with whom Justice O’Connor and Justice Souter join, concurring.
The Alabama state courts have assessed the defendant $2 million in “punitive damages” for having knowingly failed to tell a BMW automobile buyer that, at a cost of $600, it had repainted portions of his new $40,000 car, thereby lowering its potential resale value by about 10%. The Court’s opinion, which I join, explains why we have concluded that this award, in this case, was “grossly excessive” in relation to legitimate punitive damages objectives, and hence an arbitrary deprivation of life, liberty, or property in violation of the Due Process Clause. See TXO Production Corp. v. Alliance Resources Corp., 509 U. S. 443, 453, 454 (1993) (A “grossly excessive” punitive award amounts to an “arbitrary deprivation of property without due process of law”) (plurality opinion). Members of this Court have generally thought, however, that if “fair procedures were followed, a judgment that is a product of that process is entitled to a strong pre*587sumption of validity.” Id., at 457. See also Pacific Mut. Life Ins. Co. v. Haslip, 499 U. S. 1, 40-42 (1991) (Kennedy, J., concurring in judgment). And the Court also has found that punitive damages procedures very similar to those followed here were not, by themselves, fundamentally unfair. Id., at 15-24. Thus, I believe it important to explain why this presumption of validity is overcome in this instance.
The reason flows from the Court’s emphasis in Haslip upon the constitutional importance of legal standards that provide “reasonable constraints” within which “discretion is exercised,” that assure “meaningful and adequate review by the trial court whenever a jury has fixed the punitive damages,” and permit “appellate review [that] makes certain that the punitive damages. are reasonable in their amount and rational in light of their purpose to punish what has occurred and to deter its repetition.” Id., at 20-21. See also id., at 18 (“[UJnlimited jury discretion — or unlimited judicial discretion for that matter — in the fixing of punitive damages may invite extreme results that jar one’s constitutional sensibilities”).
This constitutional concern, itself harkening back to the Magna Carta, arises out of the basic unfairness of depriving citizens of life, liberty, or property, through the application, not of law and legal processes, but of arbitrary coercion. Daniels v. Williams, 474 U. S. 327, 331 (1986); Dent v. West Virginia, 129 U. S. 114, 123 (1889). Requiring the application of law, rather than a decisionmaker’s caprice, does more than simply provide citizens notice of what actions may subject them to punishment; it also helps to assure the uniform general treatment of similarly situated persons that is the essence of law itself. See Railway Express Agency, Inc. v. New York, 336 U. S. 106, 112 (1949) (Jackson, J., concurring) (“[T]here is no more effective practical guaranty against arbitrary and unreasonable government than to require that the principles of law which officials would impose upon a minority must be imposed generally”).
*588Legal standards need not be precise in order to satisfy this constitutional concern. See Haslip, supra, at 20 (comparing punitive damages standards to such legal standards as “reasonable care,” “due diligence,” and “best interests of the child”) (internal quotation marks omitted). But they must offer some kind of constraint upon a jury or court’s discretion, and thus protection against purely arbitrary behavior. The standards the Alabama courts applied here are vague and open ended to the point where they risk arbitrary results. In my view, although the vagueness of those standards does not, by itself, violate due process, see Haslip, supra, it does invite the kind of scrutiny the Court has given the particular verdict before us. See id., at 18 (“[Concerns of . . . adequate guidance from the court when the case is tried to a jury properly enter into the constitutional calculus”); TXO, supra, at 475 (“[I]t cannot be denied that the lack of clear guidance heightens the risk that arbitrariness, passion, or bias will replace dispassionate deliberation as the basis for the jury’s verdict”) (O’Connor, J., dissenting). This is because the standards, as the Alabama Supreme Court authoritatively interpreted them here, provided no significant constraints or protection against arbitrary results.
