The Indian Gaming Regulatory Act (IGRA), 25 U.S.C. § 2701 et seq., provides that a state must negotiate in good faith with its resident Native American tribes to reach compacts concerning casino-style gaming on Native American lands. Defendants-Appellants/Cross-Appellees the State of California (the State) and Governor Arnold Schwarzenegger (Governor Schwarzenegger) (collectively as parties to this litigation, the State) appeal the district court’s finding that, in violation of IGRA, 25 U.S.C. § 2710(d)(3)(A), the State negotiated in bad faith with Plaintiff-Appellee/Cross-Appellant the Rincon Band of Luiseno Mission Indians (Rincon) concerning amendments to the parties’ existing tribal-state gaming compact.
The district court based its bad faith finding on the State’s repeated insistence that Rincon pay a portion of its net revenues into the State’s general fund, which the district court determined to be an attempt by the State to impose a tax on the tribe in violation of 25 U.S.C. § 2710(d)(4).
The State challenges the district court’s characterization of its requests as an attempt to impose a tax, and argues that even if it was attempting to impose a tax, that alone is insufficient to support the finding of bad faith. We affirm.1
FACTUAL AND PROCEDURAL BACKGROUND
The 1999 Compacts
In the fall of 1999, the State (through then-governor Gray Davis) and Rincon negotiated a compact granting Rincon the right to operate casino-style (class III2 ****) gaming on its lands located near San Diego, California, subject to certain limitations.3 Simultaneously, the State’s negotiations also resulted in similar compacts *1023with dozens of other tribes across California. Although some of the 1999 compacts have since been renegotiated, the 1999 compact between Rincon and the State remains operative.
While negotiations over the 1999 compacts were pending, the California Supreme Court handed down its decision in Hotel Employees & Restaurant Employees International Union v. Davis, 21 Cal.4th 585, 88 Cal.Rptr.2d 56, 981 P.2d 990 (1999). In that case, the California Supreme Court determined that the California constitution prohibited everyone in the state, including Indian tribes, from operating Las Vegas-style casinos. As a major consideration, and in order to make the proposed 1999 compacts legally enforceable, the State sponsored a constitutional amendment — Proposition 1A — that would authorize tribal gaming in California.4
In March 2000, California voters approved Proposition 1A, thereby vivifying the 1999 compacts. Not only did Proposition 1A permit tribes to conduct class III gaming lawfully, it effectively gave tribes a state constitutional monopoly over casino gaming in California. In re Indian Gaming Related Cases (Coyote Valley II, 331 F.3d 1094, 1103 (9th Cir.2003)).
Revenue Sharing Under the 1999 Compacts
In consideration for the State’s efforts in securing the passage of Proposition 1A (without which the tribes would have been barred from conducting class III gaming in the State of California), the tribes agreed to share a portion of their expected revenues. Flynt v. Cal. Gambling Control Comm’n, 104 Cal.App.4th 1125, 129 Cal.Rptr.2d 167, 175-77 (2002). The State originally took the position that the revenue should be for general use, but abandoned that position during the negotiations in favor of tribal proposals. See Coyote Valley II, 331 F.3d at 1102-03, 1113. The tribes agreed to pay a portion of their revenues into two funds: the Revenue Sharing Trust Fund (RSTF) and the Special Distribution Fund (SDF). See id. at 1104-05. Monies paid into the RSTF are redistributed to tribes who choose not to, or are unable to, conduct their own gaming activities. Id. at 1105. Monies paid into the SDF, on the other hand, are used to fund:
(a) grants for programs designed to address gambling addiction; (b) grants for the support of state and local government agencies impacted by tribal 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 RSTF; and (e) “any other purposes specified by the legislature.”5
Id. at 1106.
In Coyote Valley II, appellants questioned whether the RSTF and SDF provisions of the 1999 compacts were lawful since IGRA, 25 U.S.C. § 2710(d)(4), precludes states from imposing taxes on Indian gaming. 331 F.3d 1094. We held *1024that the RSTF and SDF were permissible notwithstanding § 2710(d)(4) because, as more fully explained infra, the nature of the revenue sharing and the constitutional exclusivity obtained in consideration for it were primarily motivated by a desire to promote tribal interests. 351 F.3d at 1110-15. We further concluded that by virtue of the 1999 compacts and Proposition 1A, the State gave all tribes in California significant opportunities to benefit from gaming without taking anything significant for itself, beyond what was required to protect its citizens from the adverse consequences of gaming, and to fulfill other regulatory and police functions contemplated by IGRA. Id.
The 2003-2006 Compact Renegotiations
Operating under its 1999 compact, Rincon began to generate significant revenue that enabled it to improve tribal governmental functions and become economically self-sufficient. By 2003, Rincon desired to expand its operations beyond what the 1999 compact permitted. Accordingly, in March of that year, Rincon notified the State of its interest in renegotiating certain provisions of the 1999 compact.
Negotiations began in 2003 in response to Rincon’s request, but in October of that year, California voters recalled Governor Davis and elected Governor Schwarzenegger in his stead. Although negotiations eventually reconvened, they quickly assumed a decidedly different tone. Instead of requesting funds to help defray the costs of gaming, or to benefit Indian tribes, the State demanded that Rincon pay a significant portion of its gaming revenues into the State’s general fund.
The State made its first offer to Rincon on November 10, 2005.6 The State offered Rincon the opportunity to operate 900 additional devices plus the 1600 devices Rincon already operated, but only if Rincon would agree to pay the State 15% of the net win on the new devices, along with an additional 15% annual fee based on Rincon’s total 2004 net revenue. In exchange for the 15% revenue share demanded, the State offered Rincon an “exclusivity provision.” 7
*1025Rincon countered that, in order to obtain additional devices, it would agree to some per device fees. Rincon emphasized, however, that the use of any fees it paid had to be limited to paying for the costs of regulating gaming, building infrastructure needed to support gaming operations, and mitigating adverse impacts caused by gaming operations. Rincon further stated that “with all due respect, we are not asking for exclusivity and the State’s analysis does not hold water as it relates to Rincon in its current circumstance.”
Rincon also noted that Proposition 1A already provided for tribal gaming exclusivity, so it was not seeking whatever further exclusivity might provide. Rincon’s lands are located in the middle of a saturated tribal gaming market. Accordingly, no form of tribal exclusivity could shelter Rincon from substantial competition. As long as the proposed exclusivity provision related only to freedom from %o%-tribal competition, “exclusivity” would not provide Rincon with any meaningful economic advantages that would warrant the tribe making the requested payments.
The State interpreted Rincon’s counter-proposal for limited-use, per device fees and its rejection of exclusivity to be a request that the State agree to allow Rincon to operate additional devices beyond the 1999 compact limits “without offering the State anything meaningful in return.” The State held firm in its demand that a portion of tribal gaming revenues be paid into the State’s general fund, rather than into an earmarked fund.
Rincon re-countered with an offer substantially mirroring its previous offer, but offering slightly increased per device fees. Rincon also presented several expert reports on the financial impact the State’s offer would have on Rincon. By Rincon’s calculations,
the State’s offer ... would require Rincon to pay an additional $23 million in fees for the machines currently in play at Rincon’s gaming operation pursuant to the 1999 Compact.... By imposing the 15% fee on the Tribe’s net win as of Fiscal year 2004, the Tribe would be required to pay 15 to 20 times what it is paying now without adding a single machine onto the gaming floor! The State’s proposal is a poorly disguised tax, which is impermissible under IGRA.
The State made its next counteroffer on October 23, 2006. That offer included substantially the same terms as its November 10, 2005 offer, but offered that the compact term would be extended for five years, and that Rincon would pay an annual fee equal to 10% (instead of 15%) of its net win based on fiscal year 2005 (instead of 2004). The State noted that the terms it was offering Rincon were similar to those already accepted by a handful of other tribes and approved by the Department of the Interior. At Rincon’s request, on October 31, 2006, the State made an alternative offer to allow Rincon to operate 400 additional devices with no other changes to the existing compact. In exchange for the 400 additional devices, Rincon would have to pay $2 million annually to the RSTF, plus 25% of Rincon’s net win on those additional 400 devices to the State’s general fund.
