Opinion
BAXTER, J.This case presents two issues. First, we must consider when, if ever, a judicially confirmed award in an arbitration governed by California’s private arbitration law (Code Civ. Proc., § 1280 et seq.) is entitled to collateral estoppel, or “issue preclusion,” effect in favor of a nonparty to the arbitration.1 Second, we must determine whether a commercial general liability (CGL) insurance policy that provides coverage for sums the insured is “legally obligated to pay as damages” may cover losses arising from a breach of contract.
We reach the following conclusions: First, a private arbitration award, even if judicially confirmed, may not have nonmutual collateral estoppel effect under California law unless there was an agreement to that effect in the particular case.2 Second, the coverage phrase “legally obligated to pay as damages,” as used in a CGL insurance policy, may provide an insured *825defendant with coverage for losses pleaded as contractual damages. Accordingly, we will affirm the judgment of the Court of Appeal.
Factual and Procedural Background
The underlying litigation involves damage to a parcel of land that Vandenberg3 used as an automobile sales and service facility. Before 1958, owners Eugene and Kathryn Boyd4 operated an automobile dealership on the property. From 1958 to 1988, Vandenberg leased the property from Boyd under a series of leases. In 1988 Vandenberg discontinued the business and possession of the land reverted to Boyd.
To prepare the property for sale, Boyd removed three underground waste oil storage tanks. Testing revealed contamination of soils and groundwater underlying the property. Boyd filed an action against Vandenberg, alleging causes of action for breach of contract, breach of the covenant of gpod faith and fair dealing, public and private nuisance, negligence, waste, trespass, strict liability, equitable indemnity, declaratory relief, and injunctive relief. The Boyd complaint alleged Vandenberg had installed and operated the waste oil storage tanks and the tanks were the source of the petroleum contamination.
Vandenberg had obtained CGL insurance from several companies over the years, including Phoenix Assurance Company of New York (Phoenix), the Glens Falls Insurance Company (Glens Falls), Continental Insurance Company (Continental), TIG Insurance Corporation, Centennial Insurance Company (Centennial), and United States Fidelity and Guaranty Company (USF&G) (collectively insurers). The policies provided coverage to Vandenberg for sums he was “legally obligated to pay as damages” because of property damage. However, certain of the policies, including policies issued by USF&G and Centennial, also contained a so-called pollution exclusion, under which property damage caused by a pollutant or contaminant was not covered except for a “sudden and accidental” discharge.
*826Vandenberg tendered defense of the Boyd action to his insurers, but only USF&G agreed to provide a defense. During judicially supervised settlement proceedings, Vandenberg, Boyd, and USF&G reached an agreement among themselves to resolve the Boyd litigation. The agreement provided that its parties would contribute jointly to the investigation and remediation of the contamination, with USF&G bearing the largest share of the cost. Boyd agreed to release USF&G from any claims. Vandenberg agreed to release USF&G from claims for bad faith, breach of the contract, and extracontractual damages. Boyd released all claims against Vandenberg except those based on the theory that the contamination constituted a breach of the lease agreements. Boyd and Vandenberg agreed to resolve the reserved breach of lease issues through arbitration or by trial, depending upon their agreement on an arbitrator and arbitration schedule. Vandenberg conditioned his agreement to the settlement upon the arbitration being “binding.” USF&G agreed to defend Vandenberg, but the ultimate issues of USF&G’s coverage and indemnity obligations, as well as any claim by Vandenberg for Cumis counsel fees (see San Diego Federal Credit Union v. Cumis Ins. Society, Inc. (1984) 162 Cal.App.3d 358, 375 [208 Cal.Rptr. 494, 50 A.L.R.4th 913]), were “reserved for [future] resolution.” (Italics added.)5
The arbitration between Vandenberg and Boyd took place before a retired federal judge, Raul Ramirez. Formal discovery was conducted, and the transcribed proceedings included representation by counsel, and extensive evidence, briefing, and argument. In a lengthy and detailed decision, the arbitrator ruled for Boyd. Among other things, the arbitrator found that the contamination stemmed primarily from the underground waste oil tanks and was caused in part by Vandenberg’s improper installation, maintenance and use of the tanks. The arbitrator indicated the discharge of contaminants was not sudden and accidental. The arbitrator’s award of over $4 million to Boyd was confirmed by a superior court judgment.
The insurers rejected Vandenberg’s request for indemnification. He then filed the underlying action against his insurers, alleging various causes of action arising out of the failure to defend, settle, or indemnify in the Boyd action.
The insurers filed two motions for summary adjudication. In the first motion, Centennial and USF&G argued they had no duty to defend or indemnify because the pollution exclusion in their policies was triggered by *827the arbitrator’s determination in the Boyd action that the contamination was not sudden and accidental. These two insurers contended that Vandenberg’s relitigation of the “sudden and accidental” issue was precluded by principles of collateral estoppel. In the second motion, all insurers sought summary adjudication on the basis that the arbitrator awarded damages for breach of lease, a contractual cause of action, and contractual damages are not covered by the CGL insurance policies at issue.
The trial court granted both motions for summary adjudication. As to the first motion, the trial court ruled that relitigation of issues regarding the source and causation of the contamination was precluded by collateral estoppel. The court reviewed the arbitration transcript and concluded the “only reasonable inference is that the leaks or spills were occurring over a considerable period of time.” The court found, under Shell Oil Co. v. Winterthur Swiss Ins. Co. (1993) 12 Cal.App.4th 715 [15 Cal.Rptr.2d 815], that Vandenberg had no factual basis to contend the sudden and accidental exception to the pollution exclusion in Centennial and USF&G’s policies applied.6 As to the second motion, the trial court found Vandenberg had no coverage under the policies for the arbitration award because the claims submitted to the arbitrator were contractual.
