JUSTICE O'CONNOR, with whom JUSTICE WHITE joins, and with whom JUSTICE SOUTER joins as to Parts II-B-2, II-C, III, and IV, dissenting.
In Pacific Mut. Life Ins. Co. v. Has lip, 499 U. S. 1 (1991), this Court held out the promise that punitive damages awards would receive sufficient constitutional scrutiny to restore fairness in what is rapidly becoming an arbitrary and oppressive system. Today the Court's judgment renders Haslip's promise a false one. The procedures that converted this commercial dispute into a $10 million punitive verdict were wholly inadequate. Rather than producing a judgment founded on verifiable criteria, they produced a monstrous award-526 times actual damages and over 20 times greater than any punitive award in West Virginia history. Worse, the State Supreme Court of Appeals rejected petitioner's challenge with only cursory analysis, observing that petitioner, rather than being "really stupid," had been "really mean." 187 W. Va. 457, 474-475, 419 S. E. 2d 870, 887-889 (1992). The court similarly refused to consider the possibility of remittitur because petitioner "and its agents and servants failed to conduct themselves as gentlemen." Id., at 462, 419 S. E. 2d, at 875. In my view, due process does not tolerate such cavalier standards when so much is at stake. Because I believe that neither this award's size nor the procedures that produced it are consistent with the principles this Court articulated in Haslip, I respectfully dissent.
I
Our system of justice entrusts jurors-ordinary citizens who need not have any training in the law-with profoundly important determinations. Jurors decide not only civil matters, where the financial consequences may be great, but also criminal cases, where the liberty or perhaps life of the defendant hangs in the balance. Our abiding faith in the jury system is founded on longstanding tradition reflected in constitutional text, see U. S. Const., Art. III, § 2, Amdts. 6, 7, and is supported by sound considerations of justice and democratic theory. The jury system long has been a guarantor of fairness, a bulwark against tyranny, and a source of civic values. See 3 W. Blackstone, Commentaries *379-*381; Haslip, supra, at 40 (KENNEDY, J., concurring in judgment); w. Olson, The Litigation Explosion 175 (1991); Hyman & Tarrant, Aspects of American Trial Jury History, in The Jury System in America 23, 27-28 (R. Simon ed. 1975).
But jurors are not infallible guardians of the public good.
They are ordinary citizens whose decisions can be shaped by influences impermissible in our system of justice. In fact, they are more susceptible to such influences than judges. See H. Kalven & H. Zeisel, The American Jury 497-498 (1966) ("The judge very often perceives the stimulus that moves the jury, but does not yield to it…. The perennial amateur, layman jury cannot be so quickly domesticated to official role and tradition; it remains accessible to stimuli which the judge will exclude"). Arbitrariness, caprice, passion, bias, and even malice can replace reasoned judgment and law as the basis for jury decisionmaking. Modern judicial systems therefore incorporate safeguards against such influences. Rules of evidence limit what the parties may present to the jury. Careful instructions direct the jury's deliberations. Trial judges diligently supervise proceedings, watchful for potential sources of error. And courts of appeals stand ready to overturn judgments when efforts to ensure fairness have failed.
In the usual case, this elaborate but necessary judicial machinery functions well, ensuring that our jury system is an engine of liberty and justice rather than a source of oppression and arbitrary imposition. As JUSTICE KENNEDY has explained, "[e]lements of whim and caprice do not predominate when the jury reaches a consensus based upon arguments of counsel, the presentation of evidence, and instructions from the trial judge, subject to review by the trial and appellate courts." Has lip, 499 U. S., at 40 (opinion concurring in judgment). But the risk of prejudice, bias, and caprice remains a real one in every case nonetheless.
This is especially true in the area of punitive damages, where juries sometimes receive only vague and amorphous guidance. Jurors may be told that punitive damages are im posed to punish and deter, but rarely are they instructed on how to effectuate those goals or whether any limiting principles exist. See, e. g., id., at 39. Although this Court has not held such instructions constitutionally inadequate, it 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. See id., at 43, 63 (O'CONNOR, J., dissenting); id., at 41 (KENNEDY, J., concurring in judgment) ("[T]he generality of the instructions may contribute to a certain lack of predictability"); Browning-Ferris Industries of Vt., Inc. v. Kelco Disposal, Inc., 492 U. S. 257, 281 (1989) (Brennan, J., concurring) (Such "skeletal" guidance is "scarcely better than no guidance at all," creating a need for more careful review); Smith v. Wade, 461 U. S. 30, 88 (1983) (REHNQUIST, J., dissenting) (elastic standards applicable to punitive awards "giv[e] free reign to the biases and prejudices of juries"). As one commentator has explained:"Like everyone else in the court system, juries need and deserve objective rules for decision. Deprived of any fixed landmarks and guideposts, any of us can be distracted, played on, and befuddled to the point where our best guess is far from reliable." Olson, supra, at 175.
