dissenting:
In 1999, the Rincon San Luiseno Band of Mission Indians (“the Band”) negotiated a compact with the State of California to operate a casino — known as Harrah’s Rincon Hotel & Casino — in San Diego County. Under the terms of that compact, the Band was authorized to operate up to 1,600 Nevada-style gaming devices. The compact remains in effect until 2020. In *10432003, the Band requested that California renegotiate the compact so that the Band could increase the number of devices from 1,600 to 2,500. Throughout an exchange of proposals, the State demanded that the Band contribute a percentage of its gaming revenues to the State’s general fund. The negotiations failed, resulting in no increase in gaming devices for the Band and no revenues for the State. Unhappy with the course of the negotiations, the Band brought suit under the Indian Gaming Regulatory Act (“IGRA”), 25 U.S.C. § 2701 et seq., claiming that the State had negotiated in bad faith and requesting mandatory mediation.
In In re Indian Gaming Related Cases (“Coyote Valley II”), we held that when “a State offers meaningful concessions in return for fee demands----it exercises its authority to negotiate, which IGRA clearly permits.” 331 F.3d 1094, 1112 (9th Cir.2003). Going well beyond anything we said in Coyote Valley II, the majority holds (1) the State’s demand for “general fund revenue sharing [is] a bald demand for payment of a tax,” Maj. Op. at 1032 (emphasis omitted); (2) “general fund revenue sharing ... is ... not an authorized subject of negotiation” under IGRA, id. at 1034; and therefore (3) the State’s “demand for payment of a tax [is] ... evidence of [the State’s] bad faith,” id. at 1032. The majority concludes that the State has negotiated in bad faith and orders the parties to amend their compact or submit to mediation.
I respectfully disagree with all of these propositions. First, we have long held that the power to tax is defined by the sovereign’s power to impose the tax. California has exercised no such power here. The majority has confused California’s hardball negotiations with the taxing power. California has not claimed any authority to tax the Band and, if these negotiations fail, the Band will not have added one penny to the State’s coffers. Second, IGRA is silent on the particular question whether states and tribes may include revenue sharing in their negotiations. But the Act provides that the parties “may include provisions relating to ... any other subjects that are directly related to the operation of gaming activities,” 25 U.S.C. § 2710(d)(3)(C)(vii), and we held in Coyote Valley II that the parties could negotiate over “fee demands.” The majority now holds that such fee demands cannot include general revenue sharing derived from the operation of gaming activities, but are limited to fees spent on the regulation of gaming alone. This restriction takes from the State its primary incentive in negotiating gaming compacts with the tribes. Third, even if I thought California had imposed a tax on the Band, I do not believe the State has negotiated in bad faith. By its constitution, California has granted the tribes the exclusive right to offer Nevada-style gaming in the State. That structure puts the State and the Band into a bilateral monopoly situation in which each side has powerful economic incentives to negotiate at the margins. At the least, the State has offered the Band a more than fifty percent increase in the number of authorized gaming machines and a twenty-five year extension on its contract on terms comparable to other tribes with California compacts. The offered terms are well within the range of good faith negotiations. Of course, whether the Band should accept the State’s offer is a very different question from whether the State has negotiated in bad faith. But absent the majority’s intervention, both sides would have strong incentives to continue negotiations.
The impact of the majority’s decision may not be readily apparent from its opinion, but the holding that California has negotiated in bad faith because its demand for general revenue sharing is a tax and *1044not an authorized subject of negotiation does not just upset the apple cart — it derails the whole train. If the majority is correct, then there is nothing for California to do but to authorize whatever devices the Band wants. The Band wins. Everything.
Moreover, the damage is not confined to the Rincon Band. The revenue sharing provision that the majority strikes from the negotiations is found in fifteen other compacts that California has recently negotiated or renegotiated with tribes. All of those compacts were approved by the Secretary of the Interior. Those tribes now have a powerful argument that their compacts must be renegotiated (again) in light of the majority’s decision. The damage, moreover, is not confined to our circuit. Every state that has negotiated a compact in recent years — including Connecticut, Florida, Michigan, New York, New Mexico, Oklahoma, and Wisconsin-has negotiated a similar revenue sharing provision, one that was also approved by the Secretary. The result is going to be chaos as tribe after tribe seeks to reopen negotiations concluded and duly approved. Nothing in IGRA compels our intervention in these negotiations.
I respectfully dissent.
I
Since the framing of the Constitution, Indian tribes have occupied a “unique legal posture ... in relation to the federal government.” William C. Canby, Jr., American Indian Law 1 (5th ed.2009). “[Tjribes are independent entities with inherent powers of self-government,” but “the independence of the tribes is subject to exceptionally great powers of Congress,” which “has a responsibility for the protection of the tribes and their properties....” Id. at 1-2. The presence of fifty sovereign states complicates the relationship: “the power to deal with and regulate the tribes is wholly federal; the states are excluded unless Congress delegates power to them.” Id. at 2. “[I]n assessing state regulation that does not involve taxation,” the Supreme Court has taken a functional approach to try to “balance[ ] federal, state, and tribal interests.” Okla. Tax Comm’n v. Chickasaw Nation, 515 U.S. 450, 458, 115 S.Ct. 2214, 132 L.Ed.2d 400 (1995). However, state taxation of Indians or Indian tribes, where a “State attempts to levy a tax directly on an Indian tribe or its members inside Indian country,” has produced, “instead of a balancing inquiry, a more categorical approach.” Id. at 458, 115 S.Ct. 2214 (quotation marks omitted).1
The Supreme Court’s approach to tribal sovereign immunity has evolved over time. As Chief Justice Marshall explained in Worcester v. Georgia, 31 U.S. 515, 6 Pet. 515, 8 L.Ed. 483 (1832), at the time of the framing there was a “universal conviction that the Indian nations possessed a full right to the lands they occupied, until that right should be extinguished by the United States, with their consent: that their territory was separated from that of any state *1045■within whose chartered limits they might reside, by a boundary line, established by treaties: that, within their boundary, they possessed rights with which no state could interfere: and that the whole power of regulating the intercourse with them, was vested in the United States.” Id. at 560.2 By the late twentieth century, the Supreme Court had “rejected] ... the broad assertion that the Federal Government has exclusive jurisdiction over the Tribe for all purposes and that the State is therefore prohibited from enforcing its revenue laws against any tribal enterprise whether the enterprise is located on or off tribal land” in favor of “more individualized treatment of particular treaties and specific federal statutes, including statehood enabling legislation, as they, taken together, affect the respective rights of States, Indians, and the Federal Government.” Mescalero Apache Tribe v. Jones, 411 U.S. 145, 147-48, 93 S.Ct. 1267, 36 L.Ed.2d 114 (1973) (quotation marks omitted) (citing McClanahan v. Ariz. State Tax Comm’n, 411 U.S. 164, 93 S.Ct. 1257, 36 L.Ed.2d 129 (1973)). But “[ejven so, in the special area of state taxation, absent cession of jurisdiction or other federal statutes permitting it,” the Supreme Court found “no satisfactory authority for taxing Indian reservation lands or Indian income from activities carried on within the boundaries of the reservation,” holding that “such taxation [wa]s not permissible absent congressional consent.” Id. at 148, 93 S.Ct. 1267.
In short, “[ajbsent cession of jurisdiction or other federal statutes permitting it ... a State is without power to tax reservation lands and reservation Indians.” Okla. Tax Comm’n, 515 U.S. at 458, 115 S.Ct. 2214 (quotation marks omitted). A tax levied directly on persons or property is known as a “direct tax,” and “Indian tribes, like the Federal Government itself, are exempt from direct state taxation and ... this exemption is lifted only when Congress has made its intent to do so unmistakably clear.” Cotton Petroleum Corp. v. New Mexico, 490 U.S. 163, 183 n. 14, 109 S.Ct. 1698, 104 L.Ed.2d 209 (1989) (quotation marks omitted).
II
Against this intricate backdrop of federal-state-tribal relations, tribal gaming emerged.
A
In the 1970s, bingo halls began to open on tribal land, often in derogation of state gambling laws. See Coyote Valley II, 331 F.3d at 1095; Flynt v. Cal. Gambling Control Comm’n, 104 Cal.App.4th 1125, 129 Cal.Rptr.2d 167, 172-73 (2002). When states attempted to regulate or shut down tribal bingo halls, the tribes argued that their sovereign status immunized them from state gambling regulation. Id.; see Borona Group of Capitan Grande Band of Mission Indians v. Duffy, 694 F.2d 1185 (9th Cir.1982); Seminole Tribe v. Butterworth, 658 F.2d 310, 312 (5th Cir.1981), cert. denied, 455 U.S. 1020, 102 S.Ct. 1717, 72 L.Ed.2d 138 (1982); Oneida Tribe of Indians v. Wisconsin, 518 F.Supp. 712 (W.D.Wis.1981). Tribal bingo halls — and state attempts to regulate them — attracted national attention as the stakes and profit from these establishments skyrocketed. See James Coates, U.S. fears big-money bingo will draw mob to Indian tribes, Chi. Trib., Dec. 1, 1985, at 4; Howard LaFran*1046chi, States vs. Indian Bingo, Christian Sci. Monitor, Feb. 24, 1986, at 3; Lenore Look, Hearing to Air Complaints on Misuse of Tribes’ Bingo Revenue, L.A. Times, May 24, 1985, at Al; Iver Peterson, Bingo That Doesn’t Get Any Bigger, N.Y. Times, Nov. 3,1985, at E4.
California was right in the middle of the bingo battles: in the mid-1980s, the State attempted to enforce its gambling statutes against tribes operating bingo halls within its borders, and the tribes responded by filing suit in federal court, arguing that the State did not have regulatory jurisdiction over tribal lands. Coyote Valley II, 331 F.3d at 1095. In response, the State argued that Congress had expressly consented to the State’s exercise of jurisdiction over tribal gaming activities in Public Law 83-280, which granted five states, including California, jurisdiction over criminal violations on tribal land. Pub.L. 83-280 (Aug. 15, 1953), codified as amended at 18 U.S.C. § 1162, 28 U.S.C. § 1360, and 25 U.S.C. §§ 1321-1326; see Coyote Valley II, 331 F.3d at 1095. In 1987, the Supreme Court ruled in favor of the tribes, holding that Public Law 83-280 did not expressly authorize state jurisdiction over activities on tribal land that were regulated, rather than wholly prohibited, under state law. See California v. Cabazon Band of Mission Indians, 480 U.S. 202, 210-12, 107 S.Ct. 1083, 94 L.Ed.2d 244 (1987). Since “California regulate[d] rather than prohibit[ed] gambling in general and bingo in particular,” the State’s attempts to regulate tribal bingo were not authorized by Public Law 280, id. at 211, 107 S.Ct. 1083, and “[sjtate regulation would impermissibly infringe on tribal government,” id. at 222,107 S.Ct. 1083.
Legislation concerning tribal gaming had percolated in Congress for several years prior to Cabazon, but prior to the Court’s decision in that case, it was unclear that congressional action was required to allow state regulation of tribal gaming. The Cabazon decision answered this question in the affirmative, and the year after Cabazon, Congress passed the Indian Gaming Regulatory Act, Pub.L. 100-497 (Oct. 17, 1988), codified at 25 U.S.C. § 2701 et. seq.
B
In IGRA, Congress created a “cooperative federalist” framework that “balance[d] the competing sovereign interests of the federal government, state governments and Indian tribes, by giving each a role in the regulatory scheme.” Coyote Valley II, 331 F.3d at 1096 (quoting Artichoke Joe’s v. Norton, 216 F.Supp.2d 1084, 1092 (E.D.Cal.2002)). Recognizing that different types of gaming presented different regulatory burdens, IGRA outlined three classes of gaming and gave the respective sovereigns differing degrees of regulatory authority with respect to each gaming class. States were given very little regulatory authority over what Congress deemed class I and class II gaming — consisting of social games with prizes of minimal value, 25 U.S.C. § 2703(6), and bingo and certain non-banked card games, id. §§ 2703(7)(A)(B), respectively. States could not regulate class I gaming on tribal lands at all, id. § 2710(a)(1), and if a state allowed any class II gaming within its borders, it had to allow tribes to conduct such class II gaming on tribal lands subject only to tribal and federal regulation, id. §§ 2710(a)(2) and 2710(b)(l)(A)-(B). By contrast, class III gaming — consisting of banked card games such as blackjack and baccarat, slot machines and other electronic gaming machines, and other Nevada-style gaming activities, id. §§ 2710(b)(7)(B) and 2710(b)(8) — was “subject to a greater degree of federal-state regulation than either class I or class II gaming.” Coyote Valley II, 331 F.3d at 1097. In particular, IGRA allowed class *1047III gaming activities on tribal lands only (1) “in a State that permits such gaming for any purpose by any person, organization, or entity,” 25 U.S.C. § 2710(d)(1)(B), and (2) “conducted in conformance with a Tribal-State compact entered into by the Indian tribe and the State under[25 U.S.C. § 2710(d)(3) et seg.] that is in effect,” id. § 2710(d)(1)(C).
