Opinion
WERDEGAR, J.Plaintiffs, purchasers of a used automobile, sued the manufacturer, Ford Motor Company (Ford), for concealing the automobile’s history of transmission repairs and replacements when reselling the car. Plaintiffs presented evidence of corporate practices by Ford identical or closely similar to the fraud inflicted on them, practices they maintain earned Ford millions of dollars in profit in California every year. The jury found in plaintiffs’ favor and awarded them $17,811.60 in compensatory damages and $10 million in punitive damages. The Court of Appeal, holding Ford could constitutionally be punished in this case only for its fraud on plaintiffs and not for its overall course of conduct, reduced the punitive damages award to $53,435, approximately three times the compensatory damages.
We agree with the Court of Appeal that the $10 million punitive damages award may not, under the circumstances of this case, constitutionally be justified on the basis of disgorgement of profits earned by Ford through its entire course of wrongful conduct toward other consumers. In reducing the punitives to a small multiple of the relatively modest compensatory damages award, however, the Court of Appeal apparently failed to adequately consider that Ford’s fraud was more reprehensible because it was part of a repeated corporate practice rather than an isolated incident. For this reason, we reverse the Court of Appeal’s judgment and remand for that court to conduct again the independent due process review required under State Farm Mut. Auto Ins. Co. v. Campbell (2003) 538 U.S. 408 [155 L.Ed.2d 585, 123 S.Ct. 1513] *1197(State Farm) and BMW of North America v. Gore (1996) 517 U.S. 559 [134 L.Ed.2d 809, 116 S.Ct. 1589] (BMW).
Factual and Procedural Background
In February 1998, plaintiffs Greg and Jo Ann Johnson bought a used 1997 Ford Taurus from a car dealer, Decker Ford (Decker), for $17,411. When Greg Johnson asked about the previous ownership, the salesman told them only that the Taurus had been traded in for a newer model. When he asked to see the Taurus’s repair history, he was shown a computer printout that indicated there had been no significant repairs. The jury found Decker had acted as Ford’s agent in this sales transaction.
In fact, the previous drivers, the McGills, had experienced repeated and seemingly unrepairable difficulty with the car’s transmission after leasing it in late 1996. After at least four trips to the dealership for the transmission problems, one transmission replacement, and an incident in which the transmission locked in low gear on the freeway, the McGills, in July 1997, requested that Ford repurchase the car as a “lemon.”
Ford’s district customer service manager reviewed Decker’s records and decided the automobile did not qualify for mandatory repurchase under California’s lemon law (Civ. Code, §§ 1790-1795.7).1 (The jury later found to the contrary.) Instead, she approved issuance of an “owner appreciation certificate” worth $1,500 on any trade-in at Decker. Though the McGills were never told they had received an owner appreciation certificate from Ford, Decker applied the $1,500 credit to their trade of the Taurus for a new pickup truck, then recovered the $1,500 from Ford.
After Decker resold the Taurus to plaintiffs, they also experienced transmission problems with it. When, in August 1998, Greg Johnson complained that it delayed in shifting and “slammed” into gear, Decker replaced the transmission. In March 1999, the transmission would not shift into reverse; Decker again replaced it. At that point, in discussion with Decker’s service writer, Greg Johnson asked to see and was finally shown the car’s complete repair file, thus learning of the McGills’ earlier problems.
The Johnsons sued Ford and Decker for intentional and negligent misrepresentation and concealment, violations of the Song-Beverly Consumer Warranty Act (Civ. Code, §§ 1790-1795.7) (Lemon Law), the Consumer Legal Remedies Act (Civ. Code, §§ 1750-1784), the unfair competition law (Bus. & Prof. Code, §§ 17200-17210), and the prohibition on false or misleading *1198advertising (Bus. & Prof. Code, § 17500). Plaintiffs settled with Decker prior to trial and, after the jury verdict, voluntarily dismissed their unfair competition and false advertising causes of action against Ford.
In addition to the facts of the Taurus’s repair history and its sale to them, summarized above, plaintiffs at trial presented evidence of Ford’s corporate policies and practices regarding reacquisition of vehicles and issuance of owner appreciation certificates (OAC’s). Ford’s stated policy was that OAC’s—credits of up to $5,000 on trade-ins for new Ford vehicles provided as goodwill to help “satisfy the customer and to restore the customer’s confidence in Ford products”—were to be issued only for vehicles that did not meet the state’s definition of a lemon and therefore were not subject to mandatory reacquisition. But plaintiffs introduced evidence that, in evaluating eligibility, at least some Ford managers employed a narrow concept of what constituted a repair attempt for purposes of applying state lemon laws, including California's, under which a vehicle that cannot be repaired in a “reasonable number” of attempts must be reacquired or replaced. (See § 1793.2, subd. (d)(2).) Specifically, the regional customer service manager who handled the McGills’ complaint and authorized issuance of the OAC testified she interpreted the Ford training and policy materials to provide that an occasion on which the customer brought the vehicle in with a complaint, but the service staff was unable to find or confirm the problem, was not counted as a repair attempt. Ford’s former policy manager for the reacquired vehicle program similarly stated that “[i]f the technician does not replace a part or make an adjustment to the vehicle, and it’s properly documented as no problem found, then I would not count it as a repair.”2
In addition, Ford’s reacquired vehicle program looked almost exclusively to whether a vehicle met the law’s presumption of reasonable repair attempts, based on a specified number of attempts in a certain period (see § 1793.22, subd. (b)), rather than whether the number of attempts was itself reasonable regardless of the presumption (see § 1793.2, subd. (d)(2)). Thus, the 1998 reacquired vehicle program manual repeatedly instructed customer service managers that vehicles meeting “state lemon law presumption[s]” were not eligible for an OAC, stated that a used car would be eligible for an OAC if it “does not meet lemon law presumption,” and gave as examples of ineligible vehicles those with more repair attempts or days out of service than specified under a state’s lemon law presumption. In line with these written policies, the regional customer service manager testified she understood an OAC could not be offered “if the vehicle met the presumption of lemon law,” and the former policy manager explained that “we don’t determine anything by reasonable *1199repair attempts because we cannot define reasonable repair attempts.” By these narrow constructions, Ford allowed itself to issue OAC’s for dealer trade-in of vehicles that arguably should have been reacquired as lemons, thereby avoiding the title branding and additional notice requirements involved in reselling a lemon. (See § 1793.23, subd. (c).)
