Armstrong World Industries, Inc. v. Aetna Casualty & Surety Co.

45 Cal. App. 4th 1, 52 Cal. Rptr. 2d 690, 96 Daily Journal DAR 5048, 96 Cal. Daily Op. Serv. 3058, 1996 Cal. App. LEXIS 402

• Court: California Court of Appeal
• Decided: April 30, 1996
• Docket: Docket Nos. A049419, A049631, A049654, A049659, A049661, A049663, A049664, A049665, A049666, A049667, A049668, A049669, A049670, A049671, A049672, A049808, A049875
• Status: Published

Opinion

Opinion

DOSSEE, J.

This appeal raises a number of complex questions concerning insurance coverage for claims of asbestos-related bodily injuries and property damage. In the proceedings below, separate declaratory relief actions and related cross-actions involving three asbestos manufacturers—Armstrong World Industries, Inc., Fibreboard Corporation, and GAF Corporation —and their various insurance carriers were coordinated and tried in six separate phases over a five-year period. 1

On appeal, the parties submitted briefs on three major “Issue Groups,” and our opinion follows that organization. First, in the unpublished portion of the opinion, we discuss the issues of Issue Group I pertaining to a lost insurance policy. In Issue Group II we discuss the issues concerning the bodily injury claims: trigger and scope of coverage; the application of the phrase “neither expected nor intended”; the liability of premerger insurers; the effect of the Wellington Agreement. In Issue Group III, we discuss the issues surrounding the property damage claims: coverage for property damage; trigger and scope of coverage; the duties to defend and indemnify; and, in the unpublished portion of the opinion, the “drop-down" obligation of an INA-Armstrong excess policy.

After this appeal was submitted for decision, we granted a motion of certain parties to sever issues unique to them in order to facilitate a pending settlement. Accordingly, we have deferred decision upon issues pertaining to a lost Fibreboard-Pacific Indemnity insurance policy; the number of occurrences; the effect of the Fibreboard-Continental manuscript policy; and the application of the pollution exclusion clause.

Our previous opinion, filed on November 15, 1993, was vacated by the Supreme Court, and the matter was remanded to us for reconsideration in light of Montrose Chemical Corp. v. Admiral Ins. Co. (1995) 10 Cal.4th 645 [42 Cal.Rptr.2d 324, 897 P.2d 1].

Guiding Principles

At the outset, we set forth the principles guiding our review, Interpretation of an insurance policy is primarily a judicial function. When the trial court’s interpretation did not depend upon conflicting extrinsic evidence, the reviewing court makes its own independent determination of the policy’s meaning. (Masonite Corp. v. Great American Surplus Lines Ins. Co. (1990) 224 Cal.App.3d 912, 916 [274 Cal.Rptr. 206].)

In interpreting an insurance contract, the court’s fundamental goal is to give effect to the mutual intention of the parties. Such intent is inferred, if possible, solely from the written provisions of the contract. (AIU Ins. Co. v. Superior Court (1990) 51 Cal.3d 807, 821-822 [274 Cal.Rptr. 820, 799 P.2d 1253].) “If contractual language is clear and explicit, it governs.” (Bank of the West v. Superior Court (1992) 2 Cal.4th 1254, 1264 [10 Cal.Rptr.2d 538, 833 P.2d 545].) Words in an insurance policy are to be interpreted as a layperson would interpret them, in their “ ‘ordinary and popular sense.’ ” (AIU, supra, 51 Cal.3d at p. 822; Reserve Insurance Co. v. Pisciotta (1982) 30 Cal.3d 800, 807 [180 Cal.Rptr. 628, 640 P.2d 764].) A policy should not be read as it might be analyzed by an attorney or an insurance expert. (Delgado v. Heritage Life Ins. Co. (1984) 157 Cal.App.3d 262, 271 [203 Cal.Rptr. 672].) This is so even if the policyholder is a sophisticated insured. (AIU, supra, 51 Cal.3d at p. 823.)

If particular policy language is ambiguous, it is to be resolved by interpreting the ambiguous provisions in accordance with the insured’s objectively reasonable expectations. (Bank of the West v. Superior Court, supra, 2 Cal.4th at pp. 1264-1265.) Only if application of this rule does not resolve the ambiguity will the policy provision be construed in favor of the insured. (Id. at p. 1265.)

Issue Group I: Lost Insurance Policy *

Issue Group II: Bodily Injury Claims

A. Trigger and Scope of Coverage

Phase III of the coordinated proceedings below concerned the rights and obligations of insurers to indemnify and defend the manufacturers or distributors of asbestos or asbestos products that are, or have been, defendants in tens of thousands of lawsuits brought by persons who claim to have developed disabling and often fatal asbestos-related diseases as a result of exposure to asbestos products many years ago. It bears emphasizing that the issues do not pertain to the legal rights of those suffering from asbestos-related diseases to recover damages from asbestos manufacturers.

The principal issues before the trial court concerned the trigger and scope of coverage under the comprehensive general liability policies for asbestos-related bodily injury claims: What event triggers an insurer’s indemnification and defense obligations? And to what extent must policyholders share in the indemnity and defense costs?

In order to resolve these issues, the trial court heard extensive medical testimony and took documentary evidence concerning the pathogenesis of asbestos-related conditions. The trial court artfully described the insidious nature of asbestos: “Asbestos is a naturally occurring mineral which has long been known to man. Its principal use has been as an insulator against heat because it is incombustible in air. It has been used to insulate against heat since approximately 1866 and has been commercially produced since at least 1874. [Citation.] The health problem caused by asbestos is that when it is mined or used in the manufacturing process it produces quantities of asbestos dust composed of millions of tiny fibers which may be inhaled into the body by those working in and around it. Those fibers that avoid the body’s initial natural defense mechanisms are deposited in the human lung and remain there. The very quality that has made asbestos useful for so long, its indestructibility, also accounts for the problems that result in asbestos-related disease.”

The Medical Evidence

We adopt the trial court’s summary of the medical evidence: “Several diseases may result from exposure to asbestos. The most prevalent are asbestosis, bronchogenic carcinoma, and mesothelioma. Asbestosis is a form of lung disease characterized by the permanent deposition of asbestos fibers in the lungs and the resultant scarring of the lungs’ alveoli (air sacs) and interstitium (the membrane through which gas exchange occurs between the alveoli and the blood). In the context of asbestos inhalation, bronchogenic carcinoma (lung cancer) refers to a malignant condition of cells which arises as the result of tissue scarring caused by asbestos. Mesothelioma is, similarly, a cancerous condition. It arises at the site of asbestos-caused scarring within the visceral pleura (the lining which covers the outer aspect of the lung) or the peritoneum (the lining of the abdominal cavity).

“While the disease processes are distinct, they share at least one characteristic which makes this Court’s interpretation of the policy language universally applicable to these diseases, as well as to other conditions which may arise from inhaling asbestos. That common element is that the diseases and the associated pathological processes occur because of the fibrosis induced by the inhaled asbestos.

“Fibrosis refers to the formation of fibrous tissue, and is more commonly called scarring. When associated with an external cut to the skin, fibrosis may be considered a necessary and helpful form of healing which restores the body to a functional—albeit altered—state. When associated with the inhalation of asbestos, however, fibrosis results in the impairment and destruction of the alveolar/capillary gas exchange units necessary to breathe. As such, and because of the irreversible nature of the fibrotic process on the lung tissue, fibrosis caused by the inhalation of asbestos is more appropriately characterized as a form of injury than of healing or repair.

“Fibrosis within the lungs occurs as part of the body’s reaction to the inhalation of foreign particulate matter. The indestructible nature of asbestos fibers which helped make asbestos such an attractive construction material makes it equally as detrimental to the body once inhaled. Once deposited in the lungs, the fibers tend to remain in the alveolar region and the lungs’ normal clearance mechanisms are ineffective.

“One clearance mechanism—and a key to the fibrotic process—involves a specialized form of white blood cell known as a macrophage. These cells naturally respond to foreign matter within the body and attempt to eliminate this matter from the body by engulfing (i.e., phagocytozing) and digesting the matter with their own secretions and enzymes. This process occurs on the cellular level, but is frustrated and unsuccessful in the context of asbestos fibers because of the macrophages’ inability effectively to engulf and digest the fibers.

“This, in turn, leads to a further and sustained inflammatory process. The inflammation becomes chronic as more macrophages and other white blood cells are attracted to the site of the asbestos fibers caused by the release of certain chemical substances by the macrophages which responded initially to the fibers. More macrophages are summoned, further frustrated phagocytosis occurs, and the cycle continues.

“Another result of the inflammation is that other cells, called fibroblasts, are summoned to the site of inflammation by a different chemical secretion (fibronectin) from the macrophages. Fibronectin not only attracts these fibroblasts, but also causes them to proliferate. The fibroblasts, once summoned, produce the collagen in the alveolar walls and the interstitium which constitutes fibrosis.

“This process—inhalation of asbestos fibers, the inflammatory reaction, and the resulting fibrosis—characterizes the disease asbestosis. When the fibrosis is extensive enough, i.e., when enough alveolar/capillary units have become fibrosed, clinical symptoms of asbestosis become apparent. Although there is no universal threshold for when such symptoms will become apparent, it is estimated that at least 100 million of the 300 million alveolar/ capillary units in the human body must be affected for a clinical diagnosis to occur.

“Bronchogenic carcinoma and mesothelioma arise from a malignant transformation of cells. The asbestos fibers and related fibrosis do not directly cause the malignant transformation but, rather, enhance the potential of other cancerous agents to cause such a transformation. The transformation occurs at the site of the fibrosis and the cancer develops therefrom.”

1. Trigger of Coverage

The relevant language of the standard form comprehensive general liability (CGL) policy reads as follows: “The company will pay on behalf of the insured all sums which the insured shall become legally obligated to pay as damages because of bodily injury . . . caused by an occurrence . . . . [5D ‘Bodily injury’ means bodily injury, sickness, or disease sustained by any person. HQ ‘Occurrence’ means an accident, including injurious exposure to conditions, which results during the policy period in bodily injury . . . neither expected nor intended from the standpoint of the insured.” 8

A recurring problem in interpreting standard CGL policies that provide coverage for injuries “caused by an occurrence” is determining what has come to be called the “trigger of coverage”—that is, the operative event which activates the insurer’s defense and indemnity obligations. As the Supreme Court recently explained, the word “trigger” is not found in the CGL policies themselves, nor does the Insurance Code define “trigger of coverage.” Instead, “trigger of coverage” is a term of convenience used to describe what must happen in the policy period to give rise to insurance coverage. (Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, 655, fn. 2.)

Case law has long established that the operative event triggering coverage is the injury. Because occurrence policies (as distinguished from claims-made policies) cover occurrences that result in injury “during the policy period,” the courts in California and elsewhere have concluded that the policies are invoked, or “triggered,” when the injury takes place. (American Cyanamid Co. v. American Home Assurance Co. (1994) 30 Cal.App.4th 969, 979 [35 Cal.Rptr.2d 920]; Hallmark Ins. Co. v. Superior Court (1988) 201 Cal.App.3d 1014, 1017 [247 Cal.Rptr. 638]; State Farm Mut. Auto. Ins. Co. v. Longden (1987) 197 Cal.App.3d 226, 231 [242 Cal.Rptr. 726]; Schrillo Co. v. Hartford Accident & Indemnity Co. (1986) 181 Cal.App.3d 766, 773 [226 Cal.Rptr. 717]; Atlantic Mutual Ins. Co. v. Travelers Ins. Co. (1983) 147 Cal.App.3d 1054, 1056 [195 Cal.Rptr. 476]; Maples v. Aetna Cas. & Surety Co. (1978) 83 Cal.App.3d 641, 647 [148 Cal.Rptr. 80]; Remmer v. Glens Falls Indem. Co. (1956) 140 Cal.App.2d 84, 88 [295 P.2d 19]; see also Employers Casualty Co. v. Northwestern Nat. Ins. Group (1980) 109 Cal.App.3d 462, 468-469 [167 Cal.Rptr. 296] disapproved on other grounds in In re Marriage of Arceneaux (1990) 51 Cal.3d 1130, 1137 [275 Cal.Rptr. 797, 800 P.2d 1227]; Chamberlin v. Smith (1977) 72 Cal.App.3d 835 [140 Cal.Rptr. 493]; 7A Appleman, Insurance Law & Practice (rev. ed. 1979) § 4501.03, p. 256; 11 Couch on Insurance (2d ed. 1982) § 44:8, p. 193; 43 Am.Jur.2d (1982 rev.) Insurance, § 243, pp. 323-324; Annot. (1985) 37 A.L.R.4th 382.) 9

In Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, 669-670, the Supreme Court reaffirmed this rule and applied it for the first time to a case of continuous or progressively deteriorating injury. The court held that when the bodily injury or property damage continues throughout successive policy periods, all of the insured’s policies in effect during those periods are triggered. (Id. at pp. 685-689.) Coverage is not limited to the policy in effect at the time of the precipitating event or conditions. (Id. at pp. 669, 686.) Nor is coverage cut off once the injury or damage begins or becomes manifest. (See 10 Cal.4th at p. 677, fn. 17; id., at pp. 680, 686.)

Montrose involved property damage and human deaths resulting from the insured’s disposal of toxic or hazardous wastes. The time of the onset of the damage and injury was not in dispute; the underlying complaints specified, in one set of lawsuits, that the bodily injury and property damage commenced in 1956 and extended to the present, and in a second set of lawsuits, that the property contamination began in 1947 and continued throughout periods (1982-1986) when the policies of the insurer (Admiral) were in effect. (Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at pp. 656-657, 659.)

In contrast, the timing of the commencement of the injuries here is not so definite. As the trial court explained, asbestos-related diseases are “insidious diseases with delayed manifestations. The original cause of each disease is the inhalation of asbestos fibers, but a victim is generally unaware of the development of such a disease until the victim or a physician detects signs or symptoms many years after the causative exposure.” Thus, the key question before the trial court in phase III with regard to the trigger of coverage was the point in time at which the injury takes place.

a. Precedent Cases

The courts have developed several different approaches to determine when bodily injury occurs in asbestos-related bodily injury cases. 10 (See Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, 673-685; Annot. (1993) 14 A.L.R.5th 695.)

1. Under the exposure theory, bodily injury is deemed to commence upon the claimant’s first exposure to asbestos, upon the claimant’s initial inhalation of asbestos fibers. (E.g., Ins. Co. North America v. Forty-Eight Insulations (6th Cir. 1980) 633 F.2d 1212, 1218-1220, clarified 657 F.2d 814, cert. den. (1981) 454 U.S. 1109 [70 L.Ed.2d 650, 102 S.Ct. 686]; Commercial Union Ins. Co. v. Sepco Corp. (11th Cir. 1985) 765 F.2d 1543; Porter v. American Optical Corp. (5th Cir. 1981) 641 F.2d 1128, cert. den. sub nom. Aetna Cas. & Surety Co. v. Porter (1981) 454 U.S. 1109 [70 L.Ed.2d 650, 102 S.Ct. 686]; Cole v. Celotex Corp. (La. 1992) 599 So.2d 1058, 1076-1077.)

2. Pursuant to the manifestation theory, no bodily injury occurs, and thus no insurance coverage is triggered, until the “asbestos-related disease became reasonably capable of medical diagnosis.” (Eagle-Picher Industries, Inc. v. Liberty Mut. Ins. (1st Cir. 1982) 682 F.2d 12, 25, cert. den. (1983) 460 U.S. 1028 [75 L.Ed.2d 500, 103 S.Ct. 1279].) In adopting the manifestation theory, the Eagle-Picher court reasoned that the language of the policies distinguishes between the event which causes injury—the accident or exposure—and the resulting injury or disease. “[I]t is the resulting injury, not the exposure, which must take place ‘during the policy period’ in order to trigger coverage . . . .” (682 F.2d at p. 19.)

3. In Keene Corp. v. Ins. Co. of North America (D.C. Cir. 1981) 667 F.2d 1034 [215 App.D.C. 156], certiorari denied (1982) 455 U.S. 1007 [71 L.Ed.2d 875, 102 S.Ct. 1644], the court adopted a theory of continuous trigger or “triple triggers” whereby the asbestos injury is deemed a continuous process and all policies are triggered on a claim if they were in effect either during the exposure period, or at the time of manifestation, or at any time in between (the latency or “exposure-in-residence” period). 11 (See also ACandS, Inc. v. Aetna Cas. and Sur. Co. (3d Cir. 1985) 764 F.2d 968.)

4. Under the injury-in-fact rule, coverage is triggered when the actual injury is shown, retroactively, to have occurred. (Abex Corp. v. Maryland Cas. Co. (D.C. Cir. 1986) 790 F.2d 119 [252 App.D.C. 297] [asbestos]; American Home Products Corp. v. Liberty Mut. Ins. (2d Cir. 1984) 748 F.2d 760 [pharmaceuticals]; Maryland Cas. Co. v. W.R. Grace & Co. (S.D.N.Y. 1991) 794 F.Supp. 1206, 1215 [asbestos]; Aetna Cas. & Sur. Co. v. Abbott Lab., Inc. (D.Conn. 1986) 636 F.Supp. 546, 548-550 [DES].)

Like the manifestation theory, the injury-in-fact approach holds that mere exposure to asbestos during the policy period is not enough to trigger coverage: “The plain language of the definition of ‘occurrence’ used in the CGL policy requires exposure that ‘results, during the policy period, in bodily injury’ in order for an insurer to be obligated to indemnify the insured. The unambiguous meaning of these words is that an injury—and not mere exposure—must result during the policy period. The CGL policies expressly distinguish exposure from injury; to equate the two ... is to ignore this distinction. Any argument that mere exposure—without injury— triggers liability is simply unsound linguistically.” (Abex Corp. v. Maryland Cas. Co., supra, 790 F.2d at p. 127, italics in original; see also American Home Products Corp. v. Liberty Mut. Ins., supra, 748 F.2d at p. 764.)

But in contrast to the manifestation trigger, the injury-in-fact approach acknowledges that injury may occur before the injury has become apparent. Under this approach, coverage is triggered by ‘“a real but undiscovered injury, proved in retrospect to have existed at the relevant time . . . irrespective of the time the injury became [diagnosable].’ ” (American Home Products Corp. v. Liberty Mut. Ins., supra, 748 F.2d at p. 766.) That is, after an injury has been diagnosed, it may be inferred, from the nature of the gestation period and from the stage of the illness, that the harm actually began sometime earlier. (Id. at p. 765.)

As mentioned above, the California Supreme Court, after reviewing the various judicially recognized triggers, has concluded that a continuous trigger should be applied to claims of continuous or progressively detriorating damage or injury. (Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645.) Yet, the Montrose court explicitly recognized that an injury-in-fact analysis is not inconsistent with a continuous trigger: “In the context of continuous or progressively deteriorating injuries, the injury-in-fact trigger, like the continuous injury trigger, affords coverage for continuing or progressive injuries occurring during successive policy periods subsequent to the established date of the initial injury-in-fact.” (Id. at p. 676.) That is, the continuous trigger pertains to the duration of coverage, providing coverage throughout successive policy periods. The injury-in-fact trigger establishes the onset of the injury, to determine when coverage begins. (See Stonewall Ins. Co. v. Asbestos Claims Management (2d Cir. 1995) 73 F.3d 1178, 1194-1197.)

b. The Trial Court’s Decision

Based upon the extensive medical evidence, the trial court found “that bodily injury occurs during the exposure period, that it continues to occur during the latency period, even in the absence of further exposure, and that it continues to occur past the manifestation point, accompanied by sickness and disease, until the claimant’s death from the disease or other causes.” Accordingly, the court adopted a continuous trigger: “[A]ll of a policyholder’s policies in effect from first exposure to asbestos or asbestos-containing products until date of death or date of claim, whichever occurs first, are triggered with respect to an asbestos-related bodily injury claim.” 12

Yet, although the trial court concluded that a continuous trigger should apply, the court reached that conclusion through an injury-in-fact analysis. Unlike the court in Keene, supra, which deemed asbestos injury to be continuous, the trial court here relied upon medical evidence to make factual findings on the physiological processes that actually occur upon inhalation of asbestos fibers and continue until death. 13

In contrasting its decision with the decision in Zurich Ins. Co. v. Raymark Industries, supra, 118 Ill.2d 23 [514 N.E.2d 150], the trial court noted that the Zurich trial court had “concluded from the medical evidence that injury does not always occur in the absence of exposure. (See id., 514 N.E.2d at pp. 160-161.) This conclusion differs from that reached by this Court.”