The State accompanied this last counteroffer with its own expert analysis comparing the value to Rincon of continuing to operate its current 1600 devices under the 1999 compact to the value to Rincon of accepting the State’s counteroffer of 2500 devices with a 10% annual fee. The State’s expert concluded that if Rincon accepted the State’s offer, it would pay California $38 million and retain $61 million in net revenue. If Rincon maintained its operations under the 1999 compact, it would pay the State nothing and retain $59 million in net revenue. Hence, according to the State’s expert, Rincon stood to gain $2 million in additional revenues if it ac*1026cepted the amendment. In contrast, the State stood to gain $38 million. Rincon rejected the State’s counteroffer, and the record of negotiations then closed.
Having reached an impasse, the parties filed cross-motions for summary judgment in the district court. The district court granted summary judgment in favor of Rincon, and this timely appeal followed.
JURISDICTION AND STANDARD OF REVIEW
IGRA grants district courts original federal jurisdiction over tribal claims that a state has failed to negotiate in good faith concerning class III gaming rights. 25 U.S.C. § 2710(d)(7)(A)(i). California has waived its Eleventh Amendment immunity from such suits.8 Cal. Gov’t Code § 98005; Hotel Employees, 88 Cal.Rptr.2d 56, 981 P.2d at 1011. We have jurisdiction under 28 U.S.C. §§ 1291 and 1292(a)(1).
Summary judgment is appropriate if there is no genuine issue of material fact and, even making all reasonable inferences in favor of the nonmoving party, the moving party is entitled to judgment as a matter of law. Fed.R.Civ.P. 56(c); Bodett v. CoxCom, Inc., 366 F.3d 736, 742 (9th Cir.2004). The State argues that the district court erred in granting summary judgment against it on the issue of whether it negotiated in good faith. See 25 U.S.C. § 2710(d)(7)(B)(iii-iv). Whether the negotiations were conducted in good faith is a mixed question of law and fact that we review de novo. Coyote Valley II, 331 F.3d at 1107 (citing Diamond v. City of Taft, 215 F.3d 1052, 1055 (9th Cir.2000)).
DISCUSSION
From the advent of colonists in North America, the new arrivals promptly began encroaching on Indian lands, and frequently treating Indians unfairly. To protect against further “great Frauds and Abuses” perpetrated by the colonists against the Indians, and to avoid war, the *1027British Crown assumed ultimate authority-over Indian affairs. 1-1 Cohen’s Handbook on Fed. Indian Law § 1.02 (Matthew Bender 2009). When our nation was formed, the federal government essentially took the place of the Crown, with Congress being granted the power to “regulate Commerce ... with the Indian tribes,” U.S. Const, art. I, § 8, cl. 3, and the President being given the power to make treaties (including with Indian tribes) with the consent of the Senate. U.S. Const, art. II, § 2, cl. 2. According to the Supreme Court, the federal government’s relationship to the tribes was that of a “ward to his guardian.” Cherokee Nation v. Georgia, 30 U.S. (5 Pet.) 1, 17, 8 L.Ed. 25 (1831). Nevertheless, promises and treaties were repeatedly broken or ignored as Indians were swept from their lands and homes by states, hoards of settlers, and sometimes even by the “guardian” federal government itself, when they wanted the lands or resources possessed by those Indians. See Cohen’s, supra, at §§ 1.02-1.03. Recounting one such instance, the Supreme Court in United States v. Sioux Nation of Indians detailed the history of how, first by military force, then by Congressional act, the government deprived the Sioux tribe in South Dakota of much of its land because gold was discovered in the Black Hills. 448 U.S. 371, 100 S.Ct. 2716, 65 L.Ed.2d 844 (1980). Today, many tribes have struck figurative gold with casino gaming and, again, some state governments, just like their predecessors, are maneuvering to take, or at least share in, some of that figurative gold.
Mindful of this ignominious legacy, Congress enacted IGRA to provide a legal framework within which tribes could engage in gaming — an enterprise that holds out the hope of providing tribes with the economic prosperity that has so long eluded them grasp — while setting boundaries' to restrain aggression by powerful states. See S.Rep. No. 100-446, at 33 (1988) (statement of Sen. John McCain), reprinted in 1988 U.S.C.C.A.N. 3071, 3103; 134 Cong. Rec. at S12654 (statement of Sen. Evans). In passing IGRA, Congress assured tribes that the statute would always be construed in their best interests. See, e.g., S.Rep. No. 100-446, at 13-14, as repainted in 1988 U.S.C.C.A.N. at 3083-84.
Under IGRA, a tribe may conduct class III gaming only once a compact with its home state is in effect. Because the compact requirement skews the balance of power over gaming rights in favor of states by making tribes dependent on state cooperation, IGRA imposes on states the concomitant obligation to participate in the negotiations in good faith. 25 U.S.C. § 2710(d)(3)(A). If a court finds that a state has failed to negotiate in good faith, IGRA empowers the court to order additional negotiations and, if necessary, to order the parties into mediation in which a compact will be imposed. § 2710(d)(7).
In evaluating a State’s good faith, the district court:
(I) may take into account the public interest, public safety, criminality, financial integrity, and adverse economic impacts on existing gaming activities, and
(II) 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.
§ 2710(d)(7)(B)(iii) (emphasis added); see Coyote Valley II, 331 F.3d at 1108-09 (quoting at length S.Rep. No. 100-446, at 13-14, as reprinted in 1988 U.S.C.C.A.N. at 3083-84, which provides guidance on how Congress intended § 2710(d)(7)(B)(iii)(I) to be interpreted).
In addition to specifying criteria for evaluating a state’s good faith, IGRA outlines permissible tribe-state negotiation topics.
(C) Any Tribal-State compact ... may include provisions relating to—
*1028(i) the application of the criminal and civil laws and regulations of the Indian tribe or the State that are directly related to, and necessary for, the licensing and regulation of such activity;
(ii) the allocation of criminal and civil jurisdiction between the State and the Indian tribe necessary for the enforcement of such laws and regulations;
(iii) the assessment by the State of such activities in such amounts as are necessary to defray the costs of regulating such activity;
(iv) taxation by the Indian tribe of such activity in amounts comparable to amounts assessed by the State for comparable activities;
(v) remedies for breach of contract;
(vi) standards for the operation of such activity and maintenance of the gaming facility, including licensing; and
(vii) any other subjects that are directly related to the operation of gaming activities.
§ 2710(d)(3)(C). However, the list of permissible negotiation topics is circumscribed by one key limitation on state negotiating authority:
Except for any assessments that may be agreed to under [§ 2710(d)(3)(C)(iii) ], nothing in this section shall be interpreted as conferring upon a State ... authority to impose any tax, fee, charge, or other assessment upon an Indian tribe.... No State may refuse to enter into the negotiations ... based upon the lack of authority in such State ... to impose such a tax, fee, charge, or other assessment.
25 U.S.C. § 2710(d)(4). IGRA limits permissible subjects of negotiation9 in order *1029to ensure that tribal-state compacts cover only those topics that are related to gaming 10 and are consistent with IGRA’s stated purposes, see Coyote Valley II, 331 F.3d at 1111, which are:
(1) to provide a statutory basis for the operation of gaming by Indian tribes as a means of promoting tribal economic development, self-sufficiency, and strong tribal governments',
(2) to provide a statutory basis for the regulation of gaming by an Indian tribe adequate to shield it from organized crime and other corrupting influences, to ensure that the Indian tribe is the primary beneficiary of the gaming operation, and to assure that gaming is conducted fairly and honestly by both the operator and players; and
(3) to declare that the establishment of independent Federal regulatory authority for gaming on Indian lands, the establishment of Federal standards for gaming on Indian lands, and the establishment of a National Indian Gaming Commission are necessary to meet congressional concerns regarding gaming and to protect such gaming as a means of generating tribal revenue.
§ 2702 (emphasis added); see also Artichoke Joe’s, 353 F.3d at 715.
Here, the State repeatedly demanded that Rincon agree to pay into the State’s general fund 10-15% of Rincon’s annual net win, and up to 25% of Rincon’s revenue from any new devices Rincon would operate under an amended compact. Once Rincon proffered evidence suggesting that the State had acted in bad faith by attempting to impose taxation, “the burden of proof [shifted to] the State to prove that the State has negotiated with the Indian tribe in good faith to conclude a Tribal-State compact governing the conduct of gaming activities.” § 2710(d)(7)(B)(ii). We conclude that the State failed to meet its burden.
