Opinion
ARGUELLES, J.In June 1986, the voters of California approved an initiative measure, the Fair Responsibility Act of 1986 (Civ. Code, §§ 1431 to 1431.5)—popularly known as, and hereafter referred to, as Proposition 51—which modified the traditional, common law “joint and several liability” doctrine, limiting an individual tortfeasor’s liability for noneconomic damages to a proportion of such damages equal to the tortfeasor’s own percentage of fault.1 Just a few weeks after the election, the underlying *1193personal injury action in this case—which arose out of a July 1980 accident and which had been pending for nearly five years prior to the June 1986 election—was assigned for trial. Before the trial began, the parties requested the trial court to determine, inter alia, whether the newly revised joint and several liability doctrine would apply to this case. Plaintiff contended that the new legislation should not be applied for a number of reasons, maintaining (1) that Proposition 51 is unconstitutional on its face, and (2) that, in any event, the measure does not apply retroactively to causes of action which accrued prior to its effective date.2 Defendants contested both arguments.
The trial court concluded (1) that Proposition 51 is constitutional on its face and (2) that it should be applied to all cases coming to trial after its effective date, including this case, regardless of when the cause of action accrued. Reviewing the trial court’s ruling in these consolidated pretrial writ proceedings, the Court of Appeal upheld the trial court’s determination in all respects, declining—with respect to the retroactivity issue—to follow another recent Court of Appeal decision, Russell v. Superior Court (1986) 185 Cal.App.3d 810 [230 Cal.Rptr. 102], which had concluded that Proposition 51 does not apply retroactively to causes of action which arose prior to the initiative’s effective date. Because of the importance of the issues and the conflict in Court of Appeal decisions on the retroactivity question, we granted review.
As we shall explain, we have concluded that the Court of Appeal judgment should be affirmed in part and reversed in part. On the constitutional question, we agree with the Court of Appeal that plaintiff’s facial constitutional challenge to Proposition 51 is untenable. Past decisions of this court make it quite clear that the initiative measure—in modifying the common law rule governing the potential liability of multiple tortfeasors—violates neither the due process nor equal protection guaranties of the state or federal Constitution. Although the proposition’s language leaves a number of issues of interpretation and application to be decided in future cases, those unsettled questions provide no justification for striking down the measure on its face.
On the question of retroactivity, we conclude that the Court of Appeal erred in ruling that Proposition 51 applies to causes of action which accrued before the measure’s effective date. It is a widely recognized legal principle, specifically embodied in section 3 of the Civil Code, that in the absence of a clear legislative intent to the contrary statutory enactments apply *1194prospectively. The drafters of the initiative measure in question, although presumably aware of this familiar legal precept, did not include any language in the initiative indicating that the measure was to apply retroactively to causes of action that had already accrued and there is nothing to suggest that the electorate considered the issue of retroactivity at all. Although defendants argue that we should nonetheless infer a legislative intent on the part of the electorate to apply the measure retroactively from the general purpose and context of the enactment, the overwhelming majority of prior judicial decisions—both in California and throughout the country—which have considered whether similar tort reform legislation should apply prospectively or retroactively when the statute is silent on the point have concluded that the statute applies prospectively. Reflecting the commonsense notion that it may be unfair to change “the rules of the game” in the middle of a contest, these authorities persuasively demonstrate that the general legal presumption of prospectivity applies with full force to a measure, like the initiative at issue here, which substantially modifies a legal doctrine on which many persons may have reasonably relied in conducting their legal affairs prior to the new enactment.
Contrary to the extravagant rhetoric of the dissenting opinion, our conclusion that Proposition 51 must properly be interpreted to apply prospectively does not postpone or delay the operative effect of Proposition 51 and is in no way inconsistent with the fact that the measure was adopted in response to a liability crisis. As we explain, the new legal doctrine established by Proposition 51 went into effect the day following the passage of the initiative and could immediately be relied on by insurance companies to reduce insurance premiums and by potential tort defendants to resume activities they may have curtailed because of the preexisting joint and several liability rule. Indeed, although the dissenting opinion vigorously asserts that Proposition 51’s relationship to a liability crisis proves that the electorate must have intended that the measure would be applied retroactively, that assertion is clearly belied by the numerous recent tort reform statutes, adopted in other states in response to the same liability crisis, which, by their terms, are expressly prospective in operation. (See post, pp. 1219-1220.) As these statutes demonstrate, a prospective application of Proposition 51 is totally compatible with the history and purpose of the initiative measure.
I.
In July 1980, plaintiff Gregory Evangelatos, an 18-year-old high school student, was seriously injured in his home, apparently while attempting to make fireworks with chemicals purchased from a retail store. In July 1981, plaintiff filed an action for damages against the retailer (Student Science *1195Store, Inc.), the wholesale distributor (Van Waters & Rogers, Inc.), and four manufacturers of the chemicals he was using, alleging that defendants were liable for his injuries on both negligence and strict liability theories. The causes of action against three of the manufacturers were dismissed on summary judgment and plaintiff voluntarily dismissed the action against the fourth manufacturer. The case proceeded against the retailer and the wholesale distributor of the chemicals.
On June 23, 1986, almost five years after the action had been filed, the case was assigned for trial. Before the trial began, plaintiff and the two remaining defendants filed motions with the trial court seeking a determination whether Proposition 51, which had been approved by the voters just three weeks earlier at the June 3, 1986, election, would be applied in this case. The motions sought a determination of the constitutional validity of the proposition and, if valid, a resolution of various questions relating to the applicability and proper interpretation of the measure.
After briefing, the trial court issued a lengthy written statement, ruling on five separate issues. The court concluded (1) that Proposition 51 was validly enacted and is not unconstitutional on its face; (2) that the measure applies to all cases, including the present proceeding, which had not gone to trial before June 4, 1986, the date on which the initiative measure became effective, regardless of when the cause of action arose; (3) that in determining each defendant’s “several” liability for a portion of plaintiff’s noneconomic damages under the proposition, the trier of fact may consider the conduct of all persons whose fault contributed to plaintiff’s injury, not just the conduct of plaintiff and defendants who are parties to the action; (4) that future medical expenses and loss of future earnings are “economic damages” within the meaning of Proposition 51 for which defendants remain jointly and severally liable; and (5) that for purposes of apportioning fault in this case, the summary judgment that had been entered in favor of three manufacturers constituted a determination that no causative fault could properly be attributed to them.
Immediately following the ruling, plaintiff and one of the defendants (Van Waters & Rogers, Inc.) filed separate mandate petitions in the Court of Appeal, challenging different aspects of the trial court’s decision. The Court of Appeal initially denied both petitions summarily, and the parties then sought review in this court. Shortly before the petitions reached us, another Court of Appeal rendered its decision in Russell v. Superior Court, supra, 185 Cal.App.3d 810, holding Proposition 51 inapplicable to all causes of action which accrued before the measure’s effective date. On October 29, 1986, our court denied a petition for review in Russell and transferred the two petitions in this matter to the Court of Appeal with *1196directions to issue alternative writs. Our order directed the Court of Appeal’s attention to the Russell decision.
On remand, the Court of Appeal issued alternative writs, consolidated the matters for briefing and argument, and ultimately concluded that the trial court had correctly resolved all of the questions at issue, including the facial constitutionality of the measure and its applicability to the instant case. Although the Court of Appeal recognized that the Russell court had reached a contrary conclusion on the retroactivity issue, it disagreed with the Russell decision, concluding that, while the initiative measure contained no express or affirmative indication that the measure was intended to apply retroactively, in its view “the legislative intent was for the statute to take effect immediately and to apply to as many cases as feasible.” Finding that it would be unduly disruptive to require retrial of all tort cases that had been tried before the enactment of Proposition 51 but in which judgments had not yet become final, the Court of Appeal concluded that “[t]he maximum feasible application of the Act is to all cases yet to be tried, including this one.”
Both plaintiff and defendant petitioned for review, and we granted review to resolve the important questions presented by the case.
II.
Before analyzing either the constitutional or retroactivity issues, we believe it may be useful to place Proposition 51’s modification of the common law joint and several liability doctrine in brief historical perspective.
Prior to the adoption of comparative negligence principles in California in the mid-1970’s, the jury, in assessing liability or awarding damages in an ordinary tort action, generally did not determine the relative degree or proportion of fault attributable either to the plaintiff, to an individual defendant or defendants, or to any nonparties to the action. Under the then-prevailing tort doctrines, the absence of any inquiry into relative culpability had potentially harsh consequences for both plaintiffs and defendants. On the one hand, if a plaintiff was found to be at all negligent, no matter how slight, under the contributory negligence rule he was generally precluded from obtaining any recovery whatsoever. (See generally 4 Witkin, Summary of Cal. Law (8th ed. 1974) Torts, § 683, p. 2968 and authorities cited.) On the other hand, if a defendant was found to be at all negligent, regardless of how minimally, under the joint and several liability rule he could be held responsible for the full damages sustained by the plaintiff, even if other concurrent tortfeasors had also been partially, or even primarily, responsible for the injury. (See id., § 35, pp. 2333-2334.) Moreover, the governing *1197rules at that time gave the plaintiff unilateral authority to decide which defendant or defendants were to be sued (see id., § 37, p. 2335); a defendant who had been singled out for suit by the plaintiff generally had no right to bring other tortfeasors into the action, even if the other tortfeasors were equally or more responsible for the plaintiff’s injury (see id., § 46, p. 2346).3
In Li v. Yellow Cab Co. (1975) 13 Cal.3d 804 [119 Cal.Rptr. 858, 532 P.2d 1226, 78 A.L.R.3d 393], this court took an initial step in modifying this traditional common law structure, ameliorating the hardship to the plaintiff by abrogating the all-or-nothing contributory negligence doctrine and adopting in its place a rule of comparative negligence. Li held that “the contributory negligence of the person injured . . . shall not bar recovery, but the damages awarded shall be diminished in proportion to the amount of negligence attributable to the person recovering.” (13 Cal.3d at p. 829.)
In American Motorcycle Assn. v. Superior Court (1978) 20 Cal.3d 578 [146 Cal.Rptr. 182, 578 P.2d 899], our court took the next step in modifying the traditional structure, this time altering the preexisting common law doctrines to diminish the hardship to defendants. Although the American Motorcycle court concluded that the traditional common law joint and several liablity doctrine should be retained—relying, in part, on the fact that at that time the “overwhelming majority” of jurisdictions that had adopted comparative negligence had also retained the joint and several liability rule (20 Cal.3d at p. 590)—at the same time the American Motorcycle court held (1) that plaintiffs should no longer have the unilateral right to determine which defendant or defendants should be included in an action and that defendants who were sued could bring other tortfeasors who were allegedly responsible for the plaintiff’s injury into the action through cross-complaints (20 Cal.3d at pp. 604-607), and (2) that any defendant could obtain equitable indemnity, on a comparative fault basis, from other defendants, thus permitting a fair apportionment of damages among tortfeasors. (See 20 Cal.3d at pp. 591-598.)
