dissenting.
I must dissent from the holding that it is proper to enforce the insurance contract so as to relieve the insured tortfeasor of liability for punitive damages. The majority rejects both logic and precedent in order to reach what is to it a desired result.
One of the most serious problems with the majority’s holding is not even discussed in the majority opinion, and that is that the holding is directly contrary to the established law for this particular award of punitive damages. In Harrell v. Ames, 265 Or 183, 508 P2d 211, 65 ALR3d 649 (1973), involving this accident, this court upheld a judgment based upon a jury verdict that the insured, Mrs. Ames, ought to be liable for punitive damages for her conduct in the sum of $25,000. The jury awarded that amount in response to the following instruction:
"Punitive damages are awarded to the plaintiff in addition to general damages in order to discourage the defendant and others from engaging in wanton misconduct. * * *.
* * jfc *
"If you so decide to award punitive damages, you may properly consider the following three items in fixing the amount:
"First, the character of the defendant’s conduct; second, the defendant’s motive; third, the amount of damages which would be required to discourage the defendant and others from engaging in such conduct in the future. (Emphasis added.)
* * * ❖ ”
In affirming this award against a challenge that it was improper to award punitive damages under the circumstances, we justified our holding as follows:
"Indeed, the fact of common knowledge that the drinking driver is the cause of so many of the more serious automobile accidents is strong evidence in itself to support the need for all possible means of deterring persons from driving automobiles after drinking, includ*220ing exposure to awards of punitive damages in thp event of accidents.” (Emphasis added.) 265 Or at 190,
The majority now holds exactly the opposite for the same award: it holds that the insured should not be exposed to an award of punitive damages but that she should be allowed to protect herself against such exposure by insurance. One must conclude that Harrell v. Ames is now overruled. Even that conclusion, however, would not explain the majority’s result. If Harrell v. Ames was wrong and Mrs. Ames should not be punished by being required to pay $25,000 so that she and others will be deterred from similar future action, then the judgment for punitive damages should not have been entered and should not be enforced, because there is no reason to justify it. The jury was told to award an amount which it thought proper to deter, and it decided on the amount of $25,000, hot the amount of an insurance premium. Despite whát the majority may think that the law should be in the future, it offers no explanation for ignoring the law we have established for this award of punitive damages.
Insofar as the future is concerned, the majority takes a step that is illogical and contrary to our precedent. It removes liability for punitive damages from the person on whom the damages were meant to have a deterrent effect and enforces them against an insurance company on which they can have no deterrent effect. The majority does not seem to realize that logic limits its choices and that the only logical way it could reach the result it desires (allowing plaintiffs to recover punitive damages without requiring defendants to be individually responsible for them) would be to change the rationale for punitive damages. This change would require overruling literally hundreds of cases going back at least as far as Martin v. Cambas, 134 Or 257, 261, 293 P 601 (1930). There are other methods of solving the problem, such as abolishing punitive damages or holding that punitive damages do not deter drunken driving, but these solutions Would not meet with the first part of the majority’s desires.
*221The majority opinion purports to distinguish an Oregon case in which we decided a similar problem and which, on principle, cannot be distinguished from the present case. In Butler v. United Pacific Ins. Co., 265 Or 473, 509 P2d 1184 (1973), plaintiff secured a judgment against an automobile dealer for both actual and punitive damages because of fraud. Plaintiff then brought an action against the dealer’s surety, on the bond required by statute of all dealers, to recover the amount of the judgment. The issue on appeal was whether the punitive damage part of the judgment could be recovered against the surety. This court held it could not be recovered, saying:
"In the present case, unlike [Stirling v.] Dari-Delite [262 Or 359, 498 P2d 753 (1972)], punitive damages were not awarded both as a penalty and to compensate the plaintiff for any expenses, inconvenience, or other injury he suffered. Punitive damages were a penalty assessed against a fraudulent automobile dealer for the purpose of deterring that dealer and others from fraudulent conduct. Punitive damages are a common-law creation and this court has restricted their use as to this purpose. Noe v. Kaiser Foundation Hosp., 248 Or 420, 425, 435 P2d 306 (1967); Davis v. Georgia-Pacific, 251 Or 239, 245, 445 P2d 481 (1968). An award of punitive damages against the surety would not be a likely deterrent.” 265 Or at 477.