I. Taxation Demands “Shall” Be Considered Evidence of Bad Faith
Under § 2710(d)(7)(B)(iii)(II), a court must consider a “demand” for a tax to be made in bad faith.11 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 10% of its net profits into the State’s general fund. According to California Government 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 *1030into the State treasury for unrestricted use yields public revenue, and is a “tax.” Moreover, unlike what occurred in the 1999 negotiations, none of the State’s communications during the renegotiations that occurred after the change in administration in 2003 reflected a willingness to take its general fund revenue sharing demand off the table. The State repeatedly emphasized its position that it would not give Rincon more devices or time without a reciprocal benefit to the State, and the record reveals that no other “benefit” was demanded besides monetary payments into the general fund. Webster’s Third New International Dictionary 598 (2002) defines “demand” as “to call for as useful, necessary, or requisite: make imperative.” The State’s repeated and forceful insistence on monetary payments to the general fund undoubtedly constitutes a “demand.” Under the plain language of § 2710(d)(7)(B)(ii)(II), the State’s demand for the payment of a tax is evidence of the State’s bad faith.
Our dissenting colleague faults us for our characterization of the 10-15% revenue share as a “tax,” primarily because he contends we fail to appreciate the import of the word “imposed” in the definition of a “tax.” Dissent at 5934-37, 5943-45. He argues that IGRA merely creates a context for “voluntary” negotiations, and that no matter how “hard line” the State’s position is, it still has not attempted to exercise authority to “impose” a tax.
The flaw in this argument — related to a faulty assumption made throughout the dissent — is that it ignores the plain fact that neither tribes nor states enter IGRA negotiations “voluntarily” in the way parties do in all the examples cited by the dissent. IGRA negotiations are therefore distinguishable from regular contract negotiations. When private parties, or independent sovereign entities, commence contract negotiations, they generally do so because each has something of value the other wants, and each side has the right to accept or reject an offer made, based on the desirability of the terms. If negotiations fail, neither party has a right to complain. Not so in IGRA negotiations. In IGRA, Congress took from the tribes collectively whatever sovereign rights they might have had to engage in unregulated gaming activities, but imposed on the states the obligation to work with tribes to reach an agreement under the terms of IGRA permitting the tribes to engage in lawful class III gaming activities. If IGRA negotiations break down between a state and a tribe because the state does not come to the bargaining table in good faith, IGRA specifically provides that courts, and the Secretary of the Interior, can intervene to impose a gaming arrangement without the affected state’s approval. See § 2710(d)(7)(B)(iii-vii). Thus, while IGRA was designed to give states a voice in Indian gaming, it was not designed to give states complete power over tribal gaming such that each state can put the opportunity to operate casinos up for sale to the tribe willing to pay the highest price. See § 2702; see also Cheyenne River Sioux Tribe v. South Dakota, 3 F.3d 273, 281 (8th Cir.1993) (noting that IGRA imposes mandatory duties upon states and gives them incentives to negotiate, but that it also provides tribes with alternative routes to a compact if the states choose not to cooperate); Dalton v. Pataki, 5 N.Y.3d 243, 802 N.Y.S.2d 72, 835 N.E.2d 1180, 1189 (2005) (“IGRA confers a benefit on the state by allowing it to negotiate and to have some input into how class III gaming will be conducted.”).
The dissent claims that § 2710(d)(4) means only that IGRA should not be interpreted as “conferring” upon states the “authority to impose ” taxes and fees. Dissent at 1048. But nothing in IGRA can reasonably be construed as conferring on states the power to impose anything; all *1031the states are empowered to do is negotiate. The logic underlying the dissent is that there is no “imposition” when there is “negotiation.” But by that logic, not only may states demand revenue sharing like California has done here, they could take any “hard-line” stance, such as demanding that a tribe agree to waive its sovereign immunity from taxation, as a condition of obtaining more gaming devices. No one disputes that requiring a tribe to waive its sovereign immunity from state taxation in order to obtain a compact is clearly contrary to IGRA. And yet, if a tribe “agreed” to do so, the waiver of taxation immunity would be no less “negotiated” than the revenue sharing the dissent advocates here.
Exercising an authority to “impose” in the context of IGRA must therefore relate to something the state does during the negotiations process to extract an improper concession. In other words, the only conceivable way a state could “impose” something during negotiations is by insisting, over tribal objections, that the tribe make a given concession — a concession beyond those specially authorized by § 2710(d)(3)(C)12 and contrary to the tribe’s sovereign interests — in order to obtain a compact.13
The dissent acknowledges that in this case “California has insisted that the Band share its gaming revenues as a condition to receiving authorization for additional gaming devices,” but then concludes that this is simply “hard-line” negotiating for revenue sharing, not imposing a tax. Dissent at 5933. If there is a distinction between insisting on obtaining a share of Rincon’s income as a non-negotiable condition of granting it a compact, and demanding a tax or “refusing] to enter into the negotiations ... based upon the lack of authority ... to impose such ■ a tax, fee, charge, or other assessment,” § 2170(d)(4), it is a distinction without a difference. In either case, the state is using its power over negotiations to force Rincon to pay the State a portion of its income into the State’s general fund (and not for any use for the benefit of Rincon or other tribes) in order to engage in class III gaming. If § 2710(d)(4) means anything, it means that California cannot do that — whatever one calls it.14
*1032In Coyote Valley II, we explained that IGRA requires courts to consider a state’s demand for taxation as evidence of bad faith, not conclusive proof. 331 F.3d at 1112-13 (citing § 2710(d)(7)(B)(iii)(II)). However, “[depending on the nature of both the fees demanded and the concessions offered in return, such demands might, of course, amount to an attempt to impose a fee, and therefore amount to bad faith on the part of a State.” Coyote Valley II, 331 F.3d at 1112 (internal quotation marks omitted). For the reasons described in greater detail infra, 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. See § 2710(d)(3)(B)(iii)(II).
Under IGRA, the State may attempt to rebut bad faith by demonstrating that the revenue demanded was to be used for “the public interest, public safety, criminality, financial integrity, and adverse economic impacts on existing gaming activities.” § 2710(d)(3)(B)(ii). See S.Rep. No. 100-446, at 13-14, as reprinted in 1988 U.S.C.C.A.N. at 3083-84. The State’s need for general tax revenues is not in the list. Even if “the public interest” or “financial integrity” could conceivably be construed to implicate the State’s need for general funds, IGRA’s purposes do not permit such a construction. Instead, those terms clearly apply to protecting the State against the adverse consequences of gaming activities. See § 2702; S.Rep. No. 100-446, at 13-14, as reprinted in 1988 U.S.C.C.A.N. at 3083-84. Moreover, construing those terms broadly in favor of the State’s interests would be inconsistent with our obligation to construe IGRA most favorably towards tribal interests. See Artichoke Joe’s, 353 F.3d at 728-30 (discussing Montana v. Blackfeet Tribe of Indians, 471 U.S. 759, 105 S.Ct. 2399, 85 L.Ed.2d 753 (1985)); see also S.Rep. No. 100-446, at 15, as reprinted in 1988 U.S.C.C.A.N. at 3085 (stating Congressional intent that courts interpret any ambiguities regarding § 2710(d)(3)(B)(ii)(I) in the way most favorable to tribal interests).
Critically, the State does not even seek to justify its general fund revenue sharing demands directly under any of the factors in § 2710(d)(7)(B)(ii)(I).15 Rather, the State relies on its interpretation of our decision in Coyote Valley II. The State’s reliance is misplaced. Coyote Valley II is an exceptional case whose facts are readily distinguishable from those in this case.
II. Coyote Valley II
Coyote Valley II considered objections to, among other things, the RSTF and SDF provisions of the 1999 compacts. We held those funds to be authorized subjects of negotiation under 25 U.S.C. § 2710(d)(3)(C)(vii) (subjects “directly related to the operation of gaming”). The SDF was clearly “directly related” to gaming because all uses of SDF funds were earmarked for gaming-related purposes. Coyote Valley II, 331 F.3d at 1114. The RSTF funds similarly were related to gaming because, by redistributing gaming funds from gaming to non-gaming tribes, they are entirely consistent with the IGRA goal of using gaming to foster tribal economic development. Id. at 1111. Notably, we expressly declined to decide if the RSTF or SDF were “taxes,” because they were decidedly not “imposed” in bad faith. *1033Rather, the tribes themselves suggested them, and were willing to pay into them in exchange for the “meaningful concession” of constitutional exclusivity. Id. at 1112— 15.
