dissenting.
Punitive damages are a powerful weapon. Imposed wisely and with restraint, they have the potential to advance legitimate state interests. Imposed indiscriminately, however, they have a devastating potential for harm. Regrettably, common-law procedures for awarding punitive damages fall into the latter category. States routinely authorize civil juries to impose punitive damages without providing them any meaningful instructions on how to do so. Rarely is a jury told anything more specific than “do what you think best.” See Browning-Ferris Industries of Vt., Inc. v. Kelco Disposal, Inc., 492 U. S. 257, 281 (1989) (Brennan, J., concurring).
*43In my view, such instructions are so fraught with uncertainty that they defy rational implementation. Instead, they encourage inconsistent and unpredictable results by inviting juries to rely on private beliefs and personal predilections. Juries are permitted to target unpopular defendants, penalize unorthodox or controversial views, and redistribute wealth. Multimillion dollar losses are inflicted on a whim. While I do not question the general legitimacy of punitive damages, I see a strong need to provide juries with standards to constrain their discretion so that they may exercise their power wisely, not capriciously or maliciously. The Constitution requires as much.
The Court today acknowledges that dangers may lurk, but holds that they did not materialize in this case. See ante, at 18-24. They did materialize, however. They always do, because such dangers are part and parcel of common-law punitive damages procedures. As is typical, the trial court’s instructions in this case provided no meaningful standards to guide the jury’s decision to impose punitive damages or to fix the amount. Accordingly, these instructions were void for vagueness. Even if the Court disagrees with me on this point, it should still find that Pacific Mutual was denied procedural due process. Whether or not the jury instructions were so vague as to be unconstitutional, they plainly offered less guidance than is required under the due process test set out in Mathews v. Eldridge, 424 U. S. 319, 335 (1976). The most modest of procedural safeguards would have made the process substantially more rational without impairing any legitimate governmental interest. The Court relies heavily on the State’s mechanism for postverdict judicial review, ante, at 20-23, but this is incapable of curing a grant of standard-less discretion to the jury. Post hoc review tests only the amount of the award, not the procedures by which that amount was determined. Alabama’s common-law scheme is so lacking in fundamental fairness that the propriety of any specific award is irrelevant. Any award of punitive damages *44rendered under these procedures, no matter how small the amount, is constitutionally infirm.
Notwithstanding its recognition of serious due process concerns, the Court upholds Alabama’s punitive damages scheme. Unfortunately, Alabama’s punitive damages scheme is indistinguishable from the common-law schemes employed by many States. The Court’s holding will therefore substantially impede punitive damages reforms. Because I am concerned that the Court today sends the wrong signal, I respectfully dissent.
I
Due process requires that a State provide meaningful standards to guide the application of its laws. See Kolender v. Lawson, 461 U. S. 352, 358 (1983). A state law that lacks such standards is void for vagueness. The void-for-vagueness doctrine applies not only to laws that proscribe conduct, but also to laws that vest standardless discretion in the jury to fix a penalty. See United States v. Batchelder, 442 U. S. 114, 123 (1979). I have no trouble concluding that Alabama’s common-law scheme for imposing punitive damages is void for vagueness.
A
Alabama’s punitive damages scheme requires a jury to make two decisions: (1) whether or not to impose punitive damages against the defendant, and (2) if so, in what amount. On the threshold question of whether or not to impose punitive damages, the trial court instructed the jury as follows: “Imposition of punitive damages is entirely discretionary with the jury, that means you don’t have to award it unless this jury feels that you should do so.” App. 105-106 (emphasis added).
This instruction is as vague as any I can imagine. It speaks of discretion, but suggests no criteria on which to base the exercise of that discretion. Instead of reminding the jury that its decision must rest on a factual or legal predicate, the instruction suggests that the jury may do whatever *45it “feels” like. It thus invites individual jurors to rely upon emotion, bias, and personal predilections of every sort. As I read the instruction, it as much permits a determination based upon the toss of a coin or the color of the defendant’s skin as upon a reasoned analysis of the offensive conduct. This is not “discretion in the legal sense of that term, but. . . mere will. It is purely arbitrary and acknowledges neither guidance nor restraint.” Yick Wo v. Hopkins, 118 U. S. 356, 366-367 (1886).
