Plaintiffs are California card clubs and charities that are prohibited under California state law from offering casino-style gaming. They challenge the validity of compacts entered into under the Indian Gaming Regulatory Act (“IGRA”), 25 U.S.C. §§ 2701-2721, between the State of California and certain Indian tribes. Pursuant to an amendment to the California Constitution that permits casino-style gaming only on Indian lands (“Proposition 1A”), California has entered into 62 compacts (“Tribal-State Compacts”) with Indian tribes allowing such gaming. Plaintiffs brought this action, in federal district court, against various state defendants1 and the Secretary and Assistant Secretary of the United States Department of the Interior, alleging that Proposition 1A and the Tribal-State Compacts violate IGRA and their rights to equal protection guaranteed by the Fifth and Fourteenth Amendments.2
The district court granted summary judgment to both the state defendants and the federal defendants. Because we hold that Proposition 1A and the Tribal-State Compacts are consistent with IGRA and do not violate the guarantees of equal protection, we affirm.
BACKGROUND
A. The Indian Gaming Regulatory Act (“IGRA”)3
In California v. Cabazon Band of Mission Indians, 480 U.S. 202, 107 S.Ct. 1083, 94 L.Ed.2d 244 (1987), the Supreme Court invalidated an attempt by California to *715enforce California Penal Code § 326.5 (the “bingo statute”) against tribes that operated bingo halls. The Supreme Court characterized the bingo statute as regulatory, rather than criminal, and held that Public Law No. 280 prohibited the enforcement of state regulatory statutes against Indian tribes:
[I]f the intent of a state law is generally to prohibit certain conduct, it falls within Pub. L. 280’s grant of criminal jurisdiction, but if the state law generally permits the conduct at issue, subject to regulation, it must be classified as civil/regulatory and Pub. L. 280 does not authorize its enforcement on an Indian reservation.
Cabazon, 480 U.S. at 209, 107 S.Ct. 1083. Because California permitted a substantial amount of gaming activity, including bingo, the bingo statute could not be characterized as criminal or prohibitory and therefore could not be enforced on Indian lands.
As a response to the Cabazon decision, Congress enacted IGRA as a means of granting states some role in the regulation of Indian gaming. As noted in the opinion below,
IGRA was Congress’ compromise solution to the difficult questions involving Indian gaming. The Act was passed in order to provide “a statutory basis for the operation of gaming by Indian tribes as a means of promoting tribal economic development, self-sufficiency, and strong tribal governments” and “to shield [tribal gaming] from organized crime and other corrupting influences to ensure that the Indian tribe is the primary beneficiary of the gaming operation.” 25 U.S.C. § 2702(1), (2). IGRA is an example of “cooperative federalism” in that it seeks to balance the competing sovereign interests of the federal government, state governments, and Indian tribes, by giving each a role in the regulatory scheme.
Artichoke Joe’s v. Norton, 216 F.Supp.2d 1084, 1092 (E.D.Cal.2002) (alteration in original).
IGRA creates three classes of gaming, each of which is subject to a different level of regulation. Class I gaming covers “social games solely for prizes of minimal value or traditional forms of Indian gaming engaged in by individuals as part of, or in connection with, tribal ceremonies or celebrations.” 25 U.S.C. § 2703(6). Class II gaming includes bingo and card games that are explicitly authorized by a state or “not explicitly prohibited by the laws of the State and are [legally] played at any location in the State.” Id. § 2703(7)(A)(ii). Class II gaming specifically excludes banked card games and slot machines.4
At issue in this case is class III gaming, the most heavily regulated and most controversial form of gambling under IGRA. Class III gaming includes “all forms of gaming that are not class I gaming or class II gaming.” Id. § 2703(8). It includes the types of high-stakes games usually associated with casino-style gambling, as well as slot machines and parimutuel horse-wagering. Class III gaming is lawful on Indian lands only if three conditions are satisfied:5 (1) authorization by an ordi*716nance or resolution of the governing body of the Indian tribe and the Chair of the National Indian Gaming Commission (“NIGC”);6 (2) location in a state that permits such gaming for any purpose by any person, organization, or entity; and (3) the existence of a Tribal-State compact approved by the Secretary of the Interior. Id. § 2710(d)(1).
IGRA’s compacting requirement allows states to negotiate with tribes that are located within their borders regarding aspects of class III Indian gaming that might affect legitimate state interests. Id. § 2710(d)(3)(C). The compacting process gives to states civil regulatory authority that they otherwise would lack under Ca-lazón, while granting to tribes the ability to offer legal class III gaming. Keweenaw Bay Indian Cmty. v. United States, 136 F.3d 469, 472 (6th Cir.1998). IGRA also imposes on states an obligation to conduct compact negotiations in good faith, 25 U.S.C. § 2710(d)(3)(A), and allows tribes to enforce that obligation in federal court, id. § 2710(d)(7)(A).7
B. California’s Regulation of Indian Gaming Under IGRA
After the enactment of IGRA, certain Indian tribes in California sought to negotiate compacts with the State to permit the operation of class III gaming on their reservations. The class III games over which the tribes sought to negotiate — live banked or percentage card games and stand-alone electronic gaming machines (similar to slot machines) — were not permitted under California law. See Cal.Penal Code §§ 330, 330a, 330b. However, California did allow other forms of class III gaming, such as nonelectronic keno and lotto. Rumsey Indian Rancheria of Wintun Indians v. Wilson, 64 F.3d 1250, 1255 n. 1 (9th Cir.1994).
During the administration of Governor Pete Wilson, California refused to negotiate with tribes with respect to the forms of gaming that they sought to conduct. Because the State did not permit live banked or percentage card games or slot machine-like devices, it took the view that it had no obligation to negotiate with respect to those games or devices. The tribes argued that, because the State permitted other types of class III games, it could not refuse to negotiate over a particular subset of class III games. See Mashantucket Pequot Tribe v. Connecticut, 913 F.2d 1024, 1030 (2d Cir.1990) (agreeing with the tribe’s position). In Rumsey, this court rejected the tribes’ view, holding that
IGRA does not require a state to negotiate over one form of Class III gaming activity simply because it has legalized another, albeit similar form of gaming. Instead, the statute says only that, if a state allows a gaming activity “for any purpose by any person, organization, or *717entity,” then it also must allow Indian tribes to engage in that same activity. 25 U.S.C. § 2710(d)(1)(B). In other words, a state need only allow Indian tribes to operate games that others can operate, but need not give tribes what others cannot have.
64 F.3d at 1258.
The Rumsey decision meant that the State of California had no obligation under federal law to negotiate with the tribes over the class III gaming that the tribes wanted to operate. The tribes thus resorted to California’s initiative process to impose a state-law obligation on California to negotiate class III gaming compacts. A coalition of California tribes drafted Proposition 5, which required the State to enter into a model class III gaming compact covering banked card games and slot machines. The Proposition required the Governor to execute compacts within 30 days after any federally recognized Indian tribe requested such an arrangement. If the Governor took no action within 30 days, the compacts were deemed approved. Flynt v. Cal. Gambling Control Comm’n, 104 Cal.App.4th 1125, 129 Cal.Rptr.2d 167, 176 (2002), cert. denied, — U.S.-, 124 S.Ct. 398, 157 L.Ed.2d 278 (2003); Cal. Gov’t Code §§ 98000-98012. Proposition 5 also contained a provision waiving California’s sovereign immunity to suits brought under IGRA. See id. § 98005.
After the passage of Proposition 5 on the November 1998 ballot, the Hotel Employees and Restaurant Employees International Union filed a petition for a writ of mandate in the California Supreme Court, seeking to prevent the Governor from implementing the Proposition. The union alleged that Proposition 5 violated Article VI, Section 19(e) of the California Constitution, which states that the “Legislature has no power to authorize, and shall prohibit casinos of the type currently operating in Nevada and New Jersey.” The California Supreme Court agreed with the union and issued a peremptory writ of mandate, preventing the Governor from implementing Proposition 5. Hotel Employees & Rest. Employees Int’l Union v. Davis, 21 Cal.4th 585, 88 Cal.Rptr.2d 56, 981 P.2d 990, 1011 (1999).
Before the California Supreme Court ruled in Hotel Employees, Governor Gray Davis took office and sought to negotiate class III gaming compacts with several Indian tribes in California. However, once the California Supreme Court issued its ruling, the State no longer had authority to execute the compacts. To address this problem, the Davis administration proposed an amendment to the California Constitution that would exempt Indian tribes from the State’s constitutional prohibition on class III gaming. The Davis administration and various Indian tribes continued to negotiate Tribal-State Compacts, adding a provision that conditioned execution of the compacts on ratification of the proposed state constitutional amendment. In September of 1999, the Governor concluded 57 compacts with Indian tribes. The California legislature quickly ratified all 57 compacts. Artichoke Joe’s, 216 F.Supp.2d at 1096.
In March of 2000, California voters ratified Proposition 1A, amending the California Constitution to provide:
Notwithstanding subdivisions (a) and (e), and any other provision of state law, the Governor is authorized to negotiate and conclude compacts, subject to ratification by the Legislature, for the operation of slot machines and for the conduct of lottery games and banking and percentage card games by federally recognized Indian tribes on Indian lands in California in accordance with federal law. Accordingly, slot machines, lottery games, and banking and percentage *718card games are hereby permitted to be conducted and operated on tribal lands subject to those compacts.
Cal. Const, art. IV, § 19(f). Shortly after the ratification of Proposition 1A, the Assistant Secretary of Indian Affairs approved the compacts on behalf of the Secretary of the Interior (“Secretary”). He concluded that “[t]he Governor can, consistent with the State’s amended Constitution, conclude a compact giving an Indian tribe, along with other California Indian tribes, the exclusive right to conduct certain types of Class III gaming.” Upon this approval, the Tribal-State Compacts went into effect.
The compacts “create a unique opportunity for[each] Tribe to operate its Gaming Facility in an economic environment free of competition from the Class III gaming ... on non-Indian lands in California.” Tribal-State Compact Between the State of California and the Augustine Band of Mission Indians (Mar. 15, 2000), pmbl. E. The compacts condition the operation of class III gaming facilities on tribes’ agreeing to certain revenue-sharing arrangements and paying into a “Special Distribution Fund,” which is used by certain state agencies to cover expenses related to Indian gaming. Since the original 57 compacts took effect, 5 additional compacts have been entered into by the Governor and approved by the Secretary. Artichoke Joe’s, 216 F.Supp.2d at 1096. At present, 39 of the 62 tribes with compacts operate casinos with slot machines. Casinos have become the main source of revenue for Indian tribes in California. Some 44 California tribes are still without compacts, however. Id.
C. District Court Proceedings
Under Proposition 1A, Indian tribes are the only entities that are permitted to conduct class III gaming in California. California law does not bar non-Indian gaming altogether, however. Player-banked games, such as poker, are permitted in card rooms that generally charge membership fees or hourly rates. Also, California allows charities to engage in bingo for fundraising. Plaintiffs — non-Indians who are currently conducting gaming operations within those regulations— brought this action to challenge the Indians’ exclusive right to conduct class III gaming. Plaintiffs’ gaming operations are not nearly as lucrative as the gaming allowed to tribes. Plaintiffs allege that allowing only Indians to conduct casino gaming violates IGRA and Plaintiffs’ right to equal protection under the Fifth and Fourteenth Amendments.