IGRA’s compact requirement represented a unique approach to a unique situation. The prospect of class III gaming on tribal lands implicated important — and potentially divergent — interests of two sovereign parties, states and tribes. At the same time, however, Congress recognized that the respective interests of states and tribes with respect to class III gaming were neither identical nor mutually exclusive. Faced with this situation, Congress came up with a solution as old as the free market itself: allow states and tribes to bargain over the terms of their class III gaming relationship. As the Senate Report for IGRA explained:
In the Committee’s view, both State and tribal governments have significant governmental interests in the conduct of class III gaming. States and tribes are encouraged to conduct negotiations within the context of the mutual benefits that can flow to and from tribe [sic ] and States. This is a strong and serious presumption that must provide the framework for negotiations. 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. 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.
S. Rep. No. 100-446, at 13 (1988), as reprinted in 1988 U.S.C.C.A.N. 3071, 3083.
At the same time, Congress recognized that it was not drafting IGRA against a blank legal slate. The Senate Report acknowledged the “long [] and well-established principle of Federal-Indian law as expressed in the United States Constitution, reflected in Federal statutes, and articulated in decisions of the Supreme Court, that unless authorized by an act of Congress, the jurisdiction of State governments and the application of state laws do not extend to Indian lands.” Id. at 5, reprinted in 1988 U.S.C.C.A.N. at 3075. “Consistent with these principles, the Committee ... developed a framework for the regulation of gaming activities on Indian lands which provide[d] that in the exercise of its sovereign rights, unless a tribe affirmatively elects to have State laws and State jurisdiction extend to tribal lands, the Congress w[ould] not unilaterally impose or allow State jurisdiction on Indian lands for the regulation of Indian gaming activities.” Id. at 6, reprinted in 1988 U.S.C.C.A.N. at 3075. “The mechanism for facilitating the unusual relationship in which a tribe might affirmatively seek the extension of State jurisdiction and the application of state laws to activities conducted on Indian land [wa]s a tribal-State compact.” Id. at 6, reprinted in 1988 U.S.C.C.A.N. at 3076.
*1048In enacting IGRA, Congress was not only aware of the complicated legal framework governing state-tribal relations, but was legislating directly in response to a Supreme Court decision in this area Cabazon Band. Recognizing the special consideration given to tribal immunity against taxation by the states absent congressional consent, Congress took care to ensure that IGRA’s compact requirement did not “[lift] the Indians’ exemption from state taxes.” Cabazon Band, 480 U.S. at 215 n. 17, 107 S.Ct. 1083. Congress did this in two ways: First, Congress explained that IGRA’s compact provision did not constitute the “express congressional authorization” required under Cabazon Band and related precedents to abrogate tribal immunity to taxation, stating that “[e]xeept for any assessments that may be agreed to under paragraph (3)(C)(iii) of this subsection, nothing in this section shall be interpreted as conferring upon a State or any of its political subdivisions authority to impose any tax, fee, charge, or other assessment upon an Indian tribe ... to engage in a class III activity.” 25 U.S.C. § 2710(d)(4) (emphases added). Second, recognizing that states might demand that tribes abrogate their immunity to direct taxation as part of compact negotiations, Congress declared such a practice to be evidence of bad faith on behalf of the state making the demand. See id. § 2710(d) (7)(B)(iii)(II) (a court “shall consider any demand by the State for direct taxation of the Indian tribe or of any Indian lands as evidence that the State has not negotiated in good faith”).
Providing a proper context to 25 U.S.C. §§ 2710(d)(4) and (d)(7)(B)(iii)(II) clarifies the congressional intent behind these provisions. These subsections are not concerned with States negotiating a “piece of the pie” with respect to class III gaming revenue: Congress expressly authorized State “assessment ... of [class III gaming] activities,” 25 U.S.C. § 2710(d)(3)(C)(iii), used deliberately narrow language in subsections (d)(4) and (d)(7)(B)(iii)(II), and, after recognizing that “[a] state’s governmental interests with respect to class III gaming on Indian lands include ... impacts on the State’s regulatory system, including its economic interest in raising revenue for its citizens,” encouraged States and tribes “to conduct negotiations within the context of the mutual benefits that can flow to and from tribe [sic] and states.” S. Rep. No. 100-446, at 13, reprinted in 1988 U.S.C.C.A.N. 3071, 3083 (emphasis added). Rather, subsections (d)(4) and (d)(7)(B)(iii)(II) are concerned with the possibility that a State might use IGRA as pretext to impose taxes directly on an Indian tribe or its lands, where the tribe would otherwise be immune, see S. Rep. No. 100-446, at 14, reprinted in 1988 U.S.C.C.A.N. 3071, 3084 (“The Committee ... view[s] the concession to any implicit tribal agreement to the application of State law as unique and does not consider such agreement to be precedent for any other incursion of State law onto Indian lands.”), as well as the converse — using the threat of direct taxation to discriminate against tribes in favor of private gaming interests, see id. at 13 (“It is the Committee’s intent that the compact requirement for class III gaming 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.”).
In sum, §§ 2710(d)(4) and (d)(7)(B)(iii)(II) were motivated not by a concern that states might generate revenue from class III gaming conducted on Indian lands — indeed, the legislative history of IGRA contemplates this very eventuality, S. Rep. No. 100-446, at 13, reprinted in 1988 U.S.C.C.A.N. 3071, 3083 (recognizing a State’s “economic interest in raising revenue for its citizens”) — but by a con*1049cern that states might use IGRA and the promise of class III gaming to otherwise impose taxes on Indian tribes and their land.
The majority treats §§ 2710(d)(4) and (d)(7)(B)(iii)(II) as though these sections delegate to Article III judges the responsibility of conducting a broad, standardless review of the overall “fairness” of negotiated compacts, but this is simply not the case. As explained above, §§ 2710(d)(4) and (d)(7)(B)(ni)(II) were limited, carefully-worded responses to specific issues arising from the background of tribal-state relations under the Constitution. Moreover, IGRA already assigns to the Secretary of the Interior — the government official “charg[ed by Congress] ... with broad responsibility for the welfare of Indian tribes,” Udall v. Littell, 366 F.2d 668, 672 (D.C.Cir.1966) (Burger, J.), and who owes “the Indians ... [a] duty of care and protection,” Rainbow v. Young, 161 F. 835, 838 (8th Cir.1908) (Van Devanter, J.); see 25 U.S.C. § 2 (Secretary of the Interior, through the Commissioner of Indian Affairs, “shall ... have the management of all Indian affairs and of all matters arising out of Indian relations”); 43 U.S.C. § 1457 (charging the Secretary of the Interior “with the supervision of public business relating to ... Indians”) — broad authority to approve or disapprove tribal-state compacts, and enumerates specific standards against which a tribal-state compact is to be reviewed by the Secretary. See 25 U.S.C. § 2710(d)(3)(B) (“[A] Tribal-State compact ... shall take effect only when notice of approval by the Secretary [of the Interior] of such compact has been published by the Secretary in the Federal Register.”); id. § 2710(d)(8) et seq. (“The Secretary may disapprove a [Tribal-State] compact ... only if such compact violates — (i) any provision of this chapter, (ii) any other provision of Federal law that does not relate to jurisdiction over gaming on Indian lands, or (iii) the trust obligations of the United States to Indians.”).
C
As we explained in Coyote Valley II, “[t]he passage of IGRA did not end the fight over Indian gaming in California.” 331 F.3d at 1098. After IGRA’s passage, numerous California tribes sought to negotiate class III gaming compacts with the State. Id. Among the class III games the tribes sought to offer were banked card games and slot machine-like electronic gaming machines. However, at the time these particular games were not permitted under California law, even though California permitted other forms of class III gaming such as non-electronic keno and lotto. Id.; see Rumsey Indian Rancheria of Wintun Indians v. Wilson, 64 F.3d 1250, 1257-58 (9th Cir.1994). During the Administration of Governor Pete Wilson, the State refused to negotiate with the tribes with respect to banked card games and slot machine-like gaming machines, arguing that under 25 U.S.C. § 2710(d)(1)(B) (“class III gaming shall be lawful on Indian lands only if such activities are ... located in a State that permits such gaming for any purpose by any person, organization, or entity”), a state need only negotiate with tribes with respect to class III games and devices otherwise allowed under state law. Coyote Valley II, 331 F.3d at 1098. In Rumsey, we agreed with the State, holding that “IGRA does not require a state to negotiate over one form of Class III gaming simply because it has legalized another, albeit similar form of gaming.... 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.
For the next few years, numerous tribes continued to offer class III gaming activi*1050ties that were categorically prohibited under California state law; in response, the Wilson administration refused to negotiate compacts covering class III games the State did permit. Id. at 1099. In 1998, a coalition of California tribes drafted a ballot proposition that would amend state law — but not the State Constitution' — to require the State to enter into compacts with Indian tribes that allowed otherwise-illegal class III gaming activities including banked card games and slot machines. Id. at 1100. This ballot proposition passed by a wide margin, but was ruled unconstitutional by the California Supreme Court, as the California Constitution provided that the “Legislature has no power to authorize, and shall prohibit casinos of the type currently operating in Nevada and New Jersey.” CAL. CONST, art. IV, § 19(e); see Hotel Employees & Rest. Employees Int’l Union v. Davis, 21 Cal.4th 585, 88 Cal.Rptr.2d 56, 981 P.2d 990, 994 (1999).
In early 1999 — while Hotel Employees was pending before the California Supreme Court, but prior to the court’s decision in that case — the administration of newly-elected Governor Gray Davis reversed the Wilson Administration’s non-negotiation policy and began compact negotiations with tribes regarding class III gaming that would include banked card games and slot machines. Coyote Valley II, 331 F.3d at 1102-03. After the Hotel Employees decision was filed in August 1999, the Davis Administration proposed an amendment to Article IV, Section 19 of the California Constitution that would exempt tribal gaming from the prohibition on Nevada-style casinos, subject to the requirement that class III gaming be eon-ducted pursuant to a tribal-state compact; this proposed amendment was to be placed on the ballot in March 2000 as Proposition 1A. Id. at 1103 & n. 11. Also in the late summer of 1999, the United States Department of Justice announced that it planned to proceed with enforcement actions against California tribes engaged in unauthorized class III gaming if those tribes did not enter into compacts with the State before October 13,1999. Id. at 1103.
The Davis Administration and the tribes engaged in compact negotiations throughout August and September of 1999. The basic quid pro quo in these 1999 negotiations was as follows: First, the State offered the tribes the exclusive right — with this exclusivity embodied in a constitutional amendment, see CAL. CONST, art. 4, § 19(f) — to conduct Nevada-style class III gaming, including house-banked slot machines and blackjack, see 1999 Tribal-State Gaming Compact (“1999 Compact”) § 12.4, http://www.cgcc.ca.gov/enabling/tsc.pdf (last visited Apr. 6, 2010). In exchange, the tribes offered a variety of concessions including, as relevant here, (1) limits on the number of class III gaming devices a tribe could operate, see id. § 4.3; and (2) “Revenue Distribution,” see id. § 5.0-5.3, under which a gaming tribe was required to remit to the State a percentage of its average net win from class III gaming devices in operation on September 1, 1999, to be deposited in a “Special Distribution Fund” (“SDF”). Id. § 5.1(a).3 Under the 1999 Compact’s “Revenue Distribution” provisions,
[t]he State’s share of the Gaming Device revenue [was to] be placed in the Special *1051Distribution Fund, available for appropriation by the Legislature for the following purposes: (a) grants, including any administrative costs, for programs designed to address gambling addiction; (b) grants, including any administrative costs, for the support of state and local government agencies impacted by tribal government gaming; (c) compensation for regulatory costs incurred by the State Gaming Agency and the state Department of Justice in connection with the implementation and administration of the Compact; (d) payment of shortfalls that may occur in the Revenue Sharing Trust Fund; and (e) any other purposes specified by the Legislature.
Id. § 5.2. If a tribe failed to timely pay its required share of gaming device revenues to the State, the tribe would be charged interest “in addition to any other remedies the State m[ight] have,” id. § 5.3(b); if “more than two quarterly contributions” by the tribe were overdue, the tribe would be prohibited from conducting class III gaming, id. § 5.3(e).
On September 10, 1999, 57 tribes, including the Band, signed letters of intent to enter the 1999 Compact. In March 2000, Proposition 1A was ratified by California voters. Two months later, the United States Secretary of the Interior approved the Tribal-State compacts entered into between the State and 60 tribes, including the Band. Coyote Valley II, 331 F.3d at 1107; see 65 Fed.Reg. 31189 (May 16, 2000). The Band currently operates its “Harrah’s Rincon” casino pursuant to the 1999 Compact. Under this compact, which does not expire until 2020, the Band is authorized to operate 1600 gaming devices. Notably, because the Band operated fewer than 200 gaming devices on September 1, 1999, it is not required to make — and never has made — revenue sharing payments to the SDF under its existing compact. See 1999 Compact § 5.1(a) (“The Tribe shall make contributions to the Special Distribution Fund ... only with respect to the number of Gaming Devices operated by the Tribe on September 1,1999.”).