Coyote Valley II thus stands for the proposition that a state may, without acting in bad faith, request revenue sharing if the revenue sharing provision is (a) for uses “directly related to the operation of gaming activities” in § 2710(d)(3)(C)(vii), (b) consistent with the purposes of IGRA, and (c) not “imposed” because it is bargained for in exchange for a “meaningful concession.” The State’s offers in this case fail on all three prongs of that proposition.
A. “Directly Related to the Operation of Gaming Activities”
The State asserts that, like the RSTF and SDF, its revenue sharing demands were authorized under § 2710(d)(3)(C)(vii) because they involve a subject directly related to gaming. The State misunderstands.
The State’s argument that general fund revenue sharing is “directly related to the operation of gaming activities” because the money is paid out of the income from gaming activities is circular. Moreover, the very next section of the statute precludes us from interpreting § 2710(d) (3)(C) (vii) in the way the State suggests. See § 2710(d)(4) (stating explicitly that states are not authorized to use negotiations to impose assessments on tribes other than those “agreed to under paragraph (3)(C)(iii) ” (emphasis added)); see also Wisconsin v. Ho-Chunk Nation, 512 F.3d 921, 932 (7th Cir.2008) (describing revenue sharing agreements as being in tension with § 2710(d)(4)).
Crucially, in Coyote Valley II we did not conclude that § 2710(d)(3)(C)(vii) authorized the RSTF and SDF because “revenue sharing” is a subject directly related to gaming. Rather, we held that fair distribution of gaming opportunities and compensation for the negative externalities caused by gaming are subjects directly related to gaming, and the RSTF and SDF were the means chosen by the parties to the 1999 compacts to deal with those issues. See Coyote Valley II, 331 F.3d at 1111, 1114.
Whether revenue sharing is an authorized negotiation topic under § 2710(d)(3)(C)(vii) thus depends on the use to which the revenue will be put, not on the mere fact that the revenue derives from gaming activities. General fund revenue sharing, unlike funds paid into the RSTF and SDF, has undefined potential uses. See Cal. Gov.Code § 16300 (providing that the “General Fund consists of money received into the treasury and not required by law to be credited to any other fund.”). Therefore, payments into the general fund cannot be said to be directly related to gaming. Indeed, in Coyote Valley II we expressly recognized the distinction between general fund revenue sharing and the RSTF and SDF.16 We noted that *1034the RSTF “provision does not put tribal money into the pocket of the State,” id. at 1113, and reserved the question of whether the SDF would be lawful if the funds were deposited straight into the State’s general fund, id. at 1114 n. 17. Consequently, we hold that general fund revenue sharing is not “directly related to the operation of gaming activities” and is thus not an authorized subject of negotiation under § 2710(d)(3)(C)(vii). See Cabazon Band of Mission Indians v. Wilson, 37 F.3d 430, 435 (9th Cir.1994) (holding that where fees go to the state’s general fund, the relationship between the revenue payments and the costs incurred in regulating gaming activities is attenuated).
Ruling out § 2710(d)(3)(C)(vii) (authorizing negotiations over subjects directly related to gaming not otherwise listed in § 2710(d)(3)(C)) makes the applicability of § 2710(d)(4) (withholding authority to impose taxes or fees other than those permitted under § 2710(d)(3)(C)(iii)) even more apparent. The State was admittedly seeking “annual fees” and objected to Rincon’s suggestion that any fees be limited to § 2710(d)(3)(C)(iii) uses. That, combined with the general notion that IGRA negotiations are supposed to be limited to gaming regulation, convinces us that there is no statutory basis for authorizing tribe-state negotiations over general fund revenue sharing. See Ho-Chunk Nation, 512 F.3d at 932-33 (declining to decide whether general fund revenue sharing is invalid, but noting that it was apparently not a subject “contemplated by Congress as being one of the matters tribes and the states may negotiate over under the IGRA”).
B. Consistent with the Purposes of IGRA
According to § 2702, IGRA is intended to promote tribal development, prevent criminal activity related to gaming, and ensure that gaming activities are conducted fairly. In Coyote Valley II we construed the meaning of subjects “directly related to the operation of gaming” in § 2710(d)(3)(C)(vii) broadly to include revenue sharing because the RSTF is consistent with the plain language of § 2702 (listing tribal economic self-sufficiency as one of IGRA’s purposes). See Coyote Valley II, 331 F.3d at 1111. By contrast, we cannot read § 2710(d)(3)(C)(vii) broadly here to include general fund revenue sharing because none of the purposes outlined in § 2702 includes the State’s general economic interests. The only state interests mentioned in § 2702 are protecting against organized crime and ensuring that gaming is conducted fairly and honestly. § 2702(2); see also S.Rep. No. 100-446, at 2, 4, as reprinted in 1988 U.S.C.C.A.N. at 3072-73, 3075.
Because the plain language of § 2702 does not support the State’s position, the State misconstrues certain statements in Coyote Valley II to say that the State’s pursuit of its economic interests is at least not inconsistent with IGRA. As an initial matter, we are reluctant to inject into the statute a purpose not codified within it. See Exxon Mobil Corp. v. Allapattah Servs., Inc., 545 U.S. 546, 568, 125 S.Ct. 2611, 162 L.Ed.2d 502 (2005). But we need not digress into that potentially complicated statutory analysis because the State clearly misinterprets Coyote Valley II.
The State first points to our recognition in Coyote Valley II that Congress acknowledged as legitimate the State’s “economic interest in raising revenue for its citizens.” 331 F.3d at 1115 (quoting S. Rep. NO. 100-446, at 13, as reprinted in 1988 U.S.C.C.A.N. at 3083). The State takes *1035this quotation out of context. The full quotation is:
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.
Coyote Valley II, 331 F.3d at 1115 (quoting S.Rep. No. 100-446, at 13, as reprinted in 1988 U.S.C.C.A.N. at 3083).
Although a grammatical analysis of the sentence could conceivably suggest that a state’s “economic interests in raising revenue for its citizens” is one of the “impacts on the State’s regulatory system,” such a construction would represent a distortion of the text. A more common sense approach to interpreting Congress’s meaning here, as taken in Coyote Valley II, is to read a state’s “interest in raising revenue” as being “included” as one of the “State’s other interests.” 331 F.3d at 1115. And in context, the “other interests” are the states’s own gaming systems like the California Lottery that would, after IGRA, have to compete with Indian gaming. IGRA itself, as well as clearer statements in the legislative history, support this interpretation. See § 2710(d)(7)(B)(iii)(I) (listing the state’s interest concerning “adverse economic impacts on existing gaming activities” among the factors relevant to a state’s good faith); see also, e.g., S.Rep. No. 100-446, at 1-2, 14, as reprinted, in 1988 U.S.C.C.A.N. at 3071-72 (acknowledging the problem that States and the gaming industry may attempt to use the compact requirement to impede tribal competition). Thus, neither the statute nor the legislative history support interpreting the phrase “other interests ... including its economic interest in raising revenue for its citizens” to mean that states may tax tribes. In fact, all indications are to the contrary. See § 2710(d)(4); supra n. 10.
The State next directs us to a similar statement from Coyote Valley II that “Congress ... did not intend to require that States ignore their economic interests when engaged in compact negotiations.” Coyote Valley II, 331 F.3d at 1115. The State’s reliance on this statement is likewise misplaced. When we said that Congress did not intend for states to ignore their economic interests, we were not deciding whether states were allowed to pursue their own economic objectives affirmatively through compact negotiations. Rather, we decided only whether the State could require the tribes to pay into the SDF to cover the government’s costs of dealing with the fallout of gaming. It is one thing to ask the tribes to contribute funds so the State is not left bearing the costs for gaming-related expenses; it is quite another to ask the tribes to help fix the State’s budget crisis. The State is therefore incorrect that pursuit of state general economic interests is consistent with IGRA’s purposes.