In rejecting the “exposure theory,” the trial court found as follows: “[Although this Court agrees with the exposure theorists that ‘bodily injury’ occurs nearly simultaneously with inhalation, and therefore throughout the exposure period, this Court finds that new and additional ‘bodily injury’ continues to occur even past the cessation of exposure . . .

Further, in comparing its decision with that of American Home Prod. v. Liberty Mut. Ins. Co (S.D.N.Y. 1983) 565 F.Supp. 1485 affirmed as modified (2d Cir. 1984) 748 F.2d 760 (AHP), the trial court noted that the AHP court had declined to make a general declaration that every exposure to any of the drugs at issue causes injury and therefore triggers coverage. But the trial court found it could do so with respect to asbestos: “This Court, however, has received evidence which supports such a ‘general declaration’ as to when injury occurs, and has applied it generally to all claimants who suffer from asbestos-related ‘bodily injury.’ ” The trial court explained that “[t]he AHP decision, as modified, provides that coverage is triggered by ‘a real but undiscovered injury, proved in retrospect to have existed at the relevant time.’ (Supra, 748 F.2d at p. 766.) This Court simply proceeds one step further in its analysis, and applies the ‘injury in fact’ concept to the asbestos medical evidence, thereby establishing in retrospect that undiscovered injury existed during the asbestos exposure period and during the latency period in the absence of exposure.”

We find no error in the trial court’s use of an injury-in-fact analysis to apply a continuous trigger. (Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d at pp. 1195-1196.) Indeed, we note that in Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at pp. 676-677, fn. 16, the Supreme Court acknowledged that claims involving asbestos-related diseases involve “unique facts” and the injury-irnfact trigger may be appropriate. Moreover, the court referred to our earlier (now vacated) opinion and observed that our affirmance of the trial court’s decision “appears largely consistent with [the Montrose] analysis of the applicable principles of third party CGL coverage . . . .” (Ibid.)

c. Arguments of Insurers

The insurers raise two principal arguments against the trial court’s trigger decision. First, the insurers dispute the trial court’s interpretation of “bodily injury.” Second, the insurers argue that the trial court’s continuous trigger decision improperly holds insurers liable even though the claimant had no contact with the policyholder’s products during the policy period. We disagree with the first argument but find some merit in the second, as we will explain below.

(1) Subclinical Changes

The insurers contend the trial court’s interpretation of bodily injury is contrary to the plain meaning of the words in that it equates bodily injury with “imperceptible subclinical cellular changes.” Although the insurers have now disavowed the manifestation theory, the lone support for the insurers’ argument is Eagle-Picher Industries, Inc. v. Liberty Mut. Ins., supra, 682 F.2d 12, where the court held that “sub-clinical insults to the lungs” do not constitute an injury “until, if ever, they accumulate to become clinically evident or manifest.” 14

In Eagle-Picker, the court refused to adopt the exposure theory, noting that “it is uncontested that even sub-clinical injury to the lung does not occur simultaneously with the inhalation of asbestos. Nor is the existence of sub-clinical injury an inevitable by-product of exposure, since the body’s natural mechanisms may remove the fibers before they become embedded in the lungs.” (682 F.2d at p. 19, fn. omitted.) Likewise, in the present case, the insurers emphasize the medical evidence and the trial court’s finding that not every exposure to asbestos results in an asbestos-related injury.

Yet, while it may be that not every inhalation of asbestos fibers results in bodily injury, it can be said that every manifested asbestos-related injury resulted from inhalation of asbestos fibers. (See Commercial Union Ins. Co. v. Sepco Corp., supra, 765 F.2d at pp. 1545-1546.) In the present case, the trial court necessarily took a retrospective point of view. In resolving the insurance coverage questions, the court was concerned only with individuals who have actually developed asbestos-related diseases, and for such claimants the court found that the evidence permitted the inference that injury took place in the past: “[T]he asbestos medical evidence [establishes] in retrospect that undiscovered injury existed during the asbestos exposure period and during the latency period in the absence of exposure.” (Italics added.)

The trial court’s continuous trigger decision, then, is based upon factual findings that for asbestos claimants an injury-in-fact took place during each triggered policy period, even though the injury was not diagnosable and compensable during the policy period. The trial court found it “sufficient that such injuries eventually became compensable, and this is, of course, true with respect to all claims for which insurers are called upon to indemnify policyholders.” We find no error in this retrospective approach. For purposes of determining insurance coverage, absolute precision is not required as to when the injury occurred. “[A] 11 that is necessary is reasonably reliable evidence that the injury, sickness, or disease more likely than not occurred during a period of coverage . . . .” (AHP, supra, 565 F.Supp. at p. 1509; accord, Abex Corp. v. Maryland Cas. Co., supra, 790 F.2d at p. 128.)

The insurers’ argument assumes that an injury does not occur until there is an impairment capable of detection. In the present case, however, the medical evidence established and the trial court found that impairment actually occurs even earlier: “[T]he physiological processes associated with the inhalation of asbestos . . . impair the gas exchange function of the lung cells and tissue. . . . almost immediately upon the inhalation and deposition of asbestos fibers into the lung, and slowly and continuously impair new portions of lung tissue throughout one’s life, even after exposure to asbestos ceases, . . . [involving] new injury to the cells, tissue and body.” Those factual findings, of course, are binding on this court.

It bears emphasizing that whether there is coverage for “bodily injury” is not the question here; it is undisputed that the CGL policies provide coverage for the asbestos-related injuries suffered by the claimants. The question before this court is when the injuries occurred. In contrast to the situation in Eagle-Picher, supra, where the court found no factual basis for the conclusion that bodily injury occurs upon exposure (Eagle-Picher, supra, 682 F.2d at p. 19, fn. 3), the trial court’s factual findings here, made after consideration of extensive medical testimony, amply support the conclusion that injury actually occurs upon exposure and continues until death. We therefore find no error in the trial court’s adoption of a continuous trigger. (See also Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d 1178, 1196-1197; Owens-Illinois, Inc. v. United Ins. Co. (1994) 138 NJ. 437 [650 A.2d 974, 995]; J.H. France Refractories v. Allstate (1993) 534 Pa. 29, 626 A.2d 502, 506-507].)

In any event, the insurers’ approach would essentially render the asbestos manufacturers’ insurance coverage illusory, for by the time asbestos diseases caused detectable impairments (in the 1970’s), insurance companies ceased issuing policies that adequately covered asbestos-related disease. Hence, the insurers’ theory would deprive the manufacturers of coverage for product liability injuries of which they were unaware during the policy periods. (E.g., Hancock Laboratories, Inc. v. Admiral Ins. (9th Cir. 1985) 777 F.2d 520, 525; Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d at pp. 1045-1046; Ins. Co. North America v. Forty-Eight Insulations, supra, 633 F.2d at p. 1219.)

Moreover, there is nothing in the language of the policies to require as a condition of coverage that the injury be discovered at any point in time. As the Supreme Court recognized in Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at pp. 688-689, to read the CGL occurrence policies to provide coverage only when the injury becomes apparent during the policy period would unfairly transform the policies into “claims-made” policies. (See also American Home Products Corp. v. Liberty Mut. Ins., supra, 748 F.2d at p. 764; Ins. Co. North America v. Forty-Eight Insulations, supra, 633 F.2d at p. 1219; Hartford County v. Hartford Mut. Ins. (1992) 327 Md. 418 [610 A.2d 286, 294-295].)

In short, we find the trial court’s continuous trigger decision well supported both by the unique facts of asbestos-related bodily injuries and by the existing case law. We uphold that decision.

(2) Contact With Policyholder’s Product

The trial court’s judgment states that all of a policyholder’s policies are triggered from the claimant’s first exposure to any asbestos product until the date of death or claim. We agree with the insurers that this aspect of the trigger decision is overbroad. Our analysis of this point is intertwined with the analysis of scope of coverage and will be presented in part 2.b below.

2. Scope of Coverage

Under the trial court’s continuous trigger decision, multiple, successive policies of a policyholder are likely to be triggered on any single bodily injury claim. Two questions emerge concerning the extent or “scope” of coverage of these multiple insurers: Should the responsibility for indemnification be apportioned among the insurers? And should the policyholder be required to share in the indemnification and defense costs if the policyholder was uninsured or self-insured for certain periods? To facilitate analysis of these issues, we draw a distinction between (1) the obligations of successive insurance carriers toward a single manufacturer-policyholder, and (2) the obligations of successive carriers when multiple asbestos manufacturers are held liable on a single claim.

a. Obligations of Successive Insurers of a Single Asbestos Manufacturer-Policyholder

(1) Apportionment Among Insurers

The standard CGL insurance policies require the insurers to indemnify the policyholder only if bodily injury occurs during the policy period: “The company will pay on behalf of the insured all sums which the insured shall become legally obligated to pay as damages because of bodily injury . . . caused by an occurrence ....[*][] ‘Occurrence’ means an accident, including injurious exposure to conditions, which results during the policy period in bodily injury . . . .”

Thus, the question raised by the insurance companies is whether liability should be apportioned among the insurers based on their periods of coverage. As will be seen in the discussion below, the trial court ruled that the policyholder must be indemnified by one insurer for the full extent of the loss up to the policy’s limits, but with liability ultimately being apportioned among all insurers based upon the policy limits and the years of coverage. We affirm that decision.

Liable “in Full”

In phase III, the trial court concluded that each policy triggered by an asbestos-related bodily injury claim has an independent obligation to respond “in full” to a claim. In reaching that conclusion, the trial court relied primarily upon the insurers’ obligations under the CGL policies to pay for “all sums which the insured shall become liable to pay as damages.”

The trial court’s decision follows several out-of-state asbestos cases. In Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d 1034, the court held as follows: “The policies at issue in this case provide that the insurance company will pay on behalf of Keene ‘all sums’ that Keene becomes legally obligated to pay as damages because of bodily injury during the policy period. . . . As a result [of our continuous trigger decision], when Keene is held liable for an asbestos-related disease, only part of that disease will have developed during any single policy period. The rest of the development may have occurred during another policy period or during a period in which Keene had no insurance. The issue that arises is whether an insurer is liable in full, or in part, for Keene’s liability once coverage is triggered. We conclude that the insurer is liable in full, subject to the ‘other insurance’ provisions . . . .” (Id. at p. 1047.) “Once triggered, each policy covers Keene’s liability. There is nothing in the policies for a reduction of the insurer’s liability if an injury occurs only in part during a policy period. As we interpret the policies, they cover Keene’s entire liability once they are triggered.” (Id. at p. 1048; see also ACandS, Inc. v. Aetna Cas. and Sur. Co., supra, 764 F.2d at p. 974; Zurich Ins. Co. v. Raymark Industries, supra, 514 N.E.2d at p. 165; Monsanto Co. v. C.E. Heath Comp. & Liability (Del. 1994) 652 A.2d 30, 34-35; J.H. France Refractories v. Allstate, supra, 626 A.2d 502, 507-508.) 15

We believe the California Supreme Court’s decision in Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, supports this reasoning. In distinguishing third party liability policies from first party liability policies, the court observed that under third party liability policies, if coverage is ultimately established, the insurer must indemnify the insured for “all sums” which the insured becomes obligated to pay. (10 Cal.4th at p. 665.) Moreover, in concluding that a continuous trigger should be applied, the court on two occasions cited with approval the opinion in Gruol Construction Co. v. Insurance Co. of No. America (1974) 11 Wn.App. 632 [524 P.2d 427], in which the Washington Court of Appeal applied a continuous trigger in a case involving progressive property damage. (10 Cal.4th at pp. 677-678, 681.) In both references to the Gruol case, the Montrose court observed that under a continuing injury theory, an insurer may become liable for the entire loss up to the policy limits even though the continuing injury may extend over several policy periods. (Id. at pp. 678, 681.)

Furthermore, in support of its conclusion that a continuous trigger should be applied, the Montrose court relied upon existing case law holding that coverage for a manifested loss is not terminated by the expiration of the policy; coverage continues until the damage is complete. (Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at pp. 680, 686, citing California Union Ins. Co. v. Landmark Ins. Co. (1983) 145 Cal.App.3d 462, 475 [193 Cal.Rptr. 461]; Snapp v. State Farm Fire & Cas. Co. (1962) 206 Cal.App.2d 827, 831-832 [24 Cal.Rptr. 44]; and Harman v. American Casualty Co. of Reading, Pa. (S.D.Cal. 1957) 155 F.Supp. 612.) As the Montrose court put it, “an insurer on the risk when continuous or progressively deteriorating damage or injury first manifests itself remains obligated to indemnify the insured for the entirety of the ensuing damage or injury.” (10 Cal.4th at p. 686, italics added.)

Apportionment

At the same time that the trial court ruled each insurer must respond “in full,” the court also ruled that “the obligation to respond in full is subject to the operation of policy limits, deductibles, applicable exclusions, applicable ‘other insurance’ clauses, provisions which make certain policies’ coverage ‘excess’ to that of other policies, and any rights to equitable contribution from the issuers of other policies triggered by the same claim.”

That decision, too, is consistent with language in the Montrose case: “Allocation of the cost of indemnification once several insurers have been found liable to indemnify the insured for all or some portion of a continuing injury or progressively deteriorating property damage requires application of principles of contract law to the express terms and limitations of the various policies of insurance on the risk. [Citing Keene and Forty-Eight Insulations,.]” (Mon trose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at p. 681, fn. 19.)

In Keene, the court held that liability among insurers must be allocated pursuant to the “other insurance” clauses: “In any suit against Keene for an asbestos-related disease, it is likely that the coverage of more than one insurer will be triggered. Because each insurer is fully liable, and because Keene cannot collect more than it owes in damages, the issue of dividing insurance obligations arises. The only logical resolution of this issue is for Keene to be able to collect from any insurer whose coverage is triggered, the full amount of indemnity that it is due, subject only to the provisions in the policies that govern the allocation of liability when more than one policy covers an injury. . . . Our holding each insurer fully liable to Keene [ ] does not mean that a single insurer will be saddled with full liability for any injury. When more than one policy applies to a loss, the ‘other insurance’ provisions of each policy provide a scheme by which the insurers’ liability is to be apportioned. . . . These provisions of the policies must govern the allocation of liability among the insurers in any particular case of asbestos-related disease. However, the primary duty of the insurers whose coverage is triggered by exposure or manifestation is to ensure that Keene is indemnified in full.” (667 F.2d at p. 1050, fn. omitted.)

The Keene court did not specify how the “other insurance” clauses would serve to allocate liability among the insurers. In the present case, however, in phase IV, the trial court concluded that the presence of “other insurance” clauses in the policies had the effect of requiring a pro rata apportionment among multiple insurers whose policies were triggered successively on the same claim. The court employed an apportionment method based on the respective policy limits multiplied by the years of coverage: “When more than one policy is triggered by a claim, defense and indemnity costs shall be allocated among all triggered policies according to applicable ‘per occurrence’ policy limits, multiplied by years of coverage. When a policy does not contain a ‘per occurrence’ limit, the ‘per person’ limit shall be used in this calculation.

“This Court finds that the most equitable method of allocation is proration on the basis of policy limits, multiplied by years of coverage. This method is consistent with the policy language in that it takes policy limits into consideration. Typically, a pro rata ‘other insurance’ clause provides for proration according to ‘the applicable limit of liability.’ This method also reflects the fact that higher premiums are generally paid for higher ‘per person’ or ‘per occurrence’ limits. Since some policies are in effect for more than one year, and injury occurs during every year from first exposure to asbestos until death (phase III Decision at p. 42), multiplying the policy limits by years of coverage results in a more equitable allocation than proration based on policy limits alone. Thus, when a particular claim triggers more than one policy, each insurer’s share of liability shall be determined by the proportion that each policy’s applicable ‘per occurrence’ limits multiplied by years the policy was in effect bears to the sum total of the applicable ‘per occurrence’ limits of all triggered policies multiplied by the years each policy was in effect. When a policy does not contain a ‘per occurrence’ limit, the ‘per person’ limit shall be used in this calculation.”

This allocation procedure does not affect the obligation of the insurers to respond in full: “a policyholder may obtain full indemnification and defense from one insurer, leaving the targeted insurer to seek contribution from other insurers covering the same loss.”

The trial court’s ruling on the method of apportionment is not challenged on appeal. We note that although the method is nontraditional, it is nonetheless sound. The general rule, when multiple policies share the same risk but have inconsistent “other insurance” clauses, is to prorate according to the policy limits. (See Argonaut Ins. Co. v. Transport Indent. Co. (1972) 6 Cal.3d 496, 507 [99 Cal.Rptr. 617, 492 P.2d 673]; Employers Reinsurance Corp. v. Phoenix Ins. Co. (1986) 186 Cal.App.3d 545, 557 [230 Cal.Rptr. 792]; CNA Casualty of California v. Seaboard Surety Co. (1986) 176 Cal.App.3d 598, 620 [222 Cal.Rptr. 276].) Courts in other jurisdictions have taken different approaches. Most prominent among the alternatives is an allocation based upon time on the risk—i.e., the number of years an insurer

covered the continuous loss. (E.g., Ins. Co. North America v. Forty-Eight Insulations, supra, 633 F.2d 1212.) 16

The apportionment formula used by the trial court in the present case—combining the policy limit formula with the time on the risk approach—was advocated by some insurers in CNA Casualty of California v. Seaboard Surety Co., supra, 176 Cal.App.3d at pages 619-620, with respect to defense costs. The court rejected the argument and used a straight policy limit approach. But in doing so the court noted that the Supreme Court had declined to formulate a definitive rule “in light of varying equitable considerations which may arise” in particular cases. (Signal Companies, Inc. v. Harbor Ins. Co. (1980) 27 Cal.3d 359, 369 [165 Cal.Rptr. 799, 612 P.2d 889, 19 A.L.R.4th 75].) Quoting from an earlier case, the Supreme Court explained the need for “equitable” considerations: “ ‘The reciprocal rights and duties of several insurers who have covered the same event do not arise out of contract, for their agreements are not with each other .... Their respective obligations flow from equitable principles designed to accomplish ultimate justice in the bearing of a specific burden. . . (27 Cal.3d at p. 369.) The CNA court, therefore, acknowledged that in an appropriate case the scope of an insured’s coverage could be affected by such factors as the insurer’s time on the risk. (176 Cal.App.3d at p. 620.) And, indeed, in the present case, the trial court found its method of allocation, based upon both the policy limits and the time on the risk, to be the “most equitable.”

Given that the trial court’s method of apportionment is not challenged on appeal, we find no error in the decision to hold each policy responsible in full subject to such apportionment. 17 (See also Owens-Illinois, Inc. v. United Ins. Co., supra, 650 A.2d 974, 993-995, proposing the same method of allocation.) We are not persuaded otherwise by Ins. Co. North America v. Forty-Eight Insulations, supra, 633 F.2d 1212, in which the trial court employed an exposure trigger and prorated liability among the multiple, successive insurers who were on the risk while the claimant was exposed to asbestos. On appeal, no question was raised concerning this proration; the dispute focused on prorating the costs of defense. The appellate court affirmed, finding the exposure theory to provide a reasonable means of proration: “An insurer contracts to pay the entire cost of defending a claim which has arisen within the policy period. The insurer has not contracted to pay defense costs for occurrences which took place outside the policy period. . . . [H [The] exposure theory . . . establishes that a reasonable means of proration is available. . . . [Indemnity costs can be allocated by the number of years that a worker inhaled asbestos fibers.” (Id. at pp. 1224-1225.) Although other courts using an exposure theory have similarly prorated liability among the insurers (Commercial Union Ins. Co. v. Sepco Corp., supra, 765 F.2d at p. 1544; Porter v. American Optical Corp., supra, 641 F.2d at p. 1145 [the rule of proration among insurers is “logically consequent” to the exposure theory]), we agree with the Illinois Supreme Court that a pro rata approach does not apply to defense costs or indemnity if the exposure theory is not used. (Zurich Ins. Co. v. Raymark Industries, supra, 514 N.E.2d 150, 165.) In finding that asbestos injuries continue to occur even after exposure to asbestos ceases, the trial court necessarily rejected the underlying temporal premise of the exposure theory, that injury occurs and the insurers’ obligations are triggered only during the claimant’s period of exposure to asbestos. 18

Moreover, the rule of proration adopted by Forty-Eight Insulations and its progeny fails to recognize that the event which triggers coverage does not define the scope of coverage. Although each policy is triggered only by the occurrence of an injury during the policy period, once a policy is triggered, the policy obligates the insurer to pay “all sums” for which the policyholder becomes liable. There is nothing in the policies limiting the scope of coverage to that portion of a continuous injury that developed during the policy period. (Keene Corp. v. Ins. Co. North America, supra, 667 F.2d at p. 1049.) “As long as there was either inhalation exposure or exposure in residence during a policy period, and as long as [the policyholder] must pay damages as a result, the insurer must indemnify [the policyholder] for whatever damages it must pay.” (Id. at pp. 1044-1045, fn. 20; see also J.H. France Refractories v. Allstate, supra, 626 A.2d 502, 508.)