By 2003, California was awash in debt and faced a $38 billion dollar budget deficit. See generally Mark Simon, Recall Election set for Oct. 7: Fiscal Crisis and voter resentment led to recall, S.F. Chron., July 25, 2003, at Al. As Governor Davis faced recall by the California voters, the State executed three new gaming compacts in August and September of 2003. See http://www.cgcc.ca.gov/compacts.asp (last visited Apr. 6, 2010). These compacts were nearly identical to the 1999 Compacts, although they differed in one important respect: instead of requiring the compacting tribe to remit a percentage of its gaming device net win to the Special Distribution Fund, these compacts called for the revenue sharing with the State’s general fund. See, e.g., La Posta Band of Mission Indians Tribal-State Gaming Compact § 4.3.1(a), http://www.cgcc.ca.gov/ compacts/8132009/laposta.pdf (last visited Apr. 6, 2010) (“The La Posta Tribe may operate no more than 350 Gaming Devices. From and after the first day of operation of its first Gaming Facility, the La Posta Tribe shall pay five percent (5%) of its Net Win from the operation of Gaming Devices to the California Gambling Control Commission, or such other State entity as may be designated by the Governor, for deposit into the General Fund.”). All three compacts executed in 2003 calling for general fund revenue sharing were approved by the Department of the Interior, http:// www.cgcc.ca.gov/compacts.asp (last visited Apr. 6, 2010).
In October 2003, Gray Davis was recalled, and Arnold Schwarzenegger was elected governor of California. The Schwarzenegger Administration has continued the late-term practice of the Davis Administration by negotiating general *1052fund, rather than Special Distribution Fund, revenue sharing in new and amended tribal compacts. At present, fifteen Tribal-State compacts in California provide for general fund revenue sharing. See http://www.cgce.ca.gov/Tribal/2009/GF %20Paying TribesforWebsite03-04-09.pdf (last visited Apr. 6, 2010). Each of these compacts has been approved by the Secretary of the Interior.
Ill
When Congress drafted IGRA, it did so against a well-defined legal background prohibiting states from directly taxing Indian tribes and tribal lands without express congressional authorization. With this history expressly in mind, see S. Rep. No. 100-446, at 5, reprinted in 1988 U.S.C.C.A.N. 3071, 3075, Congress drafted two very specific provisions explaining (1) that IGRA’s compact provisions, which allowed states to bargain with tribes over class III gaming regulation, did not authorize the states to impose a tax on the tribe, 25 U.S.C. § 2710(d)(4); and (2) that in the context of compact negotiations, a state’s demand for otherwise-prohibited direct taxation of a tribe or its lands would be evidence of bad faith, id. § 2710(d)(7)(b)(iii)(II). These provisions were carefully worded to respond to a specific problem — the possibility that IGRA’s class III gaming provisions might be interpreted by states as an “express congressional authorization” for direct taxation or, relatedly, that states might use compact negotiations to demand abrogation of a tribe’s immunity to direct taxation.
The majority steps into these stalled negotiations with an astonishing proposition: a demand for revenue sharing by the State is “simply an impermissible demand for the payment of a tax by the tribe” prohibited by § 2710(d)(4), Maj. Op. at 1042, and under § 2710(d)(7)(B)(iii)(II) “is evidence of the State’s bad faith,” id. at 1030. The majority orders the parties either to reach a compact or submit to mediation, but in either case, there is only one outcome: The Band wins. California loses.
In my view, the majority has failed to recognize the legal and historical differences between taxation and revenue sharing. A tax is imposed by the state and may be collected by the state, under penalty of law, over the objections of its citizens. Revenue sharing, by contrast, is typically negotiated by the parties and then enforced through the law of contract. If the parties fail to agree to revenue sharing in the first instance, neither side can unilaterally compel the other to share revenues; there is no contract. Here, California has insisted that the Band share its gaming revenues as a condition to receiving authorization for additional gaming devices. To date, California has nothing to show for its negotiations and, under long-established principles, no power to compel the Band to pay it a percentage of its revenues. California and the Band are independent sovereigns; they may negotiate an agreement or not, but neither has the power to tax the other. Thus, California’s insistence on revenue sharing may be “a hard line stance,” Maj. Op. at 1039, but it is not a tax.
A
The majority has confounded two distinct and well-defined concepts: “taxation” and “revenue sharing.” As described by the majority:
Under § 2710(d)(7)(B)(iii)(II), a court must consider a “demand” for a tax to be made in bad faith. A tax is “a charge, usu[ally] monetary, imposed by the government on persons, entities, transactions, or property to yield public revenue.” Black’s Law Dictionary 1594 (9th ed.2009) (emphasis added). The State insisted that Rincon pay at least *105310% of its net profits into the State’s general fund. According to Cal. Govt. Code § 16300, “[t]he General Fund consists of money received into the treasury and not required by law to be credited to any other fund.” No amount of semantic sophistry can undermine the obvious: a non-negotiable, mandatory payment of 10% of net profits into the State treasury for unrestricted use yields public revenue, and is a “tax.”
Maj. Op. at 1029-30 (footnote omitted). Indeed, at every turn, the majority has equated revenue sharing with taxation. See id. at 1032 (“when the ‘nature of the fees’ is general fund revenue sharing — a bald demand for payment of a tax — the State faces a very difficult task to rebut the evidence of bad faith necessarily arising from that demand”); id. at 1042 (“The State’s demand for 10-15% of Rincon’s net win, to be paid into the State’s general fund, is simply an impermissible demand for the payment of a tax by the tribe.”). The majority fundamentally errs when it declares that “general fund revenue sharing [is] a bald demand for payment of a tax.” Maj. Op. at 1032. In a definition quoted by the majority itself, Black’s Law Dictionary defines a “tax” as “[a] charge, usu[ally] monetary, imposed by the government on persons, entities, transactions, or property to yield public revenue.” Black’s Law Dictionary 1594 (9th ed.2009) (emphasis added). Webster’s Third New International Dictionary and The Oxford English Dictionary feature nearly identical definitions. See Webster’s Third New International Dictionary 2345 (1993) (defining “tax” as “a usu[ally] pecuniary charge imposed by legislative or other public authority upon persons or property for public purposes: a forced contribution of wealth to meet the public needs of a government” (emphases added)); XVII The Oxford English Dictionary 677 (2d ed.1989) (defining “tax” as “[a] compulsory contribution to the support of government, levied on persons, property, income, commodities, transactions, etc., now at fixed rates, mostly proportional to the amount on which the contribution is levied” (emphases added)).4 Whether the dictionary of record is Black’s, Webster’s 3d, or The Oxford English Dictionary, there are three components to the definition of “tax”: (1) a monetary contribution; (2) imposed by the government; (3) to yield public revenue. The majority focuses on the first and third components — a monetary contribution yielding public revenue — but completely ignores the second component — the requirement that a charge be “imposed ” or “levied ” by the government before it can be deemed a “tax.” See Maj. Op. at 1029-31. Yet, the fact that a monetary payment to the government is imposed, and not bargained-for, is the essential characteristic that defines a tax.
We have had occasion to consider what constitutes a tax, and have “applied this tax definition: (1) An involuntary pecuniary burden (2) imposed by the state legislature (3) for a public purpose (4) under the state’s police or taxing power. Our point in so refining the definition of a tax was to distinguish taxes from debts for voluntarily assumed obligations ....” George v. Uninsured Employers Fund (In re George), 361 F.3d 1157, 1160 (9th Cir.2006) (emphases added). As we explained in In re Lorber Industries of California, Inc., “[i]n general, ... charges can be classified as a tax only if they constitute ‘a pecuniary burden laid upon individuals or property for the purpose of supporting the *1054Government’ or to support ‘some special purpose authorized by it.’ Taxes are not equivalent to debts, which are voluntary obligations based on express or implied contracts. Taxes are levied without the consent or voluntary action of the taxpayer.” 675 F.2d 1062, 1066 (9th Cir.1982) (emphasis added) (quoting New Jersey v. Anderson, 203 U.S. 483, 492, 27 S.Ct. 137, 51 L.Ed. 284 (1906)); see also Nat’l Cable Television Ass’n v. United States, 415 U.S. 336, 340-41, 94 S.Ct. 1146, 39 L.Ed.2d 370 (1974).5
The majority, however, simply sidesteps the “imposition” requirement by slapping the conclusory labels “nonnegotiable” and “mandatory” on proposed revenue sharing payments that are neither. If the “payment of 10% of net profits into the State treasury for unrestricted use[,] yield[ing] public revenue” that the majority derides as a “non-negotiable, mandatory” tax, Maj. Op. at 1029-30, were either non-negotiable or mandatory, how is it that the State has not, to date, received one dime of general fund revenue from the Tribe? The answer, of course, is that the State’s proposed 10% general fund revenue sharing payments are both negotiable (indeed, the 10% figure is just one of several revenue sharing proposals put forth by the State in negotiations with the Tribe) and subject to the Tribe’s agreement in exchange for additional machines.
Perhaps the majority simply means that a fixed, agreed-to “payment ... into the State treasure for unrestricted use yields public revenue, and is a tax,” but this cannot be right, either. A purchaser of a bond — whether issued by a state, a municipality, or by the federal government— makes a monetary contribution that yields public revenue. But a bond is not a “tax.” Indeed, a bond is not a “tax” even if its proceeds go to a State’s general fund. The reason for this is simple: even though a bond purchaser pays money to the government to support its functions, this payment is not “imposed” on the purchaser; rather, the bond purchaser pays money to support the government in exchange for something of value, a future interest stream. Under the majority’s definition, another “obvious” tax would be simply gifting ten percent of one’s net revenue to the federal government. Such a gift would, after all, be a monetary contribution yielding public revenue. But although helpful to the public fisc, this voluntary contribution would not be a “tax.”
B
We must contrast “taxation” with “revenue sharing,” in which two parties-whether *1055private firms or sovereign entities agree to split a certain proportion of revenues from a specified source.
Revenue sharing is ubiquitous. Private entities engage in revenue sharing in areas as diverse as movie rentals, see Brooks Barnes, Movie Studios See a Threat in Growth of Redbox, N.Y. Times, Sept. 6, 2009, at B1 (Redbox and Paramount), sports arenas, see Ken Belson and David M. Halbfinger, Possible Truce Between New Jersey Arenas Could Come at a Cost to Fans, N.Y. Times, Dec. 12, 2009, at B16 (Izod Center and Prudential Center in northern New Jersey), newspapers and web portals, see Evan Hessel, Yahool’s iDangerous Newspaper Deal?, Forbes Online, June 22, 2009, http://www.forbes.com/ 2009/0 6/22/advertising-newspapers-inter-net-businessmedia-yahoo.html (last visited Apr. 6, 2010) (five newspapers and Yahoo!), and, of course, casinos, see, e.g., Introduction to Slots and Video Gaming 76, International Game Technology, http:// media.igt.com/Marketing/ PromotionalLiterature/IntroductionToGaming.pdf (last visited Apr. 6, 2010) (slot machines and gaming venues).
Moreover, revenue sharing is hardly confined to the private realm. Since the 1880s, the federal government has contracted with private parties to operate “contract postal units” (“CPUs”), which are “postal facilities operated by private parties on private property pursuant to revenue-sharing contracts with the government.” Cooper v. United States Postal Serv., 577 F.3d 479, 485 (2d Cir.2009). In exchange for the right to operate a postal facility, a CPU submits all of its revenues to the Postal Service, which then repays a portion of the revenue to the CPU. Like the federal government, state and local governments also enter into revenue-sharing contracts with private firms. For example, under a federal program administered by the U.S. Department of Transportation, a number of state and local governments contract with a private firm to collect and analyze traffic data. See Eric Lipton, U.S. System for Tracking Traffic Flow is Faulted, N.Y. Times, Dec. 13, 2009, at A22; see also Federal Highway Administration Report No. MH-2010-030, Transportation Technology Innovation and Demonstration Program (Dec. 8, 2009), http://www.oig.dot.gov/sites/ dot/files/TTID_12_8_2009.pdf (last visited Apr. 6, 2010) (“TTID Report”). The private firm is supposed to remit to these state and local governments a percentage of the money it makes from selling its traffic data, which Congress has described as revenue sharing, not a tax. See Pub.L. 105-178, § 5117(b)(3)(B)(iii) (June 9, 1998); see generally TTID Report at 6-8.
Federal, state and local governments have long entered into revenue sharing agreements with each other. “A form of general revenue sharing was actually in operation as early as 1836[,] when nearly $28 million was deposited with State governments. In [the twentieth] century, congressional attempts to secure Federal tax sharing date back to 1949[,] when a Member of Congress from Kansas introduced a bill to transfer to State treasuries 1 percent of Federal individual and corporate income taxes raised within the State.” Council of and for the Blind of Del. County Valley, Inc. v. Regan, 709 F.2d 1521, 1523 n. 2 (D.C.Cir.1983) (en banc) (quotation marks, ellipses, brackets, and citation omitted); see also 43 U.S.C. § 1608 et seq. (requiring the United States and the State of Alaska to pay to Native Alaskans two percent of annual gross revenue from the disposition of minerals removed from that state; Congress titled this program “revenue sharing”).
