Tebo v. Havlik

418 Mich. 350 (1984) 343 N.W.2d 181

TEBO
v.
HAVLIK
BURNS
v.
CARVER

Docket Nos. 68033, 70881. (Calendar Nos. 10, 11).

Supreme Court of Michigan.

Argued April 6, 1983. Decided February 6, 1984.

Smith, Bovill, Fisher, Meyer & Borchard, P.C. (by James T. Borchard and David B. Meyer), for plaintiffs Tebo.

Philip S. Della Santina for defendants Brimmer and Forbes.

Smith & Brooker, P.C. (by Michael J. Huffman, Thomas A. Connolly, and Robert A. Jarema), for defendant Denter.

Lopatin, Miller, Freedman, Bluestone, Erlich, Rosen & Bartnick (by Stephen G. Silverman and Richard E. Shaw) for plaintiff Burns.

Grenn & Grenn (by Frederick M. Grenn) for defendants Carver and Chase.

Amicus Curiae:

Ulanoff, Ross, Summer & Wesley, P.C. (by Stuart A. Ulanoff and Jerald R. Lovell), for Stonewall Insurance Company.

Decided February 6, 1984. Rehearing denied as to Tebo, 419 Mich. 1201.

BRICKLEY, J.

These cases involve the question whether this Court's decision in Putney v Haskins, 414 Mich. 181; 324 NW2d 729 (1982), interpreting the name and retain provision of the dramshop act, is to be given retroactive application.

In Burns, plaintiff filed her complaint on February 26, 1981, alleging that plaintiff's decedent died as a result of injuries sustained in an automobile collision with Richard Piornack, II. Plaintiff also alleged that appellees Vincent Carver and Brian Chase, owners of Jock's Pub, were liable for serving intoxicating beverages to Piornack in violation of the dramshop act, MCL 436.22; MSA 18.993. On May 21, 1982, plaintiff executed a "Release and Settlement of Claim" with Piornack. The agreement provided that in exchange for $20,000 Piornack was to be "retained" in the action, but released from all liability. Following this Court's decision in Putney, appellees moved for and were granted summary judgment. On February 18, 1983, we granted plaintiff's application for leave to appeal prior to decision by the Court of Appeals. 417 Mich. 887 (1983).

In Tebo, plaintiff filed her complaint on January 17, 1977. She alleged that she suffered injuries as a result of a collision with an automobile driven by Edward Havlik. An amended complaint was filed, naming as defendants Steven Brimmer and Gerald Forbes, owners of the Long Branch Bar, and Fred Denter, owner of the Oakley Liquor Bar, predicating their liability on the provisions of the dramshop act. On August 4, 1977, plaintiff and Havlik entered into a "Covenant Not to Levy Execution on Judgment in Pending Action", under which Havlik was released of all liability in excess of $50,000, but he was expressly "retained" in the action. Denter's motion for summary judgment on the grounds that plaintiff had failed to retain Havlik as required by the dramshop act was denied on December 7, 1978. Denter subsequently brought a motion in limine to preclude plaintiff from submitting evidence of any damages for which plaintiff received compensation under the no-fault act, MCL 500.3101 et seq.; MSA 24.13101 et seq. The trial court granted this motion and was reversed on interlocutory appeal. 109 Mich. App. 413; 311 NW2d 372 (1981). Putney was decided while Denter's application for leave to appeal was pending. Denter moved for summary judgment in this Court pursuant to GCR 1963, 865.1(7). We granted leave to appeal to consider both the motion for summary judgment and the decision of the Court of Appeals. 417 Mich. 887 (1983).

In Putney v Haskins, supra, p 188, this Court construed the "name and retain" provision of the dramshop act. We found in the statutory language a legislative mandate that the allegedly intoxicated person be "retained as an interested party defendant until the litigation [is] concluded". Only by requiring the allegedly intoxicated person to remain at risk could the name and retain provision's purpose of preventing fraud and collusion be completely fulfilled. As we stated:

"One of the ways the ``name and retain' provision prevents fraud and collusion is by ensuring that the defendant will have a direct financial stake in personally testifying, examining witnesses, and arguing that he did not act in a negligent manner. Once the defendant's liability is fixed and limited, he has no incentive to produce witnesses or testimony tending to prove that he was not ``visibly intoxicated' on the date in question. The dramshop defendant may have much more difficulty in identifying, locating, and obtaining favorable testimony from the defendant's friends or acquaintances who observed him at relevant times."

The agreements in the present cases obviously do not retain the allegedly intoxicated persons within the meaning of the dramshop act. Liability has been fixed. If Putney is deemed to be retroactive, plaintiffs' actions must be dismissed.

Although it has often been stated that the general rule is one of complete retroactivity, this Court has adopted a flexible approach.^[1]

"``It is evident that there is no single rule of thumb which can be used to accomplish the maximum of justice in each varying set of circumstances. The involvement of vested property rights, the magnitude of the impact of decision on public bodies taken without warning or a showing of substantial reliance on the old rule may influence the result.' Williams v Detroit, 364 Mich. 231, 265-266; 111 NW2d 1 (1961) (opinion of Justice EDWARDS in which Justices TALBOT SMITH, T.M. KAVANAGH and SOURIS concurred).

"The benefit of flexibility in opinion application is evident. If a court were absolutely bound by the traditional rule of retroactive application, it would be severely hampered in its ability to make needed changes in the law because of the chaos that could result in regard to prior enforcement under that law." Placek v City of Sterling Heights, 405 Mich. 638, 665; 275 NW2d 511 (1979).

Appreciation of the effect a change in settled law can have has led this Court to favor only limited retroactivity when overruling prior law. Thus, when the doctrine of imputed negligence was overruled in Bricker v Green, 313 Mich. 218; 21 NW2d 105 (1946), the decision was applied only to the case before the Court and to pending and future cases. When the doctrine of charitable immunity was overruled in Parker v Port Huron Hospital, 361 Mich. 1; 105 NW2d 1 (1960), the retroactive effect of the decision was limited to the parties before the Court. Even where statutory construction has been involved, this Court has limited the retroactivity of a decision when justice so required. See Gusler v Fairview Tubular Products, 412 Mich. 270; 315 NW2d 388 (1981); Franges v General Motors Corp, 404 Mich. 590; 274 NW2d 392 (1979).

The question before us is whether our interpretation of a statute should be applied retroactively to the statute's effective date. In Putney, we found the clear import of the statute to be to require the plaintiff to name and retain the allegedly intoxicated person at risk. Were Putney a case of first impression in the Michigan courts, we would hold that the statutory language gave plaintiffs no reason to believe that the settlements entered into would comply with the "retain" portion of the statute. Putney, however, was not a case of first impression in the Michigan courts.

In Buxton v Alexander, 69 Mich. App. 507; 245 NW2d 111 (1976), lv den 399 Mich. 827 (1977), the plaintiff was struck by an automobile and brought suit against the driver and the bar owners who allegedly served the driver while he was visibly impaired. Four days prior to trial, the plaintiff entered into an "Indemnification Agreement", under which the driver's liability was limited to $19,000. The driver was expressly "retained" in the action. In analyzing the effect of this agreement on the plaintiff's action against the bar owners, the court looked to the purpose of the name and retain provision as stated by the Court of Appeals in Salas v Clements, 57 Mich. App. 367; 226 NW2d 101 (1975), rev'd 399 Mich. 103; 247 NW2d 889 (1976), to avoid collusion and perjury. Extrapolating from that purpose, the Court stated:

"Balanced against this purpose is the policy of the law to encourage settlement of litigation. The statute is not intended to prohibit settlement between a plaintiff and his intoxicated tortfeasor except where claims against tavern owners are first or simultaneously settled. On the contrary, the language of the statute indicates only an intention to require the intoxicated tortfeasor to be retained as a defendant. In enacting this ``special interest legislation' the Legislature has manifested an intention to prohibit a plaintiff from going to a jury with only a claim versus a tavern. The plaintiff must not only name but must retain the tortfeasor in the lawsuit. But that is all the statute requires. To prohibit a plaintiff from a ``hold harmless' type settlement, as was used here, would impede the prompt disposition of litigation and force trials of cases that would otherwise be settled." Buxton, supra, p 511.

Of course, Buxton was implicitly overruled by Putney. Until Putney was decided by this Court, however, Buxton remained as the uncontradicted interpretation of the name and retain provision. Nonetheless, it is argued that any reliance by plaintiffs on Buxton was misplaced because it was a decision of the Court of Appeals, rather than one of this Court. This argument fails to take into account the structure of the Michigan appellate system.

A decision of a panel of the Court of Appeals is a decision of the Court. GCR 1963, 800.4. A decision by any panel of the Court of Appeals is, therefore, controlling statewide until contradicted by another panel of the Court of Appeals or reversed or overruled by this Court. See In the Matter of Hague, 412 Mich. 532; 315 NW2d 524 (1982); Hackett v Ferndale City Clerk, 1 Mich. App. 6; 133 NW2d 221 (1965). While the possibility of reversal or contradiction may lessen a claim of reliance, it does not preclude it.^[2] And, although the possibility of the contradiction of Buxton by another panel of the Court of Appeals existed, that did not happen. Subsequent expressions by panels of the Court of Appeals on the same issue did nothing to lessen the authority of Buxton. See Cussans v Harris, 118 Mich. App. 567; 325 NW2d 793 (1982); Putney v Gibson, 94 Mich. App. 466; 289 NW2d 837 (1979), rev'd sub nom Putney v Haskins, supra; Rowan v Southland Corp, 90 Mich. App. 61; 282 NW2d 243 (1979); Denham v Bedford, 82 Mich. App. 107; 266 NW2d 682 (1978), aff'd 407 Mich. 517; 287 NW2d 168 (1980). In light of the unquestioned status of Buxton at the time Putney was decided by this Court,^[3] it would be unjust to apply Putney retroactively to persons other than those before the Court in that case.^[4]

In contrast to the harsh effect which the full retroactivity of Putney would have on injured plaintiffs, prospective application will have little effect on dramshop defendants in those pending cases where settlement agreements have been made, even though the defense of Putney will be unavailable. For them, the law will simply remain as it was from 1976 to 1982. We hold that Putney v Haskins is applicable to all cases where settlement agreements are entered into with the allegedly intoxicated person after the date of decision in Putney.

In granting leave to appeal to defendant Denter, we directed that the following question be briefed: "[W]hether the no-fault insurance act, MCL 500.3101 et seq.; MSA 24.13101 et seq., alters the common-law collateral-source rule". We answer that narrow question in the negative.

Section 3135(2) of the no-fault act provides:

"(2) Notwithstanding any other provision of law, tort liability arising from the ownership, maintenance, or use within this state of a motor vehicle with respect to which the security required by section 3101(3) and (4) was in effect is abolished except as to:

* * *

"(c) Damages for allowable expenses, work loss, and survivor's loss as defined in sections 3107 to 3110 in excess of daily, monthly, and 3-year limitations contained in those sections. The party liable for damages is entitled to an exemption reducing his or her liability by the amount of taxes that would have been payable on account of income the injured person would have received if he or she had not been injured." MCL 500.3135(2); MSA 24.13135(2).

In the trial court, Denter moved to exclude all testimony regarding plaintiff's economic losses on the grounds that the above provision barred tort damages to the extent plaintiff had received compensation under the no-fault insurance act. The trial court granted the motion, and on interlocutory appeal the Court of Appeals reversed, relying on Citizens Ins Co of America v Tuttle, 411 Mich. 536; 309 NW2d 174 (1981), rev'g 96 Mich. App. 763; 294 NW2d 224 (1980).

In Tuttle, the motorist plaintiff brought suit for damages sustained when he collided with the defendant's negligently kept cow. This Court held that because the cause of action against the defendant arose out of negligent cow-keeping, the plaintiff's action was not barred by § 3135(2) of the act. We stated:

"Only persons who own, maintain or use motor vehicles can be subject to tort liability for injuries or damage caused by the ownership, maintenance or use of a motor vehicle [under the no-fault act]. The non-motorist tortfeasor cannot be subject to tort liability for injuries or damage caused by the ownership, maintenance or use of a motor vehicle. The abolition of tort liability for injuries or damage caused by (arising from) the ownership, maintenance or use of a motor vehicle, therefore, does not abolish the tort liability of the non-motorist tortfeasor." 411 Mich. 545-546.

In the present case, Denter did not own, maintain, or use the motor vehicle that caused plaintiff's injuries. His liability, instead, arises out of his alleged sale of intoxicating liquors to a visibly intoxicated person. Just as the tort liability bar of § 3135 is inapplicable, so too is the remainder of the act inapplicable.

Notwithstanding that the language of the no-fault insurance act does not apply to the present case, Denter contends that the act manifests a legislative intent to prohibit recovery of damages in all tort actions where the plaintiff has been compensated by insurance. In short, Denter requests abolition of the collateral-source rule. We find no reason in the no-fault insurance act to comply with that request.

The common-law collateral-source rule provides that the recovery of damages from a tortfeasor is not reduced by the plaintiff's receipt of money in compensation for his injuries from other sources. Motts v Michigan Cab Co, 274 Mich. 437; 264 N.W. 855 (1936); Perrott v Shearer, 17 Mich. 48 (1868). In the context of insurance, the rationale for the rule is that the plaintiff has given up consideration and is entitled to the contractual benefits. The plaintiff's foresight and financial sacrifice should not inure to the benefit of the tortfeasor, who has contributed nothing to the plaintiff's insurance coverage. Similarly, gratuitous compensation should not inure to the benefit of the tortfeasor. The tortfeasor has contributed nothing, except the activity which caused the plaintiff's injuries.

The no-fault insurance act was a radical restructuring of the rights and liabilities of motorists. Through comprehensive action, the Legislature sought to accomplish the goal of providing an equitable and prompt method of redressing injuries in a way which made the mandatory insurance coverage affordable to all motorists. See Shavers v Attorney General, 402 Mich. 554; 267 NW2d 72 (1978), cert den 442 U.S. 934 (1979). It is in light of that goal that the setoff provisions of the no-fault insurance act must be viewed.

The obvious purpose of the setoff provisions of the act is to eliminate duplicate benefits. The act "reduces the amount that the insurance companies must pay out, making it possible for them to reduce the amount that they must charge, and it does so only in those situations where benefits are redundant". O'Donnell v State Farm Mutual Automobile Ins Co, 404 Mich. 524, 546; 273 NW2d 829 (1979). In effect, the Legislature made a trade-off. Those who were required to participate in the no-fault scheme gave up the possibility of redundant recoveries, but they were intended to receive the benefit of lower insurance rates.

Analysis of the setoff provisions in the no-fault act reveals a careful legislative effort to limit duplicative recovery only where the limitation would benefit a no-fault insurer, thus providing an incentive for lower insurance rates. Section 3116 of the act requires that personal injury protection no-fault benefits be reduced to the extent the insured has received equivalent compensation from tort judgments arising from accidents outside of the state, from accidents with uninsured motorists, and from intentionally caused harm. This reduction of benefits payable inures to the benefit of the injured party's no-fault insurer in situations where the tortfeasor would normally not be within the coverage of the no-fault act. Similarly, § 3109 of the act provides that personal injury protection no-fault benefits are to be reduced by the amount of governmental benefits provided or required by law to be provided to the injured insured. Again, when the additional compensation to the insured has come from a source outside the scope of the no-fault act, the no-fault insurer benefits by having to pay less compensation to the insured.