In Montrose, the Supreme Court criticized language in California Union Ins. Co. v. Landmark Ins. Co., supra, 145 Cal.App.3d at page 478, which held the successive insurers liable “jointly and severally” for the full amount of the damage. 19 (Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at p. 681, fn. 19.) In the present case, the trial court correctly explained that the doctrine of joint and several liability has no application to the obligations of successive insurers of a single policyholder. 20 Nevertheless, the insurance companies insist that the trial court’s decision on the scope of coverage imposes joint and several liability upon the insurers. It does not. The trial court’s decision ensures that the policyholder is indemnified by one insurer for the full extent of the loss up to the policy’s limits, but apportions liability among all insurers whose policies were triggered by the claimant’s asbestos-related bodily injury. We find nothing erroneous in that decision.

(2) Effect of Policyholder’s Self-insurance

In phase III, the trial court concluded that “the policyholders do not have an obligation to share pro rata in indemnification and defense costs because of any uninsured or self-insured periods of time simultaneous with the ‘occurrence’ of bodily injury pertaining to a claim . . . .” The insurers challenge that ruling. The insurers argue that they are obligated to pay only for injuries that took place during the policy periods; thus the manufacturers must pay for injuries that occurred during periods in which the manufacturers were uninsured or self-insured.

The leading support for the insurers’ position is provided by those cases in which the courts apportioned coverage among insurers based upon the time each policy was on the risk. Those courts then included the policyholder in the allocation scheme and held the policyholder responsible for a pro rata share for periods of self-insurance or no insurance. (Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d at pp. 1202-1204; Commercial Union Ins. Co. v. Sepco Corp., supra, 765 F.2d at p. 1544; Ins. Co. North America v. Forty-Eight Insulations, supra, 633 F.2d at p. 1225; NSP v. Fidelity & Cas. Co. of New York (Minn. 1994) 523 N.W.2d 657, 662; see also IMCERA Group, Inc. v. Liberty Mutual Ins. Co. (1996) 47 Cal.App.4th 699, 736-743 [50 Cal.Rptr.2d 583] [defense costs] review granted May 22, 1996 (S052878); Gulf Chemical & Metallurgical v. Associated Metals (5th Cir. 1993) 1 F.3d 365, 372.)

We decline to follow this approach, for we conclude that a distinction must be drawn between apportionment among multiple insurers and apportionment between an insurer and its insured. This distinction was noted in Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, 665: “In suits between an insured and an insurer to determine coverage, interpretation of the policy language . . . will typically take precedence. ... [^] In contrast, where two or more CGL carriers turn to the courts to allocate the costs of indemnity for a paid loss, different contractual and policy considerations may come into play in the effort to apportion such costs among the insurers. The task may require allocation of contribution amongst all insurers on the risk in proportion to their respective policies’ liability limits (such as deductibles and ceilings) or the time periods covered under each such policy.”

As we have already explained in sub part (1) above, the trial court apportioned the liability of the successive insurers based upon both the policy limits and the time on the risk. 21 That apportionment among multiple insurers, however, has no bearing upon the obligations of the insurers to the insured. The insurance policies obligate the insurers to pay on behalf of a policyholder “all sums” that the policyholder becomes legally obligated to pay as damages because of bodily injury during the policy period. We interpret this language to mean that once coverage is triggered, the insurer’s obligation to the policyholder is to cover the policyholder’s liability “in full” up to the policy limits. It is irrelevant that only part of the asbestos-related disease developed during any single policy period or during a period in which the manufacturer had no insurance. The logical consequence of this ruling is that the policyholder is covered (up to the policy limits) for the full extent of its liability and need not pay a pro rata share. (ACandS, Inc. v. Aetna Cas. and Sur. Co., supra, 764 F.2d at p. 974; Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d at pp. 1047-1049; Zurich Ins. Co. v. Raymark Industries, supra, 514 N.E.2d at p. 165; J.H. France Refractories v. Allstate, supra, 626 A.2d 502, 508; but see Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d at p. 1058 (conc. opn. of Wald, J.).)

We therefore affirm the trial court’s decision in phase III relieving the policyholders from any responsibility to share in the loss for periods of no insurance.

b. Obligations of Insurers Covering Multiple Tortfeasors on a Claim

When a claimant was exposed to products of more than one manufacturer such that an asbestos manufacturer-policyholder is but one of several tortfeasors held liable to an injured victim, the question arises as to how liability should be apportioned among the defendant-manufacturers, especially if the claimant was minimally exposed to one manufacturer’s product and extensively exposed to another’s.

No general statement can be made about the allocation of tort liability of multiple asbestos manufacturers. In some cases, the manufacturers may be held jointly and severally liable to the injured claimant, despite the claimant’s relatively short period of exposure to a particular defendant’s product. (Borel v. Fibreboard Paper Products Corporation (5th Cir. 1973) 493 F.2d 1076 cert. den. (1974) 419 U.S. 869 [42 L.Ed.2d 107, 95 S.Ct. 127].) 22 In other cases, upon adequate proof of the claimant’s varying exposures to different products, damages may be apportioned among the defendants. (Moore v. Johns-Manville Sales Corp. (5th Cir. 1986) 781 F.2d 1061; see Prosser & Keeton on Torts (5th ed. 1984) § 52, p. 352; 3 Harper et al., The Law of Torts (2d ed. 1986), § 10.1, p. 1.) In some jurisdictions, the market share doctrine may be applied, at least where the claimant was exposed to fungible asbestos products (Wheeler v. Raybestos-Manhattan (1992) 8 Cal.App.4th 1152 [11 Cal.Rptr.2d 109]), making the manufacturers only severally liable, based upon each manufacturer’s share of the national market at the time of the plaintiff’s exposure to the product. (Brown v. Superior Court (1988) 44 Cal.3d 1049, 1072-1075 [245 Cal.Rptr. 412, 751 P.2d 470]; Sindell v. Abbott Laboratories (1980) 26 Cal.3d 588 [163 Cal.Rptr. 132, 607 P.2d 924, 2 A.L.R.4th 1061], cert. den. (1980) 449 U.S. 912 [66 L.Ed.2d 140, 101 S.Ct. 285].)

It bears emphasizing that questions concerning the nature and extent of the tort liability of the asbestos manufacturers are not involved in this litigation. Those questions must be resolved in the underlying injury suits. The question here is the extent of the indemnity obligations of the insurers toward their policyholders. The contractual obligations of insurers to a single manufacturer-policyholder are separate and distinct from the tort liability of multiple asbestos manufacturers to an asbestos claimant. (Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d at p. 1051; Ins. Co. North America v. Forty-Eight Insulations, supra, 633 F.2d at p. 1225.) No matter what the tort liability of an asbestos manufacturer—whether joint and several, proportionate to fault or proportionate to market share—the indemnity obligations of its insurers are as set forth in part 2.a above: to respond in full to the policyholder’s liability obligations up to the policy’s limits, subject to apportionment pursuant to “other insurance” clauses. (See Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d at pp. 1050-1051, 1051 fn. 39.)

The insurers, however, have raised a trigger question which arises when multiple asbestos manufacturers are held liable on a single claim and each manufacturer is insured by multiple, successive policies: For purposes of deciding which of a manufacturer-policyholder’s successive policies cover the manufacturer’s liability on the claim, is insurance coverage triggered if the claimant was first exposed to the policyholder’s product after the insurer’s policy had expired? That is, does an insurer have any indemnity obligation if the policyholder’s product was not involved in the claimant’s injury during the policy period? 23

The trial court’s trigger decision states that all of a policyholder’s policies are triggered upon the claimant’s exposure to any asbestos product. The effect of this decision is to trigger an insurer’s indemnity obligations even if the claimant was not exposed to the policyholder’s product until after the insurer’s policy period had expired. For example, if a claimant was first exposed to asbestos products of manufacturer A in 1957 but was not exposed to manufacturer B’s asbestos products until 1967, the trial court’s decision would make policies insuring manufacturer B covering the period from 1957-1966 triggered: the 1957 policy by virtue of the claimant’s exposure to asbestos and the 1958-1966 policies by virtue of the latent development of asbestos disease. (Policies insuring manufacturer B from 1967 to the date of claim or death would, of course, also be triggered by the continuous development of asbestos disease.)

In this respect, the trial court’s decision runs counter to the decision in Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d 1034. After concluding that insurance coverage is continuously triggered from the point of exposure to the point of manifestation, and after concluding that each successive insurer must indemnify the policyholder in full, subject to apportionment under the “other insurance” provisions, the Keene court went on to hold that an insurance company has no liability if it can prove that the claimant was not exposed to the manufacturer-policyholder’s product either during the policy period or before the policy period. “If a victim sues more than one asbestos-product manufacturer, it may be impossible to prove which company’s products were used at which time. If so, it will be impossible to prove that exposure to Keene’s products—as opposed to those of another manufacturer—occurred during a particular time period. In such a case, there should be a presumption that throughout the victim’s period of exposure to asbestos he or she was exposed to Keene’s and the other manufacturers ’ products. The insurer defending Keene in the underlying tort suits may then try to show that Keene’s products could not have been involved for certain years. Similarly, if a suit arises to resolve the allocation of insurance liability, any insurance company can try to prove that there was no inhalation of Keene’s asbestos during or before its policy period. If an insurance company does so, then that company will be free of liability.” (Id. at p. 1052.)

Although the above quoted portion of the Keene decision purports to pertain to the allocation of liability among insurers, it effectively serves to qualify the trigger of coverage. The Keene court held that there is no coverage unless the injury resulted from exposure to the policyholder’s products. “If [there was no inhalation of Keene’s asbestos during or before an insurer’s policy period], then that company will be free of liability.” (667 F.2d at at p. 1052.) Indeed, despite its rejection of the exposure theory for purposes of triggering coverage, the Keene court recognized that the time of the claimant’s exposure to the policyholder’s products is relevant to the trigger of coverage: “[The full extent of the claimant’s exposure to asbestos] is essential to determining which policies cover Keene’s liability.” (Id. at at p. 1051.)

On this point, the Keene court cited and followed the decision in Ins. Co. North America v. Forty-Eight Insulations, supra, 633 F.2d 1212, even though Forty-Eight Insulations had applied an exposure trigger. (Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d at p. 1052, fn. 42.) In Forty-Eight Insulations, the court recognized the basic contract principle that an insurance policy provides coverage only for injuries resulting from the policyholder’s own products, and the court concluded that “where an insurer can show that no exposure to asbestos manufactured by its insured took place during certain years, then that insurer cannot be liable for those years. The reason is simple: no bodily injury resulting from Forty-Eight’s products, took place during the years in question.” (633 F.2d at p. 1225, italics in original.) At the same time, the court in Forty-Eight Insulations held that in asbestos cases, because of the difficulties of proof, each asbestos manufacturer’s products should be presumed to be involved upon a claimant’s exposure to asbestos, but an insurer is entitled to show that the claimant was not exposed to its manufacturer-policyholder’s products. (Id. at 1225-1226, fn. 27.)

Both Forty-Eight Insulations and Keene instruct that an insurance policy is not triggered if the claimant’s exposure to the manufacturer-policyholder’s products took place after the policy period. 24 Even though coverage is triggered continuously, upon exposure, upon manifestation, and upon exposure-in-residence, it is not enough to trigger coverage that the claimant experienced some asbestos-related injury during a policy period; the injury must have resulted from exposure to the policyholder’s products. In the hypothetical example given above, then, the policies insuring manufacturer B would not be triggered until 1967, the date of the claimant’s first exposure to B’s products. B’s policies in effect before 1967 would not be triggered on the claim, despite the claimant’s earlier exposure to manufacturer A’s products in 1957, as there was no injury from B’s products until 1967. (Of course, B’s policies in effect from 1967 onward, to the date of the claimant’s death or claim, would be triggered by virtue of the progressive development of the disease, even if exposure ceased in 1967.)

We think this conclusion is sound. The language of the CGL policies reflects the requirement of a causal connection between the claimant’s injury and the policyholder’s conduct. The insurance policies obligate the insurer to pay all sums “which the insured shall become legally obligated to pay as damages because of bodily injury . . . caused by an occurrence.” An “occurrence” is defined as injurious exposure to conditions “which results during the policy period in bodily injury . . . .” When a claimant was not exposed to the insured’s products until after the policy had ended, the causal connection is missing. If there is bodily injury during the policy period due to exposure to another manufacturer’s products, the insured does not become liable “because of’ such bodily injury. The insured does not become liable at all unless and until there is an exposure to its own products. It is only “because of’ the bodily injury caused by the later exposure to the insured’s products that the insured becomes “legally obligated to pay” damages. As a matter of common sense, an ordinary lay reader would not understand the policy provisions to extend coverage to damages arising from conduct that took place after the policy had expired. 25 (Contra, Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d 1178, 1200-1201, finding policy ambiguous and holding policies triggered on exposure to asbestos generally.)

Thus, in our hypothetical example, if a manufacturer is held liable to a claimant who was exposed to the products of other manufacturers in 1957 and who was exposed to the manufacturer-policyholder’s products in 1967, the manufacturer’s legal liability would be “because of’ that later exposure to its own products, not because of the 1957 exposure. Hence, policies in effect before 1967 would not be triggered on the claim. To construe the policy language otherwise so as to trigger coverage upon exposure to any asbestos product could have the absurd result of triggering policies in effect before the manufacturer-policyholder ever manufactured asbestos products.

In the present case, the trial court’s trigger decision fails to acknowledge this point. The decision creates in effect an irrebuttable presumption that all of an asbestos manufacturer’s policies are triggered by a claim of injurious exposure to any asbestos product. We agree with Keene and Forty-Eight Insulations that an insurer is entitled to rebut that presumption and show that its policy was not triggered because the claimant was first exposed to its manufacturer-policyholder’s products after the policy period had expired.

We therefore modify the judgment to read that a policyholder’s policies are triggered from the claimant’s first exposure to the policyholder’s products. An insurer has no liability if its policy expired before the claimant was exposed to the policyholder’s product. We emphasize, however, that pursuant to Keene and Forty-Eight Insulations the claimant will be presumed to have been exposed to asbestos products of all defendant-manufacturers, and the burden is on the insurer to prove that the claimant was not exposed to its policyholder’s product before or during the policy period.

We reiterate, too, our affirmance of the trial court’s continuous trigger decision: all of a policyholder’s policies are triggered from first exposure to the policyholder’s products until the date of claim or death, whichever occurs first. Thus, coverage is triggered if either (1) the claimant was exposed to the manufacturer’s products during the policy period or (2) the claimant was exposed to the manufacturer’s products at an earlier time such that during the policy period the claimant was experiencing latent asbestos injury from that earlier exposure.

In seeking to justify the imposition of liability upon an insurer which was on the risk before the claimant was exposed to the policyholder’s product, the policyholders contend the insurer’s liability should parallel the “joint and several” liability of the manufacturers. Although one court has said as much (ACandS, Inc. v. Aetna Cas. and Sur. Co., supra, 764 F.2d at p. 974), we think this reasoning is faulty. First, as already noted at the beginning of this section, the liability of the manufacturers is not necessarily joint and several. Second, this is a trigger issue; we are not concerned with apportionment of liability. Whether the liability of the manufacturers is joint and several, proportionate to fault, or proportionate to market share, the contractual obligation of an insurer to indemnify the manufacturer-policyholder, to pay “all sums” which its policyholder becomes legally obligated to pay, arises only if coverage is triggered. And coverage is triggered only if the claimant was exposed to the policyholder’s product either before or during the policy period.

In summary, we affirm the trial court’s continuous trigger decision, and we affirm the trial court’s allocation of indemnity and defense costs among the insurers, including the decision relieving policyholders from bearing a share of the loss for periods of no insurance. But we modify the first sentence of paragraph 8 of the judgment to read that all of a policyholder’s policies that were in effect from the date of the claimant’s first exposure to the policyholder’s asbestos product until the date of death or claim, whichever occurs first, are triggered on an asbestos-related bodily injury claim, but the claimant is presumed to have been exposed to all defendant-manufacturers’ asbestos products, and the burden is on the insurer to prove that the claimant was not exposed to its policyholder’s product before or during the policy period.

B. “Neither Expected Nor Intended”

In phase III of the proceedings, the trial court was called upon to interpret the phrase “neither expected nor intended,” which appeared in the standard CGL policy definition of an occurrence: “ ‘Occurrence’ means an accident, including injurious exposure to conditions, which results during the policy period, in bodily injury . . . neither expected nor intended from the standpoint of the insured.” The trial court determined that the phrase applies to exclude coverage “where the insured acted either wilfully, intentionally, or maliciously for the purpose of causing injury.”

In phase IV, two insurers of Armstrong, Commercial Union and Travelers, argued that their policies provided no coverage for the claims of asbestos-related injuries because Armstrong expected or intended bodily injury resulting from exposure to asbestos. The trial court, however, rejected the argument and, applying its earlier test, found that Armstrong did not act for the purpose of causing injury. Hence, the court found the injuries from exposure to asbestos were neither expected nor intended by Armstrong.

Commercial Union appeals and challenges the trial court’s interpretation of the policy language, arguing that the trial court’s interpretation focuses only on the term “intended” and fails to give effect to the term “expected.” The distinction between intended and expected takes on significance because Commercial Union’s policy (issued by its predecessor, Employers’ Liability Assurance Corporation (ELAC)), being an excess policy to the underlying policy of Continental Casualty, followed the form of that Continental manuscript policy whose language differed from the standard form policy language and defined occurrence as follows: “The term ‘Occurrence’ means an event or continuous or repeated exposure to conditions, which unexpectedly causes Personal Injury and/or Property Damage and/or Advertising Liability during the policy period. . . .” (Italics added.) (That is, in lieu of the standard phrase “neither expected nor intended,” the policy used the term “unexpectedly.”)

The Commercial Union policy was in effect from January 1, 1966, to January 1, 1969. Commercial Union contends that by 1966 Armstrong officials were well aware of asbestosis and other problems associated with inhalation of asbestos. In the view of Commercial Union, because Armstrong knew of the injurious effects of asbestos, the current claims for asbestos-related injuries are not claims for “unexpected” injuries and therefore are not covered, even though Armstrong may not have intended to cause the injuries.

1. Facts

The corporation now known as Armstrong World Industries, Inc. (hereafter Armstrong) began operating in the 1800’s as a cork company and has a long history as a manufacturer of cork products, including L.T. Cork Covering, a low temperature insulation. That product had a paper backing and in 1956, after an episode in which the paper had burned, the decision was made to include asbestos in the paper backing. From 1956 to 1959, Armstrong manufactured this asbestos-containing product.

The asbestos in L.T. Cork Covering was not friable; it was bonded into the paper. The record reveals that Armstrong officials believed L.T. Cork Covering was not a harmful product because it was not dusty. In fact, the cases now pending against Armstrong involving L.T. Cork Covering are relatively few, and in those cases Armstrong is a peripheral defendant. In 1959 Armstrong discontinued the manufacture and sale of L.T. Cork Covering and replaced it with Armaflex, a new insulation product that did not contain asbestos. 26

In the manufacturing operations, Armstrong’s safety supervisor was concerned with the inhalation of dust, especially silica dust, which was known to cause silicosis. The manufacturing plants were kept well ventilated, and in the areas where sacks of asbestos (or other dusty material) were opened and dumped into a hopper for mixing into a product, a suction device was used to draw the dust away from the workers. Respirators were also made available, although they were seldom, if ever, used, as they were bulky and uncomfortable. Employees were given regular chest X-rays, and any employee with a sign of lung disease was transferred to another area of the plant. Actually, Armstrong has received relatively few claims for asbestos-related injuries arising from its manufacturing operations.