Perhaps the most prominent example of broad-based intergovernmental revenue sharing was the State and Local Fiscal
*1056Assistance Act of 1972, Pub.L. 92-512, 86 Stat. 919, more commonly known as the Revenue Sharing Act. See generally President Richard Nixon, Statement About the General Revenue Sharing Bill (Oct. 20, 1972) (transcript available at http://www. presidency.ucsb.edu/ws/index.php?pid= 3636 (last visited Apr. 6, 2010)). Under the Revenue Sharing Act, Congress “g[ave] legal effect to a concept which had been proposed intermittently for more than 20 years — that the federal government make general revenue grants to state and local governments with no strings attached.” Council of and for the Blind of Del. County Valley, Inc., 709 F.2d at 1523. For just over a decade (the program was allowed to expire in 1983), the federal government paid billions of dollars to state and local government general funds with “no strings ... attached.” See S. Rep. No. 92-1050, at 3876 (1972), reprinted in 1972 U.S.C.C.A.N. 3874, 3876; Goolsby v. Blumenthal, 590 F.2d 1369, 1371 (5th Cir.1979) (en banc).6
As these examples demonstrate, there is a long history of revenue sharing in the United States. It has taken two forms. First, there is vertical revenue sharing, where one government decides to share its revenues with another (smaller) governmental entity. In this form of revenue sharing, the smaller entity has no continuing demand on the revenues; it appears that the decision by the larger entity to engage in revenue sharing is one of largesse and can be amended at any time. The second form is horizontal or negotiated revenue sharing, where two sovereigns, two private entities, or a sovereign and a private entity negotiate revenue sharing as a matter of contract. Horizontal revenue sharing is a voluntary matter, to be agreed upon by the parties as a part of their overall negotiations. No one has ever thought that horizontal revenue sharing is a form of taxation.
C
The ultimate question, then, is whether California’s demand for a share of the Band’s gaming revenues is direct taxation or a form of revenue sharing. The answer is critical to the resolution of this case. In my view, the answer is not a difficult one.
IGRA requires that tribes and states “enter[ ] into a Tribal-State compact governing the conduct of gaming activities.” 25 U.S.C. § 2710(d)(3)(A). Congress’s chosen words here — “ ‘enter,” “compact,” and “governing”' — -are the language of treaties and other sovereign-to-sovereign agreements. Since the states are barred from entering into “any Treaty, Alliance, or Confederation,” U.S. Const, aft. I, § 10, cl. 1, Congress chose an appropriate word: “compact” — the very definition of which implies negotiation between equal parties. See Black’s Law Dictionary 318 (9th ed.2009) (defining “compact” as “[a]n agreement or covenant between two or more parties, esp. between governments or states”). Notably, the tribal-state compacts in IGRA, which derive their existence from a congressional enactment and must be approved by the Secretary of the Interior, resemble very closely the interstate compacts described by the Constitution’s Compact Clause. See U.S. Const. art. I, § 10, cl. 3 (“No State shall, without the Consent of Congress ... enter into any Agreement or Compact with another State.... ”). An interstate compact, even if it involves the exchange of funds, is not interstate taxation, but “[a] voluntary *1057agreement between states enacted into law in the participating states upon federal congressional approval.” Black’s Law Dictionary 318 (9th ed.2009). Critically, “[interstate compacts, like treaties, are presumed to be the subject of careful consideration before they are entered into, and are drawn by persons competent to express their meaning, and to choose apt words in which to embody the purposes of the high contracting parties.” New Jersey v. Delaware, 552 U.S. 597, 615-16, 128 S.Ct. 1410, 170 L.Ed.2d 315 (2008) (quotation marks omitted).
As a sovereign (a state) negotiating a compact with another sovereign (an Indian tribe), California has no “authority to impose any tax” on the Band. 25 U.S.C. § 2710(d)(4). It may “demand” that the Band share a portion of its gaming revenues with the State, but “demands” are part and parcel of negotiations; as in negotiations between private parties, the State has no authority to impose such a requirement unilaterally. This case is the best proof of this principle. California can demand general revenue sharing until it is hoarse, but it will not get revenue from the Band until the Band agrees. The State’s strategy is aggressive, and aggression in negotiations comes with the risk that the deal will fall through. California is running that risk. For all of its posturing and demands, California has, to date, not one penny to show for its negotiations with the Band.7
In the tribal-state compacts California has negotiated since 2003, it has secured revenue sharing from other tribes. It has not imposed revenue sharing, but rather obtained it through sovereign-to-sovereign negotiations. When “two equal sovereigns,” S. Rep. No. 100-446, at 13, reprinted in 1988 U.S.C.C.A.N. 3071, 3083; Coyote Valley II, 331 F.3d at 1112, negotiate a percentage of gross revenue to be paid from one party to the other in exchange for other concessions, this is revenue sharing — even if the revenue is not earmarked for a specific fund. Revenue sharing is not a “tax,” and it is certainly not “direct taxation of [a] tribe or ... [its] lands.” 25 U.S.C. § 2710(d)(7)(B)(iii)(II). The majority errs when it holds otherwise.
IV
Although much of the majority opinion is devoted to holding that California’s revenue sharing proposal is a tax, the majority stops short of conceding that all revenue sharing provisions are prohibited. See Maj. Op. at 1037 (“[W]e do not hold that no future revenue sharing is permissible... .”). The majority qualifies its concern so long as two conditions are satisfied: First, that the State has offered “meaningful concessions,” Coyote Valley II, 331 F.3d at 1111, and second, that the purposes for which the State intends to use the revenue are directly related to the operation of gaming activities, see 25 U.S.C. § 2710(d)(3)(C) (“Any Tribal-State compact negotiated under subparagraph *1058(A) may include provisions relating to ... (vii) any other subjects that are directly related to the operation of gaming activities.”). In my view, the majority has misread both requirements.
A
In Coyote Valley II, we held that where California “offered meaningful concessions in return for its demands, it did not ‘impose’ [a putative tax] within the meaning of § 2710(d)(4).” 331 F.3d at 1111. We expressed concern that, if a state demanded revenue from a tribe without offering anything in return, then the revenue demand might constitute an “impose[d] tax, fee, charge, or other assessment,” which IGRA’s compact provisions did not authorize. 25 U.S.C. § 2710(d)(4). However, we did not actually decide whether such a demand would constitute the imposition of a tax, fee, or charge, because we found that California had offered concessions in the negotiating process:
We do not hold that the State could have, without offering anything in return, taken the position that it would conclude a Tribal-State compact with Coyote Valley only if the tribe agreed to pay into the RSTF. Where, as here, however, a State offers meaningful concessions in return for fee demands, it does not exercise “authority to impose” anything. Instead, it exercises its authority to negotiate, which IGRA clearly permits.
Coyote Valley II, 331 F.3d at 1112. Similarly, in Shoshone-Bannock Tribes, we observed that § 2710(d)(4) was not a bar where “the parties negotiated a bargain permitting such payments in return for meaningful concessions from the state (such as a conferred monopoly or other benefits).” 465 F.3d at 1101. We reasoned that “[although the state did not have authority to exact such payments, it could bargain to receive them in exchange for a quid pro quo conferred in the compact.” Id.
The majority goes one step beyond Coyote Valley II and Shoshone-Bannock Tribes. Unfortunately, it goes in the wrong direction. Instead of allowing the parties to work out their differences through the negotiating process, it finds that the State did not offer “meaningful concessions” to the Band because it “disagree[s] that the State makes ‘meaningful concessions’ whenever it offers a bundle of rights more valuable than the status quo.” Maj. Op. at 1040. The majority’s statement mangles the phrase “meaningful concessions”: it is hard to think what concession could be more meaningful than “a bundle of rights more valuable than the status quo.”
The majority’s novel conception of “meaningful concessions” finds no support in IGRA and conflicts with our explanation of “meaningful concessions” in Coyote Valley II, the Department of the Interior’s reading of the Act, and centuries of contracts jurisprudence as well.
A “meaningful concession” is something conceded or granted that is real to at least one of the parties and not merely nominal or illusory. Notably, this is exactly how we described the requisite “concessions” under 25 U.S.C. § 2710(d)(4) in Coyote Valley II: “If [ ] offered concessions by a State are real, § 2710(d)(4) does not categorically prohibit fee demands.” 331 F.3d at 1112 (emphasis added). Simply put, “meaningful” implies “non-negligible,” but no more.
The Department of Interior has also interpreted “meaningful concessions” along the same line. As described in an IGRA review letter by the Principal Deputy Assistant Secretary of the Interior:
As our previous compact decision letters have stated, in order to determine whether revenue sharing violates 25 *1059U.S.C. § 2710(d)(4), we first look to whether the State has offered meaningful concessions. We have traditionally viewed this concept as one where the State concedes something that it was otherwise not required to negotiate that provides a benefit to the Tribe, i.e. exclusivity or some other benefit.
Absentee Shawnee Tribe Oklahoma, Dept, of Interior IGRA Rev. Ltr. (Dec. 17, 2004), http://www.nigc.gOv/Portals/0/ NIGC% 20Uploads/readingroom/compacts/Absen-tee% 20 Shawnee% 20Tribe% 20of% 20Oklahoma/absenteeshawnee 121704.pdf (last visited Apr. 6, 2010) (emphasis added). DOI’s interpretation of “meaningful concession! ]” as “the State conceding] something that it was otherwise not required to negotiate” accords with the phrase’s plain meaning and also falls neatly within black-letter contract law.
This understanding of “meaningful concession! ]” dovetails nicely with one of the most venerable concepts in the common law — the doctrine of consideration. Under the common law of contracts, a court will not weigh the adequacy, in the sense of market equivalence, of bargained-for consideration, but will determine whether putative consideration is nominal or a “sham.” See 4 Joseph M. Perillo et al, Corbin on Contracts §§ 5.14, 5.17 (2d ed.1995). Thus, a court will make a threshold inquiry to make sure that what is exchanged is not a “pretense,” but will not assign its own subjective valuation to each side’s respective concessions. In the words of a leading treatise, this approach to contract law is the very cornerstone of market exchange:
We have a free market, under our common law, for the reason that the courts have left it free. They do not require that one person shall pay as much as others may be willing to pay, or that one person shall receive as little as others may be willing to receive for a like article. The contracting parties make their own contracts, agree upon their own exchanges, and fix their own values.
4 Corbin on Contracts § 5.14.
In light of our own explanation of the phrase in Coyote Valley II, DOI’s understanding of the phrase, and black letter contract law, there can be little doubt that when we used the phrase “meaningful concessions” in Coyote Valley II, we were describing exactly what the majority rejects: a State makes “meaningful concessions” whenever it offers a bundle of rights more valuable than the status quo. Courts have not traditionally tipped the scales of negotiations between two well-represented parties, and there is no warrant in IGRA or in Coyote Valley II to do so here. Coyote Valley II states the applicable test as plainly as words will allow: “If ... offered concessions by a State are real, § 2710(d)(4) does not categorically prohibit fee demands.” 331 F.3d at 1112 (emphasis added).
B
Equally infirm is the majority’s restricted reading of § 2710(d)(3)(C)(vii) and its conception of what a state may do with the revenue it has bargained for. In the end analysis, I think that the majority has turned Coyote Valley II on its head. The majority now holds that “[wjhether revenue sharing is an authorized negotiation topic ... depends on the use to which the revenue will be put, not on the mere fact that the revenue derives from gaming activities.” Maj. Op. at 1033. As described by the majority, the RSTF and SDF addressed “fair distribution of gaming opportunities and compensation for the negative externalities caused by gaming” and did not “ ‘put tribal money into the pocket of the State.’ ” Id. at 1033-34 (quoting Coyote Valley II, 331 F.3d at 1113). But whereas in Coyote Valley II we simply held that the RSTF and SDF were permissible subjects of negotiation, the majority now holds that the RSTF and SDF represented the outer *1060limits of permissible negotiation. Thus, the majority answers the question we reserved in Coyote Valley II and holds that “general fund revenue sharing is not ‘directly related to the operation of gaming activities’ and is thus not an authorized subject of negotiation.” Maj. Op. at 1034 (quoting 25 U.S.C. § 2710(d)(3)(C)(vii)); see also id. at 1039 (“[Gjeneral fund revenue sharing is not a state public policy interest directly related to gaming and is not an authorized negotiation topic .... ”).
I have two responses to the majority’s analysis. First, I do not think that § 2710(d)(3)(C) is so limited. The listing of subjects that are appropriate for State-Tribe negotiations is permissive: the compact “may include provisions relating to” the items on the list. 25 U.S.C. § 2710(d)(3)(C) (emphasis added). In United States v. Redman, 35 F.3d 437 (9th Cir.1994), we interpreted the phrases “may consider,” “shall consider,” and “should consider” in successive versions of the Sentencing Guidelines. Id. at 440-41. The 1989 amendment to § 5G1.3(c) of the Guidelines provided that the court “may consider” a reasonable incremental penalty to be a sentence for the instant offense, while the 1991 amendment to this section used the phrase “shall consider,” and the 1992 amendment to the same section used the phrase “should consider.” Id. We noted that “Webster defines the word ‘may’ as ‘have permission to[,]’ Webster’s Third New International Dictionary 1396 (1986) ... [,] ‘[s]hall’ is defined as being ‘used in laws, regulations and directives to express what is mandatory[,]’ id. at 2085[, and] ‘[s]hould’ is used to express what is ‘probable or expected!,]’ id. at 2104,” and reasoned:
The meaning of the current “should consider” language in § 5G1.3 can be discerned by reference to the earlier iterations of the Guideline. Under the 1989 [“may consider”] amendment, courts had complete discretion to employ the commentary methodology or not when deciding on a reasonable incremental penalty. The 1991 [“shall consider”] amendment stripped courts of all discretion in the matter.... The “should consider” language in the current version falls somewhere between the “may consider” language in the 1989 amendment and the “shall impose” mandate of the 1991 version.
