Opinion by Judge BEEZER; Concurrence by Judge W. FLETCHER.
BEEZER, Circuit Judge:Hawaii Governor Benjamin J. Cayetano and Attorney General Earl I. Anzai (collectively “the State”) appeal the district court’s judgment in favor of Chevron U.S.A., Inc. Chevron filed suit for declaratory and injunctive relief against enforcement of Section 3(c) of Act 257 of the 1997 Hawaii State Legislature (“Act 257”). Act 257, inter alia, proscribes the maximum rent that oil companies can collect from dealers who' lease company-owned service stations. Both parties moved for summary judgment on whether the maximum permissible regulated rent effects an unconstitutional regulatory taking. After concluding that it does, the district court granted Chevron’s motion and denied the State’s motion. The State appeals only the grant of summary judgment to Chevron; it does not appeal the denial of its own motion. We have jurisdiction under 28 U.S.C. § 1291. Because genuine issues of material fact exist, we vacate the judgment and remand for further proceedings.
I
In response to concerns about the highly concentrated gasoline market in Hawaii and the resulting high cost of gasoline to consumers, the Hawaii Legislature enacted Act 257 on June 21, 1997.1 Act 257, inter alia, regulates the maximum rent an oil company can charge dealers who lease its service stations.
Chevron is one of two gasoline refiners and one of six wholesalers in Hawaii. At the retail level, Chevron sells most of its gasoline through company-owned stations, which are leased to independent dealers. Chevron leases 64 service stations to dealers in Hawaii. From 1984 through the end of 1996, Chevron relied on estimated gasoline sales to calculate the rent owed by the lessee dealers. After determining that the amount of gross rent receipts was not satisfactory, Chevron initiated a new nationwide dealer rental program in January 1997. Chevron restructured the manner in which it calculated lease rates. This program, which the parties agree would be in effect in Hawaii absent Act 257, requires the lessee dealer to pay a monthly rent, consisting of an escalating percentage of the dealer’s gross margin on actual, rather than estimated, gasoline sales. For instance, the rent would be calculated as 18% of the gross margin up to $18,000; 32% of the portion between $18,000 and $28,000; and 38% of the portion over $28,-000. In contrast, Act 257 establishes a maximum regulated rent of 15% of gross margin.
The maximum rents Chevron projects it could receive under the statutory scheme *1033imposed by Act 257 totals $6,126,646 for 1998. Chevron’s projected expenses total $6,292,855, exceeding Chevron’s projected rental income by $166,209. Chevron concedes, however, that it has not fully recovered its expenses relating to dealer stations (including ground lease rents, real property taxes, ordinary maintenance and depreciation) from rent in any state in the last 20 years.
Instead, Chevron relies on supply contracts to earn a profit. Dealers who choose to rent a station from Chevron must, as a condition of their lease, agree to purchase from Chevron all of the fuel necessary to satisfy demand at that station for Chevron gasoline. The price under the supply contract is unilaterally set by Chevron. Both the lease agreement and the supply contract permit the dealer to transfer his or her occupancy rights upon obtaining Chevron’s written consent and paying a transfer fee set by Chevron. Act 257 does not prohibit such transfers.
In conjunction with the alienability of the leaseholds, the parties stipulated to the following facts:
34. The existing dealer at the time of the enactment of Act 257 may be able to sell his leasehold at a premium that derives from the value of the dealer’s leasehold interest, given the reduced rent imposed by Act 257, assuming Chevron does not object in good faith when the selling dealer seeks Chevron’s consent to the assignment.
35. Assuming everything else remains equal, the market value of the lessee-dealer leasehold reasonably could be expected to increase as the amount of the rent payable decreases.
Based largely on these facts, Chevron moved for summary judgment on its takings claim.2 Chevron argued that Act 257 effects a regulatory taking because it fails to substantially advance a legitimate state interest. Chevron also maintained that Act 257 prevents the company from receiving a just and reasonable return. Finally, Chevron contended that Act 257 is unconstitutional because it neither provides individualized consideration, nor contains a mechanism for obtaining relief from confiscatory rent cap provisions. Because the district court resolved the first argument in Chevron’s favor, it declined to reach the other two.
On appeal, the State challenges both the standard used by the district court to evaluate Chevron’s regulatory taking claim and the court’s application of that standard in the summary judgment context.