Most claims stem from Armstrong’s insulation installation business. Armstrong began in the 1940’s installing various asbestos-containing insulation products which were manufactured by other companies. In some cases Armstrong’s name was placed on the insulation, but Armstrong did not manufacture the insulation.

For its manufacturing operations, Armstrong employed a staff of permanent employees, but for its insulation installation business, Armstrong used temporary workers, hired from the union hall, for each job. A job might last a week; it might last a year. Different insulation products were used on different jobs. It is not entirely clear from the record what precautions were taken with respect to the insulation workers. Because the insulation installations were performed on the owners’ property, Armstrong had less control over working conditions. Respirators were apparently available, but not frequently used, except when using a spray-on product.

Armstrong received its first workers’ compensation claim for an asbestos-related injury in 1952, from an insulation installer. The worker had worked for many different insulation installation companies, but had worked for Armstrong for only 2 weeks. The claim was dismissed as to Armstrong. During the 1960’s, the number of workers’ compensation claims from insulation installers for asbestos-related injuries substantially increased. Yet, about half of the workers’ compensation claims filed against Armstrong were eventually dismissed.

In 1958 Armstrong was restructured so that its installation business became a subsidiary, Armstrong Contracting and Supply Company. In the late 1960’s, Armstrong decided to focus its business on home furnishings rather than commercial installations, and it sold the subsidiary in 1969.

One of the asbestos-containing products used by Armstrong Contracting and Supply was a spray-on insulation called Limpet, which was manufactured in England. (A spray-on product is probably the most dangerous because the asbestos particles are sent directly into the air.) By the 1960’s, Limpet was becoming expensive (due to import duties) and difficult to obtain in sufficient quantities. Moreover, the workers were not happy with the dustiness of the product. Hence, Armstrong Contracting and Supply asked its parent company, Armstrong, to come up with an alternative product, and in 1967 Armstrong began to manufacture Armaspray, a spray-on asbestos insulation. Armstrong sold it only to its subsidiary, Armstrong Contracting and Supply Company, as a test product. Ultimately, the product was deemed a failure, and it was discontinued in December 1968, although the inventory continued to be used by Armstrong Contracting and Supply into 1969 to finish jobs in progress. By that time, too, Armstrong Contracting and Supply was sold. According to the testimony, the decision to discontinue Armaspray was based not on health concerns, but on the lack of sales.

Armaspray was only 9 percent asbestos, in contrast to Limpet, which was nearly 100 percent asbestos. Armstrong’s manufacturing policy called for safety testing of a product before manufacture, and Armaspray passed the health and safety examination given it by Armstrong’s industrial hygienists. Yet, in fact, no tests had been done to determine the airborne dust levels during use of the product. Not until 1968, after a meeting with Dupont, one of Armstrong’s major customers for insulation installations, did Armstrong undertake its first tests of Armaspray to determine the dust count. But by December 1968 Armstrong had stopped manufacturing Armaspray. Test results that came in after Armaspray had been discontinued showed dust levels more than 10 times the acceptable limit.

The first asbestos-related lawsuit was filed against Armstrong in 1970, Borel v. Fibreboard, supra, in which Armstrong was one of the many named defendants. By 1987, the time of the phase IV trial, Armstrong had been sued by over 60,000 plaintiffs. Nearly all cases involve Armstrong’s contract insulation installations. Some plaintiffs were Armstrong’s own workers. (Not all states make workers’ compensation benefits an exclusive remedy.) Some were workers at the job sites where the insulation work was done. Others were household members exposed to the dust on a worker’s clothes.

2. The Trial Court’s Decision

In phase III, the trial court ruled in part as follows on the meaning of the “neither expected nor intended” clause: “This Court determines that the ‘neither expected nor intended’ clause applies where the insured acted either wilfully, intentionally, or maliciously for the purpose of causing injury. (See U.S Fid. & Guar. Co. v. American Employers Insurance Co. (1984) 159 Cal.App.3d 277 [205 Cal.Rptr. 460].) The intent behind the act in question must involve an element of wrongfulness or misconduct. (Mullen [v. Glens Falls Ins. Co. (1977)] 73 Cal.App.3d [163,] 171 [140 Cal.Rptr. 605].)

“One final clarification on the standard is in order. An insurer is not required to produce express testimony or documentation as to an insured’s subjective, wrongful intent to cause injury, but may show that reason mandates that by the very nature of the act undertaken, coupled with the knowledge actually in possession of the insured, harm must have been intended. [Citation.] This clarification accords with the language of the policy, which speaks in terms of what was ‘expected or intended,’ and not in terms of what should have been ‘expected or intended. ’ It further accords with the general duty and right of courts to make determinations based on circumstantial evidence and inferences, as well as determinations drawn directly from the evidence.” (Italics in original.)

After the trial in phase IV, the trial court concluded that “the evidence presented by Commercial Union does not satisfy [the phase III] standard. . . . that the insured acted wilfully, intentionally, or maliciously for the purpose of causing injury. Here, the Court finds that Armstrong did not act for the purpose of causing injury.” (Italics in original.) 27

3. Discussion

Prior to 1966, standard liability policies covered injuries “caused by accident.” The words “caused by accident” served to exclude coverage for wilful acts of the insured. (1 Long, The Law of Liability Insurance (1993) § 1.08[1].) Beginning in 1966, however, the standard policy language was changed to provide coverage for injuries “caused by an occurrence” so long as the injuries were “neither expected nor intended from the standpoint of the insured.” (See Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, 671-672.)

The meaning of the phrase “neither expected nor intended” has been a puzzle for the courts around the country, and the courts have given it varying interpretations. Some find the phrase “neither expected nor intended” to be ambiguous. Other courts do not. Some courts find the terms “expected” and “intended” to have separate meanings. Others find them to be synonymous. Some courts employ a subjective standard (the insured actually expected), while others employ an objective standard (the insured should have expected). It is difficult to find a clear trend in the law. (See Keeton & Widiss, Insurance Law (1988) § 5.4(d)(1) & (2), pp. 518-524; 1 Long, op. cit. supra, § 1.08[2][b]; Annot. (1984) 31 A.L.R.4th 957.) As we will explain below, we give the phrase its plain meaning and construe “expected” to mean an actual awareness that harm was practically certain even though harm was not intended.

a. High Probability

Several courts in other jurisdictions have construed “expected” within the phrase “neither expected nor intended” to mean a high degree of probability: “ ‘The term “expected” when used in association with “intended” carries the connotation of a high degree of certainty or probability . . . ‘ “practically [to] equate with “intended” . . . .’ ” (Patrons-Oxford Mut. Ins. Co. v. Dodge (Me. 1981) 426 A.2d 888, 891, quoting from State Farm Fire & Casualty Company v. Muth (1973) 190 Neb. 248 [190 Neb. 272, 207 N.W.2d 364, 366].)

In Patrons-Oxford, supra, the court found the phrase “neither expected nor intended” ambiguous and construed it against the insurer. Accordingly, the court interpreted the word “expected” so as not to enlarge the exclusion of coverage. Other courts have adopted substantially similar interpretations by examining the ordinary (dictionary) definition of “expect.” (Indiana Farmers Mut. Ins. Co. v. Graham (Ind.Ct.App. 1989) 537 N.E.2d 510, 512 [“the insured acted although he was consciously aware that the harm caused by his actions was practically certain to occur”]; Brown Foundation v. St. Paul Ins. Co. (Ky. 1991) 814 S.W.2d 273, 278 [the insured “subjectively foresaw as a practically certain or expected-to-be result of the conduct”]; Quincy Mut. Fire Ins. Co. v. Abernathy (1984) 393 Mass. 81, 469 N.E.2d 797, 800] [the insured “knew to a substantial certainty that the bodily injury would result”]; United Services Auto. Ass’n v. Elitzky (1986) 358 Pa.Super. 362, 517 A.2d 982, 991] [“the insured acted even though he was substantially certain that an injury . . . would result”]; and see Bay State Ins. Co. v. Wilson (1983) 96 Ill.2d 487 [71 Ill.Dec. 726, 451 N.E.2d 880, 883] [the insured was “consciously aware that. . . injuries were practically certain to be caused by his conduct”]; Farmers Union Oil v. Mutual Service Ins. (Minn.Ct.App. 1988) 422 N.W.2d 530, 533 [the insured subjectively “knew of the substantial risks involved, proceeded in light of this knowledge, and disregarded the known hazard”]; see also Keeton & Widiss, Insurance Law, op. cit. supra, § 5.4(e)(2), at p. 535.)

In California, the fundamental principle guiding judicial interpretation of insurance policy language is that words must be construed in their ordinary and popular sense unless the parties intended a special or technical sense. (Civ. Code, § 1644; AIU Ins. Co. v. Superior Court, supra, 51 Cal.3d at pp. 821-822.) The plain and ordinary meaning of “expect,” as reflected in dictionary definitions, is to anticipate, to consider probable or certain. 28 Hence, in Shell Oil Co. v. Winterthur Swiss Ins. Co. (1993) 12 Cal.App.4th 715 [15 Cal.Rptr.2d 815], this court concluded that “expected," as used in the language of insurance policies, means anticipation with a high degree of probability, no matter whether the degree of that probability is expressed as “substantially certain, practically certain, highly likely, or highly probable.” (Id. at p. 746.)

b. Subjective Awareness

In Shell Oil, the jury had been instructed that “ ‘the exclusionary word “expected” denotes that the actor knew or should have known that there was a substantial probability that certain consequences would result from his or her acts or omissions.’ ” (12 Cal.App.4th at p. 743.) The appellate court rejected the objective (should have known) standard, holding instead that the appropriate inquiry is what the insured actually knew or believed. (Id. at pp. 746-748; see also Waller v. Truck Ins. Exchange, Inc. (1995) 11 Cal.4th 1, 17 [44 Cal.Rptr.2d 370, 900 P.2d 619]; Titan Corp. v. Aetna Casualty & Surety Co. (1994) 22 Cal.App.4th 457, 468 [27 Cal.Rptr.2d 476].)

This interpretation conforms to a line of product liability cases in which the California courts have held that an injury is excluded from coverage only if the insured knew of the defects. The rationale underlying these product liability cases seems to be that if the insured knew of the defects the insured also knew that injuries were practically certain to occur and, hence, the injuries were expected: “[T]o bar third party liability coverage, the defect causing the postsale damage must have been known to Chu or their agents before the units were sold . . . .” (Chu v. Canadian Indemnity Co. (1990) 224 Cal.App.3d 86, 97 [274 Cal.Rptr. 20], italics in original.) “As previously discussed, the purpose of third party liability insurance is to protect the insured against injuries to third parties neither expected nor intended by the insured. To the extent Chu knew of extant defects pre-sale, any injuries flowing therefrom would not be an unexpected or unintended consequence of selling defective units. . . . [<]D However, if Chu did not have pre-sale knowledge of the defect, any injuries suffered would be an unexpected or unintended consequence of selling the units.” (Id., at pp. 98-99; see also Hogan v. Midland National Ins. Co. (1970) 3 Cal.3d 553, 560 [91 Cal.Rptr. 153, 476 P.2d 825]; Geddes & Smith, Inc. v. St. Paul Mercury Indemnity Co. (1959) 51 Cal.2d 558, 563-564 [334 P.2d 881]; Economy Lumber Co. v. Insurance Co. of North America (1984) 157 Cal.App.3d 641, 648 [204 Cal.Rptr. 135].)

Several out-of-state courts have similarly concluded a subjective test should be employed to determine whether the injury was intended or expected. (Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d 1178, 1205; Broderick Inv. Co. v. Hartford Acc. & Indem. Co. (10th Cir. 1992) 954 F.2d 601, 605-606, cert. den. 506 U.S. 865 [121 L.Ed.2d 133, 113 S.Ct. 189]; City of Johnstown v. Bankers Standard Ins. (2d Cir. 1989) 877 F.2d 1146, 1151, fn. 1; Hatco Corp. v. W.R. Grace & Co.—Conn. (D.N.J. 1992) 801 F.Supp. 1334, 1375-1376; Indiana Farmers Mut. Ins. Co. v. Graham, supra, 537 N.E.2d at p. 512; Brown Foundation v. St. Paul Ins. Co., supra, 814 S.W.2d at pp. 278-279; Patrons-Oxford Mut. Ins. Co. v. Dodge, supra, 426 A.2d 888; Quincy Mut. Fire Ins. Co. v. Abernathy, supra, 469 N.E.2d at p. 800; Queen City Farms v. Central Nat. Ins. Co. (1994) 126 Wn.2d 50 [882 P.2d 703, 712-714]; and see Farmers Union Oil v. Mutual Service Ins., supra, 422 N.W.2d at p. 533; United Services Auto. Ass’n v. Elitzky, supra, 517 A.2d at p. 991.)

The courts have used various approaches to reach that conclusion. In Patrons-Oxford Mut. Ins. Co. v. Dodge, supra, 426 A.2d 888, the court found the phrase “neither expected nor intended” to be ambiguous and therefore construed the phrase against the insurer. In United Services Auto. Ass’n v. Elitzky, supra, 517 A.2d at page 991, the court focused on the ordinary, dictionary definitions of “expected”: “Each of these definitions connotes an element of conscious awareness by the insured. None of them defines expected as events the insured should have known about.” (See also Indiana Farmers Mut. Ins. Co. v. Graham, supra, 537 N.E.2d at p. 512 [“Nothing in the definition of ‘expected’ excludes harm that the insured ‘should have anticipated.’ ”].)

Other courts have focused on the language of the exclusionary clause: “neither expected nor intended from the standpoint of the insured.” As one court put it, “The policies here state that the insurer has a duty to indemnify or defend the insured for damage if the damage was neither expected nor intended from the standpoint of the insured. They do not say from the standpoint of a reasonable person.” (Brown Foundation v. St. Paul Ins. Co., supra, 814 S.W.2d at p. 279.)

Other courts have rejected an objective standard because such a standard would extend the exclusionary clause to exclude coverage even for negligence—the very risk the insured sought coverage for. “We are also fearful that an exclusion of injuries the insured ‘should have anticipated’ might exclude from coverage, not only intentional injuries but also those caused by negligence.” (United Services Auto. Ass’n v. Elitzky, supra, 517 A.2d at p. 991; see also Queen City Farms v. Central Nat. Ins. Co., supra, 882 P.2d at pp. 712-713; Grange Mutual Casualty Company v. Thomas (Fla.Dist.Ct.App. 1974) 301 So.2d 158, 159 [declining to differentiate “expected” from “intended” because to do so would exclude coverage for gross negligence].)

In the Shell Oil decision, the court took this approach and reasoned in part that the objective standard would deny coverage for mere negligence: “By testing what Shell should have known, the instructions invited denial of coverage for conduct within the realm of negligence . . . .” (12 Cal.App.4th at p. 748.)

That decision finds support in Chu v. Canadian Indemnity Co., supra, 224 Cal.App.3d 86, where the insured was aware of certain construction defects in the condominium units at the time of sale, but other defects manifested themselves later. The insurer argued that coverage should be denied for the postsale defects, as they were the inevitable result of the known defects. The appellate court rejected the insurer’s argument and held that “third party liability coverage was not barred merely because Chu ‘should have discovered’ the defect but negligently failed to do so.” (224 Cal.App.3d at p. 97; see also Fire Ins. Exchange v. Abbott (1988) 204 Cal.App.3d 1012, 1021 [251 Cal.Rptr. 620] [“We assume that the [phrase] ‘. . . neither expected nor intended by the insured’ excludes from insurance coverage only conduct by the insured which was subjectively intended to harm or injure.”].)

The Chu court reasoned that if the insured was actually ignorant of the defects, denial of coverage on account of a negligent failure to investigate would defeat the very purpose of third party liability coverage: “[I]f Chu did not have pre-sale knowledge of the defect, any injuries suffered would be an unexpected or unintended consequence of selling the units. This is the case even though Chu may have had notice of facts which would incite investigation by a reasonably prudent person, but nevertheless negligently failed to investigate and obtain actual knowledge of the existence of such defects. [Citation.] Since a major purpose of third party liability insurance is to protect the insured from claims for negligence (Garvey v. State Farm Fire & Casualty Co., supra, 48 Cal.3d 395, 407-408 . . .), Chu’s third party coverage is not forfeited merely because they should have known of the existence of defects but negligently failed to discover such defects.” (224 Cal.App.3d at p. 99.) “In reviewing the pertinent authorities, we find no cases denying third party liability coverage to an insured who sold property when he ‘should have known’ of the defect, but who through negligence was actually ignorant of the defect.” (Id., at pp. 99-100.)

We are persuaded by the Chu court’s reasoning, and we apply it here. In our view, imposing a “should have known” standard on insureds would defeat the essential purpose of insurance agreements. What is expected or intended is different from that which was reasonably foreseeable or which should have been known. An insurance policy exclusion from manufacturing activities which carry a risk of causing environmental harm, although not known or intended to cause harm in the insured’s business conduct, would create an exclusion swallowing the entire purpose of insurance protection for unintended consequences. Insurance is purchased and premiums are paid to indemmify the insured for damages caused by accidents, that is, for conduct not meant to cause harm but which goes awry. The insured may be negligent indeed in failing to take precautions or to foresee the possibility of harm, yet insurance coverage protects the insured from his own lack of due care. If coverage is lost for damage which a prudent person should have foreseen, there would be no point to purchasing a policy of liability insurance.

c. “Expected” Differentiated from “Intended”

In the Shell Oil case some insurance policies did not contain the exclusionary phrase “neither expected nor intended.” The court therefore had to decide whether Shell’s environmental pollution came within the statutory exclusion for “wilful” acts (Ins. Code, § 533). 29 The court concluded that a “wilful” act extends beyond an act which causes harm that the insured intended and encompasses as well an act which causes harm that the insured expected: “A ‘wilful act’ under section 533 must also include a deliberate, liability-producing act that the individual, before acting, expected to cause harm. Conduct for which the law imposes liability, and which is expected or intended to result in damage, must be considered wrongful and willful. Therefore, section 533 precludes indemnification for liability arising from deliberate conduct that the insured expected or intended to cause damage.” (Shell Oil Co. v. Winterthur Swiss Ins. Co., supra, 12 Cal.App.4th 715, 743.)

By implication, then, the Shell Oil court acknowledged that “expected” injuries are different from “intended” injuries in that the insured may expect injuries—believe them to be substantially certain to occur—without having the express purpose of causing damage. (See also 12 Cal.App.4th at p. 745.)

Good reasons exist for excluding injuries that are expected though they are not intended. “[Ojrdinarily insurance does not provide indemnification for the type of economic detriments that occur so regularly that they are commonly regarded as a cost, rather than as an insurable risk, of an enterprise or activity. Closely associated with this basic principle is the view that it is fundamentally inconsistent with the legitimate purpose of an insurance arrangement for one to seek to use it as protection against calculated risks .'. . .” (Keeton & Widiss, Insurance Law, op. cit. supra, § 5.4(e), pp. 534-535.) “ ‘If the single insured is allowed through intentional or reckless acts to consciously control the risks covered by the policy, a central concept of insurance is violated. ’ ” (Farmers Union Oil v. Mutual Service Ins., supra, 422 N.W.2d at p. 533, quoting from Bituminous Cas. Corp. v. Bartlett (1976) 307 Minn. 72 [240 N.W.2d 310, 313] overruled on other grounds, Prahm v. Rupp Const. Co. (Minn. 1979) 277 N.W.2d 389, 391; italics added; see also City of Carter Lake v. Aetna Cas. and Sur. (8th Cir. 1979) 604 F.2d 1052, 1059 [where insured took calculated risk that damage would occur and elected to proceed, the results were not accidental]; 7A Appleman, Insurance Law & Practice, op. cit. supra, § 4492.01, p. 21.)