Finally, § 3135 of the act provides that tort judgments against persons arising from their ownership, maintenance, or use of a motor vehicle within the scope of the no-fault act are allowable only to the extent they exceed personal injury protection no-fault benefits received by the injured party under §§ 3107 to 3110 of the act. Although under § 3135 it is the liability insurer of the tortfeasor who benefits, it is still, normally, a no-fault insurer.

As relevant to the present action, the dramshop defendant Denter has not participated in the no-fault scheme. Denter asks us to turn around the policy of the no-fault act to benefit a tortfeasor rather than a no-fault insurer. This would not only be an unwarranted extension of the no-fault act, but it would be contrary to its intent. Abolition of the collateral-source rule as to him would do nothing to fulfill the legislative goal of lower no-fault insurance rates. The careful structure of the no-fault act's setoff provisions evinces no legislative intent to benefit dramshop defendants and their insurers.

In Burns, the summary judgment is reversed, and the cause is remanded for trial.

In Tebo, pursuant to GCR 1963, 865.1(7), the motion for summary judgment is denied, the decision of the Court of Appeals is affirmed, and the cause is remanded for trial.

CAVANAGH (participating as to Burns only) and BOYLE, JJ., concurred with BRICKLEY, J.

WILLIAMS, C.J.

I concur with my brother BRICKLEY except that I agree with my brother LEVIN as to the meaning of denial of application for leave to appeal by the Supreme Court.

RYAN, J. (concurring).

One of the functions of an appellate court is to authoritatively declare the meaning of the language of legislative enactments. One, but only one, of the purposes of that function is to enable practicing lawyers to advise their clients with a reasonable degree of certainty and predictability as to action which may be taken by those clients in their daily personal and economic affairs. In the course of construing and interpreting countless statutes, this state's appellate courts have repeatedly held that legislative language does not always mean what it seems plainly to say. Every judge and practicing lawyer of any experience knows that our courts have often told the bench and bar that the language of a legislative enactment means what the appellate courts say it means and not necessarily what the words used in the statute seem to mean.

Practicing law in such a jurisprudence involves very real risks. An appellate court may never be called upon to interpret a particular statute or, if asked, may receive the request only after years of "unauthoritative" interpretation and application by thousands of practicing lawyers who have been required to advise clients whether and how to proceed without the guidance of judicial interpretation of the statute in question. If the appellate court, when a request for construction of a statute is presented to it, construes the enactment in a way which is at odds with the reading theretofore given it by practicing lawyers, the latecoming judicial construction is nevertheless ordinarily accepted by all, including those who have "misperceived" the correct meaning of the statute, as the correct interpretation ab initio. The general rule is one of retroactivity. Such corrective judicial hindsight is understood and accepted by bench and bar as an ordinary risk indigenous to a profession in which service to a client rests largely upon the vicissitudes of analysis, scholarship, judgment, and prediction — the essence of the practice of law. Practicing lawyers live every day with the understanding that they may be reading a statute in a way with which an appellate court will subsequently disagree. They accept that risk.

However, when a Michigan appellate court does construe a statute and unequivocally declares what it means, the bench and bar are entitled, indeed obligated, to accept that construction as authoritative, and lawyers can be expected to render professional advice consistent with that interpretation in the confidence that the risk of misinterpretation of the statutory language has been lifted from their shoulders.^[1] While risk is ever present that a higher appellate court may disagree with an intermediate appellate court's interpretation of the meaning of a statute, a lawyer and his client should not be expected to assume that that will happen. When it does happen, it is the higher court's duty to determine whether the decisional error of the lower court, considered in the light of traditional doctrines of appellate review and notions of fundamental fairness, requires that the higher court decision be given retroactive, limited retroactive, or even only prospective effect. When, as in Buxton v Alexander, 69 Mich. App. 507; 245 NW2d 111 (1976), lv den 399 Mich. 827 (1977), the Supreme Court, upon being asked to review the Court of Appeals construction of the statute, declines the opportunity for corrective review, the bench and bar should be free to accept and act upon the intermediate court's interpretation, no matter how aberrant, in the confidence that the rules of the game will not be later rewritten retroactively.

While nice scholarly distinctions may be made and unimpeachably sound jurisprudential hair-splitting undertaken in defense and application of the canons that intermediate appellate court decisions are always subject to modification and reversal by higher authority, and that Supreme Court denial of leave to appeal is nonprecedential and does not amount to affirmation of the lower court decision, the fact is that practicing lawyers must give advice here and now to real clients about real claims of pain and suffering and related damages, involving real dollars. For them and their clients, the truism that nothing an intermediate appellate court ever says is set in jurisprudential concrete is no help.

In Buxton, supra, the intermediate appellate level of Michigan's one court of justice erroneously concluded that the name and retain provision of MCL 436.22; MSA 18.993 did not mean what it plainly and unequivocally said. Putney v Haskins, 414 Mich. 181; 324 NW2d 729 (1982). In so holding, it misadvised and misled bench and bar. While this Court's subsequent order denying the defendant's interlocutory application for leave to appeal should not, in any sense, have been taken as an approval of the Court of Appeals decision,^[2] the only common-sense inference a practicing lawyer could have drawn was that Buxton was surely the law. The logic of that inference was repeatedly fortified when at least four other panels of the Court of Appeals^[3] followed Buxton's patently erroneous holding, despite the fact that it was obviously at odds with the plain language of the statute. The result is that for six years, until this Court's decision in Putney, supra, the clear, direct, and unequivocal mandate of the Legislature that the actually intoxicated person was to be "retained in the action" was disregarded. There is simply no way of knowing how many litigants, acting upon the understandable advice of judicially guided counsel, followed the rule of Buxton and its progeny, or how many millions of dollars in damages have been paid in partial settlement of dramshop cases, followed by litigation of the remaining claims, in disregard of MCL 436.22; MSA 18.993 and subsequently decided Putney.

The question now is whether to give Putney the retroactive effect traditional rules of appellate review dictate, or only the prospective application plaintiffs in these cases contend fairness requires. If we give Putney merely prospective effect, we compound the mischief done in Buxton by effectively suspending for six years an otherwise valid statute, and retroactively penalizing the litigants and their counsel who followed the letter of MCL 436.22; MSA 18.993, declining the opportunity to settle claims against the actually intoxicated person separately, often going to trial with its inherently uncertain and sometimes unjust results.^[4]

If we give Putney retroactive effect, we vindicate the judgment of those attorneys and their clients who, despite judicial construction to the contrary, followed the letter of the statute, ignored the Court of Appeals decision in Buxton and its progeny and resisted the pressure to settle a portion of multiple dramshop claims separately. In following that course, however, we would unfairly penalize those who accepted the invitation of five panels of the Court of Appeals to ignore the plain language of the statute and settled a portion of a dramshop claim with the implied assurance that the balance of the claim could be safely litigated later. Included in that category are those countless lawyers and litigants who faced and succumbed to the "settlement juggernaut" to which my brother LEVIN refers, and which is indeed a part of the real-world practice of law in many circuits in this state. It is obvious, therefore, that the appellate judicial error of Buxton has created a dilemma from which there can be no extrication assuring absolute fairness to all.

Arguments abound for and against taking either course. The error which requires that we make a choice is judicially induced and on balance seems to call for whatever resolution minimizes the resultant injustice. It is enough that lawyers and their clients must run the risk and pay the price of erroneously anticipating judicial decisions interpreting the meaning of statutory language. They should not, at least in these cases, be penalized for following such decisions.

For those reasons, and those reasons only, I concur in the Court's judgment for limited prospectivity of MCL 436.22; MSA 18.993.

I likewise concur in the Court's disposition of the collateral-source rule issue.

LEVIN, J. (concurring in Burns, and partially dissenting in Tebo).

We agree that this Court's decision in Putney v Haskins, 414 Mich. 181; 324 NW2d 729 (1982), should be applied prospectively, and therefore concur in Burns and to that extent in Tebo. We would not suggest, however, that a decision of this Court disapproving or overruling an "unquestioned" or "uncontradicted interpretation"^[1] by the Court of Appeals should invariably be given prospective effect.

We also write separately because Robert and Dorothy Tebo should not be permitted to recover damages from the defendant taverns for wage loss and medical expense for which they have already been reimbursed, pursuant to the no-fault automobile liability act,^[2] through the payment of personal injury protection (PIP) benefits. We would not extend the concept of the common-law doctrine known as the collateral-source rule^[3] to include PIP benefits even where, as here, the PIP benefits are paid to the owner of an automobile by the no-fault insurer of that automobile.

The no-fault act substitutes one automobile insurer, who may be the no-fault/liability insurer of the injured person's vehicle or of some other vehicle,^[4] for another, the liability insurer of a negligently operated vehicle, as the source of reparations for wage/work loss within the statutory limit and medical expense. Before the enactment of the no-fault act, amounts paid by a liability insurer of a negligently operated vehicle for such elements of damage would, to avoid compensating an injured person in an amount greater than his loss, have been subtracted from damages payable by a non-motorist tortfeasor,^[5]e.g., an automobile or parts manufacturer. We would not ascribe to the Legislature, in changing the automobile insurer that reimburses for wage/work loss and medical expense, an intent to bring about, through extension of common-law doctrine (the collateral-source rule), double recovery of work loss and medical expense and a parallel enlargement of the liability of non-motorist tortfeasors.

Damages and reparations required to be paid by the common and statutory law of this state — damages in a tort action, PIP benefits pursuant to the no-fault act — are not generated from independent, separate, or collateral sources, but rather from the same source, namely the public policy of this state. Damages and reparations are payable now, as before the enactment of the no-fault act, out of a pool of money funded with premiums paid by owners and drivers of motor vehicles to discharge the liability imposed on them by the law of this state. It is not consistent with the policy of the common law, or with the purpose of the no-fault act to provide a more cost-effective reparations system, to read the no-fault act to permit duplicative recovery of benefits, at common law collectible but once, required to be paid by the law of this state.

The rationale of the collateral-source rule, judicially promulgated before the no-fault act was enacted, does not inexorably require extension of the rule to include PIP benefits; it should not be so extended absent evidence of a legislative purpose to provide, as part of the legislative program embodied in the no-fault act, for such double recovery and increase in the exposure of non-motorist tortfeasors.

I

This Court generally has decided the retroactivity-prospectivity question ad hoc. Sometimes the Court has stated its reasoning with regard to the conflicting considerations and, in other cases, it has stated little or no more than its conclusion.

The question has been resolved in particular cases by a process of reasoning and intuition. No rule of decision or principle for deciding the question has evolved. Perhaps the question must, as the lead opinion indicates, unavoidably be decided on the basis of the Court's somewhat generalized perception of what "justice" requires.^[6]

In the instant case, we are, with our colleagues, impressed that after this Court denied leave to appeal from the decision of the Court of Appeals in Buxton v Alexander, 69 Mich. App. 507; 245 NW2d 111 (1976),^[7] there was widespread reliance by bench and bar on that decision. And, as stated in the lead opinion, subsequent decisions of the Court of Appeals did nothing to lessen the authority of Buxton.^[8]

There is considerable institutional pressure to settle cases. Pretrial and settlement conferences and mediation exert pressure to settle. Settlement is the norm, trial the exception. In that climate, it would be difficult, although not impossible, for a lawyer to maintain, against the tide of professional and judicial opinion, that because Buxton might be overruled he could not safely advise his clients to settle a lawsuit on the authority of that decision.

In this context, therefore, we agree that the clients of lawyers who accommodated the institutional juggernaut and accepted Buxton — a decision that made it possible to settle one part of a three-cornered dispute without the consent of all the parties — as the governing rule, should not be penalized by a retroactive application of Putney and that it should be given prospective effect only.^[9]

We would not suggest, as does the lead opinion, that an "unquestioned" or "uncontradicted interpretation"^[10] by the Court of Appeals may invariably be relied on. That rationale is inconsistent with the concept that a denial of leave to appeal does not intimate anything regarding this Court's view on the merits of a decision of the Court of Appeals.^[11] Adoption of that rationale would encroach considerably on the Court's discretion whether to grant leave to appeal. In many cases that rationale would not achieve a just balancing of the rights of litigants advocating competing views of what the substantive rule should be.

This Court frequently declines to grant leave to appeal although a majority of the justices may be troubled by, or have doubts regarding, the reasoning or conclusion of the Court of Appeals. The Court may deny leave to appeal because it concludes that the decision of the Court of Appeals, although possibly erroneous, is not "clearly erroneous", or that it will not "cause material injustice".^[12] There is, by hypothesis, no conflict "with decisions of the Supreme Court or other Court of Appeals decisions". The Court, which must yearly consider some 2,500 applications for leave to appeal and requests for review (the latter from indigent persons convicted of an offense),^[13] may not on the first presentation of a question perceive that the "subject matter of the appeal involves legal principles of major significance to the jurisprudence of the State".^[14]

Some may urge that there is a need for a rule one way or the other, and that this Court should speak clearly on the question whether a decision of the Court of Appeals may be relied on by bench and bar. We do not think that it is possible for the Court to commit itself so absolutely. Nevertheless, and recognizing that ordinarily we should not go beyond what is necessary to decide the question before us, we believe it is appropriate to summarize our understanding of the pertinent principles:

(a) decisions of the Court of Appeals are binding precedent which must be followed by the trial courts unless and until there is a conflicting decision of the Court of Appeals^[15] or a disapproving or overruling decision by this Court;

(b) denial of leave to appeal by this Court does not constitute an expression of opinion regarding the analysis or conclusion of the Court of Appeals;^[16]

(c) a decision of the Court of Appeals (except possibly a decision en banc) is not a rule of law within the doctrine of stare decisis, since it binds neither other panels of the Court of Appeals^[17] nor the Supreme Court;^[18]

(d) a subsequent decision by this Court disapproving or overruling a Court of Appeals decision does not constitute a "declaration of a new rule" or the "overruling of established precedent"; and

(e) a decision of this Court is generally retroactive unless it overrules established precedent or otherwise declares a new rule, in which case the Court may limit the retroactivity of the new rule.^[19]

II

Dorothy Tebo was injured in a collision between an automobile she was driving and an automobile being driven by Edward S. Havlik. The no-fault automobile insurer of the vehicle Dorothy Tebo was driving,^[20] Detroit Automobile Inter-Insurance Exchange, paid PIP benefits in excess of $35,000 to Robert and Dorothy Tebo for lost wages and medical expense.^[21]

The Tebos commenced this action against Havlik and the owners of two taverns, the Long Branch Bar and the Oakley Liquor Bar. They alleged that Havlik was intoxicated at the time of the accident and that the taverns were subject to liability under the dramshop act.^[22] The Tebos settled with Havlik for $50,000.