Redman, 35 F.3d at 440-441 (citations omitted). In short, we have interpreted the word “may” as Webster’s Third New International Dictionary defines it: as a grant of permission rather than an prescriptive limitation. Section 2710(d)(3)(C), which enumerates a number of subjects which a Tribal-State compact “may include provisions relating to,” identifies these interests as permitted subjects of negotiation, but neither explicitly nor implicitly restricts negotiations to these interests.8 See generally Cohen’s Handbook *1061of Federal Indian Law 873 (2005 ed.) (“Tribes and states may include a wide variety of subjects in the gaming compacts .... More than 200 tribal-state compacts have been approved, and they are often subject to amendments, as well as new negotiations if they have sunset provisions. It is thus impossible to generalize about the content of provisions ... ”).
Even if the list outlined in § 2710(d)(3)(C)(i)-(vii) were exclusive — a proposition that runs contrary to the dictionary definition of “may,” our interpretation of the term “may” as used in other congressional enactments, and our longstanding interpretation of § 158(d) of the NLRA-the State’s revenue-sharing proposal would still be an authorized subject of negotiation, as it falls within § 2710(d)(3)(C)(vii), a catch-all provision authorizing negotiation of “any other subjects that are directly related to the operation of gaming activities.” The majority reads this clause very restrictively, explaining that “[wjhether revenue sharing is an authorized negotiation topic under § 2710(d)(3)(C)(vii) [ ] depends on the use to which the revenue will be put, not on the mere fact that the revenue derives from gaming activities,” Maj. Op. at 1033; see also id. at 1028-29 n. 9, but its reading of § 2710(d)(3)(c)(vii) cannot be reconciled with our decision in Coyote Valley II.
In Coyote Valley II, we rejected the tribe’s request “to read § 2710(d)(3)(C)(vii) narrowly and to hold that it does not encompass a provision like the RSTF.” 331 F.3d at 111. Instead, “we h[e]ld that § 2710(d)(3)(C)(vii) authorize^] the RSTF provision....” 331 F.3d at 1111. Critically, the RSTF did not use revenue for gaming purposes. Rather, the RSTF distributed revenue to non-gaming tribes, who could use it however they saw fit; the RSTF was effectively a contribution to the non-gaming tribes’ general revenue funds. The RSTF certainly used revenue for the benefit of Indian tribes, but it did not use revenue for any purpose directly related to gaming. Our express holding in Coyote Valley II that “the RSTF provision clearly falls within [§ 2710(d)(3)(C)(vii)’s] scope,” id., is irreconcilable with the “use” restriction relied on by the majority.
The majority also runs afoul of Coyote Valley II when it declares that “what compels a limited reading[of § 2710(d)(3)(C)(vii) ] is the canon of construction obligating us to construe a statute abrogating tribal rights narrowly and most favorably towards tribal interests.” Maj. Op. at 1028-29 n. 9. This canon, which “we regard [ ] as a mere guideline and not a substantive law,” Williams v. Babbitt, 115 F.3d 657, 663 n. 5 (9th Cir.1997) (quotation marks omitted), only applies to “am*1062biguous provisions,” Montana v. Blackfeet Tribe of Indians, 471 U.S. 759, 766, 105 S.Ct. 2399, 85 L.Ed.2d 753 (1985). Yet in Coyote Valley II, we held that § 2710(d)(3)(C)(vii) “is not ambiguous and that the RSTF provision clearly falls within its scope,” rejecting the tribe’s argument that the RSTF provision was unauthorized under § 2710(d)(3)(C)(vii) based on the principle that “we should interpret any ambiguities on these issues in a matter that will be most favorable to tribal interests.” 331 F.3d at 1111 (quotation marks omitted). Since we have expressly held that § 2710(d)(3)(C)(vii) is not ambiguous, the canon that ambiguities in statutes should be construed in favor of Indian tribes is inapplicable. Once again, the majority is simply wrong on this point.9
Instead of rigidly imposing a gaming-only restriction on how negotiated revenue sharing may be used, § 2710(d)(3)(C)(vii) simply “prevent[s] compacts from being used as a subterfuge for imposing State jurisdiction on tribes concerning issues unrelated to gaming.” Coyote Valley II, 331 F.3d at 1111. As in Coyote Valley II, “the State has not sought to engage in such a subterfuge” here. Id. Tribes such as the Band are asking for something from the State-additional gaming devices — to which they are not entitled under their current compacts. California has requested a share of the revenues generated by those devices. This seems pretty “related to the operation of gaming activities” to me. If, for example, California had asked the Band to open a water park on tribal lands located elsewhere in the state or to make a donation to the booster club at UCLA, we might have an issue.
The majority and I would probably agree that it would have been better if Congress had made it explicit in IGRA whether general revenue sharing could be a part of compact negotiations or not. But it didn’t, and that failure has generated public controversy, litigation, and a cottage *1063industry in law review articles.10 Unfortunately, the only consensus among academics appears to be that Congress should clarify the issue and address revenue sharing head-on.11 States and tribes, however, do appear to have reached a consensus, at least in practice:
Although it is clear that states may not tax gaming activities, tribes and states have devised approaches that allow some form of revenue sharing with the states.... The Secretary of the Interior has approved revenue sharing arrangements on the ground that those payments are not taxes, but exchanges of cash for significant economic value conferred by the exclusive or substantially exclusive right to conduct gaming in the state.... Despite generating high-level publicity and debate, revenue sharing provides states and tribes with the means to consummate compacts that both sides can legitimately claim will provide substantial economic benefits to their constituents.
Cohen’s Handbook of Federal Indian Law 874. Perhaps the best evidence for the practice is the fact that, after twenty years of experience with IGRA, every compact negotiated in recent years-including compacts in Connecticut,12 Florida,13 *1064Michigan,14 New York,15 New Mexico,16 Oklahoma,17 Wisconsin,18 and, of course, California—has included general revenue sharing. Equally significant, the Secretary of the Interior has sanctioned these compacts.19 Many of the payments under *1065these compacts are substantial. For example, Florida receives approximately $100 million annually in general fund revenue sharing from the Seminole Tribe, Courtney J.A. DaCosta, Note, When “Turnabout” Is Not “Fair Play”: Tribal Immunity under the Indian Gaming Regulatory Act, 97 Geo. L.J. 515, 543 (2009), while New Mexico’s general fund received nearly $63 million in revenue sharing payments between October 1, 2008 and September 30, 2009, see http://www.nmgcb. org/tribal/revsharing.html. In Connecticut, home to the nation’s largest tribal casino and two of the largest casinos in the entire world, the Mashantucket Pequot and Mohegan tribes pay twenty-five percent of gross revenue from electronic gaming machines to the State of Connecticut’s general fund. See supra n. 9. In 2008, the last full calendar year for which revenue information is available, Connecticut received $411,410,973 in general fund revenue under the Mashantucket Pequot and Mohegan gaming compacts, http://www.ct.gov/ dosr/hb/dosr/GeneraLFund_Revenues_ All_Games.pdf (last visited Apr. 6, 2010), about half of which came from Connecticut’s slot machine revenue sharing agreement with the tribes, http://www.ct.gov/ dosr/lib/ dosr/Fosltweb.pdf (last visited Apr. 6, 2010). In the years 2003-07, Connecticut received $387 million, $403 million, $418 million, $428 million, and $430 million, respectively in general fund revenue from its tribal casinos. The Secretary of the Interior, charged with seeing to the general welfare of the tribes, has approved the revenue sharing provisions of every one these compacts, taking the “view[that] revenue-sharing arrangements are lawful so long as the state offers the tribe separate consideration in exchange for a cut of the tribe’s gaming revenues.” Dacosta, 97 Geo. L.J. at 543—44; see also Cohen’s Handbook of Federal Indian Law at 876 (“The Secretary of the Interior ... has not expressed concern over how the revenues are used by a state, so long as the exclusivity provides ‘substantial economic benefit’ to the tribe.”).
V
From its flawed premises — that negotiated revenue sharing is a direct tax and that any state demands for revenue sharing must be accompanied by substantial concessions and a promise to use the revenues for gaming-related purposes only— the majority proceeds to consider whether the State has failed to negotiate with the Band in good faith. The majority concludes that it has. Among the majority’s concerns with the State’s demands, one stands above all others: in the majority’s view, “the financial benefit to Rincon from the amendments proposed would be negli*1066gible: Rincon stood to gain only about $2 million in additional revenues compared to the State’s expected $38 million.” Maj. Op. at 1038; see also id. at 1035-36 (“We [ ] find particularly persuasive the fact that the revenue sharing demanded in this case would result in $38 million in additional net revenue to the State compared to $2 million for the tribe.”); id. at 1025-26.
“IGRA’s legislative history ... makes clear that the good faith inquiry is nuanced and fact-specific, and is not amenable to bright-line rules.” Coyote Valley II, 331 F.3d at 1113. “The terms of each compact may vary extensively depending on the type of gaming, the location, the previous relationship of the tribe and State, etc.” S. Rep. No. 100-446, at 14 (1988), reprinted in 1988 U.S.C.C.A.N. 3071, 3084. Even if the State’s demand for revenue sharing could be construed as a demand for direct taxation, I think the majority’s concerns are misplaced and the State has not negotiated in bad faith. The State has taken a hardline position, one that the Band does not like, but that does not make it a bad faith negotiating position. See, e.g., Wis. Elec. Power Co. v. Union Pac. R.R. Co., 557 F.3d 504, 510 (7th Cir.2009) (“A duty of good faith does not mean that a party vested with a clear right is obligated to exercise that right to its own detriment for the purpose of benefiting another party to the contract.”); NLRB. v. Truitt Mfg. Co., 351 U.S. 149, 154-55, 76 S.Ct. 753, 100 L.Ed. 1027 (Frankfurter, J., concurring in part and dissenting in part) (“ ‘Good faith’ means more than merely going through the motions of negotiating; it is inconsistent with a predetermined resolve not to budge from an initial position. But it is not necessarily incompatible with stubbornness or even with what to an outsider may seem unreasonableness.... The previous relations of the parties, antecedent events explaining behavior at the bargaining table, and the course of negotiations constitute the raw facts for reaching such a determination.”).
For reasons I explain below, the way in which California authorized Indian gaming creates a bilateral monopoly that gives both parties an economic incentive to take hardline positions — positions that are perfectly consistent with the obligation to negotiate in good faith. I then explain why the State’s demands are reasonable and its concessions meaningful. Because I do not think it is our place to interfere in these negotiations absent bad faith, I would let the parties work it out (or not).
A
In California, unlike other states, Indian tribes’ territorial exclusivity is enshrined in the state Constitution. California prohibits all entities other than Indian tribes from operating Nevada-style gaming within its borders. Thus, when the State negotiates a class III gaming compact with a tribe, it is negotiating with the only entity legally authorized to offer such gaming in California. The State, for its part, has the sole right to authorize the operation of additional gaming machines by the tribes. The negotiation of a new compact between the State and the Band therefore presents a classic bilateral monopoly, where “two [parties] ... bargain over some thing of value only with each other [and] neither can rely on competition to determine the price of the thing.” In re Hoskins, 102 F.3d 311, 315 (7th Cir.1996), abrogated on other grounds by Assocs. Commercial Corp. v. Rash, 520 U.S. 953, 117 S.Ct. 1879, 138 L.Ed.2d 148 (1997); see generally Black’s Law Dictionary 1098 (9th ed.2009) (defining “bilateral monopoly” as “[a] hypothetical market condition in which there is only one buyer and one seller, resulting in transactional delays because either party can hold out for a better deal without fearing that the other party will turn to a third party”).
*1067The Band enjoys such a monopoly position. Although it can only offer class III gaming pursuant to a Tribal-State compact, see Cal. Const, art. 4, § 19(f), and can only increase its number of gaming devices pursuant to a new or amended compact, at the same time, the Band is the only entity that can offer class III gaming in its immediate territory under the California Constitution. By law, the State has monopoly power over the authorization of gaming machines; equally by law, the Band is the only entity that can be authorized to offer these machines.