II
“States have broad power to regulate ... the landlord-tenant relationship ... without paying compensation for all economic injuries that such regulation entails.” Loretto v. Teleprompter Manhattan CATV Corp., 458 U.S. 419, 440, 102 S.Ct. 3164, 73 L.Ed.2d 868 (1982). When a landowner decides to rent his land to tenants, the government may place ceilings on the rents the landowner can charge. See, e.g., Pennell v. San Jose, 485 U.S. 1, 11-12, 108 S.Ct. 849, 99 L.Ed.2d 1 (1988). “[Wjhile property may be regulated to a certain extent, if regulation goes too far it will be recognized as a taking.” Pennsylvania Coal Co. v. Mahon, 260 U.S. 393, 415, 43 S.Ct. 158, 67 L.Ed. 322 (1922).
The question in this case is whether Act 257 goes too far. To analyze that question, the district court concluded that “the appropriate inquiry is whether [Act 257] substantially advances a legitimate state interest.” In reaching this conclusion, the court explicitly rejected the more deferential standard urged by the State. The State argues that the courts should look only to whether “the Legislature rationally could have believed the Act would substantially advance a legitimate government purpose.”3
*1034To support this position, the State relies on a footnote in Chief Justice Rehnquist’s dissenting opinion in Keystone Bituminous Coal Association v. DeBenedictis, 480 U.S. 470, 107 S.Ct. 1232, 94 L.Ed.2d 472 (1987):
[0]ur inquiry into legislative purpose is not intended as a license to judge the effectiveness of legislation. When considering the Fifth Amendment issues presented by Hawaii’s Land Reform Act, we noted that the Act, “like any other, may not be successful in achieving its intended goals. But ‘whether in fact the provisions will accomplish the objectives is not the question: the [constitutional requirement] is satisfied if ... the ... [State] Legislature rationally could have believed that the [Act] would promote its objective.’ ” Hawaii Housing Authority v. Midkiff, 467 U.S. 229, 242, 104 S.Ct. 2321, 81 L.Ed.2d 186 (1984).
Keystone, 480 U.S. at 511 n. 3, 107 S.Ct. 1232 (alterations in original). The State’s reliance on this quote is unsound for two reasons. First, as the district court correctly noted, Midkiff dealt with a physical taking, rather than a regulatory one. In a physical taking, the government exercises its eminent domain power to take private property for “public use.” Importantly, the government intends to take the property and is willing to pay compensation to the landowner. We have recognized that a more deferential standard applies in those circumstances. See Richardson v. City and County of Honolulu, 124 F.3d 1150, 1158 (9th Cir.1997), cert. denied, 525 U.S. 871, 119 S.Ct. 168, 142 L.Ed.2d 137 (1998) (“[W]e see nothing inconsistent in applying heightened scrutiny when the taking is uncompensated, and a more deferential standard when the taking is fully compensated.”); see also Hall v. City of Santa Barbara, 833 F.2d 1270, 1280 n. 25 (9th Cir.1986) (“It makes considerable sense to give greater deference to the legislature where it deliberately resorts to its eminent domain power than where it may have stumbled into exercising it through actions that incidentally result in a taking.”).
Second, the State’s argument is foreclosed by our decision in Richardson.' In Richardson, we established that land use regulations, including rent control ordinances like Act 257 that permit the capture of a premium, do not effect a taking if the regulation “substantially furthers a legitimate state interest.” Richardson, 124 F.3d at 1164; see also Keystone, 480 U.S. at 485, 107 S.Ct. 1232 (“[L]and use regulation can effect a taking if it ‘does not substantially advance legitimate state interests, ... or denies an owner economically viable use of his land.’ ”) (quoting Agins v. Tiburon, 447 U.S. 255, 260, 100 S.Ct. 2138, 65 L.Ed.2d 106 (1980)).
Ill
WTiile we recognize the concurring opinion’s dissatisfaction with our application of the “substantially advances” test, we do not believe that our holding today either expands Richardson or contravenes Supreme Court precedent.
Relying on Pennell v. City of San Jose, 485 U.S. 1, 108 S.Ct. 849, 99 L.Ed.2d 1 (1988), the concurrence asserts that rent control can ordinarily only be challenged as violative of due process, rather than as a regulatory taking. We read Pennell differently. In that case, a landlord challenged the constitutionality of a city rent control ordinance on three grounds: (1) the Takings Clause; (2) Due Process; and (3) Equal Protection. See id. at 4, 108 S.Ct. 849. The Court summarized the Takings Clause claim as follows:
§ 5703.28 of the Ordinance establishes the seven factors that a hearing officer is to take into account in determining the reasonable rent increase. The first six of these factors are all objective, and are related either to the landlord’s costs *1035of providing an adequate rental unit, or to the condition of the rental market. Application of these six standards results in a rent that is “reasonable” by-reference to what appellants contend is the only legitimate purpose of rent control: the elimination of “excessive” rents caused by San Jose’s housing shortage. When the hearing officer then takes into account “hardship to a tenant” pursuant to § 5708.28(c)(7) and reduces the rent below the objectively “reasonable” amount established by the first six factors, this additional reduction in the rent increase constitutes a “taking.”