Giaccio v. Pennsylvania, 382 U. S. 399 (1966), offers a compelling analogy. At issue in Giaccio was a statute that left to the discretion of the jury whether or not to assess costs against an acquitted criminal defendant. The statute did not set out any standards to guide the jury’s determination. Id., at 401. The Court did not hesitate in striking down the statute on vagueness grounds. Id., at 402. It reasoned that the utter lack of standards subjected acquitted defendants to “arbitrary and discriminatory impositions of costs.” Ibid.' Justice Black wrote for the Court:
“The Act, without imposing a single condition, limitation or contingency on a jury which has acquitted a defendant simply says the jurors ‘shall determine, by their verdict, whether . . . the defendant, shall pay the costs’ .... Certainly one of the basic purposes of the Due Process Clause has always been to protect a person against having the Government impose burdens upon him except in accordance with the valid laws of the land. Implicit in this constitutional safeguard is the premise that the law must be one that carries an understandable meaning with legal standards that courts must enforce. This state Act as written does not even begin to meet this constitutional requirement.” Id., at 403.
Alabama’s common-law punitive damages scheme fails for precisely the same reason. It permits a jury to decide whether or not to impose punitive damages “without imposing a single condition, limitation or contingency” on the jury. *46Ibid. The State offers no principled basis for distinguishing those tortfeasors who should be liable for punitive damages from those who should not be liable. Instead, the State delegates this basic policy matter to individual juries “for resolution on an ad hoc and subjective basis, with the attendant dangers of arbitrary and discriminatory application.” Grayned v. City of Rockford, 408 U. S. 104, 108-109 (1972). As in Giaccio, this grant of unchanneled, standardless discretion “does not even begin to meet th[e] constitutional requirement.” Giaccio, 382 U. S., at 403.
The vagueness question is not even close. This is not a case where a State has ostensibly provided a standard to guide the jury’s discretion. Alabama, making no pretensions whatsoever, gives civil juries complete, unfettered, and unchanneled discretion to determine whether or not to impose punitive damages. Not only that, the State tells the jury that it has complete discretion. This is a textbook example of the void-for-vagueness doctrine. Alabama’s common-law scheme is unconstitutionally vague because the State entrusts the jury with “such broad and unlimited power . . . that the jurors must make determinations of the crucial issue upon their notions of what the law should be instead of what it is.” Ibid.
If anything, this is an easier case than Giaccio. There, the Court struck down on vagueness grounds a Pennsylvania law, under which the monetary penalty that could be assessed by the jury against the defendant was limited to the costs of prosecution — in that case, $230.95. Id., at 400. Our scrutiny under the vagueness doctrine intensifies, however, in proportion to the severity of the penalty imposed, see Hoffman Estates v. The Flipside, Hoffman Estates, Inc., 455 U. S. 489, 498-499 (1982), and Alabama’s punitive damages scheme places no substantive limits on the amount of a jury’s award. Pacific Mutual was found liable for punitive damages of $840,000. Ante, at 7, n. 2. Even this substantial sum pales by comparison to others handed down by juries *47in the State. See App. to Brief for Alabama Defense Lawyers Association as Amicus Curiae la-19a (listing Alabama jury verdicts including punitive damages awards as high as $10 million, $25 million, and $50 million).
It is no defense to vagueness that this case concerns a jury instruction rather than a statute. The constitutional prohibition against vagueness does not disappear simply because the state law at issue originated in the courts rather than the legislature. “[I]f anything, our scrutiny of awards made without the benefit of a legislature’s deliberation and guidance would be less indulgent than our consideration of those that fall within statutory limits.” Browning-Ferris, 492 U. S., at 281 (Brennan, J., concurring). See ante, at 20-22. Moreover, the instruction in this case was not an aberration. It tracked virtually word for word Alabama’s Pattern Jury Instruction on punitive damages. See Alabama Pattern Jury Instructions, Civil 11.03 (1974).
Nor does it matter that punitive damages are imposed by civil juries rather than criminal courts. The vagueness doctrine is not limited to criminal penalties. See Hoffman Estates, supra; City of Mesquite v. Aladdin’s Castle, Inc., 455 U. S. 283 (1982). The Court in Giaccio expressly repudiated this distinction:
“Both liberty and property are specifically protected by the Fourteenth Amendment against any state deprivation which does not meet the standards of due process, and this protection is not to be avoided by the simple label a State chooses to fasten upon its conduct or its statute. So here this state Act whether labeled ‘penal’ or not must meet the challenge that it is unconstitutionally vague.” 382 U. S., at 402.
Here, as in Giaccio, the civil/criminal distinction is blurry. Unlike compensatory damages, which are purely civil in character, punitive damages are, by definition, punishment. They operate as “private fines levied by civil juries” to advance governmental objectives. Gertz v. Robert Welch, Inc., *48418 U. S. 323, 350 (1974). Because Alabama permits juries to inflict these potentially devastating penalties wholly at random, the State scheme is void for vagueness.
B
If an Alabama jury determines that punitive damages are appropriate in a particular case, it must then fix the amount. Here, the trial court instructed the jury: “Should you award punitive damages, in fixing the amount, you must take into consideration the character and the degree of the wrong as shown by the evidence and [the] necessity of preventing similar wrong.” App. 106.