Ford managers also testified that the company regarded OAC’s as assistance to the customer, not the dealership. Ford thereby avoided the requirement of California law to notify future buyers of defects that led to a vehicle’s reacquisition by the manufacturer or the dealer with manufacturer assistance “in response to a request by the buyer or lessee that the vehicle be either replaced or accepted for restitution because the vehicle did not conform to express warranties,” a requirement that applies even if the vehicle is not reacquired as a lemon. (§ 1793.23, subd. (d) [in such cases, the manufacturer must give the subsequent buyer the “warranty buyback notice” prescribed in § 1793.24].) The reacquired vehicle policy manual specified that while OAC’s were ordinarily mailed to the dealer, in California they were to be sent instead to the customer (though, as noted earlier, no certificate was in fact mailed to the McGills). According to the former policy manager, the policy of mailing to the customer was adopted “[t]o avoid anyone getting the impression we’re trying to assist a dealer.”
Plaintiffs presented further evidence that Ford’s San Francisco and Los Angeles offices issued about 1,200 to 1,400 OAC’s per year in the year of trial and the previous year (2000 and 2001). The average face amount of OAC’s issued over the four previous years was between $2,700 and $3,200. Finally, testimony was given to the effect that the cost of reacquiring a vehicle as a lemon (i.e., the cost of repurchasing or replacing the vehicle less its resale or salvage value) was between $8,500 and $13,500, depending on the year and the method of reacquisition (refund or replacement).
Based on this evidence, plaintiffs’ attorney argued to the jury that Ford saved $6,000 to $10,000 on each OAC for a vehicle that would otherwise have had to be reacquired, and that approximately 1,000 such OAC’s were issued per year to California customers (excluding some issued out of California offices to customers in other states). Counsel estimated Ford’s savings in California from “this whole scheme of owner appreciation certificates”—that is, the practice of issuing OAC’s for vehicles that should have been reacquired as lemons, and of failing to provide warranty buyback notices on all vehicles traded in with OAC’s, thus concealing the vehicles’ defects from subsequent buyers—to be $6 to $10 million per year for 2000 and 2001. He urged the jury, in order to deter Ford from continuing that conduct, to impose punitive damages in an amount that would, at least, take from Ford all those wrongful profits.
*1200The jury, after a single-phase trial on liability and punitive damages, found that Ford, directly and through its agent Decker, had committed fraud by misrepresentation and concealment and had violated the Lemon Law and the Consumer Legal Remedies Act. On the Lemon Law cause of action, the jury found specifically that the McGills’ vehicle qualified for mandatory replacement or restitution because Ford and Decker had been unable to conform it to warranty after a reasonable number of attempts (see § 1793.2, subd. (d)(2)), that Ford reacquired or assisted Decker in reacquiring the vehicle at the McGills’ request because it did not conform to warranty (see § 1793.23, subd. (d)), and that Ford then failed to provide the Johnsons the required warranty buyback notice when they purchased the used vehicle (see § 1793.24). The jury awarded the Johnsons $17,811.60 in compensatory damages on all causes of action; further found by clear and convincing evidence that Ford, through its officers, directors or managing agents, had acted with fraud .or malice; and assessed punitive damages of $10 million.
The trial court entered judgment on the jury’s verdict, subtracting the $100,000 amount of the prior settlement with Decker, and awarded plaintiffs $379,348 in attorney fees on their statutory causes of action.
The Court of Appeal found substantial evidence not only that Ford had fraudulently concealed material facts from the Johnsons by failing to provide them the warranty buyback notice required under section 1793.24, but also that punitive damages against the corporation were justified because “defendant’s entire customer response program was structured precisely to short-circuit lemon law claims whenever defendant plausibly could,” by restrictively interpreting state lemon laws and ignoring the possibility of nonpresumptive lemons.
Though it affirmed the jury’s decision to award punitive damages, the Court of Appeal deemed the amount of the award unconstitutionally excessive. Relying on its own decision in Romo v. Ford Motor Co. (2003) 113 Cal.App.4th 738 [6 Cal.Rptr.3d 793], the court opined that federal law, particularly State Farm, supra, 538 U.S. 408, limits punitive damages to those needed to “punish only the conduct that injured the present plaintiffs.” Punitive damages designed “to punish and deter defendant’s overall course of conduct,” the Court of Appeal held, are not constitutionally permitted. Consequently, the jury could not, as it was invited to do, “cause defendant to disgorge all profit from use of owner appreciation certificates in California over a two-year period.” After reviewing the constitutional guideposts of State Farm, supra, 538 U.S. 408, and BMW, supra, 517 U.S. 559, but without any specific explanation of the amount reached, the Court of Appeal decided that “punitive damages in the amount of $53,435, three times the compensatory damages,” constituted the maximum award consistent with due process and modified the judgment accordingly.
We granted plaintiffs’ petition for review.