As already explained, IGRA’s stated purposes include ensuring that tribes are the primary beneficiaries of gaming and ensuring that gaming is protected as a means of generating tribal revenue. § 2702. We therefore 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. In such case, it is the State, not the tribe, that would be the “primary beneficiary” of the gaming rights under negotiation. See *1036Cabazon Band of Mission Indians, 37 F.3d at 433 (explaining that where the state benefits “from the tribal gaming operation to a considerably greater extent than the [tribe, the tribe would not] be described as a ‘primary benefieiary[,’ and s]uch an outcome contravenes the purposes of IGRA”).
C. “Meaningful Concessions”
Because we hold above that general fund revenue sharing is neither authorized by IGRA nor reconcilable with its purposes, it is difficult to imagine what concessions the State could offer to rebut the strong suggestion of bad faith arising from such demands. But even if it were possible to conjure up an exceptional circumstance where such would be the case, where, as here, the State demands significant taxes and fails to offer any “meaningful concessions” in return, a finding of bad faith is the only reasonable conclusion.
The relevance of “meaningful concessions” arises from § 2710(d)(4). We have interpreted § 2710(d)(4) as precluding state authority to impose taxes, fees, or assessments, but not prohibiting states from negotiating for such payments where “meaningful concessions” are offered in return. See Coyote Valley II, 331 F.3d at 1112; see also Idaho v. Shoshone-Bannock Tribes, 465 F.3d 1095, 1101 (9th Cir.2006). In other words:
[t]he theory on which [revenue sharing] payments were allowed [in Coyote Valley II] ... was that the parties negotiated a bargain permitting such payments in return for meaningful concessions from the state (such as a conferred monopoly or other benefits). 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.
Shoshone-Bannock, 465 F.3d at 1101 (internal citation omitted).
Importantly, we emphasized in Coyote Valley II that we were not holding that “the State could have, without offering anything in return, taken the position that it would conclude a Tribal-State compact with [the tribe] only if the tribe agreed to pay into the RSTF.” 331 F.3d at 1112. But, “[w]here ... 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.” Id. With this concept in mind, the State argues that it offered “meaningful concessions” and therefore merely exercised its authority to negotiate within the permissible bounds of IGRA.
The State’s analogy fails. Unlike in Coyote Valley II, in which the tribes proposed the revenue sharing provisions, see 331 F.3d at 1113, Rincon did not suggest revenue sharing; indeed, Rincon has consistently objected to it. Additionally, the State has not offered any “meaningful concessions.”
Just how “meaningful” the exclusivity provision at issue in Coyote Valley II was at the time of the 1999 compacts cannot be overstated. In 1999, the California constitution prohibited casino-style gaming, and the State was therefore under no obligation to allow tribes to conduct it, or even negotiate concerning it. Rumsey Indian Rancheria of Wintun Indians v. Wilson, 64 F.3d 1250 (9th Cir.1994). The State nonetheless negotiated the 1999 compacts with dozens of tribes, and to make the 1999 compacts fully operable, the State promoted a constitutional amendment exempting tribes, and tribes alone, from the constitutional prohibition. Flynt, 129 Cal.Rptr.2d at 175-77; see also Artichoke Joe’s, 353 F.3d at 718.
*1037The value of a monopoly is obvious, and the value of a monopoly that cannot be altered except by the extraordinary act of further constitutional amendment is even greater. Such a benefit was well beyond anything IGRA required the State to offer. See Coyote Valley II, 331 F.3d at 1111, 1115. Specifically, IGRA provides that tribes can engage in class III gaming to the same extent as others in the state. § 2710(d)(1). Thus, IGRA only requires that states treat tribes equally. However, with the strong encouragement of the then governor, California voters gave the tribes an economic opportunity denied to everyone else. The State’s agreement (with the consent of the voters) to confer such a substantial benefit on the tribes proved that its request that more successful tribes financially assist the less fortunate ones (and that the tribes agree to cover the costs of adverse impacts) was freely negotiated. Indeed, in a rare example of generosity to tribes, the State conferred a valuable economic right on the tribes in exchange for a program under which all of the significant benefits of the compact were to be enjoyed by the tribes themselves.
In short, we approved exclusivity as a “meaningful concession” in Coyote Valley II because it was exceptionally valuable and bargained for. By contrast, in the current legal landscape, “exclusivity” is not a new consideration the State can offer in negotiations because the tribe already fully enjoys that right as a matter of state constitutional law. Moreover, the benefits conferred by Proposition 1A have already been used as consideration for the establishment of the RSTF and SDF in the 1999 compact. See Flynt, 129 Cal.Rptr.2d at 177. “It is elementary law that giving a party something to which he already has an absolute right is not consideration to support that party’s contractual promise.” Salmeron v. United States, 724 F.2d 1357, 1362 (9th Cir.1983). The State asserts that it would be unfair to permit Rincon to keep the benefit of exclusivity conferred by Proposition 1A without holding the tribe to an ongoing obligation to periodically acquiesce in some new revenue sharing demand. While we do not hold that no future revenue sharing is permissible, it is clear that the State cannot use exclusivity as new consideration for new types of revenue sharing since it and the collective tribes already struck a bargain in 1999, wherein the tribes were exempted from the prohibition on gaming in exchange for their contributions to the RSTF and SDF. Flynt, 129 Cal.Rptr.2d at 177.17
The State offers various alternative arguments to support its claim that it offered more than illusory consideration. We find all of the State’s arguments unpersuasive.
1. Revised and Expanded Exclusivity
The State first claims that it offered Rincon revised and expanded exclusivity, which had even greater economic value than the exclusivity originally granted by Proposition 1A.
We first note that the State never defined the precise contours of the exclusivity provision it proposed, giving us very little upon which to base a finding that the State has met its burden to show it offered a real, meaningful concession in the form *1038of a new and improved exclusivity provision. However, we assume, as apparently the parties have done, that the revised exclusivity provision would have been similar to the one recently accepted by several other tribes, including the Pala Band of Mission Indians. The Pala Band’s exclusivity provision provides that (1) “the State shall not authorize any person or entity other than an Indian tribe ... to engage in any Gaming Activities ... within the Tribe’s core geographic market” (San Diego, Riverside, Orange, and Los Angeles Counties); and (2) if the State were to breach its obligation to ensure geographic exclusivity, “the Tribe shall have the right to enjoin such gaming” and “the right to cease the payments” due to the State. Such a revised and expanded exclusivity would be practically worthless to Rincon.
Since the passage of a constitutional amendment eliminating tribal gaming exclusivity is extremely unlikely, neither the injunctive relief18 nor the monetary19 remedies contingent upon that event have anything more than speculative value to Rincon. In addition, Rincon did not request or desire a revised local tribal gaming exclusivity provision because such a provision — which would not apply to other tribes — would not protect Rincon against the high degree of tribal competition it experiences in its core geographic market. Freedom from nontribal competition in its core geographic market therefore provides Rincon with no significant additional economic advantages over whatever value Rincon receives from the statewide exclusivity it already enjoys.
More importantly, even if there were some enhanced value in the proposed revised and expanded exclusivity provision, the 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, rather than negotiate for, a fee. See Coyote Valley II, 331 F.3d at 1112. Under IGRA and Coyote Valley II, that is bad faith.
Our conclusion is buttressed by the fact that the State never wavered from its general fund revenue sharing demands. We do not mean to suggest that the State is guilty of bad faith whenever it takes a “hard line” negotiating position. Indeed, in Coyote Valley II we upheld the State’s ability to insist on certain provisions. See Coyote Valley, 331 F.3d at 1116 (finding *1039that it was not bad faith for the State to insist on a particular labor standards provision). As already suggested supra, a “hard line” stance is not inappropriate so long as the conditions insisted upon are related to legitimate state interests regarding gaming and the purposes of IGRA. See § 2710(d)(3)(B)(iii)(I). We hold only that a state may not take a “hard line” position in IGRA negotiations when it results in a “take it or leave it offer” to the tribe to either accept nonbeneficial provisions outside the permissible scope of §§ 2710(d)(3)(C) and 2710(d)(1), or go without a compact. Cf. NLRB v. Ins. Agents’ Int’l Union, 361 U.S. 477, 485, 80 S.Ct. 419, 4 L.Ed.2d 454 (1960).