The Court concludes that this instruction sufficiently limited the jury's discretion, ante, at 19-20, but I cannot share this conclusion. Although the instruction ostensibly provided some guidance, this appearance is deceiving. As Justice Brennan said of a similar instruction: “Guidance like this is scarcely better than no guidance at all. I do not suggest that the instruction itself was in error; indeed, it appears to have been a correct statement of [state] law. The point is, rather, that the instruction reveals a deeper flaw: the fact that punitive damages are imposed by juries guided by little more than an admonition to do what they think is best.” Browning-Ferris, supra, at 281 (concurring opinion). I agree wholeheartedly. Vague references to “the character and the degree of the wrong” and the “necessity of preventing similar wrong” do not assist a jury in making a reasoned decision; they are too amorphous. They restate the overarching principles of punitive damages awards — to punish and deter — without adding meaning to these terms. For example, the trial court did not suggest what relation, if any, should exist between the harm caused and the size of the award, nor how to measure the deterrent effect of a particular award. It provided no information to the jury about criminal fines for comparable conduct or the range of punitive damages awards in similar cases. Nor did it identify the *49limitations dictated by retributive and deterrent principles, or advise the jury to refrain from awarding more than necessary to meet these objectives. In short, the trial court’s instruction identified the ultimate destination, but did not tell the jury how to get there. Due process may not require a detailed roadmap, but it certainly requires directions of some sort.
Giaccio is instructive in this inquiry. There, the State argued that even if the cost-assessment statute was imper-missibly vague as written, subsequent state court decisions had adopted meaningful standards for implementing it. The jury in Giaccio was thus instructed that it could assess costs against the defendant if it found that he was guilty of misconduct that, while not a criminal offense, warranted a penalty. See Giaccio, 382 U. S., at 404. This Court did not accept that this nebulous instruction cured the statute’s vagueness. “It may possibly be that the trial court’s charge comes nearer to giving a guide to the jury than those that preceded it, but it still falls short of the kind of legal standard due process requires.” Ibid.
The trial court’s instruction in this case fares no better. In fact, the minimal guidance it offered may well have pushed the jury further away from reasoned decisionmaking. Paraphrased slightly, the court’s terse instruction told the jury: “Think about how much you hate what the defendants did and teach them a lesson.” This is not the sort of instruction likely to produce a fair, dispassionate verdict. Like most common-law punitive damages instructions, this one has “an open-ended, anything-goes quality that can too easily stoke . . . the vindictive or sympathetic passions of juries.” P. Huber, Liability: The Legal Revolution and Its Consequences 118 (1988) (hereinafter Huber). Our cases attest to the wildly unpredictable results and glaring unfairness that characterize common-law punitive damages procedures. See infra, at 54-55.
*50One need not look far to see that these so-called standards provide no guidance to Alabama juries. Consider, for example, a recent Alabama case involving a collision between a train and a tractor-trailer truck, which resulted in the death of the driver of the truck. Notwithstanding that the truck pulled onto the tracks right in front of the train, thereby-ignoring a stop sign, three warning signs, and five speed bumps, the administratrix of decedent’s estate asked for $3 million in punitive damages. The jury, after receiving instructions no more vague than those at issue here, awarded her $15 million. Whitt v. Burlington Northern R. Co., No. CV-85-311 (Cir. Ct. Ala., Aug. 23, 1988), aff’d conditionally, 575 So. 2d 1011 (1990) (remitting award to $5 million), stay granted, No. A-408 (90-1250) (Dec. 5, 1990) (Kennedy, J., Circuit Justice).
That Alabama’s “standards” in fact provide no guidance whatsoever was illustrated quite dramatically by Alabama Supreme Court Justice Houston in his concurring opinion in Charter Hospital of Mobile, Inc. v. Weinberg, 558 So. 2d 909, 916 (1990). He pointed to two cases involving substantially the same misconduct and jury instructions, but having very different results: Washington Nat. Ins. Co. v. Strickland, 491 So. 2d 872 (Ala. 1985), and Land & Associates, Inc. v. Simmons, 562 So. 2d 140 (Ala. 1989). In both cases, an insurance agent misrepresented to a prospective insured that coverage would begin as soon as the insured paid the first premium when, in reality, the agent should have known that coverage was conditioned upon a medical examination that the insured was unlikely to pass. See Strickland, supra, at 873, 877; Simmons, supra, at 142. In one case, the jury handed down a punitive damages award of approximately $21,000 — 1572 times the compensatory damages. See Strickland, supra, at 874. In the other case, the jury penalized substantially the same conduct with a punitive damages award of $2,490,000 — 249 times the compensatory award. See Simmons, supra, at 151 (Houston, J., concurring spe-*51daily). These vastly disparate results demonstrate that, under Alabama’s common-law scheme, any case-to-case consistency among verdicts is purely fortuitous.