*1201Discussion
As we explain at greater length in the companion case of Simon v. San Paolo U.S. Holding Company, Inc. (2005) 35 Cal.4th 1159 (Simon), recent United States Supreme Court decisions require a court reviewing an award of punitive damages for constitutionality to make an independent assessment of the relationship between the award and the factual circumstances of the case. (State Farm, supra, 538 U.S. at p. 418; Cooper Industries, Inc. v. Leatherman Tool Group, Inc. (2001) 532 U.S. 424, 436-443 [149 L.Ed.2d 674, 121 S.Ct. 1678].) This review encompasses three constitutional “guideposts”: “(1) the degree of reprehensibility of the defendant’s misconduct; (2) the disparity between the actual or potential harm suffered by the plaintiff and the punitive damages award; and (3) the difference between the punitive damages awarded by the jury and the civil penalties authorized or imposed in comparable cases.” (State Farm, supra, at p. 418; see BMW, supra, 517 U.S. at p. 575.) The high court’s decisions do not preclude California from imposing civil damages awards “for the sake of example and by way of punishing the defendant” (§ 3294, subd. (a)), though constitutional review using the State Farm/BMW guideposts may, in some circumstances, limit the degree of deterrence California can achieve through awards of punitive damages. (Simon, supra, at pp. 1185-1187.)
Plaintiffs contend the $10 million punitive damages award here was justified as a deterrent measure, for “in cases such as this where the defendant’s misconduct is profit driven, punitive damages which deny a defendant its profit are uniquely appropriate to effect deterrence,” a consideration the Court of Appeal assertedly ignored in reducing the award.3 Ford contends plaintiffs’ “aggregate disgorgement” theory is foreclosed by State Farm, supra, 538 U.S. 408, “which emphasized that due process requires that punitive damages be closely tethered to the defendant’s conduct toward the plaintiffs themselves and the injury to those specific plaintiffs” and that, for the same reason, the award was properly reduced to $53,435. We conclude that the original award cannot be supported on a disgorgement theory, but that the Court of Appeal, in determining the constitutional maximum, may not have adequately considered how the scale and profitability of Ford’s repeated conduct reflects on its reprehensibility.
*1202I. Repeated Wrongful Conduct, Profitability, and Reprehensibility
Neither BMW, supra, 517 U.S. 559, which first drew the contours of the required substantive due process review, nor State Farm, supra, 538 U.S. 408, which elaborated on the reprehensibility and relationship-to-harm criteria (see Simon, supra, 35 Cal.4th at pp. 1180, 1181-1183), states precisely what role evidence of the defendant’s similar wrongful conduct to others plays in the analysis. As both decisions are nonetheless instructive on this question, we begin by reviewing them.
BMW involved an automobile distributor’s nationwide policy of not advising dealers or their customers of predelivery damage to new vehicles when the cost of repair was less than 3 percent of the retail price. (BMW, supra, 517 U.S. at pp. 563-564.) The individual Alabama plaintiff, whose new car had been repainted without his knowledge before he bought it, proved compensatory damages of only $4,000 but, on evidence that nationwide the defendant had sold about 1,000 refinished automobiles without disclosure, was awarded $4 million in punitive damages. (Id. at pp. 564-565.) The Alabama Supreme Court reduced that award to $2 million, in part based on its conclusion that conduct in other jurisdictions should not be considered. (Id. at p. 567.)
The United States Supreme Court reversed and remanded for a new due process analysis by the state court, without itself determining the maximum constitutional award. (BMW, supra, 517 U.S. at pp. 585-586.) The high court’s legal analysis, however, is instructive on the role of the defendant’s practices toward those other than the plaintiff. The court’s central holding in this regard is that instances of the defendant’s similar conduct in states other than Alabama were not properly considered in assessing punitive damages because the conduct was not illegal in all the other states, and “a State may not impose economic sanctions on violators of its laws with the intent of changing the tortfeasors’ lawful conduct in other States.” (Id. at p. 572.)
At the same time, the court made clear it regarded similar conduct by the defendant as potentially relevant to the reprehensibility of the conduct, and hence to the permissible size of an award. To its holding that the state court correctly ignored BMW’s out-of-state conduct in assessing the award, the high court added this footnote: “Of course, the fact that the Alabama Supreme Court correctly concluded that it was error for the jury to use the number of sales in other States as a multiplier in computing the amount of its punitive sanction does not mean that evidence describing out-of-state transactions is irrelevant in a case of this kind. To the contrary, as we stated in TXO Production Corp. v. Alliance Resources Corp., 509 U.S. 443, 462, n. 28 [125 L.Ed.2d 366, 113 S.Ct. 2711, 2722] (1993), such evidence may *1203be relevant to the determination of the degree of reprehensibility of the defendant’s conduct.” (BMW, supra, 517 U.S. at p. 574, fn. 21.)4
The court expanded on this point in its discussion of reprehensibility, the first and “[p]erhaps the most important” of the constitutional guideposts. (BMW, supra, 517 U.S. at p. 575.) Though it ultimately rejected the plaintiff’s contention that the defendant’s nondisclosure of the minor repairs to his car was particularly reprehensible because it was part of a nationwide pattern of tortious conduct (the point was rejected for lack of a showing the practice was generally tortious), the court acknowledged that recidivism increases the wrongfulness of a defendant’s conduct and may justify greater punishment: “Certainly, evidence that a defendant has repeatedly engaged in prohibited conduct while knowing or suspecting that it was unlawful would provide relevant support for an argument that strong medicine is required to cure the defendant’s disrespect for the law. [Citation.] Our holdings that a recidivist may be punished more severely than a first offender recognize that repeated misconduct is more reprehensible than an individual instance of malfeasance.” (Id. at pp. 576-577.)
Seven years later, in State Farm, supra, 538 U.S. 408, the high court again considered the role of a corporation’s practices and policies in an individual lawsuit seeking punitive damages. The court reiterated that tortious conduct toward others could be relevant to the reprehensibility of an individual tort and that conduct in other states where it might not be illegal should not be considered, but additionally distinguished between courses of conduct that were similar to the individual tort and those that were dissimilar. (Id. at pp. 420-424.)
The defendant insurer’s wrongful conduct toward the individual State Farm plaintiffs was its bad faith refusal to settle a third party tort suit against the plaintiffs, its insureds. (State Farm, supra, 538 U.S. at p. 413.) The plaintiffs, however, introduced evidence of other assertedly fraudulent State Farm business practices encompassing many years and many states, most of which “bore no relation to third-party automobile insurance claims.” (Id. at p. 415.) Consequently, the high court complained, the case “was used as a platform to expose, and punish, the perceived deficiencies of State Farm’s operations throughout the country.” (Id. at p. 420.)