Several out-of-state courts have determined that the two words “expected” and “intended” within the phrase “neither expected nor intended” language cannot be treated as synonymous. These courts have reasoned that the purpose of adding the phrase “neither expected nor intended from the standpoint of the insured” was to broaden the class of excluded injuries beyond intentional injuries. (Patrons-Oxford Mut. Ins. Co. v. Dodge, supra, 426 A.2d at pp. 890-891; Farm Bureau Town & Country Ins. v. Turnbo (Mo.Ct.App. 1987) 740 S.W.2d 232, 236; United Services Auto. Ass’n. v. Elitzky, supra, 517 A.2d at p. 990; see Keeton & Widiss, Insurance Law, op. cit. supra, § 5.4(e)(4), p. 538; id., § 5.4(g), pp. 544-545; 7A Appleman, op. cit. supra, § 4491, pp. 3-4.) Accordingly, the courts have concluded that unless the terms are given different meanings, “expected” would serve no purpose within the exclusionary clause. (Bay State Ins. Co. v. Wilson, supra, 451 N.E.2d at p. 882; Aetna Cas. & Sur. Co. v. Freyer (1980) 89 Ill.App.3d 617 [44 Ill.Dec. 791, 411 N.E.2d 1157, 1159]; Indiana Farmers Mut. Ins. Co. v. Graham, supra, 537 N.E.2d at p. 512; Steelman v. Holford (Mo.Ct.App. 1989) 765 S.W.2d 372, 377 [765 SW2d 372]; Farm Bureau Town & Country Ins. v. Turnbo, supra, 740 S.W.2d at p. 236.)

In dictum, this court, too, has observed that the inclusion of the term “expected” within the policy language renders the exclusionary clause broader than the exclusion for intentional or wilful acts. (See United Pacific Ins. Co. v. McGuire Co. (1991) 229 Cal.App.3d 1560, 1566, fn. 2 [281 Cal.Rptr. 375].)

In light of these authorities, we conclude that in the present case the exclusion within the Commercial Union policy for “unexpected” injuries applies to injuries that the insured subjectively knew or believed to be practically certain to occur even though the insured did not act for the purpose of causing injury. The trial court correctly used a subjective standard, but the court failed to differentiate “expected” from “intended” and did not consider whether Armstrong, the insured, though not intending to cause injury, expected the injuries because it knew of the hazards of asbestos and was aware of the substantial probability of harm from its manufacture and sale.

Armstrong contends that the insurers’ arguments come too late because the insurers failed to alert the trial court to the distinct language of the Continental policy. This contention, however, is not persuasive. The trial court expressly acknowledged that one argument before it, advanced by some insurers, was that “the term ‘expected’ must be given a meaning independent of ‘intended’ so as to bar coverage when the resultant damage is a ‘substantial probability,’ or is ‘likely,’ or is ‘highly expectable.’ ” Whether that argument was directed at the standard form language (“neither expected nor intended”) or the Continental manuscript policy (“unexpectedly”) is irrelevant. The fact remains that the trial court did not give meaning to the term “unexpected” and did not make a finding on whether the asbestos injuries were expected by Armstrong.

Although we find the trial court’s interpretation of the policy language in error (paragraphs 20 and 21 of the judgment), we find it unnecessary to remand for further findings in accordance with a broader interpretation of the exclusionary language. When a trial court fails to make a finding on a material issue, the omission is harmless error unless the evidence is sufficient to sustain a finding in favor of the complaining party. (Nunes Turfgrass, Inc. v. Vaughan-Jacklin Seed Co. (1988) 200 Cal.App.3d 1518, 1525 [246 Cal.Rptr. 823]; People ex rel. Sorenson v. Randolph (1979) 99 Cal.App.3d 183, 187 [160 Cal.Rptr. 69]; South Bay Irr. Dist. v. California-American Water Co. (1976) 61 Cal.App.3d 944, 995 [133 Cal.Rptr. 166].) From our review of the trial record, we find insufficient evidence to support a finding that Armstrong officials knew or believed the asbestos bodily injuries were practically certain to occur.

By all accounts, Armstrong executives believed that the company’s own asbestos products, L.T. Cork Covering and Armaspray, were not dangerous. There was concern about the dustiness of Armaspray—but not so much for health reasons as for reasons of cleanup and potential damage to electrical equipment. The specifications for Armaspray called for good ventilation, screening off the areas from non-users, and respirators while using the product.

There is no evidence that Armstrong officials knew that its own workers were endangered by Armstrong’s insulation installation operations. Mr. Bushnell, Armstrong’s research and development specialist, testified that he first learned of the health dangers of asbestos at a May 1968 conference. Mr. King, the marketing manager for Armaspray, testified that he was unaware of the health risk of asbestos until the 1968 meeting with Dupont officials. Other Armstrong executives, however, were aware by the early 1960’s that breathing asbestos dust could be dangerous. But general knowledge of the hazards of asbestos is not equivalent to knowledge that asbestos bodily injuries were practically certain to occur. The record indicates that the Armstrong officials believed that the workers would not be harmed as long as the dust levels were controlled. There is no evidence the Armstrong officials actually knew the dust levels at their own job sites were hazardous.

Armstrong’s insurance manager, Mr. Hofferth, knew that Armstrong’s insurers periodically sent “loss prevention engineers” to inspect Armstrong’s manufacturing plants and job sites. He relied on them to alert Armstrong to potential problems. At no time during the 1953-1973 policy periods did any loss prevention report from any of Armstrong’s carriers express a concern for the dangers of using or inhaling asbestos. (In 1977, Armstrong’s carrier inserted an exclusion for asbestosis.) In fact, Mr. Hofferth knew that Armstrong got favorable premiums because of its relatively low loss experience compared to the national average.

Commercial Union primarily relies upon the evidence that Armstrong expected workers’ compensation claims from its insulation installers. By 1961, Mr. Hofferth was “alarmed” and “concerned” at the rise in workers’ compensation claims for asbestos injuries. At trial, Mr. Hofferth explained that what had concerned him was that Armstrong was being saddled with workers’ compensation claims for injuries that occurred on other companies’ jobs. He knew that the workers hired from the union hall worked for other companies who might not take the same precautionary measures. And he knew that in many states the workers’ compensation laws make the last employer responsible for payment of workers’ compensation benefits. Mr. Hofferth proposed that the insulation workers be given preemployment chest X-rays, so that diseased employees could be screened out, but that plan was opposed by the union and was not put into effect. Mr. Hofferth therefore expected the workers’ compensation claims to increase.

But Mr. Hofferth testified that he believed that Armstrong was taking every necessary precaution to protect the workers from injury; he believed the dust levels were being controlled. Moreover, Mr. Hofferth relied on Armstrong’s carriers to investigate workers’ compensation claims. Despite the carriers’ awareness of the increasing number of claims, no carrier expressed concern over the use of asbestos products.

Finally, there is no evidence that anyone at Armstrong knew that third parties might be injured by exposure to asbestos fibers released during Armstrong’s contract installation activities. The record indicates that until the Borel lawsuit was filed in 1970, Armstrong officials had been unaware that its asbestos products were a danger to third parties. In fact, in 1973, Armstrong changed its primary insurance carrier to Liberty Mutual, and even though the Borel lawsuit had been filed, Liberty Mutual expressed no concern for Armstrong’s potential liability for asbestos injuries.

In summary, although the evidence of Armstrong’s general knowledge of asbestos dangers might support a finding that Armstrong should have expected the asbestos bodily injuries, the insurer’s burden was to prove, directly or circumstantially, that Armstrong actually did expect them. In light of the whole record, we find the evidence insufficient to support a finding that during the policy period at issue here, 1966 through 1968, Armstrong was actually aware the asbestos bodily injuries were practically certain to occur. Consequently, we affirm the trial court’s judgment (paragraph 32) that coverage under the Commercial Union policy for asbestos bodily injury claims is not excluded.

C. Premerger Liability

From May 1, 1961, to May 1, 1967, Continental Casualty and Commercial Union provided excess insurance coverage to GAF Corporation under three separate policies. The premiums were based on GAF’s gross sales and were adjusted annually to reflect changes in GAF’s operations, including corporate acquisitions, during the policy period. During the 1961-1967 policy periods, GAF did not manufacture asbestos products. On May 26, 1967—after the expiration of the Continental and Commercial Union policies—GAF merged with Ruberoid Co., which had manufactured asbestos building materials since the 1880’s. After the merger, Ruberoid ceased to exist. All of its assets were transferred to GAF, which took over its asbestos-manufacturing operations. Not until 1969 was the first claim brought against GAF for an asbestos-related bodily injury arising out of Ruberoid’s products.

In phase IV of the coordinated proceedings below, the trial court was asked to decide whether the premerger insurers of GAF (Continental and Commercial Union) provided coverage for asbestos-related injuries attributable to the products of the Ruberoid Company. The trial court concluded that the premerger policies do provide coverage, but we have concluded that ruling was erroneous.

1. The Insuring Agreements

The insuring agreements in each of the three insurance policies at issue here obligate the insurers to pay “for all sums which the insured shall be obligated to pay by reason of the liability imposed upon him by law or assumed by him under contract for damages ... on account of . . . personal injuries . . . .” In support of the trial court’s ruling, GAF emphasizes that upon the merger with Ruberoid GAF became obligated both by law (Ray v. Alad Corp. (1977) 19 Cal.3d 22, 28 [136 Cal.Rptr. 574, 560 P.2d 3]; Moe v. Transamerica Title Ins. Co. (1971) 21 Cal.App.3d 289, 304 [98 Cal.Rptr. 547]) and by contract for the liabilities of Ruberoid. 30 Hence, GAF argues that the insuring agreements extend coverage to damages for which GAF is held liable on account of asbestos-related injuries caused by exposure to Ruberoid’s products.

The trial court accepted this argument and ruled that “Coverage is mandated by the language contained in the insuring agreement of each of the three policies. . . . Courts have imposed liability for asbestos-related bodily injury damages on GAF because of its acquisition of Ruberoid. The plain language in the insuring agreements therefore provides coverage to GAF for that liability.” We cannot agree.

It is axiomatic that insurance policies must be interpreted as a whole. (Producers Dairy Delivery Co. v. Sentry Ins. Co. (1986) 41 Cal.3d 903, 916, fn. 7 [226 Cal.Rptr. 558, 718 P.2d 920]; Milazo v. Gulf Ins. Co. (1990) 224 Cal.App.3d 1528, 1536 [274 Cal.Rptr. 632].) Although, on its face, the insuring agreement may appear to extend coverage to GAF’s liabilities attributable to Ruberoid, the insuring agreement must be read in conjunction with the “named insured” provision. (Milazo, supra, at p. 1536.) As we will discuss below, Ruberoid does not qualify as a named insured. 31

2. Named Insured

The two policies issued by Commercial Union’s predecessors, Employers’ Surplus Lines Insurance Company (ESLIC) and Employers’ Liability Assurance Corporation (ELAC), contain language limiting products liability coverage to products manufactured “by the named insured or by others trading under his name.” The Continental policy follows form to the underlying ESLIC policy. The “named insured” is identified in the ESLIC and Continental policies as “[GAP] and/or its subsidiary, associated, and affiliated companies or owned and controlled companies as now existing or hereafter constituted.” The ELAC policy contains a more limited definition of named insured, insuring only those companies owned or acquired during the policy term. 32

Commercial, Union and Continental contend that the definition of “named insured” within the ESLIC and Continental policies cannot extend to Ruberoid because Ruberoid was never a subsidiary of GAP nor was it an owned and controlled company: upon the merger, Ruberoid ceased to exist. GAP, on the other hand, argues that the purpose of the language was to extend coverage to corporations acquired by GAP. The trial court agreed with GAP’s argument: “The plain meaning of ‘hereafter constituted’ indicates an intention to provide coverage to GAP despite its assumption of new liabilities resulting from the acquisition of Ruberoid.”

We conclude to the contrary. We do not doubt that the phrase “or hereafter constituted” within the named insured definition would extend coverage to a company acquired after the policy period began. 33 (See Reserve Insurance Co. v. Apps (1978) 85 Cal.App.3d 228, 231 [149 Cal.Rptr. 223] [dictum: named insured includes spouse acquired after policy took effect].) But the word “hereafter” cannot reasonably be read as referring to any time in the indefinite future. Obviously, in the abstract, the phrase “or hereafter constituted” could refer either to companies acquired at any time in perpetuity or to those acquired after the inception of the policy but before the end of the policy term. (See, e.g., Webster’s New International Diet., supra, p. 1058 [defining “hereafter” as “in some future time or state”].) As a matter of policy interpretation, however, the phrase must be read within the context of the policy as a whole, and thus must be read in conjunction with the policy period.

A liability insurance policy has a finite duration. The period of time during which the insurance policy is effective is an essential element of a liability insurance contract (Ins. Code, § 381, subd. (e); Parlier Fruit Co. v. Fireman’s etc. Ins. Co. (1957) 151 Cal.App.2d 6, 21 [311 P.2d 62]), and the reason is obvious: the insurer’s obligation to indemnify is limited to insurable events occurring during the coverage period. Unless coverage has been triggered during the policy period, there is no coverage once the policy period has ended. Logically, then, neither is there a named insured once the policy period has ended. Thus, a corporate acquisition taking place after the policy has expired can have no retroactive effect on the identity of the named insured during the policy period. 34

We therefore conclude that the named insured definition under the ESLIC and Continental policies does not include a company acquired, as here, after the policy period ended. 35 This conclusion applies as well to the ELAC policy, which expressly limits the named insured to those companies owned or acquired by GAP during the policy term.

Indeed, in all of the California cases we have located on this issue, the party who qualified as a named insured did so during the policy period. (See, e.g., Continental Cas. Co. v. Phoenix Constr. Co. (1956) 46 Cal.2d 423, 438 [296 P.2d 801, 57 A.L.R.2d 914] [negligent driver qualified as a “managing employee”]; Utley v. Allstate Ins. Co. (1993) 19 Cal.App.4th 815, 819, 823 [24 Cal.Rptr.2d 1] [adult son was additional insured as “resident relative” of insured]; Safeco Ins. Co. v. Gibson (1989) 211 Cal.App.3d 176, 182, 184 [259 Cal.Rptr. 206] [minor child of divorced parents was “resident” of the insured’s household]; Reserve Insurance Co. v. Apps, supra, 85 Cal.App.3d at p. 231 [separated spouse was “resident” of the household]; State Farm Mut. Auto. Ins. Co. v. Elkins (1975) 52 Cal.App.3d 534, 538 [125 Cal.Rptr. 139] [insured’s college student daughter was “resident” of the household]; cf. National Auto. & Cas. Ins. Co. v. Underwood (1992) 9 Cal.App.4th 31, 40 [11 Cal.Rptr.2d 316] [minor child of divorced parents was not “resident” of insured’s household].) 36

In the present case, Ruberoid had no relationship with GAP during the 1961-1967 policy periods. The merger of Ruberoid and GAP took place after the Continental and Commercial Union policies had expired. The fact that the companies became affiliated later is not enough to give Ruberoid the status of a named insured under the premerger policies. Accordingly, we reverse the trial court’s judgment (paragraph 30) on this point.

D. The Wellington Agreement

In 1985 certain parties of the coordinated proceedings, along with other asbestos manufacturers and insurers, joined in a settlement known as the Wellington Agreement. The trial court described the settlement as follows:

“Negotiations between producers of asbestos products and insurers began in 1982 in response to the problems associated with massive, nationwide litigation of asbestos bodily injury claims. During the period the Wellington Agreement was being negotiated, producers of asbestos were faced with literally tens of thousands of bodily injury claims by workers, as well as cross-claims by co-defendants in the underlying cases. In addition, there were numerous and major coverage disputes between producers and insurers. After several years of negotiations, the Wellington Agreement was executed on June 19, 1985. There were 47 original signatories to Wellington, including both insurers and producers. Any other producer or insurer could become a signatory to the agreement.

“The purposes of the Wellington Agreement were to resolve the numerous coverage disputes between and among insurers and producers, to revolve the cross-claims among producers, and to reduce the costs of litigation. According to the agreement itself, the subscribers to Wellington desired to take reasonable and practical steps ‘to ensure the expenditure of funds for the reasonable payment of meritorious claims at reasonable processing costs.’

“To this end, the subscribing members of Wellington agreed to establish a non-profit organization, the Asbestos Claims Facility, which would administer, evaluate, settle, pay or defend all asbestos-related claims against the subscribing producers and insurers. The Wellington Agreement sets forth standards for the handling of claims by the facility. The facility is governed by a board of directors which contains an equal number of producers and insurers.

“Settlement of the cross-claims among producers was essential to the consolidation of the handling of asbestos claims into a single entity. In order to achieve such a settlement, producers agreed to pay a percentage of all claims, whether or not they were named in a claim. The mechanism by which liability on each asbestos-related claim is allocated among producers is the producer allocation formula. The Court does not have before it the percentage that each producer pays, but rather the formula from which the numbers are derived. The percentage allocation is computed based on the number of open and closed claims for each producer as of September 30, 1983 and the amount paid or owing on closed claims.”

1. The Trial Court’s Findings

In phase IV of the proceedings the trial court was called upon to decide the effect of the Wellington Agreement (and other settlements) on disputes among settling and non-settling insurers: does the settlement determine the amount of “other insurance” available to the policyholder for payment of claims for purposes of contribution among insurers? The court was also called upon to decide the effect of the Wellington Agreement on disputes between settling policyholders and nonsettling insurers. Continental Casualty Company and its related companies, Columbia Casualty Company and CNA Casualty of California, (hereafter referred to collectively as the CCC Companies) challenge the trial court’s decision on the latter issue.

Reasonable Settlement

Although the former (other insurance) issue is not before us, the trial court’s decision on that issue has relevance to our analysis of the effect of the settlement upon the CCC Companies’ indemnity obligations. We therefore take note that the trial court heard evidence and made a finding that the Wellington Agreement was a reasonable, good-faith settlement, despite the charge of inaccuracy in the producer allocation formula. The trial court’s decision follows:

“Dr. Wecker testified that the producer allocation formula is inaccurate in that it does not replicate the tort system. A major source of inaccuracy, according to Dr. Wecker, is the requirement that producers pay on claims in which they are not named. Dr. Wecker testified that the formula assumes that the frequency with which a producer is named in a claim will not change over time. If new categories of claims arise which apply to a particular producer, the frequency with which that producer is named would increase, and there would be a corresponding decrease in the frequency with which other producers are named. Mr. Pulkrabek testified that there were new categories of claims and that some producers and insurers have expressed concern regarding their Wellington share.

“Assuming that the formula has resulted over time in differentials between what producers would pay under the tort system and what producers are paying under Wellington, it does not follow that the producer allocation formula is unreasonable. The Court must evaluate the settlement at the time it was made. It is clear to the Court that at the time the Wellington Agreement was executed, the producer allocation formula was intended to replicate what producers would have paid on claims outside of Wellington. Mr. Pulkrabek testified that Armstrong, for example, looked very hard at the formula and determined that Armstrong’s liability would not increase under Wellington. This is borne out by the formula itself, which determines each producer’s share on the basis of the producer’s litigation experience over the years prior to September 30, 1983. Although a producer pays on all claims, whether or not that producer is named in a particular claim, it is equally true that other producers pay on claims in which they are not named and thus pay a proportionate share of the named producer’s liability. In addition, the Court finds that defense costs were reduced substantially by the Wellington Agreement. Given the circumstances in which the Wellington Agreement was executed, the producer allocation formula clearly meets the standard of reasonableness.

“The Wellington Agreement represents a unique solution to an unprecedented litigation problem. Given the lengthy negotiations between insurers and producers preceding the execution of the agreement, the procedures and standards set up for handling claims, and the allocation formulas incorporated in the agreement, the Court is convinced that the Wellington Agreement represents a reasonable, good faith settlement among the subscribing insurers and producers.”

