome by looking to two cases this court has decided under the closely analogous Fair Housing Amendments Act (FHAA), 42 U.S.C. § 3601 eb seq.; namely, Hemisphere Bldg. Co. v. Richton Park, 171 F.3d 437 (7th Cir.1999), and Good Shepherd Manor Foundation, Inc. v. Momence, 323 F.3d 557 (7th Cir.2003). Those cases, it argues, demonstrate that there is actually no independent accommodation requirement under any of these statutes. Instead, the accommodation requirement has no independent force at all; it is merely a remedy for practices that have a disparate impact on the disabled. As the majority sees the world, “[i]n the absence of disparate impact ... there is no need for accommodation under the FHAA,” or by extension, under Title II of the ADA. Ante at 646. (It does not explain why the same logic would not apply to Title I of the ADA, except to acknowledge that Barnett forecloses this argument.)
The majority concedes that its position is inconsistent with language that appears in Good Shepherd, where the court wrote unequivocally that “ ‘[fjailure to reasonably accommodate’ is an alternative theory of liability.” Ante at 646; quoting 323 F.3d at 562. In an effort to squeeze Good Shepherd into its own theory of the FHAA and hence the ADA, it suggests that reasonable accommodation is an independent theory only insofar as it is “the means by which disparate impact is alleviated.” Ante at 646. But that is not what Good Shepherd said. There is a difference between a theory of liability and a remedy for a proven violation. The problem with the majority’s reasoning is reflected in its assumption that the disabled can suffer a deprivation of access to public services and programs only if (a) laws and regulations actively discriminate against them, or (b) laws and regulations produce disparate effects. Both of these theories, however, require a comparison between the effect of the practice on the favored group and the effect of the practice on the disfavored group — here, the disabled.
The problem is that there are many services and facilities that are of interest only to disabled people, such as the ramps and the audible elevator announcements mentioned earlier. In those situations, there would never be a way to prove either individual animus or disparate impact, unless the latter theory were applied far more broadly than it normally is. Recall that disparate impact theory was born in Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971), where the Court considered the question whether a high school diploma or GED requirement for hiring or transfer was having a disparate impact on minority candidates. Here, only the disabled would have any interest in the particular service or facility at issue; because the nondisabled are indifferent to *651it, there would never be a way to prove the disproportionate impact required by that theory. Moreover, the statistics needed to show disparate impact will often be unavailable to the disabled.
The approach I favor does not, contrary to the majority’s implication, require one to conclude that Good Shepherd silently overruled our earlier decision in Hemisphere Building. In Hemisphere Building, another FHAA case, a developer argued that he was entitled to both re-zoning and a special-use permit so that he could build two four-unit residences for wheelchair-bound persons. Without the re-zoning, he would have been free to build identical residences suitable for the same population, but only six homes instead of eight, and he would have had to charge a higher price for them. On these facts, we held that the Village was under no duty to change its zoning practices. The duty of reasonable accommodation had to be confined to rules, policies, practices or services that hurt the disabled “by reason of their handicap, rather than that hurt them solely by virtue of what they have in common with other people, such as a limited amount of money to spend on housing.” 171 F.3d at 440 (emphasis in original).
The district court’s decision here fits comfortably within the rule announced in Hemisphere Building and is equally consistent with the elaboration found in Good Shepherd. It focused on the rules, practices, etc., that hurt WCS and its clients because of their mental disabilities, as opposed to their lack of money, their physical appearance, or other characteristics that they share with many members of the general public. The desire for cheaper housing, which was what Hemisphere Building was all about in the end, is not limited to the disabled, and thus a limitation in the supply of cheap housing does not hurt the disabled by reason of their handicap. Mental health services, in contrast, are uniquely important for people with mental disabilities. If a city were to zone them entirely out of its territory, on the theory that mentally disabled people are unpleasant neighbors or visitors to commercial establishments, its action would be taken by reason of the disability and it would violate the ADA. What has happened here is something less than total exclusion, but that just means that there is a factual question to be explored. Here, the district court reasonably concluded that the exclusion was enough to trigger the duty to accommodate. Moreover, it is worth noting that the court also found that “the clinic would not conflict with City zoning” (309 F.Supp.2d 1096, 1106 (E.D.Wis.2004)); that special uses like the clinic are recognized as not inherently incompatible with the area (id.); and that not only did the zoning board have the power to make this accommodation, but that this was a power that “the Milwaukee BOZA [Board of Zoning Appeals] has exercised willingly on previous analogous occasions” (173 F.Supp.2d 842, 853 (E.D.Wis.2001)).
In summary, this court had it right in Good Shepherd when it identified three separate ways of proving a claim under the FHAA, and there is no reason to shrink that number down to two for a claim brought under Title II of the ADA and the Rehabilitation Act. The majority’s approach is inconsistent with the implementing regulations under Title II, and it risks having the unfortunate effect of barring the disabled from relief just when they need it most: when a public entity is failing to provide a service that only the disabled would need, under circumstances where intentional discrimination and disparate impact would be impossible to prove as a practical matter. The only way to avoid such a result would be to expand the understanding of disparate impact well be*652yond its historical boundaries. Perhaps that is what the majority contemplates for its remand. In my view, the district court’s understanding of the law was correct, and its findings of fact are supported by the record. I would affirm, and I therefore respectfully dissent.