The majority now suggests in a footnote1 that its decision does not affect our decision in Butler. This suggestion is particularly odd given that Butler was argued under the assumption that "in this jurisdiction liability insurance carriers are not liable for punitive damages.” 265 Or at 475, n 1. Of course in Butler we were not faced with a contract of insurance, but we were faced with a contract of surety, executed pursuant to a statutory requirement. While the statute might have controlled our decision, we found that it was not sufficiently clear. In reaching a conclusion, we relied instead on the well-established policy behind *222punitive damages, which the majority now rejects: that punitive damages, if they are to be paid at all, ought to be paid by the tortfeasor and not by a surety or insurer. We said:
"We are of the opinion, however, that the better reasoning is that the statute is not clear whether the surety should be held for punitive damages. The purpose of punitive damages is to deter. Requiring the surety to pay a judgment for punitive damages likely will not be a deterrent to automobile dealers; therefore, no recovery for punitive damages should be allowed.” 265 Or at 478.
The portion of the majority opinion which discusses the role of public policy in contract law claims that this case involves the adoption of a new rule of public policy. This simply is not true. There is nothing new about the policy of deterrence in punitive damage awards. See Martin v. Cambas, 134 Or 257, 261, 293 P 601 (1930). Nor is there anything new in the idea that to carry out that policy, punitive damage liability ought not to be shifted, Butler v. United Pacific Ins. Co., supra. The majority opinion then goes on to indicate that courts should be very reluctant to frustrate contracts on principles of public policy. This represents a major departure for this court. See, e.g., Real Good Food v. First National Bank, 276 Or 1057, 557 P2d 654 (1976), which holds on the basis of ¡public policy that a bailee for hire cannot contract against liability for negligent keeping of the goods.
The essential failing of the majority’s "public policy” discussion is its lack of any actual public policy analysis. We ordinarily enforce contracts so' as to protect the reasonable expectations of the contracting parties. In this case we may assume that the insured expected not to be liable for punitive damages, but our analysis should not stop there. "Since it is the task of the law to form and project, as well as mirror and reflect, we should not, as judges, merely recite the expectations and risks without examining the desirability of saddling them upon society.” United States v. White, 401 US 745, 786, 91 S Ct 1122, 28 L Ed 2d 453 *223(1971) (Harlan, J.). I see nothing reasonable or desirable in protecting an insured’s expectation of being above the law of punitive damages.
This court acknowledged in Harrell v. Ames that the insured in the present case could have been subject to criminal sanctions for her conduct. 265 Or at 189. The major purpose of such sanctions would be identical to the purpose of punitive damages, i.e., deterrence. Roshak v. Leathers, 277 Or 207, 211, 560 P2d 275 (1977). I would have said that if the insured’s insurance contract protected her from liability for a criminal fine for her conduct, this court would not enforce the contract, because to do so would eliminate the deterrent effect of the fine. Under the majority’s reasoning, however, the contract would have to be enforced, on the theory that the policy of enforcing the insured’s contract outweighs the policy of deterring her conduct.
The majority states that there is no evidence that insurance against punitive damage liability affects deterrence. This statement is inconsistent with the following statements in Harrell v. Ames, supra, 265 Or at 190-91, written about the same situation:
"It may be debatable whether either awards of punitive damages or the imposition of criminal penalties will effectively deter persons from driving after drinking. However, in the absence of a showing of substantial evidence to the contrary, we are not prepared to hold that law enforcement officials and courts, who have a heavy responsibility in this area, are wrong in their present apparent assumption that both criminal penalties and awards of punitive damages may have at least some deterrent effect in dealing with this serious problem. We are also not aware of any good reasons why punitive damages should not have as much deterrent effect upon this type of wanton and reckless conduct as upon other types of conduct in which awards of punitive damages are traditionally approved by the courts.” (Footnotes omitted.)