Fred Denter, doing business as the Oakley Liquor Bar, filed a motion in limine to exclude evidence of lost wages and medical expense for which the Tebos had been reimbursed through the payment of PIP benefits by the no-fault insurer of the Tebo automobile. The motion was granted. The Court of Appeals reversed.^[23] It invoked the "collateral-source rule", which declares that money received from a source, e.g., automobile insurance, other than a tortfeasor does not generally "operate to lessen damages recoverable from the wrongdoer",^[24] and held that Dorothy Tebo could recover her wage loss and medical expense from the taverns although she had been reimbursed for those elements of damage by the no-fault insurer of her automobile.^[25] We would reverse the judgment of the Court of Appeals and reinstate the circuit court's decision granting the motion to exclude evidence of lost wages and medical expense for which the Tebos were fully reimbursed through the payment of PIP benefits.

III

The no-fault automobile liability act provides that an action cannot be maintained against an owner or driver of a motor vehicle for noneconomic loss unless the injuries meet the statutory threshold of "death, serious impairment of body function or serious disfigurement".^[26] An action can be maintained against an owner or driver for economic loss in excess of the PIP benefits prescribed by the no-fault act (excess economic loss) without regard to the statutory threshold.^[27] The no-fault act does not limit the tort liability of non-motorist tortfeasors (e.g., an automobile or parts manufacturer or tavern) who may have caused or contributed to an automobile accident.^[28]

Thus, if Mrs. Tebo's physical injury was of a severity that met the statutory threshold, the Tebos could properly maintain this action against Havlik, the driver of the other vehicle, for noneconomic loss and, without regard to the statutory threshold, for excess economic loss. Also without regard to the statutory threshold, they may maintain this action for all^[29] their economic and noneconomic damages against the taverns that allegedly sold Havlik liquor while he was visibly intoxicated.

Damages are awarded in a personal injury action to compensate a person for his injury and loss.^[30] Where more than one tortfeasor contributes to a person's injury and loss, he may recover all his damages from any tortfeasor who contributed to his loss^[31] but he may not recover from them a sum aggregating more than the amount required to compensate him in full for his loss.^[32] Accordingly, an amount recovered from one tortfeasor as the result of a settlement or upon a judgment is required, by a rule of common law, to be subtracted from any verdict or judgment rendered against another tortfeasor.^[33] As between themselves, the tortfeasors may have rights of contribution or indemnification.^[34]

The common-law rules subjecting a tortfeasor to liability for all the damages suffered by an injured person, but limiting the injured person to a single recovery however many tortfeasors contributed to the loss, establish that the question whether one can seek to recover for his loss from two or more sources and the question whether a full recovery can be obtained from each of several sources are separate questions. Although the Tebos may, consistent with the no-fault act, maintain an action against the taverns under the dramshop act for all^[35] their damages, economic and noneconomic, it does not follow that a verdict for wage loss and medical expense damages against a tavern should not be reduced by whatever is recovered for work loss and medical expense through the payment of PIP benefits, as such a verdict clearly would be reduced if the work loss and medical expense had been recovered as damages from another tortfeasor.^[36] The question whether the amount of the PIP benefits for work loss and medical expense received by the Tebos should be subtracted from any verdict against the taverns for those elements of damage for which the Tebos have already been reimbursed through the payment of PIP benefits is a separate question from whether the no-fault act limits the tort liability of non-motorist tortfeasors, the question already decided by this Court adversely to non-motorist tortfeasors.

The common-law rule requiring that the amount of an earlier recovery from one tortfeasor be subtracted from a subsequent verdict against another tortfeasor continues, after the enactment of the no-fault act, to apply to recoveries from the owner or driver of an automobile for noneconomic and excess economic loss. Thus, if the defendants remaining in the instant action were not taverns but were owners or drivers of other automobiles involved in the collision, the $50,000 settlement obtained from Havlik for noneconomic or excess economic loss would be subtracted from any verdict against them. The rule also continues to apply where the tortfeasor is a non-motorist such as an automobile or parts manufacturer or the owner of a cow^[37] or (see Part VI) a tavern — the $50,000 settlement would be subtracted from any recovery from such a defendant for economic or noneconomic loss. Thus, however this case is decided, amounts recovered from an owner or driver for noneconomic or excess economic loss, or from a non-motorist tortfeasor for economic or noneconomic loss, will continue to be subtracted from a verdict for like elements of damage against another owner or driver or a non-motorist tortfeasor.

The question here is whether an injured person, who may recover for noneconomic or excess economic loss but once, however many owners, drivers, manufacturers, cows, or taverns may have contributed to his loss, may recover a second time from a non-motorist tortfeasor for economic loss fully recompensed by PIP benefits paid pursuant to the no-fault act.

IV

This is a personal injury action arising out of an automobile accident. In 1972, the Legislature enacted the no-fault automobile liability act^[38] and in separate legislation (see Part VI) amended the dramshop act.^[39] Both enactments were legislative responses to perceived deficiencies in the tort system, common law and statutory law.

Insurance proceeds have been the principal source of recovery for injury suffered in an automobile accident, whether of a severity above or below the statutory threshold, both before and after the enactment of the no-fault act. Before the enactment of the no-fault act, a person injured in an automobile accident who recovered from a tortfeasor for wage loss or medical expense ordinarily received payment from the liability insurer of the tortfeasor. If more than one tortfeasor was liable for the loss, whatever was recovered from the insurer of one tortfeasor (or the tortfeasor himself) was, under the common-law rule requiring the setoff of a recovery from another tortfeasor, subtracted from any verdict or judgment against a tortfeasor who also contributed to the loss.

The no-fault act changes the automobile insurer from which reparations may be recovered for wage/work loss within the statutory limit and medical expense. An automobile insurer, often but not necessarily the no-fault insurer of an automobile owned by the injured person^[40] rather than the automobile liability insurer of the owner or driver of another vehicle found to have been at fault, pays for lost wages or other economic loss within the statutory limits and medical expense.^[41]

Under another common-law rule, recoveries from a "collateral source" are not subtracted from a verdict or judgment awarding damages in tort.^[42] Damages, whether paid by the tortfeasor or the tortfeasor's insurer, are not deemed to have been received from a collateral source and are so subtracted.^[43] The proceeds of insurance payable to the holder or beneficiary of a policy of insurance, as distinguished from insurance proceeds paid in discharge of the liability of a tortfeasor, are generally regarded as received from a collateral source and, thus, are not subtracted from a damage award against a tortfeasor who contributed to the injured person's loss. An injured person may, therefore, retain the proceeds of insurance paid to him as a policyholder or beneficiary and recover a second time from a tortfeasor.^[44]

The Tebos contend that because wage/work loss within the statutory limit and medical expense are no longer payable by an owner or driver of a motor vehicle or his liability insurer to the injured person as damages and the person now responsible for the payment of those losses is often a no-fault insurer of a vehicle owned by the injured person, amounts paid for work loss and medical expense by a no-fault insurer as PIP benefits are received from a collateral source and may not be subtracted from a verdict for those elements of damage against a non-motorist tortfeasor.

A

A number of rationales have been advanced in justification of the collateral-source rule. The rationales have been questioned by commentators, who are generally critical of the collateral-source rule and of its heedless expansion.^[45]

The lead opinion in the instant case states:

"In the context of insurance, the rationale for the [collateral-source] rule is that the plaintiff has given up consideration and is entitled to the contractual benefits. The plaintiff's foresight and financial sacrifice should not inure to the benefit of the tortfeasor, who has contributed nothing to the plaintiff's insurance coverage. Similarly, gratuitous compensation should not inure to the benefit of the tortfeasor. The tortfeasor has contributed nothing, except the activity which caused the plaintiff's injuries."^[46]

The arguments that the Tebos exercised "foresight", made a "financial sacrifice", and gave "up consideration and are entitled to the contractual benefits", ignore that many persons who recover PIP benefits have not paid no-fault insurance premiums because they do not own an automobile and that the financial sacrifice of those who have paid no-fault insurance premiums is governmentally mandated, an exaction like a tax or assessment on owning a motor vehicle. Payment of premiums to a no-fault insurer is not made in the exercise of foresight, but in compliance with the obligation, under the no-fault act, to furnish security for the payment of PIP benefits. A different question would be presented if PIP benefits were purchased as are, say, disability or major medical expense insurance benefits — in the exercise of the personal initiative of the recipient, unconstrained by the mandate of statute — to provide the recipient with benefits not provided by law or to supplement those that are provided by law.

To be sure, an owner of an automobile who receives no-fault benefits from the no-fault insurer of that automobile may be said to have "given up consideration", and no-fault benefits may be said to be "contractual benefits", but such characterizations, again, ignore that the "consideration" is a governmentally mandated exaction and that the scope and amount of no-fault benefits are specified by statute.^[47]

The arguments that the tortfeasor should not benefit from the payment of PIP benefits because he "contributed nothing, except the activity which caused the plaintiff's injury" and "contributed nothing to the plaintiff's insurance coverage" ignore that a tortfeasor who obtains the benefit of a subtraction from a verdict against him of the amount of a recovery from another tortfeasor also could be said to have contributed nothing to the recovery from the other tortfeasor or to the cost of the tortfeasor's insurance other than his activity which, in combination with the activity of the other tortfeasor, gave rise to the plaintiff's injury.^[48]

The payment of PIP benefits by a no-fault insurer is not "gratuitous compensation"; the cost of providing such benefits is paid by a multitude of owners and drivers of motor vehicles in compliance with their obligation, under the no-fault act, to furnish security for the payment of such benefits.

B

The Court of Appeals said that although the no-fault act requires owners of motor vehicles to carry no-fault insurance^[49] and the Tebos thus did not voluntarily contract for no-fault insurance, "[n]evertheless, premiums were paid for the coverage and the insurance benefits were still obtained from a source wholly independent of and collateral to the defendants. We will not sidestep the collateral source rule on the ground that the insurance was mandatory."^[50] (Emphasis supplied.)

The decision of the Court of Appeals allowing a double recovery of wage loss and medical expense rests on (i) the generalization that insurance proceeds are a collateral source, (ii) the change brought about by the no-fault act in the insurer who is responsible to pay for wage/work loss within the statutory limit and medical expense from the automobile liability insurer of an owner or driver at fault to an automobile no-fault insurer, and (iii) the circumstance that in the instant case the no-fault insurer required to reimburse Dorothy Tebo is the insurer of her own automobile.

The collateral-source doctrine developed in respect to fire and life^[51] insurance, and was extended to other forms of insurance and other benefits and payments.^[52] Those insurance programs, whether

the premiums were paid by the insured, the decedent, a member of his family, or his employer,^[53] were not a statutorily mandated substitute form of reparations^[54] for loss which, if compensated at all, was theretofore compensated mainly through liability insurance. Parenthetically, a property owner no longer may retain the proceeds of fire insurance and seek a duplicative recovery against a tortfeasor assertedly responsible for the fire loss; the standard fire insurance policy subrogates the insurer to any right of action of the insured.^[55] This rule seems to apply generally to property insurance.^[56]

When the collateral-source doctrine developed, life, disability, health, accident, and medical insurance all were seen by the courts as ancillary or alternative sources to the then primary source of recovery in an action against a tortfeasor and collection from the tortfeasor or his liability insurer. Since the enactment of the no-fault act, however, the primary source of recovery of wage/work loss within the statutory limit and medical expense resulting from an automobile accident is PIP benefits. Because an owner or driver is only subject to tort liability for excess economic loss and because most automobile accidents do not involve non-motorist tortfeasors, an injured person will not ordinarily have a tort source of recovery for wage/work loss and medical expense now paid as PIP benefits under the no-fault act.

Insurance policies, including medical expense insurance,^[57] purchased voluntarily and not under the constraint of governmental mandate,^[58] remain collateral sources of recovery to the primary PIP benefits source of recovery. An action against a non-motorist tortfeasor (an automobile or parts manufacturer or the owner of a cow or a tavern) has now become the collateral source. To speak of the primary source of recovery as an independent, separate, collateral source of recovery is to ignore that PIP benefits payable by a no-fault insurer ordinarily are the only source of recovery for wage/work loss within the statutory limit and medical expense arising out of an automobile accident.

C

The effect of the "insurance proceeds-collateral source" rationale is to ascribe to the Legislature an intent to create two classes of PIP benefit recipients:

1) those who recover PIP benefits from a no-fault insurer to whom the recipient paid no-fault insurance premiums, who, on the rationale that PIP benefits are the proceeds of insurance and thus deemed received from a collateral source, can recover for the same loss a second time from a non-motorist tortfeasor, and

2) those who do not own an automobile and who recover PIP benefits from, say, a self-insured owner of a motor vehicle, who, because the PIP benefits are not the proceeds of insurance and thus not deemed received from a collateral source, cannot recover for the same loss a second time from a non-motorist tortfeasor.

Absent a basis for ascribing to the Legislature an intent to create two classes of PIP benefits recipients, the no-fault act should be construed to allow all recipients of PIP benefits who have a cause of action against a non-motorist tortfeasor to collect a second time or to disallow any PIP benefit recipient a duplicative recovery. The "insurance proceeds-collateral source" rationale does not justify today's decision.

V

The argument that PIP benefits, at least where paid by the no-fault insurer of a vehicle owned by the injured person, are received from a collateral source rests in large part on the change brought about by the no-fault act in the automobile insurer responsible to pay reparations for wage loss and medical expense. The question presented thus primarily depends on legislative intent and only secondarily on the scope of the common-law collateral-source doctrine.

The true question is not whether PIP benefits are or are not the proceeds of insurance within the meaning of the collateral-source rule but whether the Legislature intended to provide for or bring about double recovery by PIP benefit recipients who have a cause of action against a non-motorist tortfeasor.

The Legislature did not specifically provide whether PIP benefits paid for work loss or medical expense should be deemed received from a collateral source and on that basis not subtracted from a tort recovery reimbursing for the same wage loss or medical expense, with the result that a person injured in an automobile accident can recover both PIP benefits from a no-fault insurer and damages from a non-motorist tortfeasor to compensate for the same loss.^[59] We must, accordingly, determine

from the history and structure of the no-fault act the construction of that act on the issue presented that is most consonant with its purpose.

A

The lead opinion rests today's decision on the policy of the no-fault act as well as on the collateral-source rule. The opinion refers to a section of the no-fault act providing for a setoff from PIP benefits of benefits provided or required to be provided by federal or state law (e.g., social security survivors' benefits and workers' compensation benefits)^[60] and to a section authorizing a no-fault insurer to recoup PIP benefits from a tort recovery for wage loss or medical expense obtained from a motorist who is not relieved of tort liability,^[61] and observes that such setoff or recoupment reduces the cost of providing PIP benefits.^[62] The opinion concludes that the policy of the no-fault act is "to limit duplicative recovery only where the limitation would benefit a no-fault insurer, thus providing an incentive for lower insurance rates".^[63] The opinion continues that the taverns had not "participated in the no-fault scheme", states that the act "evinces no legislative intent to benefit dramshop defendants and their insurers", and concludes that it would be contrary to the intent of the no-fault act "to benefit a tortfeasor rather than a no-fault insurer" because to do so would "do nothing to fulfill the legislative goal of lower no-fault insurance rates".^[64]

First, the analysis of the lead opinion assumes that the cost of automobile insurance would not be reduced if duplicative recovery of wage/work loss and medical expense was eliminated. The cost would, however, be reduced if a no-fault insurer can obtain reimbursement in respect to PIP benefits paid by it from a non-motorist tortfeasor such as a tavern. Although the reimbursement question is not presented and therefore should not be decided, explicitly or by implication, a substantial argument could be made that a no-fault insurer should be able to claim reimbursement from a non-motorist tortfeasor.^[65]

Second, the analysis of the lead opinion reflects misunderstandings of the operative effect of the no-fault act and of the consequence of today's decision. The statement in the opinion that "[i]n effect, the Legislature made a trade-off. Those who were required to participate in the no-fault scheme gave up the possibility of redundant recoveries, but they were intended to receive the benefit of lower insurance rates",^[66] is based on an incorrect predicate. "Those [owners and drivers] required to participate in the no-fault scheme" did not have the possibility of redundant recoveries to "give up". Before the enactment of the no-fault act, an owner or driver injured in an automobile accident could collect but once,^[67] and then only if he was not at fault and the driver of the other vehicle or other tortfeasor was both at fault and collectible.