Id. at 9, 108 S.Ct. 849.
The Takings Clause claim in Pennell was not based on the mere existence of rent control, but was instead dependent on the hardship provision. Indeed, because rent control is not a per se taking, see FCC v. Florida Power Corp., 480 U.S. 245, 252, 107 S.Ct. 1107, 94 L.Ed.2d 282 (1987), the landlord had to argue why this particular rent control ordinance effected a taking. To do so, he focused on the hardship provision, as the other factors apparently established a reasonable rent. The Court neither explicitly nor implicitly approved or disapproved of the landlord’s argument. In fact, the Court declined to reach the merits of the Takings Clause claim because the hardship provision had never been applied. See Pennell, 485 U.S. at 9-10, 108 S.Ct. 849. The Court did proceed to analyze the landlord’s remaining constitutional arguments. It did not, however, intimate in any way that rent control provisions should only be analyzed under the Due Process Clause. Rather, the Court determined that the Takings Clause claim was premature and then analyzed the Due Process Clause claim under the Due Process “reasonableness” test.
Moreover, to the extent that “something else” is required to challenge a rent control ordinance under the Takings Clause, the existence of the premium in this case suffices. The stipulated possibility that an incumbent dealer will be able to capture the value of the decreased rent in the form of a premium separates Act 257 from an ordinary rent control situation, where such a transfer is prohibited.
Logically, it makes far more sense for us to analyze Chevron’s regulatory takings claim under the Takings Clause test than it does to review it under the Due Process Clause test. Chevron raised a Due Process Clause claim in its First Amended Complaint, but chose not to move for summary judgment on that claim. Once the district court in this case granted Chevron’s summary judgment motion on the Takings Clause claim, Chevron dismissed its remaining claims without prejudice. Thus, there is no Due Process claim for us to review, even if we were so inclined.
The concurring opinion also downplays the significance of the Supreme Court’s language in Yee v. City of Escondido, 503 U.S. 519, 112 S.Ct. 1522, 118 L.Ed.2d 153 (1992). Although the Court in Yee did not conclusively announce the applicable test for a regulatory takings challenge to a rent control statute, it did suggest that the possibility of a premium similar to the one in this case “might have some bearing on whether the ordinance causes a regulatory taking, as it may shed some light on whether there is a sufficient nexus between the effect of the ordinance and the objectives it is supposed to advance.” Id. at 530, 112 S.Ct. 1522. The Court then cited Nollan v. California Coastal Commission, 483 U.S. 825, 107 S.Ct. 3141, 97 L.Ed.2d 677 (1987), which employs the “substantially advances” test.
The Yee Court went slightly further than in Pennell and held that the facial regulatory takings challenge was ripe. See Yee, 503 U.S. at 534, 112 S.Ct. 1522. The Court did not reach the merits of the regulatory takings claim, however, because it concluded that the issue was not fairly presented in the certiorari petition. See id. at 537, 112 S.Ct. 1522. In reaching that conclusion, the Court noted that “were we to address the issue here, we would apparently be the first court in the Nation to determine whether an ordinance *1036like this one effects a regulatory taking.” Id. at 538, 112 S.Ct. 1522.
Although the Court recognized the novelty of the regulatory takings claim, it did not take any position on its merits. The ordinance in Yee is very similar to Act 257, in that the transfer of the mobile homes in Yee and the service station leaseholds here both create the possibility of a one-time transfer of wealth in the form of a premium that inures to incumbent lessees. Had the Court in Pennell rejected a regulatory takings challenge such as the one posed by Chevron in the instant case, as the concurring opinion suggests, the Court in Yee would not have needed to expressly decline to reach the issue. We believe that the Court’s treatment of the regulatory takings issue in Yee further undermines the concurrence’s reading of Pennell.