After consolidating the cases, the Court of Appeal issued peremptory writs of mandate reversing both summary adjudication orders. The appellate court first held that, absent a contrary agreement by the arbitral parties, a party to private arbitration is not barred from relitigating issues decided by the arbitrator when those issues arise in a different case involving a different adversary and different causes of action. It would not be fair to give a private arbitration decision nonmutual collateral estoppel effect without the arbitral parties’ consent, the Court of Appeal reasoned, because private arbitration lacks significant safeguards of court litigation, particularly the right to full judicial review.7 The Court of Appeal also ruled that coverage under the insurance policies in question could not be determined by reference to the “general rule” that damages for an insured’s nonperformance of a contract are not covered under CGL insurance policies. (See International Surplus *828Lines Ins. Co. v. Devonshire Coverage Corp. (1979) 93 Cal.App.3d 601, 610-611 [155 Cal.Rptr. 870] (International Surplus).) Rather, the court reasoned, when there is damage to property, the focus of the inquiry should be the nature of the risk or peril that caused the injury and the specific policy language, not the form of action brought by the injured party.
We granted the insurers’ petitions for review to consider the circumstances, if any, in which private contractual arbitration decisions may have collateral estoppel effect in favor of nonparties, and whether CGL policy language indicating coverage for sums the insured becomes “legally obligated to pay as damages” can include losses pled as breach of contract. We consider each issue in turn.
Discussion
I.
Collateral estoppel is one of two aspects of the doctrine of res judicata. In its narrowest form, res judicata “ ‘precludes parties or their privies from relitigating a cause of action [finally resolved in a prior proceeding].’ ” (Teitelbaum Furs, Inc. v. Dominion (1962) 58 Cal.2d 601, 604 [25 Cal.Rptr. 559, 375 P.2d 439] (Teitelbaum Furs), quoting Bernhard v. Bank of America (1942) 19 Cal.2d 807, 810 [122 P.2d 892] (Bernhard).) But res judicata also includes a broader principle, commonly termed collateral estoppel, under which an issue “ ‘necessarily decided in [prior] litigation [may be] conclusively determined as [against] the parties [thereto] or their privies . . . in a subsequent lawsuit on a different cause of action.’ ” (Teitelbaum Furs, supra, 58 Cal.2d at p. 604, italics added.)
Thus, res judicata does not merely bar relitigation of identical claims or causes of action. Instead, in its collateral estoppel aspect, the doctrine may also preclude a party to prior litigation from redisputing issues therein decided against him, even when those issues bear on different claims raised in a later case. Moreover, because the estoppel need not be mutual, it is not necessary that the earlier and later proceedings involve the identical parties or their privies. Only the party against whom the doctrine is invoked must be bound by the prior proceeding. (Lucido v. Superior Court (1990) 51 Cal.3d 335, 341 [272 Cal.Rptr. 767, 795 P.2d 1223, 2 A.L.R.5th 995] (Lucido); Teitelbaum Furs, supra, 58 Cal.2d 601, 604; Bernhard, supra, 19 Cal.2d 807, 810-813.)
Accordingly, the collateral estoppel doctrine may allow one who was not a party to prior litigation to take advantage, in a later unrelated matter, of *829findings made against his current adversary in the earlier proceeding. This means that the loss of a particular dispute against a particular opponent in a particular forum may impose adverse and unforeseeable litigation consequences far beyond the parameters of the original case. (See Kelly v. Trans Globe Travel Bureau, Inc. (1976) 60 Cal.App.3d 195, 202 [131 Cal.Rptr. 488] (Kelly).)
Collateral estoppel (like the narrower “claim preclusion” aspect of res judicata) is intended to preserve the integrity of the judicial system, promote judicial economy, and protect litigants from harassment by vexatious litigation. (Lucido, supra, 51 Cal.3d 335, 343.) However, even where the minimal prerequisites for invocation of the doctrine are present, collateral estoppel “ ‘is not an inflexible, universally applicable principle; policy considerations may limit its use where the . . . underpinnings of the doctrine are outweighed by other factors.’ ” (Ibid., quoting Jackson v. City of Sacramento (1981) 117 Cal.App.3d 596, 603 [172 Cal.Rptr. 826]; see also Kelly, supra, 60 Cal.App.3d 195, 202.)
Whether collateral estoppel is fair and consistent with public policy in a particular case depends in part upon the character of the forum that first decided the issue later sought to be foreclosed. In this regard, courts consider the judicial nature of the prior forum, i.e., its legal formality, the scope of its jurisdiction, and its procedural safeguards, particularly including the opportunity for judicial review of adverse rulings. (See, e.g., United States v. Utah Constr. Co. (1966) 384 U.S. 394, 421-422 [86 S.Ct. 1545, 1559-1560, 16 L.Ed.2d 642] (Utah Constr. Co.) [findings by Board of Contract Appeals were entitled to collateral estoppel where agency acted in “judicial capacity,” findings were relevant to issues presented, parties had full and fair opportunity to litigate, and judicial review of adverse findings was available]; People v. Sims (1982) 32 Cal.3d 468, 479-482 [186 Cal.Rptr. 77, 651 P.2d 321] [Utah Constr. Co. test permits collateral estoppel effect, in criminal action, for prior findings of Department of Social Services]; Sanderson v. Niemann (1941) 17 Cal.2d 563, 573-575 [110 P.2d 1025] (Sanderson) [small claims judgments not entitled to collateral estoppel effect, given informality of proceedings, including limited right to judicial review]; cf. Pacific Estates, Inc. v. Superior Court (1993) 13 Cal.App.4th 1561, 1572-1575 [17 Cal.Rptr.2d 434] [collateral estoppel against nonparty insurers inappropriate where nature of proceedings did not guarantee full litigation of facts, and nonparty insurers could not, as matter of law, seek judicial review].)
Moreover, a particular danger of injustice arises when collateral estoppel is invoked by a nonparty to the prior litigation. (See, e.g., Parklane Hosiery Co. v. Shore (1979) 439 U.S. 322, 326-333 [99 S.Ct. 645, 649-653, 58 *830L.Ed.2d 552] (Parklane Hosiery); Kelly, supra, 60 Cal.App.3d 195, 202.) Such cases require close examination to determine whether nonmutual use of the doctrine is fair and appropriate. (See Parklane Hosiery, supra, at pp. 330-331 [99 S.Ct. at pp. 651-652]; see also Imen v. Glassford (1988) 201 Cal.App.3d 898, 906 [247 Cal.Rptr. 514].)
