Justice O’Connor delivered the opinion of the Court.

On occasion, a would-be doctrinal rule or test finds its way into our case law through simple repetition of a phrase—however fortuitously coined. A quarter century ago, in Agins v. City of Tiburon, 447 U. S. 255 (1980), the Court declared that government regulation of private property “effects a taking if [such regulation] does not substantially advance legitimate state interests ….” Id., at 260. Through reiteration in a half dozen or so decisions since Agins, this language has been ensconced in our Fifth Amendment takings jurisprudence. See Monterey v. Del Monte Dunes at Monterey, Ltd., 526 U. S. 687, 704 (1999) (citing cases).

In the case before us, the lower courts applied Agins ’ “substantially advances” formula to strike down a Hawaii statute that limits the rent that oil companies may charge to dealers who lease service stations owned by the companies. The lower courts held that the rent cap effects an uncompensated taking of private property in violation of the Fifth and Fourteenth Amendments because it does not substantially advance Hawaii’s asserted interest in controlling retail gasoline prices. This case requires us to decide whether the “substantially advances” formula announced in Agins is an appropriate test for determining whether a regulation effects a Fifth Amendment taking. We conclude that it is not.

I

The State of Hawaii, whose territory comprises an archipelago of 132 islands clustered in the midst of the Pacific Ocean, is located over 1,600 miles from the U. S. mainland and ranks among the least populous of the 50 States. Because of Hawaii’s small size and geographic isolation, its wholesale market for oil products is highly concentrated. When this lawsuit began in 1997, only two refineries and six gasoline wholesalers were doing business in the State. As of that time, respondent Chevron U. S. A. Inc. was the largest refiner and marketer of gasoline in Hawaii: It controlled 60 percent of the market for gasoline produced or refined in-state and 30 percent of the wholesale market on the State’s most populous island, Oahu.

Gasoline is sold at retail in Hawaii from about 300 different service stations. About half of these stations are leased from oil companies by independent lessee-dealers, another 75 or so are owned and operated by “open” dealers, and the remainder are owned and operated by the oil companies. Chevron sells most of its product through 64 independent lessee-dealer stations. In a typical lessee-dealer arrangement, Chevron buys or leases land from a third party, builds a service station, and then leases the station to a dealer on a turnkey basis. Chevron charges the lessee-dealer a monthly rent, defined as a percentage of the dealer’s margin on retail sales of gasoline and other goods. In addition, Chevron requires the lessee-dealer to enter into a supply contract, under which the dealer agrees to purchase from Chevron whatever is necessary to satisfy demand at the station for Chevron’s product. Chevron unilaterally sets the wholesale price of its product.

The Hawaii Legislature enacted Act 257 in June 1997, apparently in response to concerns about the effects of market concentration on retail gasoline prices. See 1997 Haw. Sess. Laws no. 257, §1. The statute seeks to protect independent dealers by imposing certain restrictions on the ownership and leasing of service stations by oil companies. It prohibits oil companies from converting existing lessee-dealer stations to company-operated stations and from locating new company-operated stations in close proximity to existing dealer-operated stations. Haw. Rev. Stat. §§486H–10.4(a), (b) (1998 Cum. Supp.). More importantly for present purposes, Act 257 limits the amount of rent that an oil company may charge a lessee-dealer to 15 percent of the dealer’s gross profits from gasoline sales plus 15 percent of gross sales of products other than gasoline. §486H–10.4(c).

Thirty days after Act 257’s enactment, Chevron sued the Governor and Attorney General of Hawaii in their official capacities (collectively Hawaii) in the United States District Court for the District of Hawaii, raising several federal constitutional challenges to the statute. As pertinent here, Chevron claimed that the statute’s rent cap provision, on its face, effected a taking of Chevron’s property in violation of the Fifth and Fourteenth Amendments. Chevron sought a declaration to this effect as well as an injunction against the application of the rent cap to its stations. Chevron swiftly moved for summary judgment on its takings claim, arguing that the rent cap does not substantially advance any legitimate government interest. Hawaii filed a cross-motion for summary judgment on all of Chevron’s claims.