City of Monterey v. Del Monte Dunes at Monterey, Ltd., 526 U.S. 687, 119 S.Ct. 1624, 143 L.Ed.2d 882 (1999), gives additional support to the application of the “substantially advances” test to Chevron’s Takings Clause claim. Although the Court recognized that it has not provided a “thorough explanation of the nature or applicability of the requirement that a regulation substantially advance legitimate public interests outside the context of required dedications or exactions,” it noted that the jury instructions given by the trial court regarding the “substantially advances” test were consistent with the Court’s previous general discussions of regulatory takings liability. Id. at 704, 119 S.Ct. 1624. As support for this statement, the Court cited several land use regulatory takings cases, including Yee. See id. It is apparent to us that the Court viewed the rent control ordinance in Yee in the same manner as the land use regulations in the other cases. It is further apparent to us that when challenged as a regulatory taking, each is subject to the same “substantially advances” test.
We have great difficulty finding Supreme Court precedent to support the concurring opinion’s assertion that rent control should be viewed differently than other land use regulations. Accord Santa Monica Beach, Ltd. v. Superior Court, 19 Cal.4th 952, 81 Cal.Rptr.2d 93, 968 P.2d 993, 1021 (1999) (Baxter, J., dissenting) (“[W]e cannot assume that decisions upholding other forms of price control are authority for rejecting a takings clause challenge to a rent control ordinance.”).
The concurring opinion also does not satisfactorily explain how to deal with Richardson. Richardson is the law of our circuit and has conclusively answered the question of what test should be applied in this case. Nevertheless, the concurring opinion attempts to avoid the holding of Richardson by inventing a “certainty” requirement that does not exist. , The concurrence would limit the “substantially advances” test to eases in which “the existence of the premium capture is essentially beyond dispute.” We do not believe that Richardson may faithfully be read so narrowly.4
Finally, we note that other federal cases have applied the “substantially advances” test in considering a regulatory takings challenge to a rent control ordinance. See Federal Home Loan Mortgage Corp. v. New York State Div. of Hous. and Community Renewal, 83 F.3d 45, 48 (2d Cir.1996) (concluding that rent stabilization law does not constitute either physical or regulatory taking); Greystone Hotel Co. v. City of New York, 13 F.Supp.2d 524, 527-29 (S.D.N.Y.) (concluding that rent stabilization provision substantially advanced a legitimate state interest and thus did not effect a regulatory taking); Adamson Cos. v. City of Malibu, 854 F.Supp. 1476, 1501-02 (C.D.Cal.1994) (concluding that city’s mobile home rent control ordinance was substantially related to a legitimate interest).
*1037In sum, we disagree with the concurrence’s position that we should apply the “reasonableness” test to evaluate Chevron’s regulatory takings claim. The correct test is “whether the legislation substantially advances a legitimate state interest,” as discussed above, as suggested by the Supreme Court in Yee, as used by the district court in this case, and as established by this court in Richardson.
IV
Although the district court applied the correct standard, it should not have granted summary judgment. Notwithstanding the fact that both sides moved for summary judgment and agreed that summary judgment was appropriate one way or the other, genuine issues of material fact remain.5
We review a district court’s grant of summary judgment de novo. See Balint v. Carson City, 180 F.3d 1047, 1050 (9th Cir.1999) (en banc). We determine, viewing the evidence in the light most favorable to the nonmoving party, whether the district court correctly applied the relevant substantive law and whether there are any genuine issues of material fact. See id. Because Chevron challenges Act 257 on its face, rather than as applied, Chevron bears the burden of proving by a preponderance of the evidence that the regulation does not substantially advance a legitimate state interest. See City of Monterey v. Del Monte Dunes at Monterey, Ltd., 526 U.S. 687, 700-01, 119 S.Ct. 1624, 143 L.Ed.2d 882 (1999).
A
The State argues here that Chevron was not entitled to summary judgment because the district court ignored genuine issues of fact raised by the affidavit of its expert economist, Dr. Keith Leffler. Chevron, in its moving papers, relied solely on the stipulated fact that an existing dealer may be able to sell his leasehold at a premium that derives from the increased value of the dealer’s leasehold interest due to the reduced rent imposed by Act 257. Based on this possible premium transfer, Chevron argued that the benefits of Act 257 will inure to the incumbent dealer, rather than to consumers. Thus, according to Chevron, Act 257 does not substantially advance its purpose of lowering gasoline prices.