Third, a tortfeasor "participates", to the extent participation is required, by being subject to liability to the injured person for the same loss. While the persons under such a common burden of liability generally, at least today, have rights of contribution between themselves (see Part VI, fns 82, 83), the policy of avoiding duplicative recoveries has not depended on evidence that each person subject to liability in fact participated in contributing to reimbursement of the loss. It is not a precondition to the common-law rule precluding duplicative recovery that the judgment shall have been rendered against a tortfeasor who has participated in an insurance program for the benefit of injured persons. An uninsured motorist would be entitled to the benefit of the common-law rule if there had been a settlement with another tortfeasor who, together with the uninsured motorist, is subject to liability to the injured person for the same loss.

Fourth, a decision that PIP benefits are not received from a collateral source would not confer a benefit on taverns. Such a decision would rather maintain the status quo, namely the rule that precludes duplicative recovery by requiring that reparations obtained from another source be subtracted from a verdict against a non-motorist tortfeasor.

Fifth, it does not follow from the truism that reducing the aggregate amount paid or payable by a no-fault insurer for PIP benefits by amounts received by the PIP benefit recipient or recouped by the no-fault insurer from certain other sources reduces the cost of providing PIP benefits and thus should result in lower insurance rates, that the no-fault act reflects a policy of barring duplicative recovery only where a no-fault policyholder may benefit through lower rates. Stated conversely, it does not follow that the no-fault act reflects a policy to allow duplicative recovery unless the avoidance of double recovery reduces the cost of no-fault insurance.

The provisions of the no-fault act relied on in the lead opinion (requiring a setoff of benefits provided or required to be provided under state or federal laws and allowing for recoupment of duplicative benefits obtained from a motorist not relieved of tort liability) reflect a legislative policy parallel to the policy of the common law of allowing an injured person to be reimbursed for his loss but once. The necessary consequence of these cost-saving common-law and statutory law policies is to confer a benefit on whoever is required by the common and statutory law of this state to provide or contribute to the cost of providing reparations to injured persons. The inference drawn by the majority from this consequence, that the no-fault act reflects a separate legislative policy to benefit no-fault insurers and their insureds to the exclusion of non-motorist tortfeasors and their insureds, is a non sequitur.

The no-fault act does not, as this Court observed in Citizens Ins Co of America v Tuttle, 411 Mich. 536, 547; 309 NW2d 174 (1981), address the tort liability of non-motorists. Thus the no-fault act does not reflect any separate legislative policy respecting non-motorists. That the no-fault act reflects a policy to bar redundancy in PIP benefits, and thus to reduce the cost to motorists of providing such benefits, does not imply a separate legislative policy to provide for redundancy with the attendant increase in cost where the reparations are required to be provided by non-motorists.

Allowing double recovery may not, as the lead opinion notes, increase automobile insurance rates, but it will increase insurance rates to automobile and parts manufacturers, taverns, owners of cows, and other non-motorists. Since any increase in the exposure of non-motorist tortfeasors will result in increased insurance and other costs for those subject to such increased exposure, which in turn must be passed on to the consumer, the increased cost of double recovery will unavoidably be reflected in the cost of automobiles, parts, alcoholic beverages, and milk, and thus will ultimately be paid by the same persons who pay for automobile insurance. In the last analysis, the consumer pays for all the costs of the tort and no-fault systems and thus for all redundancy in those systems.

This Court should not ascribe to the Legislature the narrow purpose of saving the consumer the cost of redundancy only when that cost is passed on to him in the form of automobile insurance premiums, and of intending to bring about both redundancy and the attendant passing on of the cost of that redundancy to consumers in the form of increased cost of automobiles, parts, alcoholic beverages, and milk.

B

Although reparations for wage/work loss and medical expense now are generally paid as PIP benefits by an automobile insurer who may, as here, be the automobile insurer of a vehicle owned by the injured person, rather than as "damages" by the automobile insurer of the offending vehicle, the source of the funding remains the same under the no-fault act as before its enactment — a pool of money funded with premiums paid by owners and drivers of motor vehicles to discharge the liability imposed on them by the law of this state.

Judgments against individuals and business concerns — before and after the enactment of the no-fault act — generally are not collectible apart from insurance. But for the pool of money funded with insurance premiums, few judgments could be collected, and the flood of litigation that gave rise to the no-fault act would not have occurred. Before the no-fault act — and now after its enactment — owners and drivers paid premiums to insurance companies and thereby provided a pool of money from which judgments and settlements (and, now, PIP benefits) are paid.

Before the no-fault act, a person injured in an automobile accident who recovered from a tortfeasor for wage loss or medical expense ordinarily received payment from the public liability automobile insurer of the tortfeasor. Although the source of the recovery was an insurer rather than the tortfeasor, the law did not regard the payment of insurance proceeds by a tortfeasor's liability insurer as a collateral source. Whatever was recovered from the automobile insurer of one tortfeasor was, under the general rule of the common law previously discussed, subtracted from any verdict or judgment against other tortfeasors.^[68]

C

It may be argued that in making payment to the injured person, an automobile liability insurer was substituting for the tortfeasor (and payment by a tortfeasor or his liability insurer of damages is not money received from a collateral source), but that here the recipient of PIP benefits who recovers from the no-fault insurer who insured the recipient's automobile paid for the PIP benefits in much the same way as one pays for benefits payable pursuant to life or medical expense insurance, the proceeds of which are deemed received from a collateral source.^[69]

The effort to analogize PIP benefits mandated by law to the proceeds of insurance policies not mandated by governmental policy overlooks the fundamental difference between a payment mandated by governmental policy and one that the insurer and policyholder contract for apart from governmental mandate.

No-fault insurance and PIP benefits, on the one hand, and life and medical expense insurance and benefits, on the other, are distinguishable in a number of particulars. Life and medical expense insurance generally cover loss arising from multiple perils, not a single peril such as an automobile accident. PIP benefits are payable only for loss arising in respect to automobiles. Life and medical expense insurance generally will not fully compensate for the financial loss. Medical expense PIP benefits compensate for the full amount of the loss, and work-loss PIP benefits will compensate most persons^[70] for the full amount of their loss, and an action for excess economic loss may be maintained against an owner or driver at fault.^[71]

Although the injured person need not now prove fault, and payment is assured, the elements of loss for which PIP benefits are paid are the same as those that were compensable in a tort action against another owner or driver before no-fault. The substitute PIP benefits "damages" or reparations are sometimes, as here, payable pursuant to a statutory "contract" of insurance, but they are payable independently of contract. It is not open to the insurer and policyholder to agree on benefit terms other than those prescribed by statute. PIP benefits, as distinguished from benefits payable under a life or medical expense policy, are required to be paid, under the priority system of the no-fault act, without regard to whether the injured person has an insured-insurer relationship with any insurer.^[72]

The money paid by a recipient of PIP benefits to the no-fault insurer of his automobile, is, in contrast with premiums paid to a life or medical expense insurer, a governmentally mandated exaction to finance a social welfare entitlement program, administered through insurance companies and self-insured enterprises, requiring the payment of reparations for the benefit of all persons injured in automobile accidents without regard to whether they own automobiles^[73] or contribute to the cost of financing or funding the pool of money from which PIP benefits are paid.

PIP benefits are closely akin to, and function as, statutorily mandated automobile accident damages or reparations for wage loss and medical expense. Damages — reparations required to be paid by law — are, all would agree, not a collateral source. PIP benefits are reparations required to be paid by law. They are payable by an automobile insurer — now as before enactment of the no-fault act — out of the pool of money funded with premiums paid by owners and drivers of motor vehicles to discharge the liability imposed and arising by operation of the law of this state, formerly the law of torts, now the no-fault act.

While the no-fault insurer is not a tortfeasor, it substitutes for the liability insurer of the loss formerly payable by a motorist tortfeasor. Although a no-fault insurer pays without regard to fault, so too may an automobile or parts manufacturer (or its insurer), which is subject to liability for a defective product however carefully it may have sought to produce the product.

The source of the liability of a no-fault insurer (and of a self-insured owner of a vehicle), a tavern, an automobile or parts manufacturer, and an owner of a cow, are fundamentally the same: governmental policy, either a statute (the no-fault or dramshop acts) or a rule of law (the tort law of negligence and of product liability) of this state. Payments in discharge of a liability to pay PIP benefits and in discharge of a liability to pay a tort judgment are not generated from independent, separate, collateral sources, but from the same source: the law of this state, which causes persons to provide the funds (generally by the purchase of insurance) to pay the reparations required by the law.

D

There is nothing in the language, history, or purpose of the no-fault act suggesting that the Legislature intended, through the extension of the common-law collateral-source rule, that a person who is injured in an automobile accident and recovers PIP benefits from a no-fault insurer to whom insurance premiums were paid should be permitted to recover a second time for the same wage/work loss and medical expense from a tortfeasor.

The effect of extending the collateral-source rule to include the proceeds of a no-fault policy paid in the form of PIP benefits is to ascribe to the Legislature a purpose, by enacting the no-fault act, not only to substitute the no-fault insurer for the other driver and his insurer as the source of payment for wage/work loss and medical expense, but also to reward some (or, depending on how one reads the lead opinion, all) injured persons who have a cause of action against a non-motorist tortfeasor with double recovery and to deprive non-motorist tortfeasors of the reduction of the amount of their liability that had theretofore resulted from the subtraction of reparations obtained from others the law subjects to liability for the same loss.

Treating no-fault benefits as received from a collateral source is not consistent with the purpose and structure of the no-fault act. The policy of the common-law rule requiring the subtraction of duplicative benefits is also reflected, as noted in the lead opinion, in provisions of the no-fault act.^[74] Fundamentally the same common-law and statutory policy — avoidance of duplicative recovery — that requires subtraction of the $50,000 paid by Havlik from any recovery from a motorist or non-motorist tortfeasor for noneconomic and excess economic loss, also requires subtraction of the $35,000 paid as PIP benefits to the Tebos from any recovery, duplicative of the statutory PIP benefits, from a non-motorist tortfeasor.

In enacting the no-fault act, the Legislature sought to provide a more efficient and cost-effective automobile accident reparations system. It is not consistent with that purpose to hold in effect that, by changing the source of the recovery from the owner or driver of the other automobile or his liability insurer to an automobile no-fault insurer (who sometimes is the insurer of the vehicle owned by the recipient of the PIP benefits), the Legislature also intended to enable injured persons to recover for the same injury a second time from a non-motorist tortfeasor, at least in those cases where the no-fault insurer of the vehicle owned by the recipient rather than, say, a self-insured owner pays the PIP benefits. Such a construction

— increases the exposure of non-motorist tortfeasors and their insurers, who before the enactment of the no-fault system could offset against a judgment all payments by a motorist tortfeasor or his automobile liability insurer;

— provides double recovery for wage loss and medical expense theretofore recoverable but once; and

— is inconsistent with the apparent purpose of the Legislature to replace certain aspects of the tort recovery system with the payment of PIP benefits under the no-fault act.

We conclude that the PIP benefits paid to the Tebos under the no-fault act should be subtracted from a tort recovery for wage loss and medical expense from a non-motorist tortfeasor.

VI

The rule requiring the subtraction from a verdict of all amounts obtained from other sources that are required to provide reparations for the same loss by the law of this state, applicable where the verdict is against an automobile or parts manufacturer or the owner of a cow, applies as well where the verdict is against a tavern. There is no reason to suppose that the Legislature, which in 1972 enacted both the no-fault act and amendments to the dramshop act, intended that the enactment of the no-fault act would increase the exposure of taverns. The 1972 amendments to the dramshop act indicate just the opposite. In adding "visibly" before "intoxicated", eliminating exemplary damages, and adding the "name and retain" requirement,^[75] the Legislature clearly sought to reduce the cost to taverns of providing reparations pursuant to the dramshop act.

It has been argued that subtracting amounts received from other sources subject to liability for the same loss may result in a tavern paying nothing if the amount received from other sources is greater than a verdict the jury may award. That is, however, true whenever there is a subtraction.^[76]

Where there is more than one tavern, the plaintiff may settle with one tavern and proceed to trial with the others. The rule nevertheless is that whatever is recovered from one tavern in settlement is to be subtracted from any verdict against other taverns.^[77] This too could result in no actual recovery against taverns that do not settle.

After Putney v Haskins, the intoxicated person must be retained in the action and there cannot be a settlement unless the tavern agrees. If there had been a verdict against the taverns for $50,000 or another amount and against Havlik for the $50,000 paid in settlement of the claim against him, there would have been but one recovery. The $50,000 would have been subtracted by the judge^[78] from any verdict against the taverns^[79] unless a different rule should apply where the non-motorist tortfeasor is a tavern than where it is an automobile or parts manufacturer or the owner of a cow.

It is urged that a purpose of the tort law and of the dramshop act is to encourage potential tortfeasors to act with due care and that eliminating duplicative recovery does not serve that purpose. First, there are significant areas of tort liability where a tortfeasor is subject to liability even if he has exercised due care. Second, most tort liability is funded by insurance; the impact of a tort recovery on a particular tortfeasor and the likely effect on his future behavior are often marginal. Third, subtracting PIP benefits from a tort recovery would not result in the tortfeasor or his insurer escaping all liability; they would remain liable or potentially liable for noneconomic loss (frequently the greater exposure) and excess economic loss. That potential liability should encourage those who would not otherwise exercise due care, who can be so constrained, to do so. Fourth, the argument ignores that the common-law rule requires the subtraction of a recovery from another tortfeasor subject to liability for the same loss; if this argument for not requiring subtraction of a duplicative recovery is thought to be sound, then the general rule requiring subtraction of recoveries of reparations from other tortfeasors should be abolished and an injured person should be permitted to recover his damages in full as many times as there are collectible tortfeasors who contributed to his loss.