To facilitate resolution of the summary judgment motions, the parties jointly stipulated to certain relevant facts. They agreed that Act 257 reduces by about $207,000 per year the aggregate rent that Chevron would otherwise charge on 11 of its 64 lessee-dealer stations. On the other hand, the statute allows Chevron to collect more rent than it would otherwise charge at its remaining 53 lessee-dealer stations, such that Chevron could increase its overall rental income from all 64 stations by nearly $1.1 million per year. The parties further stipulated that, over the past 20 years, Chevron has not fully recovered the costs of maintaining lessee-dealer stations in any State through rent alone. Rather, the company recoups its expenses through a combination of rent and product sales. Finally, the joint stipulation states that Chevron has earned in the past, and anticipates that it will continue to earn under Act 257, a return on its investment in lessee-dealer stations in Hawaii that satisfies any constitutional standard.

The District Court granted summary judgment to Chevron, holding that “Act 257 fails to substantially advance a legitimate state interest, and as such, effects an unconstitutional taking in violation of the Fifth and Fourteenth Amendments.” Chevron U. S. A. Inc. v. Cayetano, 57 F. Supp. 2d 1003, 1014 (1998). The District Court accepted Hawaii’s argument that the rent cap was intended to prevent concentration of the retail gasoline market—and, more importantly, resultant high prices for consumers—by maintaining the viability of independent lessee-dealers. Id., at 1009–1010. The court concluded that the statute would not substantially advance this interest, however, because it would not actually reduce lessee-dealers’ costs or retail prices. It found that the rent cap would allow incumbent lessee-dealers, upon transferring occupancy rights to a new lessee, to charge the incoming lessee a premium reflecting the value of the rent reduction. Accordingly, the District Court reasoned, the incoming lessee’s overall expenses would be the same as in the absence of the rent cap, so there would be no savings to pass along to consumers. Id., at 1010–1012. Nor would incumbent lessees benefit from the rent cap, the court found, because the oil company lessors would unilaterally raise wholesale fuel prices in order to offset the reduction in their rental income. Id., at 1012–1014.

On appeal, a divided panel of the Court of Appeals for the Ninth Circuit held that the District Court had applied the correct legal standard to Chevron’s takings claim. Chevron U. S. A. Inc. v. Cayetano, 224 F. 3d 1030, 1033–1037 (2000). The Court of Appeals vacated the grant of summary judgment, however, on the ground that a genuine issue of material fact remained as to whether the Act would benefit consumers. Id., at 1037–1042. Judge William Fletcher concurred in the judgment, maintaining that the “reasonableness” standard applicable to “ordinary rent and price control laws” should instead govern Chevron’s claim. Id., at 1048.

On remand, the District Court entered judgment for Chevron after a 1-day bench trial in which Chevron and Hawaii called competing expert witnesses (both economists) to testify. 198 F. Supp. 2d 1182 (2002). Finding Chevron’s expert witness to be “more persuasive” than the State’s expert, the District Court once again concluded that oil companies would raise wholesale gasoline prices to offset any rent reduction required by Act 257, and that the result would be an increase in retail gasoline prices. Id., at 1187–1189. Even if the rent cap did reduce lessee-dealers’ costs, the court found, they would not pass on any savings to consumers. Id., at 1189. The court went on to reiterate its determination that Act 257 would enable incumbent lessee-dealers to sell their leaseholds at a premium, such that incoming lessees would not obtain any of the benefits of the rent cap. Id., at 1189–1190. And while it acknowledged that the rent cap could preclude oil companies from constructively evicting dealers through excessive rents, the court found no evidence that Chevron or any other oil company would attempt to charge such rents in the absence of the cap. Id., at 1191. Finally, the court concluded that Act 257 would in fact decrease the number of lessee-dealer stations because the rent cap would discourage oil companies from building such stations. Id., at 1191–1192. Based on these findings, the District Court held that “Act 257 effect[ed] an unconstitutional regulatory taking given its failure to substantially advance any legitimate state interest.” Id., at 1193.

The Ninth Circuit affirmed, holding that its decision in the prior appeal barred Hawaii from challenging the application of the “substantially advances” test to Chevron’s takings claim or from arguing for a more deferential standard of review. 363 F. 3d 846, 849–855 (2004). The panel majority went on to reject Hawaii’s challenge to the application of the standard to the facts of the case. Id., at 855–858. Judge Fletcher dissented, renewing his contention that Act 257 should not be reviewed under the “substantially advances” standard. Id., at 859–861. We granted certiorari, 543 U. S. ___ (2004), and now reverse.