Subsequent cases established that under the principles articulated in American Motorcycle, supra, 20 Cal.3d 578, a defendant may pursue a comparative equitable indemnity claim against other tortfeasors either (1) by filing a cross-complaint in the original tort action or (2) by filing a separate indemnity action after paying more than its proportionate share of *1198the damages through the satisfaction of a judgment or through a payment in settlement. (See, e.g., Sears, Roebuck & Co. v. International Harvester Co. (1978) 82 Cal.App.3d 492, 496 [147 Cal.Rptr. 262]; American Bankers Ins. Co. v. Avco-Lycoming Division (1979) 97 Cal.App.3d 732, 736 [159 Cal.Rptr. 70].) In addition, more recent decisions also make clear that if one or more tortfeasors prove to be insolvent and are not able to bear their fair share of the loss, the shortfall created by such insolvency should be apportioned equitably among the remaining culpable parties—both defendants and plaintiffs. (See, e.g., Paradise Valley Hospital v. Schlossman (1983) 143 Cal.App.3d 87 [191 Cal.Rptr. 531]; Ambriz v. Kress (1983) 148 Cal.App.3d 963 [196 Cal.Rptr. 417].)
Although these various developments served to reduce much of the harshness of the original all-or-nothing common law rules, the retention of the common law joint and several liablity doctrine produced some situations in which defendants who bore only a small share of fault for an accident could be left with the obligation to pay all or a large share of the plaintiff’s damages if other more culpable tortfeasors were insolvent.
The initiative measure in question in this case was addressed to this remaining issue. While recognizing the potential inequity in a rule which would require an injured plaintiff who may have sustained considerable medical expenses and other damages as a result of an accident to bear the full brunt of the loss if one of a number of tortfeasors should prove insolvent, the drafters of the initiative at the same time concluded that it was unfair in such a situation to require a tortfeasor who might only be minimally culpable to bear all of the plaintiff’s damages. As a result, the drafters crafted a compromise solution: Proposition 51 retains the traditional joint and several liability doctrine with respect to a plaintiff’s economic damages, but adopts a rule of several liability for noneconomic damages, providing that each defendant is liable for only that portion of the plaintiff’s noneconomic damages which is commensurate with that defendant’s degree of fault for the injury.4 It was this compromise measure—which drew heavily *1199upon a number of bills which had been passed by the Senate but not by the Assembly in a number of preceding legislative sessions (see Sen. Bill No. 75 (1985-1986 Reg. Sess.); Sen. Bill No. 575 (1983-1984 Reg. Sess.); Sen. Bill No. 500 (1981-1982 Reg. Sess.))—that was adopted by the electorate in the June 1986 election.
Although Proposition 51 is the first legislative modification of the joint and several liability doctrine to be enacted in California, in recent years analogous statutory alterations of the traditional common law joint and several liablity rule have been adopted by many states throughout the country, often as part of a comprehensive legislative implementation of comparative fault principles. The revisions of the joint and several liability doctrine in other jurisdictions have taken a variety of forms: several states have abolished joint and several liability entirely and replaced it with a “pure” several liability rule,5 other states have formulated various guidelines to distinguish between more culpable and less culpable tortfeasors and have adopted several liability only for the less culpable tortfeasors,6 and still others, like California, have distinguished between different categories of damages sustained in an injury, retaining some form of joint and several liability for “economic” or “medically related” damages, while adopting some form of several liability for “pain and suffering” and other noneconomic damages.7 Thus, while Proposition 51 unquestionably made a *1200substantial change in this state’s traditional tort doctrine, when viewed from a national perspective it becomes apparent that the measure’s modification of the common law joint and several liability rule was not an isolated or aberrant phenomenon but rather paralleled similar developments in the evolution and implementation of the comparative-fault principle in other states.
Having briefly reviewed the historical background of Proposition 51, we turn initially to plaintiff’s broad claim that the Court of Appeal erred in failing to strike down the initiative measure as unconstitutional on its face.
III.
Plaintiff contends that Proposition 51 is facially unconstitutional on two separate grounds, asserting (1) that the measure is “too vague and ambiguous” to satisfy the due process requirements of either the state or federal Constitutions, and (2) that the enactment violates both the state and federal equal protection clauses by establishing classifications that are not rationally related to a legitimate state interest. As we shall see, both of these constitutional claims are similar to contentions raised just a few years ago in a series of cases challenging the validity of a variety of provisions of another legislative tort reform measure, the Medical Injury Compensation Reform Act of 1975 (MICRA) (Stats. 1975, 2d Ex. Sess. 1975-1976, chs. 1, 2, pp. 3949-4007), an enactment which modified a number of common law tort doctrines in the medical malpractice area. Our decisions in the earlier MI-CRA cases clearly establish that plaintiff’s current constitutional challenges lack merit.
A.
Plaintiff initially contends that Proposition 51 is unconstitutionally vague. Relying on the United States Supreme Court’s classic statement of the vagueness doctrine in Connally v. General Const. Co. (1926) 269 U.S. 385, 391 [70 L.Ed. 322, 328, 46 S.Ct. 126]—“a statute which either forbids or requires the doing of an act in terms so vague that men of common intelligence must necessarily guess at its meaning and differ as to its application, violates the first essential of due process of law”—plaintiff maintains that Proposition 51 is subject to just such a criticism. To support his *1201contention, plaintiff catalogues a series of questions relating to the application of Proposition 51 to which he suggests the language of the measure provides no clear answer.8 He asserts that the existence of these numerous unanswered questions renders the measure unconstitutionally vague on its face and warrants the invalidation of the enactment in its entirety.
Plaintiff’s contention is plainly flawed. Many, probably most, statutes are ambiguous in some respects and instances invariably arise under which the application of statutory language may be unclear. So long as a statute does not threaten to infringe on the exercise of First Amendment or other constitutional rights, however, such ambiguities, even if numerous, do not justify the invalidation of a statute on its face. In order to succeed on a facial vagueness challenge to a legislative measure that does not threaten constitutionally protected conduct—like the initiative measure at issue here—a party must do more than identify some instances in which the application of the statute may be uncertain or ambiguous; he must demonstrate that “the law is impermissibly vague in all of its applications.” (Italics added.) (Hoffman Estates v. Flipside, Hoffman Estates (1982) 455 U.S. 489, 497 [71 L.Ed.2d 362, 371, 102 S.Ct. 1186].) Plaintiff clearly has not satisfied this burden.
Plaintiff’s vagueness claim echoes a similar constitutional argument that was raised in American Bank & Trust Co. v. Community Hospital (1984) 36 Cal.3d 359, 377-378 [204 Cal.Rptr. 671, 683 P.2d 670, 41 A.L.R.4th 233], with respect to section 667.7 of the Code of Civil Procedure, a section of MICRA which provided for the periodic payment of judgments in medical malpractice cases under certain circumstances. In American Bank, plaintiff claimed, inter alia, that the statutory provision mandating periodic payment “should ... be struck down as unconstitutionally ‘void for vagueness, ambiguity and unworkability,’ because it leaves unanswered many questions as to how a trial court is to actually formulate a comprehensive payment schedule without the benefit of very detailed special jury verdicts.” (36 Cal.3d at p. 377.) After noting that the practical problems of application *1202were by no means insurmountable, we went on to point out that “[i]n any event, plaintiff provides no authority to support its claim that the remaining uncertainties which may inhere in the statute provide a proper basis for. striking it down on its face. As with other innovative procedures and doctrines—for example, comparative negligence—in the first instance trial courts will deal with novel problems that arise in time-honored case-by-case fashion, and appellate courts will remain available to aid in the familiar common law task of filling in the gaps in the statutory scheme. [Citation.]” (Id. at p. 378.)
Precisely the same reasoning applies in this case, Although the language of Proposition 51 may not provide a certain answer for every possible situation in which the modified joint and several liability doctrine may come into play, the application of the statute in many instances will be quite clear. Thus, for example, while plaintiff cites the statute’s lack of clarity on the retroactivity issue, there is no question but that the statute applies to causes of action accruing after its effective date; similarly, although plaintiff complains that the statute is not clear as to whether it applies to causes of action based on intentional tortious conduct or how it should be applied with respect to cases involving absent tortfeasors, the statute’s application in an ordinary multiple tortfeasor comparative negligence action in which all tortfeasors are joined is not in doubt. Further, as stated in American Bank, supra, 36 Cal.3d 359, when situations in which the statutory language is ambiguous arise, the statute’s application can be resolved by trial and appellate courts “in time-honored, case-by-case fashion,” by reference to the language and purposes of the statutory scheme as a whole. The judiciary’s traditional role of interpreting ambiguous statutory language or “filling in the gaps” of statutory schemes is, of course, as applicable to initiative measures as it is to measures adopted by the Legislature. (See, e.g., Amador Valley Joint Union High Sch. Dist. v. State Bd. of Equalization (1978) 22 Cal.3d 208, 244-246 [149 Cal.Rptr. 239, 583 P.2d 1281].) Accordingly, there is no merit to plaintiff’s claim that the statute should be struck down as unconstitutionally vague on its face.
B.
Plaintiff alternatively contends that Proposition 51 violates the state and federal equal protection guaranties, allegedly because the classifications drawn by the statute are not rationally related to a legitimate state interest.9 Plaintiff claims in particular that the statute is *1203invalid under the equal protection clause (1) because it discriminates between the class of injured persons who suffer economic damage and the class of injured persons who suffer noneconomic damage, providing full protection for those who suffer economic damage but a lesser protection for those who suffer noneconomic damage, and (2) because it improperly discriminates within the class of victims who suffer noneconomic damage, permitting full recovery for victims who are injured by solvent tortfeasors, but providing only partial recovery to victims injured by insolvent tortfeasors. Both claims are clearly without merit.
Plaintiff’s challenge to the proposition’s disparate treatment of economic and noneconomic damages parallels a similar equal protection attack that was directed at Civil Code section 3333.2, a provision of MICRA which placed a $250,000 limit on the noneconomic damages which may be recovered in a medical malpractice action, but which placed no similar limit on economic damages. In rejecting that equal protection challenge in Fein v. Permanente Medical Group, supra, 38 Cal.3d 137, we explained that there is clearly a rational basis for distinguishing between economic and noneconomic damages and providing fuller protection for economic losses,10 and observed that “[t]he equal protection clause certainly does not require the Legislature to limit a victim’s recovery for out-of-pocket medical expenses or lost earnings simply because it has found it appropriate to place some limit on damages for pain and suffering and similar noneconomic losses.” (38 Cal.3d at p. 162.) In similar fashion, the equal protection clause clearly does not require a state to modify the traditional joint and several liability rule as it applies to economic damages, simply because the state has found it appropriate to limit an individual tortfeasor’s potential liability for an injured person’s noneconomic damages. Indeed, the distinction which Proposition 51 draws between economic and noneconomic damages is, in general terms, less severe than the statutory distinction upheld in Fein-, Proposition 51 places no dollar limit on the noneconomic damages a plaintiff may properly recover, but simply provides that each individual tortfeasor will be liable only for that share of the plaintiff’s noneconomic damages which is *1204commensurate with the tortfeasor’s comparative fault. There is no constitutional impediment to such differential treatment of economic and noneconomic losses.