Here, the “thing of value” over which the State and the Band are bargaining is the revenue from additional class III gaming devices at Harrah’s Rincon. According to the State’s economic expert, 900 new devices would bring in an additional $40 million in gaming revenue. The “price” being negotiated here is the proportion of new gaming revenue the Band will pay to the State in exchange for these additional devices and an extension of its compact. What economists call the “bargaining range” in a bilateral monopoly can be ascertained by determining the “corner solutions” — i.e., the situations in which one party is an unmitigated winner and the other an unmitigated loser. See generally Walgreen Co. v. Sara Creek Prop. Co., 966 F.2d 273, 278 (7th Cir.1992) (noting that in such situations “[njegotiations would be unlikely to break down completely, given such a bargaining range, but they might well be protracted and costly”). Here, the corner solutions are as follows. At one extreme, the State offers the Band some number of additional gaming devices, but the Band pays 100 percent of new revenues to the State. The State is an unmitigated winner; the Band an unmitigated loser. Assuming the Band realizes $40 million in revenues from 900 new machines, this would result in $40 million for the State, and $0 for the Band. At the other extreme, the State offers the Band some number of additional gaming devices, but the Band pays no more to the State than it does under the current compact; the State gains nothing at all from this transaction, while the Band gains 100% of new revenue. The Band is an unmitigated winner; the State an unmitigated loser. Assuming $40 million in revenues from 900 new machines, this would result in $40 million for the Band, and $0 for the State.
These corner solutions determine the bargaining range in negotiations between the State and the Band for a new compact allowing the Band to operate additional class III gaming devices. Assuming $40 million in new revenues, the bargaining range is anywhere between: (1) $0.01 for the State and $39,999,999 for the Band; and (2) $39,999,999.99 for the State and $0.01 for the Band. Given the vagaries and transaction costs involved in bilateral monopoly, “[tjhere is no way to predict where in the bargaining range the bargain w[ill] be struck,” In re Hoskins, 102 F.3d at 316, but “at any price between those figures ... both parties would be better off,” Walgreen Co., 966 F.2d at 276 (emphasis added). The reason for this “any price” principle is because here, there exists another possible “corner solution” besides the two described above: the Band refuses to share revenues and the State refuses to authorize new gaming devices, and neither party gains any additional revenue. Under the circumstances of the bilateral monopoly at issue here, any compact offer falling within the “bargaining range”— which includes the State’s offers in this case — cannot be seen as made in “bad faith,” since the Band would be offered real net gains in the millions of dollars annually under even the most parsimonious State offer.
Despite the raw economics of the situation, the majority holds that “a state may not take a ‘hard line’ position in IGRA *1068negotiations when it results in a ‘take it or leave it offer’ to the tribe....” Maj. Op. at 1039. What the majority has failed to appreciate is that the State and the Band are in precisely the same position. If the State wants additional revenue (or any additional intangible benefits from Harrah’s Rincon’s expanded operations), and the Band refuses to budge, the State must take or leave the Band’s offer. The State has no other options available to it.
B
Even negotiating within the framework of a bilateral monopoly, the State’s offer is not one-sided. The State argues that it has offered three “meaningful concessions” to the Band: (1) territorial exclusivity, (2) an extension of the compact term, and (3) authorization for additional devices. Although I agree with the majority that the State’s offer of territorial exclusivity is not a meaningful concession in light of Proposition 1A, there is no question that the State has offered a “real” concession to the Band by offering to extend the compact term and to increase the allowed number of class III gaming devices the Band can operate. I will address each in turn.
1
As an initial matter, I agree with the majority that the State’s “offer” of territorial exclusivity is not a “meaningful concession” within the meaning of Coyote Valley II. See Maj. Op. at 1037-39. The California Constitution already guarantees tribal exclusivity against private gaming interests, and under the Band’s current compact, as modified by Addendum A, the Band retains the right to terminate its compact or to continue under the compact with reduced revenue sharing requirements “[i]n the event the exclusive right of Indian tribes to operate Gaming Devices in California is abrogated .... ” The “revised and expanded” exclusivity provision provides no meaningful protection beyond that afforded by the Band’s current compact, and thus cannot serve as consideration for an amended compact. In other words, it is not a “real” concession, see Coyote Valley II, 331 F.3d at 1112, but an illusory one.
2
The State’s other offered concessions, however, are unquestionably “real.” First, the State’s offer to extend the expiration date of the Band’s compact to 2045 is quite meaningful. The Band’s current compact, which expires on December 31, 2020 (or June 30, 2022 if a party requests negotiations for an extension), contains no right to automatic renewal. Under the Supreme Court’s decision in Seminole Tribe v. Florida, 517 U.S. 44, 116 S.Ct. 1114, 134 L.Ed.2d 252 (1996), a tribe cannot sue a state in federal court under 25 U.S.C. § 2710(d)(7) et seq. absent an express waiver of sovereign immunity by the state. The current compact between California and the Band contains a limited waiver from the State of its sovereign immunity. See 1999 Compact § 9.4 et seq., http:// www.cgcc.ca.gov/enabling' tsc.pdf (last visited Apr. 6, 2010). Once the compact expires, the State will not longer be subject to suit by the Band to enforce any of IGRA’s requirements, including its good faith negotiation requirement. See Seminole Tribe, 517 U.S. at 72, 116 S.Ct. 1114. Thus, by offering to extend the Band’s compact until 2045, the State is offering to subject itself to IGRA’s requirements with respect to negotiations with the Band for an additional twenty-five years. To my mind, the State’s promise to extend the compact and waive its sovereign immunity for another twenty-five years is a substantial concession. Without it, the Band has a casino resort until 2020, but not beyond.
3
Even more “meaningful” is the State’s offer to increase the number of gaming *1069devices the Band may operate from 1600 to 2500 devices. Under the California Constitution, “[t]he Legislature has no power to authorize, and shall prohibit, casinos of the type currently operating in Nevada and New Jersey” — i.e., establishments offering Class III gaming. Cal. Const, art. 4, § 19(e). California’s Constitution provides a limited exception to this prohibition with respect to federally recognized Indian tribes on Indian lands: “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,” and “slot machines, lottery games, and banking and percentage card games are ... permitted to be conducted and operated on tribal lands subject to those compacts.” Id. art. 4, § 19(f) (emphasis added). Thus, although the California Constitution allows Indian tribes to operate slot machines, banked card games, and similar activities, it does not provide blanket authorization for their operation; the Constitution explicitly requires that casino-style gaming be operated “subject to [a] compact[ ]” between the Governor and the Tribe.
Critically, since the very inception of Indian gaming compacts in California, the State has negotiated, in gaming compacts, limits on the scope of class III gaming, including the number of gaming devices that a tribe can operate. See 1999 Model Tribal-State Gaming Compact § 4.3.1 (“The Tribe may operate no more Gaming Devices than the larger of the following: (a) A number of terminals equal to the number of Gaming Devices operated by the Tribe on September 1, 1999; or (b) Three hundred fifty (350) Gaming Devices.”), http://www.egcc.ca.gov/enabling/ tsc.pdf (last visited Apr. 6, 2010).20 As the majority itself concedes, see Maj. Op. at 1038-39 & n. 20, device licensing is a proper subject for negotiation under IGRA, see 25 U.S.C. § 2710(d)(3)(C)(vi). Since tribal exclusivity is already written into the California Constitution, the number of authorized gaming devices is, at this point, the single most important issue subject to negotiation in Tribal-State compacts in California.21
*1070When the State offered to allow the Band to operate a greater number of machines, it was not only offering a “real” or “meaningful” concession, it was offering a concession on the single most “meaningful” subject over which it retained control. Under our decision in Coyote Valley II, this means that the State did not violate § 2710(d)(4) in connection with the proposed compact. See 331 F.3d at 1112. Since “offered concessions by [the] State [we]re real, § 2710(d)(4) d[id] not categorically prohibit [the] fee demands” in this case. Id.
C
Next, I wish to address why, in light of these concessions, the disparity between the revenue sharing demanded by the State and the profits to be realized by the Band is not as great as the majority portrays it to be and why it is not, therefore, evidence of bad faith negotiations. When the Band entered into its original compact in 1999, it was operating less than two hundred gaming devices. As a consequence, the Band was exempt from making revenue payments into the SDF. In fact, the Band does not now make (and never has made) any SDF payments under its current compact.22 In stark contrast to every tribe that has negotiated, or renegotiated, a compact with California since 1999, the Band shares no revenue with the State. But this will likely change in 2020. Without a doubt, the Band’s sweetheart deal with respect to revenue sharing will not last past its current compact: every compact negotiated in California since 1999 includes either SDF or general fund revenue sharing.23 Thus, any compact negotiation between the Band and the State — whether it be for an amended compact today or for a wholly new compact a decade from now — will inevitably include a State demand for revenue sharing of some sort; to do otherwise would put the Band in a preferred position vis-a-vis the other tribes in California.
In Coyote Valley II, we approved of the SDF revenue sharing provisions in compacts negotiated by the Davis Administration. We recognized that “the contributions tribes must make to the SDF are significant,” but concluded that, in light of the State’s concession of exclusivity, it was not “inimical to the purpose or design of IGRA for the State ... to ask for a reasonable share of tribal gaming revenues.” 331 F.3d at 1114-15. Here, the majority declares that the State’s revenue sharing proposals in the negotiations with the Band differ so vastly from the “reasonable share of tribal revenues” paid under the SDF provision by the Coyote Valley II tribes that the State’s bad faith is conclusively established:
[T]he calculations presented by the State’s own expert reveal that the financial benefit to Rincon from the amendments proposed would be negligible: Rincon stood to gain only about $2 million in additional revenues compared to the State’s expected $38 million. Thus, in stark contrast to Coyote Valley II, the relative value of the demand versus the concession here strongly suggests the State was improperly using its authority over compact negotiations to impose, *1071rather than negotiate for, a fee. Under IGRA and Coyote Valley II, that is bad faith.
Maj. Op. at 1038 (citation omitted).
The majority’s analysis on this point could not diverge any more sharply from the reality of what we deemed to be a “reasonable share of tribal gaming revenues” in Coyote Valley II. 331 F.3d at 1115. Under the SDF formula — and using the same economic model used to calculate the majority’s $38 million-$2 million split— the Band would be required to pay $35.8 million to the state, and would keep approximately $4 million in net revenues.24 There is simply no logical basis for simultaneously deeming a formula leading to a $36 million-$4 million split in net revenues “a reasonable share,” and a formula leading to a $38 million-$2 million split in net revenues so substantively and objectively unreasonable as to constitute “bad faith.” I do not see how the majority’s analysis can be reconciled with Coyote Valley II.
Furthermore, although the Band is not required to make SDF payments under its current compact,25 once the Band’s current compact expires in 2020, the State will certainly insist on revenue sharing at a level at least equal to the SDF level required of other tribes and already approved by this court in Coyote Valley II. Thus, come 2020, the Band will likely be faced with the following choice: negotiate a new compact with the State that will include at least SDF-level revenue sharing, or cease class III gaming. The State’s offer here, however, not only extends the Band’s compact until 2045, but allows the Band to make more money now and over the next decade leading up to 2020. The State has offered a ten-year revenue stream of $2 million annual payments to which the Band would not otherwise be entitled; after ten years has passed, the Band will be able to conduct class III gaming under a revenue sharing agreement similar to that which the State would no doubt insist upon in negotiating a new compact. In sheer monetary terms, the State is offering concessions with a net present value of approximately $19.5 million.26 If this is not “meaningful,” I’m not sure what is.
Nothing that I have said here is an endorsement of the State’s terms. I don’t know whether the State’s offer is a “good” one or not. The point is not whether the State’s offer is “good” or “bad.” We are not the Band’s advisors or guardians ad litem. The Band has been well represented in this proceeding, and I am sure it has been well advised throughout these negotiations as well. All we have to decide is *1072whether the State has negotiated in good faith, not whether it has made the Band a good offer. I would hold that these terms have been offered in good faith.
D
Finally, there is abundant real-world evidence that the State’s offers were not made in bad faith: specifically, numerous tribes in California (none of whom are barred by the Eleventh Amendment from suing California in federal court) have agreed to compact terms effectively identical to those offered by the State. Fifteen tribes currently contribute gaming revenue to the State’s General Fund, see http:// www.cgcc.ca.gov/ Tribal/2009/GFPayingT ribesforWebsite 03-04-09.pdf (last visited Apr. 6, 2010), all pursuant to compacts negotiated or renegotiated since 2003, http://www.cgcc.ca.gov/compacts.asp (last visited Apr. 6, 2010). Notably, eleven of the tribes that currently contribute class III gaming revenues to the State’s general fund are 1999 Compact tribes (like the Band) who now contribute to the State’s general fund pursuant to renegotiated compacts that allowed these tribes to operate additional devices in exchange for general fund revenue sharing. Id.; see, e.g., Shingle Springs Band of Miwok Indians Compact' § 4.3.1, http://www.cgcc.ca.gov/ compacts/Shingle% 20Springs% 20 Compact.pdf (last visited Apr. 6, 2010) (amending 1999 Compact to increase allowed devices to 5,000 and providing for general fund revenue sharing of 20% for first $200 million in net win and 25% for net win exceeding $200 million). And whatever questions these compacts may have raised for the Secretary of the Interior, the Secretary ultimately accepted the negotiations and declined to disapprove all of the compacts and their amendments.