First, the Alabama statute that permits punitive damages does not itself contain a standard that readily distinguishes between conduct warranting very small, and conduct warranting very large, punitive damages awards. That statute permits punitive damages in cases of “oppression, fraud, wantonness, or malice.” Ala. Code § 6-ll-20(a) (1993). But the statute goes on to define those terms broadly, to encompass far more than the egregious conduct that those terms, at first reading, might seem to imply. An intentional misrepresentation, made through a statement or silence, can easily amount to “fraud” sufficient to warrant punitive damages. See §6-ll-20(b)(l) (“Fraud” includes “intentional... concealment of a material fact the concealing party had a *589duty to disclose, which was gross, oppressive, or malicious and committed with the intention ... of thereby depriving a person or entity of property”) (emphasis added); §6-11-20(b)(2) (“Malice” includes any “wrongful act without just cause or excuse . . . [w]ith an intent to injure the . . . property of another”) (emphasis added); §6-ll-20(b)(5) (“Oppression” includes “[subjecting a person to . . . unjust hardship in conscious disregard of that person’s rights”). The statute thereby authorizes punitive damages for the most serious kinds of misrepresentations, say, tricking the elderly out of their life savings, for much less serious conduct, such as the failure to disclose repainting a car, at issue here, and for a vast range of conduct in between.
Second, the Alabama courts, in this case, have applied the “factors” intended to constrain punitive damages awards in a way that belies that purpose. Green Oil Co. v. Hornsby, 539 So. 2d 218 (Ala. 1989), sets forth seven factors that appellate courts use to determine whether or not a jury award was “grossly excessive” and which, in principle, might make up for the lack of significant constraint in the statute. But, as the Alabama courts have authoritatively interpreted them, and as their application in this case illustrates, they impose little actual constraint.
(a) Green Oil requires that a punitive damages award “bear a reasonable relationship to the harm that is likely to occur from the defendant’s conduct as well as to the harm that actually has occurred.” Id., at 223. But this standard does little to guide a determination of what counts as a “reasonable” relationship, as this case illustrates. The record evidence of past, present, or likely future harm consists of (a) $4,000 of harm to Dr. Gore’s BMW; (b) 13 other similar Alabama instances; and (c) references to about 1,000 similar instances in other States. The Alabama Supreme Court, disregarding BMW’s failure to make relevant objection to the out-of-state instances at trial (as was the court’s right), held that the last mentioned, out-of-state instances did not *590count as relevant harm. It went on to find “a reasonable relationship” between the harm and the $2 million punitive damages award without “considering] those acts that occurred in other jurisdictions.” 646 So. 2d 619, 628 (1994) (emphasis added). For reasons explored by the majority in greater depth, see ante, at 574-586, the relationship between this award and the underlying conduct seems well beyond the bounds of the “reasonable.” To find a “reasonable relationship” between purely economic harm totaling $56,000, without significant evidence of future repetition, and a punitive award of $2 million is to empty the “reasonable relationship” test of meaningful content. As thus construed, it does not set forth a legal standard that could have significantly constrained the discretion of Alabama factfinders.
(b) Green Oil’s second factor is the “degree of reprehensibility” of the defendant’s conduct. Green Oil, supra, at 223. Like the “reasonable relationship” test, this factor provides little guidance on how to relate culpability to the size of an award. The Alabama court, in considering this factor, found “reprehensible” that BMW followed a conscious policy of not disclosing repairs to new cars when the cost of repairs amounted to less than 3% of the car’s value. Of course, any conscious policy of not disclosing a repair — where one knows the nondisclosure might cost the customer resale value — is “reprehensible” to some degree. But, for the reasons discussed by the majority, ante, at 575-580, I do not see how the Alabama courts could find conduct that (they assumed) caused $56,000 of relevant economic harm especially or unusually reprehensible enough to warrant $2 million in punitive damages, or a significant portion of that award. To find to the contrary, as the Alabama courts did, is not simply unreasonable; it is to make “reprehensibility” a concept without constraining force, i. e., to deprive the concept of its constraining power to protect against serious and capricious deprivations.
*591(c) Green Oil’s third factor requires “punitive damages” to “remove the profit” of the illegal activity and “be in excess of the profit, so that the defendant recognizes a loss.” Green Oil, 539 So. 2d, at 223. This factor has the ability to limit awards to a fixed, rational amount. But as applied, that concept’s potential was not realized, for the court did not limit the award to anywhere near the $56,000 in profits evidenced in the record. Given the record’s description of the conduct and its prevalence, this factor could not justify much of the $2 million award.