“Title 9 of the Code of Civil Procedure . . . represents a comprehensive statutory scheme regulating private arbitration in this state. (§ 1280 et seq.).” (Moncharsh v. Heily & Blase (1992) 3 Cal.4th 1, 9 [10 Cal.Rptr.2d 183, 832 P.2d 899] (Moncharsh).) The fundamental premise of the scheme is that “[a] written agreement to submit [either a present or a future controversy] to arbitration ... is valid, enforceable and irrevocable, save upon such grounds as exist for the revocation of any contract.” (Code Civ. Proc., § 1281.)8 The statutes set forth procedures for the enforcement of agreements to arbitrate (id., §§ 1281.2-1281.95), establish rules for the conduct of arbitration proceedings except as the parties otherwise agree (id., §§ 1282-1284.2), describe the circumstances in which arbitrators’ awards may be judicially vacated, corrected, confirmed, and enforced (id., §§ 1285-1288.8), and specify where, when, and how court proceedings relating to arbitation matters shall occur (id., §§ 1290-1294.2).
“Through this detailed statutory scheme, the Legislature has expressed a ‘strong public policy in favor of arbitration as a speedy and relatively inexpensive means of dispute resolution.’ [Citations.] Consequently, courts will ‘ “indulge every intendment to give effect to such proceedings.” ’ [Citations.] Indeed, more than 70 years ago, this court explained: ‘The policy of the law in recognizing arbitration agreements and in providing by statute for their enforcement is to encourage persons who wish to avoid delays incident to a civil action to obtain an adjustment of their differences by a tribunal of their own choosing.’ [Citation.]” (Moncharsh, supra, 3 Cal.4th 1, 9.)
“In cases involving private arbitration, ‘[t]he scope of arbitration is ... a matter of agreement between the parties’ [citation], and ‘ “[t]he powers of an arbitrator are limited and circumscribed by the agreement or stipulation of submission.” ’ [Citations.]” (Moncharsh, supra, 3 Cal.4th 1, 8-9.) An independent purpose of modem private arbitration statutes is to overcome earlier judicial hostility to arbitration agreements, and to ensure that such agreements, like other legally valid contracts, are enforced in accordance with their terms. (Dean Witter Reynolds Inc. v. Byrd (1985) 470 U.S. 213, 219-221 [105 S.Ct. 1238, 1241-1243, 84 L.Ed.2d 158] [discussing *831similar contemporaneously enacted provisions of FAA].) Accordingly, policies favoring the efficiency of private arbitration as a means of dispute resolution must sometimes yield to its fundamentally contractual nature, and to the attendant requirement that arbitration shall proceed as the parties themselves have agreed. (Ibid.)
Of course, the parties to a private arbitration need not, and sometimes may not, specify every detail, characteristic, and consequence of the proceeding they contemplate. As indicated above, the California statutes set forth certain basic, common characteristics of such proceedings, including their general relationship to the judicial system, that will apply even absent the specific agreement of the parties. These include, among others, the limited grounds and procedures for judicial review of a private arbitration award. (Code Civ. Proc., §§ 1286.2, 1286.6; Moncharsh, supra, 3 Cal.4th 1, 8-33 [finding statutory grounds for judicial review exclusive].)
Limited judicial review is a well-understood feature of private arbitration, inherent in the nature of the arbitral forum as an informal, expeditious, and efficient alternative means of dispute resolution. By choosing private arbitration, the parties “evince [their] intent to bypass the judicial system and thus avoid potential delays at the trial and appellate levels.” (Moncharsh, supra, 3 Cal.4th 1, 10.) Judicial interference with the arbitrator’s decision would thus defeat the very advantages the arbitral parties sought to achieve. (Ibid.)
Accordingly, the parties, simply by agreeing to arbitrate, are deemed to accept limited judicial review by implication, particularly where their agreement specified that the award would be “final” and “binding” upon them. (Moncharsh, supra, 3 Cal.4th. 1, 9-10.) In effect, it is appropriate to insulate a private arbitral award from close judicial scrutiny because, given the inherent nature of arbitration, “the parties have agreed that it be so.” (Id., at p. 10, italics in original.)
However, very different considerations affect the issue whether private arbitration awards should have nonmutual collateral estoppel effect. California’s statutory scheme nowhere specifies that, despite the arbitral parties’ failure so to agree, a private arbitration award may be binding in favor of nonparties in litigation involving different causes of action. Moreover, in our view, such a consequence is not an inherent or expected feature of private arbitration that is implicitly accepted by the arbitral parties.
As we recently observed, private arbitration is a process in which parties voluntarily trade the safeguards and formalities of court litigation for an expeditious, sometimes roughshod means of resolving their dispute. The traditional rule is that “ ‘[arbitrators, unless specifically required *832to act in conformity with rules of law, may base their decision upon broad principles of justice and equity, and in doing so may expressly or impliedly reject a claim that a party might successfully have asserted in a judicial action.’’ [Citations.] As early as 1852, this court recognized that, ‘The arbitrators are not bound to award on principles of dry law, but may decide on principles of equity and good conscience, and make their award ex aequo et bono [according to what is just and good].’ [Citation.] ‘As a consequence, . . . “[p]arties who stipulate in an agreement that controversies . . . shall be settled by arbitration, may expect not only to reap the advantages that flow from the use of that nontechnical, summary procedure, but also to find themselves bound by an award reached by paths neither marked nor traceable and not subject to judicial review.” [Citation.]’ ” (Moncharsh, supra, 3 Cal.4th 1, 10-11, quoting Nogueiro v. Kaiser Foundation Hospitals (1988) 203 Cal.App.3d 1192, 1195 [250 Cal.Rptr. 478], italics added.)
But while the informal and imprecise nature of private arbitration, and its insulation from judicial interference, are “ ‘the very advantages the . . . parties [seek] to achieve’ ” in arbitrating their own claims (see, ante, p. 831, quoting Moncharsh, supra, 3 Cal.4th 1, 10), these same features can be serious, unexpected disadvantages if issues decided by the arbitrator are given leveraged effect in favor of strangers to the arbitration.
An agreement to arbitrate particular claims reflects each party’s conclusion that the immediate stakes make it preferable to avoid the delay and expense of court proceedings, and instead to resolve the matter between themselves without resort to the judicial process. Under such circumstances, each party is willing to risk that the arbitration will result in a “final” and “binding” defeat with respect to the submitted claims, even though the party would have won in court, and even though the arbitrator’s errors must be accepted without opportunity for review. (See Moncharsh, supra, 3 Cal.4th 1, 10-12.) But this does not mean each arbitral party also consents that issues decided against him by this informal, imprecise method may bind him, in the same manner as a court trial, in all future disputes, regardless of the stakes, against all adversaries, known and unknown.