This evidence of wide-ranging business practices could not, consistent with due process, be used to show reprehensibility that would support a large *1204($145 million) punitive damages award for two reasons: First, “[a] State cannot punish a defendant for conduct that may have been lawful where it occurred.” (State Farm, supra, 538 U.S. at p. 421 [citing BMW].) Second, and “more fundamentally],” “[a] defendant’s dissimilar acts, independent from the acts upon which liability was premised, may not serve as the basis for punitive damages. A defendant should be punished for the conduct that harmed the plaintiff, not for being an unsavory individual or business.” (Id. at pp. 422-423.)
The punitive damages award in State Farm therefore could not be justified on grounds of recidivism. Quoting BMW s statement that “ ‘[o]ur holdings that a recidivist may be punished more severely than a first offender recognize that repeated misconduct is more reprehensible than an individual instance of malfeasance,’ ” the high court in State Farm added the qualification that “in the context of civil actions courts must ensure the conduct in question replicates the prior transgressions.” (State Farm, supra, 538 U.S. at p. 423.) The State Farm plaintiffs had produced “scant evidence of repeated misconduct of the sort that injured them,” and while “evidence of other acts need not be identical to have relevance in the calculation of punitive damages,” conduct toward others that “had nothing to do with” the tortious conduct toward the plaintiffs could not constitutionally be considered. (Id. at pp. 423-424.)
While both BMW and State Farm were cases in which the evidence state courts had considered of conduct toward others was impermissibly broad, the United States Supreme Court’s analysis in both cases makes clear that due process does not prohibit state courts, in awarding or reviewing punitive damages, from considering the defendant’s illegal or wrongful conduct toward others that was similar to the tortious conduct that injured the plaintiff or plaintiffs. We therefore join the numerous courts holding that a civil defendant’s recidivism remains pertinent to an assessment of culpability.5
*1205The appellate court below, however, opined that punitive damages may not be used “to punish and deter defendant’s overall course of conduct,” seemingly ruling out consideration of the scale and profitability of Ford’s fraudulent conduct toward California consumers. In lieu of further explanation, the lower court referred readers to its contemporaneous decision in Romo v. Ford Motor Co., supra, 113 Cal.App.4th 738 (Romo). We therefore briefly examine the Romo decision to see if the due process principles discussed there compel the Court of Appeal’s conclusion here.
Although involving the same defendant, Romo was not factually similar to this case; it was a defective product suit arising from a fatal vehicle rollover, in which the jury awarded the plaintiffs $5 million in compensatory and $290 million in punitive damages. (Romo, supra, 113 Cal.App.4th at p. 744.) On remand from the United States Supreme Court for reconsideration in light of State Farm, the Court of Appeal reduced the punitive damages award to around $23.7 million. (Romo, supra, at p. 763.) Pertinent to our inquiry is the court’s general theoretical discussion of punitive damages after State Farm.
Drawing heavily on a law review article (Colby, Beyond the Multiple Punishment Problem: Punitive Damages as Punishment for Individual, Private Wrongs (2003) 87 Minn. L.Rev. 583 (hereafter Beyond the Multiple Punishment Problem)), the Romo court distinguished between a “narrow” “historically based” view of punitive damages and the “broad” view the court believed had recently prevailed in California and other jurisdictions (Romo, supra, 113 Cal.App.4th at pp. 748-749) and concluded that the United States Supreme Court had adopted the narrow view as a matter of constitutional doctrine (id. at p. 749). Under the narrow historical approach, the Romo court opined, the punishment imposed was “for the particular affront to the plaintiff, not a broader sanction for an affront to society at large” (id. at p. 747), while the broad modem theory, developed in an era of products liability and other consumer actions, held that punitive damages served to punish and deter the affront “to all affected by the goods or services or, given the reach of the misconduct, the affront... to society as a whole” (ibid.).
Under the narrow view, as Romo explains it, the size of a permissible award was limited by the need for a reasonable relationship to the harm caused the individual plaintiff (Romo, supra, 113 Cal.App.4th at p. 747), regardless of whether such an award would actually deter repetition or imitation of the defendant’s conduct (id. at pp. 750-751). In contrast, under the broad view, “punitive damages awards needed to be based on the overall scope of the wrong in order to punish and deter the mass torts” (id. at p. 747), *1206leading to awards calculated to actually deter a corporate course of action, given the corporation’s profits and financial condition (id. at pp. 748-749).
The Romo court’s analysis does not convince us that the United States Supreme Court, in State Farm, adopted wholesale the “historical” view of punitive damages outlined in the Colby article (Beyond the Multiple Punishment Problem, supra, 87 Minn. L.Rev. 583) as a constitutional rule binding on the states. The article is not cited in the high court’s decision; nor does the decision contain any explicit references to “broad” and “narrow,” or “modem” and “historical,” theories of punitive damages. The high court does discuss the history of single-digit ratios between compensatory damages and civil penalties (State Farm, supra, 538 U.S. at p. 425), but does not relate its presumptive preference for single-digit ratios (see Simon, supra, 35 Cal.4th at p. 1182) to a requirement that the states adopt a restrictive historical view of the purposes of punitive damages.
More important, we are not convinced the high court’s precedents dictate that states take such a narrow view as to “what is to be deterred” through punitive damages (Romo, supra, 113 Cal.App.4th at p. 747) as to blind state juries and courts to the state’s public interest in deterring a wrongful course of conduct. Indeed, the court’s analysis in BMW, supra, 517 U.S. 559, expressly affirms a state’s constitutional freedom to use punitive damages as a tool to protect the consuming public, not merely to punish a private wrong. Alabama, the BMW court explained, could legitimately use punitive damages to punish and deter the defendant’s unlawful conduct, thereby furthering its interest in “protecting] its citizens [from] deceptive trade practices.” (Id. at p. 568.) To that end, a proper award of punitive damages would be one “supported by the State’s interest in protecting its own consumers and its own economy.” (Id. at p. 572.)
