American Atheists, Inc. v. City of Detroit Downtown Development Authority

OPINION SUTTON, Circuit Judge. As part of its efforts to revitalize the local economy, the City of Detroit in the late 1990s built a new stadium, Comerica Park, for its professional baseball team (the Tigers) and a new stadium, Ford Field, for its football team (the Lions). Soon after completing the stadiums (stadia if you like), Detroit sought to feature them by putting in bids to host the Super Bowl for the National Football League, the All-Star game for Major League Baseball and most recently the Men’s Basketball Final Four for the National Collegiate Athletic Association. When the first of these bids succeeded— when the NFL in 2002 agreed to hold the Super Bowl at Ford Field in 2006 — the City began to prepare for the event. In 2003, it created a development program, empowered to reimburse up to 50% of the costs of refurbishing the exteriors of downtown buildings and parking lots. The program was limited to property in a discrete section of downtown Detroit but reached *282 out to all property in that area, including property owned by religious organizations. Three churches participated in the program: a Methodist church, a Baptist church and an Episcopal church. Of the $11.5 million allocated for completed and authorized projects, 6.4% (or about $737,000) went to these churches. The question at hand is whether the Establishment Clause of the United States Constitution, or its counterpart in the Michigan Constitution, prohibits the City from including religious organizations in the program. The lead plaintiff, American Atheists, maintains that they do, and it filed this lawsuit to enjoin the agency from making any grants to religious entities. The district court rejected these arguments in large part, and we reject them in full. Everyone agrees that the program allocates benefits to a broad spectrum of entities on a neutral basis, as the City awards grants without regard to the religious, non-religious or areligious nature of the entity. The facial neutrality of the program, everyone also agrees, does not mask an intent to advance religion: Detroit sought to fix up its downtown, not to establish a religion. And as will generally be the case when a governmental program allocates generally available benefits on a neutral basis and without a hidden agenda, this program does not have the impermissible effect of advancing religion in general or any one faith in particular. By endorsing all qualifying applicants, the program has endorsed none of them, and accordingly it has not run afoul of the federal or state religion clauses. I. A. An agency of the City of Detroit, the Downtown Development Authority was created in 1976 “to halt property value deterioration in the downtown business district, to eliminate the causes of such deterioration, and to promote economic growth in the downtown business district.” JA 841. In pursuing these goals, the agency may draft, propose and implement development plans for the economic growth of downtown Detroit, Mich. Comp. Laws § 125.1657(l)(c), (e)-(f), all of which must be approved by the Detroit City Council, and all of which must relate to a defined section of downtown Detroit, id. § 125.1653(4). In 2003, the agency proposed, and the City Council approved, the Lower Woodward Facade Improvement Program. One among many “initiatives” to revitalize downtown Detroit, the program was designed to enhance the visual appearance of the downtown area by “encourag[ing] improvements to building facades and upgrades to edges of surface parking lots.” JA 1115. In addition to preparing the city for the influx of visitors that would descend on Detroit for the 2006 NFL Super Bowl, and eventually the 2005 Major League Baseball All Star Game and the 2009 NCAA Men’s Basketball Final Four, the agency believed that the program would “increase the sense of security, community, and continuity for existing and future developments” and “increase revenue at existing downtown business locations.” JA 781. Under the program, the agency allocated reimbursement grants, funded by local property-tax revenues, to property owners and tenants who owned or leased property within the program area and who made approved renovations to their buildings and parking lots. Qualified applicants were reimbursed for 50% of the costs of the renovations, subject to caps of $150,000 per building and $30,000 per parking lot. *283 Anyone who owned or leased property within the program area — mainly a nine-block area between Campus Martius Park and Grand Circus Park — could apply for a grant, so long as the applicant was current on its state and local taxes and initially could fund the project on its own. Although the program limited applicants to one grant per building or parking lot, applicants could receive multiple grants if they owned or leased multiple properties within the area. The agency established guidelines for making the grants. As to buildings, applicants could receive grants for “[pjermanent physical improvements” to “building facades generally visible from a public right-of-way,” JA 1120, such as renovations to storefront windows, exterior-lighting fixtures, masonry, brickwork, awnings, shutters, exterior doors, windows and signs. As to parking lots, applicants could receive grants for installing “fencing, brick piers, new curbs, new approaches, shrubbery, irrigation, lighting and other upgrades along the street-side edges of parking lots.” JA 1121. In reviewing grant applications, the agency ensured that proposed projects conformed to the “[pjrogram’s design guidelines and objectives.” JA 1126. Applicants had to describe the project in detail, estimate the project’s cost and prove that they initially could finance the project in full. After the program manager verified that information, the applicant selected a firm to prepare a conceptual design of the project. Once the program manager, a design-review committee and the agency’s board of directors approved the design, the applicant and the agency executed a reimbursement contract, in which the applicant agreed to complete the project and pay all of the costs of doing so, and the agency agreed to reimburse 50% of the approved design and construction costs (up to the $150,000 and $30,000 limits). Applicants became eligible for reimbursement after they completed the construction and after the agency verified that the project was constructed “in accordance with approved design documents,” JA 1129, and the applicant had paid the contractor in full. B. All told, during the two and a half years of the program’s existence, the agency received 204 grant applications, and 189 of them met the program’s eligibility requirements. The agency approved 123 projects, and the applicants completed 91 of these projects, at a total cost to the agency of $11.5 million. Of these 91 projects, nine involved renovations to buildings and parking lots owned by three downtown Detroit churches. Central United Methodist Church received four grants: one for repairs to the exterior of the church (including renovation of the building facade, entry door, outdoor sign, steeple clock and light fixtures); another for similar repairs to the church’s activities building; and two more for improvements to the church’s parking lots. Second Baptist Church received four grants: one for repairs to the exterior of the church (including renovation of the brick facade, the glass covering some stained glass windows, the exterior doors, the outdoor sign and the building trim); two more for similar repairs to two other church-owned buildings; and a fourth for improvements to the church parking lot. St. John’s Episcopal Church received one grant for repairs to the exterior of the church and to the parking lot, including funds for the renovation of a stone wall, metal fence and entranceway canopies, installation of a new church sign and replacement of a protective glazing over some stained glass windows. *284 The remaining 82 projects covered a range of improvements to downtown Detroit, including grants to a music hall, a bank, a hotel, an opera house, a theater and an apartment building. The three churches received about $737,000 from the agency — what comes to 6.4% of the $11.5 million in reimbursements. C. American Atheists, a non-profit organization that denies the existence of a deity and that “is dedicated to the separation of church and state,” JA 913, along with one individual member of the organization, another individual and one professional corporation (collectively, “American Atheists”), filed this action against the Detroit Downtown Development Authority (“Detroit” or the “City”), challenging the program as applied to the three churches. Seeking declaratory and injunctive relief (as well as attorney’s fees), they claimed that the agency’s transfer of tax-generated funds to the churches violates the Establishment Clause of the Federal Constitution and its counterpart in the Michigan Constitution. The district court permitted St. John’s Episcopal Church to intervene as a defendant. After the parties filed cross motions for summary judgment, the district court upheld all of the City’s reimbursement grants to the churches, save for two of them: (1) the costs of improving outdoor signs at the three churches and (2) the costs of replacing the storm windows covering two stained-glass windows at St. John’s Episcopal Church. See