This is not a case where more precise standards are either impossible or impractical. See Kolender, 461 U. S., at 361. Just the opposite. The Alabama Supreme Court has already formulated a list of seven factors that it considers relevant to the size of a punitive damages award:
“ ‘(1) Punitive damages should 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. If the actual or likely harm is slight, the damages should be relatively small. If grievous, the damages should be much greater.
“‘(2) The degree of reprehensibility of the defendant’s conduct should be considered. The duration of this conduct, the degree of the defendant’s awareness of any hazard which his conduct has caused or is likely to cause, and any concealment or “cover-up” of that hazard, and the existence and frequency of similar past conduct should all be relevant in determining this degree of reprehensibility.
“ ‘(3) If the wrongful conduct was profitable to the defendant, the punitive damages should remove the profit and should be in excess of the profit, so that the defendant recognizes a loss.
“ ‘(4) The financial position of the defendant would be relevant.
“ ‘(6) All the costs of litigation should be included, so as to encourage plaintiffs to bring wrongdoers to trial.
“‘(6) If criminal sanctions have been imposed on the defendant for his conduct, this should be taken into account in mitigation of the punitive damages award.
“ ‘(7) If there have been other civil actions against the same defendant, based on the same conduct, this should be taken into account in mitigation of the punitive dam*52ages award.’” Green Oil Co. v. Hornsby, 539 So. 2d 218, 223-224 (1989), quoting Aetna Life Ins. Co. v. Lavoie, 505 So. 2d 1050, 1062 (Ala. 1987) (Houston, J., concurring specially).
In my view, these standards — the “Green Oil factors”— could assist juries to make fair, rational decisions. Unfortunately, Alabama courts do not give the Green Oil factors to the jury. See 539 So. 2d, at 224 (Maddox, J., concurring specially). Instead, the jury has standardless discretion to impose punitive damages whenever and in whatever amount it wants. The Green Oil factors play a role only after the jury has rendered its verdict. The trial court and other reviewing courts may — but are not required to — take these factors into consideration in determining whether a punitive damages award is excessive. Id., at 223.
Obviously, this post hoc application of the Green Oil factors does not cure the vagueness of the jury instructions. Cf. Baggett v. Bullitt, 377 U. S. 360, 373 (1964) (“[judicial safeguards do not neutralize the vice of a vague law”). See also Roberts v. United States Jaycees, 468 U. S. 609, 629 (1984). As respondents candidly admit, judicial review in Alabama is limited to the amount of the award. The void-for-vagueness doctrine, on the other hand, is concerned with the procedures by which the amount is determined. After-the-fact review of the amount in no way diminishes the fact that the State entrusts its juries with standardless discretion. It thus does not matter that the amount settled upon by the jury might have been permissible under a rational system. Even a wholly irrational process may, on occasion, stumble upon a fair result. What is crucial is that the existing system is not rational. “[Pjrocedural due process rules are shaped by the risk of error inherent in the truth-finding process as applied to the generality of cases, not the rare exceptions.” Mathews v. Eldridge, 424 U. S., at 344. The state court justice who devised the Green Oil factors, Justice Houston, has recognized this. Addressing a vagueness chai-*53lenge to the State’s punitive damages procedures, he wrote: “We have attempted to deal with the issue of the reliability of punitive damages assessments by post-trial review only. That attempt does not really address the issue.” Charter Hospital, 558 So. 2d, at 915 (opinion concurring specially) (emphasis added; citations omitted).
II
For the reasons stated above, I would hold that Alabama’s common-law punitive damages scheme is void for vagueness. But the Court need not agree with me on this point in order to conclude that Pacific Mutual was denied procedural due process. Whether or not the Court agrees that the jury instructions were so vague as to be unconstitutional, there can be no doubt but that they offered substantially less guidance than is possible. Applying the test of procedural due process set out in Mathews v. Eldridge, supra, more guidance was required. Modest safeguards would make the process significantly more rational without impairing any legitimate governmental interest.
A
In Mathews v. Eldridge, supra, at 334, we recognized that “‘“[d]ue process,” unlike some legal rules, is not a technical conception with a fixed content unrelated to time, place and circumstances. Cafeteria Workers v. McElroy, 367 U. S. 886, 895 (1961).’’’ “[D]ue process is flexible and calls for such procedural protections as the particular situation demands.” Morrissey v. Brewer, 408 U. S. 471, 481 (1972). Accordingly, Mathews described a sliding-scale test for determining whether a particular set of procedures was constitutionally adequate. We look at three factors: (1) the private interest at stake; (2) the risk that existing procedures will wrongly impair this private interest, and the likelihood that additional procedural safeguards can effect a cure; and (3) the governmental interest in avoiding these additional procedures. Mathews, supra, at 335.