The State relied on Dr. Leffler to rebut Chevron’s argument. In his affidavit, Dr. Leffler asserted that Act 257 is likely to lessen the adverse competitive effects that result from the highly concentrated gasoline market in Hawaii and thereby benefit consumers. This is so, according to Dr. Leffler, because lower monthly rent payments reduce a dealer’s cost of continued operation, meaning more dealers are likely to stay in business. The presence of more dealers means a greater supply of fuel to consumers and a greater supply leads to lower prices. Therefore, Dr. Leffler predicted that the rent cap would act to lower retail gasoline prices. Although Dr. Lef-fler acknowledged that an oil company could raise the wholesale cost of gasoline to dealers in order to make up for lost rent, he believed that wholesale oil suppliers would be unlikely to do so because such action would lead directly to reductions in volume as dealers would react by raising their street price. He also believed that while a premium transfer was *1038possible, an incumbent dealer could not be expected to capture the net present value of the reduced rent because the net present value would depend on too many unknown variables, such as the oil company’s rent policy, the future level of gasoline margins, and the sales of goods other than gasoline. Even if new lessee dealers could estimate the present value of rent reductions accurately, Dr. Leffler stated that any cost reductions would be fully captured by incumbent dealers only if the market for the sale of lessee dealer rights was perfectly competitive. Relying on the stipulated facts, which indicate that less than three sales of dealer rights occur per year, Dr. Leffler concluded that the market is relatively thin, rather than perfectly competitive. Finally, because Act 257 encourages dealers to stay in business longer by lowering fixed costs, Dr. Leffler asserted that Act 257 will benefit consumers even if the premium is fully captured.
In response to Dr. Leffler, Chevron proffered an affidavit from its own expert economist Dr. John Umbeck. Dr. Um-beck’s opinion differs significantly from that of Dr. Leffler. Dr. Umbeck believes that Act 257 will reduce the net present value of an oil company’s rental revenue and thereby discourage oil companies from building new stations or maintaining old ones. Accordingly, Dr. Umbeck asserted that Act 257 will actually increase the concentration of Hawaii’s gasoline market, rather than decrease it. Because Dr. Um-beck believes that there will be fewer stations in the long run, he states that the demand facing surviving dealers will increase, thus motivating and allowing them to raise fuel prices. Finally, Dr. Umbeck disputes Dr. Leffler’s conclusion that the market for the sale of lessee dealer rights is thin. Dr. Umbeck believes that the premium value of the capped rent will be fully captured by the incumbent dealer.
The conflicting affidavits establish that genuine issues of fact remain as to whether Act 257 will result in lower gasoline prices. Whether, and to what extent, Chevron will raise its wholesale price of fuel to compensate for lost rent, and whether, and to what extent, incumbent dealers will capture the value of the capped rent in the form of a premium upon transfer of the leasehold, remain as unanswered questions. Moreover, to the extent that the district court purported to answer these questions, such resolution was premature.6
*1039Admittedly, these are questions of predictive fact, rather than historical fact. Nevertheless, summary judgment is no more appropriate when predictive facts are disputed. See Washington Post Co. v. United States Dep’t of Health and Human Servs., 865 F.2d 320, 326 (D.C.Cir.1989) (reversing summary judgment based on conflicting experts’ predictions). In Washington Post, the plaintiff sought access under the Freedom of Information Act, 5 U.S.C. § 552, to certain financial disclosure forms filed by National Cancer Institute consultants. See id. at 321. The district court granted summary judgment in favor of the government. On appeal, the D.C. Circuit held that a genuine issue of material fact existed as to the effect of disclosure on the government’s ability to obtain the information it needed from its scientist-consultants. See id. In so holding, the court reviewed the evidence in the record and concluded that the conflicting affidavits created a genuine issue of fact regarding the consequences of disclosing the sought-after information. The court explicitly rejected the trial court’s explanation that
[r]esolution of this dispute involves an analysis of the potential, hypothetical impairment the government might suffer.... Due to the nature of the inquiry, there is no definitive proof that may be adduced by either side in support of their respective contentions. At best, the parties may provide the Court with speculation from individuals who speak with varying degrees of authority.
Id. at 324. The court held that the district court had “short-circuited the fact-finding process.” Id. at 326. The court further explained that:
“Factual” issues like those presented here are rarely susceptible to definitive proof. Rather, “factual” issues that involve predictive facts almost always require a court to survey the available evidence, to credit certain pieces of evidence above others, and to draw cumulative inferences until it reaches a judgmental conclusion. In the end, the court makes its best assessment about what is most likely to happen in the future. In such an inquiry, the ultimate “facts” in dispute are most successfully approached when all relevant evidentiary underpinnings are fully developed.
Id. at 326.