(d) Green Oil’s fourth factor is the “financial position” of the defendant. Ibid. Since a fixed dollar award will punish a poor person more than a wealthy one, one can understand the relevance of this factor to the State’s interest in retribution (though not necessarily to its interest in deterrence, given the more distant relation between a defendant’s wealth and its responses to economic incentives). See TXO, 509 U. S., at 462, and n. 28 (plurality opinion); id., at 469 (Kennedy, J., concurring in part and concurring in judgment); Haslip, 499 U. S., at 21-22; Browning-Ferris Industries of Vt., Inc. v. Kelco Disposal, Inc., 492 U. S. 257, 300 (1989) (O’Connor, J., concurring in part and dissenting in part). This factor, however, is not necessarily intended to act as a significant constraint on punitive awards. Rather, it provides an open-ended basis for inflating awards when the defendant is wealthy, as this case may illustrate. That does not make its use unlawful or inappropriate; it simply means that this factor cannot make up for the failure of other factors, such as “reprehensibility,” to constrain significantly an award that purports to punish a defendant’s conduct.
(e) Green Oil’s fifth factor is the “costs of litigation” and the State’s desire “to encourage plaintiffs to bring wrongdoers to trial.” 539 So. 2d, at 223. This standard provides meaningful constraint to the extent that the enhancement it authorized is linked to a fixed, ascertainable amount approximating actual costs, even when defined generously to reflect *592the contingent nature of plaintiffs’ victories. But as this case shows, the factor cannot operate as a constraint when an award much in excess of costs is approved for other reasons. An additional aspect of the standard — the need to “encourage plaintiffs to bring wrongdoers to trial” — is a factor that does not constrain, but enhances, discretionary-power — especially when unsupported by evidence of a special need to encourage litigation (which the Alabama courts here did not mention).
(f) Green Oil’s sixth factor is whether or not “criminal sanctions have been imposed on the defendant for his conduct.” Ibid. This factor did not apply here.
(g) Green Oil’s seventh factor requires that “other civil actions” filed “against the same defendant, based on the same conduct,” be considered in mitigation. Id., at 224. That factor did not apply here.
Thus, the first, second, and third Green Oil factors, in principle, might sometimes act as constraints on arbitrary behavior. But as the Alabama courts interpreted those standards in this case, even taking those three factors together, they could not have significantly constrained the court system’s ability to impose “grossly excessive” awards.
Third, the state courts neither referred to, nor made any effort to find, nor enunciated any other standard that either directly, or indirectly as background, might have supplied the constraining legal force that the statute and Green Oil standards (as interpreted here) lack. Dr. Gore did argue to the jury an economic theory based on the need to offset the totality of the harm that the defendant’s conduct caused. Some theory of that general kind might have provided a significant constraint on arbitrary awards (at least where confined to the relevant harm-causing conduct, see ante, at 570-574). Some economists, for example, have argued for a standard that would deter illegal activity causing solely economic harm through the use of punitive damages awards that, as a whole, would take from a wrongdoer the total cost of the *593harm caused. See, e.g., S. Shavell, Economic Analysis of Accident Law 162 (1987) (“If liability equals losses caused multiplied by . . . the inverse of the probability of suit, insurers will act optimally under liability rules despite the chance that they will escape suit”); Cooter, Punitive Damages for Deterrence: When and How Much, 40 Ala. L. Rev. 1143, 1146-1148 (1989). My understanding of the intuitive essence of some of those theories, which I put in crude form (leaving out various qualifications), is that they could permit juries to calculate punitive damages by making a rough estimate of global harm, dividing that estimate by a similarly rough estimate of the number of successful lawsuits that would likely be brought, and adding generous attorney’s fees and other costs. Smaller damages would not sufficiently discourage firms from engaging in the harmful conduct, while larger damages would “over-deter” by leading potential defendants to spend more to prevent the activity that causes the economic harm, say, through employee training, than the cost of the harm itself. See Galligan, Augmented Awards: The Efficient Evolution of Punitive Damages, 51 La. L. Rev. 3, 17-20, 28-30 (1990). Larger damages might also “double count” by including in the punitive damages award some of the compensatory, or punitive, damages that subsequent plaintiffs would also recover.