Am. Atheists, Inc. v. City of Detroit Downtown Dev. Auth., 503 F.Supp.2d 845, 849 (E.D.Mich. 2007). The district court granted attorney’s fees to American Atheists. American Atheists appeals the district court’s decision to allow any of the church-related grants. The City and St. John’s cross-appeal the district court’s decision to strike the portion of the grants attributable to restoring the signs and to replacing the covers of two stained-glass windows. The City also cross-appeals the attorney’s fees award. II. Before addressing the parties’ competing positions about the constitutionality of this program, we must address two preliminary constitutional issues of a different sort: Does American Atheists have standing to bring this claim? And, even if the group properly filed this lawsuit in the first instance, is this dispute still a live case or controversy given that the agency paid the approved reimbursement grants to each of the churches after the district court’s decision? A. The City argues that we need not, indeed cannot, decide the merits of this dispute because American Atheists premises standing on its members’ status as taxpayers and because taxpayer standing is a long-disfavored basis for enlisting the federal courts to resolve constitutional challenges to the validity of state or federal programs. See Frothingham v. Mellon, 262 U.S. 447, 487-89, 43 S.Ct. 597, 67 L.Ed. 1078 (1923); see also DaimlerChrysler Corp. v. Cuno, 547 U.S. 332, 343-46, 126 S.Ct. 1854, 164 L.Ed.2d 589 (2006); Valley Forge Christian Coll. v. Ams. United for Separation of Church & State, Inc., 454 U.S. 464, 476-82, 102 S.Ct. 752, 70 L.Ed.2d 700 (1982). As a general rule, a taxpayer concerned that government officials have misspent his tax dollars does not have a personal stake in the dispute — any more than any other taxpayer does — and thus lacks the “concrete and particularized” injury that Article III requires. Cuno, 547 U.S. at 344, 126 S.Ct. 1854 (internal quotation marks omitted). *285 Yet that general rule gives way to an exception for municipal taxpayers. Since the Court first articulated the general prohibition against taxpayer standing in Frothingham, it has maintained that the “peculiar relation of the corporate taxpayer to the corporation” sets municipal taxpayers apart from their federal (and state) counterparts: Like a shareholder of a private corporation, a municipal taxpayer has an immediate interest in how the municipality spends resources that reflect his contributions. Frothingham, 262 U.S. at 487, 43 S.Ct. 597; see also Cuno, 547 U.S. at 349, 126 S.Ct. 1854; ASARCO, Inc. v. Kadish, 490 U.S. 605, 613, 109 S.Ct. 2037, 104 L.Ed.2d 696 (1989) (plurality opinion). In applying this distinction, we have held that, so long as the challenged government action involves the expenditure of municipal funds (or the loss of municipal revenue), Frothingham’s bar on taxpayer suits does not apply. See Hawley v. City of Cleveland, 773 F.2d 736, 741-42 (6th Cir.1985); see also Taub v. Kentucky, 842 F.2d 912, 917-18 (6th Cir.1988). Only if the challenged local governmental action involves neither an appropriation nor expenditure of city funds will the municipal taxpayer lack standing, for in that case he will have suffered no “direct dollars-and-cents injury.” Doremus v. Bd. of Educ. of Hawthorne, 342 U.S. 429, 433-35, 72 S.Ct. 394, 96 L.Ed. 475 (1952); compare Pelphrey v. Cobb County, 547 F.3d 1263, 1280-81 (11th Cir.2008) (municipal taxpayers had standing because the challenged practice involved the expenditure of public funds), Koenick v. Felton, 190 F.3d 259, 263 (4th Cir.1999) (same), United States v. City of New York, 972 F.2d 464, 470-71 (2d Cir.1992) (same), Cammack v. Waihee, 932 F.2d 765, 770-71 (9th Cir.1991) (same), D.C. Common Cause v. District of Columbia, 858 F.2d 1, 4-5, 8-9 (D.C.Cir.1988) (same), and Donnelly v. Lynch, 691 F.2d 1029, 1030-32 (1st Cir.1982) (same), rev’d on other grounds, 465 U.S. 668, 104 S.Ct. 1355, 79 L.Ed.2d 604 (1984), with ACLU-NJ v. Twp. of Wall, 246 F.3d 258, 262-64 (3d Cir.2001) (plaintiffs did not have municipal-taxpayer standing because they did not show the city spent any funds on the challenged activity), Doe v. Duncanville Indep. Sch. Dist., 70 F.3d 402, 408 (5th Cir.1995) (same), Foremaster v. City of St. George, 882 F.2d 1485, 1489 n. 5 (10th Cir.1989) (same), Pulido v. Bennett, 848 F.2d 880, 885-86 (8th Cir.1988) (same), modified on other grounds, 860 F.2d 296 (8th Cir.1988), and Freedom from Religion Found., Inc. v. Zielke, 845 F.2d 1463, 1470 (7th Cir.1988) (same). American Atheists satisfies these modest requirements. The City funded its grants through local property taxes, including taxes paid by at least one member of American Atheists. Because American Atheists challenges direct expenditures of tax dollars its members paid and because an association may premise standing on the interests of its individual members, see Friends of the Earth, Inc. v. Laidlaw Envtl. Servs., Inc., 528 U.S. 167, 181, 120 S.Ct. 693, 145 L.Ed.2d 610 (2000), the group clears Article Ill’s injury-in-fact hurdle. Not so, the City rejoins: Hein v. Freedom from Religion Foundation, Inc., 551 U.S. 587, 127 S.Ct. 2553, 168 L.Ed.2d 424 (2007), changes everything because it categorically barred taxpayer suits attacking executive-branch spending decisions that violate the Establishment Clause. Because the Downtown Development Authority (an executive-branch-like entity), not the Detroit City Council (a legislative-branch-like entity), awarded the grants to the churches, the City argues that Hein eliminates standing to pursue this lawsuit. Yet this argument overreads Hein. The decision did not erect a new barrier to *286 taxpayer suits; it marked the boundaries of an existing exception to the rule against federal and state taxpayer standing. Flast v. Cohen held that the general bar on taxpayer lawsuits did not apply to complaints that congressional appropriations violated the Establishment Clause. 392 U.S. 83, 102-04, 88 S.Ct. 1942, 20 L.Ed.2d 947 (1968). Hein in turn explained that the Flast exception does not extend to suits challenging executive-branch expenditures of unearmarked funds. See 127 S.Ct. at 2565-68 (plurality opinion); see also id. at 2573-74 (Scalia, J., concurring in the judgment). While Flast opened the federal-court door to suits attacking Congress’s decisions to spend money in violation of the Establishment Clause, it never applied (as Hein makes clear) to suits attacking decisions by the executive branch. Flast and Hein have no application here. As Frothingham itself explained, and as we have since held, the general rule against taxpayer standing does not extend to municipal-taxpayer suits. See 262 U.S. at 487, 43 S.Ct. 597; Hawley, 773 F.2d at 741-42. Flast’s exception to Frothing-ham— and thus Hein’s exception to Flast — does not affect American Atheists’ standing to challenge the municipal agency’s grants. Cf. Pelphrey, 547 F.3d at 1280-81. Hein’s potential application to state taxpayers, cf. Hinrichs v. Speaker of the House, 506 F.3d 584, 598-600 (7th Cir.2007), in the end makes no difference to a case filed by municipal taxpayers. Nor need we decide whether the city council’s actions approving the program would satisfy Hein’s specific-legislative-appropriation requirement if Flast applied in this setting. All that matters is that American Atheists alleges that Detroit — a city — misspent taxes its members have paid. The City may or may not be right that the trajectory of Hein places the municipal-taxpayer standing rule in jeopardy. And it may or may not be right that drawing a line for standing purposes between a citizen’s status as a federal/state taxpayer and a city taxpayer makes less sense today than it did when the Court decided Frothingham in 1923. Why, the City suggests, should Article III bar a taxpayer from challenging spending decisions made by his State, yet allow him to challenge those of the same State’s political subdivisions? Cf. City of New York, 972 F.2d at 470-71. Population statistics do not explain the distinction. Thirty-two cities, including Detroit, have populations larger than at least one State. And the nation’s largest municipality, New York City, has a population larger than those of 39 States. See Population Division, U.S. Census Bureau, Annual Estimates of the Resident Population for the United States, Regions, States & Puerto Rico (2007), available at http://www.census.gov/popest/states/NST-ann-est2007.html; Population Division, U.S. Census Bureau, Annual Estimates of the Population for Incorporated Places over 100,000 (2007), available at http://www.census.gov/popest/cities/SUBEST2007.html. Why residents of New York City necessarily have a greater stake in the way city officials spend their share of the city budget than North Dakotans have in the way state officials spend their share of the state budget may indeed be a head-scratcher. But the Court has drawn a line between “state” action by States and “state” action by cities before, and the municipal-taxpayer doctrine may simply be one more area reflecting this distinction. Compare, e.g., Mt. Healthy City Sch. Dist. Bd. of Educ. v. Doyle, 429 U.S. 274, 280, 97 S.Ct. 568, 50 L.Ed.2d 471 (1977) (Eleventh Amendment’s bar on suits “against one of the United States” does not apply “to counties and similar municipal corporations”), with, e.g., Avery v. Midland County, 390 U.S. 474, 479-80, 88 S.Ct. 1114, 20 *287 L.Ed.2d 