It is therefore no surprise that, time and again, this Court and its Members have expressed concern about punitive damages awards" 'run wild,'" inexplicable on any basis but caprice or passion. Has lip, supra, at 9-12, 18 (discussing cases); see also Gertz v. Robert Welch, Inc., 418 U. S. 323, 350 (1974) ("[J]uries assess punitive damages in wholly unpredictable amounts bearing no necessary relation to the actual harm caused").
Influences such as caprice, passion, bias, and prejudice are antithetical to the rule of law. If there is a fixture of due process, it is that a verdict based on such influences cannot stand. See Haslip, supra, at 41 (KENNEDY, J., concurring in judgment) ("A verdict returned by a biased or prejudiced jury no doubt violates due process"). Of course, determining whether a verdict resulted from improper influences is no easy matter. By tradition and necessity, the circumstances in which jurors may impeach their own verdict are quite limited. See Tanner v. United States, 483 U. S. 107, 117-121, 127 (1987); 11 C. Wright & A. Miller, Federal Practice and Procedure § 2810, pp. 71-72 (1973); 2 W. Tidd, Practice of Courts of King's Bench and Common Pleas *908-*909. But fundamental fairness requires that impermissible influences such as bias and prejudice be discovered nonetheless, by inference if not by direct proof. As a result, courts at common law in England traditionally would strike any award that appeared so grossly disproportionate as to evidence caprice, passion, or bias.1 This practice long has been followed in this Nation as well.2 Indeed, the New Hampshire Supreme Court emphasized its importance over a century ago, observing that a court's duty to interfere with a disproportionate jury verdict "is absolutely necessary to the safe administration of justice, and ought, in all proper cases, to be asserted and exercised." Belknap v. Boston & Maine R. Co., 49 N. H. 358, 372 (1870). Accord, Gough v. Farr, 1 Y. & J. 477, 479-480, 148 Eng. Rep. 759, 760 (Ex. 1827) (Vaughan, B.) ("It is essential to the due administration of justice, that the Courts should exercise a salutary control over Juries" by requiring retrial where the amount of the verdict indicates that the jury "acted improperly, or upon a gross misconception of the facts"); id., at 478-479, 148 Eng. Rep., at 759-760 (Alexander, L. C. B.) (Where damages are so excessive that "the Courts are of opinion… that the Jury have acted under the influence of undue motives, or of misconception, it is their duty to interfere"); Travis v. Barger, 24 Barb. 614, 629 (N. Y. 1857) (reciting Lord Ellenborough's view that, "if it appeared from the amount of damages given, as compared with the facts of the case laid before jury, that the jury must have acted under the influence either of undue motives, or some gross error or misconception of the subject, the court would have thought it their duty to submit the question to the consideration of a second jury"); Flannery v. Baltimore & Ohio R. Co., 15 D. C. 111, 125 (1885) (When the punitive damages award is disproportionate, "we feel it our duty to interfere").
Judicial intervention in cases of excessive awards also has the critical function of ensuring that another ancient and fundamental principle of justice is observed-that the punishment be proportionate to the offense. As we have observed, the requirement of proportionality is "deeply rooted and frequently repeated in common-law jurisprudence." Solem v. Helm, 463 U. S. 277, 284-285 (1983). See, e. g., Le Gras v. Bailiff of Bishop of Winchester, Y. B. Mich. 10 Edw. II, pI. 4 (C. P. 1316), reprinted in 52 Selden Society 3, 5 (1934) (amercement vacated and bailiff ordered to "take a moderate amercement proper to the magnitude and manner of that offence"); First Statute of Westminster, 3 Edw. I, ch. 6 (1275). Because punitive damages are designed as punishment rather than compensation, Browning-Ferris, 492 U. S., at 297 (O'CONNOR, J., concurring in part and dissenting in part) (citing cases), courts historically have required that punitive damages awards bear a reasonable relationship to the actual harm imposed.3 This Court similarly has recognized that the requirement of proportionality is implicit in the notion of due process. We therefore have held that an award that is "plainly arbitrary and oppressive," Southwestern Telegraph & Telephone Co. v. Danaher, 238 U. S. 482, 491 (1915), "grossly excessive," Waters-Pierce Oil Co. v. Texas (No.1), 212 U. S. 86, 111 (1909), or "so severe and oppressive as to be wholly disproportioned to the offense and obviously unreasonable," St. Louis, 1. M. & S. R. Co. v. Williams, 251 U. S. 63, 66-67 (1919), offends the Due Process Clause and may not stand.