2. More Devices and Time
The State’s next argument is that, at the very least, it is entitled to some new consideration in exchange for giving Rincon expanded gaming rights, and the increased revenue share it requested was the only possible new consideration it could seek.
The State is correct that general contract principles dictate that new or additional consideration for a compact amendment is required. However, as already explained, IGRA does not permit the State and the tribe to negotiate over any subjects they desire; rather, IGRA anticipates a very specific exchange of rights and obligations, defined in §§ 2710(d)(3)(C) and (d)(4). Cf, e.g., Cal. Const., art. XV, § 1 (limiting the permissible scope of loan contract negotiations by prohibiting, subject to penalty, negotiation for usurious interest rates).
As held above, general fund revenue sharing is not a state public policy interest directly related to gaming and is not an authorized negotiation topic under § 2710(d)(3)(C). Therefore, gaming rights that tribes are entitled to negotiate for under IGRA, like device licensing and time, see § 2701(d)(3)(C)(vi),20 cannot serve as consideration for general fund revenue sharing; the consideration must be for something “separate” than basic gaming rights. See Dissent at 5957 (quoting Courtney J.A. DaCosta, Note, When “Turnabout” Is Not “Fair Play”: Tribal Immunity under the Indian Gaming Regulatory Act, 97 Geo. L.J. 515, 543-44 (2009)). In order to obtain additional time and gaming devices, Rincon may have to submit, for instance, to greater State regulation of its facilities or greater payments to defray the costs the State will incur in regulating a larger facility. See 25 U.S.C. § 2710(d)(3)(C)(i, iii). Rincon need not, however, submit to demands that it assist the State in addressing its budget crisis.
We are further influenced by the fact that the Department of the Interior, the executive agency charged with approving gaming compacts, also interprets IGRA in this way. As stated by the Assistant Secretary of Indian Affairs,
It is the position of the Department to permit revenue-sharing payments in exchange for quantifiable economic benefits over which the State is not required to negotiate under IGRA, such as substantial exclusive rights to engage in Class III gaming activities. We 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.21
*1040To hold otherwise would effectively mean that states could put gaming rights “up for sale.”22 That would be inconsistent with IGRA’s spirit, and its express refusal to allow states to use their right to engage in compact negotiations as a means to extract fees. § 2710(d)(4).
3. “Exclusive Gaming Rights ”
Next, the State argues that the value of its offers during compact negotiations should be analyzed as a whole, not piecemeal. Specifically, the State contends that we should evaluate its offer not simply by considering the value of the exclusivity provision itself, but rather the value of the entire bundle of “exclusive gaming rights” that Rincon would obtain under an amended compact. Viewing its offer in this way, the State contends it negotiated in good faith.
If there is a meaningful distinction between “exclusivity” and “gaming rights” as stand alone terms and “exclusive gaming rights,” it is too minuscule to see. Because exclusivity exists independent of the current compact negotiations, the negotiations concern only the extent of gaming rights, which are, by nature as a result of California constitutional law, exclusive. Were we to accept the State’s view that whenever it negotiates with a tribe, it offers “meaningful concessions” because the gaming rights offered will, as a matter of law, be exclusive, we would effectively be holding that, as a matter of law, the State is entitled to insist on significant general fund revenue sharing whenever a tribe wants to renegotiate basic terms. We reject the State’s view. As explained in Part II.C.2, supra, IGRA entitles tribes to negotiate for basic class III gaming rights without being forced to accept revenue sharing.
Further, we disagree that the State makes “meaningful concessions” whenever it offers a bundle of rights more valuable than the status quo.23 As previously explained, IGRA endows states with limited negotiating authority over specific items. Accepting the State’s “holistic” view of negotiations would permit states to lump together proposals for taxation, land use restrictions, and other subjects along with IGRA class III gaming rights. Such a construction of IGRA would violate the purposes and spirit of that law.24 See supra n. 10.
*1041III. Other Evidence of Good Faith
The State raises one final argument in support of its position. Specifically, the State contends that it genuinely believed its revenue sharing demands were authorized by Coyote Valley II, approved by the Department of the Interior, and fair because other tribes had accepted them. The State therefore urges us to find that its demands, even though herein held improper under IGRA, were nonetheless made in good faith.
IGRA does not provide express guidance about whether good faith is to be evaluated objectively or subjectively. However, we are influenced by the factors outlined in § 2710(d)(7)(B)(iii), which lend themselves to objective analysis and make no mention of unreasonable beliefs. Further, the structure and content of § 2710(d) make clear that the function of the good faith requirement and judicial remedy is to permit the tribe to process gaming arrangements on an expedited basis, not to embroil the parties in litigation over their subjective motivations. We therefore hold that good faith should be evaluated objectively based on the record of negotiations, and that a state’s subjective belief in the legality of its requests is not sufficient to rebut the inference of bad faith created by objectively improper demands.25 See Mashantucket Pequot Tribe v. Connecticut, 913 F.2d 1024, 1033 (2d Cir.1990) (“The statutory terms are clear, and provide no exception for sincere but erroneous legal analyses.”).
Here, the State’s belief that IGRA permitted the revenue sharing it sought was objectively unreasonable. IGRA expressly condemns state attempts to compel fees for purposes other than those specified in § 2710(d)(3)(C)(iii). The State does not even attempt to fit its general fund revenue sharing demand within § 2710(d)(3)(C)(iii). While Coyote Valley II held that § 2710(d)(3)(C)(vii) might also authorize revenue sharing, that case involved exceptional circumstances and turned on the specialized, limited uses of the revenue. As explained supra, the rationale for permitting revenue sharing under 2710(d)(3)(C)(vii) in Coyote Valley II is not present in this case, and the State was unreasonable to rely on Coyote Valley II for propositions not decided, or expressly reserved. See supra pp. 1032-33. Further, the Department of the Interior has frequently noted its concerns about the legitimacy of general fund revenue sharing. See, e.g., supra nn. 21 & 22 and accompanying text; New Mexico v. Pueblo of Pojoaque, 30 Fed.Appx. 768, 768 (10th Cir.2002) (order); Mescalero Apache Tribe v. New Mexico, 131 F.3d 1379, 1382 n. 2 (10th Cir.1997); Pueblo of Sandia v. Babbitt, 47 F.Supp.2d 49, 51 (D.D.C.1999). The Department of the Interior has approved compacts with general fund revenue sharing provisions agreed to by other California tribes, but has done so reluctantly, and only after the tribes themselves confirmed the desirability of the amendments.26 Often, the Secretary simply per*1042mits compacts with revenue sharing provisions to go into effect “only to the extent they are consistent with IGRA,” thus leaving open the precise question at issue. The State therefore could not reasonably have relied on the Department of the Interior’s approval of certain other compacts as proof that its demands to Rincon were lawful. This is especially true since IGRA anticipates that, even if some tribes agree to a waiver of their rights not to be taxed by a state, such a waiver cannot be a basis for the State expecting the same from Rincon. See Shoshone-Bannock, 465 F.3d at 1101-02; S.Rep. No. 100-446, at 5, as reprinted in 1988 U.S.C.C.A.N. at 3075-76 (explaining that “it is the Committee’s intention that to the extent tribal governments elect to relinquish rights in a tribal-State compact that they might have otherwise reserved, the relinquishment of such rights shall be specific to the tribe so making the election, and shall not be construed to extend to other tribes”).
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. See § 2710(d)(3)(B)(iii); § 2710(d)(4). None of the State’s arguments suffices to rebut the inference of bad faith such an improper demand creates.
In so holding, we are mindful that many states, and especially California, are currently writhing in the financial maw created by the clash of certain mandatory state expenditures at a time when state revenues have plummeted from historic levels. However, we are also keenly aware of our nation’s too-frequent breach of its trust obligations to Native Americans when some of its politically and economically powerful citizens and states have lusted after what little the Native Americans have possessed. In developing IGRA, Congress anticipated that states might abuse their authority over compact negotiations to force tribes to accept burdens on their sovereignty in order to obtain gaming opportunities. See, e.g., 134 Cong. Rec. S12643-01, at S12651 (1988) (statement of Sen. Evans); see also S.Rep. No. 100-446, at 33 (statement of Sen. John McCain), reprinted in 1988 U.S.C.C.A.N. at 3103. That is why the good faith requirement exists, and why IGRA condemns state taxation demands.