Plaintiffs sought declaratory and injunc-tive relief on four counts of their complaint. Count One charged the federal defendants, namely, the Secretary and Assistant Secretary of the Interior, with violating IGRA and the Fifth Amendment’s guarantee of equal protection by approving California’s Tribal-State Compacts.8
Count Two sought similar relief under 42 U.S.C. § 1983 against Governor Gray Davis, the Director of the California Division of Gambling Control, and the Chair and Members of the California Gambling Control Commission. Plaintiffs alleged that Proposition 1A and the Tribal-State Compacts violate both IGRA and the Fourteenth Amendment’s Equal Protection Clause. Plaintiffs sought to invalidate the current Tribal-State Compacts and to bar the Governor from entering into any *719future Tribal-State compacts (“Prospective Compacts”). Count Three also targeted Prospective Compacts; Plaintiffs sought to enjoin the state defendants from passing legislation thought to precede a compact with the Lytton Band, a tribe that Plaintiffs expected would build a casino near San Francisco.
Finally, in Count Four, Plaintiffs sought to enjoin California’s Attorney General, the Director of the California Division of Gambling Control, and the Chair and Members of the California Gambling Control Commission from prosecuting Plaintiffs under the relevant sections of the penal code that proscribe casino-style gaming, so long as the Tribal-State Compacts remain in effect.
The district court held that it had “jurisdiction to resolve the claims against the federal defendants, the claims against the Governor related to existing compacts, and the claims against the State Attorney General and the Director of the California Division of Gambling Control as to the enforcement of state gaming laws against plaintiffs.” Artichoke Joe’s, 216 F.Supp.2d at 1090. However, the court dismissed Count Three and Plaintiffs’ other claims as to the Prospective Compacts on the ground that Plaintiffs had failed to demonstrate that they faced an “immediate and imminent” threat of harm from the Prospective Compacts. Id. at 1103. The court also dismissed the California Gambling Control Commission from Counts Two and Four, finding that Plaintiffs had not demonstrated that the threat of enforcement was fairly traceable to the Commission.9
On the merits, the district court determined that Proposition 1A satisfies IGRA’s requirement that a state “permit” class III gaming “for any purpose, by any person, organization, or entity.” Id. at 1128. The court went on to decide that granting Indian tribes a monopoly on class III gaming does not violate the Fifth or the Fourteenth Amendment’s guarantee of equal protection. Id. at 1132-33. The district court therefore granted Defendants’ motion for summary judgment.10 Id. at 1133.
This timely appeal followed.
STANDARD OF REVIEW
We review de novo a grant of summary judgment. Robi v. Reed, 173 F.3d 736, 739 (9th Cir.1999). We also review de novo questions of statutory in*720terpretation, Lopez v. Wash. Mut. Bank, 302 F.3d 900, 903 (9th Cir.2002), and questions of the constitutionality of a federal statute, Mayweathers v. Newland, 314 F.3d 1062, 1066 (9th Cir.2002), cert. denied, — U.S. -, 124 S.Ct. 66, 157 L.Ed.2d 20 (2003).
ANALYSIS
A. Validity of Proposition 1A and the Tribal-State Compacts Under IGRA and the Johnson Act11
Assessing whether Proposition 1A and the Tribal-State Compacts violate IGRA requires us to interpret 25 U.S.C. § 2710(d)(1)(B), IGRA’s provision regulating class III gaming on Indian lands. “We interpret a federal statute by ascertaining the intent of Congress and by giving effect to its legislative will.” Ariz. Appetito’s Stores, Inc. v. Paradise Vill. Inv. Co. (In re Ariz. Appetito’s Stores, Inc.), 893 F.2d 216, 219 (9th Cir.1990). “We begin, as always, with the language of the statute.’ ” Navajo Nation v. Dep’t of Health & Human Servs., 325 F.3d 1133, 1136 (9th Cir.2003) (en banc) (quoting Duncan v. Walker, 533 U.S. 167, 172, 121 S.Ct. 2120, 150 L.Ed.2d 251 (2001)). “When the words of a statute are unambiguous, ... judicial inquiry is complete.” Conn. Nat’l Bank v. Germain, 503 U.S. 249, 254, 112 S.Ct. 1146, 117 L.Ed.2d 391 (1992) (internal quotation marks omitted). ‘Where the language is not dispositive, we look to the congressional intent revealed in the history and purposes of the statutory scheme.” United States v. Buckland, 289 F.3d 558, 565 (9th Cir.2002) (en banc) (internal quotation marks omitted).
1. Statutory Text
The parties dispute the proper construction of 25 U.S.C. § 2710(d)(1)(B), which provides that “[cjlass III gaming activities shall be lawful on Indian lands only if such activities are ... (B) located in a State that permits such gaming for any purpose by any person, organization, or entity.” Construing this provision involves two separate but related questions of interpretation. First, we must decide whether Proposition 1A, which permits class III gaming only on Indian lands, is consistent with § 2710(d)(l)(B)’s requirement that a State be one that “permits such gaming” before a class III gaming compact may take effect. Then we must decide whether “any person, organization, or entity” should be read to exclude Indian tribes. To resolve these questions, we examine the text both in isolation and in the context of the statute as a whole.
(a) “permits such gaming”
In construing the phrase “permits such gaming” in § 2710(d)(1)(B), we are mindful of cases that characterize as “patent bootstrapping” the notion that a Tribal-State compact can satisfy both the “permits such gaming” requirement under § 2710(d)(1)(B) and the compact requirement under § 2710(d)(1)(C). See Citizen Band Potawatomi Indian Tribe of Okla. v. Green, 995 F.2d 179, 181 (10th Cir.1993) (holding that a Tribal-State compact allowing the importation of gambling devices onto a tribe’s lands violated the Johnson *721Act because, standing alone, the compact could not satisfy § 2710(d)(l)(B)’s “permits such gaming” requirement); United States v. Santa Ynez Band of Chumash Mission Indians of Santa Ynez Reservation, 33 F.Supp.2d 862 (C.D.Cal.1998) (describing games that are illegal under state law as “uncompactable”). However, Proposition 1A distinguishes the present controversy from the “bootstrapping” cases. Proposition 1A does more than authorize the Governor to enter into Tribal-State compacts. It explicitly states that “slot machines, lottery games, and banking and percentage card games are hereby permitted to be conducted and operated on tribal lands” subject to the regulations embodied in the Tribal-State compact. Proposition 1A (emphasis added). Thus, there is law— separate from the compact itself — that “permits such gaming” in certain circumstances. 25 U.S.C. § 2710(d)(1)(B).
Plaintiffs argue that, because California has no authority to “permit” gaming on Indian lands under Cabazon, “permits such gaming” can refer only to the State’s ability to regulate gaming operations conducted on non-Indian lands. When enacting IGRA, Plaintiffs contend, Congress would have known that the State, acting alone, could not “permit” class III gaming on Indian lands and, thus, that a state law providing for class III gaming conducted only by Indian tribes on Indian lands could not satisfy the requirement of § 2710(d)(1)(B). This argument is plausible but does not foreclose an alternative understanding of the verb “permit,” for three reasons: (1) Under Cabazon, California did not lack jurisdiction to apply its prohibition on class III gaming on Indian lands before IGRA was enacted; (2) at the same time Congress passed IGRA, it gave California the regulatory authority that Cabazon took away, 18 U.S.C. § 1166; and (3) we have not construed IGRA’s use of the word “permit” to require a legally binding affirmative act, see Rumsey, 64 F.3d at 1257.
Under Cabazon, the Supreme Court applied the long-standing general rule that a state has jurisdiction over Indian lands only if Congress has explicitly ceded that jurisdiction. 480 U.S. at 207, 107 S.Ct. 1083. But see id. at 214-15,107 S.Ct. 1083 (explaining exceptions in which a state may exercise jurisdiction without an express congressional mandate). California claimed that it had jurisdiction over gaming activities on Indian lands pursuant to authority ceded by Public Law No. 280, which conferred “broad criminal jurisdiction over offenses committed by or against Indians within all Indian country within the State.” Id. at 207, 107 S.Ct. 1083. Because California permitted some of the gaming at issue, however, with violations of the law deemed to be mere misdemeanors, the Supreme Court concluded that the statutes actually were regulatory rather than prohibitory (or criminal) in nature. Public Law No. 280 did not confer jurisdiction on the State to regulate gaming and, thus, California could not base its jurisdiction on that statute. Nonetheless, the general criminal jurisdiction that California exercises under Public Law No. 280 allowed California to prohibit gaming for Indian tribes, if the scheme was prohibitory rather than regulatory. Id. at 208, 107 S.Ct. 1083. Therefore, pre-IGRA and post-Cabazon, California still had some jurisdiction over Indian lands pertaining to gaming.
IGRA changed the landscape but did not divest California of its general criminal jurisdiction over Indian lands. Rather, it devised a method to give back some of the regulatory authority that the Supreme Court had held inapplicable to Indian lands in Cabazon. One of the bases of the holding in Cabazon was that Congress had not explicitly ceded regulatory authority *722for gaming to the states in Public Law No. 280 or otherwise. IGRA responded by creating a statutory basis for gaming regulation that introduced the compacting process as a means of sharing with the states the federal government’s regulatory authority over class III gaming. Cabazon, 480 U.S. at 207-14, 107 S.Ct. 1083; 25 U.S.C. §§ 2702(1), 2710(d)(1). Simultaneously, IGRA put into effect 18 U.S.C. § 1166, which provides that “all State laws pertaining to the licensing, regulation, or prohibition of gambling, including but not limited to criminal sanctions applicable thereto, shall apply in Indian country in the same manner and to the same extent as such laws apply elsewhere in the State.” 18 U.S.C. § 1166(a). The federal government retained the power to prosecute violations of state gambling laws in Indian country, so as to preserve the delicate balance of power between the States and the tribes. See Sycuan Band of Mission Indians v. Roache, 54 F.3d 535, 538 (9th Cir.1994) (explaining the limit on a state’s “jurisdiction” to enforce the gambling laws that extend into Indian lands). However, the fact that the federal government retained that power does not change the fact that California may enact laws and regulations concerning gambling that have an effect on Indian lands via § 1166. See United States v. E.C. Invs., Inc., 77 F.3d 327, 330-31 (9th Cir.1996) (discussing the relationship between state law and exclusive federal enforcement power in § 1166(d)).