II

A

The Takings Clause of the Fifth Amendment, made applicable to the States through the Fourteenth, see Chicago, B. & Q. R. Co. v. Chicago, 166 U. S. 226 (1897), provides that private property shall not “be taken for public use, without just compensation.” As its text makes plain, the Takings Clause “does not prohibit the taking of private property, but instead places a condition on the exercise of that power.” First English Evangelical Lutheran Church of Glendale v. County of Los Angeles, 482 U. S. 304, 314 (1987). In other words, it “is designed not to limit the governmental interference with property rights per se, but rather to secure compensation in the event of otherwise proper interference amounting to a taking.” Id., at 315 (emphasis in original). While scholars have offered various justifications for this regime, we have emphasized its role in “bar[ring] Government from forcing some people alone to bear public burdens which, in all fairness and justice, should be borne by the public as a whole.” Armstrong v. United States, 364 U. S. 40, 49 (1960); see also Monongahela Nav. Co. v. United States, 148 U. S. 312, 325 (1893).

The paradigmatic taking requiring just compensation is a direct government appropriation or physical invasion of private property. See, e.g., United States v. Pewee Coal Co., 341 U. S. 114 (1951) (Government’s seizure and operation of a coal mine to prevent a national strike of coal miners effected a taking); United States v. General Motors Corp., 323 U. S. 373 (1945) (Government’s occupation of private warehouse effected a taking). Indeed, until the Court’s watershed decision in Pennsylvania Coal Co. v. Mahon, 260 U. S. 393 (1922), “it was generally thought that the Takings Clause reached only a ‘direct appropriation’ of property, or the functional equivalent of a ‘practical ouster of [the owner’s] possession.’ ” Lucas v. South Carolina Coastal Council, 505 U. S. 1003, 1014 (1992) (citations omitted and emphasis added; brackets in original); see also id., at 1028, n. 15 (“[E]arly constitutional theorists did not believe the Takings Clause embraced regulations of property at all”).

Beginning with Mahon, however, the Court recognized that government regulation of private property may, in some instances, be so onerous that its effect is tantamount to a direct appropriation or ouster—and that such “regulatory takings” may be compensable under the Fifth Amendment. In Justice Holmes’ storied but cryptic formulation, “while property may be regulated to a certain extent, if regulation goes too far it will be recognized as a taking.” 260 U. S., at 415. The rub, of course, has been—and remains—how to discern how far is “too far.” In answering that question, we must remain cognizant that “government regulation—by definition—involves the adjustment of rights for the public good,” Andrus v. Allard, 444 U. S. 51, 65 (1979), and that “Government hardly could go on if to some extent values incident to property could not be diminished without paying for every such change in the general law,” Mahon, supra, at 413.

Our precedents stake out two categories of regulatory action that generally will be deemed per se takings for Fifth Amendment purposes. First, where government requires an owner to suffer a permanent physical invasion of her property—however minor—it must provide just compensation. See Loretto v. Teleprompter Manhattan CATV Corp., 458 U. S. 419 (1982) (state law requiring landlords to permit cable companies to install cable facilities in apartment buildings effected a taking). A second categorical rule applies to regulations that completely deprive an owner of “ all economically beneficial us[e]” of her property. Lucas, 505 U. S., at 1019 (emphasis in original). We held in Lucas that the government must pay just compensation for such “total regulatory takings,” except to the extent that “background principles of nuisance and property law” independently restrict the owner’s intended use of the property. Id., at 1026–1032.