On the contrary, common sense weighs against the assumption that parties contemplate such remote and collateral ramifications when they agree to arbitrate controversies between themselves. Logic equally suggests that conscious agreements to give arbitrators’ decisions nonmutual collateral estoppel effect would not be routine. The very fact that arbitration is by nature an informal process, not strictly bound by evidence, law, or judicial oversight, suggests reasonable parties would hesitate to agree that the arbitrator’s findings in their own dispute should thereafter bind them in cases involving different adversaries and claims. Even where, as here, the arbitral *833parties have imposed some formality on their proceedings, have aired their dispute thoroughly, and have received a detailed decision, there is no reason to assume they agreed to “issue preclusive” effect in favor of nonparties. In the usual case, tactical considerations would weigh against such an agreement. Most often, the effect would be to burden whichever party lost the arbitration, while affording no corresponding benefit to either arbitral party. (See Shell, Res Judicata and Collateral Estoppel Effects of Commercial Arbitration (1988) 35 UCLA L.Rev. 623, 667.).
Accordingly, there is little basis to surmise that mere silence implies the arbitral parties’ acceptance of nonmutual collateral estoppel. A general rule that confirmed that private arbitation awards may have such effect would thus violate the fundamental premise that private arbitration is a contractual proceeding whose scope and effect are defined and limited by the parties’ consent. For similar reasons, such a rule would chill, rather than promote, the voluntary use of the arbitral forum as an efficient and informal alternative means of resolving particular controversies.
Under these circumstances, the public policy reasons against applying the collateral estoppel doctrine (see, e.g., Lucido, supra, 51 Cal.3d 335, 343; Kelly, supra, 60 Cal.App.3d 195, 202) well outweigh those in favor of doing so. In fact, the traditional justifications for collateral estoppel—factors which “strongly influence whether its application in a particular circumstance would be fair to the parties and constitute[] sound [public] policy” (Lucido, supra, at p. 343)—have diminished force when the nonmutual prong of the doctrine is applied to private arbitration without the arbitral parties’ specific consent.
As noted above, the primary purposes of collateral estoppel are to “preserv[e] the integrity of the judicial system, promot[e] judicial economy, and protect[] litigants from harassment by vexatious litigation.” (Lucido, supra, 51 Cal.3d 335, 343.) But because a private arbitrator’s award is outside the judicial system, denying the award collateral estoppel effect has no adverse impact on judicial integrity. Moreover, because private arbitration does not involve the use of a judge and a courtroom, later relitigation does not undermine judicial economy by requiring duplication of judicial resources to decide the same issue. Finally, when collateral estoppel is invoked by a nonparty to the private arbitration, the doctrine does not serve the policy against harassment by vexatious litigation. In such cases, the doctrine is asserted not to protect one who has already once prevailed against the same opponent on the same cause of action, but simply to gain vicarious advantage from a litigation victory won by another.
We therefore face a situation in which the policies underlying the doctrine of collateral estoppel must yield to the contractual basis of private arbitration, i.e., the principle that the scope and effect of the arbitration are for the *834parties themselves to decide. Accordingly, we are compelled to conclude that a private arbitration award, even if judicially confirmed, can have no collateral estoppel effect in favor of third persons unless the arbitral parties agreed, in the particular case, that such a consequence should apply.
We realize that some commentators, and most other courts addressing the issue, have taken a contrary approach. The predominant view is that unless the arbitral parties agreed otherwise, a judicially confirmed private arbitration award will have collateral estoppel effect, even in favor of nonparties to the arbitration, if the arbitrator actually and necessarily decided the .issue sought to be foreclosed and the party against whom estoppel is invoked had full incentive and opportunity to litigate the matter. (E.g., Witkowski v. Welch (3d Cir. 1999) 173 F.3d 192, 198-205 (Witkowski) [applying federal and Pennsylvania law]; Mandich v. Watters (8th Cir. 1992) 970 F.2d 462, 465-467 [applying Minnesota law]; Cities Service Co. v. Gulf Oil Corp. (1999) 1999 OK 14 [980 P.2d 116, 123-130 ] [applying federal and Oklahoma law]; Konieczny v. Micciche (1997) 305 N.J.Super. 375 [702 A.2d 831, 836-837]; Western Indus. v. Kaldveer Associates (1994) 126 Idaho 541 [887 P.2d 1048, 1050-1052] (Kaldveer); Aufderhar v. Data Dispatch, Inc. (Minn. 1990) 452 N.W.2d 648, 650-654; Clemens v. Apple (1984) 102 A.D.2d 236 [477 N.Y.S.2d 774, 775-776], affd. (1985) 65 N.Y.2d 746 [492 N.Y.S.2d 20, 481 N.E.2d 560, 561] (Clemens); see Rest.2d Judgments, § 84; id., com. h, p. 292; see also American Ins. Co. Messinger (1977) 43 N.Y.2d 184 [401 N.Y.S.2d 36, 371 N.E.2d 798, 803-804] (Messinger) [applying issue preclusion in later litigation between former arbitral parties themselves, court specifically rejects argument that parties did not agree to such consequence, finding no contrary agreement]; cf. Kerins v. Prudential Property & Cas. (1992) 185 A.D.2d 403 [585 N.Y.S.2d 637, 639] [finding, in the arbitral parties’ “agreement and the explicit rules governing their arbitration process,” a “clear” intent not to allow collateral estoppel in favor of nonparties].) To determine whether particular arbitration proceedings provided a full and fair opportunity to litigate, courts typically proceed case by case, assessing multiple factors, including the extent to which the arbitration resembled a court trial. (Motomura, Arbitration and Collateral Estoppel: Using Preclusion to Shape Procedural Choices (1988) 63 Tul. L.Rev. 29, 32-36, 52-53 (Motomura) & cases cited; see, e.g., Clemens, supra, 481 N.E.2d 560, 561 [nature of forum, importance of claim in arbitration, incentive to litigate, extent of actual litigation].)