45 (1968) (Fourteenth Amendment’s mandate that “no State ... deny to any person ... the equal protection of the laws” applies to local governments). Either way, the city-state distinction in the context of taxpayer standing is one that the Supreme Court has maintained for 86 years and one we have no authority to second guess. B. That one of American Atheists’ members is a City of Detroit taxpayer, then, allowed the group to satisfy Article Ill’s case-or-controversy requirement at the outset of this case. But does the group satisfy these requirements at the tail end of the case? After the district court issued its decision, the City paid most of the reimbursement grants to the three churches, and that, we are told, moots the case. By preventing the federal courts from resolving moot disputes, Article III ensures that courts respect the case-or-controversy requirement at “all stages of federal judicial proceedings, trial and appellate.” Lewis v. Cont’l Bank Corp., 494 U.S. 472, 477, 110 S.Ct. 1249, 108 L.Ed.2d 400 (1990); see Cuno, 547 U.S. at 352, 126 S.Ct. 1854. A federal court must dismiss a case as moot if “the issues presented are no longer ‘live’ or the parties lack a legally cognizable interest in the outcome.” Powell v. McCormack, 395 U.S. 486, 496, 89 S.Ct. 1944, 23 L.Ed.2d 491 (1969). But so long as the parties maintain “a personal stake in the outcome of the lawsuit,” Lewis, 494 U.S. at 478, 110 S.Ct. 1249 (internal quotation marks omitted), and we can afford some kind of “meaningful relief’ to a prevailing party, a case is not moot, Church of Scientology v. United States, 506 U.S. 9, 12, 113 S.Ct. 447, 121 L.Ed.2d 313 (1992). The remedy need not satisfy a claimant’s every expectation: “[T]he availability of a partial remedy is sufficient to prevent a case from being moot.” Calderon v. Moore, 518 U.S. 149, 150, 116 S.Ct. 2066, 135 L.Ed.2d 453 (1996) (per curiam) (internal quotation marks omitted). The first answer to the mootness question is that the City and St. John’s — parties in the case — remain unsatisfied with the district court’s judgment. Although the district court upheld most of the reimbursement grants to St. John’s, it did not uphold all of them, including one to upgrade a building sign and one to replace storm windows covering two stained-glass windows. Because the district court enjoined these grants, because the City has not reimbursed the church for them and because American Atheists maintains that the state and federal religion clauses bar the grants, St. John’s and the City retain a live dispute with American Atheists over the permissibility of these payments. That proves that one of the requests for relief in this case remains live — American Atheists’ efforts to enjoin additional payments to religious entities. But what about its other request for relief — to return the grants already made to the City’s fisc? Generally speaking, American Atheists must show that each request for relief is not moot. See Powell, 395 U.S. at 496 n. 8, 89 S.Ct. 1944; cf. Friends of the Earth, 528 U.S. at 185, 120 S.Ct. 693. That may or may not make a difference here because there is a necessary overlap between the merits issues implicated by each requested form of relief in this case. At any rate, did this second request for relief become moot when the City paid these other grants and when most of them went to entities that have never participated in this case — making it difficult, if not judicially impossible, to get the money back should American Atheists prevail on appeal? Cf. Laskowski v. Spellings, 443 F.3d 930, 933-35 (7th Cir.2006), vacated on *288 other grounds by Univ. of Notre Dame v. Laskowski, — U.S. —, 127 S.Ct. 3051, 168 L.Ed.2d 755 (2007); Guidry v. Sheet Metal Workers Int’l Ass’n, 10 F.3d 700, 705 (10th Cir.1993). Here again St. John’s enters the picture: As a party, it may be ordered to return the grants already made to it. Great-West Life & Annuity Ins. Co. v. Knudson, 534 U.S. 204, 213-14, 122 S.Ct. 708, 151 L.Ed.2d 635 (2002); United States v. Lane Labs-USA Inc., 427 F.3d 219, 225 (3d Cir.2005). If American Atheists prevails on appeal, it thus may obtain at least a “partial remedy” — the return of St. John’s grants — which suffices to prevent this aspect of the case from being moot. Church of Scientology, 506 U.S. at 13, 113 S.Ct. 447; see Ams. United for Separation of Church & State v. Prison Fellowship Ministries, Inc., 509 F.3d 406, 421 (8th Cir.2007) (“[T]he dispute as to restitution, premised on the unconstitutionality of the performed contracts, is a live controversy.”); In re Zenith Elecs. Corp., 329 F.3d 338, 340-41 n. 1 (3d Cir.2003); cf. Roemer v. Bd. of Pub. Works of Md., 426 U.S. 736, 767-68 n. 23, 96 S.Ct. 2337, 49 L.Ed.2d 179 (1976) (plurality opinion). Laskowski v. Spellings, 546 F.3d 822 (7th Cir.2008), we realize, held that federal taxpayers cannot pursue similar relief once funds have been disbursed to intermediary party recipients of the grants after they released the funds to non-party beneficiaries. Allowing a federal-taxpayer plaintiff to “proceed against a private grant recipient for restitution to the Treasury as a remedy in an otherwise moot Establishment Clause case,” the court reasoned, would “extend the Flast exception beyond the limits of the result in Flast.” Id. at 827. And because Hein “strictly confined” Flast “to the result in Flast,” the court concluded that Hein prohibited any such extension. Id. Laskowski does not affect this case for at least two reasons. American Atheists, to repeat, premises standing on its members’ status as municipal, not federal, taxpayers, and accordingly any order requiring St. John’s to repay the funds it received would not extend Flast — would not, indeed, even implicate Flast. And because Laskowski involved a party intermediary who had received federal funds, but not an end recipient of the funds who was a party to the case, the decision raised restitution questions (and restitution complexities) not presented here. No party to this case, notably, has questioned the federal district court’s authority to require St. John’s to return the funds it already has received for its qualifying and (thus far) court-approved reimbursement grants. The case is not moot. III. The First Amendment to the United States Constitution, as applied to the States through the Fourteenth Amendment, see Cantwell v. Connecticut, 310 U.S. 296, 303, 60 S.Ct. 900, 84 L.Ed. 1213 (1940), provides that “Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof,” U.S. Const, amend. I. From the outset, the Court has construed the give- and-take language of the two clauses to forbid government from using its power either to “favor” or to “handicap” any one religion or religion in general. Everson v. Bd. of Educ. of Ewing, 330 U.S. 1, 18, 67 S.Ct. 504, 91 L.Ed. 711 (1947). The Religion Clauses command the States to “be a neutral” in their “relations with groups of religious believers and non-believers” and not the “adversary” of either side. Id. To these ends, the Establishment Clause requires government to enact laws that are neutral as to religion, do not have the purpose of advancing religion and do not *289 have the primary effect of advancing religion. See Mitchell v. Helms, 530 U.S. 793, 838-39, 120 S.Ct. 2530, 147 L.Ed.2d 660 (2000) (O’Connor, J., concurring in the judgment); Aogstini v. Felton, 521 U.S. 203, 222-23, 233-35, 117 S.Ct. 1997, 138 L.Ed.2d 391 (1997); see also Lemon v. Kurtzman, 403 U.S. 602, 612-13, 91 S.Ct. 2105, 29 L.Ed.2d 745 (1971); cf. Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 531-34, 113 S.Ct. 2217, 124 L.Ed.2d 472 (1993). Detroit’s downtown refurbishment program satisfies all three requirements. A. The most essential hurdle that a government-aid program must clear is neutrality — that the program allocates benefits in an evenhanded manner to a broad and diverse spectrum of beneficiaries. See Good News Club v. Milford Cent. Sch., 533 U.S. 98, 114, 121 S.Ct. 2093, 150 L.Ed.2d 151 (2001); Mitchell, 530 U.S. at 809-13, 120 S.Ct. 2530 (plurality opinion); id. at 838, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment). Phrased as an interrogatory: Does the program determine a recipient’s eligibility for benefits in spite of, rather than because of, its religious character? See Mitchell, 530 U.S. at 809-10, 120 S.Ct. 2530 (plurality opinion); Rosenberger v. Rector & Visitors of Univ. of Va., 515 U.S. 819, 839-40, 115 S.Ct. 2510, 132 L.Ed.2d 700 (1995). Since its earliest explorations of the Establishment Clause, the Court has underscored neutrality as a central, though not dispositive, consideration in sizing up state-aid programs. See Mitchell, 530 U.S. at 809-10, 120 S.Ct. 2530 (plurality opinion); id. at 838, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment); Rosenberger, 515 U.S. at 839, 115 S.Ct. 2510; id. at 846, 115 S.Ct. 2510 (O’Connor, J., concurring); Everson, 330 U.S. at 17-18, 67 S.Ct. 504. What the Court has said matches what it has done. Programs that allocate benefits based on distinctions among religious, non-religious and areligious recipients are generally doomed from the start. See, e.g., Texas Monthly, Inc. v. Bullock, 489 U.S. 1, 14-15, 109 S.Ct. 890, 103 L.Ed.2d 1 (1989) (plurality opinion) (invalidating state sales-tax exemption “for periodicals published or distributed by a religious faith and consisting wholly of writings promulgating the teaching of the faith”); Larson v. Valente, 456 