*54Applying the Mathews test to Alabama’s common-law punitive damages scheme, it is clear that the state procedures deprive defendants of property without due process of law. The private property interest at stake is enormous. Without imposing any legislative or common-law limits, Alabama authorizes juries to levy civil fines ranging from zero to tens of millions of dollars. Indeed, a jury would not exceed its discretion under state law by imposing an award of punitive damages that was deliberately calculated to bankrupt the defendant. Unlike compensatory damages, which are tied to an actual injury, there is no objective standard that limits the amount of punitive damages. Consequently, “‘the impact of these windfall recoveries is unpredictable and potentially substantial.’” Bankers Life & Casualty Co. v. Crenshaw, 486 U. S. 71, 87 (1988) (O’CONNOR, J., concurring in part and concurring in judgment), quoting Electrical Workers v. Foust, 442 U. S. 42, 50 (1979).
Compounding the problem, punitive damages are quasi-criminal punishment. Unlike compensatory damages, which serve to allocate an existing loss between two parties, punitive damages are specifically designed to exact punishment in excess of actual harm to make clear that the defendant’s misconduct was especially reprehensible. Hence, there is a stigma attached to an award of punitive damages that does not accompany a purely compensatory award. The punitive character of punitive damages means that there is more than just money at stake. This factor militates in favor of strong procedural safeguards.
The second Mathews prong focuses on the fairness and reliability of existing procedures. This is a question we have spoken to before. Over the last 20 years, the Court has repeatedly criticized common-law punitive damages procedures on the ground that they invite discriminatory and otherwise illegitimate awards. E. g., Gertz, 418 U. S., at 350 (common-law procedures leave juries “free to use their discretion selectively to punish expressions of unpopular *55views”); Electrical Workers, supra, at 50-51, and n. 14 (“[Pjunitive damages may be employed to punish unpopular defendants”); Rosenbloom v. Metromedia, Inc., 403 U. S. 29, 84 (1971) (Marshall, J., dissenting) (“This discretion allows juries to penalize heavily the unorthodox and the unpopular and exact little from others”); Smith v. Wade, 461 U. S. 30, 59 (1983) (Rehnquist, J., dissenting) (“[PJunitive damages are frequently based upon the caprice and prejudice of jurors”). For this reason, the Court has forbidden the award of punitive damages in certain defamation suits brought by private plaintiffs, Gertz, supra, at 349-350, and in unfair representation suits brought against labor unions under the Railway Labor Act, Electrical Workers, supra, at 52.
Although our cases have not squarely addressed the due process question before us today, see Browning-Ferris, 492 U. S., at 276-277, we have strongly hinted at the answer. See ante, at 9-12. Justice Brennan and Justice Marshall joined the Courtis opinion in Browning-Ferris, but wrote separately to express their “understanding that it leaves the door open for a holding that the Due Process Clause constrains the imposition of punitive damages in civil cases brought by private parties.” 492 U. S., at 280 (Brennan, J., concurring). In a separate opinion that Justice Stevens joined, I voiced strong concerns “regarding the vagueness and procedural due process problems presented by juries given unbridled discretion to impose punitive damages.” Id., at 283 (opinion concurring in part and dissenting in part). This echoed my earlier statement, with which Justice Scalia joined, in Bankers Life, supra, at 88: “This grant of wholly standardless discretion to determine the severity of punishment appears inconsistent with due process” (opinion concurring in part and concurring in judgment).
As explained above, see supra, at 52-53, Alabama’s grant of standardless discretion to juries is not remedied by post hoc judicial review. At best, this mechanism tests whether the award is grossly excessive. This is an important sub*56stantive due process concern, but our focus here is on the requirements of procedural due process. Cf. Santosky v. Kramer, 455 U. S. 745, 757 (1982) (“Retrospective case-by-case review cannot preserve fundamental fairness when a class of proceedings is governed by a constitutionally defective evi-dentiary standard”).
Even if judicial review of award amounts could potentially minimize the evils of standardless discretion, Alabama’s review procedure is not up to the task. For one thing, Alabama courts cannot review whether a jury properly applied permissible factors, because juries are not told which factors are permissible and which are not. See Wheeler, The Constitutional Case for Reforming Punitive Damages Procedures, 69 Va. L. Rev. 269, 290 (1983) (hereinafter Wheeler). Making effective review even more unlikely, the primary component of Alabama’s review mechanism is deference. The State Supreme Court insists that a jury’s award of punitive damages carries a “presumption of correctness” that a defendant must overcome before remittitur is appropriate. Green Oil, 539 So. 2d, at 222, 224. Reviewing courts are thus required to uphold the jury’s exercise of unbridled, unchanneled, standardless discretion unless the amount happened upon by the jury cannot be reconciled with even the most generous application of the Green Oil factors.