Outside these two relatively narrow categories (and the special context of land-use exactions discussed below, see infra, at 16–18), regulatory takings challenges are governed by the standards set forth in Penn Central Transp. Co. v. New York City, 438 U. S. 104 (1978). The Court in Penn Central acknowledged that it had hitherto been “unable to develop any ‘set formula’ ” for evaluating regulatory takings claims, but identified “several factors that have particular significance.” Id., at 124. Primary among those factors are “[t]he economic impact of the regulation on the claimant and, particularly, the extent to which the regulation has interfered with distinct investment-backed expectations.” Ibid. In addition, the “character of the governmental action”—for instance whether it amounts to a physical invasion or instead merely affects property interests through “some public program adjusting the benefits and burdens of economic life to promote the common good”—may be relevant in discerning whether a taking has occurred. Ibid. The Penn Central factors—though each has given rise to vexing subsidiary questions—have served as the principal guidelines for resolving regulatory takings claims that do not fall within the physical takings or Lucas rules. See, e.g., Palazzolo v. Rhode Island, 533 U. S. 606, 617–618 (2001); id., at 632–634 (O’Connor, J., concurring).

Although our regulatory takings jurisprudence cannot be characterized as unified, these three inquiries (reflected in Loretto, Lucas, and Penn Central ) share a common touchstone. Each aims to identify regulatory actions that are functionally equivalent to the classic taking in which government directly appropriates private property or ousts the owner from his domain. Accordingly, each of these tests focuses directly upon the severity of the burden that government imposes upon private property rights. The Court has held that physical takings require compensation because of the unique burden they impose: A permanent physical invasion, however minimal the economic cost it entails, eviscerates the owner’s right to exclude others from entering and using her property—perhaps the most fundamental of all property interests. See Dolan v. City of Tigard, 512 U. S. 374, 384 (1994); Nollan v. California Coastal Comm’n, 483 U. S. 825, 831–832 (1987); Loretto, supra, at 433; Kaiser Aetna v. United States, 444 U. S. 164, 176 (1979). In the Lucas context, of course, the complete elimination of a property’s value is the determinative factor. See Lucas, supra, at 1017 (positing that “total deprivation of beneficial use is, from the landowner’s point of view, the equivalent of a physical appropriation”). And the Penn Central inquiry turns in large part, albeit not exclusively, upon the magnitude of a regulation’s economic impact and the degree to which it interferes with legitimate property interests.

B

In Agins v. City of Tiburon, a case involving a facial takings challenge to certain municipal zoning ordinances, the Court declared that “[t]he application of a general zoning law to particular property effects a taking if the ordinance does not substantially advance legitimate state interests, see Nectow v. Cambridge, 277 U. S. 183, 188 (1928), or denies an owner economically viable use of his land, see Penn Central Transp. Co. v. New York City, 438 U. S. 104, 138, n. 36 (1978).” 447 U. S., at 260. Because this statement is phrased in the disjunctive, Agins ’ “substantially advances” language has been read to announce a stand-alone regulatory takings test that is wholly independent of Penn Central or any other test. Indeed, the lower courts in this case struck down Hawaii’s rent control statute as an “unconstitutional regulatory taking,” 198 F. Supp. 2d, at 1193, based solely upon a finding that it does not substantially advance the State’s asserted interest in controlling retail gasoline prices. See supra, at 6–7. Although a number of our takings precedents have recited the “substantially advances” formula minted in Agins, this is our first opportunity to consider its validity as a freestanding takings test. We conclude that this formula prescribes an inquiry in the nature of a due process, not a takings, test, and that it has no proper place in our takings jurisprudence.

There is no question that the “substantially advances” formula was derived from due process, not takings, precedents. In support of this new language, Agins cited Nectow v. Cambridge, 277 U. S. 183, a 1928 case in which the plaintiff claimed that a city zoning ordinance “deprived him of his property without due process of law in contravention of the Fourteenth Amendment,” id., at 185. Agins then went on to discuss Village of Euclid v. Ambler Realty Co., 272 U. S. 365 (1926), a historic decision holding that a municipal zoning ordinance would survive a substantive due process challenge so long as it was not “clearly arbitrary and unreasonable, having no substantial relation to the public health, safety, morals, or general welfare.” Id., at 395 (emphasis added); see also Nectow, supra, at 188 (quoting the same “substantial relation” language from Euclid ).