U.S. 228, 246-47 & n. 23, 255, 102 S.Ct. 1673, 72 L.Ed.2d 33 (1982) (striking down state law exempting only certain “well-established churches” from various registration and reporting requirements) (internal quotation marks omitted); Sch. Dist. of Abington Twp. v. Schempp, 374 U.S. 203, 205, 83 S.Ct. 1560, 10 L.Ed.2d 844 (1963) (invalidating programs mandating daily Bible reading in public school). Yet programs that evenhandedly allocate benefits to a broad class of groups, without regard to their religious beliefs, generally will withstand scrutiny. See Good News Club, 533 U.S. at 114, 121 S.Ct. 2093; Mitchell, 530 U.S. at 809-14, 120 S.Ct. 2530 (plurality opinion); Agostini, 521 U.S. at 230-31, 234-35, 117 S.Ct. 1997; Rosenberger, 515 U.S. at 840-43, 115 S.Ct. 2510; Zobrest v. Catalina Foothills Sch. Dist., 509 U.S. 1, 10, 113 S.Ct. 2462, 125 L.Ed.2d 1 (1993); Lamb’s Chapel v. Ctr. Moriches Union Free Sch. Dist., 508 U.S. 384, 395, 113 S.Ct. 2141, 124 L.Ed.2d 352 (1993); Witters v. Wash. Dep’t of Svcs. for the Blind, 474 U.S. 481, 487-88, 106 S.Ct. 748, 88 L.Ed.2d 846 (1986); Mueller v. Allen, 463 U.S. 388, 398-99, 103 S.Ct. 3062, 77 L.Ed.2d 721 (1983); Widmar v. Vincent, 454 U.S. 263, 273-75, 102 S.Ct. 269, 70 L.Ed.2d 440 (1981); Bd. of Educ. v. Allen, 392 U.S. 236, 238, 243-44, 88 S.Ct. 1923, 20 L.Ed.2d 1060 (1968); Everson, 330 U.S. at 17-18, 67 S.Ct. 504; see also Walz v. Tax *290 Comm’n, 397 U.S. 664, 672-73, 90 S.Ct. 1409, 25 L.Ed.2d 697 (1970). Detroit’s program satisfies this neutrality requirement. It makes grants available to a wide spectrum of religious, nonreligious and areligious groups alike and employs neutral, secular criteria to determine an applicant’s eligibility, what projects may be reimbursed and how much each grantee receives. That the program includes, rather than excludes, several churches among its many other recipients helps “ensure neutrality, not threaten it.” Good News Club, 533 U.S. at 114, 121 S.Ct. 2093; see also Rosenberger, 515 U.S. at 836, 839, 115 S.Ct. 2510; cf. Steele v. Indus. Dev’t Bd., 301 F.3d 401, 410-11, 414-15 (6th Cir.2002). B. Although the facial neutrality of a program frequently will provide refuge from attack in this area, it is not a safe haven. A government program, even an ostensibly neutral program, adopted with the aim of advancing one religion or all religions generally will not survive an Establishment Clause challenge. See Wallace v. Jaffree, 472 U.S. 38, 55-56, 105 S.Ct. 2479, 86 L.Ed.2d 29 (1985). Not just an overt objective of furthering religion will undermine a program, so too will other proof of an impermissible objective, including evidence gleaned from the language of the program, its context and its history. See Edwards v. Aguillard, 482 U.S. 578, 594-95, 107 S.Ct. 2573, 96 L.Ed.2d 510 (1987); see also McCreary County, Ky. v. ACLU of Ky., 545 U.S. 844, 863-66, 125 S.Ct. 2722, 162 L.Ed.2d 729 (2005); cf. Caleb Nelson, Judicial Review of Legislative Purpose, 83 N.Y.U. L.Rev. 1784, 1852-53 (2008). The implementation of a program also may reveal that what purports to be evenhanded is not. An aid program on its face may offer benefits to all comers but may in reality favor only religious groups — say, a program providing roof repairs only for buildings with steeples, or a program refurbishing large auditoriums in a neighborhood where the only buildings that fit the bill are houses of worship. In Committee for Public Education & Religious Liberty v. Nyquist, for example, the Court struck down a program that provided aid to private schools where “all or practically all” of those schools eligible to receive grants were “related to the Roman Catholic Church and [taught] religious doctrine to some degree.” 413 U.S. 756, 768 (1973) (internal quotation marks omitted); cf. Lukumi, 508 U.S. at 534-38, 113 S.Ct. 2217 (considering law’s “adverse impact,” which affected one religious group’s practices almost exclusively, as evidence of purpose to target the religion for detrimental treatment in violation of the Free Exercise Clause). None of these problems exists here. American Atheists has never alleged, much less demonstrated, that Detroit’s program has any such purpose — overt or masked. Nor does anything in the record suggest that the agency proposed, or the city council approved, the program with the aim of advancing the fortunes of the City’s houses of worship. What the record shows instead is that the program was designed to revitalize the downtown area across the board by giving all of its buildings (and parking lots) a facelift. Fostering that kind of improvement, the agency thought, not only would improve “the sense of security, community, and continuity for existing and future developments” but also would brighten the economic prospects of downtown businesses. JA 781; cf. Joseph L. Price, An American Apotheosis: Sports as Popular Religion, in Religion and Popular Culture in America 195 (Bruce David Forbes & Jeffrey H. Mahan eds., rev. ed. *291 2005) (“Fans, in fact, spend more money on the Super Bowl — making a pilgrimage to the game; attending parties bedecked with official Super Bowl paraphernalia; placing bets in office pools and elsewhere — than Americans spend on traditional religious practices and institutions throughout the entire month.”). The evenhanded language of the program, the religion-neutral purposes behind the program and the wide array of entities that obtained facade-improvement grants under it — secular and religious, public and private, for-profit and not-for-profit — foreclose any claim that the program was implemented with the purpose of advancing religion. C. That leaves the possibility that Detroit’s program, though written in evenhanded terms, though not masking a hidden agenda, nonetheless oversteps the Establishment Clause because in application it has the primary effect of advancing religion. See Agostini, 521 U.S. at 234-35, 117 S.Ct. 1997; Bowen v. Kendrick, 487 U.S. 589, 609, 108 S.Ct. 2562, 101 L.Ed.2d 520 (1988). The Court has identified several hallmarks of laws with this forbidden effect, but this program suffers from none of them. First, a program may have the primary effect of advancing religion if it employs skewed selection criteria that stack the deck in favor of groups that engage in religious indoctrination, encouraging potential recipients to take part in religious activity by rewarding them for doing so. Agostini, 521 U.S. at 231, 117 S.Ct. 1997. But because Detroit’s program makes aid “available to both religious and secular beneficiaries on a nondiscriminatory basis” and employs “neutral, secular criteria that neither favor nor disfavor religion,” it presents no such problem and creates no such incentive. Id.; see also Mitchell, 530 U.S. at 829-30, 120 S.Ct. 2530 (plurality opinion). Nor does American Atheists argue that the program grants administrative discretion to the agency that could be used — or has been used — to skew benefits in favor of religious entities. Second, a program may have the primary effect of advancing religion if it leads to “religious indoctrination” that “could reasonably be attributed to governmental action.” Mitchell, 530 U.S. at 809, 120 S.Ct. 2530 (plurality opinion); id. at 843, 860, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment); see also Agostini, 521 U.S. at 226, 230, 117 S.Ct. 1997. Although three churches received reimbursement grants under the' program and although the churches engage in what could be called (uncharitably) “religious indoctrination,” there is no cognizable basis for “attributing]” that activity to the City. By providing these benefits to religious and secular entities alike and by allocating the benefits based on criteria that have nothing to do with religion, the City demonstrated that it neither endorsed nor approved of the churches’ religious teachings. A comparison to other permitted benefits proves the point. A city may extend sewers and sidewalks to churches, synagogues and mosques. See Everson, 330 U.S. at 17-18, 67 S.Ct. 504. It may provide police and fire-protection services to them. See id. And it may extend property-tax exemptions for charitable organizations to them. See Walz, 397 U.S. at 680, 90 S.Ct. 1409. All of these generally available benefits facilitate the operation of the religious institution, either by saving them money directly or by sparing them the expense of providing the services on their own, and in the process they make it more likely that newcomers will attend their re *292 ligious services. Yet the Court has long approved the extension of these general government benefits to religious entities. The same is true of school-bus services, books and school technology offered to religious schools, even though the assistance facilitates the religious and secular missions of the recipients. See Everson, 330 U.S. at 17-18, 67 S.Ct. 504; Allen, 392 U.S. at 238, 244-48, 88 S.Ct. 1923; Mitchell, 530 U.S. at 809-14, 120 S.Ct. 2530 (plurality opinion); id. at 836-37, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment). If a city may save the exterior of a church from a fire, it is hard to understand why it cannot help that same church with peeling paint or tuckpointing — at least when it provides the same benefit to all downtown buildings on the same terms. No reasonable, reasonably informed observer, see Good News Club, 533 U.S. at 118, 121 S.Ct. 2093; see also Zelman v. Simmons-Harris, 536 U.S. 639, 655, 122 