That is precisely what happened here. When Pacific Mutual challenged the State’s procedures governing awards of punitive damages, the trial court simply deferred to the jury. The judge noted that he “would in all likelihood have rendered a lesser amount,” App. to Pet. for Cert. A-15, but that the verdict was not excessive or unfair because “[t]he jury was composed of male and female, white and black and . . . acted conscientiously throughout the trial.” Ibid. Relying on the trial judge’s refusal to disturb the verdict, the State Supreme Court afforded it a double dose of deference, stating that “jury verdicts are presumed correct, and that presump*57tion is strengthened when the presiding judge refuses to grant a new trial.” 553 So. 2d 537, 543 (1989).
This strong deference is troubling given that the Alabama Supreme Court has explicitly acknowledged that its current procedures provide for “‘unguided discretion,” Green Oil, 539 So. 2d, at 222, and in no way dictate a rational jury verdict: “ ‘The current system furnishe[s] virtually no yardstick for measuring the amount of the award over against the purpose of the award. ’ ” Ibid., quoting Ridout’s-Brown Service, Inc. v. Holloway, 397 So. 2d 125, 127-128 (Ala. 1981) (Jones, J., concurring specially). “[I]t is possible for a jury to hear the evidence in the case, make findings of fact, correctly apply the law, and still, albeit unwittingly, assess damages that bear no reasonable relationship to the accomplishment of [punishment and deterrence] goals.” 539 So. 2d, at 222. Thus, the State Supreme Court recognizes that its common-law procedures produce irrational results, yet insists on deferring to these results. Blind adherence to the product of recognized procedural infirmity is not judicial review as I understand it. It is an empty exercise in rationalization that creates only the appearance of evenhanded justice.
Crucial to Mathews’ second prong, the procedural infirmities here are easily remedied. The Alabama Supreme Court has already given its approval to the Green Oil factors. By giving these factors to juries, the State would be providing them with some specific standards to guide their discretion. This would substantially enhance the fairness and rationality of the State’s punitive damages system. Other procedural safeguards might prove equally effective. For example, state legislatures could establish fixed monetary limits for awards of punitive damages for particular kinds of conduct. So long as the legislatively determined ranges are sufficiently narrow, they could function as meaningful constraints on jury discretion while at' the same time permitting juries to render individualized verdicts.
*58Another possibility advocated by several commentators, see ante, at 23, n. 11; Wheeler 300-301, is that States could bifurcate trials into liability and punitive damages stages. At the punitive damages stage, clear and convincing evidence that the defendant acted with the requisite culpability would be required. This would serve two goals. On a practical level, the clear-and-convincing-evidence requirement would constrain the jury’s discretion, limiting punitive damages to the more egregious cases. This would also permit closer scrutiny of the evidence by trial judges and reviewing courts. See Ellis, Punitive Damages, Due Process, and the Jury, 40 Ala. L. Rev. 975, 995-996 (1989). On a symbolic level, the higher evidentiary standard would signal to the jury that it should have a high level of confidence in its factual findings before imposing punitive damages. Id., at 995; Wheeler 297-298. Any of these rudimentary modifications would afford more meaningful guidance to juries, thereby lessening the chance of arbitrary and discriminatory awards, without impairing the State’s legitimate interests in punishment and deterrence. Given the existence of several equally acceptable methods, concerns of federalism and judicial restraint counsel that this Court should not legislate to the States which particular method to adopt. I would thus leave it to individual States to decide what method is most consistent with their objectives.
The final Mathews factor asks whether the State has a legitimate interest in preserving standardless jury discretion that is so compelling as to render even modest procedural reforms unduly burdensome. The Court effectively answered this question in Gertz, 418 U. S., at 349, announcing that “the States have no substantial interest in securing for plaintiffs . . . gratuitous awards of money damages far in excess of actual injury.” (Emphasis added.)
Respondents do not give up easily. They point out that the State has a substantial interest in deterring wrongful conduct and draw from this a peculiar argument. They con*59tend that, by making jury awards more predictable, procedural safeguards will tend to diminish the deterrent effect of punitive damages. If award amounts are predictable, they argue, corporations will not avoid wrongdoing; instead, they will merely calculate the probability of a punitive damages award and factor it in as a cost of doing business. Accordingly, to best advance the State’s interest in deterrence, juries must be given unbridled discretion to render awards that are wildly unpredictable.