When viewed in historical context, the Court’s reliance on Nectow and Euclid is understandable. Agins was the Court’s first case involving a challenge to zoning regulations in many decades, so it was natural to turn to these seminal zoning precedents for guidance. See Brief for United States as Amicus Curiae in Agins v. City of Tiburon, O. T. 1979, No. 602, pp. 12–13 (arguing that Euclid “set out the principles applicable to a determination of the facial validity of a zoning ordinance attacked as a violation of the Takings Clause of the Fifth Amendment”). Moreover, Agins ’ apparent commingling of due process and takings inquiries had some precedent in the Court’s then-recent decision in Penn Central. See 438 U. S., at 127 (stating in dicta that “[i]t is … implicit in Goldblatt [v. Hempstead, 369 U. S. 590 (1962),] that a use restriction on real property may constitute a ‘taking’ if not reasonably necessary to the effectuation of a substantial public purpose, see Nectow v. Cambridge, supra ”). But see Goldblatt, supra, at 594–595 (quoting “ ‘reasonably necessary’ ” language from Lawton v. Steele, 152 U. S. 133, 137 (1894), a due process case, and applying a deferential “ ‘reasonableness’ ” standard to determine whether a challenged regulation was a “valid exercise of the … police power” under the Due Process Clause). Finally, when Agins was decided, there had been some history of referring to deprivations of property without due process of law as “takings,” see, e.g., Rowan v. Post Office Dept., 397 U. S. 728, 740 (1970), and the Court had yet to clarify whether “regulatory takings” claims were properly cognizable under the Takings Clause or the Due Process Clause, see Williamson County Regional Planning Comm’n v. Hamilton Bank of Jefferson City, 473 U. S. 172, 197–199 (1985).

Although Agins ’ reliance on due process precedents is understandable, the language the Court selected was regrettably imprecise. The “substantially advances” formula suggests a means-ends test: It asks, in essence, whether a regulation of private property is effective in achieving some legitimate public purpose. An inquiry of this nature has some logic in the context of a due process challenge, for a regulation that fails to serve any legitimate governmental objective may be so arbitrary or irrational that it runs afoul of the Due Process Clause. See, e.g., County of Sacramento v. Lewis, 523 U. S. 833, 846 (1998) (stating that the Due Process Clause is intended, in part, to protect the individual against “the exercise of power without any reasonable justification in the service of a legitimate governmental objective”). But such a test is not a valid method of discerning whether private property has been “taken” for purposes of the Fifth Amendment.

In stark contrast to the three regulatory takings tests discussed above, the “substantially advances” inquiry reveals nothing about the magnitude or character of the burden a particular regulation imposes upon private property rights. Nor does it provide any information about how any regulatory burden is distributed among property owners. In consequence, this test does not help to identify those regulations whose effects are functionally comparable to government appropriation or invasion of private property; it is tethered neither to the text of the Takings Clause nor to the basic justification for allowing regulatory actions to be challenged under the Clause.

Chevron appeals to the general principle that the Takings Clause is meant “ ‘to bar Government from forcing some people alone to bear public burdens which, in all fairness and justice, should be borne by the public as a whole.’ ” Brief for Respondent 17–21 (quoting Armstrong, 364 U. S., at 49). But that appeal is clearly misplaced, for the reasons just indicated. A test that tells us nothing about the actual burden imposed on property rights, or how that burden is allocated cannot tell us when justice might require that the burden be spread among taxpayers through the payment of compensation. The owner of a property subject to a regulation that effectively serves a legitimate state interest may be just as singled out and just as burdened as the owner of a property subject to an ineffective regulation. It would make little sense to say that the second owner has suffered a taking while the first has not. Likewise, an ineffective regulation may not significantly burden property rights at all, and it may distribute any burden broadly and evenly among property owners. The notion that such a regulation nevertheless “takes” private property for public use merely by virtue of its ineffectiveness or foolishness is untenable.

Instead of addressing a challenged regulation’s effect on private property, the “substantially advances” inquiry probes the regulation’s underlying validity. But such an inquiry is logically prior to and distinct from the question whether a regulation effects a taking, for the Takings Clause presupposes that the government has acted in pursuit of a valid public purpose. The Clause expressly requires compensation where government takes private property “ for public use.” It does not bar government from interfering with property rights, but rather requires compensation “in the event of otherwise proper interference amounting to a taking.” First English Evangelical Lutheran Church, 482 U. S., at 315 (emphasis added). Conversely, if a government action is found to be impermissible—for instance because it fails to meet the “public use” requirement or is so arbitrary as to violate due process—that is the end of the inquiry. No amount of compensation can authorize such action.