The record before us, however, contains nothing suggesting that the Alabama Supreme Court, when determining the allowable award, applied any “economic” theory that might explain the $2 million recovery. Cf. Browning-Ferris, supra, at 300 (noting that the Constitution “does not incorporate the views of the Law and Economics School,” nor does it “ 'require the States to subscribe to any particular economic theory’ ”) (O’Connor, J., concurring in part and dissenting in part) (quoting CTS Corp. v. Dynamics Corp. of America, 481 U. S. 69, 92 (1987)). And courts properly tend to judge the rationality of judicial actions in terms of the reasons that were given, and the facts that were before the court, cf. TXO, *594509 U. S., at 468 (Kennedy, J., concurring in part and concurring in judgment), not those that might have been given on the basis of some conceivable set of facts (unlike the rationality of economic statutes enacted by legislatures subject to the public’s control through the ballot box, see, e. g., FCC v. Beach Communications, Inc., 508 U. S. 307, 315 (1993)). Therefore, reference to a constraining “economic” theory, which might have counseled more deferential review by this Court, is lacking in this case.
Fourth, I cannot find any community understanding or historic practice that this award might exemplify and which, therefore, would provide background standards constraining arbitrary behavior and excessive awards. A punitive damages award of $2 million for intentional misrepresentation causing $56,000 of harm is extraordinary by historical standards, and, as far as I am aware, finds no analogue until relatively recent times. Amici for Dr. Gore attempt to show that this is not true, pointing to various historical cases which, according to their calculations, represented roughly equivalent punitive awards for similarly culpable conduct. See Brief for James D. A. Boyle et al. as Amici Curiae 4-5 (hereinafter Legal Historians’ Brief). Among others, they cite Wilkes v. Wood, Lofft 1, 98 Eng. Rep. 489 (C. R 1763) (£1,000 said to be equivalent of $1.5 million, for warrantless search of papers); Huckle v. Money, 2 Wills. 205, 95 Eng. Rep. 768 (K. B. 1763) (£300, said to be $450,000, for 6-hour false imprisonment); Hewlett v. Cruchley, 5 Taunt. 277, 128 Eng. Rep. 696 (C. P. 1813) (£2,000, said to be $680,000, for malicious prosecution); Merest v. Harvey, 5 Taunt. 442, 128 Eng. Rep. 761 (C. P. 1814) (£500, said to be $165,000, for poaching). But amici apparently base their conversions on a mathematical assumption, namely, that inflation has progressed at a constant 3% rate of inflation. See Legal Historians’ Brief 4. In fact, consistent, cumulative inflation is a modern phenomenon. See McCusker, How Much Is That in Real Money? A Historical Price Index for Use as a Deflator *595of Money Values in the Economy of the United States, 101 Proceedings of American Antiquarian Society 297, 310, 323-332 (1992). Estimates based on historical rates of valuation, while highly approximate, suggest that the ancient extraordinary awards are small compared to the $2 million here at issue, or other modern punitive damages figures. See Appendix to this opinion, infra, at 597-598 (suggesting that the modern equivalent of the awards in the above cases is something like $150,000, $45,000, $100,000, and $25,000, respectively). And, as the majority opinion makes clear, the record contains nothing to suggest that the extraordinary size of the award in this case is explained by the extraordinary wrongfulness of the defendant’s behavior, measured by historical or community standards, rather than arbitrariness or caprice.
Fifth, there are no other legislative enactments here that classify awards and impose quantitative limits that would significantly cabin the fairly unbounded discretion created by the absence of constraining legal standards. Cf., e. g., Tex. Civ. Prac. & Rem. Code Ann. §41.008 (Supp. 1996) (punitive damages generally limited to greater of double damages, or $200,000, except cap does not apply to suits arising from certain serious criminal acts enumerated in the statute); Conn. Gen. Stat. § 52-240b (1995) (punitive damages may not exceed double compensatory damages in product liability cases); Fla. Stat. §768.73(1) (Supp. 1993) (punitive damages in certain actions limited to treble compensatory damages); Ga. Code Ann. §51-12-5.1(g) (Supp. 1995) ($250,000 cap in certain actions).
The upshot is that the rules that purport to channel discretion in this kind of case, here did not do so in fact. That means that the award in this case was both (a) the product of a system of standards that did not significantly constrain a court’s, and hence a jury’s, discretion in making that award; and (b) grossly excessive in light of the State’s legitimate punitive damages objectives.