Further, our decision in Rumsey supports a construction of the phrase “permits such gaming” under which California could “permit” gaming on Indian lands both before and after the ratification of Proposition 1A. Rumsey held that, in the statute at issue, “permit” had a clear and unambiguous meaning. 64 F.3d at 1257. “In United States v. Launder, 743 F.2d 686[, 689] (9th Cir.1984), we adopted a [Black’s Law Dictionary definition of the term ‘permit’ as meaning ‘ “[t]o suffer, allow, consent, let; to give leave or license; to acquiesce, by failure to prevent, or to expressly assent or agree to the doing of an act.” ’ ” Rumsey, 64 F.3d at 1257. In other words, under Rumsey, the word “permit” in this statute does not necessarily require an affirmative act of legal authority in order to “permit” conduct. California may “permit” class III gaming within the meaning of IGRA even if it “acquiesces, by failure to prevent” class III gaming. Under Rumsey, mere tolerance of class III gaming might be enough to satisfy § 2710(d)(l)(B)’s requirement that a state “permit[] such gaming for any purpose by any person, organization, or entity.” § 2710(d)(1)(B).
Plaintiffs are quick to point out that Rumsey dealt with a state’s obligation to negotiate under IGRA, rather than the limits of a state’s negotiating authority in the compacting process. Indeed, most cases interpreting § 2710(d)(1)(B) have done so in the context of addressing a state’s duty to negotiate with a tribe concerning class III gaming. See, e.g, Rumsey, 64 F.3d at 1258; Cheyenne River Sioux Tribe v. South Dakota, 3 F.3d 273, 279 (8th Cir.1993) (construing the phrase “such gaming” in § 2710(d)(1)(B) in deciding whether the state had to negotiate with the tribes about traditional keno); Mashantucket Pequot Tribe, 913 F.2d at 1029-31 (construing the same provision in deciding which games were subject to negotiation). However material, this distinction does not compel a reading of “permits such gaming” that requires California to legalize non-Indian class III gaming before executing valid compacts under IGRA.
We are still left, then, with a statutory provision that is susceptible to more than one interpretation. Neither the statutory *723text read in isolation nor judicial constructions of it resolve the ambiguity.
(b) “any person, organization, or entity”
We reach a similar conclusion with respect to the requirement of § 2710(d)(1)(B) that a state permit class III gaming by “any person, organization, or entity.” There is nothing in the text itself that definitively resolves whether Congress intended Indian tribes to fall within the scope of “any person, organization, or entity” under this provision.
Plaintiffs interpret this phrase in § 2710(d)(1)(B) to exclude Indian tribes from “any person, organization, or entity,” because they read the subsection simply as a “most-favored nations” clause. If no other class III gaming is permitted in a state, they contend, then federal law bars Indians from conducting such gaming. Under this construction, § 2710(d)(1)(B) merely places Indians on an equal footing with other gambling businesses in the state. As Rumsey put it, after citing § 2710(d)(1)(B), “a state need only allow Indian tribes to operate games that others can operate, but need not give tribes what others cannot have.” 64 F.3d at 1258; see also 18 U.S.C. § 1166(a) (providing that state laws on gambling “shall apply in Indian country in the same manner and to the same extent as such laws apply elsewhere in the State”). There is no statutory provision expressly allowing States to enter into exclusive arrangements with Indian tribes. Plaintiffs reasonably interpret IGRA as a statute that is intended to maintain a competitive balance between Indian and non-Indian gaming interests under California and federal law.
Defendants, on the other hand, construe “any person, organization, or entity” to include Indian tribes. Looking only to the text itself, the breadth of the provision does suggest that it is meant to be read inclusively; “permits such gaming for any purpose by any person, organization, or entity” does not lend itself to easy circumscription. 25 U.S.C. § 2710(d)(1)(B) (emphasis added). The statute does not explicitly exclude Indians from its scope. And Rumsey did not hold that a state may not give tribes what others do not have, but only that a state need not do so. 64 F.3d at 1258.
As the district court noted, Congress employed “capacious language” to define those situations in which it would be legal for Indian tribes to conduct class III gaming operations on Indian lands. Artichoke Joe’s, 216 F.Supp.2d at 1121. Nonetheless, the parties’ textual dispute ultimately reduces to what the statute does not say. Because the statute does not explicitly exclude Indians from the phrase “any person, organization, or entity,” Defendants read IGRA to allow state-law provisions legalizing class III gaming monopolies on Indian lands to satisfy the requirements of § 2710(d)(1)(B). Because the statute does not expressly permit Indian tribes to run class III gaming enterprises as a monopoly, Plaintiffs read § 2710(d)(1)(B) to require states to permit class III gaming for any purpose by any non-Indian person, entity, or organization.
Looking to past judicial constructions of the text, Defendants have somewhat the better of the argument. In the context of deciding whether California could condition approval of a Tribal-State compact on a tribe’s agreement to certain revenue-sharing and employment provisions, we said that the core of the compact between Indian tribes and the State is the exchange of “the exclusive right to conduct lucrative Las Vegas-style class III gaming, free from non-tribal competition” for the tribes’ agreement “to a number of restrictions and obligations concerning their gaming *724enterprises.”12 In re Indian Gaming Related Cases, 331 F.3d at 1104. An Arizona district court reached a similar result in a case deciding that a Tribal-State compact, standing alone, cannot legalize Indian gaming under IGRA. The court read § 2710(d)(1)(B) to require a state to “first legalize a game, even if only for tribes, before it can become a compact term.” Am. Greyhound Racing, Inc. v. Hull, 146 F.Supp.2d 1012, 1067 (D.Ariz.2001), vacated on other grounds, 305 F.3d 1015 (9th Cir.2002) (emphasis added). Similarly, the California Supreme Court, adopting the reasoning of the district court in the present case, has construed § 2710(d)(1)(B) to allow tribal monopolies of class III gaming activities.13 See Flynt, 129 Cal.Rptr.2d at 178. Although the trend of judicial construction- of § 2710(d)(1)(B) slightly favors Defendants’ view, none of the cases controls the issue before us.
(c) The subsection and the statute as a whole
Reading 25 U.S.C. § 2710(d)(1)(B) in the light of IGRA as a whole also fails to resolve the ambiguity conclusively. Looking both to similar text used elsewhere in the statute and to other operational text addressing the application of state law on tribal lands, we find potentially relevant provisions supporting each party’s interpretation of § 2710(d)(1)(B).
IGRA uses identical text to govern whether tribes may engage in class II gaming under § 2710(b)(1)(A). A tribe may conduct class II gaming operations only if that tribe is “located within a State that permits such gaming for any purpose by any person, organization or entity.” Plaintiffs argue that the absence of a compacting process for class II gaming shows that the verb “permits” must refer to what states allow on non-Indian lands. They then cite the familiar canon of construction that identical text appearing more than once in the same statute is presumed to have the same meaning. See Batjac Prods., Inc. v. GoodTimes Home Video Corp., 160 F.3d 1223, 1228-29 (9th Cir.1998). However, this argument begs the question whether a state could permit class II gaming only on Indian lands, even without a compacting process.
Other sections of the statute that employ text similar to the phrase “any person, organization, or entity” in § 2710(d)(1)(B) lend support both to Plaintiffs’ and Defendants’ interpretations of that text. Section 2710(d)(2)(A) provides that, “[i]f any Indian tribe proposes to engage in, or to authorize any person or entity to engage in, a class III gaming activity on Indian lands of the Indian tribe, the governing body of the Indian tribe shall adopt and submit to the Chairman an ordinance or resolution that meets the requirements of’ § 2710(b). (Emphasis added.) In this provision, the term “any person or entity” implicitly excludes Indian tribes. Other subsections, however, qualify the phrase “any person or entity” when it is meant to exclude Indian tribes. Thus, *725for example, § 2710(d)(4) restricts a state’s ability to tax gaming activities. It states that IGRA should not be read to authorize the imposition of “any tax, fee, charge, or other assessment upon an Indian tribe or upon any other person or entity authorized by an Indian tribe to engage in class III activity.” (Emphasis added.) Section 2710(b)(4)(A) also qualifies “any person or entity” with the phrase “other than the Indian tribe” when distinguishing between Indian and non-Indian entities. These qualifiers suggest that the unadorned phrase “any person or entity” includes Indian tribes.
Looking to other operative text in the statute that addresses the applicability of state law on tribal lands, we again find support for both of the proposed interpretations. Section 2710(b)(4)(A) states that tribal licensing requirements for class II gaming operations must be “at least as restrictive as those established under State law governing similar gaming within the jurisdiction of the State within which such Indian lands are located.” (Emphasis added.) Read narrowly, this text supports the notion that Congress was legislating with the assumption in mind that existing state law would apply to tribal gaming operations — at least insofar as class II gaming is concerned. Read broadly, the text supports Plaintiffs’ interpretation that IGRA confers “most favored nation” status on Indian tribes.
Section 2710(d)(5) seems to carry the same assumption to the context of class III gaming, stating that
[n]othing in this subsection shall impair the right of an Indian tribe to regulate class III gaming on its Indian lands concurrently with the State, except to the extent that such regulation is inconsistent with, or less stringent than, the State laws and regulations made applicable by any Tribal-State compact....
In establishing consistency between state and tribal gaming regulations, Congress apparently assumed the existence of state regulation of class III gaming.
Finally, 18 U.S.C. § 1166 establishes federal jurisdiction over violations of state gaming laws on Indian lands. Section 1166(a) states that, “for purposes of Federal law, all State laws pertaining to the licensing, regulation, or prohibition of gambling, including but not limited to criminal sanctions applicable thereto, shall apply in Indian country in the same manner and to the same extent as such laws apply elsewhere in the State.” This provision suggests that existing state law is expected to apply to both Indian and non-Indian gaming activities. However, § 1166(c)(2), by excluding class III gaming operations conducted under Tribal-State compacts from the term “gambling” in § 1166(a), expressly carves out such operations from this jurisdictional scheme.
In summary, an examination of IGRA as a whole fails to resolve the ambiguities inherent in 25 U.S.C. § 2710(d)(1)(B). We thus turn to sources outside the operative statutory text.
2. Congressional Findings, Purpose, and Legislative History
As is true of the operative text, Congress’ statements of intent in enacting IGRA, and IGRA’s legislative history, leave us in equipoise between the parties’ competing interpretations of the statute. Both Plaintiffs and Defendants find support for their respective positions, but no party has directed our attention to, nor can we find, any statement of intent or legislative history that clearly demonstrates that Congress even considered the question before us.
The official legislative history of IGRA — the formal record of the Select *726Committee Report and the subsequent floor debates — is silent on the specific issue of tribal monopolies on class III gaming. This silence does not definitively favor either interpretation of IGRA although, as the district court noted, silence does suggest that tribal gaming monopolies were not at “the forefront of what Congress had in mind.” Artichoke Joe’s, 216 F.Supp.2d at 1121. “Drawing inferences as to congressional intent from silence in legislative history is always a precarious business.” Symons v. Chrysler Corp. Loan Guarantee Bd., 670 F.2d 238, 242 (D.C.Cir.1981).
The legislative history conveys quite clearly that Congress devised the Tribal-State compacting process as a means to resolve the most contentiously debated issue in the legislation: which authority— Tribal, State, or Federal — would regulate class III gaming. S.Rep. No. 100-446, at 5-6 (1988), repnnted in 1988 U.S.C.C.A.N. 3071, 3075-76. Indian tribes resisted ceding any authority over gaming to the states. Id. at 13, reprinted at 3083. The states, on the other hand, resisted tribally or federally regulated class III gaming, especially because many states already had in place regulatory systems for such gaming. Id. IGRA’s drafters conceived of the Tribal-State compact as “the best mechanism to assure that the interests of both sovereign entities are met with respect to the regulation of complex gaming enterprises.” Id.