Liability of Policyholders

The nonsettling insurers argued below that their obligations to indemnify the asbestos manufacturers should not be defined by the manufacturers’ Wellington payments because the producer allocation formula requires the manufacturers to pay on claims for which they are not legally liable. The trial court rejected the argument: “The Court concludes that the amounts paid by the policyholders pursuant to the producer allocation formula are presumptive evidence of the legal liability of the policyholders for asbestos-related claims. . . . [U Since the insurers have failed to offer sufficient evidence to rebut this presumption, the Court concludes that the insurers are obligated to reimburse the policyholders for their liability for asbestos-related claims as defined by the producer allocation formula, subject to the contribution principles set forth in this decision.”

2. Discussion

Fibreboard and Armstrong became parties to the Wellington Agreement, but the CCC Companies declined to join the settlement. In this segment of the appeal, the CCC Companies argue that the trial court’s finding that the policyholders’ Wellington payments are “presumptive evidence” of the policyholders’ liability contravenes basic principles of insurance law.

The CCC Companies rely on the rule that it is the policyholder who has the initial burden of proving that a claim comes within the scope of coverage. (Merced Mutual Ins. Co. v. Mendez (1989) 213 Cal.App.3d 41, 47 [261 Cal.Rptr. 273]; Royal Globe Ins. Co. v. Whitaker (1986) 181 Cal.App.3d 532, 537 [226 Cal.Rptr. 435].) And the CCC Companies emphasize the policy language which triggers coverage only if the occurrence “results during the policy period in bodily injury” and which obligates the insurers to indemnify the policyholders for amounts the policyholders become “legally obligated to pay as damages.” The CCC Companies argue, therefore, that the policyholders have the burden of establishing the policies were triggered by the claims paid. Since no evidence was presented on the facts of any of the claims paid through Wellington and, indeed, since the Wellington payments admittedly went toward all claims regardless of whether they were claims for which the policyholders were liable, the CCC Companies argue they have no obligation to reimburse the policyholders for their Wellington payments.

We cannot agree. The general rule placing the burden on the policyholder to establish facts to trigger coverage is subject to the exception explained in Isaacson v. California Ins. Guarantee Assn. (1988) 44 Cal.3d 775 [244 Cal.Rptr. 655, 750 P.2d 297]: When the insurer refuses to accept a settlement and the insured meets its burden of proving the settlement was reasonable, then the insured is entitled to a presumption in his favor—a presumption that the insured is indeed liable to the claimant and that the amount of his liability is the amount of the settlement.

In Isaacson, the Supreme Court reiterated the rule that if an insurance carrier breaches its contract with the insured and erroneously denies coverage or refuses to defend, then the insured is entitled to make a reasonable settlement with the claimant and to sue the carrier to recover the amount of the settlement. (44 Cal.3d at p. 791; see also Clark v. Bellefonte Ins. Co. (1980) 113 Cal.App.3d 326, 335 [169 Cal.Rptr. 832].) Further, in such an action for reimbursement of the settlement, the settlement is presumptive evidence of the insured’s legal liability on the third party’s claim and the amount of the insured’s liability. (Isaacson, supra, 44 Cal.3d 775; see also Peter Culley & Associates v. Superior Court (1992) 10 Cal.App.4th 1484, 1493-1494, 1497 [13 Cal.Rptr.2d 624]; Kershaw v. Maryland Casualty Co. (1959) 172 Cal.App.2d 248, 256-257 [342 P.2d 72]; Lamb v. Belt Casualty Co. (1935) 3 Cal.App.2d 624, 631 [40 P.2d 311].)

The Isaacson court went on to say that even if the insurer has not denied coverage or refused to defend, the insurer has a duty to accept a reasonable settlement, and the insurer’s refusal to settle may give rise to the insured’s action for reimbursement of the settlement. (44 Cal.3d at p. 792; see also Crisci v. Security Ins. Co. (1967) 66 Cal.2d 425, 430 [58 Cal.Rptr. 13, 426 P.2d 173]; Comunale v. Traders & General Ins. Co. (1958) 50 Cal.2d 654, 659 [328 P.2d 198].) In such a case, the insured has the burden of showing the settlement was reasonable and if it meets that burden, then again the act of settlement raises two presumptions: that the claim was legitimate and that the amount of the settlement was the amount of the insured’s liability. (Isaacson v. California Ins. Guarantee Assn., supra, 44 Cal.3d at pp. 793-794.)

In Isaacson, the court held the insureds in that case were not entitled to reimbursement for their contribution toward the settlement because the insurer had neither denied coverage nor refused to defend and the insured had failed to prove the settlement was reasonable. In the present case, the trial court distinguished Isaacson on these grounds: “In this case, no determination has as yet been made as to whether the non-settling insurers wrongfully refused to defend or indemnify claims. However, it is clear that the insurers were disputing defense and coverage obligations at the time that the policyholders entered into the Wellington Agreement. The policyholders in this case have ‘offered evidence of sufficient substantiality’ that the Wellington Agreement is reasonable, and that the non-settling insurers had the opportunity to join Wellington but declined to do so.”

Further, the trial court reiterated that the Wellington Agreement is a reasonable settlement: “In Isaacson, the court evaluated the reasonableness of the settlement of the underlying claim in determining whether the insureds were entitled to a presumption of liability. Here, there is no evidence before the Court regarding the specifics of the settlements of the underlying claims. However, in the unique context of this case, the Court finds that the principles of Isaacson are applicable to the producer allocation formula and to the ongoing process of claims handling by the Asbestos Claims Facility as set forth in the Wellington Agreement. It would place an unreasonable burden on the policyholders and on the judicial system to allow the non-settling insurers to revisit the merits of the many claims which have been settled by the facility since its inception. Moreover, the insureds have offered no evidence of bad faith or unreasonableness in the facility’s handling of the underlying claims. [*]0 Under the circumstances of this case, the Court finds that the amounts paid by the policyholders pursuant to the producer allocation formula ... are presumptive evidence of the policyholders’ liability.”

The trial court’s finding that the Wellington Agreement was a reasonable settlement is not challenged on appeal, and it is determinative of the issue. 37 Once the trial court found the settlement was reasonable, despite the inaccuracy of the producer allocation formula, the Isaacson rule became operative and justified the trial court’s treatment of the settlement as presumptive evidence of liability. (See Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d 1178, 1206-1208.) The judgment on the effect of the Wellington Agreement (paragraph 19) is affirmed.

Issue Group III: Property Damage

In phase V of the coordinated proceedings, the trial court was asked to determine the obligations of Armstrong’s insurers to defend and indemnify Armstrong in the so-called “building cases”—the myriad of property damage lawsuits filed against Armstrong on account of the presence of asbestos-containing building material (ACBM) in buildings. 38 Armstrong is facing liability primarily for its manufacture of asbestos-containing floor tile and insulation materials.

The underlying complaints in the building cases, taken as a whole, reveal that the presence of ACBM in buildings may have various consequences to the buildings’ owners. The ACBM may pose a health hazard to those who use the building in that asbestos fibers may be released into the air or onto building surfaces (walls, upholstery, fixtures, etc.) or settled releases may be disturbed and “reentrained” into the air. Whether or not the ACBM has released asbestos fibers, the building owner may decide to remove or encapsulate the asbestos to eliminate the potential health risk. Or the building owner may incur costs for inspecting, assessing, maintaining and repairing in-place ACBM. And the market value of the property may fall as a result of the presence of asbestos.

Beginning in the early 1980’s, after the federal government started to voice concern about the safety of ACBM, numerous lawsuits were brought against Armstrong and other asbestos producers for property damage to buildings in which ACBM had been installed. At the start of phase V there were 163 building cases, including a number of class actions, pending against Armstrong in courts across the country. Although the complaints advance various legal theories, the plaintiffs in the building cases generally seek compensation for the sums they must expend to eliminate the alleged health hazard in their buildings and for the diminished value of their buildings resulting from the presence of asbestos.

A. Coverage for “Property Damage”

We begin with the question whether the injuries allegedly suffered by the building owners constitute “property damage” as defined by the insurance policies. Many of Armstrong’s policies are standard CGL policies; others have substantially the same provisions. The insuring agreements of the CGL policies obligate the insurers to pay “all sums which the insured shall become legally obligated to pay as damages because of . . . property . . . damage caused by an occurrence.” 39 Since 1973, the standard CGL policy has defined property damage as follows: “i) physical injury to or destruction of tangible property which occurs during the policy period, including the loss of use thereof at any time resulting therefrom, or ii) loss of use of tangible property which has not been physically injured or destroyed provided such loss of use is caused by an occurrence during the policy period.”

Before 1973, under the 1966 revision to the standard CGL occurrence policy, “physical injury” was not a necessary element of property damage; property damage was defined as “injury to or destruction of tangible property.” Before 1966, the standard CGL policy had no requirement that the property be “tangible.” Because the post-1973 policies contain the most restrictive definition of property damage, we confine our analysis to those policies, for if there is coverage under the post-1973 policies, there will be coverage under the pre-1973 policies as well.

The trial court concluded that all claims, whether for release of asbestos fibers or for mere installation of ACBM, are for covered “property damage” under all of Armstrong’s policies. The trial court reasoned that the release of asbestos fibers is an act of contamination that amounts to physical injury and, even without a release of fibers, the diminished value resulting from the incorporation of ACBM in a building constitutes property damage. Although we employ slightly different reasoning, we agree with the trial court’s conclusion that the building claims allege “property damage” within the meaning of the insurance policies.

1. Injury Is Assumed

Relying upon the rule that in a coverage dispute the burden is on the insured to prove coverage (Royal Globe Ins. Co. v. Whitaker (1986) 181 Cal.App.3d 532 [226 Cal.Rptr.2d 435]), the insurers argue that Armstrong failed to prove that the buildings have suffered physical injury. The insurers complain that Armstrong relied on the allegations of injury, without actual evidence of such injury. Indeed, the trial court noted that “Armstrong introduced little evidence independent of the underlying allegations to support its position.”

We find no merit in the insurers’ argument. This is a declaratory relief action, held before the determination of Armstrong’s liability for property damage. None of the 163 building cases filed against Armstrong have yet gone to trial. In such circumstances, the trial court may properly determine questions of insurance coverage on the basis of the underlying pleadings and such other evidence as is available. (See Underwriters Ins. Co. v. Purdie (1983) 145 Cal.App.3d 57, 64 [193 Cal.Rptr. 248]; State Farm Fire & Cas. Co. v. Kohl (1982) 131 Cal.App.3d 1031, 1034 [182 Cal.Rptr. 720].) Here, in ruling upon the meaning of “property damage,” the trial court looked to the “nature of the insured’s potential liability,” taken from the allegations in the various complaints in the underlying building cases, together with the “totality of the evidence.” We see no error in this approach.

We emphasize that there is nothing in the trial court’s decision, nor in our own, which resolves whether the various effects of ACBM upon a building will give rise to liability of the asbestos manufacturer for property damage. In fact, we note that in some circumstances tort liability is uncertain. (See, e.g., Anthony v. Kelsey-Hayes Co. (1972) 25 Cal.App.3d 442, 446-447 [102 Cal.Rptr. 113] [mere depreciation in value caused by safety concern not compensable property damage]; San Francisco Unified School Dist. v. W.R. Grace & Co. (1995) 37 Cal.App.4th 1318, 1324-1335 [44 Cal.Rptr.2d 305] [mere presence of ACBM, without contamination from released fibers, not compensable property damage]; Adams-Arapahoe School Dist. No. 28-J v. GAF Corp. (10th Cir. 1992) 959 F.2d 868, 872 [same].)

The trial court’s conclusion that the claims of injury from ACBM are covered “property damage” as defined by the insurance policies was necessarily based upon the assumption that there have been legally compensable injuries to the buildings for which Armstrong will be held liable, for if it is ultimately determined that there have been no such injuries, then there will be no need for insurance coverage. In this appeal, too, for purposes of deciding the coverage dispute, we assume, as did the trial court, the buildings have been injured as alleged in the complaints.

2. Admissibility of Evidence

In a separate brief, four insurers challenge the admissibility of certain deposition testimony during the trial held on Armstrong’s declaratory relief action. 40 We treat the issue rather summarily, as the evidence was not prejudicial.

The challenged evidence consists of the deposition testimony of experts designated by Reliance Insurance Co., which the trial court had initially excluded from Armstrong’s case-in-chief, but which the trial court eventually admitted after various insurers had moved for judgment on the ground that Armstrong had failed to prove property damage. The four challenging insurers claim that without the deposition testimony Armstrong’s case was devoid of evidence and the trial court would have granted the insurers’ motion for judgment.

We reject the argument. Although the deposition testimony supported Armstrong’s position that buildings are injured by the presence of ACBM, that position was founded in the allegations of the underlying complaints. As we have explained in part A.l above, the trial court relied primarily upon the allegations in the underlying building cases and assumed, for purposes of the declaratory relief action, that the buildings suffered damage for which Armstrong will be held liable. We can discern no prejudice to the insurers from the admission of the deposition testimony.

3. The Injury Is Physical

Because we assume there has been an injury, the question we must decide is whether injury to a building from ACBM qualifies as a “physical” injury. In cases alleging releases of asbestos fibers into a building’s air supply and onto building surfaces, the issue is relatively easy to resolve. As the trial court found, upon a release of asbestos fibers within a building, “[t]he area becomes hazardous and certain measures must be taken to restore the surface to its prior condition.” The courts have held that contamination of buildings and their contents from released fibers constitutes a physical injury and, hence, property damage covered under the terms of the insurance policies. (Dayton Independent School D. v. National Gypsum (E.D.Tex. 1988) 682 F.Supp. 1403, 1407, revd. on jurisdictional grounds sub nom. W.R. Grace & Co. v. Continental Cas. Co. (5th Cir. 1990) 896 F.2d 865, 875; USF&G v. Wilkin Insulation Co. (1991) 144 Ill.2d 64 [161 Ill.Dec. 280, 578 N.E.2d 926, 931-932]; see AIU Ins. Co. v. Superior Court (1990) 51 Cal.3d 807, 829, 842 [274 Cal.Rptr. 820, 799 P.2d 1253] [contamination of environment from release of hazardous waste]; Abraham, Environmental Liability Insurance Law (1991), pp. 80-81, 88.) This conclusion finds support in a number of tort cases in which the courts have held, under the tort doctrine allowing recovery only upon physical injury to property, that contamination from the release of asbestos fibers constitutes a physical injury. (City of Greenville v. W. R. Grace & Co. (4th Cir. 1987) 827 F.2d 975, 980; City of Manchester v. National Gypsum Co. (D.R.I. 1986) 637 F.Supp. 646, 651 - 652; Town of Hooksett School Dist. v. W.R. Grace Co. (D.N.H. 1984) 617 F.Supp. 126, 130-131; 80 S. 8th St. Ltd. Ptsp. v. Carey-Canada (Minn. 1992) 486 N.W.2d 393, 399 modified 492 N.W.2d 256; Sch. Dist. of Independence v. U.S. Gypsum (Mo.Ct.App. 1988) 750 S.W.2d 442, 456-457; Northridge Co. v. W.R. Grace and Co. (1991) 162 Wis.2d 918 [471 N.W.2d 179, 186].)

The insurers complain that “contamination” is not a legally defined term, and the trial court’s ruling makes the release of even a single asbestos fiber property damage. We find the complaint unfounded, as the insurers have failed to distinguish between the question whether Armstrong is liable for asbestos property damage and the question whether the insurance policies provide coverage. As to the former, whether and to what extent a release of asbestos fibers has damaged the buildings are factual issues for the underlying building cases. It may be that the trier of fact will conclude in a particular case that a low level of contamination was not damaging, and Armstrong then will have no liability and no need for insurance coverage. As we have explained in part A.l above, however, for purposes of determining the separate question of insurance coverage for the property damage claims, we assume, as did the trial court, that damage has occurred for which Armstrong will be held liable. We hold that as long as Armstrong is held liable for the release of asbestos fibers, whatever the level of contamination, the injury is a physical injury covered by the insurance policies.

Some of the underlying complaints in the building cases, however, allege that the mere presence of ACBM in buildings is a health hazard because of the potential for future releases of asbestos fibers. The complaints allege that common daily activities may cause asbestos fibers to be released from the ACBM and thus the ACBM poses a threat of harm. We conclude that even in such cases, when there have been no releases of asbestos fibers, if Armstrong is held liable for the mere presence of ACBM, the injury to the buildings is a physical one.

Once installed, the ACBM, whether in the form of insulating pipe coverings, fireproof floor tile, accoustical ceiling finishes, or the like, is physically linked with or physically incorporated into the building and therefore physically affects tangible property. We agree with the formulation put forth by the Seventh Circuit Court of Appeals that the term “physical injury” covers “a loss that results from physical contact, physical linkage, as when a potentially dangerous product is incorporated into another and, because it is incorporated and not merely contained (as a piece of furniture is contained in a house but can be removed without damage to the house), must be removed, at some cost, in order to prevent the danger from materializing.” (Eljer Mfg., Inc. v. Liberty Mut. Ins. Co. (7th Cir. 1992) 972 F.2d 805, 810 cert. den. (1993) 507 U.S. 1005 [123 L.Ed.2d 267, 113 S.Ct. 1646] [defective plumbing systems]; see also American Motorists Ins. Co. v. Trane Co. (7th Cir. 1983) 718 F.2d 842, 844 [defective heat exchangers, a component of a natural gas plant].)

The physical incorporation of ACBM into the buildings distinguishes the present case from those cases involving hazardous waste leaks or spills from containers. In the latter cases, the courts have held that remedial costs incurred in cleaning up contaminated waste sites are covered by CGL policies, but “prophylactic” costs—costs incurred in advance of any release of hazardous waste, to prevent threatened future pollution—are not incurred because of property damage. (AIU Ins. Co. v. Superior Court, supra, 51 Cal.3d 807, 843; Aerojet-Corp. v. Superior Court (1989) 211 Cal.App.3d 216, 237-238 [258 Cal.Rptr. 684]; Maryland Cas. Co. v. Armco, Inc. (4th Cir. 1987) 822 F.2d 1348, 1353, cert. den. (1988) 484 U.S. 1008 [98 L.Ed.2d 654, 108 S.Ct. 703].) In contrast, in the present case, because the potentially hazardous material is physically touching and linked with the building, and not merely contained within it, the injury is physical even without a release of toxic substances into the building’s air supply.

The insurers rely upon the rule that physical incorporation of a defective product into another does not constitute property damage unless there is physical harm to the whole. (St. Paul Fire & Marine Ins. Co. v. Coss (1978) 80 Cal.App.3d 888, 892 [145 Cal.Rptr. 836]; Hamilton Die Cast, Inc. v. United States F. & G. Co. (7th Cir. 1975) 508 F.2d 417, 419-420; see Maryland Casualty Co. v. Reeder (1990) 221 Cal.App.3d 961, 969-970 [270 Cal.Rptr. 719]; Fresno Economy Import Used Cars, Inc. v. United States Fid. & Guar. Co. (1977) 76 Cal.App.3d 272, 284 [142 Cal.Rptr. 681]; General Ins. Co. v. Intern. Sales Corp. (1977) 18 Wn.App. 180, 566 P.2d 966, 968-969].) In our view, however, that rule is designed to limit the liability coverage of contractors against claims of defective materials or poor workmanship, for such claims are a commercial risk which is not passed on to the liability insurer. (See Maryland Casualty Co. v. Reeder, supra, 221 Cal.App.3d at p. 967; Economy Lumber Co. v. Insurance Co. of North America (1984) 157 Cal.App.3d 641, 649-651 [204 Cal.Rptr. 135]; Rafeiro v. American Employers’ Ins. Co. (1970) 5 Cal,App.3d 799, 808 [85 Cal.Rptr. 701].) Here, Armstrong is facing liability not as a contractor but as a manufacturer or supplier of ACBM. The claims against Armstrong go beyond allegations of defective work or materials and allege injury to other property. (See discussion in part G, post.)