State Farm, in turn, did not bar deterrence of future public injuries as a goal of punitive damages. The court reiterated its statement in BMW that “ ‘[pjunitive damages may properly be imposed to further a State’s legitimate interests in punishing unlawful conduct and deterring its repetition’ ” (State Farm, supra, 538 U.S. at p. 416) and did not limit the concept to punishment and deterrence purely on behalf of the plaintiff. In elaborating on BMWs reprehensibility guidepost, the court in State Farm noted that conduct involving “repeated actions” was worse than, and could be punished more severely than, conduct limited to “an isolated incident.” (State Farm, supra, at p. 419.)6
*1207To be sure, State Farm requires reasonable proportionality between punitive damages and actual or potential harm to the plaintiff. But what ratio is reasonable necessarily depends on the reprehensibility of the conduct, “ ‘the most important indicium of the reasonableness of the award’ ” (State Farm, supra, 538 U.S. at p. 419), which in turn is influenced by the frequency and profitability of the defendant’s prior or contemporaneous similar conduct. As the high court has recognized, that a defendant has repeatedly engaged in profitable but wrongful conduct tends to show that “strong medicine is required” to deter the conduct’s further repetition. (BMW, supra, 517 U.S. at p. 577; see Kemp v. American Telephone & Telegraph Company (11th Cir. 2004) 393 F.3d 1354, 1363 [“large-scale corporate malfeasance,” involving collection of almost $300,000 in illegal gambling debts, “merited a substantial penalty” under high court’s guideposts].)
In certain cases, as we explain in Simon, supra, 35 Cal.4th at page 1187, “the state may have to partly yield its goals of punishment and deterrence to the federal requirement that an award stay within the limits of due process.” The scale and profitability of a course of wrongful conduct by the defendant cannot justify an award that is grossly excessive in relation to the harm done or threatened, but scale and profitability nevertheless remain relevant to reprehensibility and hence to the size of award warranted, under the guideposts, to meet the state’s interest in deterrence. BMW and State Farm limit the size of individual awards but leave undisturbed the states’ “discretion” (State Farm, supra, 538 U.S. at p. 416) in use of punitive damages generally. Nothing the high court has said about due process review requires that California juries and courts ignore evidence of corporate policies and practices and evaluate the defendant’s harm to the plaintiff in isolation.
California law has long endorsed the use of punitive damages to deter continuation or imitation of a corporation’s course of wrongful conduct, and hence allowed consideration of that conduct’s scale and profitability in determining the size of award that will vindicate the state’s legitimate interests.7 We do not read the high court’s decisions, which specifically *1208acknowledge that states may use punitive damages for punishment and deterrence, as mandating the abandonment of that principle.
II. Disgorgement of Aggregate Profits from Repeated Conduct
To recognize that recidivism remains relevant is not to approve plaintiffs’ aggregate disgorgement theory of punitive damages. We must consider directly the basis and fairness of plaintiffs’ approach.
Plaintiffs’ aggregate disgorgement theory should be distinguished from simple return of ill-gotten gains earned from an individual plaintiff. Removal of any profits the defendant has earned by a wrongful act is a logical step toward deterring its repetition or imitation. “A gain-based measure of this sort sends a clear signal to defendants that such misconduct does not pay and, thus, serves the deterrent function of punitive damages.” (Cummings Medical Corp. v. Occupational Medical Corp. (1992) 10 Cal.App.4th 1291, 1300 [13 Cal.Rptr.2d 585, 13 Cal.Rptr. 2].)8 But an approach calculating punitive damages in an individual tort case by the profits made through similar torts against hundreds or thousands of other individuals creates possibilities for *1209unfairness—to the defendant and other possible claimants both—which may be of constitutional dimension.
The high court has observed that an award punishing the defendant for dissimilar hypothetical claims of others “creates the possibility of multiple punitive damages awards for the same conduct; for in the usual case nonparties are not bound by the judgment some other plaintiff obtains.” (State Farm, supra, 538 U.S. at p. 423.)9 Critics of aggregate disgorgement or aggregate harm as theories of punitive damages argue the same danger exists when the hypothetical claims by others are for conduct similar to that which injured the individual plaintiff. (See, e.g., Owens-Coming Fiberglas Corp. v. Malone (Tex. 1998) 972 S.W.2d 35, 50 [41 Tex. Sup. Ct. J. 877] [concluding that “repeatedly imposing punitive damages on the same defendant for the same course of wrongful conduct may implicate substantive due process constraints”].)10
Plaintiffs argue that over-punishment may be avoided by permitting a defendant to present evidence of past punitive damages awards for the same conduct, which might be considered either by the jury or by courts reviewing the jury’s award.11 But the exact basis for a prior punitive damages award *1210will not always be clear, and even where it is proven that the defendant has already been punished severely for a course of conduct that included harm to the current plaintiff, there is no guarantee the jury or court will agree to deny the plaintiff before them recovery of punitive damages simply because another plaintiff, in another court, has already recovered. (See Roginsky v. Richardson-Merrell, Inc. (2d Cir. 1967) 378 F.2d 832, 840.) Permitting an aggregate recovery followed by credits in future cases could, moreover, unfairly deprive subsequent claimants of their own recoveries, as well as present a problem of “successive prosecution” in which a defendant that loses a single case would also lose the benefit of all previous victories against the same claim of misconduct. (Beyond the Multiple Punishment Problem, supra, 87 Minn. L.Rev. at pp. 594—597.)
Nor does an aggregate disgorgement theory fit easily within the BMW!State Farm guideposts. Although the scale and profitability of a corporate practice are related to its reprehensibility, gains made over some period of time and the harm or potential harm to an individual plaintiff are not necessarily related. An award of disgorgement of all profits from a group of transactions similar to that which harmed the plaintiff (but not defined through the procedural limits of a class action) is therefore likely to be disproportionate to the individual plaintiff’s compensatory award.