Although the instant case concerns setoff and not contribution, arguments have been made^[80] based on an extension of a concept expressed in Virgilio v Hartfield, 4 Mich. App. 582; 145 NW2d 367 (1966), where the Court of Appeals held that taverns sued under the dramshop act could not maintain an action for contribution against the intoxicated driver.^[81]Virgilio was decided before Moyses v Spartan Asphalt Paving Co, 383 Mich. 314; 174 NW2d 797 (1970),^[82] and before the adoption of the revised Uniform Contribution Among Tortfeasors Act.^[83] It is now clear that contribution can be enforced although tortfeasors are separate and concurrent rather than "joint" wrongdoers, although there is not a common theory of liability, and although the liability of one tortfeasor may be based on fault while the liability of the other is based on a statute or rule of the common law allowing recovery without regard to fault. All that is now necessary to enforce contribution is that there be a common burden of liability.^[84]

A no-fault insurer, albeit not a tortfeasor, is under a common burden of liability with a tavern for the payment of wage/work loss within the statutory limit and medical expense to a person injured in an automobile accident. Although an action against a tavern is statutory and liability does not depend on proof of wrongdoing, the basis of liability does not differ substantially from the basis of liability of an automobile or parts manufacturer for product liability, which also may arise without proof of wrongdoing.

It also has been asserted that the dramshop act is penal and that a different rule should therefore apply. Although the dramshop act contains penal as well as remedial provisions, punitive damages have not been recoverable under the dramshop act.^[85] The Legislature amended the statute in 1972 to eliminate the recovery of exemplary damages.^[86] It is now clear that only compensatory damages are recoverable.^[87]

We conclude that the rule requiring the subtraction from verdicts against tortfeasors of amounts received from other sources that are required to provide reparations for the same loss by the law of this state should apply to a verdict against a tavern as well as a verdict against an automobile or parts manufacturer or the owner of a cow.

VII

The Oakley Liquor Bar moved to exclude evidence of wage loss and medical expense. Arguably, such evidence should be admitted, and the amount paid as PIP benefits then subtracted by the judge^[88] from any judgment against the tavern. The elements of damage for which PIP benefits are payable are not, however, recoverable in an action against the allegedly intoxicated driver,^[89] who must be named and retained under the amended dramshop act as construed in Putney v Haskins. It seems, therefore, that it would simplify the trial to exclude the evidence with such advice to the jury in that regard, if any, as may be thought appropriate.

We concur in Burns, and in Tebo would remand for further proceedings consistent with this opinion.

KAVANAGH, J., concurred with LEVIN, J.

NOTES

[1] For a review of this Court's treatment of retroactivity see Moody, Retroactive Application of Law-Changing Decisions in Michigan, 28 Wayne L Rev 439 (1982).

[2] That is not to say that a denial of leave to appeal fixes the state of the law or strengthens the case for the prospectivity of a later overruling case. A denial of leave to appeal has no precedential value. See Frishett v State Farm Mutual Automobile Ins Co, 378 Mich. 733 (1966). We, therefore, ascribe no weight to our denial of leave to appeal in Buxton. What must be given some weight, however, is that an uncontradicted decision of the Court of Appeals must be followed by the circuit courts.

[3] As our brother Justice LEVIN notes, the institutional pressure towards settlement is great. In light of the oft-stated policy of this Court favoring settlement, see, e.g., Brewer v Payless Stations, Inc, 412 Mich. 673, 679; 316 NW2d 702 (1982); Ogden v George F Alger Co, 353 Mich. 402, 407; 91 NW2d 288 (1958), a party's reliance on Buxton can hardly be deemed unreasonable.

[4] We find United States v Estate of Donnelly, 397 U.S. 286; 90 S. Ct. 1033; 25 L. Ed. 2d 312 (1970), cited in support of full retroactivity, to be distinguishable. There, in 1960, the respondents purchased property subject to federal tax liens in reliance on the Court of Appeals decision in Youngblood v United States, 141 F2d 912 (CA 6, 1944), where the 6th Circuit had held certain federal tax liens invalid for failing to comply with Michigan filing requirements. In United States v Rasmuson, 253 F2d 944 (CA 8, 1958), however, the 8th Circuit held that such federal tax liens need not comply with local law. The Supreme Court ultimately agreed with the position taken by the 8th Circuit. United States v Union Central Life Ins Co, 368 U.S. 291; 82 S. Ct. 349; 7 L. Ed. 2d 294 (1961). In holding that Union Central was to be retroactively applied to the respondents in Donnelly, the Supreme Court stated:

"Deviant rulings by circuit courts of appeals, particularly in apparent dictum, cannot generally provide the ``justified reliance' necessary to warrant withholding retroactive application of a decision construing a statute as Congress intended it. In rare cases, decisions construing federal statutes might be denied full retroactive effect, as for instance where this Court overrules its own construction of a statute, cf. Simpson v Union Oil Co, 377 U.S. 13, 25; 84 S. Ct. 1051; 12 L. Ed. 2d 98 (1964), but this is not such a case." Donnelly, supra, 397 U.S. 295.

Even assuming that we would have decided Donnelly as did the United States Supreme Court, and ignoring the structural differences between the federal courts and the courts of Michigan, Buxton cannot be considered a "deviant" decision. It was uncontradicted.

[1] People v Phillips, 416 Mich. 63, 74; 330 NW2d 366 (1982); People v George, 399 Mich. 638; 250 NW2d 491 (1977), explicate the niceties of the point at which an intermediate appellate court judgment is final, binding, and precedential.

[2] Orders denying leave to appeal express no opinion on the merits of the case. Frishett v State Farm Mutual Automobile Ins Co, 378 Mich. 733 (1966) (order); Great Lakes Realty Corp v Peters, 336 Mich. 325, 328-329; 57 NW2d 901 (1953); Malooly v York Heating & Ventilating Corp, 270 Mich. 240, 247; 258 N.W. 622 (1935).

[3] See Cussans v Harris, 118 Mich. App. 567; 325 NW2d 793 (1982); Putney v Gibson, 94 Mich. App. 466; 289 NW2d 837 (1979), rev'd sub nom Putney v Haskins, supra; Rowan v Southland Corp, 90 Mich. App. 61; 282 NW2d 243 (1979); Denham v Bedford, 82 Mich. App. 107; 266 NW2d 682 (1978), aff'd 407 Mich. 517; 287 NW2d 168 (1980).

[4] Similarly disadvantaged are those who, in obedience to the statute, settled the claim against the actually intoxicated person and were thus forced to abandon any claim against the dramshop.

[1] Ante, pp 363 and 362.

[2] 1972 PA 294, MCL 500.3101 et seq.; MSA 24.13101 et seq.

[3] See fns 24 and 52.

[4] The other vehicle may or may not be involved in the accident. See MCL 500.3114; MSA 24.13114, concerning priority of claims of occupants of motor vehicles, MCL 500.3115; MSA 24.13115, concerning priority of claims of persons not occupants of motor vehicles, and MCL 500.3171 et seq.; MSA 24.13171 et seq., concerning assigned claims facility and plan.

[5] See fn 32.

[6] Ante, p 361.

[7] 399 Mich. 827 (1977).

[8] Ante, p 363.

[9] We would further set down for argument whether we should grant rehearing in Putney, rehearing denied November 22, 1982, reconsideration denied April 26, 1983, 414 Mich. 1111, because it appears that it might be "unjust" to make the rule there stated applicable retroactively in only that case.

[10] See fn 1.

[11] Before the adoption of the 1963 Constitution providing for the Court of Appeals, this Court declared that "[t]he denial of an application for leave to appeal is * * * not equivalent of an affirmation of the decree sought to be reviewed". Malooly v York Heating & Ventilating Corp, 270 Mich. 240, 247; 258 N.W. 622 (1935). See also Great Lakes Realty Corp v Peters, 336 Mich. 325, 328-329; 57 NW2d 901 (1953).

This rule was extended to denials of applications for leave to appeal from decisions of the Court of Appeals in Frishett v State Farm Mutual Automobile Ins Co, 378 Mich. 733, 734 (1966), where it was said that a denial of leave to appeal "means that the Supreme Court expresses no present view with respect to the legal questions dealt with in the opinion of the Court of Appeals". See also State Bar of Michigan v Brotherhood of Railroad Trainmen, 383 Mich. 201, 208; 174 NW2d 811 (1970); Walters v Arenac Circuit Judge, 377 Mich. 37, 42; 138 NW2d 751 (1966).

[12] "Appeal may be taken to the Supreme Court only upon application and leave granted, in the discretion of the Supreme Court, from any decision of the Court of Appeals, interlocutory or final, upon a showing of a meritorious basis for appeal and any one of the following grounds.

"(1) The subject matter of the appeal involves legal principles of major significance to the jurisprudence of the State.

"(2) The decision of the Court of Appeals is clearly erroneous and will cause material injustice.

"(3) The decision is in conflict with decisions of the Supreme Court or other Court of Appeals decisions.

"(4) In any appeal of an interlocutory order of the Court of Appeals, it must be shown that appellant would suffer substantial harm by awaiting final judgment before taking appeal." GCR 1963, 853.1.

[13] Administrative Order No. 1977-4, 400 Mich. lxvii; Administrative Order No. 1983-4, 417 Mich. li.

[14] See fn 12.

[15] A decision of the Court of Appeals is binding on the trial courts; if there are conflicting decisions, a trial court may follow the view that it finds most persuasive. See In the Matter of Hague, 412 Mich. 532, 552; 315 NW2d 524 (1982).

[16] See fn 11.

[17] A decision of the Court of Appeals is not binding on other panels of the Court of Appeals. See Hackett v Ferndale City Clerk, 1 Mich. App. 6, 11; 133 NW2d 221 (1965); In the Matter of Hague, supra. Accordingly, a decision of the Court of Appeals is not a rule of law within that aspect of the doctrine of stare decisis that declares that "a decision on an issue of law embodied in a final judgment is binding on the court that decided it" even though, under another aspect of that doctrine, it would be binding on "such other courts as owe obedience to its decisions, in all future cases". 1B Moore, Federal Practice (2d ed), § 0.402[1], p 5. See also 20 Am Jur 2d, Courts, § 183, pp 519-520. As to its binding effect on "such other courts as owe obedience to its decisions", see also fn 15.

A decision by an appellate court on the merits of an issue establishes the "law of the case" for any court of equal or lower rank to which subsequent proceedings in the same case are addressed and, to that extent, a decision of the Court of Appeals would be binding on another panel of that Court. 5 Am Jur 2d, Appeal and Error, §§ 744, 747, pp 188, 191. See also CAF Investment Co v Saginaw Twp, 410 Mich. 428, 454; 302 NW2d 164 (1981). However, the Court of Appeals law of the case does not bind the Supreme Court on a subsequent appeal in the same case; see fn 18.

[18] The denial of an application for leave to appeal does not constitute an expression of opinion regarding the analysis or conclusion of the Court of Appeals. See fn 11. In Jones v Keetch, 388 Mich. 164; 200 NW2d 227 (1972), this Court rejected the argument that a decision of the Court of Appeals could become the law of the case and thereby bind the Supreme Court if the losing party in the Court of Appeals opted to proceed on remand rather than to seek review in the Supreme Court. See also Hack v Concrete Wall Co, 350 Mich. 118, 130; 85 NW2d 109 (1957).

Thus, whether a losing party in the Court of Appeals does not appeal or this Court denies an application for leave to appeal, the judgment of the Court of Appeals does not preclude review by this Court at a subsequent stage of the proceedings.

The United States Supreme Court has similarly held: (1) a denial of certiorari does not express any view concerning the merits of the lower court's judgment, Hamilton-Brown Shoe Co v Wolf Brothers & Co, 240 U.S. 251, 258; 36 S. Ct. 269; 60 L. Ed. 629 (1916); Maryland v Baltimore Radio Show, Inc, 338 U.S. 912, 917; 70 S. Ct. 252; 94 L. Ed. 562 (1950) (opinion of Frankfurter, J.); Hathorn v Lovorn, 457 U.S. 255, 262, fn 11; 102 S. Ct. 2421; 72 L. Ed. 2d 824 (1982); (2) a denial of certiorari does not render the Court of Appeals decision the law of the case for purposes of subsequent Supreme Court review, Mercer v Theriot, 377 U.S. 152, 153-154; 84 S. Ct. 1157; 12 L. Ed. 2d 206 (1964); Hathorn v Lovorn, supra; and (3) the failure of the losing party in the United States Court of Appeals to petition for certiorari does not render that court's judgment the law of the case for purposes of subsequent Supreme Court review, United States v United States Smelting, Refining & Mining Co, 339 U.S. 186, 198-199; 70 S. Ct. 537; 94 L. Ed. 750 (1950). See also Panama R Co v Napier Shipping Co, 166 U.S. 280, 284; 17 S. Ct. 572; 41 L. Ed. 1004 (1897); Messenger v Anderson, 225 U.S. 436, 444; 32 S. Ct. 739; 56 L. Ed. 1152 (1912); Hughes Tool Co v Trans World Airlines, Inc, 409 U.S. 363, 365, fn 1; 93 S. Ct. 647; 34 L. Ed. 2d 577 (1973); White v Higgins, 116 F2d 312, 317 (CA 1, 1940).

A majority of the state supreme courts which have explicitly addressed this question have also held that a lower appellate court's decision cannot become the law of the case, binding upon the supreme court of the state, even if the losing party on the first appeal did not seek review in the supreme court or the supreme court had denied permission to appeal after the first appellate court judgment. Sjostrom v Sproule, 33 Ill 2d 40, 41; 210 NE2d 209 (1965); Via v Asbestos Textile Co, Inc, 335 Mass 210, 212; 139 NE2d 393 (1957); Fantony v Fantony, 21 NJ 525, 532; 122 A2d 593 (1956); Hornstein v Podwitz, 254 NY 443, 450; 173 N.E. 674; 84 A.L.R. 1 (1930); Pengelly v Thomas, 151 Ohio St 51, 60; 84 NE2d 265 (1949). Contra: Gore v Bingaman, 20 Cal 2d 118, 119-120; 124 P2d 17 (1942); Berry v Howe, 39 Wash 2d 235, 238; 235 P2d 170 (1951); ROA Motors, Inc v Taylor, 220 Ga 122, 126; 137 SE2d 459 (1964); Life & Casualty Ins Co v Jett, 175 Tenn 295, 298; 133 S.W.2d 997 (1939). See also Baker v Baker, 207 S.W.2d 244, 250 (Tex Civ App, 1947), holding that Supreme Court's dismissal of application for review constitutes "approval" of lower court's action.

[19] See generally United States v Johnson, 457 U.S. 537; 102 S. Ct. 2579; 73 L. Ed. 2d 202 (1982); Los Angeles Dep't of Water & Power v Manhart, 435 U.S. 702; 98 S. Ct. 1370; 55 L. Ed. 2d 657 (1978); Hankerson v North Carolina, 432 U.S. 233; 97 S. Ct. 2339; 53 L. Ed. 2d 306 (1977); People v Stephens, 416 Mich. 252, 265-266; 330 NW2d 675 (1982); People v Phillips, 416 Mich. 63, 67-75; 330 NW2d 366 (1982).

[20] The record indicates that the automobile Dorothy Tebo was driving was owned by her or her husband or both of them.

[21] Under the no-fault act, PIP benefits for work loss within the statutory limit and medical expense are payable by a no-fault insurer without regard to the severity of the injury. MCL 500.3107; MSA 24.13107. Where the injury is of a severity meeting the statutory threshold (death, serious impairment of body function, or permanent serious disfigurement), the no-fault act permits a person injured in an automobile accident to maintain an action for noneconomic loss against the owner or driver of a vehicle involved in the collision. MCL 500.3135; MSA 24.13135. See fn 26.