The insurers further argue that the mere presence of ACBM results only in economic losses—e.g., diminished property value, abatement costs, or costs of responding to the presence of asbestos—and not in a physical injury. They urge us to follow those cases which have construed the phrase “physical injury” so as to differentiate economic losses: e.g., Waller v. Truck Ins. Exchange, Inc. (1995) 11 Cal.4th 1, 17-18 [44 Cal.Rptr.2d 370, 900 P.2d 619]; New Hampshire Ins. Co. v. Vieira (9th Cir. 1991) 930 F.2d 696, 697-701; Federated Mut. Ins. Co. v. Concrete Units (Minn. 1985) 363 N.W.2d 751, 756; Wyoming Sawmills v. Transportation Ins. Co. (1978) 282 Or. 401 [578 P.2d 1253]. In our view, however, the damages allegedly suffered by the building owners from the presence of ACBM cannot be considered solely economic losses. Diminished market value or abatement costs or costs of inspecting, assessing, and maintaining the in-place ACBM are not the “property damage.” They are “damages because of property damage.” That is, they are the alternative measures of the physical injury to the building. (Maryland Cas. Co. v. W.R. Grace and Co. (2d Cir. 1993) 23 F.3d 617, 627, cert. den. (1994) _ U.S. _ [130 L.Ed.2d 559, 115 S.Ct. 655]; see Geddes & Smith, Inc. v. St. Paul Mercury Indemnity Co. (1959) 51 Cal.2d 558, 565 [334 P.2d 881], quoting from Hauenstein v. Saint Paul-Mercury Indem. Co. (1954) 242 Minn. 354 [65 N.W.2d 122, 125]; see also Geddes & Smith, Inc. v. St. Paul Mercury Indem. Co. (1965) 63 Cal.2d 602, 609 [47 Cal.Rptr. 564, 407 P.2d 868]; Mozzetti v. City of Brisbane (1977) 67 Cal.App.3d 565, 576 [136 Cal.Rptr. 751].) The fact that the measure of damages is economic does not preclude a physical injury. (Northridge Co. v. W.R. Grace and Co., supra, 471 N.W.2d 179, 184; see U.S. Fid. & Guar. v. Specialty Coatings (1989) 180 Ill.App.3d 378 [129 Ill.Dec. 306, 535 N.E.2d 1071, 1081].)

At trial, the insurers introduced evidence to show that the mere presence of ACBM is not necessarily injurious: that ACBM’s do not spontaneously emit asbestos fibers, nor do releases occur more frequently as the ACBM’s deteriorate with age; that left undisturbed, ACBM’s pose no health hazard to building occupants; and that even if the ACBM’s are occasionally disturbed such that asbestos fibers are released, the fibers are removed quickly by normal air circulation. That evidence, however, has no bearing on the insurance coverage issue before us. Once again, the insurers have failed to distinguish between questions of liability and questions of insurance coverage. Whether ACBM has actually caused harm is a question for the underlying building cases, and if the trier of fact finds that the mere presence of Armstrong’s products does not cause damage, then Armstrong will have no liability and no need for insurance coverage. The posture of the present case is such that for purposes of determining the separate question of insurance coverage we assume that the presence of ACBM is injurious and that Armstrong will be held liable for injuring the buildings. Given that assumption, we conclude that the alleged injury from installation of ACBM qualifies as “physical injury to . . . tangible property” under the terms of the policies.

B. Trigger of Coverage

1. Multiple Trigger

Having concluded that there is property damage, we must next decide when the property damage takes place, for purposes of determining which policies are responsible for indemnifying Armstrong. As we have discussed in Issue Group II, part A, ante, the courts have devised several approaches for determining when asbestos-related bodily injuries occur, and we have upheld the trial court’s use of a continuous trigger. In property damage cases, too, some courts have applied a continuous trigger. (E.g., California Union Ins. Co. v. Landmark Ins. Co. (1983) 145 Cal.App.3d 462 [193 Cal.Rptr. 461] [leaking swimming pool]; Hatco Corp. v. W.R. Grace & Co.—Conn., supra, 801 F.Supp. 1334, 1345 [hazardous waste]; Hirschberg v. Lumbermens Mut. Cas. (N.D.Cal. 1992) 798 F.Supp. 600, 603 [same]; New Castle County v. Continental Cas. Co. (CNA) (D.Del. 1989) 725 F.Supp. 800, 813, affd. in part and revd. in part (3d Cir. 1991) 933 F.2d 1162 [same]; Dayton Independent School D. v. National Gypsum, supra, 682 F.Supp. at pp. 1409-1410, reversed sub nom. W.R. Grace & Co. v. Continental Cas. Co., supra, 896 F.2d at pp. 875-876 [asbestos]; Lac d’Amiante du Quebec v. Am. Home Assur. (D.N.J. 1985) 613 F.Supp. 1549 [asbestos]; Owens-Illinois, Inc. v. United Ins. Co., supra, 650 A.2d 974, 983-984, 995 [asbestos]; Gottlieb v. Newark Ins. Co. (1990) 238 N.J.Super. 531 [570 A.2d 443] [toxic insecticides].)

In Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, our Supreme Court has held that a continuous trigger should be applied to claims of continuous or progressively deteriorating bodily injuries or property damage. The Montrose court observed, however, that proper resolution of a trigger of coverage issue may depend on whether the CGL policy insures against liability to third parties for bodily injury, property damage, or both. (Id. at pp. 665-666.) In Montrose, the coverage clauses in the CGL policies did not distinguish between the nature of the underlying harm, whether bodily injury or property damage. Accordingly, the parties in Montrose did not dispute that “under a plain reading of that unambiguous aspect of the policy language, whatever be the circumstances (or timing of the circumstances) that will potentially trigger liability coverage under the policies, coverage will apply uniformly under such circumstances whether the claims be for bodily injury, or property damage, alleged in the underlying third party lawsuits.” (10 Cal.4th at p. 666.)

Yet, in Montrose, the property damage and the bodily injuries were all alleged to be continuous or progressive. As the Montrose court put it, “. . . we are dealing both with claims of continuous or progressively deteriorating bodily injury . . . and progressively deteriorating property damage ... all arising from continuous or repeated exposure to hazardous waste contamination over time . . . .” (10 Cal.4th at p. 666.)

In Lac d’Amiante du Quebec v. Am. Home Assur., supra, 613 F.Supp. 1549, 1561 the court found that contamination of buildings from released asbestos fibers was both continuous and progressive: “release of [asbestos] fibers . . . may occur by a slow continuous degradation of the insulating surface which may be accelerated by the air movement and vibration which occurs in most buildings. . . . [T]he injury to property caused by asbestos is both continous and progressive and certainly not complete upon the act of installation.” Reasoning that it would be “illogical” to apply the continuous trigger to asbestos bodily injury claims but not to asbestos property damage, the court ruled that coverage was triggered at the time of installation, at the time of removal, and at all points in between.

In the present case, in contrast, the record indicates that releases of asbestos fibers, if they occur at all, occur sporadically, as a result of episodic disturbances such as accidental striking, vandalism, water damage, and the like. The trial court declined to apply a continuous trigger to the property damage claims, finding that asbestos property damage is not necessarily continuous: “the evidence presented to this court indicates that property damage from ACBM is not always continuous. The continuous trigger adopted by this Court in the bodily injury cases was based upon a finding that bodily injury from asbestos exposure was a continuous process beginning with first exposure. Because property damage from ACBM is not always continuous, the Court cannot adopt a comprehensive ‘continuous trigger’ approach as to which policies owe a duty to defend in the building cases. . . . [T]he Court declines to adopt a comprehensive rule stating that all policies from the time of installation until the time of removal of ACBM owe a duty to indemnify for property damage. The evidence presented in this phase as to the nature of property damage does not support the ‘continuous’ trigger approach adopted by this Court in the bodily injury case.”

Instead of a continuous trigger, the trial court adopted a multiple trigger: “Aside from property damage occurring at the time ACBM is installed, property damage also happens at any time asbestos fiber or material is released from ACBM into the air or on surfaces of the building, and when settled releases are disturbed and reentrained into the air. . . . The Court therefore holds that indemnity obligations are triggered if it is shown that ACBM was installed in the building or buildings in question, that ACBM released fiber or material into the air or on surfaces of the buildings in question, or that settled releases of ACBM were disturbed and reentrained into the air, during any portion of the period that the policy was in effect.”

Unlike the court in Lac d’Amiante, we do not find it illogical to apply different triggers for asbestos-related bodily injuries and asbestos-related property damage. The same legal rule applies in both instances: coverage is triggered when the injury actually occurred. (Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at pp. 669-670.) But the triggers are different because the injury to the human body upon inhalation of asbestos fibers is not the same as the injury to a building from the presence of ACBM. (See Maryland Casualty Co. v. W.R. Grace & Co., supra, 23 F.3d at p. 627.) We interpret the trial court’s decision to mean that in contrast to the continuous, progressive physiological process involved in the inhalation of asbestos, asbestos property damage is episodic, with measurable intervals between episodes, so that the process of injuries cannot be deemed continuous.

As the Supreme Court recognized in Montrose (10 Cal.4th at p. 694), whether the underlying damage or injury is in fact continuous is a matter for determination by the trier of fact. (Carey Canada, Inc. v. California Union Ins. Co. (D.D.C. 1990) 748 F.Supp. 8; see also Triangle Publications v. Liberty Mut. Ins. Co. (E.D.Pa. 1989) 703 F.Supp. 367, 371.) In light of the trial court’s factual finding that asbestos property damage is not always continuous, good reason exists for adopting a trigger different from the continuous trigger adopted for asbestos-related bodily injuries. (See Home Ins. Co. v. Landmark Ins. Co. (1988) 205 Cal.App.3d 1388, 1394-1395 [253 Cal.Rptr. 277] [property damage cases differ from asbestos bodily injury cases]; Amess & Eliason, Insurance Coverage for “Property Damage” in Asbestos and Other Toxic Tort Cases (1986) 72 Va. L.Rev. 943, 972-973.)

2. Release and Reentrainment as Triggers

In Maryland Cas. Co. v. W.R. Grace and Co., supra, 23 F.3d 617, 627-628, the Second Circuit Court of Appeals concluded that property damage occurs upon installation of asbestos products in a building, but the damage does not continue afterward. “Once installed, the damage that asbestos inflicts is complete. ... If [asbestos fibers are constantly released and re-entrained into a building’s atmosphere], its damaging effect concerns solely the health of those persons who breathe the contaminated air. No further property damage occurs because the need to remove or encapsulate the asbestos, which occurred upon the product’s installation, remains unchanged.” (Id. at p. 628.) Hence, the court applied a single trigger, holding that only the insurers on the risk at the time of installation were obligated to defend and indemnify the insured. (See also Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d 1178, 1210.)

As we have already said, the trial court concluded otherwise, that “property damage . . . happens at any time asbestos fiber or material is released from ACBM into the air or on surfaces of the building, and when settled releases are disturbed and reentrained into the air.” In deciding the trigger of coverage issue, the trial court ruled that the duty to indemnify is triggered “when it is shown that property damage occurred during any portion of the period that the policy was in effect. . . . The Court therefore holds that indemnity obligations are triggered if it is shown that ACBM was installed in the building or buildings in question, that ACBM released fiber or material into the air or on surfaces of the buildings in question, or that settled releases of ACBM were disturbed and reentrained into the air, during any portion of the period that the policy was in effect.”

The insurers complain that the trial court’s decision, taken literally, makes any release—even the release of a single asbestos fiber—enough to trigger coverage, despite the fact that the release was inconsequential and, by itself, would not constitute compensable property damage. 41 The insurers contend that coverage should not be triggered unless the fibers released during the policy period were “sufficiently appreciable.”

We reject the argument. The trigger question—which policies provide coverage—must be distinguished from the question whether the policies provide coverage. As to the latter issue, we reiterate that the trial court assumed, as we must do for purposes of deciding insurance coverage, that compensable property damage has occurred for which Armstrong will be held liable. As we have explained in part A.3 above, as long as Armstrong is held liable for contamination from the release of asbestos fibers, no matter what the level of contamination, the insurance policies provide coverage.

With respect to the separate question of when the property damage occurred for purposes of determining which policies are triggered, the fact that a particular release or reentrainment of asbestos fibers, by itself, might not give rise to liability is irrelevant. The property damage for which Armstrong may be held liable and for which the policies provide coverage is the contamination of the buildings from the introduction of asbestos fibers into the building air supply and onto building surfaces. We understand the trial court’s decision to mean that each release or reentrainment contributes to the state of contamination of the building: “[Property damage . . . happens at any time asbestos fiber or material is released from ACBM into the air or on surfaces of the building, and when settled releases are disturbed and reentrained into the air. The release of a harmful substance onto an area is a ‘physical injury to tangible property.’ [Citation.] The area becomes hazardous and certain measures must be taken to restore the surface to its prior condition. Release of asbestos material and fiber and reentrainment are all property damage events because each is an act of contamination which makes the building more hazardous.” Each release or reentrainment, then, is a part of the overall property damage giving rise to Armstrong’s liability.

We infer from the trial court’s multiple trigger decision that the total property damage may take place across several policy periods. That is, although property damage from release and reentrainment of asbestos fibers is episodic and not continuous, the property damage is continual in that new episodes of releases or reentrainments of asbestos fibers may occur repeatedly over time. There is nothing in the policies to preclude coverage from being triggered simply because only a part of the total damage occurred during any particular policy period. We therefore conclude that when, as here, property damage takes place during several policy periods, the trial court correctly ruled that insurance coverage is triggered when any part of the damage—any release or reentrainment—takes place.

In sum, we find no error in the trial court’s decision that if Armstrong is held liable for contamination of a building from released asbestos fibers, each policy is triggered if any part of the contamination damage, no matter how small the quantity of the released fibers, took place when the policy was in effect. 42

The insurers further argue that triggering coverage upon reentrainment violates the loss-in-progress rule, as resuspension of settled fibers is a continuation of the loss that began when the fibers were released from the ACBM. The argument is unsound. The loss-in-progress rule provides that an insurer can insure only against a contingent or unknown loss. (Ins. Code, §§ 22, 250) The rule does not apply if the damage was unknown or contingent or the insured’s liability was uncertain at the time the policy was issued, even if the damage was inevitable. (Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at pp. 689-693.) As long as Armstrong’s liability for the reentrainments was uncertain when the policy was issued, the loss-in-progress rule does not preclude coverage.

3. Installation as Trigger

The insurers contend that the selection of the installation of ACBM as an event triggering insurance coverage violates the rule that coverage is triggered by “the event causing the actual injury and not an earlier event which created the potential for future injury.” (Hallmark Ins. Co. v. Superior Court (1988) 201 Cal.App.3d 1014, 1018 [247 Cal.Rptr. 638]; Maples v. Aetna Cas. & Surety Co. (1978) 83 Cal.App.3d 641, 647-648 [148 Cal.Rptr. 80]; see also Whittaker Corp. v. Allianz Underwriters, Inc. (1992) 11 Cal.App.4th 1236 [14 Cal.Rptr.2d 659]; Millers Mut. Fire Ins., etc. v. Ed Bailey (1982) 103 Idaho 377 [647 P.2d 1249].) The insurers argue that installation of the ACBM did not inflict injury at the moment of installation; rather, there was a time gap between the installation and the injury, and coverage should be triggered when costs were actually incurred for removing or neutralizing and maintaining the asbestos materials or when the building’s market value was actually reduced.

We conclude, however, that the trial court’s decision fully conforms to the “actual injury” requirement. The underlying complaints allege that the installation of ACBM in a building created a health hazard to the building occupants—a hazard that has come to light only in recent years as a result of governmental reports and increased knowledge. This hazard occurred no less at installation than upon realization of the dangers of asbestos. The damage was done as soon as the ACBM was installed. (Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d at p. 1209; Maryland Cas. Co. v. W.R. Grace and Co., supra, 23 F.3d at p. 627.)

The Seventh Circuit Court of Appeals has held that in light of the purposes of insurance, property damage occurs in the policy year in which a defective product is installed, rather than the policy year in which it fails or is replaced in anticipation of failure or causes the market value of the building to diminish. “[T]he incorporation of a defective product into another product inflicts physical injury in the relevant sense on the latter at the moment of incorporation—here, the moment when the defective Qest [plumbing] systems were installed in homes.” (Eljer Mfg., Inc. v. Liberty Mut. Ins. Co., supra, 972 F.2d at p. 814; see also Colonial Gas Co. v. Aetna Cas. & Sur. Co. (D.Mass. 1993) 823 F.Supp. 975 [installation of “UFFI” insulation].) We are persuaded by that view and find it applicable to asbestos products.

Our conclusion is supported by the line of property damage cases in which the courts have held that injury occurs and coverage is triggered immediately upon first exposure to the hazardous material or latent defect. (California Union Ins. Co. v. Landmark Ins. Co., supra, 145 Cal.App.3d 462 [water seepage from leaking swimming pool]; Ray Industries, Inc. v. Liberty Mut. Ins. Co. (6th Cir. 1992) 974 F.2d 754, 764-766 [dumping of hazardous waste]; Continental Ins. v. N.E. Pharm. & Chem. Co. (8th Cir. 1987) 811 F.2d 1180, 1189-1192, vacated on other grounds upon rehearing en banc 842 F.2d 977, cert. den. 488 U.S. 821 [102 L.Ed.2d 43, 109 S.Ct. 66] [same]; Trizec Properties v. Biltmore Const. Co. (11th Cir. 1985) 767 F.2d 810, 813 [latent defects in roof]; Lac d’Amiante du Quebec v. Am. Home Assur., supra, 613 F.Supp. at pp. 1560-1561 [release of asbestos]; Gruol Construction Co. v. Insurance Co. of No. Amer. (1974) 11 Wn.App. 632 [524 P.2d 427] [dry rot of foundation from defective backfilling].) 43

We emphasize that for purposes of determining insurance coverage, we have assumed, as did the trial court, that Armstrong will be held liable for the damages alleged in the underlying complaints. Some of those complaints do not allege contamination from released asbestos fibers, but allege only that asbestos is present in the buildings, posing a threat of future releases. We hold that in the event that Armstrong is held liable for the mere presence of ACBM, without evidence of contamination from released asbestos fibers, coverage is triggered at the time ACBM was installed in the building. (As we explain in footnote 42 ante, where the basis of liability is the mere presence of ACBM, not contamination from released fibers, coverage would not also be triggered by subsequent, incidental releases of asbestos fibers.)

Moreover, even in those jurisdictions in which the mere presence is not enough to give rise to liability, if Armstrong is held liable for contamination of the building from released asbestos fibers, coverage is triggered at the time of installation as well as at the time of release. Our reasoning parallels that set forth in part B.2 above. The fact that the mere presence of ACBM, by itself, might not give rise to liability for property damage is of no consequence if the insured has, in fact, been held liable for property damage. The installation of ACBM obviously contributes to the state of contamination of the building; it is a part of the overall property damage for which the insured is liable. The trial court correctly concluded that insurance coverage is triggered when any part of the damage—installation, release, or reentrainment—took place.

The insurers urge us to reject the trial court’s multiple trigger and to follow Prudential-LMI Com. Insurance v. Superior Court (1990) 51 Cal.3d 674, 694-699 [274 Cal.Rptr. 387, 798 P.2d 1230], to conclude instead that property damage occurs on a single date of loss, the date of “manifestation.” The Supreme Court, however, has held that the rationale of Prudential-LMI, adopting a manifestation trigger of coverage for first party cases, is inapposite in the context of third party liability policies. (Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, 663-666, 683-685, 687-689.) Indeed, here the unique facts surrounding property damage from ACBM provide good reasons to reject a manifestation trigger.

The manifestation rule deems damage to occur only when the damage has become apparent. Yet, the health hazard allegedly created by the presence of ACBM occurred no less at installation than upon the realization of the dangers of asbestos. That the building owners were unaware of the dangers until years after installation of the ACBM does not mean that the hazard was not present or that damage did not occur. Nor does the fact that the damage was not susceptible of measurement until the dangers of asbestos were known obviate the occurrence of injury at an earlier time. (Maryland Casualty Co. v. W.R. Grace and Co., supra, 23 F.3d at pp. 626-627.)

As one court has made clear, the manifestation rule creates a legal fiction: “There are situations ... in which the existence or scope of damage remains concealed or uncertain for a period of time even though damage is occurring. The leakage of hazardous wastes . . . is a clear example. Determining exactly when damage begins can be difficult, if not impossible. In such cases we believe that the better rule is that the occurrence is deemed to take place when the injuries first manifest themselves.” (Mraz v. Canadian Universal Ins. Co., Ltd. (4th Cir. 1986) 804 F.2d 1325, 1328.) With respect to asbestos building cases, however, as we have explained, the alleged injury to the buildings first occurred when the ACBM was installed. That date is ascertainable; there is no need for a fictional date of injury. (See Maryland Cas. Co. v. W.R. Grace and Co., supra, 23 F.3d at pp. 627-628.)