Chevron’s challenge to the Hawaii statute in this case illustrates the flaws in the “substantially advances” theory. To begin with, it is unclear how significantly Hawaii’s rent cap actually burdens Chevron’s property rights. The parties stipulated below that the cap would reduce Chevron’s aggregate rental income on 11 of its 64 lessee-dealer stations by about $207,000 per year, but that Chevron nevertheless expects to receive a return on its investment in these stations that satisfies any constitutional standard. See supra, at 4. Moreover, Chevron asserted below, and the District Court found, that Chevron would recoup any reductions in its rental income by raising wholesale gasoline prices. See supra, at 5. In short, Chevron has not clearly argued—let alone established—that it has been singled out to bear any particularly severe regulatory burden. Rather, the gravamen of Chevron’s claim is simply that Hawaii’s rent cap will not actually serve the State’s legitimate interest in protecting consumers against high gasoline prices. Whatever the merits of that claim, it does not sound under the Takings Clause. Chevron plainly does not seek compensation for a taking of its property for a legitimate public use, but rather an injunction against the enforcement of a regulation that it alleges to be fundamentally arbitrary and irrational.

Finally, the “substantially advances” formula is not only doctrinally untenable as a takings test—its application as such would also present serious practical difficulties. The Agins formula can be read to demand heightened means-ends review of virtually any regulation of private property. If so interpreted, it would require courts to scrutinize the efficacy of a vast array of state and federal regulations—a task for which courts are not well suited. Moreover, it would empower—and might often require—courts to substitute their predictive judgments for those of elected legislatures and expert agencies.

Although the instant case is only the tip of the proverbial iceberg, it foreshadows the hazards of placing courts in this role. To resolve Chevron’s takings claim, the District Court was required to choose between the views of two opposing economists as to whether Hawaii’s rent control statute would help to prevent concentration and supracompetitive prices in the State’s retail gasoline market. Finding one expert to be “more persuasive” than the other, the court concluded that the Hawaii Legislature’s chosen regulatory strategy would not actually achieve its objectives. See 198 F. Supp. 2d, at 1187–1193. Along the way, the court determined that the State was not entitled to enact a prophylactic rent cap without actual evidence that oil companies had charged, or would charge, excessive rents. See id., at 1191 . Based on these findings, the District Court enjoined further enforcement of Act 257’s rent cap provision against Chevron. We find the proceedings below remarkable, to say the least, given that we have long eschewed such heightened scrutiny when addressing substantive due process challenges to government regulation. See, e.g., Exxon Corp. v. Governor of Maryland, 437 U. S. 117, 124–125 (1978); Ferguson v. Skrupa, 372 U. S. 726, 730–732 (1963). The reasons for deference to legislative judgments about the need for, and likely effectiveness of, regulatory actions are by now well established, and we think they are no less applicable here.

For the foregoing reasons, we conclude that the “substantially advances” formula announced in Agins is not a valid method of identifying regulatory takings for which the Fifth Amendment requires just compensation. Since Chevron argued only a “substantially advances” theory in support of its takings claim, it was not entitled to summary judgment on that claim.

III

We emphasize that our holding today—that the “substantially advances” formula is not a valid takings test—does not require us to disturb any of our prior holdings. To be sure, we applied a “substantially advances” inquiry in Agins itself, see 447 U. S., at 261–262 (finding that the challenged zoning ordinances “substantially advance[d] legitimate governmental goals”), and arguably also in Keystone Bituminous Coal Assn. v. DeBenedictis, 480 U. S. 470, 485–492 (1987) (quoting “ ‘substantially advance[s]’ ” language and then finding that the challenged statute was intended to further a substantial public interest). But in no case have we found a compensable taking based on such an inquiry. Indeed, in most of the cases reciting the “substantially advances” formula, the Court has merely assumed its validity when referring to it in dicta. See Tahoe-Sierra Preservation Council, Inc. v. Tahoe Regional Planning Agency, 535 U. S. 302, 334 (2002); Del Monte Dunes, 526 U. S., at 704; Lucas, 505 U. S., at 1016; Yee v. Escondido, 503 U. S. 519, 534 (1992); United States v. Riverside Bayview Homes, Inc., 474 U. S. 121, 126 (1985).