This argument goes too far. While the State has a legitimate interest in avoiding rigid strictures so that a jury may tailor its award to specific facts, the Due Process Clause does not permit a State to classify arbitrariness as a virtue. Indeed, the point of due process — of the law in general — is to allow citizens to order their behavior. A State can have no legitimate interest in deliberately making the law so arbitrary that citizens will be unable to avoid punishment based solely upon bias or whim. The procedural reforms suggested here in no way intrude on the jury’s ability to exercise reasoned discretion, nor do they preclude flexible decision-making. Due process requires only that a jury be given a measurable degree of guidance, not that it be straitjacketed into performing a particular calculus.
Similarly, the suggested procedural safeguards do not impair the State’s punishment objectives. Admittedly, the State has a strong interest in punishing wrongdoers, but it has no legitimate interest in maintaining in pristine form a common-law system that imposes disproportionate punishment and that subjects defendants guilty of similar misconduct to wholly different punishments. Due process requires, at some level, that punishment be commensurate with the wrongful conduct. See Solem v. Helm, 463 U. S. 277, 284-290 (1983); id., at 311, n. 3 (Burger, C. J., dissenting). The State can therefore have no valid objection to procedural measures that merely ensure that punitive damages awards *60are based on some factual or legal predicate, rather than the personal predilections and whims of individual jurors.
B
In his concurrence, Justice Sc alia offers a very different notion of what due process requires. He argues that a practice with a long historical pedigree is immune to reexamination. Ante, at 38. The Court properly rejects this argument. Ante, at 18. A static notion of due process is flatly inconsistent with Mathews, 424 U. S., at 334-335, in which this Court announced that the requirements of the Due Process Clause are “‘flexible’” and may vary with “‘time, place and circumstances.’” We have repeatedly relied on the Mathews analysis, and our recent cases leave no doubt as to its continued vitality. See, e. g., Washington v. Harper, 494 U. S. 210, 229 (1990); Brock v. Roadway Express, Inc., 481 U. S. 252, 261-262 (1987); Walters v. National Assn. of Radiation Survivors, 473 U. S. 305, 320-321 (1985); Cleveland Bd. of Ed. v. Loudermill, 470 U. S. 532, 542-543 (1985); Ake v. Oklahoma, 470 U. S. 68, 77 (1985); Schall v. Martin, 467 U. S. 253, 274 (1984).
Due process is not a fixed notion. Procedural rules, “even ancient ones, must satisfy contemporary notions of due process.” Burnham v. Superior Court of Cal., County of Marin, 495 U. S. 604, 630 (1990) (Brennan, J., concurring in judgment). Although history creates a strong presumption of continued validity, “the Court has the authority under the [Fourteenth] Amendment to examine even traditionally accepted procedures and declare them invalid.” Id., at 628 (White, J., concurring in part and concurring in judgment), citing Shaffer v. Heitner, 433 U. S. 186 (1977).
The Court’s decision in Williams v. Illinois, 399 U. S. 235 (1970), is also instructive. In Williams, the Court invalidated on equal protection grounds the time-honored practice of extending prison terms beyond the statutory maximum *61when a defendant was unable to pay a fine or court costs. The Court’s language bears repeating:
“[Njeither the antiquity of a practice nor the fact of steadfast legislative and judicial adherence to it through the centuries insulates it from constitutional attack ....
“The need to be open to reassessment of ancient practices other than those explicitly mandated by the Constitution is illustrated by the present case since the greatly increased use of fines as a criminal sanction has made nonpayment a major cause of incarceration in this country.” Id., at 239-240.
Punitive damages are similarly ripe for reevaluation. In the past, such awards “merited scant attention” because they were “rarely assessed and likely to be small in amount.” Ellis, Fairness and Efficiency in the Law of Punitive Damages, 56 S. Cal. L. Rev. 1, 2 (1982). When awarded, they were reserved for the most reprehensible, outrageous, or insulting acts. See F. Pollock, Law of Torts (1887); Huber 119. Even then, they came at a time when compensatory damages were not available for pain, humiliation, and other forms of intangible injury. Punitive damages filled this gap. See K. Redden, Punitive Damages § 2.3(A) (1980); Note, Exemplary Damages in the Law of Torts, 70 Harv. L. Rev. 517, 519-520 (1957).
Recent years, however, have witnessed an explosion in the frequency and size of punitive damages awards. See RAND Institute for Civil Justice, M. Peterson, S. Sarma, & M. Shanley, Punitive Damages — Empirical Findings iii (1987) (hereinafter RAND). A recent study by the RAND Corporation found that punitive damages were assessed against 1 of every 10 defendants who were found liable for compensatory damages in California. Id., at viii. The amounts can be staggering. Within nine months of our decision in Browning-Ferris, there were no fewer than six punitive damages awards of more than $20 million. Crovitz, Absurd Punitive Damages Also “Mock” Due Process, Wall St. *62Journal, Mar. 14, 1990, p. A19, col. 3. Medians as well as averages are skyrocketing, meaning that even routine awards are growing in size. RAND vi, ix, 65. The amounts “seem to be limited only by the ability of lawyers to string zeros together in drafting a complaint.” Oki America, Inc. v. Microtech Int’l, Inc., 872 F. 2d 312, 315 (CA9 1989) (Kozinski, J., concurring).