CONCLUSION
We AFFIRM the district court’s finding that the State of California negotiated with Rincon in bad faith by conditioning its agreement to expand Rincon’s class III gaming rights on Rincon’s agreement to pay a percentage of its revenues to the State’s general fund. The district court’s order compelling the parties to reach a compact or submit their best offers to a mediator pursuant to § 2710(d)(7) is effective forthwith.
AFFIRMED.
. Because we affirm the district court's finding of bad faith on the issue of the State’s demands for revenue sharing, we do not reach any of the alternative grounds for a finding of bad faith asserted by Rincon on cross-appeal.
. IGRA divides gaming into three classes. Class I involves traditional tribal gaming; class II involves bingo and non-banked card games; class III involves gambling not covered in class I or class II (such as casino-style gaming, including slot machines and banked card games). 25 U.S.C. § 2703.
. The compact included several terms that are irrelevant to the issues addressed in this opinion. For context, however, we note that under the 1999 compact, Rincon was permitted to operate a certain number of devices, plus draw additional device licenses from a limited statewide pool. The 1999 compact thus reflected a system of limited gaming.
. For a more detailed history of Proposition 1A and related state/tribe negotiations concerning the 1999 compacts, see In re Indian Gaming Related Cases, 331 F.3d 1094 (9th Cir.2003) and Artichoke Joe’s Cal. Grand Casino v. Norton, 353 F.3d 712 (9th Cir.2003).
. In this opinion, we focus only on subsections (a), (b) and (e) of the SDF. Subsection (c) is expressly authorized by § 2710(d)(3)(C)(iii), and the State does not rely upon it in its quest because it seeks to deposit funds into its general fund, not one with earmarked uses. Subsection (d) is effectively part of the RSTF so it need not be analyzed separately. We have previously construed subsection (e) to cover only those purposes directly related to gaming. Coyote Valley II, 331 F.3d at 1113-14.
. In June 2004, Rincon filed the present suit in order to force the State to expedite negotiations. The offers described herein were both IGRA negotiations and on-the-record settlement negotiations.
. Specifically, the State offered:
1. The State would agree to allow the Tribe to operate an additional 900 Gaming Devices outside of the licensing pool established by the Tribe's existing compact as long as the total number of Gaming Devices in operation by the Tribe do [sic] not exceed 2500 Gaming Devices[;]
2. The Tribe would be required to maintain its existing Gaming Device licenses, but the parties would negotiate over the amount of the contributions made by the Tribe to the [RSTF] in connection therewith;
3. The Tribe would pay annually to the State 15% of the average net win for each of the additional Gaming Devices outside of the licensing system that it operates pursuant to the compact amendment, provided that the average net win is calculated on the basis of all Gaming Devices operated by the Tribe;
4. The Tribe would pay to the State, for the duration of the compact term, an annual fee equal to 15% of the net win in Fiscal Year 2004 from the Gaming Devices in operation at the Tribe’s casino;
5. The term of the amended compact would be the same as that of the existing compact;
6. A portion of the Tribe's payment to the State could be designed for San Diego County and CalTrans, which amount would be negotiated between the Tribe and the State ... [;]
7. Except as set forth in paragraphs 5 and
8. the amendment would contain the same non-economic provisions as the Pala Compact Amendment;
8. The Tribe [would] be afforded an exclusivity provision, the terms of which [would] be subject to further negotiation ... [but] the exclusivity provisions would be ''similar” to the Pala compact amendment....
. Many states have not waived their Eleventh Amendment immunity under IGRA as California has. The dissent's reliance on the prevalence of compacts containing revenue sharing provisions is therefore suspect because that reliance ignores the fact that many of the states involved in those compacts would not permit tribes to challenge state demands as made in bad faith. See, e.g., Seminole Tribe of Florida v. Florida, 517 U.S. 44, 116 S.Ct. 1114, 134 L.Ed.2d 252 (1996) (holding that IGRA did not abrogate state Eleventh Amendment immunity, so Florida state actors could not be sued by tribe to force good faith negotiations); Mescalero Apache Tribe v. New Mexico, 131 F.3d 1379, 1384-85 (10th Cir.1997) (holding that New Mexico has not waived its Eleventh Amendment immunity to IGRA suits); Ponca Tribe of Okla. v. Oklahoma, 89 F.3d 690 (10th Cir.1996) (same for Oklahoma); Santee Sioux Tribe of Neb. v. Nebraska, 121 F.3d 427, 431 (8th Cir.1997) (same for Nebraska); Sault Ste. Marie Tribe of Chippewa Indians v. Michigan, 800 F.Supp. 1484 (W.D.Mich.1992) (same for Michigan). Tribes in states that have not waived their Eleventh Amendment immunity for IGRA suits have no recourse to challenge the validity of revenue sharing, and some therefore choose to accept revenue sharing rather than go without a compact. See Pueblo of Sandia v. Babbitt, 47 F.Supp.2d 49, 51, 56-57 & n. 7 (D.D.C.1999) (explaining that the Department of the Interior believed the revenue sharing provision was illegal, but also believed that it had no choice but to allow the compact to go into effect because the tribe would have no recourse against the state to obtain a legal compact). Moreover, this reality — for better or worse — will prevent the proliferation of lawsuits feared by our dissenting colleague. The dissent suggests that 25 C.F.R. §§ 291.1 et seq. is a potential vehicle for tribes to challenge state demands, Dissent at 1073. n. 27, but the only circuit court to consider the question has held the regulations invalid. Texas v. United States, 497 F.3d 491 (5th Cir.2007), cert. denied sub nom. Kickapoo Traditional Tribe of Tex. v. Texas, - U.S. -, 129 S.Ct. 32, 172 L.Ed.2d 18 (2008). The validity of the regulations is not before us, and we therefore do not find it appropriate to rely on them, or express any opinion as to their validity.
. Our dissenting colleague suggests that § 2710(d)(3)(C) is not exhaustive, and then goes on to say that even if it is, reading the "catch-all” provision, § 2710(d)(3)(C)(vii), restrictively conflicts with Coyote Valley II. Dissent at 1059-62. The language and structure of § 2710(d)(3)(C) suggests it is exhaustive. There are seven categories of what "may” be negotiated. Although "may” indicates permissiveness as the dissent explains, to grant permission is not necessarily to grant carte blanche. What is “permitted” is limited. Section 2710(d)(C)(vii) explicitly addresses unenumerated topics, but limits them to those “directly related” to gaming. See Wisconsin v. Ho-Chunk Nation, 512 F.3d 921, 933-34 (7th Cir.2008) (reading § 2710(d)(3)(C) as exhaustive for jurisdictional purposes); see also Coyote Valley II, 331 F.3d at 111 (noting that Congress "limit[ed1 the proper topics for compact negotiations to those that bear a direct relationship to the operation of gaming activities” (emphasis added)). Significantly, what compels a limited reading of the permitted topics is the canon of construction obligating us to construe a statute abrogating tribal rights narrowly and most favorably towards tribal interests. See United States v. Winans, 198 U.S. 371, 381, 25 S.Ct. 662, 49 L.Ed. 1089 (1905) (stating that "the treaty was not a grant of rights to the Indians, but a grant of right from them, — a reservation of those [rights] not granted”); Bryan v. Itasca County, 426 U.S. 373, 392, 96 S.Ct. 2102, 48 L.Ed.2d 710 (1976) ("[I]n construing this admittedly ambiguous statute, we must be guided by that eminently sound and vital canon that statutes passed for the benefit of dependent Indian tribes are to be liberally construed, doubtful expressions being resolved in favor of the Indians.” (internal quotation marks, citations, and ellipsis omitted)). The dissent is correct that Coyote Valley II held that § 2710(d)(3)(C)(vii) is not ambiguous, and we do not hold otherwise. Whether revenue sharing fits into the unambiguous phrase "directly related to gaming” however is a subject of significant dispute between the parties. To help resolve the dispute, we, like the Coyote Valley II court, consider relevant the Congressional directive to construe ambiguities related to the issues covered by IGRA most favorably to tribal interests. Coyote Valley II, 331 F.3d at 1111 (quoting S.Rep. No. 100446, at 15, reprinted in 1988 U.S.C.C.A.N. at 3085). Even if the dissent is correct that resolution of this issue features a clash of "dueling canons,” Congress’ specific invocation of the tribal canon persuades us which canon must triumph here. *1029persuades us which canon must triumph here.