The legislative history thus shows that Congress looked to the compacting process primarily as a means of balancing state and tribal interests. The Select Committee’s articulation of these interests refers only obliquely to the economic concerns of third parties:
In the Committee’s view, both State and tribal governments have significant governmental interests in the conduct of class III gaming.... A tribe’s governmental interests include raising revenues to provide governmental services for the benefit of the tribal community and reservation residents, promoting public safety as well as law and order on tribal lands, realizing the objectives of economic self-sufficiency and Indian self-determination, and regulating activities of persons within its jurisdictional borders. A State’s governmental interests with respect to class III gaming on Indian lands include the interplay of such gaming with the State’s public policy, safety, law and other interests, as well as impacts on the State’s regulatory system, including its economic interest in raising revenue for its citizens.
Id. (emphasis added). Recognition of the interests of non-Indians is implied, but only as a consideration for the state, not for Congress directly.
Plaintiffs direct our attention to parts of the legislative history showing that Congress intended to preserve an environment of free-market competition among entities engaged in class III gaming. Although the notion of free-market competition does appear in the legislative history, it figures most prominently in the context of warning states not to abuse the compacting process to protect non-Indian gaming interests: “It is the Committee’s intent that the compact requirement for class III not be used as a justification by a State for excluding Indian tribes from such gaming or for the protection of other State-licensed gaming enterprises from free market competition with Indian tribes.” Id.
This is not to say that the statute’s effect on existing gaming interests was entirely absent from Congress’ deliberations. The Select Committee on Indian Affairs stated that IGRA “allows States to consider negative impacts on existing gaming activities.” Id. at 14, repnnted at *7273084. However, the legislative history seems simply to assume (rather than affirmatively to require) existing, competitive gaming activities. Congress intended that IGRA not “allow States to reject Indian gaming on the mere showing that Indian gaming will compete with non Indian games. Rather, States must show that economic consequences will be severe and that they will clearly outweigh positive economic consequences.” Id.
Other portions of the legislative history suggest that Congress enacted IGRA with the assumption that States would use laws and regulatory systems that police extant class III gaming operations to regulate similar operations on tribal lands. One goal of IGRA is to “foster a consistency and uniformity in the manner in which laws regulating the conduct of gaming activities are applied.” Id. at 6, re-printed at 3076. The Justice Department opposed the creation of a federal agency to regulate class II and class III gaming activity because, in the Department’s view, “the expertise to regulate gaming activities and to enforce laws related to gaming could be found in state agencies.” Id. at 5, reprinted at 3075.
Congress shared the Justice Department’s concern about the effect of tribal class III gaming operations on existing state regulatory systems. The Select Committee noted “the strong concerns of states that state laws and regulations relating to sophisticated forms of class III gaming be respected on Indian lands.” Id. at 13, reprinted at 3083. Its admonition to the states not to use compact negotiations to protect “other State-licensed gaming enterprises,” id., from competition with Indian tribes also reveals the assumption that the class III gaming compacts would be concluded in states with laws and agencies that regulate existing non-Indian class III gaming operations.
Other portions of the legislative history support Defendants’ interpretation. We first note that any express congressional consideration of non-Indian gaming interests in the text of IGRA seems to have fallen out as Congress worked through competing drafts of the legislation. As introduced, Senate Bill 555 contained text that provided explicitly for the interpretation that Plaintiffs urge here. Section 11(d)(1) and (2) provided:
(1) Except as provided in paragraph (2) of this subsection, class III gaming shall be unlawful on any Indian lands under section 1166 of title 18, United States Code.
(2)(A) A gaming activity on Indian lands that is otherwise legal tuithin the State where such lands are located may be exempt from the operation of paragraph (1) of this subsection where the Indian tribe requests the Secretary to consent to the transfer of all civil and criminal jurisdiction, except for taxing authority, pertaining to the licensing and regulation of gaming over the proposed gaming enterprise to the State within which such gaming enterprise is to be located and the Secretary so consents.
133 Cong. Rec. 3740 (1987) (emphasis added). The final version of S.B. 555, however, uses the current, broader text. The current text originated in S.B. 1303, a bill that ultimately failed in committee. Id. at 14332, § 10(b). The fact that the “permits such gaming” text was taken from another bill suggests that the substitution was deliberate, and the particular substitution that the drafters chose implies that Congress intended a broader meaning than the one proposed by Plaintiffs.
Further, at least one individual legislator did recognize a preference for Indian gaming. In his separate statement following the Select Committee Report, Senator Evans observed that
*728Indian tribes may have a competitive economic advantage because, rightly or wrongly, many states have chosen not to allow the same types of gaming in which tribes are empowered to engage. Ironically, the strongest opponents of tribal authority over gaming on Indian lands are from States whose liberal gaming policies would allow them to compete on an equal basis with the tribes.
S.Rep. No. 100-446, at 36, reprinted at 3105. Although not dispositive as to Congress’ view of class III gaming monopolies, Senator Evans’ comments do show that at least one legislator, and perhaps other interested parties as well, viewed Indian monopolies as a possibility under IGRA. IGRA’s stated purposes also tend slightly to favor Defendants’ interpretation. Section 2702 provides that IGRA’s purpose is
(1) to provide a statutory basis for the operation of gaming by Indian tribes as a means of promoting tribal economic development, self-sufficiency, and strong tribal governments;
(2) to provide a statutory basis for the regulation of gaming by an Indian tribe adequate to shield it from organized crime and other corrupting influences, to ensure that the Indian tribe is the primary beneficiary of the gaming operation, and to assure that gaming is conducted fairly and honestly by both the operator and players[.]
25 U.S.C. § 2702.14 Those stated purposes address the Indian tribes’ interests in matters related to their autonomy and the State’s interest in the prevention of crime. Nowhere is there any reference to the idea that IGRA serves as a means of policing equality between Indian and non-Indian gaming operations in the context of class III gaming. At the same time, Congress’ interests in tribal self-sufficiency and self-government do not necessarily translate into approval of state-wide gaming monopolies.
The statement of congressional findings in § 2701 is similarly ambiguous:
Indian tribes have the exclusive right to regulate gaming activity on Indian lands if the gaming activity is not specifically prohibited by Federal law and is conducted within a State which does not, as a matter of criminal law and public policy, prohibit such gaming activity.
Id. § 2701(5). One way to interpret this section is to note that Congress wanted to grant to Indian tribes an “exclusive right,” a right unavailable to non-Indian gaming interests, and to read “such gaming activity” to refer back to the entire phrase “gaming activity on Indian lands.” This reading favors Defendants. It is equally possible, however, to read this section to say that the only “exclusive right” granted is the right “to regulate gaming activity on Indian lands” and thus to read “such gaming activity” to refer back only to the remaining phrase “the gaming activity” which, in general, must be a type of gaming activity that a state does not prohibit. This reading favors Plaintiffs.
In sum, an examination of the legislative history and IGRA’s stated purposes does not resolve the ambiguities in the operative text. Both Plaintiffs’ interpretation and Defendants’ interpretation remain plausible. We therefore turn to other interpretive aids.
3. The Blackfeet Tribe Presumption
Ambiguity in a statute that is enacted for the benefit of Indians implicates a well-known canon of construction. In Montana *729v. Blackfeet Tribe of Indians, 471 U.S. 759, 105 S.Ct. 2399, 85 L.Ed.2d 753 (1985), the Supreme Court held that “the standard principles of statutory construction do not have their usual force in cases involving Indian law. As we said earlier this Term, ‘[t]he canons of construction applicable in Indian law are rooted in the unique trust relationship between the United States and the Indians.’” Id. at 766, 105 S.Ct. 2399 (quoting Oneida County v. Oneida Indian Nation, 470 U.S. 226, 247, 105 S.Ct. 1245, 84 L.Ed.2d 169 (1985)). The Court then described the two canons of construction that apply specially in Indian law, one of which is that “statutes are to be construed liberally in favor of the Indians, with ambiguous provisions interpreted to their benefit.” Id.
This presumption, also known as the trust doctrine, grew out of the trust obligation that Congress owes to Indian tribes. Initially, the presumption was applied in the context of Indian treaties as a counterbalance to Congress’ ability to abrogate a treaty obligation unilaterally without a tribe’s consent. See Felix S. Cohen’s Handbook of Federal Indian Law 221-23 (2d ed. 1982) (collecting cases). Later decisions extended this presumption beyond the context of Indian treaties to land disputes, United States v. Santa Fe Pac. R.R., 314 U.S. 339, 62 S.Ct. 248, 86 L.Ed. 260 (1941); state jurisdiction to tax, Bryan v. Itasca County, 426 U.S. 373, 96 S.Ct. 2102, 48 L.Ed.2d 710 (1976); and the application of the Bill of Rights to tribes in matters of self-government, Santa Clara Pueblo v. Martinez, 436 U.S. 49, 98 S.Ct. 1670, 56 L.Ed.2d 106 (1978).15
The Blackfeet presumption simply requires that, when there is doubt as to the proper interpretation of an ambiguous provision in a federal statute enacted for the benefit of an Indian tribe, “the doubt [will] benefit the Tribe, for ‘[a]mbiguities in federal law have been construed generously in order to comport with ... traditional notions of sovereignty and with the federal policy of encouraging tribal independence.’ ” Merrion v. Jicarilla Apache Tribe, 455 U.S. 130, 152, 102 S.Ct. 894, 71 L.Ed.2d 21 (1982) (quoting White Mountain Apache Tribe v. Bracker, 448 U.S. 136, 143-44, 100 S.Ct. 2578, 65 L.Ed.2d 665 (1980)); see also Native Vill. of Venetie I.R.A. Council v. Alaska, 944 F.2d 548, 553 (9th Cir.1991) (applying the Blackfeet presumption when choosing between two interpretations of the Indian Child Welfare Act of 1978). This presumption is subject to two implicit limitations. First, the presumption applies only to federal statutes that are “passed for the benefit of dependent Indian tribes.” Hoonah Indian Ass’n v. Morrison, 170 F.3d 1223, 1228-29 (9th Cir.1999) (internal quotation marks omitted). Second, ambiguity is a prerequisite for any application of the Blackfeet presumption. Thus, in Rumsey, we refused to apply the presumption favoring tribes where doing so would contradict the plain words of the statute. 64 F.3d at 1257.
Neither of those limitations prevents us from turning to the Blackfeet presumption *730in this case. As discussed above, the relevant text of the statute is ambiguous. Further, IGRA is undoubtedly a statute passed for the benefit of Indian tribes. IGRA’s declaration of policy, providing that the statute is intended “as a means of promoting tribal economic development, self-sufficiency, and strong tribal governments,” 25 U.S.C. § 2702(1), firmly places the statute in the category of legislation to which the Blackfeet presumption applies.