As in Washington Post, the experts’ conflicting predictions here created genuine issues of fact. A proper evaluation of Act 257 depends upon further development of the experts’ underlying facts. Although both experts based their opinions on the stipulated facts, they also rely on other unproven factual assumptions. In order to determine whose predictions are more accurate, there needs to be a better understanding of the competitiveness of the market for the sale and purchase of lessee dealer rights and the elasticity of demand for gasoline. This can be attained only through additional factual development and cross-examination of the parties’ expert witnesses. Cross-examination is particularly appropriate here because it is necessary for the court to evaluate the witnesses’ credibility in order to evaluate their expert opinions. By adopting the predictions of Chevron’s expert without the benefit of this needed information, the district court “short-circuited the fact-finding process.”
B
Issues of fact do not preclude summary judgment unless they are material to the substantive claim at issue, that is, unless they “might affect the outcome of the suit under the governing, law.” Moreland v. Las Vegas Metro. Police Dep’t, 159 F.3d 365, 369 (9th Cir.1998) (quoting Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 106 S.Ct. 2505, 91 *1040L.Ed.2d 202 (1986)). The factual disputes here are material only if their resolution would alter whether Act 257 “substantially advances a legitimate state interest.”
In order to determine materiality, we first determine what is required by the “substantially advances” test. Obviously, a connection of some sort must be established between Act 257’s means (that is, a maximum regulated rent) and the intended end (lower gas prices). The question arises whether the mere possibility that the rent cap will not achieve its purpose is sufficient to destroy this nexus. The district court, relying on Richardson, believed this to be the case. If that belief is correct, the factual disputes are not material because the parties concede that that possibility exists. If, however, the standard requires a closer evaluation of the likelihood that the means will achieve the end, then these disputes are material.
In Richardson, we held that a Honolulu ordinance that imposed a cap on rent for land under condominium units was an unconstitutional regulatory taking. The court concluded that:
The absence of a mechanism that prevents lessees from capturing the net present value of the reduced land rent in the form of a premium, means that the Ordinance will not substantially further its goal of creating affordable owner-occupied housing in Honolulu. Incumbent owner occupants who sell to those who intend to occupy the apartment will charge a premium for the benefit of living in a rent controlled condominium. The price of housing ultimately will remain the same. The Ordinance thus effects a regulatory taking.
124 F.3d at 1166. Importantly, our conclusion in Richardson was based on the district court’s findings that incumbent owners will charge a premium and that the price of housing will remain the same. Apparently, these findings were not contested on appeal.
Whether the price of gasoline will remain the same here is vigorously contested. The mere possibility that it will does not satisfy Chevron’s burden of proving that Act 257 does not substantially further its purpose. Logically, there must be a determination of the likelihood of that possibility. Otherwise, regulatory legislation would be unconstitutional any time it was not absolutely guaranteed to achieve its purpose. Would a statute that was 95% likely to be effective fail to substantially advance its interest simply because there was a 5% chance that it would not achieve its goal? What if the statute was 99% effective? Surely the 1% chance that it would be ineffective does not, as a matter of law, mean that the statute does not “substantially advance” its purpose.
Not only does such a reading of Richardson fail the logic test, but it also fails to consider the language. Had the test been intended to require an absolute cause-and-effect relationship, the word “substantially” would have no meaning. For these reasons, the district court improperly relied on Richardson to conclude that the absence of a mechanism that prevents a premium transfer necessarily destroys the constitutionally-required connection.
The State argues that the existence of a premium is not relevant. As discussed in Section III above, however, the Supreme Court recognized this relevance in Yee, 503 U.S. at 530, 112 S.Ct. 1522. While the existence of a premium transfer had nothing to do with a physical taking, the Court stated that “[t]his effect might have some bearing on whether the ordinance causes a regulatory taking, as it may shed some light on whether there is a sufficient nexus between the effect of the ordinance and the objectives it is supposed to advance.” Id. The Supreme Court did not hold that the possibility of a premium transfer necessarily defeated that connection; it recognized only that it “might have some bearing” on it. Id.
The question remains as to what exactly the connection does require. In a candid statement, the Supreme Court recently acknowledged that is has not provided a “thorough explanation of the nature or ap*1041plicability of the requirement that a regulation substantially advance legitimate public interests outside the context of required dedications or exactions.... ” City of Monterey, 526 U.S. at 704, 119 S.Ct. 1624. In City of Monterey, the Court affirmed judgment in favor of Del Monte Dunes, a property owner who contended that the city’s repeated rejections of its plans for development effected an unconstitutional regulatory taking. The Court held that the case was properly submitted to a jury. Although the Court refused to consider the City of Monterey’s challenge to the jury instructions, as the City itself had proposed their essence, the Court nonetheless noted that the following jury instructions were consistent with the Court’s previous discussions of regulatory takings liability:
“Public bodies, such as the city, have the authority to take actions which substantially advance legitimate public interests .... So one of your jobs as jurors is to decide if the city’s decision here substantially advanced any such legitimate public purpose.”