It might be argued that this formula played a role in our decisions in Nollan v. California Coastal Comm’n, 483 U. S. 825 (1987), and Dolan v. City of Tigard, 512 U. S. 374 (1994). See Brief for Respondent 21–23. But while the Court drew upon the language of Agins in these cases, it did not apply the “substantially advances” test that is the subject of today’s decision. Both Nollan and Dolan involved Fifth Amendment takings challenges to adjudicative land-use exactions—specifically, government demands that a landowner dedicate an easement allowing public access to her property as a condition of obtaining a development permit. See Dolan, supra, at 379–380 (permit to expand a store and parking lot conditioned on the dedication of a portion of the relevant property for a “greenway,” including a bike/pedestrian path); Nollan, supra, at 828 (permit to build a larger residence on beachfront property conditioned on dedication of an easement allowing the public to traverse a strip of the property between the owner’s seawall and the mean high-tide line).

In each case, the Court began with the premise that, had the government simply appropriated the easement in question, this would have been a per se physical taking. Dolan, supra, at 384; Nollan, supra, at 831–832. The question was whether the government could, without paying the compensation that would otherwise be required upon effecting such a taking, demand the easement as a condition for granting a development permit the government was entitled to deny. The Court in Nolan answered in the affirmative, provided that the exaction would substantially advance the same government interest that would furnish a valid ground for denial of the permit. 483 U. S., at 834–837. The Court further refined this requirement in Dolan, holding that an adjudicative exaction requiring dedication of private property must also be “ ‘rough[ly] proportiona[l]’ … both in nature and extent to the impact of the proposed development.” 512 U. S., at 391; see also Del Monte Dunes, supra, at 702 (emphasizing that we have not extended this standard “beyond the special context of [such] exactions”).

Although Nollan and Dolan quoted Agins ’ language, see Dolan, supra, at 385; Nollan, supra, at 834, the rule those decisions established is entirely distinct from the “substantially advances” test we address today. Whereas the “substantially advances” inquiry before us now is unconcerned with the degree or type of burden a regulation places upon property, Nollan and Dolan both involved dedications of property so onerous that, outside the exactions context, they would be deemed per se physical takings. In neither case did the Court question whether the exaction would substantially advance some legitimate state interest. See Dolan, supra, at 387–388; Nollan, supra, at 841. Rather, the issue was whether the exactions substantially advanced the same interests that land-use authorities asserted would allow them to deny the permit altogether. As the Court explained in Dolan, these cases involve a special application of the “doctrine of ‘unconstitutional conditions,’ ” which provides that “the government may not require a person to give up a constitutional right—here the right to receive just compensation when property is taken for a public use—in exchange for a discretionary benefit conferred by the government where the benefit has little or no relationship to the property.” 512 U. S., at 385. That is worlds apart from a rule that says a regulation affecting property constitutes a taking on its face solely because it does not substantially advance a legitimate government interest. In short, Nollan and Dolan cannot be characterized as applying the “substantially advances” test we address today, and our decision should not be read to disturb these precedents.

***

Twenty-five years ago, the Court posited that a regulation of private property “effects a taking if [it] does not substantially advance [a] legitimate state interes[t].” Agins, supra, at 260. The lower courts in this case took that statement to its logical conclusion, and in so doing, revealed its imprecision. Today we correct course. We hold that the “substantially advances” formula is not a valid takings test, and indeed conclude that it has no proper place in our takings jurisprudence. In so doing, we reaffirm that a plaintiff seeking to challenge a government regulation as an uncompensated taking of private property may proceed under one of the other theories discussed above—by alleging a “physical” taking, a Lucas -type “total regulatory taking,” a Penn Central taking, or a land-use exaction violating the standards set forth in Nollan and Dolan. Because Chevron argued only a “substantially advances” theory in support of its takings claim, it was not entitled to summary judgment on that claim. Accordingly, we reverse the judgment of the Ninth Circuit and remand the case for further proceedings consistent with this opinion.

It is so ordered.