Nor is Proposition 51 vulnerable to constitutional attack on the basis of plaintiff’s claim that it improperly discriminates within the class of plaintiffs who have suffered noneconomic harm. Plaintiff asserts that the statute draws an arbitrary distinction between persons with noneconomic damages who have been injured by solvent tortfeasors and those who have been injured by insolvent defendants, permitting full recovery of noneconomic damages by the former class but only partial recovery by the latter class. The terms of the proposition itself, however, reflect no legislative intent to discriminate between injured victims on the basis of the solvency of the tortfeasors by whom they are injured; instead, the measure quite clearly is simply intended to limit the potential liability of an individual defendant for noneconomic damages to a proportion commensurate with that defendant’s personal share of fault.
Although one consequence of the statute’s adoption of several liability for noneconomic damages will be that persons who are unfortunate enough to be injured by an insolvent tortfeasor will not be able to obtain full recovery for their noneconomic losses, that consequence does not render the provision unconstitutional. Under any tort liability scheme, a plaintiff who is injured by a single tortfeasor who proves to be insolvent is, of course, worse off than a plaintiff who is injured by a single tortfeasor who can pay an adverse judgment. Such “differential treatment” flowing from the relative solvency of the tortfeasor who causes an injury, however, has never been thought to render all tort statutes unconstitutional or to require the state to compensate plaintiffs for uncollectible judgments obtained against insolvent defendants. And while the common law joint and several liability doctrine has in the past provided plaintiffs a measure of protection from the insolvency of a tortfeasor when there are additional tortfeasors who are financially able to bear the total damages, plaintiff has cited no case which suggests that the joint and several liability doctrine is a constitutionally mandated rule of law, immune from legislative modification or revision. As with other common law tort doctrines—like the doctrines at issue in the recent line of MICRA decisions (see, e.g., American Bank & Trust Co. v. Community Hospital, supra, 36 Cal.3d 359, 366-374 [modification of common law doctrine providing for payment of judgment in lump sum]; Barme v. Wood (1984) 37 Cal.3d 174 [207 Cal.Rptr. 816, 689 P.2d 446] [modification of collateral source rule]; Fein v. Permanente Medical Group, supra, 38 Cal.3d 137 [limitation of noneconomic damages])—the allocation of tort damages among multiple tortfeasors is an entirely appropriate subject for legislative resolution. In this regard, it is worth recalling that Propo*1205sition 51 does not require the injured plaintiff to bear the entire risk of a potential tortfeasor’s insolvency; solvent defendants continue to share fully in such risk with respect to a plaintiff’s economic damages.
In sum, although reasonable persons may disagree as to the wisdom of Proposition 51’s modification of the common law joint and several liability doctrine, the measure is not unconstitutional on its face.
IV.
Plaintiff’s second major contention is that even if the lower courts were correct in upholding the constitutionality of the proposition, the trial court and Court of Appeal were nonetheless in error in concluding that the newly enacted statute should apply retroactively to causes of action—like the present action—which accrued prior to the effective date of the initiative measure. Plaintiff points out that prior to the enactment of Proposition 51 many individuals—both plaintiffs and defendants—relied on the then-existing joint and several liability doctrine in deciding which parties to join in litigation and whether to accept or reject settlement offers relating to such preexisting claims, and plaintiff contends that because there is nothing in the terms of the proposition which indicates that it is to apply retroactively to defeat such reliance, the lower courts erred in giving it such an application. In response, defendants contend that retroactive application is warranted in light of the nature and purposes of the initiative measure.
A.
Before analyzing the retroactivity principles and precedents discussed by both parties, we must address a threshold contention, raised by a number of amici, who assert that there is no need to consider the retroactivity issue at all in this case. Although defendants themselves do not suggest that application of Proposition 51 to causes of action which accrued prior to its effective date but which did not come to trial until after such effective date would constitute only a prospective, rather than a retroactive, application of the measure, several amici have put forth that suggestion, arguing that by confining the measure’s operation to trials conducted after the initiative’s effective date the Court of Appeal simply applied Proposition 51 prospectively. The Court of Appeal did not rest its conclusion on this theory and, as we explain, the governing cases do not support amici’s contention.
In Aetna Cas. & Surety Co. v. Ind. Acc. Com. (1947) 30 Cal.2d 388 [182 P.2d 159]—perhaps the leading modem California decision on the subject—the same argument was raised by injured parties who contended that a new statute, increasing workers’ compensation benefits, should be applied *1206to awards made by the workers’ compensation board after the effective date of the new statute, even though the awards pertained to injuries which the workers had suffered before the new legislation was enacted. The injured employees argued that such an application of the statute to future awards would constitute a prospective, rather than a retroactive, application of the statute.
In Aetna Cas., this court, speaking through Chief Justice Gibson, emphatically rejected the argument, explaining that “ ‘[a] retrospective law is one which affects rights, obligations, acts, transactions and conditions which are performed or exist prior to the adoption of the statute.’ ” (30 Cal.2d at p. 391.) “Since the industrial injury is the basis for any compensation award, the law in force at the time of the injury is to be taken as the measure of the injured person’s right of recovery.” (Id. at p. 392.) Decisions of both the United States Supreme Court and the courts of our sister states confirm that the application of a tort reform statute to a cause of action which arose prior to the effective date of the statute but which is tried after the statute’s effective date would constitute a retroactive application of the statute. (See, e.g., Winfree v. Nor. Pac. Ry. Co. (1913) 227 U.S. 296 [57 L.Ed. 518, 33 S.Ct. 273]; Joseph v. Lowery (1972) 261 Or. 545 [495 P.2d 273].) Accordingly, amici’s argument that the legal principles relating to the retroactive application of statutes are not relevant in this case is clearly without merit.
B.
The fact that application of Proposition 51 to the instant case would constitute a retroactive rather than a prospective application of the statute is, of course, just the beginning, rather than the conclusion, of our analysis. Although plaintiff maintains that a retroactive application of the statute would be unconstitutional (cf. In re Marriage of Buol (1985) 39 Cal.3d 751, 759-764 [218 Cal.Rptr. 31, 705 P.2d 354]), defendants properly observe that in numerous situations courts have upheld legislation which modified legal rules applicable to pending actions. (See, e.g., San Bernardino County v. Indus. Acc. Com. (1933) 217 Cal. 618, 627-629 [20 P.2d 673].) Because the question whether a statute is to apply retroactively or prospectively is, in the first instance, a policy question for the legislative body which enacts the statute, before reaching any constitutional question we must determine whether, as a matter of statutory interpretation, Proposition 51 should properly be construed as prospective or retroactive. If, as a matter of statutory interpretation, the provision is prospective, no constitutional question is presented.
In resolving the statutory interpretation question, we are guided by familiar legal principles. In the recent decision of United States v. Security *1207Industrial Bank (1982) 459 U.S. 70, 79-80 [74 L.Ed.2d 235, 243-244, 103 S.Ct. 407], Justice (now Chief Justice) Rehnquist succinctly captured the well-established legal precepts governing the interpretation of a statute to determine whether it applies retroactively or prospectively, explaining: “The principle that statutes operate only prospectively, while judicial decisions operate retrospectively, is familiar to every law student. [Citations. ] This court has often pointed out-. ‘[7]/ze first rule of construction is that legislation must be considered as addressed to the future, not to the past. . . . The rule has been expressed in varying degrees of strength but always of one import, that a retrospective operation will not be given to a statute which interferes with antecedent rights . . . unless such be “the unequivocal and inflexible import of the terms, and the manifest intention of the legislature.” ’ [Citation.]” (Italics added.)
California authorities have long embraced this general principle. As Chief Justice Gibson wrote for the court in Aetna Cas. & Surety Co. v. Ind. Acc. Com., supra, 30 Cal.2d 388—the seminal retroactivity decision noted above—“[i]t is an established canon of interpretation that statutes are not to be given a retrospective operation unless it is clearly made to appear that such was the legislative intent.” (30 Cal.2d at p. 393.) This rule has been repeated and followed in innumerable decisions. (See, e.g., White v. Western Title Ins. Co. (1985) 40 Cal.3d 870, 884 [221 Cal.Rptr. 509, 710 P.2d 309]; Glavinich v. Commonwealth Land Title Ins. Co. (1984) 163 Cal.App.3d 263, 272 [209 Cal.Rptr. 266]. See generally 5 Witkin, Summary of Cal. Law (8th ed. 1974) Constitutional Law, § 288, pp. 3578-3579.)
Indeed, Civil Code section 3, one of the general statutory provisions governing the interpretation of all the provisions of the Civil Code—including the provision at issue in this case—represents a specific legislative codification of this general legal principle, declaring that “[ri\o part of [this Codé\ is retroactive, unless expressly so declared.” (Italics added.)11 Like similar provisions found in many other codes (see, e.g., Code Civ. Proc., *1208§ 3; Lab. Code, § 4), section 3 reflects the common understanding that legislative provisions are presumed to operate prospectively, and that they should be so interpreted “unless express language or clear and unavoidable implication negatives the presumption.” (Glavinich v. Commonwealth Land Title Ins. Co., supra, 163 Cal.App.3d 263, 272.)
The dissenting opinion—relying on passages in a few decisions of this court to the effect that the presumption of prospectivity is to be “subordinated ... to the transcendent canon of statutory construction that the design of the Legislature be given effect . . . [and] is to be applied only after, considering all pertinent factors, it is determined that it is impossible to ascertain the legislative intent” (Marriage of Bouquet, supra, 16 Cal.3d 583, 587 [italics deleted]; Mannheim v. Superior Court (1970) 3 Cal.3d 678, 686-687 [91 Cal.Rptr. 585, 478 P.2d 17]; In re Estrada, supra, 63 Cal.2d 740, 746)—apparently takes the position that the well-established legal principle which Justice Rehnquist suggested was “familiar to every law student” (see United States v. Security Industrial Bank, supra, 459 U.S. 70, 79 [74 L.Ed.2d 235, 243]) is inapplicable in this state and that Civil Code section 3 and other similar statutory provisions have virtually no effect on a court’s determination of whether a statute applies prospectively or retroactively. The language in the decisions relied on by the dissent, however, generally has not been, and should not properly be, interpreted to mean that California has embraced a unique application of the general prospectivity principle, distinct from the approach followed in other jurisdictions (see generally 2 Sutherland on Statutory Construction (4th ed. 1986) § 41.04, pp. 348-350), so that the principle that statutes are presumed to operate prospectively ordinarily has no bearing on a court’s analysis of the retroactivity question and may properly be considered by a court only as a matter of last resort and then only as a tie-breaking factor.