Finally, and most pertinent to this case, an individual plaintiff resting his or her claim for a large punitive damages award on profits earned from transactions with a large class of similar claimants, but proceeding without the formalities of a class action, can hope to recover without ever proving the specifics of those “hypothetical claims.” (State Farm, supra, 538 U.S. at p. 423.) In a class action, once the issues common to the class have been tried, and assuming some individual issues remain, each plaintiff must still by some means prove up his or her claim, allowing the defendant an opportunity to contest each individual claim on any ground not resolved in the trial of common issues. (See Sav-on Drug Stores, Inc. v. Superior Court (2004) 34 Cal.4th 319, 334-335, 339-340 [17 Cal.Rptr.3d 906, 96 P.3d 194].) Here, the Johnsons, as individual plaintiffs, proved only the facts of Ford’s tortious transaction with them, yet they sought and obtained disgorgement of Ford’s estimated earnings on a thousand or more other transactions without proof that each of the others was also tortious.
Plaintiffs claim to have justified the punitive damages award by proving that “[i]n the years 2000-2001, Ford issued about 1,300 . . . OACs per year in California,” that “each OAC transaction represents a potentially dangerous, defective and unrepaired vehicle which is resold to an unsuspecting consumer without disclosure of material facts,” and that by issuing OAC’s instead of *1211replacing or repurchasing these vehicles Ford saves “$6,000 to $10,000 per vehicle.” Plaintiffs’ proof, however, suffers from several major deficiencies.
First, plaintiffs nowhere explain the pertinence of the two-year period, 2000-2001, they use to estimate profits. Ford issued an OAC to the McGills for their Taurus in 1997, and plaintiffs bought that vehicle in 1998. Plaintiffs apparently assume, but do not point the court to any evidence, that Ford’s OAC and reacquired vehicle policies and practices were uniform in nature and number from 1997 to 2001. Nor do plaintiffs explain why Ford’s profits should be estimated over a two-year period.
More important, plaintiffs offered no proof that all OAC transactions—in any period—involved defective vehicles subject to California’s Lemon Law, much less that all such vehicles were “dangerous” or “unrepaired.”12 To the contrary, Ford introduced evidence that OAC’s were used to address a variety of customer dissatisfactions. According to Ford’s former policy manager for reacquired vehicles, “we have customers that are concerned about a lot of things that aren’t defects. They could be concerned about a normal attribute of the vehicle . . . and in their perception that’s a concern or problem. It’s not necessarily a defect.” For a vehicle reacquired through use of an OAC for reasons other than defect, our law demands no notice to a subsequent purchaser. (§ 1793.23, subds. (c)-(e).)
Even a vehicle with a defect is not necessarily a lemon. A “nonconformity” requiring the vehicle’s refund or replacement under our law must “substantially impair[] the use, value, or safety of the new motor vehicle.” (§ 1793.22, subd. (e)(1).) Not every customer complaint about a new car, or even every valid customer complaint, rises to that level. And customers and manufacturers frequently disagree about whether a defect has been repaired' or a reasonable number of attempts have been made. A Ford manager testified such a disagreement was “a good opportunity to use an owner appreciation certificate.” Plaintiffs, seemingly, would have us assume the customer is always right in such disputes, an assumption we cannot make.
A vehicle that is not a lemon but that is reacquired in response to customer warranty complaints must carry a warranty buyback notice (§ 1793.23, subds. (d), (e)), but its reacquisition under the Lemon Law is not required. Plaintiffs’ $6,000 to $10,000 savings estimate, which is derived from comparison of lemon buyback costs with OAC face values, is therefore inapplicable to such a vehicle. Plaintiffs introduced no evidence as to how many unrepaired, defective OAC trade-ins fell into this category.
*1212Nor can we assume that in every other case in which a vehicle traded in with an OAC was resold, the new buyer was kept entirely in the dark regarding previous repairs and repair attempts. In plaintiffs’ case, Ford’s own fraudulent concealment—its failure to provide the required historical notices—was successful because Decker’s salesman also concealed, and affirmatively misrepresented, the Taurus’s repair history. But plaintiffs did not show that California Ford dealers always, or generally, conceal and lie about the repair history of used cars they sell.
We do not mean to suggest Ford’s fraud on plaintiffs was unique. Ford’s reacquired vehicle manual stated that the principal use for OAC’s was to “satisfy customers . . . who have lost confidence in a repaired vehicle.” A large number of OAC’s therefore probably involved vehicles with serious defects. When taken in conjunction with the evidence that Ford maintained its OAC’s did not assist dealers to reacquire vehicles and interpreted other Lemon Law requirements narrowly, this stated use supports an inference that in some number of cases in addition to plaintiffs’ own, perhaps a large number of other cases, OAC’s were used to evade Lemon Law requirements, and sales of OAC-traded vehicles were made without the proper historical disclosures, resulting in significant savings to Ford.13 In some subset of those cases, a dealer may also have concealed and lied about the vehicle’s history, preventing the new buyer from learning the truth. But one cannot infer that this fraudulent practice occurred in all cases of OAC-traded vehicles. As plaintiffs’ estimate of Ford’s savings on each OAC transaction depends on assumptions that each such transaction was for a vehicle that should have been reacquired as a lemon and thus should have carried with it a statutory notice, and that each subsequent buyer of an OAC-traded vehicle was defrauded, predicates plaintiffs failed to prove, their attempt to estimate aggregate profits from fraudulent transactions similar to theirs also fails.
We need not decide whether a plaintiff could ever, consistent with due process, justify the size of an award on a total profits basis. Our independent, de novo review of the record, required by due process (State Farm, supra, 538 U.S. at p. 418; Cooper Industries, Inc. v. Leatherman Tool Group, Inc., supra, 532 U.S. at pp. 436-443; Simon, supra, 35 Cal.4th at p. 1172), demonstrates that these plaintiffs’ attempt to calculate punitive damages on this basis was fatally deficient. The Court of Appeal correctly determined the award here could not be upheld on a disgorgement theory.