In other words, the State has regularly proposed, the tribes have regularly accepted, and the Secretary has regularly approved contract terms similar to those proposed by the State in this case. Although it is true, as the majority posits, that the fact that numerous tribes have agreed to general fund revenue sharing would not insulate such revenue sharing from IGRA’s substantive requirements, it defies logic to hold, as does the majority, that a contract proposal readily agreed to by a dozen other tribes was so outrageously one-sided that it could only have been made in “bad faith.” See Maj. Op. at 1029-30,1041-42; see also Wis. Elec. Power Co., 557 F.3d at 510 (“the duty of good faith does not require your putting one of your customers ahead of the others”). Yet this is what 25 U.S.C. § 2710(d)(7)(B)(iii) requires for court intervention. The provision does not prohibit “hard bargaining” within the negotiating range of the Tribal-State compact bilateral monopoly, but is instead concerned solely with “bad faith” based on an “inquiry [that] is nuaneed and fact-specific, and is not amenable to bright-line rules,” Coyote Valley II, 331 F.3d at 1113, and which takes into account “the previous relationship of the tribe and State,” S. Rep. No. 100-446, at 14 (1988), reprinted in 1988 U.S.C.C.A.N. 3071, 3084. The negotiations at issue here, even though inconclusive, have been well within those parameters.
As we explained in Coyote Valley II:
Congress did not intend to allow States to invoke their economic interests as a justification for excluding Indian tribes from class III gaming; nor did Congress intend to permit States to use the compact requirement as a justification for the protection of other State-licensed gaming enterprises from free market competition with Indian tribes. By the same token, however, Congress also did not intend to require that States ignore their economic interests when engaged in compact negotiations. See [S. Rep. No. 100-446, at 2, reprinted in 1988 U.S.C.C.A.N. 3071, 3071] (“An objective *1073inherent in any government regulatory scheme is to achieve a fair balancing of competitive economic interests.”). Indeed, § 2710(d)(7)(B)(iii)(I) expressly provides that we may take into account the “financial integrity” of the State and “adverse economic impacts on the State’s existing gaming activities” when deciding whether the State has acted in bad faith, and IGRA’s legislative history explains that a State’s governmental interests with respect to class III gaming on Indian lands include its economic interest in raising revenue for its citizens.
331 F.3d at 1115 (quotation marks, brackets, ellipses, and internal citation omitted).
VI
The majority’s legal errors carry grave- and widespread — practical repercussions. The majority’s decision will call into question Tribal-State gaming compacts not just in cash-strapped California, where no fewer than fifteen Tribal-State compacts provide for general fund revenue sharing, see http://www.cgcc.ea.gov/Tribal/2009/ GF% 20Paying% 20Tribes% 20for% 20We bsite% 2003-04-09.pdf (last visited Apr. 6, 2010), but throughout the country.27 The Second Circuit has never addressed a legal challenge to the Connecticut compacts governing the behemoth Foxwoods and Mohegan Sun Casinos, but the majority decision here will inevitably spur such challenges in Connecticut and in New York. The Sixth, Tenth, and Eleventh Circuits have yet to consider the validity of general revenue sharing under IGRA, but it can reasonably be expected that district court clerks in Michigan, New Mexico, Oklahoma, and Florida will be docketing legal challenges sometime soon. These lawsuits — based on a misreading of IGRA by the majority in the first federal appellate decision to squarely address the legality of negotiated general fund revenue sharing — will eat up State, tribal, and federal resources and will unsettle dozens of mutually beneficial revenuesharing provisions that have fed both tribal coffers and revenue-hungry state treasuries.
I respectfully dissent.
. Like the unique relationship between Congress and the tribes, see U.S. Const, art. I, § 8, cl. 3 (giving Congress the power "[t]o regulate commerce with foreign nations, and among the several states, and with the Indian Tribes”), the special tax immunity enjoyed by Indian tribes has its roots in the Constitution, see id. art. I, § 2, cl. 3 ("Representatives and direct Taxes shall be apportioned among the several States which may be included within this Union, according to their respective Numbers, which shall be determined by adding to the whole Number of free Persons, including those bound to Service for a Term of Years, and excluding Indians not taxed, three fifths of all other Persons.” (emphasis added)); see also id. amend. XIV, § 2 (“Representatives shall be apportioned among the several states according to their respective numbers, counting the whole number of persons in each State, excluding Indians not taxed.”).
. “The source of federal authority over Indian matters has been the subject of some confusion, but it is now generally recognized that the power derives from federal responsibility for regulating commerce with Indian tribes and for treaty making.” McClanahan v. Ariz. State Tax Comm’n, 411 U.S. 164, 172 n. 7, 93 S.Ct. 1257, 36 L.Ed.2d 129 (1973) (citing U.S. Const, art. I, § 8, cl. 3; id. art. II, § 2, cl. 2 (additional citations omitted)).
. The 1999 compact also provided for revenue sharing between gaming and non-gaming tribes through the creation of a "Revenue Sharing Trust Fund” ("RSTF”). Id. § 4.3.2 et seq. The RSTF provision required gaming tribes to pay, on a quarterly basis, a per-device annual fee into the RSTF for distribution to non-gaming tribes in California. Id. § 4.3.2.2(2). The RSTF provisions of the 1999 Compact were solely directed at gaming and non-gaming tribes — "[i]n no event [was] the State's General Fund [ ] obligated to make up any shortfall or pay any unpaid claims.” Id. § 4.3.2.1(b).
. "Levied” is simply a tax term of art for "imposed.” See Black's Law Dictionary 991 (9th ed.2009) (defining "levy" as “[t]o impose or assess (a fine or tax) by legal authority”); Webster’s Third New International Dictionary 1301 (1993) (defining "levy” as "to impose or collect (as a tax or tribute) by legal process or by authority”).
. IGRA also recognized the distinction between a tax and a direct tax. Compare 25 U.S.C. § 2710(d)(4) (discussing the "imposfition] of any tax, fee, charge, or other assessment upon an Indian tribe”) with 25 U.S.C. § 2710(d)(7)(B)(iii)(II) (discussing “any demand by the State for direct taxation of the Indian tribe or of any Indian lands”). A direct tax is
[a] tax that is imposed on property, as distinguished from a tax on a right or privilege. A direct tax is presumed to be borne by the person upon whom it is assessed, and not “passed on” to some other person.
Black’s Law Dictionary 1595 (9th ed.2009). See also Pollock v. Farmers' Loan & Trust Co., 157 U.S. 429, 588, 15 S.Ct. 673, 39 L.Ed. 759 (1895) (Field, L, concurring) ("Direct taxes. in a general and large sense, may be described as taxes derived immediately from the person, or from real or personal property, without any recourse therefrom to other sources for reimbursement.”), superseded by U.S. Const. amend. XVI; Quarty v. United States, 170 F.3d 961, 970 (9th Cir.1999) ("In contrast to an excise or indirect tax, which is levied upon the use or transfer of property even though it might be measured by the property’s value, a direct tax is levied upon the property itself.”) (quotation marks and citations omitted); Laurence H. Tribe, American Constitutional Law 843 (3d ed. 2000) (“A direct tax is one imposed upon property as such, rather than on the performance of an act.”).
. States have their own forms of intergovernmental revenue sharing. For example, California has decided to share its cigarette tax revenues with its counties and cities, but not with the Indian tribes within its borders. Cal. Rev. & Tax § 30462; see Chemehuevi Indian Tribe v. Cal. State Bd. of Equalization, 800 F.2d 1446, 1450 (9th Cir.1986).
. I do not doubt for a moment that a tribe would regard a demand for revenue sharing with the same disdain it would the imposition of a tax. That is, a tribe may equate revenue sharing and taxation because they have the same economic effect on its welfare. But the legal difference between the two is significant, for reasons I have described above.
A simple example illustrates my point about the difference between an economic and a legal effect. Whether my landlord raises my rent by $100 a month or the government increases my income taxes by $100 a month, the economic impact on me is the same: I am out an additional $100 a month. But there are legal and political consequences where government imposes additional taxes that are not found in the rent increase. No one would mistake the rent increase for a tax, even though my checkbook can't tell the difference between the two.
. The majority opines that "[t]he language and structure of § 2710(d)(3)(C) suggest it is exhaustive.” Maj. Op. at 1028 n. 9. But the majority's analysis on this point runs contrary to our longstanding interpretation of the compact requirement’s closest legal analog: the "good faith bargaining” requirement in the National Labor Relations Act. See In re Indian Gaming Related Cases ("Coyote Valley I"), 147 F.Supp.2d 1011, 1020-21 (N.D.Cal.2001) ("Like IGRA, the NLRA imposes a duty to bargain in good faith, but does not expressly define 'good faith.' ... The Court does not intend to import federal case law interpreting the NLRA wholesale into its interpretation of the IGRA.... However, the Court considers the NLRA case law for guidance in interpreting a standard undefined by the IGRA.”), aff'd, Coyote Valley II, 331 F.3d 1094.
Under section 158 of the NLRA, 29 U.S.C. § 158 et seq., there is an "[ojbligation to bargain collectively,” defined as "the performance of the mutual obligation of the employer and the representative of the employees to meet at reasonable times and confer in good *1061faith with respect to wages, hours, and other terms and conditions of employment....” Id. § 158(d). The list "wages, hours, and other terms and conditions of employment” could easily be read to comprise the entirety of permissible subjects that could be collectively bargained; i.e., the list could be read as "exhaustive,” an interpretation the majority believes is compelled by the language and structure of § 2710(d)(3)(C) in IGRA. Like the subjects in 25 U.S.C. § 2710(d)(3)(C), the subjects in 29 U.S.C. § 158(d) include a broad final clause — in IGRA, "any other subjects that are directly related to the operation of gaming activities,” and in the NLRA, "other terms and conditions of employment.” Yet 29 U.S.C. § 158(d) has not been interpreted to establish an "exhaustive” list of subjects that may be negotiated in collective bargaining. Instead, we have interpreted § 158(d) as setting out "[mjandatory subjects — 'wages, hours, and other terms and conditions of employment’ — [that] must be bargained collectively in good faith,” Retlaw Broad. Co. v. NLRB, 172 F.3d 660, 665 (9th Cir.1999), but which are not the only subjects subject to collective bargaining. Rather, "all other subjects are permissive subjects,” which "deal with a wide range of matters deemed to fall outside the framework of mandatory terms.” /¿.(quotation marks omitted).
. To the extent the majority relies on this canon beyond its interpretation of § 2710(d)(3)(C)(vii) (to which we have already held the canon inapplicable), even this reliance is flawed. As we explained in Coyote Valley II and as the Supreme Court made clear in Seminole Tribe, IGRA involved two countervailing sovereign interests-states and Indian tribes-and it granted authority to each. See Coyote Valley II, 331 F.3d at 1096 ("IGRA ... 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.”); Seminole Tribe, 517 U.S. at 58, 116 S.Ct. 1114 ("[T]he Act grants the States a power that they would not otherwise have....”). Although states have not traditionally had the power to regulate Indian tribes, states have traditionally had the power to regulate the scope of gambling within their territorial borders. As the Supreme Court has explained, "Congress should make its intention clear and manifest if it intends to pre-empt the historic powers of the States.... This plain statement rule is ... an acknowledgment that the States retain substantial sovereign powers under our constitutional scheme, powers with which Congress does not readily interfere.” Gregory v. Ashcroft, 501 U.S. 452, 461, 111 S.Ct. 2395, 115 L.Ed.2d 410 (1991) (quotation marks and internal citation omitted); see also, e.g., Owasso Indep. Sch. Dist. No. I-011 v. Falvo, 534 U.S. 426, 435-36, 122 S.Ct. 934, 151 L.Ed.2d 896 (2002) (rejecting a proposed interpretation of the ambiguous term "education records” in the Family Education Rights and Privacy Act where it was "doubt[ful] Congress meant to intervene in this drastic fashion with traditional state functions.... The Congress is not likely to have mandated this result, and we do not interpret the statute to require it.”). Thus, to the extent that IGRA's provisions other than § 2710(d)(3)(C)(vii) are ambiguous, this is a "dueling canons” situation in which it is far from clear that the federalism canon should be shoved aside in favor of the tribal sovereignty canon. See, e.g., Fla. Dept. of Revenue v. Piccadilly Cafeterias, Inc., 554 U.S. 33, 128 S.Ct. 2326, 2338, 171 L.Ed.2d 203 (2008) ("We agree with Florida that the federalism canon ... obliges us to construe [a statutory exemption from state taxation] narrowly.”).
. See, e.g., Katie Eidson, Note, Will States Continue to Provide Exclusivity in Tribal Gaming Compacts or Will Tribes Bust on the Hand of the State in Order to Expand Indian Gaming?, 29 Am. Indian L. Rev. 319, 325-39 (2004); Courtney J.A. DaCosta, Note, When “Turnabout" Is Not "Fair Play”: Tribal Immunity under the Indian Gaming Regulatory Act, 97 Geo. L.J. 515, 542-546 (2009); Ezekiel J.N. Fletcher, Negotiating Meaningful Concessions from States in Gaming Compacts to Further Tribal Economic Development: Satisfying the “Economic Benefits" Test, 54 S.D. L. Rev. 419, 431-39 (2009); Matthew L.M. Fletcher, Bringing Balance to Indian Gaming, 44 Harv. J. on Legis. 39, 57-95 (2007); Eric S. Lent, Note, Are States Beating the House?: The Validity of Tribal-State Revenue Sharing under the Indian Gaming Regulatory Act, 91 Geo. L.J. 451, 453-74 (2003); Matthew Murphy, Note, Betting the Ranchería: Environmental Protections as Bargaining Chips under the Indian Gaming Regulatory Act, 36 B.C. Envtl. Aff. L. Rev. 171, 181-84 (2009); Kathryn R.L. Rand, Caught in the Middle: How State Politics, State Law, and State Courts Constrain Tribal Influence over Indian Gaming, 90 Marq. L. Rev. 971, 985-87 (2007); Kathryn R.L. Rand and Steven Andrew Light, How Congress Can and Should "Fix” the Indian Gaming Regulatory Act: Recommendations for Law and Policy Reform, 13 Va. J. Soc. Pol’y & L. 396, 462-65 (2006); Joshua L. Sohn, The Double-Edged Sword of Indian Gaming, 42 Tulsa L. Rev. 139, 150-56 (2006).