“The regulatory actions of the city or any agency substantially advanc[e] a legitimate public purpose of the action bears a reasonable relationship to that objective.”
“Now, if the preponderance of the evidence establishes that there was no reasonable relationship between the city’s denial of the ... proposal and legitimate public purpose, you should find in favor of the plaintiff. If you find that there existed a reasonable relationship between the city’s decision and a legitimate public purpose, you should find in favor of the city. As long as the regulatory action by the city substantially advances their [sic] legitimate public purpose, ... its underlying motives and reasons are not to be inquired into.”
526 U.S. at 700-01, 119 S.Ct. 1624. Based on these instructions, a challenged regulatory action “substantially advances” its interest if it bears a reasonable relationship to that interest.
Understanding what the “substantially advances” test requires also depends to some extent on understanding what it does not require. In City of Monterey, the Court declined to extend the “rough proportionality” test beyond the special context of exactions — land use decisions conditioning approval of development on the dedication of property to public use. See id. at 702-03, 119 S.Ct. 1624. Thus, City of Monterey teaches that a reasonable relationship does not depend on the State’s action being roughly proportional to its asserted interests.
Whether Act 257’s rent cap is reasonably related to its objective of lowering fuel prices certainly depends on whether it will in fact lead to lower fuel prices. The factual issues that remain in dispute are material to the ultimate determination required in this action. We hold that the district court should not have granted summary judgment on Chevron’s first argument.
C
Because the judgment can be affirmed on any basis supported in the record, we briefly consider the second and third arguments made by Chevron in its summary judgment motion, even though the district court did not. There are two ways in which a party can challenge a regulatory action on its face. The first has been thoroughly discussed above. The second depends on whether the regulation deprives an owner of the economic viability of property. An owner is not denied the economic viability of property, unless there remains no permissible or beneficial use for that property after the regulatory action. See City of Monterey, 526 U.S. at 700, 119 S.Ct. 1624 (quoting jury instruction to that effect). Chevron has made no such showing. In fact, the evidence shows that Act 257 allows Chevron to charge approximately $1.1 million more than it would otherwise have charged under its own rental program. Although this evidence may be useful to Chevron in demon*1042strating the ineffectiveness of Act 257, it belies its claim of loss of economic viability.
Chevron also argues that Act 257 is unconstitutional because it fails to provide for any individualized consideration and contains no mechanism for obtaining relief from the confiscatory rent limitation provisions. In support of this argument, Chevron relies on two state court cases and one of the two district court orders in Richardson. See Cromwell Assocs. v. Mayor and Council of Newark, 211 N.J.Super. 462, 511 A.2d 1273, 1275 (1985); Birkenfeld v. City of Berkeley, 17 Cal.3d 129, 130 Cal.Rptr. 465, 550 P.2d 1001, 1028-1033 (1976); Richardson v. City and County of Honolulu, 802 F.Supp. 326, 336-37 (D.Haw.1992) (“Richardson II”). The precedents established in these cases do not bind us. Although addressed by the district court in Richardson II, in Richardson we declined to reach the issue on appeal. Likewise, the Supreme Court in Permian Basin Area Rate Cases, 390 U.S. 747, 770, 88 S.Ct. 1344, 20 L.Ed.2d 312 (1968), refused to decide whether individualized consideration and administrative relief were “constitutionally imperative.” Earlier Supreme Court cases, however, have suggested that there are no such constitutional requirements. In Bowles v. Willingham, 321 U.S. 503, 518, 64 S.Ct. 641, 88 L.Ed. 892 (1944), the Court stated that otherwise valid price-fixing was not improper because it was on a class rather than an individual basis. Justice Brandéis stated for a unanimous Court in Phillips v. Commissioner, 283 U.S. 589, 596-97, 51 S.Ct. 608, 75 L.Ed. 1289 (1931), that “[w]here only property rights are involved, mere postponement of the judicial enquiry is not a denial of due process, if the opportunity given for the ultimate judicial determination of the liability is adequate.” As evidenced by the instant action, Chevron was not denied an opportunity to seek judicial review of Act 257.
Finally, the cases relied upon by Chevron — Cromwell Associates, Birkenfeld, and Richardson II — are distinguishable because they involved ordinances that regulated a landlord’s sole source of revenue. Imposing a maximum rent removed the only mechanism by which a landlord could increase revenues in the event of an increase in costs. Chevron’s lessee dealer stations provide Chevron with two sources of revenue. For the past 20 years, Chevron has relied on its lessee dealers’ contractually-required purchase of gasoline to assure a reasonable rate of return on its service stations. Act 257 does not regulate this revenue source. Accordingly, Act 257’s failure to contain an administrative relief provision does not result in a facial taking.