In the years since Estrada, supra, 63 Cal.2d 740, Mannheim, supra, 3 Cal.3d 678, and Marriage of Bouquet, supra, 16 Cal.3d 583, both this court and the Courts of Appeal have generally commenced analysis of the question of whether a statute applies retroactively with a restatement of the fundamental principle that “legislative enactments are generally presumed to operate prospectively and not retroactively unless the Legislature expresses a different intention.” (See, e.g., Fox v. Alexis (1985) 38 Cal.3d 621, 637 [214 Cal.Rptr. 132, 699 P.2d 309]; White v. Western Title Co., supra, 40 Cal.3d 870, 884; Hoffman v. Board of Retirement (1986) 42 Cal.3d 590, 593 [229 Cal.Rptr. 825, 724 P.2d 511]; Baker v. Sudo (1987) 194 Cal.App.3d 936, 943 [240 Cal.Rptr. 38]; Sagadin v. Ripper (1985) 175 Cal.App.3d 1141, 1156 [221 Cal.Rptr. 675]; Glavinich v. Commonwealth Land Title Ins. Co., supra, 163 Cal.App.3d 263, 272.) These numerous precedents demonstrate that California continues to adhere to the time-honored principle, codified *1209by the Legislature in Civil Code section 3 and similar provisions, that in the absence of an express retroactivity provision, a statute will not be applied retroactively unless it is very clear from extrinsic sources that the Legislature or the voters must have intended a retroactive application. The language in Estrada, Mannheim, and Marriage of Bouquet should not be interpreted as modifying this well-established, legislatively-mandated principle.
Applying this general principle in the present matter, we find nothing in the language of Proposition 51 which expressly indicates that the statute is to apply retroactively.12 Although each party in this case attempts to stretch the language of isolated portions of the statute to support the position each favors,13 we believe that a fair reading of the proposition as a whole makes it clear that the subject of retroactivity or prospectivity was simply not addressed. As we have explained, under Civil Code section 3 and the general principle of prospectivity, the absence of any express provision directing retroactive application strongly supports prospective operation of the measure. Although defendants raise a number of claims in an attempt to escape the force of this well-established principle of statutory interpretation, none of their contentions is persuasive.
C.
Defendants initially contend that even though there is no express language in the statute calling for retroactive application, an intent that the provision should apply retroactively can clearly be inferred from the objectives of the legislation, as reflected in the stated “findings and declaration of purpose” accompanying the provision and in the ballot arguments which *1210were before the voters at the time the measure was adopted.14 As defendants correctly point out, on a number of occasions in the past we have found that even when a statute did not contain an express provision mandating retroactive application, the legislative history or the context of the enactment provided a sufficiently clear indication that the Legislature intended the statute to operate retrospectively that we found it appropriate to accord the statute a retroactive application. (See, e.g., Marriage of Bouquet, supra, 16 Cal.3d 583; Mannheim, supra, 3 Cal.3d 678, 686.)15
Defendants assert that consideration of the factors deemed relevant to the inquiry into legislative intent in those cases—e.g., “ ‘[the] context [of the legislative enactment], the object in view, the evils to be remedied, the history of the times and of legislation upon the same subject’ ” (Marriage of *1211Bouquet, supra, 16 Cal.3d 583, 587)—supports retroactive application of the legislation at issue here. As we shall explain, we cannot agree.
To begin with, unlike Marriage of Bouquet or Mannheim, there is nothing in either the statutory “findings and declaration of purpose” or the brochure materials which suggests that, notwithstanding the absence of any express provision on retroactivity, the retroactivity question was actually consciously considered during the enactment process. In Marriage of Bouquet, the court, in concluding that the statute at issue in that case should be applied retroactively, relied, in part, on the Legislature’s adoption of a resolution, shortly after the enactment of the measure, indicating that the retroactivity question was specifically discussed during the legislative debate on the measure and declaring that the provision was intended to apply retroactively (see Marriage of Bouquet, supra, 16 Cal.3d at pp. 588-591); in Mannheim, the statute in question incorporated by reference a separate statutory scheme which had expressly been made retroactive, and the Mannheim court reasoned that the Legislature must have intended the later statute to have a parallel application to the provision on which it was expressly fashioned. (See Mannheim, supra, 3 Cal.3d at pp. 686-687.) Defendants can point to nothing in the election brochure materials which provide any comparable confirmation of an actual intention on the part of the drafters or electorate to apply the statute retroactively.
Indeed, when “ ‘the history of the times and of legislation upon the same subject’ ” (Marriage of Bouquet, supra, 16 Cal.3d at p. 587) is considered, it appears rather clear that the drafters of Proposition 51, in omitting any provision with regard to retroactivity, must have recognized that the statute would not be applied retroactively. As we have noted briefly above, the tort reform measure instituted by Proposition 51 paralleled somewhat similar tort reform legislation—MICRA—which was enacted in the mid-1970’s in response to a liability insurance crisis in the medical malpractice field. In Bolen v. Woo (1979) 96 Cal.App.3d 944, 958-959 [158 Cal.Rptr. 454] and Robinson v. Pediatric Affiliates Medical Group, Inc. (1979) 98 Cal.App.3d 907, 911-912 [159 Cal.Rptr. 791], two separate panels of the Court of Appeal addressed the question whether one of the tort reform provisions of MICRA should apply retroactively to a cause of action that accrued prior to MICRA’s enactment but which was tried after the act went into effect. In both Bolen and Robinson, the courts held that in the absence of a specific provision in the legislation calling for such retroactive application, the general presumption of prospective application should apply; the Bolen court observed that if the Legislature had intended the statute to apply retroactively it “could very easily have inserted such language in the statute itself. It chose not to do so.” (96 Cal.App.3d at p. 959.) Because at least one of the principal institutional proponents and drafters of Proposition 51 was very *1212much involved in the post-MICRA litigation,16 it appears inescapable that—given the Bolen and Robinson decisions—the drafters of Proposition 51 would have included a specific provision providing for retroactive application of the initiative measure if such retroactive application had been intended. (Cf. Aetna Cas. & Surety Co., supra, 30 Cal.2d 388, 396 [“it must be assumed that the Legislature was acquainted with the settled rules of statutory interpretation, and that it would have expressly provided for retrospective operation of the amendment if it had so intended.”].) Since the drafters declined to insert such a provision in the proposition—perhaps in order to avoid the adverse political consequences that might have flowed from the inclusion of such a provision—it would appear improper for this court to read a retroactivity clause into the enactment at this juncture.
D.
Defendants contend, however, that whether or not the drafters of the proposition intended that the measure would apply retroactively, it is the intent of the electorate that is controlling, and they maintain that, in light of the purposes of the proposition, it is evident that the voters must have intended a retroactive application.
This argument, while novel, is flawed in a number of fundamental respects. To begin with, although the intent of the electorate would prevail over the intent of the drafters if there were a reliable basis for determining that the two were in conflict, in the present case there is simply no basis for finding any such conflict. Neither the Legislative Analyst’s analysis of Proposition 51 nor any of the statements of the proponents or opponents that were before the voters in the ballot pamphlet spoke to the retroactivity question, and thus there is no reason to believe that the electorate harbored any specific thoughts or intent with respect to the retroactivity issue at all. Because past cases have long made it clear that initiative measures are subject to the ordinary rules and canons of statutory construction (see, e.g., Carter v. Seaboard Finance Co. (1949) 33 Cal.2d 564, 579-582 [203 P.2d 758]; Amador Valley Joint Union High Sch. Dist. v. State Bd. of Equalization, supra, 22 Cal.3d 208, 244-246), informed members of the electorate who happened to consider the retroactivity issue would presumably have concluded that the measure—like other statutes—would be *1213applied prospectively because no express provision for retroactive application was included in the proposition.
Furthermore, defendants’ claim that the “remedial” purpose of the measure necessarily demonstrates that the electorate must have intended that the proposition apply retroactively cannot be sustained. Although the “findings and declaration of purpose” included in the proposition clearly indicate that the measure was proposed to remedy the perceived inequities resulting under the preexisting joint and several liablity doctrine and to create what the proponents considered a fairer system under which “defendants in tort actions shall be held financially liable in closer proportion to their degree of fault” (Civ. Code, § 1431.1), such a remedial purpose does not necessarily indicate an intent to apply the statute retroactively. Most statutory changes are, of course, intended to improve a preexisting situation and to bring about a fairer state of affairs, and if such an objective were itself sufficient to demonstrate a clear legislative intent to apply a statute retroactively, almost all statutory provisions and initiative measures would apply retroactively rather than prospectively. In light of the general principles of statutory interpretation set out above, and particularly the provisions of Civil Code section 3, the contention is clearly flawed. (See, e.g. Aetna Cas. & Surety Co. v. Ind. Acc. Com., supra, 30 Cal.2d at p. 395.)17
What defendants’ contention overlooks is that there are special considerations—quite distinct from the merits of the substantive legal change embodied in the new legislation—that are frequently triggered by the *1214application of a new, “improved” legal principle retroactively to circumstances in which individuals may have already taken action in reasonable reliance on the previously existing state of the law. Thus, the fact that the electorate chose to adopt a new remedial rule for the future does not necessarily demonstrate an intent to apply the new rule retroactively to defeat the reasonable expectations of those who have changed their position in reliance on the old law. The presumption of prospectivity assures that reasonable reliance on current legal principles will not be defeated in the absence of a clear indication of a legislative intent to override such reliance.
The Oregon Supreme Court’s decision in Joseph v. Lowery, supra, 495 P.2d 273 illustrates the point quite well, in a context closely related to the instant case. The question at issue in Joseph was whether a newly enacted comparative-negligence statute should be applied retroactively to a cause of action which accrued before the passage of the statute but which did not come to trial until after the new law went into effect. The plaintiff in that case, like defendants in this case, argued forcefully that the court should infer from the remedial nature of the legislative change that the Legislature intended to apply the newly enacted, more equitable comparative negligence rule to all cases tried after the passage of the new legislation, even when the cause of action accrued prior to the enactment; the plaintiff emphasized, in this regard, that the defendant’s “primary conduct” at the time of the accident was obviously not undertaken in reliance on the contributory negligence doctrine.
The Oregon Supreme Court rejected the plaintiff’s argument for retroactive application of the statute, explaining: “Certainly, no one has an accident upon the faith of the then existing law. However, it would come as a shock to someone who has estimated his probable liability arising from a past accident, and who has planned his affairs accordingly, to find that his responsibility therefor is not to be determined as of the happening of the accident but is also dependent upon what the legislature might subsequently do. Every day it is necessary in the conduct of the affairs of individuals and of businesses to make a closely calculated estimate of the responsibility or lack thereof resulting from an accident or from other unforeseen and unplanned circumstances and to act in reliance on such estimate. We believe there is merit in the prior view of this court, as demonstrated by its decisions, that, in the absence of an indication to the contrary, legislative acts should not be construed in a manner which changes legal rights and responsibilities arising out of transactions which occur prior to the passage of such acts.” (495 P.2d at p. 276.) The vast majority of other courts—including the United States Supreme Court—which have faced the question whether a remedial statute replacing the all-or-nothing contributory negligence doc*1215trine with a more equitable comparative negligence rule should be applied retroactively to causes of action which accrued prior to the date of the comparative negligence statute, when the enactment is silent on the retroactivity issue, have reached the same conclusion as the Joseph court, applying the new remedial statute prospectively only.18
Although, as we have noted, there is no indication that the voters in approving Proposition 51 consciously considered the retroactivity question at all, if they had considered the issue they might have recognized that retroactive application of the measure could result in placing individuals who had acted in reliance on the old law in a worse position than litigants under the new law. We briefly examine why retroactive application of the proposition could have such a consequence.