When justification for these rules is offered, it centers on one or more of three premises. The first is the general policy against relitigation of issues already decided. (E.g., Messinger, supra, 371 N.E.2d 798, 803.) The second is that collateral estoppel causes no injustice when the party to be bound had a full and fair opportunity to litigate the issues to be foreclosed. The third, *835and most sophisticated, is that “final” and “binding” arbitration necessarily implies the possibility of collateral estoppel, particularly when (as in California) the law gives judicially confirmed arbitration awards the force and effect of civil judgments. (See, e.g., Witkowski, supra, 173 F.3d 192, 199-200 [arbitration award confirmed by federal district court is “final judgment” on the merits under FAA which can have issue-preclusive effect under Pennsylvania law]; Kaldveer, supra, 887 P.2d 1048, 1051 [Idaho’s version of Uniform Arbitration Act provides that confirmed arbitration award shall be enforced as any other judgment or decree].)
Respectfully, we find these rationales unpersuasive. As our earlier discussion suggests, we believe they give insufficient consideration and weight to the voluntary, contractual, and informal nature of private arbitration, and to the consequent reasonable expectations of the arbitral parties.
In particular, we reject the notion, strongly urged by the concurring and dissenting opinion, that California’s statutory provision giving confirmed private arbitration awards the force and effect of civil judgments (see Code Civ. Proc., § 1287.4)9 automatically implies that such awards “may have nonmutual collateral estoppel effect with or without the [specific] consent of the arbitral parties.” (Conc. & dis. opn. of Brown, J., post, at p. 842, italics in original.) Under California law, collateral estoppel will apply in any setting only where such application comports with fairness and sound public policy. (Lucido, supra, 51 Cal.3d 335, 343.) In keeping with that principle, Sanderson, supra, 17 Cal.2d 563, denied collateral estoppel effect to issues decided by a final judgment of the small claims court.
As we have indicated, California’s private arbitration statutes, including Code of Civil Procedure section 1287.4, do not warn parties who choose arbitation over court litigation that the arbitrator’s award may be used against them by third persons to resolve different causes of action. The contractual nature of private arbitration dictates that the scope and effect of an arbitral award must derive from the parties’ consent. Yet the informal nature of arbitration, the usual reasons for its use, the potentially disproportionate consequences of nonmutual collateral estoppel, and the fact that such consequences may not be immediately apparent to the arbitral parties, all suggest that their silence on the subject does not imply consent. Fairness and public policy thus counsel against application of nonmutual collateral estoppel in this setting, unless the parties specifically agree thereto. That California law treats confirmed arbitration awards as judgments does not compel us to conclude otherwise.
*836Moreover, we believe the case-by-case approach prevalent in other jurisdictions has particular adverse effects on the choice, use, and effectiveness of private arbitration as a faster, cheaper alternative to litigation in court. Under the prevailing rule, parties who agree to arbitrate, but neglect or fail to negotiate a specific disclaimer of collateral estoppel effect, cannot know in advance whether a court may later find the arbitration binding in favor of third parties on different claims. If the issue ever arises in future litigation, its resolution will depend on a judicial opinion whether the prior arbitration afforded the losing party a full and fair opportunity to litigate the matter to be foreclosed.
This “ad hoc, post hoc” standard eliminates a prime benefit of choosing the arbitral over the judicial forum—the right to shape, control, know, and predict at the outset the scope and effect of the arbitrator’s decision. Moreover, the case-by-case approach encourages both the arbitral parties, and the arbitrator himself, to “hedge” against the future possibility that the decision will be deemed binding in different litigation with other parties. “By creating the possibility that some arbitral findings will [have] collateral estoppel [effect], the case-by-case approach puts subtle but strong pressure on the arbitration process to conform its perspective and methods to those of litigation, in order to justify the confidence in its findings that collateral estoppel represents. [A] compelling reason to abandon the case-by-case approach to arbitral collateral estoppel is to ease or eliminate this pressure, in order to maintain arbitration as a useful, distinct alternative to litigation.” (Motomura, supra, 63 Tul. L.Rev. 29, 71.)10
Accordingly, we adopt, for California purposes, the rule that a private arbitration award cannot have nonmutual collateral estoppel effect unless the *837arbitral parties so agree. It remains to apply this rule to the facts before us.
We note first that neither USF&G nor Centennial was a party to the arbitration between Vandenberg and Boyd.11 Moreover, policyholder Vandenberg’s current insurance coverage claims against USF&G and Centennial are entirely distinct from the breach of lease claims decided in favor of lessor Boyd, and against lessee Vandenberg, in the Boyd/Vandenberg arbitration. Hence, USF&G and Centennial seek to give the arbitrator’s decision nonmutual collateral estoppel effect against Vandenberg. They may not do so unless Vandenberg and Boyd so agreed.
We find no such agreement. On the contrary, the terms of the three-way settlement among Boyd, Vandenberg, and USF&G strongly suggest the parties’ intent that, while the arbitration would be “binding” between Vandenberg and Boyd, it should not have collateral estoppel effect in favor of Vandenberg’s insurers. Under the express provisions of the settlement, some of Boyd’s causes of action against Vandenberg would be dismissed; the remaining, or “reserved,” claims in the Boyd action would be decided by “binding” arbitration; Boyd and Vandenberg would release certain potential causes of action against USF&G; and USF&G would provide Vandenberg a defense in the Boyd action under a “full reservation of rights”; while “all questions regarding ultimate coverage and indemnity obligations between USF&G and . . . Vandenberg . . . [were] reserved for resolution.” (Italics added.)
The clear import of these provisions is that the arbitration would finally resolve the Boyd action, but that ancillary insurance issues between Vandenberg and USF&G, aside from those issues specifically addressed in the settlement, would be unaffected. Under the settlement’s terms, both Vandenberg and USF&G retained the right, in future litigation between them, to address de novo “all questions” pertinent to USF&G’s “coverage and indemnity obligations.”
Though Centennial was not a party to the settlement, there is no basis to infer that the parties intended or agreed the arbitrator’s decision would have *838broader effect in Centennial’s favor than it would have in USF&G’s. Accordingly, we conclude there was no agreement by the arbitral parties allowing Vandenberg’s insurers to assert, against Vandenberg, the collateral estoppel effect of issues decided in the Boyd/Vandenberg arbitration.