The majority offers no explanation of how, if punitive damages have a deterrent effect, the ability to escape *224punitive damages through insurance does not interfere with that deterrence.
The majority apparently fails to understand what it calls the "shift of the burden” argument. The argument was stated by Judge Wisdom in Northwestern National Casualty Company v. McNulty, 307 F2d 432, 440-41 (5th Cir 1962), as follows:
"The policy considerations in a state where, as in Florida and Virginia [and Oregon], punitive damages are awarded for punishment and deterrence, would seem to require that the damages rest ultimately as well [as] nominally on the party actually responsible for the wrong. If that person were permitted to shift the] burden to an insurance company, punitive damages would serve no useful purpose. Such damages do not compensate the plaintiff for his injury, since compensatory damages already have made the plaintiff whole. And there is no point in punishing the insurance company; it has done no wrong. In actual fact, of course, and considering the extent to which the public is insured, the burden would ultimately come to rest not on the insurance companies but on the public, since the added liability to the insurance companies would be passed along to the premium payers. Society would then be punishing itself for the wrong committed by the insured.” .
The point is not, as the majority seems to think, that punitive damages would "punish” insurance companies that contract to pay them. The point is that they would not "punish” anybody and therefore would serve no purpose. The majority does not answer this basic point.
The majority provides an economic analysis of the effect of punitive damages on "well-established” businesses which must be labeled, in the majority’s own words, "naive at least, if not pure fiction.” The majority assumes that such businesses will be able to shift the burden of punitive damage awards to their customers. If the businesses are faced with price competition, this will simply not be true. Even well-established businesses must keep their expensies at a minimum. It is for this reason that considerable effort *225is spent on avoiding additional taxation. I would expect comparable effort to be spent on avoiding punitive damage awards; in short, I would expect punitive damages to have a deterrent effect. If they would not, there is no reason for their existence and they should be abolished.
The majority’s expressions of concern about the possible financial ruin of less affluent defendants seem exaggerated, if not disingenuous. For just this reason the law of punitive damages requires that the wealth of a defendant is a matter to be considered by a jury in setting such damages. Phelan v. Beswick, 213 Or 612, 617-18, 326 P2d 1034 (1958). This consideration tends to protect less affluent defendants. The majority’s holding, however, has no tendency to protect less affluent defendants. Protection by insurance coverage of punitive damages will be available only to those who contract for it, that is, to those who know they need it and are able to pay for it. The majority holding will therefore tend to protect people who have wealth and sophistication in legal matters. Such people will now be able to place themselves above the law of punitive damages.
Perhaps the most puzzling part of the majority opinion is its discussion of the scope of conduct subject to punitive damages. Apparently, this part of the opinion is intended to point out the practical considerations which justify, in the majority’s view, an abandonment of logic and precedent. It cites as examples several cases in which this court approved awards of punitive damages for various conduct. It is not clear whether the majority wishes to overrule these cases. If the majority does not disapprove of these cases, the only explanation of them consistent with the majority opinion is that in each case this court approved an award of punitive damages, presumably for the purpose of deterrence, without intending the defendant to pay it.
In its concluding summary, the majority pays lip *226service to the idea of limiting or eliminating punitive damages. It suggests that this might be a matter best left to the legislature, presumably in its next session two years from now. If there is a problem with the scope of punitive damages in Oregon, it has been created by this court, and this court should attempt to solve it. In the past this court has at least limited punitive damages to cases in which they served the public policy of deterrence. In the present case it abandons that limitation. Instead, it is satisfied to allow the foresighted and affluent to place themselves above the policy of the law of punitive damages through the purchase of insurance, while the less foresighted and less affluent remain subject to the law as before. This is a solution which I cannot, in good conscience, accept.