. "Gaming by its very nature is a unique form of economic enterprise and the Committee is strongly opposed to the application of the jurisdictional elections authorized by this bill to any other economic or regulatory issue that may arise between tribes and States in the future.” S.Rep. No. 100-446, at 14, as reprinted in 1988 U.S.C.C.A.N. 3071, 3084. See also 134 Cong. Rec. S12643-01, at S12651 (1988) ("There is no intent on the part of Congress that the compacting methodology be used in such areas such as taxation, water rights, environmental regulation, and land use .... The exigencies caused by the rapid growth of gaming in Indian country and the threat of corruption and infiltration by criminal elements in class III gaming warranted the utilization of existing State regulatory capabilities in this one narrow area.”) (statement of Sen. Inouye) (emphasis added).
. In Coyote Valley II we were convinced that any inference of bad faith had been rebutted, so we did not need to address this threshold inquiry into whether the RSTF and SDF were taxes demanded by the State. Because the bad faith question is more difficult in this case, we commence with the threshold inquiry-
. Indeed, the dissent overlooks the significance of § 2710(d)(4)’s introductory phrase: "Except for any assessments that may be agreed to under [§ 2710(d)(3)(C)(iii)]." Thus, § 2710(d)(4) clearly contemplates which fees may be "agreed to,” and subsequently imposed by the state, in exchange for basic gaming rights: only those described in § 2710(d)(3)(C)(iii).
. Under § 2710(d)(4), it is not only "taxes” that are precluded, it is any "tax, fee, charge, or other assessment.” Even if the dissent were correct that the fees are not "taxes,” we fail to see how the State's demands, which the State itself described as "annual fees,” do not run afoul of this provision. The importance of the fact that the "demand” was for a “direct tax” only matters as to the question of whether the court is required to take the demand as evidence of bad faith under § 2710(d)(7)(B)(iii)(II), which refers to direct taxation but not the other sorts of extracted payments named in § 2710(d)(4). But under § 2710(d)(7)(B)(iii)(II), the language is clear that what matters is the State's "demand” for the tax. This shows that the analysis must focus on what states "demand,” not what they may think they have the authority to "impose” in the way the dissent interprets that term.
. Even if we were to accept the dissent’s interpretation of the meaning of §§ 2710(d)(3)(C) and 2710(d)(4), that would only mean that IGRA is silent on the issue of revenue sharing. The dissent takes silence as authorization, but in doing so forgets that this is a statute affecting Indian tribes. As such, we are obligated to construe ambiguities in the statute most favorably towards tribal interests, which means that we are obligated to construe the silence as a withholding of state authority to negotiate for that term. See Bryan, 426 U.S. at 392, 96 S.Ct. 2102.
. The State’s offers include reference to its desire to limit the number of gaming devices operating in the State. Although a desire to prevent excessive proliferation of casinos and gambling devices would likely be a legitimate interest justifying State refusal to permit a tribe to expand its gaming operations, such an interest is not at issue here. The State does not rely on that interest in this case, but cf. supra n.3, nor could the State credibly do so since it has shown its willingness to permit nearly unlimited gaming if the price is right.
. This distinction also gave the Seventh Circuit pause in Ho-Chunk Nation.
While we decline to use the case before us to weigh in on [the contentious issue of revenue sharing], we do note that the terms of the revenue-sharing agreements at issue in In re Indian Gaming are distinct from the one contained in the Compact between the Nation and the State. In In re Indian Gaming, the state’s use of the payments made by the tribes was heavily restricted, with all payments placed in two funds, one of which distributed gaming revenue amongst non-gaming tribes, with the other designed to fund programs to treat gambling addiction, support local agencies impacted by Indian gaming, and finance other costs directly related to gaming operations. Here, however, the Nation’s payments to the State are made without any restrictions or limits on the manner in which the State may use those funds.
*1034512 F.3d at 932 (internal citations omitted).
. We are aware that a few tribes have renegotiated their compacts with the State and accepted general fund revenue sharing in exchange for revised exclusivity. We express no opinion concerning the validity of those compacts. Those tribes agreed that the revised exclusivity offered, and the financial benefits they would receive from amending their 1999 compacts, were satisfactory to them. In this case, to the contrary, the general fund revenue sharing demanded by the State in exchange for a revised exclusivity was not freely accepted by Rincon, and was demonstrably not of significant value to Rincon.
. Even if a constitutional amendment eliminating the nontribal casino ban were imminent, the value of the proposed injunctive remedy is questionable. We are unaware of any authority that would permit a court to enjoin a constitutional amendment on the grounds that it violated Rincon’s contract.
. The State suggests the amended monetary remedy is superior to the monetary remedy provided under the 1999 compact because, although both compacts provide for the end of revenue sharing in the event exclusivity is lost, the Pala compact amendment permits continued gaming, whereas the 1999 compact requires termination. We are unconvinced that the continuation of gaming regime is better for Rincon than the termination regime. The Pala compact amendment would permit termination of revenue sharing if tribal exclusivity is lost in Rincon's core geographic market (which is a benefit Rincon would not enjoy in any event, see supra pp. 1025, 1037-38). The 1999 compact provides for termination whenever tribal exclusivity is lost anywhere statewide. Also, because IGRA requires the State to negotiate with Rincon in good faith, termination may not actually result in a loss of gaming-Rincon could obtain a new compact. Thus, under the 1999 termination option, Rincon has greater opportunities for essentially the same monetary relief.
. "[L]icensing issues under clause vi may include agreements on days and hours of operation, wage and pot limits, types of wagers, and size and capacity of the proposed facility.” S.Rep. No. 100-446, at 13, reprinted in 1988 U.S.C.C.A.N. at 3084.
. We take judicial notice of this statement by the Assistant Secretary pursuant to Federal Rule of Evidence 201.
. Congress considered amending IGRA in 2003, although no definitive action has since been taken. During the discussions, a representative from the Department of the Interior stated to Congress that the Department was seriously concerned about having to make decisions about the legitimacy of revenue sharing agreements, and wanted Congress to give it clear guidance so that it could avoid seeing "more and more of these revenues going to the States under these compacts.” S. Hrg. 108-475, at 33-34 (statement of Acting Deputy Assistant Sec’y for Policy and Econ. Dev. for the Dep't of the Interior Skirbine). The Department representative summarized, "we do not believe it was the intent of IGRA to have all the provisions up for sale.” Id. at 33.
. The dissent takes this statement out of context. At oral argument, the State argued that as long as the tribe would be better off with the new compact concerning "exclusive” gaming rights than the tribe was under its old one, the State could demand revenue sharing as its reward. Although we do not inquire into the adequacy of consideration as a general rule, when the consideration must necessarily be divided into two parts — that which IGRA contemplates and that which is outside of IGRA — we cannot bundle the rights being negotiated and compare the whole to the status quo as our method for determining whether the concessions are meaningful. The consideration in exchange for the revenue sharing must be independently meaningful in comparison to the status quo, i.e. not illusory (or illegal) if standing alone.
. "I hope the States will be fair and respectful of the authority of the tribes in negotiating these compacts and not take unnecessary advantage of the requirement for a compact.” *1041134 Cong. Rec. at S12651 (statement of Sen. Inouye).
. Interestingly, on the question of the scope of discovery permissible in IGRA negotiations, the State has taken the position that good faith should be proved based on the objective course of negotiations. See also Fort Independence Indian Cmty. v. California, No. Civ. S-08-432, 2009 WL 1283146, at *3 (E.D.Cal. May 7, 2009) (agreeing with the State that good faith should be based on objective factors). The State cannot have it both ways. If the State wants to avoid discovery and limit review of good faith to the official record of negotiations, the State cannot defend itself on the good faith question by claiming its objectively improper demands were made with an innocent intent.
. In approving the compact amendments referenced by the State, the Secretary noted that "the Department has sharply limited the *1042circumstances under which Indian tribes can make direct payments to a State for purposes other than defraying the costs of regulating gaming activities,” but agreed to approve the compacts at issue because the tribes themselves had confirmed that the exclusivity provision constituted "meaningful geographical exclusivity" and that the "amount of payment to the State [was] appropriate in light of the exclusivity right conferred.”