Work loss and survivors' loss is limited to $1,000 per 30-day period, but by reason of cost of living adjustment is now $2,252 per 30-day period. MCL 500.3107, 500.3108; MSA 24.13107, 24.13108.

The Tebos had medical expense coverage with Blue Cross-Blue Shield of Michigan. When no-fault automobile insurance was purchased they did not elect to coordinate the Blue Cross-Blue Shield coverage with the no-fault coverage. The no-fault insurer of the Tebo automobile was required by the no-fault act to reimburse them for medical expense without regard to whatever benefits were paid or payable to the Tebos by Blue Cross-Blue Shield. See MCL 500.3109a; MSA 24.13109(1).

[22] 1972 PA 196, MCL 436.22; MSA 18.993.

[23] Tebo v Havlik, 109 Mich. App. 413; 311 NW2d 372 (1981).

[24] Id., p 415. See also cases discussed in fn 52 and 2 Harper & James, Torts, § 25.22, pp 1343-1344; Dobbs, Remedies, § 8.10, p 581; 4 Restatement Torts, 2d, § 920A, Comments b and c, pp 514-515.

[25] The Court said:

"This rule has been applied to bar a setoff against damages recoverable from a tortfeasor of payments received by the injured plaintiff under an insurance policy providing coverage for the losses caused by the defendant. Perrott v Shearer, 17 Mich. 48; (1868); Motts v Michigan Cab Co, 274 Mich. 437, 443-446; 264 N.W. 855 (1936)." Tebo v Havlik, fn 23 supra, p 415.

Perrott and Motts are discussed in fn 52.

The Court said that the rationale for the rule is that it would constitute an "unjust enrichment" of the tortfeasor to allow him to reduce his liability because plaintiffs, exercising "a contract right of recovery against their insurer" for which they had paid consideration in the form of premiums, had already been reimbursed for their loss. Tebo v Havlik, supra, p 415. Commentators have countered that insurance is generally purchased to assure prompt payment and to cover many perils so that only a fraction of the premium is paid to cover the peril which occurs, and that any unjust enrichment could be avoided by requiring the tortfeasor to reimburse or contribute to the loss of the insurer. Professor Dobbs has written:

"Perhaps the weakest argument made in support of the collateral source rule is the one that has been most mentioned in the courts — the wrongdoing defendant should not get the benefit of any reduction in the plaintiff's damages by a collateral source, since this would be a ``windfall' and since a windfall should be given to the plaintiff rather than to the wrongdoer. There are many answering arguments. It is possible to regard anything that diminishes or prevents a tortfeasor's potential liability as a windfall, but since there is no standard amount payable for a tort, and since some torts involving little fault are costly while others involving much fault cause little or no harm, it does not seem very meaningful to refer to diminished liability as a windfall. Furthermore, no one has suggested that the collateral source rule should be adjusted as fault increases or decreases, and in fact it is applied without regard to the tortfeasor's fault at all, and even in cases where liability is strict and there is no fault. It is now widely agreed, in addition, that many tort cases based on ``fault' involve no moral fault at all and that liability in a substantial number of cases is in fact strict liability, scantily clad in the rhetoric of negligence. To whatever extent this may be so, the ``wrongdoer' argument for the collateral source rule fails. Finally, whatever may be said of the individual defendant as a wrongdoer, the truth is that in most cases a collectible judgment is insured against and it is the insurer, not the individual defendant, who pays. Not only does this deprive the ``wrongdoer' thesis of any support where there is insurance, it also means that the collateral source rule is responsible for higher insurance premium costs." Dobbs, fn 24 supra, § 8.10, pp 586-587.

[26] See MCL 500.3135; MSA 24.13135, abolishing tort liability for economic loss, within the statutory limits, arising from the ownership, maintenance, or use of a motor vehicle with respect to which security (no-fault insurance or, where authorized, self insurance) is in effect but retaining tort liability for noneconomic loss where the injured person has suffered death, serious impairment of body function, or permanent serious disfigurement. See fn 21.

[27] See fn 26.

[28] Citizens Ins Co of America v Tuttle, 411 Mich. 536; 309 NW2d 174 (1981).

[29] The injured person may have failed to seek PIP benefits within the time limited by the no-fault act. MCL 500.3145; MSA 24.13145. In such a case, the tardy claimant may seek to obtain a full recovery of economic loss from a tortfeasor otherwise subject to liability for the loss.

There may be a factual dispute respecting the amount of work loss or medical expense or whether it was caused by an automobile accident. The injured person may settle the dispute with the no-fault insurer and seek to obtain the amount in dispute in an action against a tortfeasor.

[30] "The primary aim in measuring damages is compensation, and this contemplates that the damages for a tort should place the injured person as nearly as possible in the condition he would have occupied if the wrong had not occurred, and that the damages for breach of contract should place the plaintiff in the position he would be in if the contract had been fulfilled." McCormick, Damages, § 137, p 560.

See Allison v Chandler, 11 Mich. 542, 550 (1863); 2 Harper & James, Torts, § 25.1, p 1299.

[31] Prosser, Torts (4th ed), § 47, pp 296-297; 1 Dooley, Modern Tort Law, § 18, p 429; 2 Restatement Torts, 2d, § 433A, Comment i, pp 439-441.

[32] The Restatement of Torts provides:

"A payment by any person made in compensation of a claim for a harm for which others are liable as tortfeasors diminishes the claim against the tortfeasors, at least to the extent of the payment made, whether or not the person making the payment is liable to the injured person and whether or not it is so agreed at the time of payment or the payment is made before or after judgment." 4 Restatement Torts, 2d, § 885, subsection (3), p 333.

"A payment made by a tortfeasor or by a person acting for him to a person whom he has injured is credited against his tort liability, as are payments made by another who is, or believes he is, subject to the same tort liability." 4 Restatement Torts, 2d, § 920A, subsection (1), p 513.

See Manuel v Weitzman, 386 Mich. 157; 191 NW2d 474 (1971); Wood v Vroman, 215 Mich. 449, 465; 184 N.W. 520 (1921).

The rule is so well established that in Brewer v Payless Stations, Inc, 412 Mich. 673; 316 NW2d 702 (1982), the parties agreed that a $150,000 settlement paid by General Motors for alleged negligent design of an automobile fuel tank was required to be subtracted from any recovery from a gas station for negligent design of the station and maintaining a nuisance. The only question presented was whether the jury or the judge should subtract the $150,000 already recovered by the plaintiff. The Court held that the jury should not be informed of the settlement or of the amount that had been paid, unless the parties stipulated otherwise, and that following a jury verdict the court should calculate the amount by which the jury verdict would be reduced.

[33] See fn 32.

[34] See MCL 600.2925a; MSA 27A.2925(1); Moyses v Spartan Asphalt Paving Co, 383 Mich. 314; 174 NW2d 797 (1970); Caldwell v Fox, 394 Mich. 401; 231 NW2d 46 (1975).

[35] See fn 29.

[36] See fn 32.

[37] Citizens Ins Co of America v Tuttle, supra, which held that § 3135 of the no-fault act did not abolish the tort liability of non-motorist tortfeasors, involved a collision between an automobile and an errant cow. The question presented and decided adversely to the owner of the cow was whether a non-motorist was relieved of liability for negligence by the no-fault act.

[38] See fn 2.

[39] 1972 PA 196, MCL 436.22; MSA 18.993.

[40] One does not always recover PIP benefits from his own no-fault insurer. A person who does not own an automobile and does not reside with a relative who owns an automobile insured for PIP benefits recovers from the insurer of the owner or driver of a vehicle occupied or involved in the accident or from the assigned claims facility or, where applicable, the insurer of a vehicle operated in the business of transporting passengers. MCL 500.3114, 500.3115; MSA 24.13114, 24.13115. An employee and relatives residing with him injured while an occupant of the employee's vehicle receive PIP benefits "to which the employee is entitled from the insurer of the furnished vehicle". MCL 500.3114; MSA 24.13114.

A person who owns an automobile and is uninsured may recover PIP benefits, as if he did not own an automobile, if he suffers injuries in an automobile accident in which his vehicle is not involved. Heard v State Farm Mutual Automobile Ins Co, 414 Mich. 139; 324 NW2d 1 (1982).

There may be no insurance, and yet PIP benefits may be paid. Large corporations, common carriers, or employers may be self-insured. MCL 500.3101; MSA 24.13101.

[41] See fn 21.

[42] See fns 24 and 52.

[43] "Payments made by one who is not himself liable as a joint tortfeasor will go to diminish the claim of the injured person against others responsible for the same harm if they are made in compensation of that claim, as distinguished from payments from collateral sources such as insurance, sick benefits, donated medical or nursing services, voluntary continuance of wages by an employer, and the like." 4 Restatement Torts, 2d, § 885, Comment on subsection (3), pp 335-336.

See also 4 Restatement Torts, 2d, § 920A, p 513, and Comments b and c, p 514. See fn 32.

[44] See fn 52.

[45] The collateral-source doctrine and the distinction between fire (see fns 55 and 56 and accompanying text) and accident insurance in this regard have generated considerable critical comment. Professors Harper and James have written:

"Because the rule in accident insurance cases is so often urged as a model for social insurance legislation, it invites careful analysis. It is often justified on the ground that the insurance money is the product of the claimant's own thrift and foresight. But precisely the same thing is true of property insurance in respect to each of the notions that may underlie the thrift and foresight argument. Let us examine them. (1) Accident insurance is sometimes likened to an investment and so to life insurance with its heavy investment features. But the accident insurance premium buys only ``straight protection' just as the fire or collision insurance premium does. If the event insured against does not happen during the term of the policy the premium has still been fully spent for the protection. Nothing comes back. There is no saving or investment feature at all. (2) The rule of double recovery may be thought a desirable inducement to insure. But the certainty and speed of insurance payments, and the fact that they will be made in many situations where there is no tort at common law, operate as strongly here as in property insurance. Moreover it may be noted that this consideration is irrelevant in any scheme of insurance where participation is compulsory. (3) It may be thought that a man should be free to bargain and pay for double recovery if he wishes; and that he has done so in the case of accident insurance and has not in the case of property insurance. It is true that this may reflect the intentions of the parties today and that modern policies are so written. But it is not as simple as that. The concept of strict indemnity in property insurance is not rooted in the intentions of the parties but in the policy to prohibit wagering that has found expression in the doctrine of insurable interest, and this limits the sphere within which parties are free to contract. (Emphasis supplied.)

* * *

"The implications of the foregoing discussion for the problem of social insurance seem fairly clear. To the extent that welfare legislation is concerned with compensation or indemnity, accident insurance furnished a weak analogy, and there seems to be no basis for double recovery to the claimant whether the legislation is one providing for a contribution on his part or not. Presumably such schemes are not intended to provide windfalls. Certainly there is little if any trace of such intent in the philosophy of the American legislation, and it is submitted that the courts should be exceedingly reluctant to find it. The only theory that would justify such a result is one which would seek to redistribute the wealth beyond the point that indemnity or compensation calls for, and we probably do not want to do that in our country, though such a notion may well lie behind the current British solution that lets a claimant get social insurance benefits in full and makes him credit only half of them against a possible tort recovery. Moreover even if the wealth is to be redistributed, this would be a haphazard and capricious way to do it." 2 Harper & James, Torts, § 25.22, pp 1351-1354.

Professor Dobbs has written:

"In many instances of the collateral source rule it is possible to argue that the plaintiff paid for the benefit he is now receiving and that the defendant ought not benefit from that payment. This argument can be made both in cases of insurance benefits and in cases of employment benefits, since in the case of insurance the plaintiff has paid directly and in the case of fringe employment benefits he has paid indirectly by taking a lower salary. But if plaintiff paid for these benefits, it was not, surely, in hope of a double recovery but merely to gain minimal security, and if a double recovery is denied him his expectations will not be frustrated. And, although the plaintiff paid premiums for his insurance, he did not necessarily pay an amount equal to the benefits, and he did not pay alone, for the fund was created by many others in similar situations. In fact, both the plaintiff and others who contribute to the insurance (or fringe benefit) fund can get the security they seek more cheaply if the fund can be relieved of the obligation to pay when that will result in double compensation. Thus what the plaintiff has ``paid for' is really a function of the legal rule: if the collateral source rule were abolished, the plaintiff will have paid for security and not for the opportunity of a double recovery. He has paid for more only because the law, by allowing double recovery, in effect requires him to pay for more. The law could just as easily deny him double recovery with the result that his premiums could go down." Dobbs, Remedies, § 8.10, pp 584-585. (Emphasis supplied.)

See also Note, Unreason in the Law of Damages: The Collateral Source Rule, 77 Harv L Rev 741 (1964); Flemming, The Collateral Source Rule and Contract Damages, 71 Cal L Rev 56, 56-57, 85 (1983).

[46] Ante, p 366.

[47] The language of a no-fault policy is prescribed pursuant to law and is not negotiable and hence is not the subject of agreement or contract.

[48] "One court has observed the fallacy behind the assumption that ``there is some norm of damages which a defendant who has acted without reasonable care ought to pay for his careless act * * *.' No such norm is available. Even the most flagrant wrongdoer ordinarily is not liable for damages unless harm in fact results from his conduct; and when harm does occur, his liability depends, not on the degree of fault, but on the extent of injury, the physical idiosyncracies and earning capacity of the injured person, and on the latter's own degree of culpability. Furthermore, there is no indication that refusals to mitigate have been sensitive to varying degrees of moral fault; the intentional wrongdoer, the merely negligent tortfeasor, and he who is subject to absolute liability are treated in the same fashion. * * * Not only is the relationship between fault and the amount of damages too attenuated to be useful, but attempts to hold the defendant responsible for the ``harm' he has caused beg the question. The very issue in controversy is what constitutes the ``harm' to be compensated; does it include salary not in fact discontinued and medical expenses never actually incurred, or is it exclusive of insurance proceeds or expenses from which the injury relieves the victim?" Note, Unreason in the Law of Damages: The Collateral Source Rule, 77 Harv L Rev 741, 748-749 (1964).

[49] MCL 500.3101; MSA 24.13101.

[50] Tebo v Havlik, fn 23 supra, p 416.

[51] Life insurance is frequently an investment. The proceeds of life insurance are paid for a loss which cannot readily be quantified in dollars. While damage awards for loss of life must nevertheless be kept within reasonable bounds, the rule that the proceeds of life insurance are not to be subtracted from a damage award rarely results in redundancy.

It has been observed:

"Consistent with the consideration theory, a court might exclude evidence of life insurance proceeds on the ground that the wrong merely accelerated payment, but nevertheless admit evidence of accident insurance proceeds. The suggested distinction turns on an inquiry which has not been made by the courts: why does one purchase accident insurance? If the insured is speculating on the possibility of double recovery in the event of an accident attributable to the wrong of another, exclusion of the evidence may be warranted; mitigation would give him less than he bargained for — although an enlightened jurisdiction might regard such a bargain as contrary to public policy. But it seems more likely that the insured is purchasing security — prompt and sure payments without the necessity of litigation and without regard to the liability and financial resources of prospective defendants. The motives behind purchase of hospitalization insurance seem susceptible to similar characterization. If this view is correct, there exists no compelling reason why the insured should get more than he bargained for." See Note, Unreason in the Law of Damages: The Collateral Source Rule, 77 Harv L Rev 741, 750-751 (1964).