S.Ct. 2460, 153 L.Ed.2d 604 (2002), moreover, would infer from the churches’ participation in this program, alongside and on equal terms with dozens of secular entities, that the agency endorsed or approved of the churches’ religious views. The program’s breadth, evenhandedness and eminently secular objectives help to break the link between the government and religious indoctrination. See Mitchell, 530 U.S. at 809-10, 830-31, 120 S.Ct. 2530 (plurality opinion); id. at 838-39, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment). Excluding the churches from taking part in the program, by contrast, would send a far stronger message- — -a message not of endorsement but of disapproval. It may be that the First Amendment did not compel Detroit to include religious groups in this downtown-revitalization project, cf. Locke v. Davey, 540 U.S. 712, 719-21, 124 S.Ct. 1307, 158 L.Ed.2d 1 (2004), but neither did it prohibit Detroit from including these groups in the project — either to enhance the success of the program (by revitalizing all of downtown) or to avoid sending a message of hostility to people of faith, cf. Van Orden v. Perry, 545 U.S. 677, 699-700, 704, 125 S.Ct. 2854, 162 L.Ed.2d 607 (2005) (Breyer, J., concurring in the judgment). Third, a program may have the primary effect of advancing religion if the benefit itself has an inherently religious content. See Mitchell, 530 U.S. at 820-22, 831-32, 120 S.Ct. 2530 (plurality opinion). Governments may not dole out crosses or Torahs to their citizens, even if they give them to all citizens, without running into an Establishment Clause problem. Cf. id. at 834-35, 120 S.Ct. 2530; Allen, 392 U.S. at 244-45, 88 S.Ct. 1923. Yet no such aid was distributed here. In the first place, the government gave monetary grants, not religious symbols, to participating entities. In the second place, the district court determined, and American Atheists does not challenge on appeal, that the vast majority of the reimbursed repairs — the renovation of exterior lights, pieces of masonry and brickwork, outdoor planters, exterior doors, concrete ramps, entrance ways, overhangs, building trims, gutters, fencing, curbs, shrubbery and irrigation systems— lack any content at all, much less a religious content. The thrust of the program goes to facade, not to substance, to giving the exterior of the buildings a clean, up-to-date appearance. There is no such thing as a Potemkin church, synagogue or mosque. That leaves the two types of grants that the district court invalidated: the church signs and the storm windows covering St. John’s stained-glass windows. The church signs contain no religious content in and of themselves. The point of the signs for churches and non-churches alike is to allow the entities to say what they are. The *293 churches, no surprise, identified themselves as churches, but that fact has no more religious content than the government-owned, national-historic-site markers that adorn the front exterior of each church’s sanctuary and that identify them as churches. See JA 910-11. Nor is the sign at St. John’s, as the district court mistakenly concluded, decorated with “crosses.” Am. Atheists, 503 F.Supp.2d at 868. The record shows that St. John’s sign is ornamented with eight wrought-iron fleurs-de-lis. As St. John’s points out, and as American Atheists does not dispute, a fleur-de-lis is an “architectural flourish” that has no inherently religious content and a well-established secular significance. Webster’s Third New International Dictionary 869 (2002) (defining a “fleur-de-lis” as “the iris chosen for the royal emblem of France by Charles V” and a “device common in artistic design and heraldry”). Nor is there a constitutional impediment to the reimbursement for work done on the storm windows covering the two stained glass windows at St. John’s. The storm windows by themselves do not have religious content. And although replacing the old, yellowing storm windows made the stained glass (and the religious iconography on it) easier to see, that did not transform the content-free storm windows into religious artifacts any more than removing plywood covering the windows would have made the wood a religious symbol. Fourth, a program may have the primary effect of advancing religion if the recipient “divertfs]” secular aid to further its religious mission. Mitchell, 530 U.S. at 840-41, 857, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment); compare id. ivith id. at 832-33 & nn. 14-17, 120 S.Ct. 2530 (plurality opinion). But the record reveals no diversion that would prompt a reasonable observer to impute to Detroit the religious mission that the churches undertake. For the most part, there was no possibility of diversion: Most of the aid the churches received — cosmetic repairs to walls, doors, awnings and parking lots, as well as limited landscaping — does not have dual uses capable of diversion from a secular to a religious purpose. Unlike a teacher, a sign-language interpreter or even an overhead projector — all of which conceivably can be used to communicate secular and religious messages — a brick, gutter or bush (unless burning) cannot be coopted to convey a religious message. One form of aid under the program, we recognize, raises a more serious diversion question. Two of the outdoor signs at issue allow the churches to display customized messages: Central United Methodist’s sign includes a panel on which it may place letters to create a message, and St. John’s sign includes a digital display that the church may program to display a variety of messages. Even if we assume that Central United and St. John’s will use their signs to display religious messages — even if we assume that the signs are divertible and will be diverted, see id. at 840-41, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment) — that does not justify enjoining the grants. If a city subsidizes a medium of communication on religion-neutral terms to a wide spectrum of speakers, the Establishment Clause does not bar a private group from using a government-provided medium to espouse its own message, even a religious message. See Rosenberger, 515 U.S. at 842, 115 S.Ct. 2510; see also Good News Club, 533 U.S. at 114, 121 S.Ct. 2093; Lamb’s Chapel, 508 U.S. at 394, 113 S.Ct. 2141; Widmar, 454 U.S. at 264-65, 270-73, 102 S.Ct. 269. So long as Detroit provides the same support to every other speaker, no reasonable observer could attribute a religious message to the City any more than he could attribute messages conveyed *294 by other downtown signs to the City. See Good News Club, 533 U.S. at 113-14, 121 S.Ct. 2093; Rosenberger, 515 U.S. at 841-42, 115 S.Ct. 2510; Lamb’s Chapel, 508 U.S. at 395, 113 S.Ct. 2141; Widmar, 454 U.S. at 274, 102 S.Ct. 269. When the government endorses everything, it endorses nothing. In this instance, Detroit offered to reimburse any eligible applicants for the repair or replacement of their signs. The broad sweep of the program alleviates “any mistaken impression” that the City endorsed any one message and “respects the critical difference between government speech endorsing religion, which the Establishment Clause forbids, and private speech endorsing religion, which the Free Speech and Free Exercise Clauses protect.” Rosenberger, 515 U.S. at 841, 115 S.Ct. 2510 (internal quotation marks omitted). By repairing and replacing Central United’s and St. John’s signs, the City of Detroit no more endorsed their messages than it did the message of any other grant recipient, be it the Michigan Opera Theatre, the First Independence Bank or the Detroit Athletic Club Foundation. Just as a reasonable observer would not view the State as endorsing the message of a religious publication when the State provides the paper on which the message is printed, see id. at 841-43, 115 S.Ct. 2510, so too a reasonable observer would not view the City as endorsing the churches’ religious messages — or the bank’s promotion of certificates of deposit — when it partially reimburses them for the medium on which the messages appear. See Good News Club, 533 U.S. at 115, 121 S.Ct. 2093; Rosenberger, 515 U.S. at 841, 115 S.Ct. 2510. Any such diversion, moreover, makes up at most a de minimis part of the reimbursement grants distributed under the program. See Mitchell, 530 U.S. at 861-67, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment). The reimbursement grants distributed for St. John’s and Central United’s signs make up “less than one percent of the total allocation” of aid under the program. Id. at 866, 120 S.Ct. 2530 (O’Connor, J. concurring in the judgment). Such evidence is “too insignificant ... to affect the constitutional inquiry.” Id. at 867, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment). Fifth, a program may have the primary effect of advancing religion if it excessively entangles the government in religious affairs. See Agostini, 521 U.S. at 232-33, 117 S.Ct. 1997. Yet all agree that the agency’s monitoring extended only to ensuring that each applicant’s conceptual design conformed to the program’s guidelines and that the construction conformed to that design. The program did not require city officials to make judgments about the religious content of particular projects, nor did it impose a scheme of “comprehensive, discriminating, and continuing state surveillance” that required the city to stay enmeshed in the churches’ affairs on an ongoing basis. See Lemon, 403 