II
The plurality does not retreat today from our prior statements regarding excessive punitive damages awards. Nor does it deny that our prior decisions have a strong basis in historical practice and the common law. On the contrary, it reaffirms our precedents once again, properly rebuffing respondents' attempt to denigrate them as Lochner-era aberrations. Ante, at 455. It is thus common ground that an award may be so excessive as to violate due process. Ibid. We part company, however, on how to determine if this is such an award.
In Solomonic fashion, the plurality rejects both petitioner's and respondents' proffered approaches, instead selecting a seemingly moderate course. See ante, at 456-458. But the course the plurality chooses is, in fact, no course at all. The plurality opinion erects not a single guidepost to help other courts find their way through this area. Rather, quoting Haslip's observation that there is no "'mathematical bright line between the constitutionally acceptable and the constitutionally unacceptable,'" ante, at 458 (quoting 499 U. S., at 18), the plurality abandons all pretense of providing instruction and moves directly into the specifics of this case.
I believe that the plurality errs not only in its result but also in its approach. Our inability to discern a mathematical formula does not liberate us altogether from our duty to provide guidance to courts that, unlike this one, must address jury verdicts such as this on a regular basis. On the contrary, the difficulty of the matter imposes upon us a correspondingly greater obligation to provide the most coherent explanation we can. I agree with the plurality that we ought not adopt TXO's or respondents' suggested approach as a rigid formula for determining the constitutionality of punitive damages verdicts. But it does not follow that, in the course of deciding this case, we should avoid offering even a clue as to our own.
TXO's suggestion that this Court should rely on objective criteria has much to commend it. As an initial matter, constitutional judgments "'should not be, or appear to be, merely the subjective views of individual Justices.''' Rum mel v. Estelle, 445 U. S. 263, 274 (1980) (quoting Coker v. Georgia, 433 U. S. 584, 592 (1977) (opinion of WHITE, J.)). Without objective criteria on which to rely, almost any decision regarding proportionality will be a matter of personal preference. One judge's excess very well may be another's moderation. To avoid that element of subjectivity, our "'judgment[s] should be informed by objective factors to the maximum possible extent.'" 445 U. S., at 274-275 (quoting same). As the plurality points out, ante, at 455-456, TXO directs our attention to various objective indicators, including the relationship between the punitive damages award and compensatory damages, awards of punitive damages upheld against other defendants in the same jurisdiction, awards upheld for similar torts in other jurisdictions, and legislatively designated penalties for similar misconduct. While these factors by no means exhaust the due process inquiry, they are quite probative. It is to their proper application that I now turn.
A
In my view, due process at least requires judges to engage in searching review where the verdict discloses such great disproportions as to suggest the possibility of bias, caprice, or passion. As JUSTICE STEVENS observed in a different context, "[o]ne need not use Justice Stewart's classic definition of obscenity-'I know it when I see it' -as an ultimate standard for judging" the constitutionality of a punitive damages verdict "to recognize that the dramatically irregular" size and nature of an award "may have sufficient probative force to call for an explanation." Cf. Karcher v. Daggett, 462 U. S. 725, 755 (1983) (concurring opinion) (footnotes omitted).
This $10 million punitive award, returned in a case involving only $19,000 in compensatory damages, is a dramatically irregular, if not shocking, verdict by any measure. At the very least it should raise a suspicious judicial eyebrow. Not only does the punitive award represent over 500 times actual damages, but it also exceeds economic harm by over $9.98 million. Thus, it cannot be accepted as bearing the "understandable relationship to compensatory damages," 499 U. S., at 22, the Court found sufficient in Haslip. Indeed, in Has lip the Court observed that an $840,000 punitive award, representing four times compensatory damages, may have been "close to the line" of "constitutional impropriety." Id., at 23-24. If the quadruple damages, $840,000 award in Haslip was "close to the line," absent a convincing explanation, this $10 million award-over 500 times actual damages-surely must cross it.