The Court of Appeal therefore correctly reversed the trial court order granting the motion of USF&G and Centennial for summary adjudication.
II.
The CGL insurance policies issued by all the insurers in this case provided coverage for sums Vandenberg was legally obligated to pay as damages because of property damage. We next consider whether this coverage language in a CGL insurance policy necessarily precludes coverage for losses pleaded as contractual damages.
In holding that coverage for property damage losses is not necessarily precluded because they are pled as contractual damages, the Court of Appeal properly focused on the property itself and the nature of the risk causing the injury. Acknowledging the line of decisions espousing a general rule of noncoverage for contractual damages, the Court of Appeal nevertheless concluded “the general rule is not a universal bar to insurance coverage whenever a contract is involved. Rather, the focus of coverage for property damage is the property itself. (Waller v. Truck Ins. Exchange, Inc. [(1995)] 11 Cal.4th [1,] 17 [44 Cal.Rptr.2d 370, 900 P.2d 619].)” Coverage under a CGL insurance policy is not based upon the fortuity of the form of action chosen by the injured party. Thus, as the Court of Appeal stated, determination of coverage must be made individually by considering “the nature of [the] property, the injury, and the risk that caused the injury, in light of the particular provisions of each applicable insurance policy.”
The insurers contend CGL insurance policies limiting coverage to amounts the insured is “legally obligated to pay as damages,” or using similar language, refer to tort liability and not contractual liability. A long line of decisions supports their position, holding that any liability arising ex contractu, as opposed to ex delicto, is not covered under such policies.12 The underlying reasoning of these cases is that the phrase “legally obligated to *839pay as damages” describes liability based upon a breach of a duty imposed by law, i.e., tort, rather than by contract. (See Wilmington, supra, 53 Cal.App.4th 186, 193.)
We disagree. The nature of the damage and the risk involved, in light of particular policy provisions, control coverage. Moreover, we reject the ex contractu/ex delicto distinction, which derives from a misreading of the seminal case, Ritchie v. Anchor Casualty Co. (1955) 135 Cal.App.2d 245 [286 P.2d 1000] (Ritchie). In Ritchie, the court analyzed whether the term “liability imposed by law,” the precursor to “legally obligated to pay,” included coverage for liability arising from contract. (Id. at p. 254.) This phrase had usually been construed to mean liability imposed in a definite sum by a final judgment against the assured. (Ibid.) But the policy before the Ritchie court contained a distinction; coverage A applied to “ ‘liability imposed ... by law or by written contract,’ ” whereas coverage B applied to “ ‘liability imposed ... by law.’ ” (Ibid.) The court concluded that the omission of the term “ ‘or by written contract’ ” in coverage B, the portion at issue in Ritchie, “is persuasive that the phrase ‘imposed upon him by law’ as used in this policy . . . relates to the nature of the liability to be defended rather than the result of the lawsuit . . . .” (Ibid., italics added.)
In International Surplus, supra, 93 Cal.App.3d 601, 611, the phrase at issue was “ ‘legally obligated to pay as damages,’ ” which the court found synonymous with “ ‘damages for a liability imposed by law,’ ” the coverage language in the Ritchie case. Without further discussion, the court then held that the “latter phrase has been uniformly interpreted as referring to a liability arising ex delicto as distinguished from ex contractu.” (Ibid., citing Ritchie, supra, 135 Cal.App.2d 245, italics added.) This brief statement led to a string of cases relying upon International Surplus for the purported distinction between tort and contract damages. These later cases fail to consider, however, the particular and explicit coverage language before the Ritchie court, and thus create an arbitrary distinction that ignores otherwise settled principles of insurance contract interpretation.
Insurance policies are contracts construed in accordance with the parties’ mutual intent at the time of contract formation, as inferred from the written provisions. (Civ. Code, §§ 1636, 1639; Montrose Chemical Corp. v. Admiral Ins. Co. (1995) 10 Cal.4th 645, 666 [42 Cal.Rptr.2d 324, 913 P.2d 878].) The “clear and explicit” meaning of the provisions, interpreted in their *840“ordinary and popular sense,” controls judicial interpretation unless “used by the parties in a technical sense or a special meaning is given to them by usage.” (Civ. Code, §§ 1638, 1644.) If the meaning a layperson would ascribe to insurance contract language is not ambiguous, courts will apply that meaning. (AIU Ins. Co. v. Superior Court (1990) 51 Cal.3d 807, 822 [274 Cal.Rptr. 820, 799 P.2d 1253] (AIU).)
Even if a provision raises doubts as to coverage in the minds of legally trained observers due to a sophisticated legal distinction, courts will not assume the distinction was incorporated into the policy. (See AIU, supra, 51 Cal.3d 807, 825.) Whatever ambiguity a phrase possesses due to a party’s legal knowledge is resolved in favor of coverage. (Id. at p. 825.) In AIU, we rejected the claim of noncoverage based upon the fact that the forms of relief sought against the insured by third party suits were equitable and therefore the insured was not legally obligated to pay under the terms of the insurance policy. “Because California has generally abandoned the traditional distinction between courts of equity and courts of law [citations], even a legally sophisticated policyholder might not anticipate that the term ‘legally obligated’ precludes coverage of equitably compelled expenses.” (Ibid.)
Nothing in the respective policies between Vandenberg and any of the insurers suggests any special or legalistic meaning to the phrase “legally obligated to pay as damages.” A reasonable layperson would certainly understand “legally obligated to pay” to refer to any obligation which is binding and enforceable under the law, whether pursuant to contract or tort liability. Further, a reasonable layperson, cognizant that he or she is purchasing a “general liability” insurance policy, would not conclude such coverage term only refers to liability pled in tort, and thus entirely excludes liability pled on a theory of breach of contract. Under general insurance principles, we must interpret the phrase “legally obligated to pay as damages” in accordance with the ordinary and popular sense, not the legalistic, and erroneously premised, interpretation of the language urged by insurers.