Application of the Blackfeet presumption is straightforward. We are confronted by an ambiguous provision in a federal statute that was intended to benefit Indian tribes. One construction of the provision favors Indian tribes, while the other does not. We faced a similar situation in the context of Indian taxation in Quinault Indian Nation v. Grays Harbor County, 310 F.3d 645 (9th Cir.2002). In choosing between two characterizations of a tax law “plagued with ambiguity,” we adopted the construction that favored the Indian Nation over the one that favored Grays Harbor County, noting that “it is not enough to be persuaded that the County’s is a permissible or even the better reading.” Id. at 647.
Here, we must follow a similar approach. We adopt Defendants’ construction, not because it is necessarily the better reading, but because it favors Indian tribes and the statute at issue is both ambiguous and intended to benefit those tribes.
Neither of the two exceptions to the application of the Blackfeet presumption causes us pause. The first exception is that deference to an agency’s interpretation can overcome the presumption in favor of Indian tribes. Haynes v. United States, 891 F.2d 235, 239 (9th Cir.1989). But see Navajo Nation, 325 F.3d at 1136 n. 4 (noting a circuit split as to whether deference to an agency’s interpretation takes priority over the Blackfeet presumption). Assuming, without deciding, that the Secretary’s interpretation of § 2710(d)(1)(B) is entitled to deference under Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), and United States v. Mead Corp., 533 U.S. 218, 121 S.Ct. 2164, 150 L.Ed.2d 292 (2001), that interpretation likewise adopts Defendants’ construction of the statute and favors Indian tribes. In other words, the Blackfeet presumption and the doctrine of agency deference point to the same result.
The second exception involves the avoidance of constitutionally doubtful interpretations of a statute. In Williams v. Babbitt, 115 F.3d 657 (9th Cir.1997), we declined to adopt the Department of the Interior’s interpretation of the Reindeer Industry Act of 1937, 25 U.S.C. §§ 500-500n. Although the agency’s interpretation was “not unreasonable” and was owed Chevron deference, we ruled that the agency’s interpretation raised severe constitutional concerns that prevailed over any deference owed to the agency. Williams, 115 F.3d at 661-62. We reasoned that, because Chevron deference trumps the Blackfeet presumption, and the Supreme Court held in Edward J. DeBartolo Corp. v. Florida Gulf Coast Building & Constr. Trades Council, 485 U.S. 568, 108 S.Ct. 1392, 99 L.Ed.2d 645 (1988), that the doctrine of constitutional avoidance trumps Chevron deference,' it follows that constitutional avoidance also trumps statutory constructions favoring Indians. Williams, 115 F.3d at 663 n. 5.
The doctrine of constitutional avoidance requires that “ ‘every reasonable construction must be resorted to, in order to save a statute from unconstitutionality.’ ” Rust v. Sullivan, 500 U.S. 173, 190, 111 S.Ct. 1759, 114 L.Ed.2d 233 (1991) (quoting DeBartolo Corp., 485 U.S. at 575, *731108 S.Ct. 1392 (emphasis omitted)). Here, we need not invalidate California’s grant to Indian tribes of exclusive class III gambling privileges in order to save the statute from unconstitutionality because, as we are about to explain, Plaintiffs’ constitutional arguments “do not carry the day.” Id. at 191, 111 S.Ct. 1759. Moreover, the privileges granted to Indian tribes do not raise constitutional concerns that are sufficiently “grave” to trigger the application of the doctrine. Id. Rather, the award of exclusive class III gaming franchises simply furthers the federal government’s longstanding trust obligations to Indian tribes and helps promote their economic self-development. As the Supreme Court has cautioned elsewhere, “[statutes should be interpreted to avoid serious constitutional doubts, not to eliminate all possible contentions that the statute might be unconstitutional.” Reno v. Flores, 507 U.S. 292, 314 n. 9, 113 S.Ct. 1439, 123 L.Ed.2d 1 (1993) (citation omitted). Thus, the doctrine of constitutional avoidance does not apply here. See Almendarez-Torres v. United States, 523 U.S. 224, 239, 118 S.Ct. 1219, 140 L.Ed.2d 350 (1998) (stating that “the ‘constitutional doubt’ doctrine does not apply mechanically whenever there arises a significant constitutional question the answer to which is not obvious”).
4. Conclusion
IGRA is a statute that Congress enacted for the benefit of Indian tribes. Section 2710(d)(1)(B) remains ambiguous after consulting its text, context, purpose, and legislative history. We therefore apply the Blackfeet presumption and hold that California’s Proposition 1A “permits” class III gaming within the meaning of IGRA by legalizing such gaming operations only when conducted by the “entity” of an Indian tribe.
B. Equal Protection
Having held that IGRA allows the State of California to grant to the Indian tribes a monopoly on class III gaming, we must decide the second question that Plaintiffs pose: whether that monopoly violates Plaintiffs’ right to equal protection under the laws. The equal protection claim requires us to answer two questions. First, we must decide whether the distinction between Indian and non-Indian gaming interests is a political or a racial classification, so we can determine the proper level of deference that is owed to the classification. Second, we must decide whether, under the applicable standard of review, legitimate state interests justify the grant to Indian tribes of a monopoly on class III gaming.
1. Nature of the Classification in IGRA and in Proposition 1A
“In areas of social and economic policy, a statutory classification that neither proceeds along suspect fines nor infringes fundamental constitutional rights must be upheld against equal protection challenge if there is any reasonably conceivable state of facts that could provide a rational basis for the classification.” FCC v. Beach Communications, Inc., 508 U.S. 307, 313, 113 S.Ct. 2096, 124 L.Ed.2d 211 (1993). Plaintiffs seek to avoid this deferential standard of review by arguing that the tribal monopoly on class III gaming amounts to a racial preference for Native Americans, which would be subject to strict scrutiny. Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 227, 115 S.Ct. 2097, 132 L.Ed.2d 158 (1995). A racial preference violates equal protection guarantees unless it is “narrowly tailored” to “further compelling governmental interests.” Id.; see also Rice v. Cayetano, 528 U.S. 495, 120 S.Ct. 1044, 145 L.Ed.2d 1007 *732(2000) (striking down a race-based voting limitation).
In the context of a challenge to legislative classifications relating to Indians or Indian tribes, the starting point for our analysis is Morton v. Mancari, 417 U.S. 535, 94 S.Ct. 2474, 41 L.Ed.2d 290 (1974). In Mancari, the Supreme Court upheld an employment preference for Native Americans seeking positions in the Bureau of Indian Affairs (“BIA”). The class action plaintiffs, who were non-Indian applicants for BIA employment, argued that the preference amounted to invidious racial discrimination that violated their right to equal protection.. The Supreme Court noted that, if credited, the plaintiffs’ argument would call into question the entirety of Congress’ regulation of Indian affairs under Article I, Section 8 of the U.S. Constitution:
Literally every piece of legislation dealing with Indian tribes and reservations, and certainly all legislation dealing with the BIA, single out for special treatment a constituency of tribal Indians living on or near reservations. If these laws, derived from historical relationships and explicitly designed to help only Indians, were deemed invidious racial discrimination, an entire Title of the United States Code (25 U.S.C.) would be effectively erased and the solemn commitment of the Government toward the Indians would be jeopardized.
Mancari, 417 U.S. at 552-53, 94 S.Ct. 2474.
The Court concluded that strict scrutiny did not apply because the preference for Indians relied on a political, rather than a racial, classification. The hiring preference was not directed toward “a ‘racial’ group consisting of ‘Indians’; instead, it applie[d] only to members of ‘federally recognized’ tribes.” Id. at 554 n. 24, 94 S.Ct. 2474. Thus, even if an applicant demonstrated the required quantum of Indian blood, he or she still would have to show membership in a federally recognized tribe in order to qualify for the hiring preference.
The Court tied its deferential standard of review to Congress’ interest in furthering tribal sovereignty. The Indians who qualified for the preference were not racially defined, but rather were “members of quasi-sovereign tribal entities whose lives and activities are governed by the BIA in a unique fashion.” Id. at 554, 94 S.Ct. 2474. The preference was also an important part of the Indian Reorganization Act of 1934. “The overriding purpose of that particular Act was to establish machinery whereby Indian tribes would be able to assume a greater degree of self-government, both politically and economically.” Id. at 542, 94 S.Ct. 2474. The Court held that legislative classifications furthering that same purpose were political and, thus, did not warrant strict scrutiny instead of ordinary, rational-basis scrutiny: “As long as the special treatment can be tied rationally to the fulfillment of Congress’ unique obligation toward the Indians, such legislative judgments will not be disturbed. Here, where the preference is reasonable and rationally designed to further Indian self-government, we cannot say that Congress’ classification violates due process.” Id. at 555, 94 S.Ct. 2474.
Supreme Court cases decided shortly after Mancari emphasized the federal government’s interest in tribal sovereignty and the unique relationship between tribal and federal governments. In Fisher v. District Court, 424 U.S. 382, 96 S.Ct. 943, 47 L.Ed.2d 106 (1976) (per curiam), the Court upheld the exclusive jurisdiction of the Northern Cheyenne Tribal Court to handle adoptions, even though tribal litigants could not gain access to Montana’s court system. Citing Mancari, the Court *733rejected the argument that denying tribal litigants access to the state-court system constituted racial discrimination:
The exclusive jurisdiction of the Tribal Court does not derive from the race of the plaintiff but rather from the quasi-sovereign status of the Northern Cheyenne Tribe under federal law. Moreover, even if a jurisdictional holding occasionally results in denying an Indian plaintiff a forum to which a non-Indian has access, such disparate treatment of the Indian is justified because it is intended to benefit the class of which he is a member by furthering the congressional policy of Indian self-government.
Fisher, 424 U.S. at 390-91, 96 S.Ct. 943 (citing Mancari, 417 U.S. at 551-55, 94 S.Ct. 2474).
In United States v. Antelope, 430 U.S. 641, 97 S.Ct. 1395, 51 L.Ed.2d 701 (1977), two Indian defendants argued that prosecuting them under a federal criminal law, for a crime committed on Indian lands, constituted racial discrimination because the federal law allowed a lower burden of proof than the corresponding state law. Exposing them to federal prosecution, they argued, violated their right to equal protection because a non-Indian who had committed the same crime would have been prosecuted under state law, not federal law. The Court rejected that argument, elaborating on its opinion in Mancari:
[T]he principles reaffirmed in Mancari point ... to the conclusion that federal regulation of Indian affairs is not based upon impermissible classifications. Rather, such regulation is rooted in the unique status of Indians as “a separate people” with their own political institutions. Federal regulation of Indian tribes, therefore, is governance of once-sovereign political communities; it is not to be viewed as legislation of a “ ‘racial’ group consisting of ‘Indians’.... ”
Antelope, 430 U.S. at 645-46, 97 S.Ct. 1395 (quoting Mancari, 417 U.S. at 553 n. 24, 94 S.Ct. 2474).