*1213Conclusion
Although the Court of Appeal correctly rejected the aggregate disgorgement approach, in concluding that the maximum award consistent with due process is $53,435, an amount approximately three times the compensatory damages, the lower court appears not to have properly considered the evidence of Ford’s policies and practices, and their scale and profitability. As we explained earlier (pt. I, ante), a defendant’s recidivism is relevant to the reprehensibility of its conduct. To the extent the evidence shows the defendant had a practice of engaging in, and profiting from, wrongful conduct similar to that which injured the plaintiff, such evidence may be considered on the question of how large a punitive damages award due process permits. Although the lower court discussed Ford’s policies in addressing reprehensibility—noting “it is reprehensible for a regulated manufacturer to implement a scheme that intentionally undermines the protections granted consumers by state law”—the court gave no express weight, in its assessment of the constitutional maximum, to the profitability of that scheme to Ford or the scale at which Ford pursued it. The Court of Appeal’s reliance on its Romo decision suggests it incorrectly believed such weighing was constitutionally precluded.
Nor does the lower court’s discussion of the remaining two State Farm/BMW guideposts explain the drastic reduction ordered. Regarding the ratio guidepost, the court merely observed that “a higher ratio of punitive to compensatory damages” was constitutional here because the compensatory damages were strictly economic. Concerning the comparable-penalties guidepost, the lower court held only that the specification in the Lemon Law of a maximum twice-damages civil penalty for willful violations (§ 1794, subd. (c)) did not so limit punitive damages for fraudulent misrepresentation. In short, the Court of Appeal’s discussion of the last two State Farm/BMW guideposts lacks a justification for restricting punitive damages to three times the compensatory award. This, together with the court’s reliance on its Romo decision, which incorrectly suggests that due process requires appellate review that is blind to the state’s interest in punishing and deterring a wrongful corporate practice, leads us to conclude the lower court may have given insufficient, if any, weight to the scale and profitability of Ford’s fraudulent conduct. As we cannot be sure the lower court made its decision under a correct understanding of the law, a remand for a new determination of the maximum constitutional award is appropriate.
Disposition
The judgment of the Court of Appeal is reversed, and the matter is remanded to that court for further proceedings consistent with our opinion.
George, C. J., Kennard, J., Chin, J., and Moreno, J., concurred.
All further statutory references are to the Civil Code unless otherwise specified.
Ford’s narrow understanding of a repair attempt has been rejected by an appellate court in a Lemon Law case against another manufacturer. (Oregel v. American Isuzu Motors, Inc. (2001) 90 Cal.App.4th 1094, 1103-1104 [109 Cal.Rptr.2d 583].)
Although plaintiffs also introduced evidence at trial of Ford’s worldwide net income, they rely in this court solely on the theory of disgorgement of profits from Ford’s OAC practices in California.
In the cited passage from TXO Production Corp. v. Alliance Resources Corp., supra, 509 U.S. at page 462, footnote 28, the court observed that in an earlier decision, Pacific Mut. Life Ins. Co. v. Haslip (1991) 499 U.S. 1, 21-22 [113 L.Ed.2d 1, 111 S.Ct. 1032], it had approved Alabama punitive damages law permitting consideration of “the existence and frequency of similar past conduct.”
See, e.g., Diamond Woodworks, Inc. v. Argonaut Ins. Company (2003) 109 Cal.App.4th 1020, 1054, footnote 34 [135 Cal.Rptr.2d 736] (jury may consider conduct “similar or bearing a relationship to” that which injured the plaintiff); Williams v. ConAgra Poultry Company (8th Cir. 2004) 378 F.3d 790, 797 (to be properly considered, “recidivist conduct must be factually as well as legally similar to the plaintiff’s claim”); Continental Trend Resources v. OXY USA, Inc. (10th Cir. 1996) 101 F.3d 634, 638-639 (evidence showing the defendant “used some of the same tactics” on others as on the plaintiffs properly considered under BMW); Planned Parenthood of the Columbia/Willamette, Inc. v. American Coalition of Life Activists (D.Or. 2004) 300 F.Supp.2d 1055, 1060, footnote 3 (repeated misconduct “demonstrates defendants’ willingness to violate the law and ignore court rulings”); Bocci v. Key Pharmaceuticals, Inc. (Or.Ct.App. 2003) 189 Ore.App. 349 [76 P.3d 669, 674], modified on other grounds and adhered to as modified, 190 Ore.App. 407 [79 P.3d 908] (the defendant’s nationwide misconduct in disseminating false and misleading information, similar to its conduct that injured the plaintiff, was *1205properly considered); Trinity Evangelical Lutheran Church v. Tower Ins. Co. (2003) 2003 WI 46 [661 N.W.2d 789, 801-802, 261 Wis.2d 333] (insurer’s repeated violation of particular duty to insureds enhances reprehensibility).
To consider the defendant’s entire course of conduct in setting or reviewing a punitive damages award, even in an individual plaintiffs lawsuit, is not to punish the defendant for its conduct toward others. An enhanced punishment for recidivism does not directly punish the earlier offense; it is, rather, “ ‘ “a stiffened penalty for the last crime, which is considered to be *1207an aggravated offense because a repetitive one.” ’ ” (Ewing v. California (2003) 538 U.S. 11, 25-26 [155 L.Ed.2d 108, 123 S.Ct. 1179].) In response to constitutional challenges to recidivist punishment, for example as ex post facto laws, “[t]he uniform answer has been that it is the second or subsequent offense which is punished, not the first.” (People v. Biggs (1937) 9 Cal.2d 508, 512 [71 P.2d 214].) By placing the defendant’s conduct on one occasion into the context of a business practice or policy, an individual plaintiff can demonstrate that the conduct toward him or her was more blameworthy and warrants a stronger penalty to deter continued or repeated conduct of the same nature.