Moreover, the arbitrariness of the distinction between contract and tort in the International Surplus line of cases is evident when we consider the same act may constitute both a breach of contract and a tort. (See Eads v. Marks (1952) 39 Cal.2d 807, 809-811 [249 P.2d 257].) Predicating coverage upon an injured party’s choice of remedy or the form of action sought is not the law of this state. (See AIU, supra, 51 Cal.3d 807, 824-825.) The International Surplus rule would thus permit the injured third party to determine insurance coverage. Instead, courts must focus on the nature of the risk and the injury, in light of the policy provisions, to make that determination. In AIU, for example, in rejecting a “form of remedy” approach for a determination of coverage, we focused on the nature of the injury and the specific *841policy language to decide whether there was coverage under the general liability policy for cleanup and other response costs under the Comprehensive Environmental Response, Compensation and Liability Act (CERCLA) (42 U.S.C. § 9601 et seq.) and related state and federal laws. (AIU, supra, 51 Cal.3d at pp. 819, 824-843; see Haralambos, supra, 195 Cal.App.3d 1308, 1317 [entitlement to defense if recovery sought damages for property damage; no property damage and thus, no coverage]; International Surplus, supra, 93 Cal.App.3d 601, 611 [insured’s conduct did not result in property damage or personal injury under policy].)
Insurance treatises concur with this approach. “[Wjhether a particular claim falls within the coverage afforded by a liability policy is not affected by the form of the legal proceeding. Accordingly, the legal theory asserted by the claimant is immaterial to the determination of whether the risk is covered.” (9 Couch, Insurance (3d ed. 1997) § 126:3, p. 126-8.) Insurance commentators explain: “The expression ‘legally obligated’ connotes legal responsibility that is broad in scope. It is directed at civil liability .... [which] can arise from either unintentional (negligent) or intentional tort, under common law, statute, or contract.” (Malecki & Flitner, Commercial General Liability (6th ed. 1997) p. 6, italics added.) “The coverage agreement [which] embraces ‘all sums which the insured shall become legally obligated to pay as damages is intentionally broad enough to include the insured’s obligation to pay damages for breach of contract as well as for tort, within limitations imposed by other terms of the coverage agreement (e.g. bodily injury and property damage as defined, caused by an occurrence) and by the exclusions . . . .” (Tinker, Comprehensive General Liability Insurance—Perspective and Overview (1975) 25 Fed. Ins. Coun. Q. 217, 265.)
We therefore conclude that the International Surplus rationale, distinguishing contract from tort liability for purposes of the CGL insurance coverage phrase “legally obligated to pay as damages,” is incorrect.13 Accordingly, we uphold the Court of Appeal’s determination that Vandenberg’s insurers cannot avoid coverage for damages awarded against Vandenberg in the Boyd action solely on grounds the damages were assessed on a contractual theory.
*842Disposition
The judgment of the Court of Appeal is affirmed.
Mosk, Acting C. J., Kennard, J., and Werdegar, J., concurred.
Tor reasons that will appear, “issue preclusion” effect in favor of a nonparty is sometimes hereafter referred to as nonmutual collateral estoppel.
Our holding is narrowly circumscribed. Nothing in our decision imposes or implies any limitations on the strict res judicata, or “claim preclusive,” effect of a California law private arbitration award. (See, e.g., Thibodeau v. Crum (1992) 4 Cal.App.4th 749, 756-761 [6 Cal.Rptr.2d 27] [unconfirmed award in private arbitration between homeowner and general contractor is res judicata barring homeowner’s identical claim against subcontractor]; Sartor v. Superior Court (1982) 136 Cal.App.3d 322, 327-328 [187 Cal.Rptr. 247] [confirmed private arbitration award in favor of architectural firm is res judicata barring homeowner’s identical causes of action against firm’s employees].) We also do not address the circumstances, if any, in which a private arbitration award may have “issue preclusive” effect in subsequent litigation between the same parties on different causes of action. No party has suggested the arbitration here at issue is governed by the Federal Arbitration Act (9 U.S.C.A. §§ 1-14) (FAA), and we have no occasion to consider whether application of the FAA would alter our ruling. We are not concerned with the collateral estoppel implications of arbitrations conducted pursuant to collective bargaining agreements within the purview of federal or state labor relations laws (see, e.g., Kelly v. Vons Companies, Inc. (1998) 67 Cal.App.4th 1329, 1339-1341 [79 Cal.Rptr.2d 763] [union arbitration to enforce collective bargaining agreement held preclusive against union members asserting similar issues in subsequent suit against employer]; but cf. Lehto v. Underground Constr. Co. (1977) 69 Cal.App.3d 933, 939-944 [138 Cal.Rptr. 419] [union arbitration held not preclusive against union members’ subsequent claim of unfair representation]). Nor do we express views on the collateral estoppel effects of arbitrations conducted under various other California statutory schemes. (See, e.g., Code Civ. Proc., § 1141.10 et seq. [mandatory judicial arbitration of civil actions with low amounts in controversy]; State Farm Mut. Auto. Ins. Co. v. Superior Court (1989) 211 Cal.App.3d 5, 12-15 [259 Cal.Rptr. 50] [judgment against insured after mandatory judicial arbitration in auto accident case can have collateral estoppel effect against auto liability insurer in third party bad faith action]; but see Flynn v. Gorton (1989) 207 Cal.App.3d 1550, 1554-1556 [255 *825Cal.Rptr. 768] [final decision in mandatory judicial arbitration should not have collateral estoppel effect]; see also Ins. Code, §§ 11580.2, subd. (f), 11580.5 [statutory arbitration of injured person’s claim for uninsured motorist coverage has no collateral estoppel effect in subsequent tort lawsuit against uninsured motorist]; Bus. & Prof. Code, § 6204, subd. (e) [statutory arbitration of attomey/client fee dispute has no res judicata or collateral estoppel effect in other proceedings].)
The plaintiffs in the underlying insurance litigation are John B. Vandenberg, Jeanette B. Vandenberg, James A. Keil, Bonnie J. Keil, Vandenberg & Keil, a general partnership, and Vandenberg Motors, Inc. These persons had various interests in the business known as Vandenberg Motors and we shall refer to them collectively as Vandenberg.
The Boyds are not involved in the indemnification litigation, the writ proceedings or in this appeal. Eugene died during the course of the litigation of their complaint. We shall refer to their interests by the singular Boyd.
The settlement agreement does not appear in the record as a signed contract. Instead, the record reflects that the terms of the settlement, including its arbitration provisions, were read in open court by USF&G’s counsel, after which the court questioned the individual parties to ascertain their understanding and assent. (See Code Civ. Proc., § 664.6.)