[52] Perrott v Shearer, 17 Mich. 48, 55 (1868), may be the oldest Michigan case on the collateral-source doctrine. The plaintiff collected under a policy of fire insurance. Defendant claimed that plaintiff should not recover from him for the loss of the goods because plaintiff had been made whole by the receipt of the proceeds of insurance. This Court ruled that although it seemed somewhat anomalous, plaintiff was entitled to recover twice. Defendant was a wrongdoer, and was not relieved of responsibility by the payment of insurance proceeds. He had no equities as against the plaintiff and no concern with any contract the plaintiff might have had. The owner of the policy did not buy insurance for the benefit of the tortfeasor. He had a right of action and had not given it away by taking out insurance. He paid an "equivalent" in the premium and was entitled to the benefit of it. The tortfeasor paid nothing for it and was entitled to no return. In answer to the argument that plaintiff recovered twice for his property, the Court said:

"[H]e recovers but once for the wrong done him, and he receives the insurance money upon a contract to which the defendant is in no way privy, and in respect to which his own wrongful act can give him no equities." Id., p 56.

The holding in Perrott was subsequently superseded by a statute providing a fire insurer rights of subrogation (see fn 55), but was followed in Baylis v Cronkite, 39 Mich. 413, 416 (1878), Hagan v Chicago, D & C G T J R Co, 86 Mich. 615; 49 N.W. 509 (1891), and Peter v Chicago & W M R Co, 121 Mich. 324; 80 N.W. 295 (1899). In Lebel v Swincicki, 354 Mich. 427, 433-434; 93 NW2d 281 (1958), the Court said that the insurance had not been provided with the intention of benefitting a tortfeasor. See also Cawood v Earl Paige & Co, 239 Mich. 485; 214 N.W. 402 (1927); Squires v Kalamazoo County Road Comm'rs, 378 Mich. 613; 147 NW2d 65 (1967) (uninsured motorist insurance).

In Motts v Michigan Cab Co, 274 Mich. 437; 264 N.W. 855 (1936), the court held that a tort victim is entitled to recover for lost wages, although wages had been paid. The trial court had charged the jury that there could be no recovery for lost wages unless the salary payments made by the employer over the period of disability were mere gratuities or gifts. While noting that there is a split of authority on the question, the Court followed what it said was the "prevailing doctrine" and reversed. The Court quoted from Roth v Chatlos, 97 Conn 282; 116 A 332 (1922):

"[W]hether the motive impelling their payment be affection, philanthropy, or contract, the injured is the beneficiary of their bounty and not him who caused the injury. * * * Such service or such payment is for the benefit of the injured person. It is a gift to him. But since it is one of the elements of injury, he is entitled to recover the reasonable value of the service." 274 Mich. 444-445.

Motts was followed in Royer v Eskovitz, 358 Mich. 279, 283; 100 NW2d 306 (1960), where the rule was held applicable to an employer's payment of his employee's medical expenses. In Canning v Hannaford, 373 Mich. 41; 127 NW2d 851 (1964), the Court extended the rationale of Motts to a case involving partnership profits.

Although no decision of this Court appears to be directly on point, several contain dicta suggesting that receipt of social security benefits (see fn 54) does not bar recovery from the tortfeasor. See, e.g., McCullough v Ward Trucking Co, 368 Mich. 108, 115-116; 117 NW2d 167 (1962) ("In the few cases in which the question has thus far engaged the attention of the courts, it has been uniformly held that the amount of recovery for death by wrongful act should not be diminished by the receipt of social security benefits," quoting Anno: 84 ALR2d 764, 765); Lebel v Swincicki, supra, p 434 ("It may be suggested also that such [social security survivors'] payments were made solely for the benefit of the recipients. In this respect they are analogous to the payments made under the insurance policies, above discussed."); Leitelt Iron Works v De Vries, 369 Mich. 47; 119 NW2d 101 (1963) (all social security and workers' compensation benefits were paid for by the decedent, his employer, or both and were not intended to benefit the tortfeasor). See also Wright v Delray C R Co, 361 Mich. 619; 106 NW2d 247 (1960).

In Kurta v Probelske, 324 Mich. 179, 188; 36 NW2d 889 (1949), the Court rejected defendant's claim that the damages awarded were excessive and said that plaintiff's receipt of unemployment benefits did not serve to mitigate his damages.

[53] As originally enacted, the workers' compensation act precluded an injured worker from both accepting workers' compensation benefits and maintaining an action against a third-party tortfeasor. The worker was put to an election of remedies. If he elected to receive workers' compensation, the employer was subrogated to the worker's claim against a third-party tortfeasor. That concept of barring duplicative recovery has been continued since the amendment of the workers' compensation act to eliminate the election of remedies and permit the injured worker to seek workers' compensation and yet sue a third-party tortfeasor; the first portion of the recovery goes to the employer, or his insurer, to reimburse the employer for workers' compensation benefits that have been paid, and any excess paid to the worker is treated as an advance payment, to that extent, of future workers' compensation benefits. See 1912 (1st Ex Sess) PA 10, part III, § 15; 1952 PA 155; 1969 PA 317, MCL 418.827; MSA 17.237(827).

In Wood v Vroman, fn 32 supra, the employer's insurer paid under the mistaken impression that it was liable under the workers' compensation law. The Court said that double recovery was permissible, since the insurer in fact had no obligation to pay. But see 4 Restatement Torts, 2d, §§ 885, 920A.

[54] Social security benefits are not analogous to PIP benefits. Social security benefits provide income maintenance upon the occurrence of any of countless different events which give rise to statutory eligibility. PIP benefits are provided only for loss arising out of the ownership, operation, maintenance, or use of a motor vehicle as a motor vehicle. MCL 500.3105; MSA 24.13105. While both social security and PIP benefits provide income maintenance and may become payable because of a disabling automobile accident, social security benefits, in contrast with PIP benefits, are not provided by the law of this state as reparations or as a substitute or replacement for a tort action or recovery from a tortfeasor or his insurer. The source of funding of social security benefits is a tax enacted for the first time when the social security program was instituted. The source of funding of PIP benefits, and a tort recovery which it replaces — a pool of money funded by premiums paid by owners and operators of automobiles — remains essentially unchanged after the enactment of the no-fault act.

[55] See MCL 500.2832; MSA 24.12832.

[56] 2 Harper & James, Torts, § 25.22, p 1351, see fn 45.

[57] See third paragraph of fn 21.

[58] The rationale of LeBlanc v State Farm Mutual Automobile Ins Co, 410 Mich. 173; 301 NW2d 775 (1981), was that the Legislature intended that the governmental health benefits there involved be included under § 3109a (MCL 500.3109a; MSA 24.13109[1]) and not that they were not governmental benefits under § 3109 (MCL 500.3109; MSA 24.13109).

[59] The Legislature has dealt with the question whether certain duplicative recoveries should be deducted from or offset against PIP benefits. Medical expense and other recoveries from certain sources are subtracted where the owner of the vehicle has elected to coordinate benefits in exchange for a reduced premium. MCL 500.3109(3); MSA 24.13109(3), MCL 500.3109a; MSA 24.13109(1). Benefits provided or required to be provided under the laws of any state, including workers' compensation benefits, Mathis v Interstate Motor Freight System, 408 Mich. 164; 289 NW2d 708 (1980), or the federal government, including social security survivors' benefits, O'Donnell v State Farm Mutual Automobile Ins Co, 404 Mich. 524; 273 NW2d 829 (1979), are required to be subtracted. See fn 60.

The Legislature has provided that a no-fault insurer may not recoup PIP benefits from a tort recovery obtained from a motorist except a duplicative recovery obtained from a motorist who is not relieved of tort liability under the provisions of the no-fault act. See MCL 500.3116; MSA 24.13116. Workman v DAIIE, 404 Mich. 477; 274 NW2d 373 (1979).

Section 3116, as originally enacted, literally required a person who received PIP benefits to repay the no-fault insurer out of a tort claim recovery a sum equal to the benefits received, but not more than the recovery, exclusive of attorney fees and other reasonable expenses incurred in effecting the recovery. In Workman, pp 509-510, this Court held it to be beyond the intent and purpose of the Legislature to permit the subtraction of PIP benefits (paid for non-excess economic loss) from a tort recovery for noneconomic and excess economic loss:

"However, if this statutory language [§ 3116] is read in the light of § 3135, we are left with an apparent and patent absurdity: the Legislature, on the one hand, provides an injured person limited tort remedy in § 3135 of the act while, on the other hand, providing that any tort recovery achieved pursuant to § 3135 will be effectively taken away under § 3116 of the act.

* * *

"Accordingly, in light of § 3135, we construe § 3116 to mean that an insurance carrier paying personal injury protection benefits is entitled to reimbursement from the tort recovery of a person injured as a result of a motor vehicle accident only if, and to the extent that, the tort recovery includes damages for losses for which personal injury protection benefits were paid. [Emphasis added.] Thus, since § 3135 abolishes tort remedy for losses covered under the personal injury protection insurance provisions of the act, an injured plaintiff should recover nothing for which the insurance carrier will have a right of reimbursement under § 3116. We believe this interpretation of § 3116 not only gives full effect to § 3135, but it also effectuates the essential purposes of this section, namely, to prevent double recovery of economic loss by those persons who retain their right to sue in tort for economic loss under the act." (Emphasis in original.)

In a separate opinion, we expressed the view that to provide the no-fault insurer with rights of "subrogation" against a tort recovery was neither remarkable (see Foremost Life Ins Co v Waters, 415 Mich. 303; 329 NW2d 688 [1982]) nor unconstitutional, but concurred in the result because the Legislature, in 1978 PA 461, had amended § 3116 to "clarify" the meaning and we were satisfied that the Legislature intended that the new version of § 3116 would control cases arising both before and after its enactment:

"Because it now appears that the Legislature intends that § 3116 require reimbursement out of a third-party tort recovery only where the defendant is uninsured, the accident occurs out of state, or the injury was intentionally inflicted, we concur in the Court's result on this issue." Workman, supra, pp 522-523.

Reimbursement of no-fault insurers is provided by amended § 3116(1) where, because the abolition of tort liability is limited to accidents with an insured vehicle within this state that are not intentionally caused (MCL 500.3135; MSA 24.13135), the injured person realizes a duplicative recovery upon a tort claim arising from an accident involving an uninsured vehicle or occurring outside this state or where the injury is intentionally caused.

To argue that § 3116 as amended should be read as barring a no-fault insurer from being reimbursed for PIP benefits from a duplicative recovery of economic loss obtained from a non-motorist would be to ascribe to the Legislature an intent and purpose beyond the "evil" sought to be remedied by the amendment of § 3116.

The no-fault act, including § 3116, addresses the tort liability of owners and drivers and provides a substitute system of reparations in lieu of a motorist's tort liability for work loss within the statutory limit and medical expense. The provisions of the no-fault act do not affect actions involving non-motorists. See Citizens Ins Co of America v Tuttle, supra, p 547, where this Court (in considering § 3135 of the no-fault act, which, like amended § 3116, uses the word "only") said that the non-motorist is "outside the basic no-fault system of allocating the costs of accidents". The amendment of § 3116 does not address the question whether a duplicative recovery for economic loss may be obtained by an injured person from a non-motorist.

[60] MCL 500.3109; MSA 24.13109; see fn 59. The questions whether certain other social security benefits are governmental benefits have been argued and are awaiting decision. Jarosz v DAIIE, 109 Mich. App. 86; 310 NW2d 903 (1981), lv gtd 414 Mich. 872 (1982); Thompson v DAIIE, 107 Mich. App. 256; 309 NW2d 228 (1981), lv gtd 414 Mich. 873 (1982).

[61] MCL 500.3116; MSA 24.13116; see fn 59.

[62] Ante, p 367.

[63] Ante, p 367.

[64] Ante, p 368.

[65] A holding that a no-fault insurer can maintain such an action would engender further litigation, the primary purpose of which would be to reallocate losses between insurers. MCL 500.3116; MSA 24.13116 provides for such further litigation where the injured person obtains a duplicative recovery from a motorist who is not relieved of tort liability under the no-fault act. It is thus arguable that it would not be inconsistent with the policy of the act to allow a no-fault insurer to seek to recover PIP benefits from a non-motorist.

The provisions of the Uniform Contribution Among Tortfeasors Act (MCL 600.2925a; MSA 27A.2925[1]) would not be applicable because a no-fault insurer is not a tortfeasor. Nevertheless, a no-fault insurer and a non-motorist tortfeasor are under a common burden of liability for the payment of wage/work loss within the statutory limit and medical expense. It is arguable that an action for equitable contribution may be maintained by a no-fault insurer against a non-motorist tortfeasor. Cf. Moyses v Spartan Asphalt Paving Co, 383 Mich. 314, 331, 334; 174 NW2d 797 (1970); Caldwell v Fox, 394 Mich. 401, 420, fn 5, 421; 231 NW2d 46 (1975); United States Fidelity & Guaranty Co v Liberty Mutual Ins Co, 127 Mich. App. 365, 373; 339 NW2d 185 (1983).

[66] Ante, p 367.

[67] See fn 32.

[68] See fn 32.

[69] Where a no-fault insured elects to coordinate medical expense insurance (see third paragraph of fn 21), the medical expense insurance substitutes, to the extent of the benefits thereby provided, for the statutorily mandated PIP benefits. We would, on that basis, conclude that in such a case medical expense insurance functions to that extent as no-fault security and as PIP benefits and hence is not received from a collateral source.

[70] By reason of the cost of living adjustment, work loss is now recoverable up to $2,252 per 30-day period. Because PIP benefits are not taxable income, the amount recoverable as PIP benefits is ordinarily 85% of the actual wage loss. MCL 500.3107; MSA 24.13107. It thus appears that unless the injured person's income from work he would have performed exceeds approximately $29,000, the work-loss benefits will fully compensate him for his lost wages.

[71] The no-fault act requires public liability insurance of $20,000/$40,000 where the injury or damage occurs in this state. MCL 500.3131; MSA 24.13131.

[72] See fn 4.

[73] See fn 4.

[74] See fn 59 discussing MCL 500.3109; MSA 24.13109, providing for the subtraction from PIP benefits of benefits provided or required to be provided by federal or state law, and MCL 500.3116; MSA 24.13116, allowing a no-fault insurer to recover PIP benefits out of a duplicative tort recovery from an owner or driver of a motor vehicle who is not relieved of tort liability by the no-fault act.

[75] See fn 22.

[76] The $150,000 paid by General Motors in Brewer v Payless Stations, Inc, 412 Mich. 673; 316 NW2d 702 (1982) could have been greater than any amount awarded by a jury against Payless. See fn 32.

[77] See Manuel v Weitzman, fn 32 supra, p 167. Larabell v Schuknecht, 308 Mich. 419, 423; 14 NW2d 50 (1944).

[78] See fn 32.

[79] The deduction of an amount obtained in settlement from an intoxicated driver from a verdict against a tavern is not likely to occur in the future as there can be no settlement unless the intoxicated driver and the tavern both settle; the tavern would be released from liability if the settlement is only with the intoxicated driver. There will henceforth be no setoff as such, but only contribution (see MCL 600.2925a; MSA 27A.2925[1]) and possibly indemnification where the owner and the driver are different persons.