To begin with, plaintiffs whose causes of action arose long before Proposition 51 was enacted will often have reasonably relied on the preexisting joint and several liability doctrine in deciding which potential tortfeasors to sue and which not to sue. Given the joint and several liability rule, plaintiffs may reasonably have determined that while there may have been other tortfeasors—in addition to the defendants named in their complaint—who might also be responsible for their injuries, there was no reason to go to the added expense and effort to attempt to join such other tortfeasors, since plaintiffs could recover all of their damages—economic and noneconomic— from the named defendants. Such plaintiffs would have understood, of course, that under the then-governing rules, the named defendants could bring any such additional tortfeasors into the suit through cross-complaints if the defendants desired.
While Proposition 51 itself, of course, does not bar a plaintiff from joining additional tortfeasors—indeed, its effect in the future well may be to encourage plaintiffs to join every conceivable responsible party—the *1216retroactive application of the measure to preexisting causes of action would frequently have the effect of depriving plaintiffs of any opportunity to recover the proportion of noneconomic damages attributable to absent tortfeasors, because in many cases the statute of limitations on the plaintiff’s preexisting cause of action against such an absent tortfeasor will have run before the enactment of Proposition 51.19 Thus, while there is nothing in the language or legislative history of Proposition 51 to suggest that the electorate intended to cut off a plaintiff’s opportunity to obtain full recovery for noneconomic damages, the retroactive application of the measure would frequently have just such an effect.
In similar fashion, retroactive application of the proposition to actions which were pending prior to the adoption of the measure would frequently defeat the reasonable expectations of parties who entered into settlement agreements in reliance on the preexisting joint and several liability rule. Acting on the assumption that any nonsettling defendants would remain fully liable for both economic and noneconomic damages, plaintiffs in preProposition 51 actions may frequently have settled with some defendants for a lesser sum than they would have accepted if they were aware that the remaining defendants would only be severally liable for noneconomic damages. By contrast, plaintiffs who settle causes of action accruing after Proposition 51 would be fully aware of the applicable principles.
Furthermore, retroactive application of Proposition 51 could also have unanticipated, adverse consequences for settling defendants as well. As noted above, under pre-Proposition 51 law, a defendant could choose to enter into a settlement agreement with the plaintiff which settled the plaintiff’s entire claim against all defendants, and could thereafter bring an equitable comparative indemnity action against other tortfeasors to compel them to bear their fair share of the amount which the settling defendant had paid in settlement of the plaintiff’s claim. (See, e.g., Sears, Roebuck & Co. v. International Harvester Co., supra, 82 Cal.App.3d 492, 496; American Bankers Ins. Co. v. Avco-Lycoming Division, supra, 97 Cal.App.3d 732, 736.) Under preexisting law, if a settling defendant pursued such a course of action and if one or more of the culpable tortfeasors proved to be insolvent, the shortfall caused by such insolvency would be shared on an equitable basis by all of the solvent tortfeasors. (See, e.g., Paradise Valley Hospital v. Schlossman, supra, 143 Cal.App.3d 87, 93.) If Proposition 51 were applied *1217retroactively to causes of action that accrued prior to its enactment, however, a nonsettling tortfeasor who was faced with an indemnity claim brought by a settling tortfeasor would be able to limit his liability for noneconomic damages to a percentage equal to his own personal degree of fault, and the settling tortfeasor—who had entered into the settlement in reliance on the preexisting state of the law—would be left to absorb by himself any proportion of the noneconomic damages that was attributable to an insolvent tortfeasor or tortfeasors.
Thus, retroactive application of the measure to past litigation could have unexpected and potentially unfair consequences for all parties who acted in reliance on the then-existing state of the law. Prospective application of the measure, while withholding the remedial benefits of the provision from defendants in pending actions, would assure that all parties to litigation were aware of the basic “ground rules” when they decided whom to join in the action and on what terms the case should be settled.
Of course, we do not suggest that most or even many voters were aware of the consequences that would result from the retroactive application of Proposition 51. A review of these consequences does indicate, however, that a voter who supported the remedial changes embodied in Proposition 51 would not necessarily have supported the retroactive application of those changes to defeat the reasonable expectations of individuals who had taken irreversible actions in reliance on the preexisting state of the law.
To avoid misunderstanding, a caveat is in order. It is no doubt possible that an informed electorate, aware of the consequences of retroactive application, would nonetheless have chosen to make the statute retroactive if the retroactivity or prospectivity issue had been directly presented to it. The crucial point is simply that because Proposition 51 did not address the retroactivity question, we have no reliable basis for determining how the electorate would have chosen to resolve either the broad threshold issue of whether the measure should be applied prospectively or retroactively, or the further policy question of how retroactively the proposition should apply if it was to apply retroactively: i.e., whether the new rule should apply to cases in which a complaint had not yet been filed, to cases which had not yet come to trial, to cases in which a trial court judgment had not yet been entered, or to cases which were not yet final on appeal.20
*1218As we have explained above, the well-established presumption that statutes apply prospectively in the absence of a clearly expressed contrary intent gives recognition to the fact that retroactive application of a statute often entails the kind of unanticipated consequences we have discussed, and ensures that courts do not assume that the Legislature or the electorate intended such consequences unless such intent clearly appears. Because in the present matter there is nothing to suggest that the electorate considered these results or intended to depart from the general rule that statutory changes operate prospectively, prospective application is required.21
E.
Defendants next argue that even if the remedial nature of Proposition 51 is not sufficient to indicate an intent on the part of the electorate to apply the measure retroactively, this court should infer such an intent from the fact that the measure’s statement of purpose and the election brochure arguments demonstrate that the proposition was adopted to meet a liability insurance crisis. Defendants maintain that because it will be years before causes of action which accrue after the effective date of the proposition actually come to trial, a prospective application of the measure would not effectuate the purpose of alleviating the insurance crisis and thus could not have been intended by the electorate. For a number of reasons, we conclude that this argument cannot be sustained.
To begin with, defendants’ account of the consequences of prospective application of the measure is inaccurate in a number of significant respects. First, because liability insurance premiums are based in part, if not exclusively, on the damages that the insurance company anticipates it will incur for the risks which will be covered by the policy, any anticipated reduction in damages to be awarded in the future for causes of action which arise *1219during policy periods following the act should logically be reflected in an immediate reduction in the premiums which potential defendants pay for post-act insurance coverage. Thus, prospective application of the proposition could reasonably have been expected to afford immediate benefits to potential defendants. Similarly, to the extent governmental or other activities had been curtailed because of the fear of the anticipated financial consequences of future accidents, the knowledge that any such future incidents would be governed by the provisions of Proposition 51 would logically support prompt resumption of the activities.
Moreover, because the insurance premiums which potential defendants had paid prior to the enactment of Proposition 51 for coverage of preProposition 51 accidents were presumably computed, at least in part, on the assumption that the then-prevailing joint and several liability doctrine would apply to the covered incidents, a retroactive application of the measure might be expected to provide a windfall to defendants’ insurers, rather than a direct benefit to the insureds themselves because the initiative contained no provision requiring insurers to return any portion of previously collected premiums to their insureds. Indeed, this potential consequence of retroactive application may have been one reason the drafters of the measure chose not to include an express retroactivity provision in the measure; if this potential insurance company windfall from retroactive application had been brought to the attention of the electorate, it might well have detracted from the popularity of the measure.
Finally, defendants’ suggestion that a prospective application of Proposition 51 will mean that it will be years before the measure will affect the actual damages paid by defendants in tort cases overlooks the fact that the vast majority of tort actions are resolved by settlement rather than by trial. Because the amounts at which cases are settled reflect the defendant’s potential liability at trial, the effects of Proposition 51 on damages actually paid by defendants are likely to be felt at a much earlier date than defendants predict even if the measure is applied prospectively.
Thus, we cannot agree that prospective application is inconsistent with the objective of alleviating a liability-insurance crisis.
Indeed, a review of other statutory provisions, similar to Proposition 51, which were enacted in other states at approximately the same time as Proposition 51 and in response to the same concerns over the effects of high liability insurance premiums,22 demonstrates that this factor does not neces*1220sarily evidence an intent to apply the statute retroactively to all cases tried after the effective date of the enactment. In the numerous statutes altering the joint and several liability rule which were enacted throughout the country in 1986 and 1987, the various state legislatures not only adopted different substantive variants of several liability (see fns. 5, 6, 7, ante), but also arrived at differing conclusions as to whether the newly enacted statutes should be applied retroactively to preexisting causes of action. Several of the new statutes were explicitly made applicable only to causes of action accruing after the date of the new legislation (Fla.Stat.Ann. § 768.71(2) (West Supp. 1987); Mo.Ann.Stat. § 538.235 (Vernon Supp. 1987); Ill.Ann.Stat, ch. 110, note following ¶¶ 2-1117, 2-1118 (Smith-Hurd Supp. 1987); 1987 Nev.Stat., ch. 709, § 2), some of the enactments apply only to cases filed on or after the effective date of the statute (1986 Colo.Sess. Laws, ch. 108, § 7; 1986 Wash. Laws, ch. 305, § 910; 1986 N.Y. Laws, ch. 682, § 12; 1987 Tex. Acts, 70th Leg., 1st C.S., ch. 2, § 4.05, in Tex.Civ.Prac. & Rem. Code Ann., note following § 9.001 (Vernon 1988)), and only one of the statutes—which adopted a several liablity rule limited to less culpable governmental defendants—applies to cases “pending on or commenced on or after” the date of the enactment (1986 Minn. Laws, ch. 455, § 95). These varying responses, of course, are relevant to the question before us only inasmuch as they demonstrate that other legislative bodies which enacted statutes in response to the same liability crisis that precipitated Proposition 51 and which consciously focused on the retroactivity question arrived at different conclusions of whether, and to what extent, such a statutory modification should apply to preexisting causes of action. Because the provision before us is silent on the question, the general presumption which dictates a prospective application in the absence of a clear contrary intent must control.
The California decision most closely on point directly supports this conclusion. As noted above, in Bolen v. Woo, supra, 96 Cal.App.3d 944, 958-959, the Court of Appeal addressed the question whether one of the tort reform provisions of MICRA should apply retroactively to a cause of action that accrued prior to MICRA’s enactment but that was tried after the act went into effect. The defendant in Bolen, like defendants in this case, relied heavily on the fact that the preamble of MICRA demonstrated that the measure was adopted in response to a crisis caused by “skyrocketing” liability insurance costs23 and argued that that purpose established an intent *1221to apply the act retroactively. The Bolen court rejected the contention, relying on the general principle of prospectivity discussed above and emphasizing that if the Legislature had intended the statute to apply retroactively it “could very easily have inserted such language in the statute itself. It chose not to do so.” (96 Cal.App.3d at p. 959.)
In light of Bolen, if the proponents of Proposition 51 felt that the liability crisis necessitated a retroactive application of the measure’s provisions, it seems evident that they would have included an express retroactivity provision in the proposition.
F.