V
Genuine issues of material fact exist as to whether Act 257 will benefit consumers. Specifically, the record contains conflicting evidence as to whether incumbent dealers will capture a premium based on the increased value of their leaseholds due to the imposition of a maximum permissible regulated rent, thereby depriving new dealers and consumers from reaping the benefits of Act 257. Questions also remain as to whether oil companies will raise the wholesale price of fuel and thereby unilaterally offset the benefits of the Act. Because resolution of these factual issues is necessary to determine whether Act 257 substantially advances, or bears a reasonable relationship to, the State’s interest in lowering gasoline prices, the district court erred in granting summary judgment. Summary judgment is likewise inappropriate on the other two grounds urged by Chevron.
We VACATE the judgment of the district court and REMAND for further proceedings consistent with this opinion.
. The relevant portions of Act 257 were later codified as Hawaii Revised Statute § 486H-10.4.
. Chevron’s other three claims (42 U.S.C. § 1983, due process and equal protection) were dismissed without prejudice by slipulation of the parties and are not at issue in this appeal.
. Both the test used by the district court and *1034that suggested by the State require a legitimate state interest. In this case, the district court found that the purpose of Act 257 is to "reduc[e] gasoline prices for Hawaii's consumers.” On appeal, the parties do not contest this finding. Likewise, the parties do not dispute the legitimacy of this interest.
. We also do not believe that our decision today expands the holding of Richardson. We save for another day the question of whether the “substantially advances" test applies outside the context of rent control statutes that permit the capture of a premium.
. At least one court has held that, under similar circumstances involving cross-motions for summary judgment, a party that failed to raise the existence of genuine issues of material fact before the district court waived the right to do so on appeal. See Shrink Missouri Gov’t PAC v. Maupin, 71 F.3d 1422, 1423 (8th Cir.1995). Even then, however, the court proceeded to analyze whether genuine issues of material fact remained. The better-reasoned approach holds that the district court is responsible for determining whether the requirements of Federal Rule of Civil Procedure 56 are met, whether or not the parties believe that they are. See, e.g., William W Schwarzer et al., The Analysis and Decision of Summary Judgment Motions, 139 F.R.D. 441, 499 (Feb. 1992) (“The filing of cross-motions does not ensure that summary judgment is in order.”). *1039no clear indication, as there was in Starsky, that all of the relevant facts are contained in the written record. In fact, in the Stipulation itself, the parties clearly, reserve the right to rely on other statements. Therefore, we hold that the district court erred in resolving factual disputes at this stage, even though it must do so later.
. That the district court engaged in improper fact-finding is apparent from its summary judgment order. For instance, the court found, despite evidence to the contrary, that "[bjecause oil companies can simply raise their wholesale prices to the same extent that rent decreases, the dealer is likely to be unaffected by the rent cap, and gas prices will remain unchanged.”
Because this case will be tried to the court, rather than a jury, the question arises whether it was improper for the district court to engage in fact-finding, since it will eventually be called upon to perform that veiy task. This court has held that if the parties agree that all of the underlying material facts are reflected in the written record, a judge may decide factual issues and essentially convert cross-motions for summary judgment into submission of the case for trial on the written record. See Starsky v. Williams, 512 F.2d 109, 111 (9th Cir.1975). The court noted, however, that that result was justified by the unique circumstances of that case. See id. Moreover, the court later cautioned that such fact-finding is not appropriate on an undeveloped factual record. See TransWorld Airlines, Inc. v. American Coupon Exchange, Inc., 913 F.2d 676 (9th Cir.1990). There, the court stated:
[Wjhere the ultimate fact in dispute is destined for decision by the court rather than by a jury, there is no reason why the court and the parties should go through the motions of a trial if the court will eventually end up deciding on the same record. However, just as the procedural shortcut must not be disfavored, courts must not rush to dispose summarily of cases-especially novel, complex, or otherwise difficult cases of public importance-unless it is clear that more complete factual development could not possibly alter the outcome and that the credibility of the witnesses’ statements or testimony is not at issue. Even when the expense of further proceedings is great and the moving party’s case seems to the court quite likely to succeed, speculation about the facts must not take the place of investigation, proof, and direct observation.
Id. at 684-85. Although the parties in this case submitted a Stipulation of Facts, there is