. Much of the academic literature disapproves of tribal-state revenue sharing as a normative matter, but the legal consensus— particularly among professorial, as opposed to student, authors — appears to be that IGRA is simply silent on the issue of revenue sharing. Commentators have suggested that Congress clarify the issue. See Fletcher, 44 Harv. J. On Legis. at 94 ("argu[ing] that a legislative package to ratify and authorize revenue sharing ... [would] alleviate the major problems in IGRA ... ”); Rand and Light, 13 Va. J. Soc. Pol'y & L. at 462 (recommending that Congress "[a]mend IGRA to create a revenue-sharing structure that encourages cooperative tribal-state policymaking while ensuring tribes remain the primary beneficiaries of Indian gaming by creating a presumptive rebut-table cap”); see also Wisconsin v. Ho-Chunk Nation, 512 F.3d 921, 932 (7th Cir.2008) (“The validity, under the IGRA, of revenue-sharing agreements in tribal-state compacts has been a contentious issue. ... [T]he legitimacy of these revenue-sharing provisions is far from a settled issue.”).
. See Mashantucket Pequot Tribe Memorandum of Understanding (Apr. 30, 1993) (providing for revenue sharing of twenty-five percent of gross operating revenues from electronic gaming machines), http://www.ct. gov/ dosr/hb/dosr/Memorandum_Of_Understanding_Foxwoods.pdf (last visited Apr. 6, 2010); Mohegan Tribe Memorandum of Understanding (May 17, 1994), http://www.ct. gov/dosr/lib/dosr/Memorandum_0£_ Understanding_Mohegan.pdf (last visited Apr. 6, 2010) (same); see also http://www.ct. gov/dos r/hb/dosr/tribal_state_compact_foxwoods.pdf (last visited Apr. 6, 2010) (stating that assessment pursuant to 25 U.S.C. *1064§ 2710(d)(3)(C)(iii) is to be paid into State’s general fund); http://www.ct.gov/dosr/lib/ dosr/tribal_state_compact_mohegan.pdf (last visited Apr. 6, 2010) (same).
. See Seminole Tribe of Florida Compact Pt. XI and Exh. A (Nov. 14, 2007), http://www. flgov.com/pdfs/2007114compact — exhibita.pdf (last visited Apr. 6, 2010) (providing for revenue sharing of between ten and twenty-five percent of the tribe's net win, and noting that "[t]he appropriation of any Payments received by the State pursuant to this Compact lies within the exclusive prerogative of the Legislature”).
. See, e.g., Match-E-Be-Nash-She-Wish Band of Pattawatomi Indians of Michigan Compact § 15-17 (May 9, 2007), http://www. michigan.gov/ documents/mgcb/Gunlake_Compac1_276443_7.pdf (last visited Apr. 6, 2010) (providing for eight to twelve percent of net win to be paid into Michigan Strategic Fund, a state economic development fund in no way related to gaming or its effects, providing for additional two percent revenue sharing between the tribe and affected local governments, and explaining that ”[b]y entering into this agreement neither the Tribe nor the State of Michigan intend to create any new authority, nor to expand or diminish any existing authority, on the part of the State of Michigan to impose taxes upon the Tribe, its members, or any person or entity doing business with the Tribe pursuant to this Compact”).
. See Seneca Nation of Indians Compact § 12(b) (April 12, 2002), http://www.ncai.org/ n cai/resource/agreements/ny_gaming-seneca_nation-4-12-2002.pdf (last visited Apr. 6, 2010) (providing for general fund revenue sharing of eighteen to twenty-five percent, depending on year).
. See Tribal-State Class III Gaming Compact § 11 (2007), http://www.nmgcb.org/ tribal/2007%20compact.pdf (last visited Apr. 6, 2010) (providing for payment by tribe of a percentage of gaming machine net win "to the State Treasurer for deposit into the General Fund of the State”).
. See Oklahoma Tribal-State Gaming Act Model Tribal Gaming Compact, 3A Okla. Stat. § 281 Part 11 (outlining percentage of adjusted gross gaming revenues that are to be paid to the State by tribes, and providing: "Payments of such fees shall be made to the Treasurer of the State of Oklahoma. Nothing herein shall require the allocation of such fees to particular state purposes, including, but not limited to, the actual costs of performing the state’s regulatory responsibilities hereunder.”).
. See, e.g., Amendments to the Menominee Indian Tribe of Wisconsin Compact § 42 (Apr.2003), http://www.doa.state.wi.us/ docview.asp? docid=2147 (last visited Apr. 6, 2010) (amending tribal-state compact to provide for general fund, rather than earmarked, revenue sharing).
. The majority has erred to the extent it implies that the Eleventh Amendment prevents the Secretary from disapproving compacts he considers illegal under IGRA. See Maj. Op. at 1026 n. 8. Under IGRA, the Secretary may disapprove any compact he determines "violates — (i) any provision of this Act, (ii) any other provision of Federal law that does not relate to jurisdiction over gaming on Indian lands, or (iii) the trust obligations of the United States to Indians.” 25 U.S.C. § 2710(d)(8)(B)(i)-(iii) (emphasis added). The Secretary, who is "charg[ed by Congress] ... with broad responsibility for the welfare of Indian tribes,” Udall, 366 F.2d at 672, and owes “the Indians ... [a] duty of care and protection,” Rainbow, 161 F. at 838, is charged by IGRA with conducting an independent review of the legality of compacts, irrespective of whether the tribes can challenge those compacts in federal court. Cf. Knight v. United States Land Ass’n, 142 U.S. 161, 12 S.Ct. 258, 35 L.Ed. 974 (1891) (“The Secretary[of the Interior] is the guardian of the people of the United States over the public lands. The obligations of his oath of office oblige him to see that the law is carried out ....”).
Moreover, as the majority itself explains, where the Secretary has had concerns about a compact but nonetheless felt the compact should not be disapproved, he has voiced these concerns explicitly. See Pueblo of Sandia v. Babbitt, 47 F.Supp.2d 49, 51 (D.D.C.1999). Yet the Secretary has approved, or *1065deemed approved without any cautionary language, dozens of compacts or compact amendments providing for general revenue sharing, including all fifteen compacts calling for general revenue sharing by California tribes. See, e.g., 69 Fed.Reg. 76,004 (Dec. 20, 2004) (approving compact between California and Coyote Valley Band of Porno Indians); 72 Fed Reg. 36,717 (July 5, 2007) (approving compacts between State of New Mexico and the Pueblo of Isleta, Pueblo of Nambe, Pueblo of Picuris, Pueblo of San Felipe, Pueblo of Sandia, Pueblo of Santa Ana, Pueblo of Tesuque, Pueblo of Taos, Pueblo of Santa Clara and Ohkay Owingeh); 72 Fed.Reg. 58,333 (Oct. 15, 2007) (approving compact between New Mexico and Pueblo of Laguna); 72 Fed. Reg. 71,939-40 (Dec. 19, 2007) (deeming approved, without cautionary language, compacts between California and Morongo Band of Mission Indians, Pechanga Band of Luiseno Indians, and Agua Caliente Band of Cahuilla Indians); 73 Fed.Reg. 1229 (Jan. 7, 2008) (deeming approved compact between Florida and Seminole Tribe of Florida); 73 Fed.Reg. 3480 (Jan. 18, 2008) (approving compact between California and San Manuel Band of Mission Indians); 74 Fed.Reg. 18397 (Apr. 22, 2009) (deeming approved, without cautionary language, compact between Michigan and Match-E-Be-Nash-She-Wish Band of Pottawatomi Indians of Michigan).
. The Band operates its casino subject to a compact effectively identical to this 1999 Model Compact. See http://www.cgcc.ca.gov/ Tribal/ 1999CompactTribesGamingándÑon-Gaming% 201ist.pdf (last visited Apr. 6, 2010).
. The majority relies on a DOI letter to the Forest County Potawatomi Community of Wisconsin for the proposition that negotiations over the number of gaming devices cannot serve as adequate consideration for revenue sharing. Maj. Op. at 1039-40. Yet this letter, and the views expressed therein, are irrelevant here. In the letter, the Assistant Secretary of Indian Affairs explains that "[w]e have not, nor are we disposed to, authorize revenue-sharing payments in exchange for compact terms that are routinely negotiated by the parties as part of the regulation of gaming activities, such as duration, number of gaming devices, hour of operation, and wager limits.” Maj. Op. at 1039 (emphasis added). However, in Wisconsin, unlike in California, the number of gaming devices a tribe may operate has not been limited by compact, but has been subject only to reporting requirements and tribal self-regulation. See, e.g., Bad River Band of the Lake Superior Tribe of Chippewa Indians Gaming Compact § XV(C)(3), http:// www.doa.state.wi.us/docview.asp?docid= 2127 (last visited Apr. 6, 2010); Amendments to the Bad River Band of the Lake Superior Tribe of Chippewa Indians Gaming Compact § 7, http://www.doa.state.wi.us/ doc-view.asp?docid=2129 (last visited Apr. 6, 2010). Since a Wisconsin tribe could already increase the number of devices it operated without a compact amendment, the DOI opined that an offer to increase the number of allowed devices was not a meaningful concession by the State of Wisconsin. This is in stark contrast to California, where the num*1070ber of allowed gaming devices has always been set by compact, and therefore is not now, and never has been, “routinely negotiated by the parties as part of the regulation of gaming activities.”
. See List of SDF Paying Tribes, http://www. cgcc.ca.gov/Tribal/2008/ SDFPayingTribesforWebsite2008.pdf (enumerating twenty-one tribes making SDF payments under their current compacts (last visited Apr. 6, 2010), of which the Band is not one).
. See generally Tribal-State Gaming Compacts, http://www.cgcc.ca.gov/compacts.asp (last visited Apr. 6, 2010).
. These calculations are based on the State’s expert testimony referred to in the majority opinion. Using a net win of $369 per day per machine for 2,500 machines gives the following payments to the State:
(1) Under the State's 10 percent/15 percent revenue sharing proposal deemed to be “bad faith” by the majority:
(0.10 2005 Gross Revenues of $197 million) + (0.15 $369/day 365 days 900 machines) = $37.9 million
(2) Under the SDF structure we deemed to be “a reasonable share of tribal revenues” in Coyote Valley II:
$369/day 365 days (0.07 300 machines + 0.10 500 machines + 0.13 1500 machines) = $35.8 million
. Tribes that did not operate 200 or more gaming devices as of September 1, 1999 were not required to contribute to the SDF under the 1999 Compact. The Band was not operating 200 or more devices as of September 1, 1999 and is therefore exempt from SDF payments under its current compact.
. This net present value calculation assumes the federal funds discount rate remains at a near-record low 0.50. If the rate were to rise to a more normal level — averaging, say, 2.25 over the next decade — the net present value of the bargained-for income stream would decrease to $17.7 million, still a "meaningful” sum.
. The majority is correct to point out that although California has waived its sovereign immunity to suits under IGRA, not all states have done so. But the majority is premature to conclude that tribes in those states could not challenge the legality of compact negotiations. See Maj. Op. at 1026 n. 8. After the Seminole Tribe decision, the Secretary promulgated 25 C.F.R. Part 291, which “establish[es] procedures that the Secretary will use to promulgate rules for the conduct of Class III Indian gaming when: (a) A State and an Indian tribe are unable to voluntarily agree to a compact and; (b) The State has asserted its immunity from suit brought by an Indian tribe under 25 U.S.C. § 2710(d)(7)(B).” 25 C.F.R. § 291.1. In 2007, a divided Fifth Circuit panel held that IGRA did not authorize 25 C.F.R. Part 291. See Texas v. United States, 497 F.3d 491 (5th Cir.2007), cert. denied, -U.S. -, 129 S.Ct. 32, 172 L.Ed.2d 18 (2008). It appears that the Department of the Interior has refused to acquiesce in the Fifth Circuit’s decision, see Alabama v. United States, 630 F.Supp.2d 1320, 1332 (S.D.Ala.2008), indicating that the invalidity of the regulations is hardly a settled issue outside the Fifth Circuit. Notably, no other circuit court — including the Second, Sixth, Tenth, and Eleventh Circuits (home to Connecticut, Michigan, New Mexico and Oklahoma, and Florida, respectively) — has held the Part 291 regulations to be invalid.