It is true that some other courts have reached the same result as the majority. Numerically, the cases in the United States are equally divided on the question. See Annotation, 20 ALR3d 343 (1968). The following cases permit the shift of liability for punitive damages from a wrongdoer to an insurer: Price v. Hartford Accident and Indemnity Company, 108 Ariz 485, 502 P2d 522 (1972); Southern Farm Bureau Casualty Ins. Co. v. Daniel, 246 Ark 849, 440 SW2d 582 (1969); Abbie Uriguen Olds. Buick, Inc. v. United States F.I. Co., 95 Idaho 501, 511 P2d 783 (1973); Continental Insurance Companies v. Hancock, 507 SW2d 146 (Ky 1974); Lazenby v. Universal Underwriters Ins. Co., 214 Tenn 639, 383 SW2d 1 (1964); and Dairyland County Mutual Ins. Co. v. Wallgren, 477 SW2d 341 (Tex Civ App 1972). The following cases hold that public policy prohibits such a shift of liability: American Surety Company of New York v. Gold, 375 F2d 523, 20 ALR3d 335 (10th Cir 1966); Northwestern National Casualty Company v. McNulty, 307 F2d 432 (5th Cir 1962); Norfolk & W. Ry. Co. v. Hartford Acc. & Indem. Co., 420 F Supp 92 (ND Ind 1976); Tedesco v. Maryland Casualty Co., 127 Conn 533, 18 A2d 357, 132 ALR 1259 (1941); Nicholson v. American Fire and Casualty *227Ins. Co., 177 So2d 52 (Fla App 1965); Crull v. Gleb, 382 SW2d 17 (Mo App 1964); Padavan v. Clemente, 43 App Div 2d 729, 350 NYS2d 694 (1973); Teska v. Atlantic National Insurance Co., 59 Misc2d 615, 300 NYS2d 375 (1969); and Esmond v. Liscio, 209 Pa Super 200, 224 A2d 793 (1966).
The legal writers are almost unanimous in supporting the position that punitive damage liability should not be shifted by insurance. Gonsoulin, Is an Award of Punitive Damages Covered Under an Automobile or Comprehensive Liability Policy?, 22 SW L J 433 (1968); Logan, Punitive Damages in Automobile Cases, 11 Federation of Insurance Counsel 59 (1960); Note, Exemplary Damages in the Law of Torts, 70 Harv L Rev 517 (1957); Comment, Insurer’s Liability for Punitive Damages, 14 Mo L Rev 175 (1949); Note, Insurance: Liability Insurance: Recovery of Punitive Damages, 14 Okla L Rev 220 (1961); Note, Insurance Coverage and the Punitive Award in the Automobile Accident Suit, 19 U Pitt L Rev 144 (1957); Note, Automobile Liability Insurance and Punitive Damages, 39 Temple L Q 459 (1966); Comment, Punitive Damages and Their Possible Application in Automobile Accident Litigation, 46 Va L Rev 1036 (1960); 46 Iowa L Rev 645, 650 (1961); 40 Mich L Rev 128 (1941).
Upon an examination of all authority, I can find no persuasive reasoning for allowing insurance to shift punitive damage liability in Oregon. Upon an examination of the cases holding that it can be shifted, it is apparent that either the rationale for such damages (when given) is not applicable under Oregon law or the reasoning is not cogent. I shall proceed to examine all rationales advanced by those cases and, in doing so, enlarge in some instances on comments already made.
The first rationale is that punitive damages are really compensatory. As indicated, Oregon has a plethora of cases which hold that the sole basis for punitive damages is deterrence.
*228The second rationale is that punitive damages do not actually deter reckless or wanton driving after drinking, and, therefore, insurance coverage for such damages does not actually interfere with publicjpolicy. As previously indicated, we rejected the basic premise of this reasoning in Harrell v. Ames, supra, 265 Or at 190-91.
The third rationale is that criminal sanctions without punitive damages provide adequate deterrence. We recently rejected this suggestion in Roshak v. Leathers, 277 Or 207, 560 P2d 275 (1977). j
The fourth rationale is that a judgment for punitive damages will deter reckless or wanton driving even if the judgment is paid by an insurance company, since increased insurance rates will punish and detier the offending driver. This is an extremely inefficient way to carry out the policy of determent, involving as it does the imposition on insurance companies (and thus on the public) of financial liability which does not directly further any deterrence purpose. It is patent that little deterrence would result, and we have recognized that minimal deterrence is insufficient. See the previous quotation from Harrell v. Ames, supra, 265 Or at 190.