Defendants next argue that, despite the absence of any express retroactivity provision, Proposition 51 should be applied retroactively by analogy to this court’s retroactive application of the decisions in Li v. Yellow Cab, supra, 13 Cal.3d 804, and American Motorcycle Association v. Superior Court, supra, 20 Cal.3d 578, to at least some cases that were pending at the time those decisions were rendered. (See Li, supra, 13 Cal.3d 804, 829; Safeway Stores, Inc. v. Nest-Kart (1978) 21 Cal.3d 322, 333-334 [146 Cal.Rptr. 550, 579 P.2d 441].) For a number of reasons, those decisions do not support defendants’ claim.
First, both Li, supra, 13 Cal.3d 804, and American Motorcycle, supra, 20 Cal.3d 578, involved changes in common law tort doctrine that were made by judicial decision, not statutory enactment. As the earlier quotation from Chief Justice Rehnquist makes clear, as a general rule there is a fundamental difference between the retroactivity of statutes and the retroactivity of judicial decisions: “The principle that statutes operate only prospectively, while judicial decisions operate retrospectively, is familiar to every law student. [Citations.]” (United States v. Security Industrial Bank, supra, 459 U.S. 70, 79 [74 L.Ed.2d 235, 243].) It is because of this difference in the governing legal principles that in most states in which the comparative negligence rule has been adopted through judicial decision—like California—the newly adopted rule has been applied to at least some pending cases (see Schwartz, Comparative Negligence (2d ed. 1986) § 8.2, pp. 140-143), while in those states in which comparative negligence has been established by statute, the change has almost uniformly been applied prospectively. (See id., §§ 8.3, 8.4, pp. 143-149; see also fn. 17, ante.) Thus, the fact that the *1222judicial modifications of tort doctrines in Li and American Motorcycle were accorded some retroactive application provides no support for defendants’ claim that the subsequent legislative modification of a tort doctrine in Proposition 51 should apply retroactively.
Second, defendants’ argument overlooks a related, but somewhat more fundamental, point. Because in the Li, supra, 13 Cal.3d 804, and American Motorcycle, supra, 20 Cal.3d 578, cases it was the court which made the policy decision that the common law rules at issue in those cases should be changed, the court was the appropriate body to determine whether or not the new rule should be applied retroactively and, if so, how retroactively. (See generally Gt. Northern Ry. v. Sunburst Co. (1932) 287 U.S. 358 [77 L.Ed. 360, 53 S.Ct. 145, 85 A.L.R. 254]; Peterson v. Superior Court (1982) 31 Cal.3d 147, 151-153 [181 Cal.Rptr. 784, 642 P.2d 1305].) In the present case, by contrast, it was the electorate who made the policy decision to implement a change in the traditional common law rule, and thus it was the voters who possessed the authority to decide the policy question of whether the new statute should be applied retroactively. Unlike in Li or in American Motorcycle, in this case our court has no power to impose its own views as to the wisdom or appropriateness of applying Proposition 51 retroactively. Because, as we have discussed above, the proposition is silent on the retroactivity question, Civil Code section 3 and well-founded principles of statutory interpretation establish that the statute must be interpreted to apply prospectively.
G.
Finally, defendants contend that Proposition 51 should be applied retroactively by analogy to a line of California cases, beginning with Tulley v. Tranor (1878) 53 Cal. 274, which have applied a number of statutory amendments, which modified the legal measure of damages recoverable in an action for wrongful conversion of personal or real property, to all trials conducted after the effective date of the revised statute. (See also Feckenscher v. Gamble (1938) 12 Cal.2d 482 [85 P.2d 885]; Stout v. Turney (1978) 22 Cal.3d 718, 727 [150 Cal.Rptr. 637, 586 P.2d 1228].)24
*1223To begin with, we believe defendants clearly overstate the scope of the Tulley line of cases in suggesting that those decisions establish a broad rule that in California any statutory provision which affects the amount of damages which an injured person may recover is presumptively retroactive. As we have seen, the seminal decision in Aetna Cas. & Surety Co., supra, 30 Cal.2d 388—decided long after Tulley, supra, 53 Cal. 274—applied the general presumption of prospective application to a statutory provision which increased the damages or benefits recoverable in a workers’ compensation action. Similarly, the two relatively recent MICRA cases noted above (Bolen v. Woo, supra, 96 Cal.App.3d 944; Robinson v. Pediatrics Affiliates Medical Group, Inc., supra, 98 Cal.App.3d 907) applied the traditional principle of prospective application to a provision of MICRA which affected the damages which a plaintiff could recover in a medical malpractice action. (Civ. Code, § 3333.1 [modification of collateral source rule].) Indeed, in our even more recent decision in White v. Western Title Ins. Co., supra, 40 Cal.3d 870, 884, this court, after noting that “'“[i]t is a general rule of construction . . . that, unless the intention to make it retrospective clearly appears from the act itself, a statute will not be construed to have that effect” ’ [citations],” went on to observe that “[t]his rule is particularly applicable to a statute which diminishes or extinguishes an existing cause of action.” (Italics added.) (Ibid.) Thus, it is not accurate to suggest that the ordinary presumption of prospectivity is inapplicable to any statute which modifies damages; after all, Civil Code section 3, which codifies the common law presumption of prospectivity with respect to provisions of the Civil Code, contains no exception for statutes relating to damages.
Instead, Tulley, supra, 53 Cal. 274, and its progeny were primarily concerned with an entirely separate issue. In Aetna Cas. & Surety Co., supra, 30 Cal.2d 388, our court, in discussing Feckenscher v. Gamble, supra, 12 Cal.2d 482—one of the cases in the Tulley line—observed that in Feckenscher the court had found that the language of the statute in question showed that the Legislature intended the measure to be applied retroactively, and that “the court was concerned mainly with the question of whether the Legislature has power to give those laws such retroactive effect.” (30 Cal.2d at p. 393.) The Tulley decision, too—after finding that the statutory *1224language left “no reasonable doubt that the amendment was intended to be applicable to a case in which the conversion had occurred prior to its passage” (53 Cal. at p. 278)25—focused primarily on the question of whether the Legislature had the constitutional authority to apply a new measure of damages to causes of action which accrued prior to the enactment of the new statute but which came to trial after the enactment, concluding that the Legislature did have such authority. (See 53 Cal. at pp. 279-280.) Thus, while Tulley and its progeny do provide support for the claim that it is not necessarily unconstitutional for the Legislature to alter the measure of damages with respect to preexisting causes of action, those decisions do not purport to reject the ordinary presumption of prospectivity or to adopt a new legal standard for determining whether the Legislature intended a statute to be retroactive or prospective; the decisions simply found that the language of the statutes at issue in those cases demonstrated that the measures were intended to apply retroactively.
As we have noted above, of course, the question whether Proposition 51 may constitutionally be applied retroactively is quite distinct from the question whether the proposition should be properly interpreted as retroactive or prospective as a matter of statutory interpretation. The Aetna Cas. & Surety Co. decision makes it clear that the Tulley line of cases cannot properly be interpreted as displacing ordinary principles of statutory interpretation with regard to the question of retroactivity. (See Aetna Cas. & Surety Co., supra, 30 Cal.2d at pp. 393-394.) Other jurisdictions have also generally applied the traditional presumption of prospective application to statutes which modify the amount of damages recoverable in tort actions. (See generally Annot. (1964) 98 A.L.R.2d 1105; Annot. (1977) 80 A.L.R.3d 583, 601-602.)
In any event, Proposition 51 is quite unlike the statutory provisions at issue in Tulley, supra, 53 Cal. 274, or its progeny in a number of important respects. First of all, unlike the statutes in those cases, Proposition 51 does not purport to alter either the measure or the total amount of damages that a plaintiff may recover for a particular tort. Although Proposition 51 does affect the amount of noneconomic damages a particular tortfeasor may be required to pay when more than one tortfeasor is responsible for an injury, and may have the effect of reducing a plaintiff’s ultimate recovery if one or more tortfeasors are insolvent, nothing in the measure evidences a legisla*1225tive objective of denying a plaintiff the opportunity to obtain full recovery for both economic and noneconomic damages by joining all responsible tortfeasors and collecting the appropriate proportion of noneconomic damages from each tortfeasor. As we have discussed above, however, retroactive application of the measure would often have the effect of placing plaintiffs in pending actions in a worse position than plaintiffs in future actions, since plaintiffs in pending actions may no longer have the ability to join all potentially liable tortfeasors because of the statute of limitations. Thus, whereas application of the statutory provisions at issue in the Tulley line of cases to both pending and future actions at least accorded like treatment to current and future plaintiffs, retroactive application in this case would not have an equalizing effect, but would impose a unique detriment on one class of plaintiffs. Accordingly, it is more difficult to assume in this case, than it was in the Tulley cases, that retroactive application was intended.
Second, given the nature of the statutory revision at issue in the Tulley line of cases, it was unlikely that the parties in pending actions had taken any irreversible actions or changed their position in reliance on the preexisting measure of damages. By contrast, as discussed above, many plaintiffs and defendants in pending actions undoubtedly relied on the preexisting joint and several liability rule in conducting their litigation prior to enactment of Proposition 51. On this ground, too, there is more reason in this case than in the Tulley decisions to question whether a retroactive application of the statute was intended.
Finally, it is impossible to ignore that the statutory change at issue here, modifying a long-standing common law doctrine applicable to all negligence actions, represents a much more substantial and signficant change in the law than the narrow statutory modifications at issue in the Tulley cases. Because of the widespread impact of retroactive application of Proposition 51, the need for an express statement of legislative intent becomes all the more essential.
Accordingly, the Tulley line of cases does not support the retroactive application of Proposition 51.26
*1226H.
Having reviewed defendants’ numerous arguments, we think it may be useful, in conclusion, to take a last look at one particularly instructive precedent. In Winfree v. Nor. Pac. Ry. Co. (1913) 227 U.S. 296 [57 L.Ed. 518, 33 S.Ct. 273], the United States Supreme Court was faced with a question of statutory interpretation very similar to the question which is before us today. In 1908, the Federal Employers Liability Act—which granted railroad workers who had been injured in the course of their employment the right to bring a negligence action in federal court against the employer—had been amended to replace the doctrine of contributory negligence with comparative negligence. In Winfree, the plaintiff claimed that although the injury in that case had preceded the 1908 act, the comparative negligence doctrine should nonetheless be applied because the matter had not gone to trial until after the act had gone into effect. The plaintiff maintained that because even before the 1908 enactment the defendant railroad should have known that it could be held liable if its negligence resulted in a worker’s injury, there was no reason to deny the plaintiff the benefit of the new comparative negligence rule.
In Winfree, the Supreme Court rejected the plaintiff’s contention and held that the statute could not properly be applied to preexisting causes of action. In reaching its conclusion, the court relied on “the almost universal rule that statutes are addressed to the future, not to the past. They usually constitute a new factor in the affairs and relations of men and should not be held to affect what has happened unless, indeed, explicit words be used or by clear implication that construction be required.” (227 U.S. at p. 301 [57 L.Ed. at p. 520].) Because the 1908 amendment “introduced a new policy and quite radically changed the existing law,” the court emphasized that it was particularly the kind of statute that “should not be construed as retrospective.” (Id. at p. 302 [57 L.Ed. at p. 520].)