U.S. at 619-20, 91 S.Ct. 2105. D. Insisting that the program nonetheless has the “primary effect” of advancing religion, American Atheists raises a series of questions about its operation. Yet the answers to none of these questions show that the program crosses this forbidden line. 1. Does it matter, first of all, that the churches received reimbursement from the City as a result of decisions by the agency rather than as a result of “independent choices of private individuals”? Zelman, 536 U.S. at 649, 122 S.Ct. 2460. No doubt, channeling aid through “independent private choice” — as opposed to apportioning *295 aid based on government decisions — is one way to break the “circuit between government and religion.” Id. at 652, 122 S.Ct. 2460; see also Rosenberger, 515 U.S. at 886, 115 S.Ct. 2510 (Souter, J., dissenting); cf. Teen Ranch, Inc. v. Udow, 479 F.3d 403, 408-09 (6th Cir.2007). When aid reaches a religious organization through the independent decisions of private recipients, even if those recipients redirect it to religious activities, “[n]o reasonable observer is likely to ... infer[ ] that the State itself is endorsing a religious practice or belief.” Witters, 474 U.S. at 493, 106 S.Ct. 748 (O’Connor, J. concurring in part and concurring in the judgment); see Zelman, 536 U.S. at 652-53, 122 S.Ct. 2460. Although private choice is one way to break the link between government and religion, it is not the only way. The Court has sustained a number of neutral aid programs that distributed aid directly to religious organizations — without filtering the aid through private choice — where the aid itself had no religious content and any actual diversion was de minimis. See, e.g., Mitchell, 530 U.S. at 842-844, 867, 120 S.Ct. 2530 (O’Connor, J. concurring in the judgment); Comm. for Pub. Educ. & Religious Liberty v. Regan, 444 U.S. 646, 656-59, 100 S.Ct. 840, 63 L.Ed.2d 94 (1980); Roemer, 426 U.S. at 736, 766-67, 96 S.Ct. 2337; id. at 768, 96 S.Ct. 2337 (White, J. concurring in the judgment); Hunt v. McNair, 413 U.S. 734, 736, 744, 93 S.Ct. 2868, 37 L.Ed.2d 923 (1973); Bradfield v. Roberts, 175 U.S. 291, 295, 298-300, 20 S.Ct. 121, 44 L.Ed. 168 (1899); see also Good News Club, 533 U.S. at 112-13, 121 S.Ct. 2093; Rosenberger, 515 U.S. at 842-45, 115 S.Ct. 2510. Here, the aid lacked religious content, was not actually diverted in a way that would lead a reasonable observer to impute religious endorsement to the City and was distributed on a neutral basis to a broad spectrum of recipients. The absence of private choice by itself does not undermine the program. 2. Does it matter that the City directly provide the aid to pervasively sectarian institutions or that the aid took the form of cash reimbursements? No on both fronts. In some settings, the Court has presumed that direct aid to a pervasively sectarian institution will inevitably result in government sponsored indoctrination because the institution is so “subsumed in the religious mission” that it is impossible to separate the religious from the secular and to channel aid only to the latter. Hunt, 413 U.S. at 743, 93 S.Ct. 2868; see also Meek v. Pittenger, 421 U.S. 349, 365-66, 95 S.Ct. 1753, 44 L.Ed.2d 217 (1975), overruled by Mitchell, 530 U.S. at 808, 120 S.Ct. 2530 (plurality opinion); id. at 837, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment). In other settings, the Court has not indulged that presumption: That an aid recipient is a church or other distinctly religious institution has not been controlling (or even perceived as relevant) in these other cases because other aspects of the program rebutted the presumption that any aid flowing to the institution invariably would result in proselytization. See Mitchell, 530 U.S. at 867, 120 S.Ct. 2530 (O’Connor, J. concurring in the judgment); Agostini, 521 U.S. at 234-35, 117 S.Ct. 1997; Rosenberger, 515 U.S. at 842-43, 115 S.Ct. 2510; Regan, 444 U.S. at 661-62, 100 S.Ct. 840. In recent years, six members of the Court have questioned whether the religious nature of an aid recipient is helpful, or even appropriate, in evaluating neutral government aid programs under the Establishment Clause. Pour justices in Mitchell rejected the notion that “the Establishment Clause requires the exclusion *296 of pervasively sectarian schools from otherwise permissible aid programs.” 530 U.S. at 829, 120 S.Ct. 2530 (plurality opinion). And although Justice O’Connor, joined by Justice Breyer, did not join the plurality’s opinion in Mitchell, they concluded that proof of actual diversion of government aid for religious indoctrination, not the possibility that aid will be diverted, was needed to invalidate a program. Id. at 857-58, 860, 120 S.Ct. 2530 (O’Connor, J. concurring in the judgment). That was because “presumptions of religious indoctrination,” which at most point to probabilities, are typically unhelpful and “normally inappropriate” in evaluating neutral government aid programs. Id. at 858, 120 S.Ct. 2530; cf. Colo. Christian Univ. v. Weaver, 534 F.3d 1245, 1251-52, 1258, 1261 (10th Cir.2008); Columbia Union Coll. v. Oliver, 254 F.3d 496, 502-04 (4th Cir.2001). Yet no majority opinion of the Court has overruled the pervasively sectarian doctrine, and that means we must follow it. See Agostini, 521 U.S. at 237, 117 S.Ct. 1997; Steele, 301 F.3d at 407-09. In considering whether the three religious entities at issue must be excluded from the program because they are “pervasively” religious — they are churches after all — we must ask whether the record rebuts the presumption they will use the aid to engage in proselytization. In this case, as in Mitchell, Rosenberger, Agostini and Regan, there is ample reason to think that the distributed aid did not (and will not) result in government-sponsored faith-based activities. In addition to the neutrality of the program, the breadth of beneficiaries and the secular nature of the aid provided — all relevant factors under these four cases — the mechanics of the program ensured that the aid would go just to the approved uses. Cf. Regan, 444 U.S. at 659-61, 100 S.Ct. 840. Not only did eligible recipients have to satisfy three levels of review verifying that the project conformed to the program’s guidelines and criteria, but they also had to finance 100% of the project on their own at the outset and 50% of the program by the end. Whatever “special dangers” exist when a city makes open-ended money grants to religious recipients, which may spend the funds on whatever they wish, Mitchell, 530 U.S. at 855, 120 S.Ct. 2530 (O’Connor, J., concurring in the judgment), those dangers do not exist when a city releases funds for completed work on approved projects that the recipient initially finances on its own. Nor does it matter that, by providing funds for facade improvements, the agency may have freed up the recipients’ resources for other activities. The Supreme Court has repeatedly rejected “the recurrent argument that all aid is forbidden because aid to one aspect of an institution frees it to spend its other resources on religious ends.” Hunt, 413 U.S. at 743, 93 S.Ct. 2868; see Roemer, 426 U.S. at 747 & n. 14, 96 S.Ct. 2337 (plurality opinion). On this record, indeed, the program may well have encouraged eligible participants, the churches included, to divert monetary resources away from their core missions (e.g., holding worship services, performing operas, selling certificates of deposit) because they could reasonably conclude that the city’s one-time offer was just that — a one-time offer that enticed them to commit resources to their buildings today that they might otherwise not have committed at all. Either way, these carefully regulated reimbursements did not have the primary effect of advancing religion. The form of the aid — cash reimbursements rather than in-kind painting and tuck-pointing services — also does not invalidate the program. See Regan, 444 U.S. at 657-59, 100 S.Ct. 840. When a city reim *297 burses a religious entity for specific items, what matters is not the form the reimbursement takes but the benefit it represents. See id. at 657, 100 S.Ct. 840. Just as a State may reimburse a religious school for the costs incurred in providing state tests in cash payments, id., so a city may do the same in covering half of the costs of a building-refurbishment program. A downtown church, like a theater or a bank, reaps the same reward from government-sponsored repairs whether the city performs the work itself, using its own employees and equipment, whether it sends an independent contractor to do the work or whether it reimburses the church after the fact. Id. at 658, 100 S.Ct. 840. “[D]rawing a constitutional distinction,” id., between city-provided services and city-reimbursed services also might create puzzling, if not perverse, incentives. Would we be better off if the City had sent government workers to each church, wearing uniforms identifying them as agents of the government? And how would the City implement such a program if it wished, as in this case, to perform just half of the work? Would one group do half of the work before the other? Or would they work together? This “solution” would not only be mightily inefficient, but it also would seem to create more perception risks than partially reimbursing each church after the fact for approved work performed by private contractors. Cf. Rosenberger, 515 U.S. at 844, 115 S.Ct. 2510. Nor, contrary to the suggestion