Vandenberg sought clarification that the absence of a pollution exclusion in the policies of Continental, Phoenix and Glens Falls prevented their attempted joinder in Centennial’s motion based on collateral estoppel. The trial court did not expressly rule on this request.
At all court levels, the parties have focused on whether fairness permits the collateral estoppel use of confirmed contractual arbitration awards. The briefs do not address what significance should attach to the arbitral parties’ agreement or failure to agree that their arbitration should have nonmutual collateral estoppel effect. In response to a question at oral argument, counsel for USF&G suggested that specific agreement on the issue was irrelevant. But our ruling that contractual arbitration- awards may never have nonmutual collateral estoppel effect unless the arbitral parties so agree can produce no unfair surprise. As noted above, the Court of Appeal reached exactly that conclusion.
Though, as noted above (see fn. 5, ante), the arbitration agreement in this case was formalized in open court, rather than by an executed writing, no party has suggested the agreement is exempt from California’s private contractual arbitration statutes because there was no “written agreement.”
As amended in 1998, section 1287.4 provides: “If [a private arbitration] award is confirmed, judgment shall be entered in conformity therewith. The judgment so entered has the same force and effect as, and is subject to all the provisions of law relating to, a judgment in ' a civil action of the same jurisdictional classification; and it may be enforced like any other judgment of the court in which it is entered, in an action of the same jurisdictional classification.”
The concurring and dissenting opinion asserts that by leaving the nonmutual collateral estoppel effect of a private arbitration award to the agreement of the arbitral parties themselves, we enhance the dangers of inconsistent arbitral and judicial rulings, improperly hinder the arbitration of arbitrable disputes, encourage “procedural gamesmanship” (cone. & dis. opn. of Brown, J., post, at p. 844), and offer arbitral parties unfair “second bite[s] at the apple” {ibid.), particularly in complex cases where arbitrable disputes are closely related to broader litigation that includes arbitral nonparties. But as the concurring and dissenting opinion effectively concedes (and at least where, as here, the arbitration agreement is governed exclusively by California law), this state’s arbitration statutes give courts specific and ample means of assuring that private arbitrations will not impact unfairly on judicial decisionmaking, or on third party rights, when a party to an arbitration agreement is also a litigant against a third person in closely related court proceedings involving “a common issue of law or fact.” (Code Civ. Proc., § 1281.2, subd. (c)) [court may deny arbitration, impose forced joinder of some or all parties or issues in a single proceeding, and use stay power to determine whether arbitration or court litigation will proceed first]; cf. Volt Info. Sciences v. Leland Stanford Jr. U. (1989) 489 U.S. 468, 472-479 [109 S.Ct. 1248, 1252-1256, 103 L.Ed.2d 488] [FAA does not preempt Code of Civil Procedure section 1281.2, subdivision (c) where arbitral parties agreed to follow California arbitration law].)
In any event, as we have explained, the fundamental policy of California’s arbitration laws is to ensure that arbitration agreements will be enforced in accordance with their terms. Nothing prevents arbitral parties, aware that their arbitrable dispute bears closely on related *837litigation with others, from agreeing that the arbitrator’s award will bind them in the related case. But the concerns raised by the concurring and dissenting opinion—many speculative in the extreme—do not justify imposing such a consequence in the absence of the arbitral parties’ consent.
Although not a party to the Boyd action, USF&G, in its capacity as Vandenberg’s liability insurer, did participate in the settlement agreement that led to the Boyd/V andenberg arbitration. As part of that agreement, USF&G provided Vandenberg’s legal defense in the arbitration. This does not mean, however, that USF&G was a party to the arbitration itself. USF&G’s legal relationship to the Boyd/V andenberg arbitration was the same as it would have been to a court trial between those parties.
(See Old Republic Ins. Co. v. Superior Court (1998) 66 Cal.App.4th 128, 145-146 [77 Cal.Rptr.2d 642] [“ ‘legally obligated’ ” refers to liability imposed by law]; Wilmington Liquid Bulk Terminals, Inc. v. Somerset Marine Inc. (1997) 53 Cal.App.4th 186, 193 [61 Cal.Rptr.2d 727] (Wilmington) [“ ‘legal liability’ ” refers to tort liability only]; Bernstein v. Consolidated American Ins. Co. (1995) 37 Cal.App.4th 763, 771-772 [43 Cal.Rptr.2d 817] [“legally obligated” in contractual liability endorsement only refers to the contractually assumed tort liability of others]; Fragomeno v. Insurance Co. of the West (1989) 207 Cal.App.3d 822, 828 [255 Cal.Rptr. 111] [“ ‘legally obligated to pay as damages’ ” limits coverage to tort liability only, no indemnification for unlawful detainer against insured]; Insurance Co. of the West v. Haralambos Beverage Co. (1987) 195 Cal.App.3d 1308, 1317 [241 Cal.Rptr. 427] *839(Haralambos), disapproved on other grounds in Buss v. Superior Court (1997) 16 Cal.4th 35 [65 Cal.Rptr.2d 366, 939 P.2d 766] [“ ‘legally obligated to pay as damages’ ” covers tort but not contract liability]; Fireman’s Fund Ins. Co. v. City of Turlock (1985) 170 Cal.App.3d 988, 995 [216 Cal.Rptr. 796] [“ ‘legally obligated to pay as damages’ ” covers only tort liabilities and not contract liabilities]; International Surplus, supra, 93 Cal.App.3d 601, 610-611 [“ ‘legally obligated to pay’ ” refers to liability arising ex delicto].)
The Court of Appeal decisions adhering to the holding of International Surplus are disapproved to the extent they are inconsistent with this opinion. (Old Republic Ins. Co. v. Superior Court, supra, 66 Cal.App.4th 128; Wilmington, supra, 53 Cal.App.4th 186; Bernstein v. Consolidated American Ins. Co., supra, 37 Cal.App.4th 763; Fragomeno v. Insurance Co. of the West, supra, 207 Cal.App.3d 822; Haralambos, supra, 195 Cal.App.3d 1308; Fireman’s Fund Ins. Co. v. City of Turlock, supra, 170 Cal.App.3d 988; International Surplus, supra, 93 Cal.App.3d 601.)