As we have explained, precisely the same principle is applicable here. Proposition 51 “introduced a new policy” which will have a *1227broad effect on most tort actions in California. Under Civil Code section 3 and the general principles of statutory interpretation, if the measure was intended to be applied retroactively, a provision directing retroactive application should have been included. In the absence of such an express declaration of retroactivity, we conclude that the proposition must be interpreted as prospective.
V.
Because we have concluded that the Court of Appeal erred in finding that Proposition 51 applies retroactively to this case, there is no need to reach the additional issues, relating to the interpretation and application of various portions of the proposition, which were discussed by the Court of Appeal.
The decision of the Court of Appeal is affirmed insofar as it upholds the constitutionality of Proposition 51, but is reversed insofar as it holds that Proposition 51 applies to causes of action that accrued prior to the effective date of the initiative measure.
Each party shall bear its own costs in these proceedings.
Mosk, Acting C. J., Broussard, J., and Panelli, J., concurred.
The complete text of Proposition 51 and all relevant portions of the election pamphlet, including the Legislative Analyst’s analysis and the arguments of the proponents and opponents, are set forth in an appendix to this opinion.
■ Under article II, section 10, subdivision (a) of the California Constitution, the measure went into effect on June 4, 1986, the day after the election.
The Contribution Act of 1957 (Code Civ. Proc., §§ 875-880) ameliorated the situation somewhat by permitting a pro rata division of damages when the plaintiff sued more than one defendant and a joint judgment was entered against the defendants. That act only applied, however, in instances in which a judgment had been entered against multiple defendants, and, if a plaintiff chose not to join a principally culpable tortfeasor in the action, the defendant or defendants who had been singled out for suit had no right to contribution.
Civil Code section 1431.2, which constitutes the heart of Proposition 51, provides in full: “(a) In any action for personal injury, property damage, or wrongful death, based upon principles of comparative fault, the liability of each defendant for non-economic damages shall be several only and shall not be joint. Each defendant shall be liable only for the amount of non-economic damages allocated to that defendant in direct proportion to that defendant’s percentage of fault, and a separate judgment shall be rendered against that defendant for that amount, [fl] (b)(1) For purposes of this section, the term ‘economic damages’ means objectively verifiable monetary losses including medical expenses, loss of earnings, burial costs, loss of use of property, costs of repair or replacement, costs of obtaining substitute domestic services, loss of employment and loss of business or employment opportunities, [fl] (2) For the purposes of this section, the term ‘non-economic damages’ means subjective, non-monetary losses including, but not limited to, pain, suffering, inconvenience, mental suffering, *1199emotional distress, loss of society and companionship, loss of consortium, injury to reputation and humiliation.”
At least five states apply a “pure” several liability rule. (See, e.g., Kan.Stat.Ann. § 60-258a(d) (1983); Vt.Stat.Ann. tit. 12, § 1036 (Supp. 1987); Ohio Rev.Code Ann. § 2315.19 (Page 1981); Utah Code Ann. §§ 78-27-38, 78-27-40 (1987); Colo.Rev.Stat. § 13-21-111.5 (1987). See also Wash.Rev. Code Ann. § 4.22.070 (West Supp. 1987) [adopting several liability as a general rule, but retaining joint and several liability in several, specified areas]; Nev.Rev.Stat.Ann. § 41.141 (Supp. 1987) [same].)
At least four states have adopted such an approach. (See, e.g., Iowa Code Ann. § 668.4 (West 1987) [joint and several liability does not apply to defendants who bear less that 50 percent of fault]; Minn.Stat.Ann. § 604.02(1) (West Supp. 1988) [if state or municipal defendant’s fault is less than 35 percent, “it is jointly and severally liable for an amount no greater than twice the amount of fault”]; Mo.Ann.Stat. § 538.230 (Vernon Supp. 1987) [in medical malpractice cases “any defendant against whom an award of damages is made shall be jointly liable only with those defendants whose apportioned percentage of fault is equal to or less than such defendant”]; Tex.Civ.Prac. & Rem. Code Ann. § 33.013 (Vernon 1988) [defendant severally liable unless percentage of fault is greater than 20 percent, or, in specified actions, defendant’s fault is greater than plaintiff’s].)
At least four states, in addition to California, have embraced such a rule. (See, e.g., N.Y. Civ.Prac.L. & R. § 1601 (McKinney Supp. 1987) [when defendant’s liability is less than 50 percent, defendant’s liability for plaintiff’s noneconomic loss shall not exceed that of defendant’s equitable share; numerous categories of cases excepted]; Fla.Stat.Ann. § 768.81(3) (West Supp. 1987) [joint and several liability abolished, except where a defendant’s percentage of fault equals or exceeds that of a particular claimant, the defendant is jointly and severally liable for the claimant’s economic damage]; Ore.Rev.Stat. § 18.485 (1983) [defendants severally liable for noneconomic damages, and jointly and severally liable for economic *1200damages unless defendant is less at fault than plaintiff or less than 15 percent at fault in which case defendant only severally liable for economic damages]; Ill.Ann.Stat. ch. 110, 2-1117, 2-1118 (Smith-Hurd Supp. 1987) [all defendants jointly and severally liable for medical expenses, defendants who are less than 25 percent at fault severally liable for all other damages, defendants who are more than 25 percent at fault jointly and severally liable for all other damages].)
Plaintiff’s petition for review lists the following allegedly unanswered questions as to the proposition’s application:
“1. Does it retroactively apply to this case?
“2. Does it apply if the jury finds Gregory 0% at fault? .
“3. Does it apply if the jury finds Van Waters & Rodgers liable based on strict products liability?
“4. [Does it] apply if the jury finds Student Science acted intentionally?
“5. If the jury finds Gregory more than 0% at fault how is his recovery adjusted?
“6. Who bears the burden of naming and serving other parties?
“7. Can the special verdict form contain a catch-all ‘other’ box or must such parties or non-parties be specified and limited to the evidence adduced at trial?”
Although plaintiff also suggests that the proposition’s classifications should be evaluated under a more stringent, “strict scrutiny” standard, the controlling decisions make it clear that the traditional “rational relationship” equal protection standard is applicable here. (See, *1203e.g., American Bank & Trust Co., supra, 36 Cal.3d 359, 373, fn. 12; Fein v. Permanente Medical Group (1985) 38 Cal.3d 137, 161-164 [211 Cal.Rptr. 368, 695 P.2d 665].)
In Fein, the court pointed out that legal commentators had long questioned whether sound public policy supported the comparable treatment of economic and noneconomic damages, explaining that “[tjhoughtful jurists and legal scholars have for some time raised serious questions as to the wisdom of awarding damages for pain and suffering in any negligence case, noting, inter alia, the inherent difficulties in placing a monetary value on such losses, the fact that money damages are at best only imperfect compensation for such intangible injuries and that such damages are generally passed on to, and borne by, innocent consumers. While the general propriety of such damages is, of course, firmly imbedded in our common law jurisprudence [citation], no California case of which we are aware has ever suggested that the right to recover for such noneconomic injuries is constitutionally immune from legislative limitation or revision.” (Footnote omitted.) (38 Cal.3d at pp. 159-160.)
In In re Marriage of Bouquet (1976) 16 Cal.3d 583, 587, footnote 3 [128 Cal.Rptr. 427, 546 P.2d 1371], the court specifically recognized that “[sjection 3 of the Civil Code embodies the common law presumption against retroactivity,” and numerous decisions of this court have recognized that comparable provisions in other codes represent legislative embodiments of this general legal principle. (See, e.g., Aetna Cas. & Surety Co. v. Ind. Acc. Com., supra, 30 Cal.2d 388, 395 [Lab. Code]; In re Estrada (1965) 63 Cal.2d 740, 746 [48 Cal.Rptr. 172, 408 P.2d 948] [Pen. Code]. See also DiGenova v. State Board of Education (1962) 57 Cal.2d 167, 172-173 [18 Cal.Rptr. 369, 367 P.2d 865].) To the extent that dictum in a footnote in the Court of Appeal decision in Andrus v. Municipal Court (1983) 143 Cal.App.3d 1041, 1045-1046, footnote 1 [192 Cal.Rptr. 341], discussing a similar provision of the Code of Civil Procedure, suggests that such a provision has no application to amendments to such codes and applies only to the original provisions of the codes, that dictum is contrary to the numerous Supreme Court decisions noted above and must be disapproved. (See also Estate of Frees (1921) 187 Cal. 150, 155-156 [201 P. 112] and cases cited.)
The full text of Proposition 51 is set out in the appendix to this opinion.
Plaintiff, taking his cue in part from a portion of the Court of Appeal decision in Russell v. Superior Court, supra, 185 Cal.App.3d 810, 818-819, suggests that the use of the word “shall” in various passages in the statute indicates that the drafters intended only a future operation. As defendants contend, however, in context we think it is more likely that the use of “shall” was intended to reflect the mandatory nature of the provision, rather than to refer to its temporal operation.
Defendants, in turn, rely on the initial clause of Civil Code section 1431.2, which states simply that the provision is to apply “[i]n any action. . . .” That familiar language, however, merely negates any implication that the new several liability rule was to apply only to a specific category of tort cases—like the earlier medical malpractice tort legislation—and provides no indication that a retroactive application was contemplated. Similar broad, general language in other statutory provisions has not been considered sufficient to indicate a legislative intent that the statute is to be applied retroactively. (See, e.g., United States v. Security Industrial Bank, supra, 459 U.S. 70, 82, fn. 12 [74 L.Ed.2d 235, 245] [“ ‘[a] few words of general connotation appearing in the text of statutes should not be given a wide meaning contrary to a settled policy, “excepting as a different purpose is plainly shown.’ ” [Citation]”]; Un. Pac. R.R. v. Laramie Stock Yards (1913) 231 U.S. 190, 199-202 [58 L.Ed. 179, 182-183, 34 S.Ct. 101].)
Civil Code section 1431.1, the introductory section of Proposition 51 which sets forth various “findings” and a “declaration of purpose,” provides in full: “The People of the State of California find and declare as follows: [j[] (a) The legal doctrine of joint and several liability, also known as ‘the deep pocket rule’, has resulted in a system of inequity and injustice that has threatened financial bankruptcy of local governments, other public agencies, private individuals and businesses and has resulted in higher prices for goods and services to the public and in higher taxes to the taxpayers. [j|] (b) Some governmental and private defendants are perceived to have substantial financial resources or insurance coverage and have thus been included in lawsuits even though there was little or no basis for finding them at fault. Under joint and several liability, if they are found to share even a fraction of the fault, they often are held financially liable for all the damage. The People—taxpayers and consumers alike—ultimately pay for these lawsuits in the form of higher taxes, higher prices and higher insurance premiums. [1J] (c) Local governments have been forced to curtail some essential police, fire and other protections because of the soaring costs of lawsuits and insurance premiums. Therefore, the People of the State of California declare that to remedy these inequities, defendants in tort actions shall be held financially liable in closer proportion to their degree of fault. To treat them differently is unfair and inequitable. [fí¡ The People of the State of California further declare that reforms in the liability laws in tort actions are necessary and proper to avoid catastrophic economic consequences for state and local governmental bodies as well as private individuals and businesses.”