A comparison of this award and prior ones in West Virginia confirms its unusual nature: It is 20 times larger than the highest punitive damages award ever upheld in West Virginia history for any misconduct. See App. to Brief for Petitioner 1a-3a (listing punitive damages awards affirmed on appeal in West Virginia). That figure is particularly surprising if one considers the nature of the offense at issue. This is not a case involving grave physical injury imposed on a helpless citizen by a callous malefactor. Rather, it is a business dispute between two companies in the oil and gas industry. TXO was accused of slandering respondents' title to a tract of land-that is, impugning their claim of ownership-in an attempt to win concessions on a pre-existing contract. Although TXO's conduct was clearly wrongful, calculated, and improper, the award in this case cannot be upheld as a reasoned retributive response. Not only is it greatly in excess of the actual harm caused, but it is 10 times greater than the largest punitive damages award for the same tort in any jurisdiction, id., at 5a-8a (listing all recorded punitive damages awards for slander of title affirmed on appeal), and orders of magnitude larger than authorized civil and criminal penalties for similar offenses, see Brief for Petitioner 19, nn. 17-18, and App. to Brief for Petitioner 9a-21a (collecting statutes). By any "objective criteria," Haslip, 499 U. S., at 23, the award is "grossly out of proportion to the severity of the offense" and bears no "understandable relationship to compensatory damages," id., at 22. It is, at first blush, an "extreme resul[t] that jar[s] one's constitutional sensibilities." Id., at 18. That these disproportions might implicate due process concerns the plurality does not deny. Nonetheless, it refuses to "enshrine petitioner's comparative approach in a 'test' for assessing the constitutionality of punitive damages awards." Ante, at 458. I agree with the plurality that, although it might be convenient to establish a multipart test and impose it upon the States, the principles of federalism counsel against such a course. The States should be permitted to "experiment with different methods" of ferreting out impermissible awards "and to adjust these methods over time." Has lip, supra, at 64 (O'CONNOR, J., dissenting). Nonetheless, I see no reason why this Court or any other would wish to disregard such probative evidence. For example, although retribution is a permissible consideration in assessing punitive damages awards, it is quite difficult to determine whether a particular award can be attributed to that goal; retribution resists quantification. Nonetheless, jury awards in similar cases and the civil and criminal penalties created by the legislature for like conduct can give us some idea of the limits on retribution. Thus, a $5,000 punitive damages award on actual damages of $1 may not seem well proportioned at first blush; but if the legislature has seen fit to impose a $50,000 penalty for that very same conduct, the award might be deemed a reasoned retributive response.
This approach, of course, has its limits. Because no two cases are alike, not all comparisons will be enlightening. See ante, at 457-458 (plurality opinion). But recognizing the limits of an approach does not compel us to discard it entirely. I do not see what can be gained by blinding ourselves to the few clear guideposts in an area so painfully bereft of objective criteria. Indeed, JUSTICE STEVENS joined in proposing precisely such an approach to punitive damages under the Eighth Amendment in Browning-Ferris, see 492 U. S., at 301 (O'CONNOR, J., joined by STEVENS, J., concurring in part and dissenting in part). Moreover, courts at common law engaged in similar comparisons. See, e. g., Travis v. Barger, 24 Barb. 614, 629 (N. Y. 1857) (comparing verdicts for similar torts); International & Great Northern R. Co. v. Telephone & Telegraph Co., 69 Tex. 277, 282, 5 S. W. 517, 518 (1887) (comparing ratios). In any event, what the comparisons demonstrate in this case is what one might have suspected from the beginning. This award cannot be justified as a reasoned retributive response, for it is notably out of line with the punishment previously imposed by juries or established by statute for similar conduct.
B
That, however, does not end our inquiry. In some cases, the unusual nature of the award will be explained by the peculiar considerations placed before the jury. Indeed, the plurality asserts that such an explanation exists in this case. The award, the plurality explains, may have been based on the profit TXO anticipated or the harm TXO would have imposed on respondents had its scheme been successful. Ante, at 459-462.