of some of the amici does Madison’s Memorial and Remonstrance Against Religious As sessments support a bright-line rule banning any public assistance for religious buildings. See 2 Writings of James Madison 183 (Gaillard Hunt ed.1901); Everson, 330 U.S. at 28-63, 67 S.Ct. 504 (Rutledge, J., dissenting); see, e.g., Am. Jewish Comm. Br. at 9-10. The Memorial responded to Patrick Henry’s proposal to tax all property owners and channel the proceeds to the (Christian) denomination of each taxpayer’s choosing, and it required each recipient sect (with a few exceptions) to spend the money either to support a “Minister or Teacher of the Gospel” or to “provid[e] places of divine worship,” Everson, 330 U.S. at 74, 67 S.Ct. 504. The very purpose of the proposal was to promote religion, and to use the coercive taxing- and-spending power of the State in doing so. See Douglas Laycock, The Underlying Unity of Separation and Neutrality, 46 Emory L.J. 43, 49 (1997). Henry’s proposal thus was the antithesis of neutrality, and by any measure of constitutionality it would have failed: It was not neutral as between religion and irreligión; its express purpose was to advance religion; and its primary effect — its only effect— was to advance religion. Reliance on the Memorial here gives historical analogy a bad name. The historical question posed today is not whether Madison would have remonstrated against Detroit’s program — which is highly unlikely given that the Memorial underscored, rather than undermined, the importance of government neutrality toward religion. See Madison, supra, at 186-88. The better historical question is to imagine what Madison (and the ratifiers) would have thought of the claimants’ position in this case: That the Bill of Rights prevents Detroit from extending its downtown-refurbishment program to all buildings — whether owned by a bank, a theater, an opera, a group devoted to worshiping a deity, a group that denies the existence of a deity or a group agnostic about the existence of a deity. The Establishment Clause requires neutrality toward religion, not hostility. “[W]e must be careful, in protecting the citizens of [Detroit] against state-established churches, to be sure that we do not inadvertently prohibit *298 [Detroit] from extending its general state law benefits to all its citizens without regard to their religious belief.” Everson, 330 U.S. at 16, 67 S.Ct. 504. 3. Do Nyquist and Tilton require us to invalidate this program? No. Although these decisions invalidated portions of government programs providing aid to construct, repair or maintain the buildings of religious schools, see Nyquist, 413 U.S. at 774-80, 93 S.Ct. 2955; Tilton v. Richardson, 403 U.S. 672, 682-84, 91 S.Ct. 2091, 29 L.Ed.2d 790 (1971) (plurality opinion), they are at least two steps removed from this program. In Nyquist, the Court struck down a maintenance-and-repair grant program for secondary and primary schools in New York City. But “all or practically all” of the schools eligible for the grants were not merely religious; they also were from the same denomination, 413 U.S. at 768, 93 S.Ct. 2955, which itself suggested a forbidden purpose. An essential point of that program, the Court has since explained, was to provide support for religious schools. See Zelman, 536 U.S. at 661, 122 S.Ct. 2460; see also Mitchell, 530 U.S. at 819 n. 8, 120 S.Ct. 2530 (plurality opinion) (explaining that the narrow group of recipients in Nyquist gave rise to “serious concerns about whether the payments were truly neutral”). The same cannot be said of Detroit’s program: Only 6% of the aid distributed by the city went to reimbursement grants for religious organizations, and even that went to three distinct denominations. Nor was the Nyquist program a onetime grant limited to exterior, cosmetic repairs. It provided an array of ongoing basic services designed to sustain the schools’ operation: The program paid up to 50% of the cost of providing “heat, light, water, ventilation and sanitary facilities; cleaning, janitorial and custodial services; snow removal; necessary upkeep and renovation of buildings, grounds and equipment; fire and accident protection.” 413 U.S. at 763 (internal quotation marks omitted). Unlike the one-time surface-level improvements designed to spruce up downtown Detroit, the state program in Nyquist kept the lights on at each religious school. At stake in Tilton was a federal program that provided grants and loans to religious and non-religious colleges to construct academic buildings. 403 U.S. at 675, 91 S.Ct. 2091 (plurality opinion). Recipients pledged not to use the federally financed buildings for “sectarian instruction” or “religious worship” for 20 years, and the federal government retained an interest in the property to enforce the requirement. Id. at 676, 683, 91 S.Ct. 2091 (plurality opinion) (internal quotation marks omitted). The Court upheld grants made under the program to four “church-related” colleges, but it excised the statutory provision limiting the religious-use prohibition to 20 years. Id. at 684, 91 S.Ct. 2091 (plurality opinion). At the expiration of that period, the plurality reasoned, sectarian schools would effectively hold the keys to the buildings, paid for by the federal government, and could use the buildings for any purpose, sacred or secular, to hold seminars or to give sermons. See id. at 683, 91 S.Ct. 2091. The first takeaway from Tilton is that the Court upheld neutral one-time grants to construct buildings at religious and nonreligious colleges. The greater authority to pay for all of the costs of a new building for a religious entity would seem to include the lesser authority to reimburse half of the costs (up to $150,000) for refurbishing the exterior of an old building or parking lot for a religious entity. No doubt, the *299 Court’s invalidation of the 20-year-limita-tion on the government’s right to seek a full recovery of the costs of construction (in the event the college used the building for religious purposes) gives us pause, but it does not change matters. Detroit’s program did not construct the buildings by paying for them in full; it merely improved the exterior of the buildings and did so by paying just half of the approved costs. Even then, Detroit’s program applies to a far wider circle of beneficiaries— namely, anyone with a downtown building — not just educational buildings. These distinctions by themselves explain why the Constitution does not require Detroit to prevent any of the beneficiaries of the program from using their refurbished buildings (or parking lots) for religious purposes. Reading this aspect of Tilton as broadly as American Atheists reads it also would bring the decision into tension, if not outright conflict, with later cases. Since Til-ton, the Court repeatedly has held that the Establishment Clause does not require the government to exclude religious groups from participating in open-access programs that make state-owned buildings available to all comers, even if such groups use the property for “religious worship and religious discussion.” Widmar, 454 U.S. at 265, 270-75, 102 S.Ct. 269; see Good News Club, 533 U.S. at 113-14, 119; Lamb’s Chapel, 508 U.S. at 394-95, 113 S.Ct. 2141; see also Rosenberger, 515 U.S. at 839-46, 115 S.Ct. 2510. What mattered in those cases was not that religious activity took place in facilities that the State had built and paid to maintain, but that the government provided access to those facilities on equal terms to all, ensuring that whatever use the groups made of them could not be chalked up to the State. American Atheists, like the lower court in Widmar, is wrong to “derive[ ] the proposition that state funds may not be used to provide or maintain buildings used by religious organizations.” 454 U.S. at 272-73 n. 12, 102 S.Ct. 269. Tilton cannot “be read so broadly,” id., and indeed later eases, Widmar included, show that a public building’s complete ban on allowing the property to be used for religious purposes, while allowing it to be used for non-religious purposes, would itself likely be unconstitutional. We see no difference, and American Atheists has identified none, between the government paying some of the repair costs of a building it allows any group — • including religious groups — to use on a neutral and evenhanded basis, and the government paying out limited, specifically defined maintenance-and-repair grants to a broad class of citizens defined without reference to religion. In either case, the activities that a religious group — or any other group — engages in cannot be attributed to the State. If we construed Tilton (and for that matter Nyquist) as establishing a categorical no-aid rule to religious entities, we also would place a host of other programs at risk. If the Establishment Clause excludes all religious institutions from programs that support their physical buildings, that would mean, as American Atheists acknowledges, that the government could not preserve a number of national landmarks, such as Ebenezer Baptist Church in Atlanta or the Old North Church in Boston, both of which benefit from direct federal aid under the “Save America’s Treasures” program. See U.S. Br. at 2; Ira C. Lupu & Robert W. Tuttle, Historic Preservation Grants to Houses of Worship: A Case Study in the Survival of Separationism, 43 B.C. L.Rev. 1139, 1164-65 (2002). The same is true, as the district court pointed out, of