[80] The holding in Virgilio v Hartfield, 4 Mich. App. 582; 145 NW2d 367 (1966), was adverted to in Barton v Benedict, 39 Mich. App. 517, 523; 197 NW2d 898 (1972), rev'd on other grounds sub nom Podbielski v Argyle Bowl, Inc, 392 Mich. 380; 220 NW2d 397 (1974), where the Court said, on the authority of Virgilio:

"There is no question that the dramshop defendant cannot require contribution from the driver-defendant, since the theory of recovery in the wrongful death action is different from the theory of recovery in the dramshop action."

The Court said that the holding in Virgilio was not there applicable because "[t]he question here is not one of contribution but rather one of mitigation by a prior settlement with the intoxicated driver". Id. In Reno v Heineman, 56 Mich. App. 509; 224 NW2d 687 (1974), the Court of Appeals, after considering Virgilio, followed Barton in holding that a settlement with the intoxicated driver was admissible in evidence in mitigation of the injured person's damages in an action against a tavern under the dramshop act.

Other panels of the Court of Appeals have held, however, that no subtraction should be made. See Putney v Gibson, 94 Mich. App. 466, 486; 289 NW2d 837 (1979) (relying on Virgilio), rev'd on other grounds sub nom Putney v Haskins, 414 Mich. 181; 324 NW2d 729 (1982); Bohannon v Campbell, 38 Mich. App. 422; 196 NW2d 836 (1972). See text following fn 67.

[81] In Virgilio, the Court of Appeals held that there was no right of contribution between an intoxicated driver and a tavern, because the statute at that time provided for contribution between "joint tortfeasors", 1961 PA 236, § 2925, and this Court had previously declared that tortfeasors whose independent concurring acts of negligence combined to cause the plaintiff's loss are not joint tortfeasors. Geib v Slater, 320 Mich. 316; 31 NW2d 65 (1948).

The Court in Virgilio also said that the liability of the intoxicated driver and of the tavern did not arise under a common theory of liability. The basis of liability against the intoxicated driver is negligence. The tavern's liability "is not grounded on any theory of negligence, but arises because of violation of the statute". Virgilio, p 585.

[82] The rule set forth in Virgilio (see fn 81) has been superseded by this Court's decision in Moyses v Spartan Asphalt Paving Co, 383 Mich. 314, 331, 334; 174 NW2d 797 (1970), as elucidated in Caldwell v Fox, 394 Mich. 401, 420-421, fn 5; 231 NW2d 46 (1975), and by changes (see fn 83) in the Revised Judicature Act (1974 PA 318 added MCL 600.2925a; MSA 27A.2925[1] and MCL 600.2925b; MSA 27A.2925[2], since amended by 1982 PA 147) to provide for contribution among tortfeasors without regard to whether they are joint tortfeasors or whether there is a common theory of liability. It is now sufficient that the tortfeasors are under a common burden of liability to the plaintiff.

In Moyses, p 331, the Court reviewed the history of § 2925 of the Revised Judicature Act and acknowledged that "joint tortfeasors" would not include those who may be joined as defendants under the court rules "and held responsible to [the injured person] for damages sustained on account of their causally cooperating but non-joint acts or omissions, say by the negligence of one, the violation of a statute like the dramshop act by another, and the breach by still another of an express or legally implied warranty". (Emphasis supplied.) The Court abolished the common-law limitations on contributions among wrongdoers, leaving the "issue of contribution to equitable principles as known hitherto and now in matters of express or implied contract". Moyses, p 335. The Court said that it was stimulated to take such action "by the national advancement of principles set forth in the Uniform Contribution Among Tortfeasors Act, supra, the tendency of other courts to provide by judicial action the right of contribution on behalf of all but intentional wrongdoers, and the compelling admonitions of modern writers like [Dean William L.] Prosser". Moyses, p 334.

In Caldwell v Fox, supra, the Court said that the doctrine of contribution among non-intentional wrongdoers had been returned by Moyses "to the original equitable rules".

In Mason v Lovins, 24 Mich. App. 101, 117-119; 180 NW2d 73 (1970), relied on in Putney v Gibson, supra, p 485, "[n]o basis [had] been suggested for allocating the $9,500 settlement between the pecuniary loss of the widow and child and the other elements of damage for which the administrator could seek to recover". Here, the elements of loss for which PIP benefits are payable and for which offset against a recovery from a tavern is sought are identical. See Reno v Heineman, supra, p 512.

[83] MCL 600.2925a et seq.; MSA 27A.2925(1) et seq. The 1974 change in the Revised Judicature Act (see fn 82), adopted at the urging of the Law Revision Commission (Michigan Law Revision Commission, Second Annual Report, 1967, p 57), repealed § 2925 and substituted as §§ 2925a-2925d of the Revised Judicature Act, MCL 600.2925a-600.2925d; MSA 27A.2925(1)-27A.2925(4), the substance of the 1955 Uniform Contribution Among Tortfeasors Act (12 ULA 63). Section 2925a provides for contribution between "persons" "jointly or severally liable in tort". (Emphasis supplied.) A tortfeasor who has paid more than his pro-rata share of the common liability has a right of contribution. There is no requirement that there be a common theory of liability. It is enough that the tortfeasor seeking contribution and the one from whom contribution is sought both commonly share a burden of some sort of liability to the plaintiff. In 1982, the Revised Judicature Act was amended to provide that the relative degrees of fault of tortfeasors should be considered "as between themselves only and without affecting the rights of the injured party to a joint and several judgment". 1982 PA 147, MCL 600.2925b; MSA 27A.2925(2).

[84] The 1939 or 1955 Uniform Contribution Among Tortfeasors Act has been adopted in twenty jurisdictions: Alaska, Arkansas, Colorado, Delaware, Florida, Hawaii, Maryland, Massachusetts, Mississippi, Nevada, New Jersey, New Mexico, North Carolina, North Dakota, Ohio, Pennsylvania, Rhode Island, South Dakota, Tennessee and Wyoming. See 12 Uniform Laws Annotated, § 1, 1983 Cumulative Supplement, p 58.

Of the twenty states that have adopted either the 1939 or 1955 uniform act, the courts of seven have said that a common theory of liability is not a prerequisite to contribution. Clark v Brooks, 377 A2d 365 (Del Super, 1977), aff'd sub nom Blackshear v Clark, 391 A2d 747 (Del, 1978); Wolfe v Ford Motor Co, 386 Mass 95; 434 NE2d 1008 (1982); White v McKenzie Electric Cooperative, Inc, 225 F Supp 940 (D ND, 1964) (applying North Dakota law); Adler's Quality Bakery, Inc v Gaseteria, Inc, 32 NJ 55; 159 A2d 97 (1960); Neveroski v Blair, 141 NJ Super 365; 358 A2d 473 (1976); Cartel Capital Corp v Fireco of New Jersey, 81 NJ 548; 410 A2d 674 (1980); Sanchez v City of Espanola, 94 NM 676; 615 P2d 993 (Ct App, 1980), rev'd on other grounds sub nom Aalco Mfg Co v City of Espanola, 95 NM 66; 618 P2d 1230 (1980); Chamberlain v Carborundum Co, 485 F2d 31 (CA 3, 1973) (applying Pennsylvania law); City of Kingsport v SCM Corp, 429 F Supp 96 (ED Tenn, 1976) (applying Tennessee law, but not mentioning the uniform act); see also Southern R Co v Foote Mineral Co, 384 F2d 224 (CA 6, 1967) (Tennessee law before adoption of the uniform act did not require a common theory).

Six states that have not adopted a uniform act do not require a common theory as a prerequisite to contribution. Safeway Stores, Inc v Nest-Kart, 21 Cal 3d 322; 146 Cal Rptr 550; 579 P2d 441 (1978); Skinner v Reed-Prentice Div Package Machinery Co, 70 Ill 2d 1; 374 NE2d 437 (1977), cert den 436 U.S. 946 (1978); Morgan v Kirk Bros, Inc, 111 Ill App 3d 914; 444 NE2d 504 (1982); Federated Mutual Implement & Hardware Ins Co v Dunkelberger, 172 NW2d 137 (Iowa, 1969), partially overruled on other grounds Lewis v State, 256 NW2d 181, 189 (Iowa, 1977); see also Chicago, R I & P R Co v Chicago & NW R Co, 280 F2d 110 (CA 8, 1960), cert den 364 U.S. 931 (1961) (applying Iowa law and a federal statute); Guillard v Niagara Machine & Tool Works, 488 F2d 20 (CA 8, 1973) (applying Minnesota law); Anderson v Comardo, 107 Misc. 2d 821; 436 NYS2d 669 (1981); Bristol-Myers Co v Gonzales, 548 S.W.2d 416 (Tex Civ App, 1976), rev'd on other grounds 561 S.W.2d 801 (Tex, 1978); see also Avery v Maremont Corp, 628 F2d 441 (CA 5, 1980) (applying Texas law).

In the following states, which have not expressly disclaimed the common theory requirement, it appears that a common theory would not be required. Tamashiro v DeGama, 51 Haw. 74; 450 P2d 998 (1969); Albert v Dietz, 283 F Supp 854 (D Hawaii, 1968) (applying Hawaiian law; Hawaii has adopted the uniform act); Bedard v Greene, 409 A2d 676 (Me, 1979); Testa v Winquist, 451 F Supp 388 (D RI, 1978) (applying Rhode Island law, which includes a uniform act, and not questioning contribution between two defendants liable on different theories).

A common theory is not a prerequisite to contribution under certain federal statutes. Reynolds v Southern R Co, 320 F Supp 1141 (ND Ga, 1969) (Federal Employers' Liability Act, collecting cases from state and federal courts); Fischbach & Moore International Corp v Crane Barge R-14, 476 F Supp 282 (D Md, 1979), aff'd 632 F2d 1123 (CA 4, 1980) (maritime law).

The following is an arrangement of the cases according to the bases of liability involved:

Separate negligence of both an employee and an employer: Clark.

Negligence of a vehicle manufacturer and breach of implied warranty of merchantability by a vehicle assembler-seller: Wolfe.

Strict liability and negligence: Adler's, Safeway Stores, Chamberlain, Skinner, Bristol-Myers, and Sanchez.

Statutory violation (fraud and deception acts) and common-law fraud: Neveroski.

Misrepresentation, active negligence, and passive negligence: City of Kingsport.

Common-law negligence and dramshop violation: Dunkelberger and Anderson.

Negligence and FELA violation: Chicago, R I & P R Co; Reynolds.

[85] Actions brought under the civil damage (dramshop) statutes of other states are generally regarded as penal in nature. 45 Am Jur 2d, Intoxicating Liquors, § 608, p 891. Early decisions of this Court adopted the view that exemplary damages could be awarded under the dramshop act to punish the defendant. Larzelere v Kirchgessner, 73 Mich. 276, 283; 41 N.W. 488 (1889); Rouse v Melsheimer, 82 Mich. 172, 176; 46 N.W. 372 (1890); Peacock v Oaks, 85 Mich. 578, 582; 48 N.W. 1082 (1891).

However, the Court's subsequent decisions distinctly repudiated this view. Bowden v Voorheis, 135 Mich. 648, 652; 98 N.W. 406 (1904); Manzer v Phillips, 139 Mich. 61, 68; 102 N.W. 292 (1905). These decisions held that exemplary damages under the statute could be awarded only to compensate injury to the feelings of the plaintiff caused by the wilful, wanton, or reckless invasion of plaintiff's rights by the defendant. Bowden v Voorheis, 135 Mich. 652; Hink v Sherman, 164 Mich. 352, 360; 129 N.W. 732 (1911). See also Haviland v Chase, 116 Mich. 214, 217; 74 N.W. 477 (1898) (dicta). Moreover, these decisions expressly rejected the contention that exemplary damages could be imposed to punish a saloon keeper or "to make an example [of the tavern keeper] for the public good". Boydan v Haberstumpf, 129 Mich. 137, 140; 88 N.W. 386 (1901). See also Ford v Cheever, 105 Mich. 679, 685; 63 N.W. 975 (1895). Thus "[t]he damages [under the dramshop act] are not punitive; they may be exemplary as provided for by the statute". Larabell v Schuknecht, 308 Mich. 419, 423; 14 NW2d 50 (1944).

There have been two references in the years since 1900 to the statute as a penal measure. See Bailey v Briggs, 143 Mich. 303, 304; 106 N.W. 863 (1906) (this statutory section is "in part penal and in part remedial"); Barton v Benedict, supra, p 524 ("The ``actual' damages would appear to be remedial and compensatory in nature, while the ``exemplary' damages would appear to be penal in nature."), subsequently relied on in Baldridge v Eastman's, Inc, 52 Mich. App. 1, 6-7; 216 NW2d 615 (1974). The penal aspects referred to by the Court in Bailey, though, are provisions prohibiting the sale of liquor to minors and the sale of liquor on Sunday; the provision granting a civil remedy would, therefore, still fall within the "remedial" rubric. The only authority cited by the Barton Court in support of its speculative conclusion is Bailey v Briggs; the Court did not address the earlier decisions of this Court holding that exemplary damages were not to be awarded under the dramshop act for the purpose of punishing or making a public example of the offending saloon keeper. See, e.g., Bowden v Voorheis, supra; Manzer v Phillips, supra; Hink v Sherman, supra.

All doubt whether the pre-1972 statute was penal or punitive in nature was removed in 1972 when the Legislature amended the statute and eliminated the exemplary damages provision. Since 1972, the act limits the plaintiff to a recovery of actual or compensatory damages.

[86] See fn 22.

[87] The dramshop act has been liberally construed, and this Court has ruled that the legislative intent is not to be ascertained by a strained or narrow construction of the words employed. Eddy v Courtright, 91 Mich. 264, 267; 51 N.W. 887 (1892); Podbielski v Argyle Bowl, Inc, 392 Mich. 380, 384-385; 220 NW2d 397 (1974). See also Dickerson v Heide, 69 Mich. App. 303, 309; 244 NW2d 459 (1976). Thus, where statutory language is ambiguous, this Court may tend to adopt the more liberal of the competing alternatives. See, e.g., LaBlue v Specker, 358 Mich. 558; 100 NW2d 445 (1960) (illegitimate and nonviable fetus is a "child" or "other person" within protection of statute); Salas v Clements, 399 Mich. 103; 247 NW2d 889 (1976) ("name and retain" provision does not prohibit action if the identity of the intoxicated person is not known).

This Court has also held, however, that because the Legislature has statutorily imposed liability upon liquor licensees in derogation of the common-law rule of no liability, the statute must be strictly construed despite its remedial nature. Holland v Eaton, 373 Mich. 34, 39; 127 NW2d 892 (1964); Browder v International Fidelity Ins Co, 413 Mich. 603; 321 NW2d 668 (1982); Putney v Haskins, supra. A court may not add provisions excluded altogether by the Legislature. Here, there is no ambiguity to be resolved; the Legislature in 1972 explicitly excluded any recovery of exemplary damages.

[88] See fn 32.

[89] See fn 26.