I have no quarrel with the plurality that, in the abstract, punitive damages may be predicated on the potential but unrealized harm to the victim, or even on the defendant's anticipated gain. Linking the punitive award to those factors not only substantially furthers the State's weighty interests in deterrence and retribution, but also can be traced well back in the common law. See, e. g., Benson v. Frederick, 3 Burr. 1846, 97 Eng. Rep. 1130 (K. B. 1766) (Wilmot, J.) (damages for ordering the plaintiff flogged by two drummers not excessive even though disproportionate to plaintiff's actual suffering, as "it was rather owing to the lenity of the drummers than of the [defendant] that the [plaintiff] did not suffer more"). The plurality's theory, however, bears little relationship to what actually happened in this case.
1
The record demonstrates that the potential harm theory is little more than an after-the-fact rationalization invented by counsel to defend this startling award on appeal. The $5 to $8.3 million estimate of potential loss that respondents proffer today appears nowhere in the record. No expert or lay witness testified to the jury about any such figure. No one directed the jury's attention to the technical documents or scattered testimony on which respondents now rely. See ante, at 450-451, n. 10 (plurality opinion). No one told the jury how to pull all those numbers together to calculate such a figure. In fact, the jury never was told that it was permitted to do so.
Respondents did not even present their $5 to $8.3 million estimate to defend the verdict before the West Virginia Supreme Court of Appeals. Nor did that court rely on such an estimate. Its opinion, which the plurality applauds as "thorough," ante, at 465, nowhere suggests that the jury might have based the award on the potential harm to respondents or on TXO's anticipated profit. Rather, its sole reference to potential harm is the "millions of dollars of damages" that might result if TXO repeated its misdeeds against "other victims." 187 W. Va., at 476, 419 S. E. 2d, at 889 (emphasis added). Virtually any tort, however, can cause millions of dollars of harm if imposed against a sufficient number of victims.
Respondents' $5 to $8.3 million estimate appeared for the first time after this Court granted certiorari, having been produced exclusively for our consumption. As the plurality notes, there is every reason to believe that the figure, derived as it is from a series of extrapolations and economic assumptions never presented to the jury and yet untested by adversary presentation, is unrealistic. See ante, at 461. Consequently, the plurality refuses to rely on the figure, instead offering a series of its own estimates. See ante, at 462. These estimates also are speculative, however, as the plurality does not indicate how they were derived or where they are supported in the record. The little evidence regarding potential harm the record does yield, it turns out, is so uncertain and ambiguous that the plurality cannot rely on it, either; to the extent it demonstrates anything at all, it shows respondents' estimate to be exaggerated. See Tr. 100, 103-104.
2
But even if we assume that the plurality's estimates of potential harm are plausible or supported by the evidence, they are, on this record, entirely irrelevant. The question is not simply whether this Court might think the award appropriate in light of its estimate of potential harm. The question is also whether the jury might have relied on such an estimate rather than some impermissible factor, such as a personal preference for the primarily local plaintiffs as compared to the unsympathetic and wealthy out-of-state defendant, as TXO contends. After all, due process does not simply require that a particular result be substantively acceptable; it also requires that it be reached on the basis of permissible considerations. See Has lip, 499 U. S., at 41 (KENNEDY, J., concurring in judgment). In this case, the jury instructions precluded the jury from relying on the potential harm theory the plurality endorses. As a result, that theory can neither explain nor justify the otherwise astonishing verdict the jury returned.
At trial, the jury was instructed to consider numerous factors when setting the punitive damages award, including "'the nature of the wrongdoing, the extent of the harm inflicted, the intent of the party committing the act, the wealth of the perpetrator, as well as any mitigating circumstances.'" Ante, at 463, n. 29 (plurality opinion) (quoting App. 34-35). Nowhere do the instructions mention the alternative measure of potential harm to respondents upon which the plurality relies today.
Of course, the instructions do mention that the goal of punitive damages is deterrence. One therefore might hypothesize that a particularly sophisticated jury would realize that imposing damages in an amount linked to potential harm or the defendant's expected gain might provide appropriate deterrence. One might even go so far as to suppose that the jury would be daring enough to apply that measure, even though the trial court listed numerous factors, including ac tual harm, but made no mention of potential harm. But such speculation has no application in this case, for the jury instructions made it quite clear that deterrence was linked not to an unmentioned factor like potential gain but to a factor the trial court did mention-TXO's wealth:
"'The object of [punitive damages] is to deter TXO Production Corp. and others from committing like offenses in the future. Therefore the law recognizes that to in fact deter such conduct may require a larger fine upon one of large means than it would upon one of ordinary means under the same or similar circumstances.'" Ante, at 463, n. 29 (plurality opinion) (quoting App. 35) (emphasis added).