Although discussions of Mancari arise primarily in the context of federal statutes relating to Indian tribes, in Washington v. Confederated Bands & Tribes of Yakima Indian Nation, 439 U.S. 463, 99 S.Ct. 740, 58 L.Ed.2d 740 (1979) (“Yakima III”), the Supreme Court described the circumstances in which rational-basis review applies to Indian-related state laws. The Court applied rational-basis review to an equal protection challenge to a state law where federal legislation extended that law into Indian country. Washington had enacted a law, Chapter 36, 1963 Washington Laws, to assert partial civil and criminal jurisdiction over Indian lands pursuant to the grant of authority from Congress in Public Law No. 280. Yakima III, 439 U.S. at 501, 99 S.Ct. 740. Chapter 36 asserted such jurisdiction depending on the nature of the land (e.g., tribal or reservation) and the subject matter of the underlying law (e.g., compulsory school attendance and motor vehicle registration). Id. at 475-76, 99 S.Ct. 740. The Yakima Nation challenged the resulting “checkerboard” jurisdictional pattern, claiming that it lacked a rational foundation and therefore violated the Nation’s right to equal protection. The Court sustained the classifications, applying rational-basis review:
It is settled that “the unique legal status of Indian tribes under federal law” permits the Federal Government to enact legislation singling out tribal Indians, legislation that might otherwise be constitutionally offensive. States do not enjoy this same unique relationship with Indians, but Chapter 36 is not simply another state law. It was enacted in response to a federal measure explicitly designed to readjust the allocation of jurisdiction over Indians. The jurisdic*734tion permitted under Chapter 36 is, as we have found, within the scope of the authorization of Pub. L. 280. And many of the classifications made by Chapter 36 are also made by Pub. L. 280.... For these reasons, we find the argument that such classifications are “suspect” an untenable one.... In enacting Chapter 36, Washington was legislating under explicit authority granted by Congress in the exercise of that federal power.
Id. at 500-01, 99 S.Ct. 740 (citations omitted).
Thus, when a state law applies in Indian country as a result of the state’s participation in a federal scheme that “readjusts” jurisdiction over Indians, that state law is reviewed as if it were federal law. If rationally related to both Congress’ trust obligations to the Indians and legitimate state interests, the state law must be upheld.
Our early discussions of Mancan suggested that, so long as a federal statute evinced a rational relationship to Congress’ trust obligations toward the Indians, it involved a political classification, so rational-basis review was appropriate. See, e.g., Alaska Chapter, Associated Gen. Contractors of Am., Inc. v. Pierce, 694 F.2d 1162, 1167 (9th Cir.1982) (upholding a contracting preference for native Alaskan-owned businesses for construction contracts in Alaskan villages). More recently, however, we have suggested that the political-versus-racial classification is not always easy to identify. As we noted above, Williams rejected the Secretary’s construction of the Reindeer Act. 115 F.3d at 664. Under the Secretary’s interpretation, the Reindeer Act granted a preference to individual native Alaskans, who belonged to no tribal organization, that allowed them to engage in reindeer herding anywhere in Alaska, free from competition. So construed, the Act would not relate “to Indian land, tribal status, self-government or culture.” Id. Consequently, rational-basis review would not apply and the statute most likely would constitute an impermissible racial preference. We applied the canon of constitutional avoidance to reject the Secretary’s interpretation. Id. at 663 &n. 5.
Plaintiffs’ suggestion that Williams controls the outcome of the present case ignores the obvious distinctions between an unqualified preference for individual native Alaskans and the limited preference for tribes reflected in the text of IGRA. The operative terms of IGRA expressly relate only to tribes, not to individual Indians. Only tribes, not individual Indians, may enter into compacts with other sovereign governments. Indeed, as in Mancari itself, only federally recognized tribes are covered. 25 U.S.C. § 2703(5). Further, through IGRA’s compacting process, and through its reliance on tribal governments and tribal ordinances to regulate class III gaming, the statute relates to tribal status and tribal self-government. The very nature of a Tribal-State compact is political; it is an agreement between an Indian tribe, as one sovereign, and a state, as another. The statute contemplates that the tribes must exercise their sovereign will in deciding whether to participate in class III gaming. See 25 U.S.C. § 2710(d)(1)(A) (providing for class III gaming only if authorized by a tribal ordinance or resolution).
Moreover, Congress’ express purpose was to promote “tribal economic development, self-sufficiency, and strong tribal governments.” 25 U.S.C. § 2702(1). The regulation of “gaming by an Indian tribe” also was intended to ensure that “the Indian tribe is the primary beneficiary of the gaming operation.” Id. § 2702(2). In its findings, Congress recognized that Indian tribes had been conducting gaming activi*735ties on Indian lands as a means of generating revenue for tribal governments. Id. § 2701(1). Congress created the mechanism of Tribal-State compacts to resolve the conflicting interests of the tribes and the states, which it acknowledged as “two equal sovereigns.” S.Rep. No. 100-446, at 13, reprinted at 3083.
Additionally, unlike the legislation construed in Williams, IGRA pertains only to Indian lands. 25 U.S.C. § 2710(d)(1). Like the vast majority of statutes by which Congress fulfills its obligations to the Indian tribes, IGRA regulates activities only on Indian lands. See Williams, 115 F.3d at 664 n. 6 (collecting statutes). This limitation is critical given the well-established connection between tribal lands and tribal sovereignty. “Indian tribal territory has always held a separate status under federal law. Tribes exercise substantial governing powers within their territory, they have important economic and property rights, and a number of federal laws also govern other relationships, all to the exclusion of state law.” Cohen at 27 (footnotes omitted). These governing powers and economic rights extend only as far as the borders of Indian lands. Once outside, the tribes shed their sovereignty and are fully amenable to state law. Under IGRA, for example, individual Indians (or even Indian tribes) could not establish a class III gaming establishment on non-Indian lands.16
Accordingly, IGRA falls squarely within the rule of Mancari. Williams continued to recognize that a statute relating to tribal self-government, to tribal status, or to Indian lands is subject to rational-basis review. 115 F.3d at 664. IGRA is just such a statute, notwithstanding the dictum in Williams that doubted whether Congress could give “Indians a complete monopoly on the casino industry.” Id. at 665. As our lengthy discussion of the statute has made clear, IGRA does not give “Indians” a monopoly; it neither relates to “Indians” (as distinct from federally recognized tribes) nor, itself, creates a monopoly.
In arguing against rational-basis review, Plaintiffs also rely on cases applying federal law to tribe-run businesses to argue that tribal gaming operations do not involve uniquely Indian interests.17 Their reliance is misplaced. The test used to determine whether rational-basis review applies to a federal statute’s classification — which we have been discussing here — and the test used to determine whether tribes are immune from otherwise generally applicable federal laws are distinct, with the latter being much more demanding. See Donovan v. Coeur d’Alene Tribal Farm, 751 F.2d 1113, 1116 (9th Cir.1985) (listing the three exceptions to the principle that federal statutes that are silent ón the issue of applicability to Indian tribes apply with equal force to Indians and non-Indians). Plaintiffs, in other words, conflate two distinct concepts.18
*736We conclude, then, that IGRA pertains to Indian lands and to tribal self-government and tribal status of federally recognized tribes. Accordingly, under Mancan, rational-basis review applies.
Turning next to Proposition 1A, we must apply Yakima III. Proposition 1A was enacted in response to IGRA, a federal law explicitly designed to readjust the regulatory authority of various sovereigns over class III gaming on the lands of federally recognized Indian tribes. The classifications in Proposition 1A echo those made in IGRA. In ratifying Proposition 1A, the people of California were legislating with reference to the authority that Congress had granted to the State of California in IGRA. Accordingly, rational-basis review applies to Proposition 1A as well.19
2. Rational-Basis Review
Having identified the level of deference that we owe to the legislation at issue, we turn to the merits of Plaintiffs’ equal protection challenge.
a. IGRA and the Tribal-State Compacts
IGRA is rationally related to Congress’ stated purposes of encouraging tribal autonomy and economic development. We recognized, in Alaska Chapter, 694 F.2d at 1170, that the furtherance of “an economic community” on Indian lands was a goal related to Congress’ special trust obligations. IGRA and the Tribal-State Compacts further that goal by authorizing gaming. Congress recognized that the revenue generated from pre-IGRA tribal gaming operations “often means the difference between an adequate governmental program and a skeletal program that is totally dependent on Federal funding.” S.Rep. No. 10CM46, at 3, reprinted at 3072. Thus, IGRA and the Tribal-State Compacts also are rationally related to the federal government’s interest in fostering tribal self-government.
b. Proposition 1A
Proposition 1A presents a more difficult question because it establishes the monopoly of which Plaintiffs principally complain. We conclude, nonetheless, that Proposition 1A passes the rational-basis test.
States have wide latitude in establishing classifications to balance interests and remedy perceived problems:
“The problem of legislative classification is a perennial one, admitting of no doctrinaire definition. Evils in the same field may be of different dimensions and proportions, requiring different remedies. Or so the legislature may think. Or the reform may take one step at a time, addressing itself to the phase of the problem which seems most acute to the legislative mind. The legislature may select one phase of one field and apply a remedy there, neglecting the others. The prohibition of the Equal Protection Clause goes no further than the invidious discrimination.”
Beach Communications, 508 U.S. at 316, 113 S.Ct. 2096 (quoting Williamson v. Lee Optical of Okla., Inc., 348 U.S. 483, 489, 75 S.Ct. 461, 99 L.Ed. 563 (1955)). Where there exists an appropriate connection to the state’s police power, even the grant of *737a monopoly does not, in itself, offend equal protection principles. See, e.g., City of New Orleans v. Dukes, 427 U.S. 297, 300, 96 S.Ct. 2513, 49 L.Ed.2d 511 (1976) (per curiam) (reversing an appellate court’s holding that a “grandfather clause” exemption from a ban on pushcart vendors, which amounted to the “creation of a protected monopoly for the favored class member,” violated equal protection (internal quotation marks omitted)); Pac. States Box & Basket Co. v. White, 296 U.S. 176, 184, 56 S.Ct. 159, 80 L.Ed. 138 (1935) (holding that “the grant of a monopoly, if otherwise an appropriate exercise of the police power, is not void as denying equal protection of the law”).
California has two legitimate interests to which Proposition 1A bears a rational connection. The first is the regulation of “vice” activity — a function that lies at the heart of a state’s police powers — by permitting certain forms of gambling only on the lands of sovereign tribal entities that enter into government-to-government compacts with the State. The second is to promote cooperative relationships between the tribes and the State by fostering tribal sovereignty and self-sufficiency.
(i) Local Prohibitions and Exceptions for “Vice” Activities
The circuits that have given significant attention to equal protection challenges to state gambling laws have, by and large, held that “the regulation of gambling lies at the heart of the state’s police power.” Helton v. Hunt, 330 F.3d 242, 246 (4th Cir.), cert. denied, — U.S.-, 124 S.Ct. 436, 157 L.Ed.2d 312 (2003) (internal quotation marks omitted). In Helton, the court held that statutes providing for the destruction of certain gaming machines and the prosecution of their owners did not violate equal protection principles. See also Casino Ventures v. Stewart, 183 F.3d 307, 310 (4th Cir.1999) (noting that, because gambling restrictions “are aimed at promoting the welfare, safety, and morals of South Carolinians, they represent a well-recognized exercise of state police power”); United States, v. Williams, 124 F.3d 411, 423 (3d Cir.1997) (upholding a Pennsylvania statute prohibiting certain gaming activities). Our own precedent, although not discussing the issue in great detail, is largely in accord. See Jacobson v. Hannifin, 627 F.2d 177, 180 (9th Cir.1980) (upholding a Nevada law giving the state’s gaming commission the power to deny licenses for gaming establishments “for any cause deemed reasonable by such commission”).