Much of this is attributable to changes in the law. For 200 years, recovery for breach of contract has been limited to compensatory damages. In recent years, however, a growing number of States have permitted recovery of punitive damages where a contract is breached or repudiated in bad faith. See, e. g., Seaman’s Direct Buying Serv., Inc. v. Standard Oil Co., 36 Cal. 3d 752, 686 P. 2d 1158 (1984). Unheard of only 30 years ago, bad faith contract actions now account for a substantial percentage of all punitive damages awards. See RAND iv. Other significant legal developments include the advent of product liability and mass tort litigation. “As recently as a decade ago, the largest award of punitive damages affirmed by an appellate court in a products liability case was $250,000. . . . Since then, awards more than 30 times as high have been sustained on appeal.” Browning-Ferris, 492 U. S., at 282 (O’Connor, J., concurring in part and dissenting in part). “Today, hardly a month goes by without a multimillion-dollar punitive damages verdict in a product liability case.” Wheeler, A Proposal for Further Common Law Development of the Use of Punitive Damages in Modern Product Liability Litigation, 40 Ala. L. Rev. 919 (1989).
As in Williams, the time has come to reassess the constitutionality of a time-honored practice. The explosion in the frequency and size of punitive damages awards has exposed the constitutional defects that inhere in the common-law system. That we did not discover these defects earlier is regrettable, but it does not mean that we can pretend that they do not exist now. “[N]ew cases expose old infirmities which *63apathy or absence of challenge has permitted to stand. But the constitutional imperatives . . . must have priority over the comfortable convenience of the status quo.” Williams, 399 U. S., at 245. Circumstances today are different than they were 200 years ago, and nothing in the Fourteenth Amendment requires us to blind ourselves to this fact. See Wheeler 277. Just the opposite is true. The Due Process Clause demands that we possess some degree of confidence that the procedures employed to deprive persons of life, liberty, and property are capable of producing fair and reasonable results. When we lose that confidence, a change must be made.
Ill
“ ‘The touchstone of due process is protection of the individual against arbitrary action of government.’’’ Daniels v. Williams, 474 U. S. 327, 331 (1986), quoting Dent v. West Virginia, 129 U. S. 114, 123 (1889). Alabama’s common-law scheme for awarding punitive damages provides a jury with “such skeletal guidance,” Browning-Ferris, supra, at 281 (Brennan, J., concurring), that it invites — even requires — arbitrary results. It gives free reign to the biases and prejudices of individual jurors, allowing them to target unpopular defendants and punish selectively. In short, it is the antithesis of due process. It does not matter that the system has been around for a long time, or that the result in this particular case may not seem glaringly unfair. The common-law scheme yields unfair and inconsistent results “in so many instances that it should be held violative of due process in every case.” Burnham, 495 U. S., at 628 (White, J., concurring in part and concurring in judgment).
I would require Alabama to adopt some method, either through its legislature or its courts, to constrain the discretion of juries in deciding whether or not to impose punitive damages and in fixing the amount of such awards. As a number of effective procedural safeguards are available, we need not dictate to the States the precise manner in which *64they must address the problem. We should permit the States to experiment with different methods and to adjust these methods over time.
This conclusion is neither groundbreaking nor remarkable. It reflects merely a straightforward application of our Due Process Clause jurisprudence. Given our statements in recent cases such as Browning-Ferris, supra, and Bankers Life & Casualty Co. v. Crenshaw, 486 U. S. 71 (1988), the parties had every reason to expect that this would be the Court’s holding. Why, then, is it consigned to a dissent rather than a majority opinion? It may be that the Court is reluctant to afford procedural due process to Pacific Mutual because it perceives that such a ruling would force us to evaluate the constitutionality of every State’s punitive damages scheme. I am confident, though, that if we announce what the Constitution requires and allow the States sufficient flexibility to respond, the constitutional problems will be resolved in time without any undue burden on the federal courts. Indeed, it may have been our hesitation that has inspired a flood of petitions for certiorari. For more than 20 years, this Court has criticized common-law punitive damages procedures, see supra, at 54-55, but has shied away from its duty to step in, hoping that the problems would go away. It is now clear that the problems are getting worse, and that the time has come to address them squarely. The Court does address them today. In my view, however, it offers an incorrect answer.