In In re Estrada, supra, 63 Cal.2d 740, the court also held that a statutory enactment should be applied retroactively despite the absence of an express retroactivity clause, but that case involved considerations quite distinct from the ordinary statutory retroactivity question. In Estrada, the Legislature had amended a criminal statute to reduce the punishment to be imposed on violators; the amendment mitigating punishment was enacted after the defendant in Estrada had committed the prohibited act but before his conviction was final. Following the rule applied by the United States Supreme Court and a majority of states (see 63 Cal.2d at p. 748), the Estrada court concluded that the defendant should receive the benefit of the mitigated punishment “because to hold otherwise would be to conclude that the Legislature was motivated by a desire for vengeance, a conclusion not permitted in view of modem theories of penology.” (63 Cal.2d at p. 745.)
Although some of the broad language in Estrada was subsequently invoked in the civil context in the Mannheim, supra, 3 Cal.3d 678, and Marriage of Bouquet, supra, 16 Cal.3d 583, decisions, the rationale for the Estrada ruling bears little relationship to the determination of the retroactivity of most nonpenal statutes, and, as noted below, other jurisdictions have not applied the special rule applicable to ameliorative penal provisions in determining the retroactivity of a general tort reform measure like Proposition 51. We similarly conclude that the Estrada decision provides no guidance for the resolution of this case.
The Association for California Tort Reform (ACTR) is one of numerous organizations that have filed amici curiae briefs in this case. In its brief, ACTR states that it sponsored the legislation that was “the precursor to and model for Proposition 51” and that its chairman “was the official proponent who filed Proposition 51 with the California Attorney General requesting preparation of a title and summary for placement on the ballot.” ACTR participated as an amicus in many of the leading MICRA cases. (E.g., American Bank & Trust Co. v. Community Hospital, supra, 36 Cal.3d 359; Fein v. Permanente Medical Group, supra, 38 Cal.3d 137.)
Justice Gibson’s opinion in Aetna Cas. & Surety Co., supra, clearly demonstrates the untenability of defendants’ claim that the remedial nature of a statute is sufficient to support an inference that the statute was intended to apply retroactively. As noted above, in Aetna the question before the court was whether a statute which increased workers’ compensation benefits should be applied to workers who had sustained work-related injuries prior to the enactment of the new law but who were not awarded benefits until after the new statute took effect. In that case, unlike the present matter, of course, it was the injured parties who sought retroactive application of the statute; the workers argued that in light of the remedial nature of the increased benefits and the statutory mandate that provisions of the workers’ compensation law be liberally construed to extend benefits to injured workers (Lab. Code, § 3202), the court should infer an intent on the part of the Legislature to apply the act retroactively even though the act contained no express provision to that effect.
In rejecting the argument, the Aetna court observed: “No authority is cited for the novel doctrine which would require the court to ignore the rule against retroactive operation with respect to statutes increasing benefits to persons favored by remedial legislation. The rule of liberal construction and the rule that statutes should ordinarily be construed to operate prospectively are neither inconsistent nor mutually exclusive. ... It would be a most peculiar judicial reasoning which would allow one such doctrine to be invoked for the purpose of destroying the other. It seems clear, therefore, that the legislative intent in favor of the retrospective operation of a statute cannot be implied from the mere fact that the statute is remedial and subject to the rule of liberal construction.” (Italics added.) (Aetna Cas. & Surety Co., supra, 30 Cal.2d at p. 395.)
See, e.g., Winfree v. Nor. Pac. Ry. Co., supra, 227 U.S. 296; Brewster v. Ludtke (1933) 211 Wis. 344 [247 N.W. 449, 450]; Edwards v. Walker (1973) 95 Idaho 289 [507 P.2d 486, 488]; Dunham v. Southside National Bank (1976) 169 Mont. 466 [548 P.2d 1383]; Rice v. Wadkins (1976) 92 Nev. 631 [555 P.2d 1232, 1233]; Smith v. Shreeve (Utah 1976) 551 P.2d 1261, 1262, footnote 2; Scammon v. City of Saco (Me. 1968) 247 A.2d 108, 110; Costa v. Lair (1976) 241 Pa.Super. 517 [363 A.2d 1313, 1314-1315]; Viers v. Dunlap (1982) 1 Ohio St.3d 173 [438 N.E.2d 881]; contra, Godfrey v. State (1975) 84 Wash.2d 959 [530 P.2d 630].
Many of the recent comparative negligence statutes are not silent on the point, but specifically address the prospective/retroactive question. (See generally Schwartz, Comparative Negligence (2d ed. 1986) §§ 8.3-8.5, pp. 143-152.) Of the numerous statutes which expressly speak to the issue, all but two specifically provide for prospective operation. (Ibid.) The Uniform Comparative Fault Act, drafted by the National Conference of Commissioners on Uniform State Laws as a model for state laws on the subject, similarly contains a provision which mandates prospective application, declaring that “[t]his Act applies to all [claims for relief] [causes of action] which accrue after its effective date.” (§ 10.)
Although in the present case we do not know the additional parties plaintiff may have chosen to sue if Proposition 51 had been in effect at the outset of the litigation, defendants— in connection with their post-Proposition 51 filings—have suggested that some responsibility for the accident may lie either with some of plaintiff’s friends or with plaintiff’s parents. The statute of limitations on any cause of action plaintiff may have had against such individuals has, of course, long since run.
The dissenting opinion asserts that in light of the remedial purposes of Proposition 51, “the inference is virtually inescapable” that the electorate intended the proposition to apply to all trials conducted after the effective date of the measure. (See, post, at pp. 1232-1233.) The dissenting opinion apparently overlooks the fact, however, that most states which enacted tort reform measures similar to Proposition 51 in response to the same liability crisis which precipitated Proposition 51, and which specifically addressed the retroactivity issue in *1218their statutes, did not provide for retroactive application of the newly enacted reforms to all cases tried after the new enactment. (See, post, at pp. 1219-1220.) In light of these other enactments, it is difficult to understand how the dissent can find it “inescapable” from the context and purpose of the enactment that such a retroactive application must have been intended.
The dissenting opinion discusses a number of cases which it suggests support the proposition that remedial statutes are generally intended to apply retroactively. (See post, pp. 1233-1235.) The cases discussed by the dissent, however, did not involve general tort reform statutes, like Proposition 51, but rather concerned statutory enactments implementing procedural changes in circumstances in which it was unlikely that retroactive application would defeat a party’s reasonable reliance on the displaced procedural rule.
In its discussion of the proper interpretation of remedial statutes, the dissent makes no mention of the numerous decisions of both the United States Supreme Court and of state courts throughout the country which have overwhelmingly concluded that a tort reform statute, which is silent on the retroactivity question, should be applied prospectively to causes of action accruing after the effective date of the new statute. (See fn. 18, ante, p. 1215.)
The preambles of a number of the 1986 and 1987 statutes closely track the “Findings and Declaration of Purpose” in Proposition 51. (See, e.g., 1986 Wash. Laws, ch. 305, § 100; Tex. *1220Acts 1987, 70th Leg., 1st C.S., ch. 2, § 1.01, in Tex.Civ.Prac. & Rem. Code Ann., note following § 9.001 (Vernon 1988).)
The preamble to MICRA read in part: “The Legislature finds and declares that there is a major health care crisis in the State of California attributable to skyrocketing malpractice premium costs and resulting in a potential breakdown of the health delivery system, severe hardships for the medically indigent, a denial of access for the economically marginal, and *1221depletion of physicians such as to substantially worsen the quality of health care available to citizens of this state. The Legislature, acting within the scope of its police powers, finds the statutory remedy herein provided is intended to provide an adequate and reasonable remedy within the limits of what the foregoing public health and safety considerations permit now and into the foreseeable future.” (Stats. 1975, 2d Ex. Sess. 1975-1976, ch. 2, § 12.5, p. 4007.)
In Tulley, supra, 53 Cal. 274, the question at issue was the application of the amended version of Civil Code section 3336, setting forth the measure of damages for wrongful conversion of personal property. At the time the cause of action in Tulley arose, section 3336 provided, inter alia, that “[t]he detriment caused by the wrongful conversion of personal property is presumed to be the value of the property at the time of conversion, with the interest from that time, or, where the action has been prosecuted with reasonable diligence, the highest market value of the property at any time between the conversion and the verdict, without interest, at the option of the injured party...” (italics added); prior to the trial of the action, the section was amended to delete the emphasized portion of the statute.
*1223In Feckenscher, supra, 12 Cal.2d 482, the statutory change at issue involved a revision of Civil Code section 3343, pertaining to the measure of damages in a real estate fraud action. Although the opinion does not quote the version of section 3343 in effect at the time the action arose, it appears that at that point the statute permitted a defrauded plaintiff to recover a sum equal to the difference between defendant’s representation as to the value of the property which plaintiff received and the actual value of that property; as revised, section 3343 permitted recovery of “the difference between the actual value of that with which the defrauded person parted and the actual value of that which he received . . . .”
Stout, supra, 22 Cal.3d 718, like Feckenscher, supra, 12 Cal.2d 482, dealt with a revision of Civil Code section 3343, setting forth the measure of damages in a real estate fraud action.
In reaching its conclusion on the statutory interpretation issue, the Tulley court relied on the fact that the section in question provided that “[t]he detriment caused by the wrongful conversion of personal property is presumed tobe . . . ” (italics added), reasoning that “[t]he expression ‘is presumed to be’ indicates that it was intended to establish a legal presumption to operate, and which could only operate, at the trial of the cause . . . .” (53 Cal. at pp. 278-279.)
Although defendants in this case have not embraced the argument, several amici contend that Proposition 51 should be applied retroactively on the ground that the measure is “procedural” rather than “substantive.” The Court of Appeal, while concluding that retroactive application was warranted, nonetheless expressly rejected this argument, reasoning that because the provision could have a substantial effect on a defendant’s liability or a plaintiff’s recovery, “its substantive effect is evident.”
We agree with the Court of Appeal that retroactive application cannot be supported by characterizing Proposition 51 as merely a “procedural” statute. In addressing the question whether the retroactivity question may be resolved by denominating a statute as “substantive” or “procedural,” the court in Aetna Cas. & Surety, supra, 30 Cal.2d 388, 394, explained: *1226“In truth, the distinction relates not so much to the form of the statute as to its effects. If substantial changes are made, even in a statute which might ordinarily be classified as procedural, the operation on existing rights would be retroactive because the legal effects of past events would be changed, and the statute will be construed to operate only in futuro unless the legislative intent to the contrary clearly appears.” As explained above, retroactive application of Proposition 51 to preexisting causes of action would have a very definite substantive effect on both plaintiffs and defendants who, during the pending litigation, took irreversible actions in reasonable reliance on the then-existing state of the law. (See also 3 Harper et al., Law of Torts (2d ed. 1986) § 10.1, p. 7 [“The joint and several liability imposed on joint tortfeasors or independent concurrent tortfeasors producing an indivisible injury is a ‘substantive liability* to pay entire damages. This differs from what might be described as a ‘procedural liability’ to be joined with other tortfeasors as defendants in a single action.”].)