See section 3295, subdivision (a)(1) (contemplating introduction by a plaintiff seeking punitive damages of evidence as to “[t]he profits the defendant has gained by virtue of the wrongful course of conduct of the nature and type shown by the evidence”); Adams v. Murakami (1991) 54 Cal.3d 105, 116, footnote 7 [284 Cal.Rptr. 318, 813 P.2d 1348] *1208(recognizing “the profitability of the defendant’s misconduct” as one measure of the ability to pay punitive damages); Egan v. Mutual of Omaha Ins. Co. (1979) 24 Cal.3d 809, 820 [169 Cal.Rptr. 691, 620 P.2d 141] (a principal purpose of punitive damages is “to deter acts deemed socially unacceptable and, consequently, to discourage the perpetuation of objectionable corporate policies”); Neal v. Farmers Ins. Exchange (1978) 21 Cal.3d 910, 928 [148 Cal.Rptr. 389, 582 P.2d 980] (size of deterrent needed depends on the defendant’s financial condition); id. at page 929, footnote 14 (award large enough to cause the defendant insurer “to lose business to those whose practices have not been” tortious serves deterrence goal, “resulting in an ultimate benefit to insurance consumers as a whole”); Grimshaw v. Ford Motor Co. (1981) 119 Cal.App.3d 757, 819-820 [174 Cal.Rptr. 348] and footnote 14 (award not excessive in light of, among other considerations, the fact that Ford’s malicious conduct “endangered the lives of thousands of Pinto purchasers” and “the profitability of the conduct”).
Under recent legislation, in effect only until July 1, 2006, 75 percent of any punitive damages awarded by final judgment is to be deposited in a state fund, to be appropriated for public purposes. (§ 3294.5, subds. (b), (i).) This would appear to confirm that punitive damages in California are imposed to address a public, not merely a private, wrong. The Legislature, however, has declared that the statute “shall not be construed or interpreted in any way to establish any policy . . . regarding the.award of punitive damages.” (Id., subd. (a).)
See also Pacific Mut. Life Ins. Co. v. Haslip, supra, 499 U.S. at page 22 (factors considered by state reviewing court, including “the profitability to the defendant of the wrongful conduct [][]... provide for a rational relationship in determining whether a particular award is greater than reasonably necessary to punish and deter”); Mallor & Roberts, Punitive Damages: Toward a Principled Approach (1980) 31 Hastings L.J. 639, 667 (“Where the defendant has engaged in wrongful conduct for a profit, the award of punitive damages should remove the profit incentive”); Model Punitive Damages Act (Final Draft 1996) section 7(a)(5), page 20 (available at http://www.law.upenn.edu/bll/ulc/mpda/finaldft.pdf [as of June 16, 2005]) (factors relevant to amount of punitive damages should include “any profit or gain, obtained by the defendant through the wrongful conduct, in excess of that likely to be divested by this and any other actions against the defendant for compensatory damages or restitution”).
The high court in BMW, in a footnote to its discussion of the Alabama Supreme Court’s analysis, observed that the $2 million figure reached by the state court appeared anomalous in light of that court’s reasoning and evidence showing only 14 of the refinished vehicles were sold in Alabama: “In light of the Alabama Supreme Court’s conclusion that (1) the jury had computed its award by multiplying $4,000 by the number of refinished vehicles sold in the United States and (2) that the award should have been based on Alabama conduct, respect for the error-free portion of the jury verdict would seem to produce an award of $56,000 ($4,000 multiplied by 14, the number of repainted vehicles sold in Alabama).” (BMW, supra, 517 U.S. at p. 567, fn. 11.)
Contrary to the argument of an amicus curiae, we do not read this footnote passage, appearing not in the high court’s legal discussion but in its recitation of the procedural background, as a definitive endorsement of using total in-state profits to calculate punitive damages. The high court’s point seems to be that the state court’s choice of a $2 million award makes little sense under the state court’s own reasoning; it ultimately remanded to the state court not for imposition of a $56,000 punitive damages award but instead for “an independent determination by the Alabama Supreme Court of the award necessary to vindicate the interests of Alabama consumers,” or, if necessary, a new trial. (BMW, supra, 517 U.S. at p. 586.)
See also Williams v. ConAgra Poultry Company, supra, 378 F.3d at page 797 (“Punishing systematic abuses by a punitive damages award in a case brought by an individual plaintiff, however, deprives the defendant of the safeguards against duplicative punishment that inhere in the class action procedure”); cf. Korea Supply Co. v. Lockheed Martin Co. (2003) 29 Cal.4th 1134, 1151 [131 Cal.Rptr.2d 29, 63 P.3d 937] (permitting a remedy of nonrestitutionary disgorgement under the unfair competition law “would expose defendants to multiple suits and the risk of duplicative liability without the traditional limitations on standing”).
See Stevens v. Owens-Corning Fiberglas Corp. (1996) 49 Cal.App.4th 1645, 1661 [57 Cal.Rptr.2d 525]; Grimshaw v. Ford Motor Co., supra, 119 Cal.App.3d at page 812; see, e.g., Owens-Corning Fiberglas Corp. v. Malone, supra, 972 S.W.2d at pages 38-40, 52-54 (the *1210defendant offered posttrial evidence of previously paid awards and other costs of asbestos litigation, which the appellate court considered in an aggregate excessiveness analysis).
While Ford did not order dealerships that acquired trade-ins using OAC’s to make repairs on the vehicles, plaintiffs presented no evidence that dealerships generally failed to do so.
We express no view regarding the Court of Appeal’s statements that Ford’s reacquired vehicle program was “structured precisely to short-circuit lemon law claims” and that the present transaction was a “typical” use of an OAC. Ford did not petition for review of the lower court’s holding that sufficient evidence of fraud existed to justify an award of punitive damages, in support of which the quoted statements were made.