Upholding a local advertising ban on casino gambling in Puerto Rico, the Supreme Court placed gambling in a category of “products or activities deemed harmful, such as cigarettes, alcoholic beverages, and prostitution,” which state legislatures have dealt with by means ranging from location restrictions to outright prohibition. Posadas de P.R. Assocs. v. Tourism Co. of P.R., 478 U.S. 328, 346-47, 106 S.Ct. 2968, 92 L.Ed.2d 266 (1986).20 Using the shorthand term “vice activity” to refer to these products and activities, the Court has re*738jected both First Amendment and equal protection challenges to legislative classifications aimed at reducing specific harms associated with vice activity. United States v. Edge Broad. Co., 509 TJ.S. 418, 426, 113 S.Ct. 2696, 125 L.Ed.2d 345 (1993). Such classifications have withstood scrutiny whether they distinguish between States, political subdivisions within a state, or establishments within the same locality.
In Salsburg v. Maryland, 346 U.S. 545, 74 S.Ct. 280, 98 L.Ed. 281 (1954), the Supreme Court upheld a Maryland statute that excepted only certain localities from the State’s exclusionary rule of evidence. The exception applied only to “ ‘the use of such evidence in Anne Arundel, Wicomico and Prince George’s Counties in the prosecution of any person for a violation of the gambling laws contained in [state statutes].”’ Id. at 548 n. 2, 74 S.Ct. 280 (quoting Md. Ann.Code of 1951, art. 35, § 5). Drawing from local-option laws regulating the sale of alcohol, Salsburg held that Maryland
could validly grant home rule to each of its 23 counties and to the City of Baltimore to determine this rule of evidence by local option. It is equally clear, although less usual, that a state legislature may itself determine such an issue for each of its local subdivisions, having in mind the needs and desires of each.
Id. at 552 & n. 7, 74 S.Ct. 280 (footnote omitted). The Court held that the state had acted reasonably in singling out the counties excepted from the evidentiary rule because that selection was related to population concentration. Particularly in the field of criminal law, states have long-established discretion to enact distinctions of this kind. See id. at 553-54, 74 S.Ct. 280.
Local-option laws in the area of liquor control have shown particular resilience against equal protection challenges. Cases upholding such laws teach that a state constitutionally can allow a local subdivision to elect a more stringent regulatory approach or outright prohibition, or conversely can exempt a local subdivision from a state-wide ban on vice activity. Although the analysis has changed somewhat since Rippey v. Texas, 193 U.S. 504, 24 S.Ct. 516, 48 L.Ed. 767 (1904), and Eberle v. Michigan, 232 U.S. 700, 34 S.Ct. 464, 58 L.Ed. 803 (1914), the judicial trend still favors the validity of local-option laws as exercises of the state’s police power. For example, the Seventh Circuit, in Philly’s, Inc. v. Byrne, 732 F.2d 87 (7th Cir.1984), upheld a local-option law allowing voters within a given Chicago city precinct to vote that precinct “dry.” The court rejected a due process challenge by restaurant owners in Chicago who had lost their liquor licenses by popular vote. Id. at 90.
The Sixth Circuit followed a similar approach in rejecting an equal protection challenge to a local-option law. In 37712, Inc. v. Ohio Department of Liquor Control, 113 F.3d 614, 621-22 (6th Cir.1997), the court applied rational-basis review to uphold an Ohio statute allowing counties to go “dry” by local-option elections. It held that, so long as the police power is “not ... abused to violate a person’s federal constitutional rights,” drawing distinctions between political subdivisions or allowing those subdivisions to distinguish themselves by an exercise of their own political will is permissible state action that survives rational-basis scrutiny. Id. at 618 (citing 44 Liquormart, Inc. v. Rhode Island, 517 U.S. 484, 515-16, 116 S.Ct. 1495, 134 L.Ed.2d 711 (1996)).
The plaintiffs in 37712, Inc., were owners of bars and carry-out stores. They challenged the local-option law because the State of Ohio had issued licenses exempting certain breweries from the law. The *739plaintiffs asserted that their establishments were “similarly situated” to the breweries exempted from the ban and that the legislative distinction was therefore arbitrary. Rejecting their claim, the Sixth Circuit credited the legislative judgment that beer sales from the licensed establishments did not involve the same risks of disorderly conduct and crime posed by sales from the unlicensed establishments. Id. at 621. Although the safer course would be to ban all sales, the court noted that the “ ‘all or nothing’ ” approach was not the only one that would survive rational-basis review. Id. at 622.
The same analysis applies where states, or local voters, choose to exempt a political subdivision from state-wide prohibitions on vice activity. Quite recently, the Georgia Supreme Court upheld state statutes that give local governments the option of excepting certain businesses and venues from a state-wide ban on the sale of alcohol on Sunday. State v. Heretic, Inc., 277 Ga. 275, 588 S.E.2d 224 (2003). The plaintiffs, who were bar owners whose establishments did not qualify for the exception, alleged that the discriminatory application of the exception amounted to a denial of equal protection.
Rejecting that challenge, the Georgia Supreme Court relied in part on McGowan v. Maryland, 366 U.S. 420, 81 S.Ct. 1101, 6 L.Ed.2d 393 (1961). In that case the United States Supreme Court upheld a state-law exception of certain retail establishments in only one county from a state-wide ban on the sale of specified merchandise on Sunday. In addition to challenging the discriminatory treatment of different establishments in the same county, the plaintiffs in McGowan challenged the statute because the exemption applied to only one county in the entire state. The Supreme Court rejected both arguments. As to the distinction between vendors, the Court held that, given the state’s legitimate concerns, such as enforcement problems, the classification was reasonable. Id. at 428, 81 S.Ct. 1101. As to the preferential treatment of vendors in Anne Arundel County, the Court held that “territorial uniformity is not a constitutional prerequisite” and that “the prescription of different substantive offenses in different counties is generally a matter for legislative discretion.” Id. at 427, 81 S.Ct. 1101.
In addition to relying on McGowan, the state court in Heretic rested its decision on a distinction between vice activities and generic economic enterprises, rejecting the plaintiffs’ argument that their claim was controlled by the court’s prior holding that no rational basis supported a state law prohibiting furniture stores from operating on Sunday. The Heretic court was unpersuaded by an analogy between alcohol and furniture: “The present case is distinguishable because the state lacks the interest in regulating the sale of furniture that it has in regulating the sale of alcohol, which poses significant risks to the health and safety of the general public.” 588 S.E.2d at 225-26.
Similarly, we are unpersuaded by Plaintiffs’ argument that allowing California to grant to tribes a monopoly on class III gaming operations, restricted to Indian lands, necessarily will lead to Indian monopolies on other forms of economic activity. As a practical matter, Congress viewed gambling as a “unique form of economic enterprise” and was “strongly opposed to the application of the jurisdictional elections authorized by this bill to any other economic or regulatory issue that may arise between tribes and States in the future.” S.Rep. No. 100^146, at 14, reprinted at 3084. As a constitutional matter, the state interests that justify, as a valid exercise of a state’s police power, California’s restriction of class III gaming *740operations to those conducted by Indian tribes on Indian lands are absent in the field of generic commercial activities. Most economic activities historically have not been deemed harmful. Thus, we do not believe that Plaintiffs’ parade of horri-bles — tribal monopolies on automobile dealerships, for example — is a likely consequence of our conclusion that legitimate state interests support a restriction of casino-style gambling to Indian lands. Of course, in this case we need not and do not decide whether states constitutionally could grant tribes exclusive rights to conduct enterprises other than casino gambling.
Were the tribal lands a political subdivision of the State, California’s exemption of tribal lands from its state-wide prohibition on class III gaming activities easily would withstand constitutional scrutiny. When enacting substantive regulations or prohibitions of vice activities, the interests implicated lie “at the heart of the state’s police power.” Helton, 330 F.3d at 246 (internal quotation marks omitted). With regard to these activities, a state is free to enact legislation that accords different treatment to different localities, and even to different establishments within the same locality, so long as that differentiation is tied to a legitimate interest in the health, safety, or welfare of its citizens. The state may make such distinctions by local-option laws, or by making the distinction between different areas itself. It may impose more stringent regulations by way of local restrictions, or it may exempt an area entirely from a state-wide ban on products and activities falling in the Supreme Court’s “vice activity” category. Unless such legislative distinctions infringe fundamental rights or involve suspect classifications, they generally survive equal protection analysis.
In the light of its stated concerns, California’s exception of operations on tribal lands from its long-standing prohibition on class III gaming passes constitutional muster. Before Proposition 1A was ratified, California absolutely banned casino-style gaming. See Hotel Employees, 88 Cal.Rptr.2d 56, 981 P.2d at 996 (tracing the history of California’s gambling prohibitions back to 1849). Thus, its regulation of gambling does not involve the State’s attaching a “vice” label to an activity without a corresponding prohibition. See 44 Liquormart, 517 U.S. at 514, 116 S.Ct. 1495. California has expressed its legislative judgment that “[ujnregulated gambling enterprises are inimical to the public health, safety, welfare, and good order.” Cal. Bus. & Prof.Code § 19801(c)(1). By limiting class III gaming to tribal lands, Proposition 1A, and the compacts negotiated pursuant to it, foster California’s “legitimate sovereign interest in regulating the growth of Class III gaming activities in California.” Tribal-State Compact, pmbl. F.
Further, limiting class III gaming operations to those run by tribes is reasonably designed to defend against the criminal infiltration of gaming operations. By restricting large-scale gambling enterprises to carefully limited locations, California furthers its purpose of ensuring that such gaming activities “are free from criminal and other undesirable elements.” Id.
Thus, a rational basis exists for California’s decision to restrict class III gaming operations to those conducted by Indian tribes on their own lands. California could, of course, pursue these interests even more effectively by banning class III gaming altogether. However, as discussed above, rational-basis review does not require states to choose an all-or-nothing approach. It requires only that the means chosen are reasonable.
*741(ii) Indian Sovereignty
The Indian tribes with whom the State of California has entered into compacts for class III gaming are sovereigns. By allowing Indian tribes to make their own decisions about the desirability of gaming operations on Indian lands, and by allowing Indian tribes to regulate those operations by tribal ordinance, the State of California sought “a mutually respectful government-to-government relationship that will serve the mutual interests of [the tribes and the State].” Tribal-State Compact § 1.0(a).