*596The first of these reasons has special importance where courts review a jury-determined punitive damages award. That is because one cannot expect to direct jurors like legislators through the ballot box; nor can one expect those jurors to interpret law like judges, who work within a discipline and hierarchical organization that normally promotes roughly uniform interpretation and application of the law. Yet here Alabama expects jurors to act, at least a little, like legislators or judges, for it permits them, to a certain extent, to create public policy and to apply that policy, not to compensate a victim, but to achieve a policy-related objective outside the confines of the particular case.
To the extent that neither clear legal principles nor fairly obvious historical or community-based standards (defining, say, especially egregious behavior) significantly constrain punitive damages awards, is there not a substantial risk of outcomes so arbitrary that they become difficult to square with the Constitution’s assurance, to every citizen, of the law’s protection? The standards here, as authoritatively interpreted, in my view, make this threat real and not theoretical. And, in these unusual circumstances, where legal standards offer virtually no constraint, I believe that this lack of constraining standards warrants this Court’s detailed examination of the award.
The second reason — the severe disproportionality between the award and the legitimate punitive damages objectives— reflects a judgment about a matter of degree. I recognize that it is often difficult to determine just when a punitive award exceeds an amount reasonably related to a State’s legitimate interests, or when that excess is so great as to amount to a matter of constitutional concern. Yet whatever the difficulties of drawing a precise line, once we examine the award in this case, it is not difficult to say that this award lies on the line’s far side. The severe lack of proportionality between the size of the award and the underlying punitive damages objectives shows that the award falls into the cate*597gory of “gross excessiveness” set forth in this Court’s prior cases.
These two reasons taken together overcome what would otherwise amount to a “strong presumption of validity.” TXO, 509 U. S., at 457. And, for those two reasons, I conclude that the award in this unusual case violates the basic guarantee of nonarbitrary governmental behavior that the Due Process Clause provides.
APPENDIX TO OPINION OF BREYER, J.
Although I recognize that all estimates of historic rates of inflation are subject to dispute, including, I assume, the sources below, those sources suggest that the value of the 18th and 19th century judgments cited by amici is much less than the figures amici arrived at under their presumption of a constant 3% rate of inflation.
In 1763, £1 (Eng.) was worth £1.73 Pennsylvania currency. See U. S. Bureau of the Census, Historical Statistics of the United States: Colonial Times to 1970, Series Z-585, p. 1198 (Bicentennial ed. 1975). For the period 1766-1772, £1 (Penn.) was worth $45.99 (U. S. 1991). See McCusker, How Much Is That in Real Money? A Historical Price Index for Use as a Deflator of Money Values in the Economy of the United States, 101 American Antiquarian Society 297, 333 (1992). Thus, £1 (Eng. 1763) is worth about $79.56 (U. S. 1991). Accounting for the 12% inflation of the U. S. dollar between 1991 and 1995 (when amici filed their brief), see Economic Indicators, 104th Cong., 2d Sess., p. 23 (Feb. 1996), £1 (Eng. 1763) is worth about $89.11 (U. S. 1995).
Calculated another way, £1 (Eng. 1763) is worth about £72.84 (Eng. 1991). See McCusker, supra, at 312, 342, 350. And £1 (Eng. 1991) is worth $1.77 (U. S. 1991). See 78 Fed. Reserve Bulletin A68 (Feb. 1992). Thus, £1 (Eng. 1763) amounts to about $128.93 (U. S. 1991). Again, accounting for inflation between 1991 and 1995, this amounts to about $144.40 (U. S. 1995).
*598Thus, the above sources suggest that the £1,000 award in Wilkes in 1763 roughly amounts to between $89,110 and $144,440 today, not $1.5 million. And the £300 award in Huckle that same year would seem to be worth between $26,733 and $43,320 today, not $450,000.
For the period of the Hewlett and Merest decisions, £1 (Eng. 1813) is worth about £25.3 (Eng. 1991). See McCusker, supra, at 344, 350. Using the 1991 exchange rate, £1 (Eng. 1813) is worth about $44.78 (U. S. 1991). Accounting for inflation between 1991 and 1995, this amounts to about $50.16 (U. S. 1995).
Thus, the £2,000 and £500 awards in Hewlett and Merest would seem to be closer to $100,320 and $25,080, respectively, than to amici’s estimates of $680,000 and $165,000.