The fifth rationale is that the possibility of liability in excess of policy limits will adequately deter. Such deterrence would be far less than it would otherwise be if punitive damages could not be covered by insurance, and we have made it clear that the slaughter on the highways makes maximum deterrence necessary. Again, see, Harrell v. Ames, supra, 265 Or at 190. Moreover, this and the other arguments which seek to minimize the deterrence aspect of punitive damages, or to substitute deterrence from some other source, fail to recognize that if punitive damages are not themselves a deterrent, there is no rationale to support them in the first place.
The sixth rationale is that because the terms of the insurance policy are broad enough to protect against *229punitive damages, the insured expected to be so protected, and this expectation should be honored. This reasoning misses the point. We may assume that the parties intended to contract for the coverage of punitive damages, but the issue remains whether public policy permits such a contract to be enforced. In determining public policy, we determine not what people expect but what the public interest demands that they should expect. Were we bound by the insured’s expectation of protection against punitive damages, we could not have affirmed the judgment for punitive damages in Harrell v. Ames, supra. We would have been required to hold that there was no point in awarding a judgment for punitive damages against defendant’s insured since insurance coverage made it impossible to punish her with such a judgment. Cf. Ashcraft v. Saunders, 251 Or 139, 144, 444 P2d 924 (1968).
Plaintiff cites 15 Couch on Insurance § 56:27 (2d ed 1966) and 7 Appleman, Insurance Law and Practice § 4312, in support of this rationale for permitting coverage for punitive damages. Language purporting to be from the 1972 Supplement to the section of Appleman is quoted by the Supreme Court of Idaho in Abbie Uriguen Olds. Buick, Inc. v. United States F. I. Co., supra, 511 P2d at 787:
"* * * It seems strangely inconsistent for an insurer, in one breath, to admit liability for compensatory damages, and then to deny liability for that part of an award claimed attributable to reckless or wanton conduct. He Hí *
I see no strange inconsistency. It appears that the commentator is confused about what makes punitive damages different from compensatory damages for purposes of the issue before us. The difference lies not in the conduct of the insured, as the commentator seems to think. The same conduct gives rise to both compensatory and punitive damages. The difference lies rather in the purpose of the two kinds of damages. The purpose of compensatory damages is to compen*230sate, and this purpose is carried out no matter who is held ultimately responsible for payment. The pürpose of punitive damages, on the other hand, is to deter, and this purpose is not carried out if the one who ultimately pays is an insurer rather than the wrongdoer.
The seventh rationale appears in some states in which respondeat superior liability attaches to employers for punitive damages awarded against their employees. Cf. Farris v. U S. Fidelity & Guaranty, 273 Or 628, 636, 542 P2d 1031 (1975). The rationale is that since courts allow employers to insure against such liability, they should allow individuals to have the same protection against punitive damage awards. Courts which use this reasoning do not seem to realize that the cases which allow employers such protection are based on the premise that there is no point in attempting to deter employers who are not able to prevent wanton or reckless acts by their employees. It does not follow that employers or others should be able to insure themselves against punitive damagtes for their own culpable conduct.
The final rationale appears in Dairyland County Insurance Co. v. Wallgren, supra, in which the Texas court relied on the prescription and approval of insurance policy language by the state insurance commission under a delegation of legislative authority. The court treated the approval of the policy language as a statement of public policy by the legislature through the commission favoring coverage for punitive damages. In Oregon there is no such approval or other indication of a legislatively established public policy.
The majority opinion is a typical example of what happens when a result is forced. The specter of what juries may do under the guise of punitive damages to persons who really have not greatly sinned is úsed to reach a completely illogical result. If the majority opinion is correct that this great danger does exijst, the remedy is not to allow people to insure against the contingency, so as to give windfalls to plaintiffs with *231no deterrent effect, but to abolish punitive damages. I dissent.
Denecke, C. J., concurs in this dissenting opinion.Footnote 22 in the majority opinion.