The selective application of state law to Indian lands is not unprecedented. In Yakima III, the Supreme Court expressly approved of jurisdictional schemes that balance a state’s interest in protecting its citizens with a tribe’s interest in exercising sovereignty over its lands:
The land-tenure classification made by the State is neither an irrational nor arbitrary means of identifying those areas within a reservation in which tribal members have the greatest interest in being free of state police power. Indeed, many of the rules developed in this Court’s decisions in cases accommodating the sovereign rights of the tribes with those of the States are strikingly similar. In short, checkerboard jurisdiction is not novel in Indian law, and does not, as such, violate the Constitution.
439 U.S. at 502, 99 S.Ct. 740.
Like Public Law No. 280, IGRA’s jurisdictional framework gives the State of California a role in identifying those areas where Indian tribes have a substantial interest in being free of state police power. It is rational for Californians to be willing to recognize the separate sovereign interests of the tribes and to allow the tribes to make a different moral and economic choice than is made by the State as a whole. By executing the Tribal-State Compacts, California has sought to bring about a “new era of tribal-state cooperation in areas of mutual concern.” Tribal-State Compact, pmbl. D. California’s decision to grant to tribes a monopoly on class III gaming activities is rationally related to both the State’s interest in protecting its citizens from the particular harms associated with large-scale gaming operations and the State’s interest in fostering relations with tribes as separate sovereigns.
In addition, the compacts that California has negotiated with the tribes are designed to further the purposes of IGRA, which include “promoting tribal economic development, self-sufficiency, and strong tribal governments.” 25 U.S.C. § 2702(1). By providing exclusive rights to engage in class III gaming, California gives Indian tribes valuable tools to promote the general welfare of their members. Class III gaming helps generate jobs and revenues to support the governmental services and programs of the tribes that enter into compacts. Further, California’s regulatory scheme benefits nongaming tribes because they receive distributions from the funds that the State requires gaming tribes to allocate to the Indian Gaming Revenue Sharing Trust. “This mechanism is intended to ensure that the wealth generated by gambling is not just limited to tribes who have entered into compacts, but will be redistributed for the benefit of all tribes in California.” Flynt, 129 Cal.Rptr.2d at 183. Thus, we agree with the district court’s finding that “California’s compacts with the tribes are rationally related to the furtherance of Congress’ unique obligation to the tribes.” Artichoke Joe’s, 216 F.Supp.2d at 1129.
In summary, Congress acted rationally in balancing the sovereign interests of tribes and states. The State of California, which historically had banned casino gambling altogether, acted rationally in limit*742ing the placement and concentration of class III gaming operations to Indian lands and in recognizing the sovereign interest of federally recognized Indian tribes to choose a different path on Indian lands. Accordingly, we find no violation of equal protection.
CONCLUSION
We hold that IGRA, Proposition 1A, and the Tribal-State Compacts are consistent with IGRA.
We also hold that Proposition 1A and the Tribal-State Compacts do not violate Plaintiffs’ rights to equal protection of the laws.
Accordingly, the judgment of the district court in favor of Defendants is
AFFIRMED.
. The complaint names the following state defendants: the Governor of California, Attorney General of California, the Director of the California Division of Gambling Control, and the members of the California Gambling Control Commission.
. The Equal Protection Clause of the Fourteenth Amendment provides a basis for Plaintiffs’ equal protection claim against the state defendants. The Due Process Clause of the Fifth Amendment applies to the actions of the federal defendants. Under the “congruence principle” in equal protection jurisprudence, "[e]qual protection analysis in the Fifth Amendment area is the same as that under the Fourteenth Amendment.” Buckley v. Valeo, 424 U.S. 1, 93, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976) (per curiam). We use the term “equal protection” to refer to Plaintiffs' equal protection claims against both the federal and the state defendants.
.Recently, we decided Chemehuevi Indian Tribe v. California (In re Indian Gaming Related Cases), 331 F.3d 1094 (9th Cir.2003), petition for cert. filed, 72 U.S.L.W. 3407 (U.S. Dec. 1, 2003) (No. 03-804). That opinion presents a detailed history of Indian gaming in California both before and after IGRA was enacted. We paraphrase those portions of that history necessary to an understanding of the present dispute.
. “In banked or percentage card games, players bet against the 'house' or the casino. In ‘nonbanked’ or 'nonpercentage’ card games, the 'house' has no monetary stake in the game itself, and players bet against one another.” Artichoke Joe's, 216 F.Supp.2d at 1092 n. 3.
. IGRA provides that class III Indian gaming must be:
(A) authorized by an ordinance or resolution that—
(i) is adopted by the governing body of the Indian tribe having jurisdiction over such lands,
*716(ii) meets the requirements of subsection (b) of this section, and
(iii) is approved by the Chairman,
(B) located in a State that permits such gaming for any purpose by any person, organization, or entity, and
(C) conducted in conformance with a Tribal-State compact entered into by the Indian tribe and the State under paragraph
(3) that is in effect.
25 U.S.C. § 2710(d)(1).
. The NIGC is a federal regulatory agency, created by IGRA, that oversees the business of Indian gaming in order to ensure its lasting integrity. The NIGC performs a variety of functions, such as the review of management contracts between tribes and outside parties to run tribal casinos. 25 U.S.C. § 2704.
. Although the Supreme Court invalidated this provision of IGRA in Seminole Tribe of Florida v. Florida, 517 U.S. 44, 116 S.Ct. 1114, 134 L.Ed.2d 252 (1996), on Eleventh Amendment grounds, California has consented to such suits by waiving sovereign immunity expressly. Cal. Gov’t Code § 98005.
. Count One of Plaintiffs' complaint also included a claim that the Secretary's approval of the Tribal-State Compacts violated the Administrative Procedure Act, 5 U.S.C. §§ 701-706. The district court granted Defendants' motion for summary judgment on that claim, and Plaintiffs do not appeal that ruling.
. No party appeals the district court’s rulings regarding justiciability. We agree with the district court's cogent application of U.S. Supreme Court precedent regarding constitutional standing, Artichoke Joe’s, 216 F.Supp.2d at 1100-09, and thus do not address Plaintiffs' claims as to the Prospective Compacts or their claims against the Gambling Control Commission.
. In the district court, amicus curiae California Nations Indian Gaming Association argued that Plaintiffs' complaint must be dismissed for failure to join California’s Indian tribes as indispensable parties under Federal Rule of Civil Procedure 19. In the absence of exceptional circumstances, which are not present here, we do not address issues raised only in an amicus brief. Swan v. Peterson, 6 F.3d 1373, 1383 (9th Cir.1993).
We note, however, that this case is distinguishable from an earlier challenge to the validity of gaming compacts entered into by the Governor of Arizona pursuant to IGRA. Am. Greyhound Racing, Inc. v. Hull, 305 F.3d 1015 (9th Cir.2002). We held there that the State of Arizona could not adequately represent the tribes because their interests were potentially adverse and because the state owed no trust responsibility to Indian tribes. Id. at 1024 n. 5. By contrast, the Secretary is a party to this case. The Secretary's interests are not adverse to the tribes' interests and the Department of Interior has the primary responsibility for carrying out the federal government's trust obligation to Indian tribes.
. The Johnson Act, 15 U.S.C. §§ 1171-1178, prohibits the possession or use of “gambling device[s],” including slot machines, within Indian country. Id. § 1171(a). IGRA waives application of the Johnson Act if the slot-machine gaming is conducted under an effective Tribal-State compact that "is entered into ... by a State in which gambling devices are legal.” 25 U.S.C. § 2710(d)(6). Whether slot machines are legal in California depends on the validity of Proposition 1A under IGRA, so the Johnson Act analysis is subsumed by the analysis, below, of the “permits such gaming” provision of § 2710(d)(1)(B).
. Exclusivity figures prominently in this exchange. Section 12.4 of the compact grants the tribes a right to terminate the compact "in the event the exclusive right of Indian tribes to operate Gaming Devices in California is abrogated by the enactment, amendment, or repeal of a state statute or constitutional provision” or by judicial decision. See In re Indian Gaming Related Cases, 331 F.3d at 1104 n. 13.
. Connecticut and New York have also granted to Indian tribes monopolies on class III gaming, conditioned on revenue-sharing agreements. See Gatsby Contreras, Note, Exclusivity Agreements in Tribal-State Compacts: Mutual Benefit Revenue-Sharing or Illegal Taxation?, 5 J. Gender Race & Just. 487, 496-500 (2002) (canvassing various tribal-state exclusivity and revenue-sharing models).
. A third staled purpose is to set up the National Indian Gaming Commission. 25 U.S.C. § 2702(3).
. Recently, we applied a Blackfeet presumption in deciding whether the Age Discrimination in Employment Act of 1967 (“ADEA”) applies to the relationship between a tribal housing authority, which supplies affordable housing solely on tribal trust lands, and its employees. EEOC v. Karuk Tribe Hous. Auth., 260 F.3d 1071, 1082 (9th Cir.2001). Traditional rules of statutory construction were consistent with the claim that the ADEA applies to Indian tribes. Id. But we reached the opposite conclusion, applying a doctrine unique to Indian law, under which a federal statute silent as to its applicability to Indian tribes will not be found to intrude on a tribe's right to self-governance in purely intramural affairs. Id.
. The present case concerns class III gaming operations that are located on Indian reservations or Indian trust lands. Thus, we need not and do not decide whether lands that are purchased specifically for the purpose of conducting class III gaming activities are "Indian lands" within the meaning of IGRA.
. See, e.g., Dawavendewa v. Salt River Project Agric. Improvement & Power Dist., 154 F.3d 1117 (9th Cir.1998) (holding that Title VII applies to the hiring practices of a tribe-run corporation); Reich v. Mashantucket Sand & Gravel, 95 F.3d 174 (2d Cir.1996) (holding that OSHA regulations apply to construction company owned and operated by an Indian tribe).
.Plaintiffs' reliance on Confederated Tribes of Siletz Indians v. Oregon, 143 F.3d 481 (9th Cir.1998), also is misplaced. Although we rejected the argument that Indian casinos "involve[d] only tribal members involved in on-reservation conduct,” we did so in the context *736of determining that, as a matter of contract law, a state reporting statute applied to the tribe's gaming operations by virtue' of the Tribal-State compact itself. Id. at 486-87.
. That Proposition 1A, unlike IGRA, expressly creates a tribal monopoly on class III gaming activities does not alter our conclusion. California's grant of a monopoly to the tribes is relevant to the merits, discussed below, but has no bearing on identifying the appropriate standard under which to review the state law.
. A decade later, the Court disavowed Posa-das ’ First Amendment holding but otherwise left its analysis of casino gaming intact. 44 Liquormart, Inc. v. Rhode Island, 517 U.S. 484, 511, 116 S.Ct. 1495, 134 L.Ed.2d 711 (1996). In Greater New Orleans Broadcasting Ass’n v. United States, 527 U.S. 173, 119 S.Ct. 1923, 144 L.Ed.2d 161 (1999), the Court held that 18 U.S.C. § 1304, which bans the broadcast of lottery and gambling advertisements in certain circumstances, could not be applied to ban broadcast advertisements for private casinos in states where such casinos are legal. The Court noted the availability of "practical and nonspeech-related forms of regulation,” including "location restrictions," to "alleviate some of the social cost of casino gambling.” Id. at 192, 119 S.Ct. 1923 (emphasis added).