government programs designed to provide one-time emergency assistance *300 through FEMA and other public agencies to churches devastated by natural disasters. Cf. Lupu & Tuttle, supra, at 1162-64. Detroit’s three downtown churches may not have been subjected to a natural disaster and they may not have the pedigree of Ebenezer Baptist Church or the Old North Church. Yet they are national historic sites, and while they may not be as well known as these other churches, “there are those who love [them]” no less, Daniel Webster, Peroration to the Darthmouth College Case, in 15 The Writings and, Speeches of Daniel Webster 9, 11 (1903). If the national government may regulate the exterior of these three churches due to their historical significance and if the City may apply generally applicable public health regulations to the exterior of these buildings due to public safety concerns, the City ought to be able to help fix up their exteriors through generally applicable, neutral aid programs. What the government may regulate, as a general rule, it presumptively ought to be able to assist. It would be strange to read the Religion Clauses to say that churches may be subjected to neutral and generally applicable laws, see Employment Div., Dep’t of Human Res. v. Smith, 494 U.S. 872, 877-82, 110 S.Ct. 1595, 108 L.Ed.2d 876 (1990), but may not receive neutral and generally applicable benefits. See Bd. of Educ. of Kiryas Joel Vill. Sch. Dist. v. Grumet, 512 U.S. 687, 696, 114 S.Ct. 2481, 129 L.Ed.2d 546 (1994) (respect for both Religion Clauses requires “a course of neutrality”) (internal quotation marks omitted); see also Lupu & Tuttle, supra, at 1152. 4. Will a government aid program that includes churches among its beneficiaries invariably create entanglement problems? No. As American Atheists and their amici see it, the prohibition on government aid that is inherently religious poses serious problems for a program that extends aid to repair the exteriors of religious buildings: For the only way to honor the ban on using the public fisc for uniquely religious symbols is to engage in a line-by-line inquiry into each expenditure to determine the religious or areligious nature of the refurbished item, an inquiry that necessarily slights the ban on excessive church- and-state entanglement. At one level, they argue, every feature of a church, synagogue or mosque, from its structure to its windows to the colors of its doors, conveys religious meaning, and courts are ill-suited to decide whether this fixed-up item or that one has sufficient religious meaning to violate the Establishment Clause. As we see it, however, this is simply another way of advocating an absolute no-aid rule and of overruling the endorsement test in the process. The Court has not adopted a no-aid rule, and the endorsement test requires us to make these very inquiries. Doubtless, an item-by-item review of specific aid projects may create hard cases. But this is not one of them. Reviewing the validity of a refurbished storm window and a sign — to say nothing of painting, tuck-pointing and parking-lot fixing — has not enmeshed us in any great controversies of religious symbolism, and neither American Atheists nor its amici argue otherwise — at least as to these projects. As to other cases that the future may bring, more difficult issues assuredly may arise. Time will tell. All we can say for now is that there occasionally is no way to avoid the “hard task of judging — sifting through the details” of each program to determine whether its application offends the First Amendment. Rosenberger, 515 U.S. at 847, 115 S.Ct. 2510 (O’Connor, J., *301 concurring). We have done that today, and if need be we can do it tomorrow. IV. Detroit’s program also does not violate the Michigan Constitution. Under the State’s Constitution, “[n]o person shall be compelled ... to contribute to the erection or support of any place of religious worship,” Mich. Const, art. I, § 4, and “[n]o money shall be appropriated or drawn from the treasury for the benefit of any religious sect,” id. The words of these guarantees and their national counterpart do not have a lot in common — either as a matter of text or history. The state and federal provisions share just a few words — “no,” “shall,” “of,” “the” and “religion/religious” — most of them free of content. And the Michigan guarantee grows out of the Blaine Amendments, the product of a mid-nineteenth century political movement with no roots in the Religion Clauses of the United States Constitution. See Mark Edward DeForrest, An Overview and Evaluation of State Blaine Amendments: Origins, Scope, and First Amendment Concerns, 26 Harv. J.L. & Pub. Pol’y 551, 556-76, 588-89 (2003). Notwithstanding these differences, the Michigan courts have construed the state and federal guarantees in the same way. See Scalise v. Boy Scouts of Am., 265 Mich.App. 1, 692 N.W.2d 858, 868 (2005); see also In re Legislature’s Request for an Opinion on the Constitutionality of Chapter 2 of Amendatory Act No. 100, 384 Mich. 82, 180 N.W.2d 265, 274 (1970). The same test for ascertaining the validity of a law under the Establishment Clause, the Michigan courts have held, governs the validity of challenges under the state constitution. See Scalise, 692 N.W.2d at 868; see also In re Legislature’s Request, 180 N.W.2d at 274. Because we have held that the challenged reimbursement grants do not violate the Establishment Clause, and because the Michigan courts have the final say over what their State’s charter means, we necessarily must conclude that the grants do not violate the Michigan Constitution. V. Having reversed the district court’s partial grant of summary judgment to American Atheists, we also must reverse its decision to award attorney’s fees to the organization. Only a “prevailing party” may receive attorney’s fees under the statute, 42 U.S.C. § 1988(b), and to qualify as a prevailing party a plaintiff “must obtain at least some relief on the merits of his claim,” Farrar v. Hobby, 506 U.S. 103, 111, 113 S.Ct. 566, 121 L.Ed.2d 494 (1992). Because all nine of the challenged grants satisfy the Establishment Clause, American Atheists is not entitled to any relief on the merits — and necessarily is not entitled to any fees. See id. at 111-12, 113 S.Ct. 566. sfc s¡« # :¡í ^ "When all is said and done, we are left with three basic choices in ruling on the claimants’ Establishment Clause claim. The first option is to embrace a bright-line rule premised on the facial neutrality of the government-aid program. So long as the benefit on its face is allocated in an evenhanded way to the religious and nonreligious, it will satisfy the Establishment Clause. The second option is a bright-line rule premised on the receipt by a religious entity of monetary aid from the government. So long as the government program passes money from the public fisc directly to a religious entity, it will violate the Establishment Clause. Both options would be easy for courts to administer and easy for governments to follow, but a majority of the Court has never embraced *302 either option. The first option runs the risk of being underinclusive by failing to account for neutrally written laws that mask a hidden agenda to advance religion. The second option runs the risk of being overinclusive by failing to account for what the courts are “walling in or walling out,” Robert Frost, Mending Wall, in North of Boston 11, 12 (3d ed.1915), and most notably by frequently walling out sensible and eminently neutral benefits programs that have nothing whatsoever to do with the establishment of religion — such as Detroit’s modest efforts to improve the appearance of its downtown. The third option, the one we have followed, starts with neutrality, but it does not end there. It begins with the key concern of both Religion Clauses — that the government will not act neutrally toward religion or among religions. See Everson, 330 U.S. at 18, 67 S.Ct. 504; Lukumi, 508 U.S. at 532-33, 113 S.Ct. 2217. In the Free Exercise context, that means rational-basis review applies to neutral, generally applicable laws, Smith, 494 U.S. at 877-79, 110 S.Ct. 1595, but that strict scrutiny applies to laws that “target[] religious conduct for distinctive treatment,” Lukumi 508 U.S. at 534, 113 S.Ct. 2217; see id. at 531-32, 113 S.Ct. 2217. In the Establishment Clause context, that means evenhanded, neutral laws generally (though not invariably) will be upheld. So long as the government benefit is neutral and generally applicable on its face, it presumptively will satisfy the Establishment Clause. For if the Free Exercise Clause permits the government to regulate religious entities through neutral, generally applicable laws, the Establishment Clause typically will permit the government to provide neutral, generally applicable benefits to these same groups. At the same time, much of what limits the government in Free Exercise cases often will limit the government in Establishment cases. Even though a program may appear to be neutral on its face, it may betray the purpose or primary effect of advancing religion in general or one religion in particular. And when that is true, the program must be invalidated. None of that happened here, however, and accordingly Detroit’s downtown refurbishment program must be upheld. For these reasons, we affirm in part and reverse in part, and remand the case to the district court for further proceedings.