A reasonable juror hearing these instructions would not have felt free to consider the potential harm or expected gain measures the plurality proposes today.
The two passages the plurality excerpts from closing arguments, see ante, at 461, do not support the plurality's theory. Respondent Tug Fork Land Company's closing argument does mention that TXO thought the wells would produce "'lot[s] of money.'" Ibid. (quoting Tr. 748-749). But that remark had nothing to do with punitive damages. Instead, counsel was addressing the issue of liability: According to him, TXO's desire to obtain all the royalties was the motive for its bad faith conduct. See Tr. 746-749 (TXO slandered respondents' title to lower the value of the property so it could exact concessions or win 100% of royalties by means of a lawsuit). When counsel did discuss the appropriate measure of punitive damages, not once did he mention the potential harm to respondents. Instead, he relied exclusively on TXO's vast wealth:
"His Honor has instructed you that you may award punitive damages and I've indicated to you what punitive damages [are]. Now, just consider the wealth of this corporation. [T]he reason for putting in [expert evidence on TXO's resources] is that's how a jury considers the amount of punitive damages. This is a multi-million dollar corporation-even a billion dollars in assets…. [Think about imposing a punitive award in the range of a] million, twelve million dollars. Those kinds of numbers are not out of line when you talk about a corporation that has assets of something like a billion dollars." Id., at 757-758 (emphases added).
Counsel for respondent Alliance Resources Corp. similarly did not argue that punitive damages should be linked to potential harm. He did mention that TXO anticipated a large profit from its nefarious scheme. See id., at 779-780; ante, at 461 (plurality opinion). But counsel once again made no attempt to quantify TXO's potential gain. Nor did he encourage the jury to base the punitive damages award on TXO's expected profit. Instead, counsel argued only one measure for punitive damages-TXO's wealth:
"A two billion dollar company. Ha[s] earnings of $225,000,000, average. Last year made $125,000,000.00 alone. Last year. Now, what's a good fine for a company like that? A hundred thousand? A million? You can do that if you think it's fair…. " Tr. 781.
The portion of counsel's argument the plurality relies upon, ante, at 461, turns out to be a transition between a discussion of TXO's conduct and a plea for the jury to award punitive damages based exclusively on TXO's wealth. Immediately after delivering the portion of the argument the plurality reproduces-in which counsel told the jury that the punishment should" 'fit'" the scheme and" 'fit the wealth,'" ibid. he asked rhetorically, "Now, how much is the wealth?" Tr. 780. It was then that he told the jury, in great detail, about TXO's vast resources. At no point, however, did counsel ask rhetorically, "Now, how much was the potential profit?" At no point did he answer that question. Nor did he ever suggest that the jury calculate potential harm or base its punitive damages award thereon. Instead, like co counsel before him, he relied exclusively on TXO's wealth. See id., at 781-782.
I am therefore unpersuaded by the plurality's assertion that this award may be upheld based on the potential harm to respondents or TXO's potential gain. That theory was not available to the jury under the court's instructions. It was not one supported by evidence on which the jury might have relied. And it is not one that trial counsel chose to promote. It was instead an after-the-fact rationalization invented by appellate counsel who could not otherwise explain this disproportionate award.
C
There is another explanation for the verdict, but it is not one that permits affirmance. As I read the record in this case, it seems quite likely that the jury in fact was unduly influenced by the fact that TXO is a very large, out-of-state corporation.
In Haslip, this Court considered jury instructions that differed from those used here in two material respects. First, unlike the instructions in Has lip, which did not permit the jury to consider the defendant's wealth, the instructions in this case specifically directed the jury to take TXO's wealth into account. The plurality concedes that introducing TXO's wealth into the calculus "increased the risk that the award may have been influenced by prejudice against large corporations, a risk that is of special concern when the defendant is," as here, "a nonresident." Ante, at 464. Second, the instructions directed the jury to impose punitive damages" 'to provide additional compensation for the conduct to which the injured parties have been subjected.'" Ante, at 463, n. 29 (plurality opinion) (quoting App. 34). The latter instruction, of course, is without legal meaning. Ante, at 464 (plurality opinion) (We do "not understand the reference… to 'additional compensation' "). Plaintiffs are compensated for injuries they have suffered; one cannot speak of