Finally, a manifestation trigger would place the entire burden for property damage claims upon those insurers who were on the risk in later years, when the dangers from ACBM were perceived. Despite the fact that Armstrong paid insurance premiums to a number of companies over the years to insure it against the risk of property damage, only a small group of insurers would be liable.

Although the manifestation rule does serve to promote certainty in the insurance industry and to avoid the problems of apportionment among insurers (Prudential-LMI Com. Insurance v. Superior Court, supra, 51 Cal.3d at p. 699; Home Ins. Co. v. Landmark Ins. Co., supra, 205 Cal.App.3d at pp. 1395-1396), we reject the manifestation rule as an appropriate trigger for insurance coverage of asbestos property damage. We conclude that in asbestos property damage cases coverage is triggered pursuant to the rule applicable generally in third party liability cases, i.e., when the claimant’s injury actually took place.

We affirm the trial court’s decision that insurance coverage is triggered if any part of the underlying property damage—installation, release, or reentrainment—took place during a policy period.

C. Accident Policies

From 1942 to 1951, Armstrong was insured by Standard Accident Insurance Company, the predecessor of Reliance Insurance Company, whose policies provided coverage for property damage “caused by accident.” Reliance argued below that its policies provide no coverage for damage from ACBM because, unlike an occurrence policy, an accident policy requires a sudden, unexpected event. The trial court rejected the argument for two reasons: (1) there is no requirement of suddenness; and (2) the release and reentrainment of asbestos fibers qualify as sudden events.

We disagree with the trial court on the first point. The trial court reasoned that “[i]n ordinary language, an unexpected and unintended event is viewed as an accident." Although we agree that the word “accident” carries the meaning of unexpected and unintended (Hogan v. Midland National Ins. Co. (1970) 3 Cal.3d 553, 559-561 [91 Cal.Rptr. 153, 476 P.2d 825]; Geddes & Smith, Inc. v. St. Paul Mercury Indemnity Co., supra, 51 Cal.2d 558), we conclude that an accident policy covers only unexpected and unintended events which are also sudden. 44 (Geddes & Smith, Inc. v. St. Paul Mercury Indemnity Co., supra, 51 Cal.2d at pp. 563-564; Shell Oil Co. v. Winterthur Swiss Ins. Co., supra, 12 Cal.App.4th 715, 751-752; see Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, 672; 1 Long, The Law of Liability Insurance (1993) § 1.08[1], p. 1-50.)

The trial court found, however, that the suddenness requirement was fulfilled by the releases and reentrainments of asbestos fibers: “Reliance’s arguments assume a finding that . . . damage occurs solely upon installation. However, as discussed, property damage from ACBM occurs episodically upon release and reentrainment of material and fibers. Episodic ACBM releases fulfill the more restrictive sudden and fortuitous requirement Reliance suggests and therefore the events are ‘caused by accident.’ These releases trigger the policies immediately after the ACBM is installed.”

Reliance does not challenge the trial court’s finding that the releases qualify as sudden events. Instead, Reliance argues that past releases or reentrainments of asbestos fibers cannot trigger coverage because no actual injury occurred at the time of those events; the injury occurred later, when costs were actually incurred and property value actually diminished. Insofar as this argument advocates a manifestation trigger, we reject the argument for the reasons explained in part B.3. Moreover, as we have explained in part A.3, response costs and diminished property values are not the “property damage”; they are the measures of the property damage. The property damage is the contamination of the buildings from the introduction of asbestos fibers into the building air supply and onto building surfaces. Each release or reentrainment contributes to the state of contamination and forms a part of the overall property damage. Accordingly, for the reasons we have explained in part B.2 above, insofar as Armstrong is held liable for contamination of a building from released asbestos fibers, the Reliance policies are triggered if any part of the contamination damage took place when the policies were in effect. 45

D. Scope of Coverage

Under the trial court’s multiple trigger decision, several successive policies from the date of installation to the date of manifestation may be triggered on a single claim. Thus the question arises how the coverage should be allocated among the insurers whose policies are triggered.

In the present case, the trial court concluded that “ [although the factual basis of the property damage claims differs from that of the bodily injury claims, the Court finds no reason to distinguish property damage from bodily injury for the purposes of the scope and allocation issues raised in phase V.” Hence, the court reached the same conclusions it reached with respect to the bodily injury claims (discussed in Issue Group II, part A, ante): The court held each of the triggered policies responsible “in full” for the losses, subject to the “no stacking” qualification (only one policy’s limits can apply to each claim) and subject to apportionment among the insurers based upon “other insurance” clauses. And the court ruled that Armstrong has no obligation to share pro rata in indemnification or defense costs because of any uninsured or self-insured periods. The insurers object to the latter ruling, that Armstrong need not pay a pro rata share of the damages.

The basic insuring agreement of the CGL policies obligates the insurer to pay “all sums which the insured shall become legally obligated to pay as damages” because of property damage. The occurrence policies define an “occurrence” as “an accident, including injurious exposure to conditions, which results during the policy period in . . . [property damage].” In addition, the 1973 policies define property damage as “physical injury to . . . tangible property which occurs during the policy period.” The insurers argue that each individual insurance policy can be called upon to pay only for property damage that took place during the policy period; hence, Armstrong must pay for any damage that took place while it was uninsured or self-insured. 46

We disagree. The insurers have confused the trigger of coverage and the scope of coverage. As we have explained in Issue Group II, part A, ante, the event which triggers an insurance policy’s coverage does not define the extent of the coverage. Although a policy is triggered only if property damage takes place “during the policy period,” once a policy is triggered, the policy obligates the insurer to pay “all sums” which the insured shall become liable to pay as damages for bodily injury or property damage. The insurer is responsible for the full extent of the insured’s liability (up to the policy limits), not just for the part of the damage that occurred during the policy period. (Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d 1034, 1047-1048; see also Hatco Corp. v. W.R. Grace & Co.—Conn., supra, 801 F.Supp. at pp. 1345-1347; New Castle County v. Continental Cas. Co. (CNA), supra, 725 F.Supp. 800, 817; Dayton Independent School D. v. National Gypsum, supra, 682 F.Supp. at p. 1410; Lac d’Amiante du Quebec v. Am. Home Assur., supra, 613 F.Supp. 1549, 1562; Monsanto Co. v. C.E. Heath Comp. & Liability, supra, 652 A.2d 30, 35; J.H. France Refractories v. Allstate, supra, 626 A.2d 502, 508.) It follows, then, that the insured need not pay a pro rata share for periods during which it had no insurance. (Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d 1034.)

We recognize that there is language in some cases to suggest that the insurers’ obligations to pay for the full extent of the policyholder’s liability apply only if the injury was continuous and indivisible; otherwise, apportionment will be allowed. (Hatco Corp. v. W.R. Grace & Co.—Conn., supra, 801 F.Supp. at pp. 1345-1347; Lac d’Amiante du Quebec v. Am. Home Assur., supra, 613 F.Supp. at pp. 1562-1563; Sandoz, Inc. v. Employer’s Liability Assur. Corp. (D.N.J. 1983) 554 F.Supp. 257, 266; Diamond Shamrock Chemicals v. Aetna (1992) 258 N.J.Super. 167 [609 A.2d 440, 467].)

As we have explained in Issue Group II, part A, ante, however, apportionment among multiple insurers must be distinguished from apportionment between an insurer and its insured. When multiple policies are triggered on a single claim, the insurers’ liability is apportioned pursuant to the “other insurance” clauses of the policies (Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d 1034, 1049) or under the equitable doctrine of contribution (Signal Companies, Inc. v. Harbor Ins. Co. (1980) 27 Cal.3d 359, 369 [165 Cal.Rptr. 799, 612 P.2d 889]; CNA Casualty of California v. Seaboard Surety Co. (1986) 176 Cal.App.3d 598, 619-620 [222 Cal.Rptr. 276]). That apportionment, however, has no bearing upon the insurers’ obligations to the policyholder. (See Dayton Independent School D. v. National Gypsum, supra, 682 F.Supp. at pp. 1410-1411 and fn. 21.) A pro rata allocation among insurers “does not reduce their respective obligations to their insured.” (Sandoz, Inc. v. Employer’s Liability Assur. Corp., supra, 554 F.Supp. at p. 267.) The insurers’ contractual obligation to the policyholder is to cover the full extent of the policyholder’s liability (up to the policy limits).

Moreover, in our view, the scope of coverage does not depend upon the continuous, indivisible nature of the damage or the application of a continuous trigger. A continuous injury merely gives rise to the scope of coverage issue by triggering several successive policies on an individual claim. We believe the insurers have the same obligation to respond in full when several successive policies are triggered by continual, episodic property damage. In both situations the claimant’s overall damage for which the insured is liable is unitary. It is irrelevant that the damage took place across several policy periods and only a part of the damage occurred during any particular policy period. The plain language of the policies requires that each triggered policy respond in full.

The Keene court said as much in dictum: “If each exposure is considered a separate ‘injury,’ under the terms of the policies, one might be able to argue that each insurer is responsible only for the ‘injuries’ that occurred during its policy periods . . . . It is clear, however, that such a result would be contrary to the terms of the insurance policies, which explicitly state that the insurer will pay ‘all sums which the insured shall become legally obligated to pay as damages because of bodily injury [during the policy period].’ As long as there was either inhalation exposure or exposure in residence during a policy period, and as long as Keene must pay damages as a result, the insurer must indemnify Keene for whatever damages it must pay. . . .” (Keene Corp. v. Ins. Co. of North America, supra, 667 F.2d at pp. 1044-1045, fn. 20.)

In the present case, we construe the trial court’s decision to mean that although asbestos property damage is composed of a series of discrete injuries, each injury constitutes an ingredient of the overall property damage for which Armstrong is liable. Each release or reentrainment of asbestos fibers, along with the installation of the ACBM, forms a part of the unitary property damage for which Armstrong is alleged to be liable. We follow the

ruling in Keene and conclude that as long as there was property damage to a building during a policy period, whether from installation of ACBM or from releases or reentrainments of asbestos fibers from existing ACBM, and as long as Armstrong must pay damages as a result of that property damage, the policies provide coverage (up to the policy limits) for whatever damages Armstrong must pay. 47 Armstrong need not pay a pro rata share.

E. Duty to Indemnify

The uncertainty of Armstrong’s underlying liability in the building cases prompted us to pose the question in a request for supplemental briefing whether declaratory relief is appropriate here, whether the determination of coverage for property damage should await resolution of Armstrong’s liability in the underlying actions and a determination of the actual types of damages for which Armstrong is liable. 48 We find it unnecessary to analyze that issue in depth, as both sides have agreed there is an actual controversy on the meaning of the policies’ language and no reason to defer a ruling on questions of interpretation. 49

The insurers argue, however, that the duty to indemnify, as distinct from the coverage of the insurance policies, cannot be determined in advance of the insured’s underlying liability. The insurers correctly point out that the duty to indemnify is different from the duty to defend. As the court explained in Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th 645, 659, fn. 9, the duty to defend arises when there is a potential for indemnity. (Horace Mann Ins. Co. v. Barbara B. (1993) 4 Cal.4th 1076, 1081 [17 Cal.Rptr.2d 210, 846 P.2d 792]; Gray v. Zurich Insurance Co. (1966) 65 Cal.2d 263, 276 [54 Cal.Rptr. 104, 419 P.2d 168].) It may exist even when coverage is in doubt and ultimately does not develop. (Horace Mann, supra; Saylin v. California Ins. Guarantee Assn. (1986) 179 Cal.App.3d 256, 263 [224 Cal.Rptr. 493].) The duty to indemnify, on the other hand, arises when the insured’s underlying liability is established. (Civ. Code, § 2778, subd. 1; Clark v. Bellefonte Ins. Co. (1980) 113 Cal.App.3d 326, 336-337 [169 Cal.Rptr. 832]; Alberts v. American Casualty Co. (1948) 88 Cal.App.2d 891, 899-900 [200 P.2d 37]; see 1 Long, The Law of Liability Insurance, supra, § 1.03[4], p. 1-11.) The duty to indemnify on a particular claim is determined by the actual basis of liability imposed on the insured. (City of Laguna Beach v. Mead Reinsurance Corp. (1990) 226 Cal.App.3d 822, 829-832 [276 Cal.Rptr. 438]; International Surplus Lines Ins. Co. v. Devonshire Coverage Corp. (1979) 93 Cal.App.3d 601, 610 [155 Cal.Rptr. 870].) Although an insurer may have a duty to defend, it may ultimately have no duty to indemnify—either because no damages were awarded or because the actual judgment was for damages not covered by the policy. (See City of Laguna Beach v. Mead Reinsurance Corp., supra, 226 Cal.App.3d at p. 830; 1 Cal. Liability Insurance Practice: Claims & Litigation (Cont.Ed.Bar 1992) § 4.6, p. 4-6.2.)

Thus, the question whether an insurer has a duty to indemnify the insured on a particular claim is ripe for consideration only if the insured has already incurred liability in the underlying action. (Aetna Cas. & Sur. Co. v. PPG Industries, Inc. (D.Ariz. 1983) 554 F.Supp. 290, 296; Outboard Marine v. Liberty Mut. Ins. (1992) 154 Ill.2d 90 [180 Ill.Dec. 691, 607 N.E.2d 1204, 1221]; USF&G v. Wilkin Insulation Co., supra, 578 N.E.2d at p. 930.) In a declaratory relief action held before the insured’s liability has been established, the trial court cannot determine the amount of the insured’s indemnity obligation; it must limit its declaration to whether the claim is covered by the policy. (Aitchison v. Founders Ins. Co. (1958) 166 Cal.App.2d 432, 439 [333 P.2d 178]; see Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at p. 659, fn. 9.) 50

In the present case, the insurers argue that the trial court should have denied Armstrong’s request for a declaration of the insurers’ duty to indemnify either because the request was premature or because Armstrong failed to prove its actual liability. We reject the argument for several reasons.

First, the trial court’s declaration of the insurers’ obligations to indemnify acknowledges the prematurity of Armstrong’s request; the declaration is conditional: “[Ijndemnity obligations are triggered if it is shown that ACBM was installed in the building or buildings in question, that ACBM released fiber or material into the air or on surfaces of the buildings in question, or that settled releases of ACBM were disturbed and reentrained into the air, during any portion of the period that the policy was in effect.” (Italics added.) “The building claims trigger the indemnity obligations of any policy that has indemnity obligations if it is shown that covered property damage occurred during any portion of the period that the policy was in effect.” (Italics added.)

From this language we infer that the trial court contemplated future proceedings between Armstrong and its insurers, after Armstrong’s liability has been established, in which Armstrong would provide proof that the conditions for triggering coverage had been met in a particular case. We note that the trial court made no determination as to which policies would cover any particular claims; that issue was obviously left for future proceedings.

Second, we reject the insurers’ contention that Armstrong failed to prove that its liability will necessarily result from covered property damage. As we have explained in part A.l above, for purposes of deciding the coverage dispute, the trial court properly looked to the allegations of the underlying complaints and assumed Armstrong would be held liable for the damages alleged therein. The underlying complaints allege liability arising either from the release of asbestos fibers or from the mere installation of ACBM. Armstrong was not required to prove more.

The gist of the insurers’ argument is not that Armstrong failed to prove its case but that the trial court erroneously found the building claims to constitute covered property damage. Because we have upheld the trial court’s decision and have concluded that installation of ACBM and releases of asbestos fibers do qualify as “physical injury to tangible property,” we find no error in the trial court’s declaration that Armstrong is entitled to indemnification if Armstrong is held liable for the damages alleged in the underlying complaints.

F. Duty to Defend

The trial court’s decision in phase V on property damage addressed the indemnification obligations of the insurers but said little about the duty to defend: “Each triggered policy has an independent obligation to pay in full any indemnity costs on an asbestos building claim and, if the policy contains a defense obligation, to also pay in full any defense costs on the claim.” 51

In a separate brief, Liberty Mutual Insurance Company raises the issue of the duty to defend and argues that mere conclusory allegations of “property damage” in the underlying complaints should not be enough to give rise to a duty to defend. Liberty Mutual asserts that the duty to defend should arise only if the underlying complaints allege actual contamination of the building and not merely a potential health hazard from the presence of ACBM.

Liberty Mutual is correct that as a general rule conclusory allegations are not enough to give rise to a duty to defend. (See Fire Ins. Exchange v. Jiminez (1986) 184 Cal.App.3d 437, 443, fn. 2 [229 Cal.Rptr. 83].) 52 But we construe Liberty Mutual’s argument as an alternate attack on the trial court’s ruling that the mere presence of ACBM in a building constitutes “physical injury to tangible property.” For the reasons explained in part A(3) above, physical injury occurs even in those cases in which there have not yet been any releases of asbestos fibers. Allegations of damages from the presence of ACBM in a building are therefore sufficient to show a potential for coverage and to give rise to a duty to defend.

G. Exclusions

The trial court was also asked to determine the applicability of various policy exclusions—the so-called “business risk” exclusions. One commentator has explained the rationale for the business risk exclusions as follows: “In every business venture there is an element of risk; the product fails to perform as expected; it lacks appeal to the consumer and does not sell; or it actively malfunctions and a third party suffers a loss. Products liability coverage is designed to protect only the bodily injury or property damage of others and not the business risks that accompany every commercial venture.” (7A Appelman, Insurance Law & Practice (rev. ed. 1979) § 4508.01, p. 353.)

The trial court concluded that none of the business risk exclusions apply to the asbestos building cases. The insurers first contend that the trial court erred in issuing such a sweeping ruling. They argue that a decision on the applicability of the exclusions must await a determination of the actual basis for Armstrong’s liability, for it may turn out that a particular plaintiff will recover damages which are excluded by the policy. (See Central Mutual Ins. Co. v. Del Mar Beach Club Owners Assn. (1981) 123 Cal.App.3d 916, 930-931 [176 Cal.Rptr. 895].)

We draw the same distinction here that we drew with respect to the coverage dispute discussed in part A above and the duty to indemnify discussed in part E above. For purposes of determining whether the property damage claims are covered or excluded under the insurance policies, we assume that Armstrong will be held liable for the damages alleged in the complaints in the underlying building cases. (Cf. Maryland Cas. Co. v. W.R. Grace & Co., supra, 794 F.Supp. at pp. 1227, 1228 [summary judgment inappropriate to determine application of exclusions in the absence of factual development of the underlying claims].) We express no opinion on whether the damages alleged are sufficient to give rise to liability, nor do we determine whether the exclusions apply in the event Armstrong is actually held liable for damages other than those alleged in the underlying complaints. (See Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at p. 694.)

1. Insured’s Own Products

The first exclusion put forth by the insurers bars coverage for damage to the insured’s own products. The 1966 and 1973 standard form CGL policies exclude “property damage to . . . the named insured’s products arising out of such products or any part of such products.” The earlier version (the 1947 standard form) excluded “injury or destruction of. . . any goods or products manufactured, sold, handled, or distributed or premises alienated by the named insured, or work completed by or for the named insured, out of which the accident arises.”

The earlier version was explained by the Court of Appeal to mean that “ ‘if the insured becomes liable to replace or repair any “goods or products” . . . after the same has caused an accident because of a defective condition, the cost of such replacement or repair is not recoverable under the policy. However, if the accident also caused damage to some other property or caused personal injury, the insured’s liability for such damage or injury becomes a liability of the insurer under the policy, and is not excluded.’ ” (Central Mutual Ins. Co. v. Del Mar Beach Club Owners Assn., supra, 123 Cal.App.3d at p. 929, quoting Liberty Bldg. Co. v. Royal Indem. Co. (1960) 177 Cal.App.2d 583, 587 [2 Cal.Rptr. 329], italics in original.)

The trial court ruled the “own products” exclusion inapplicable because the underlying building cases allege damage to the remainder of the building, not damage to the ACBM. A similar conclusion was reached by the Illinois Supreme Court in USF&G v. Wilkin Insulation Co., supra, 578 N.E.2d 926, 934. The court reasoned that the underlying complaints seeking recovery for the costs of removal, repair and/or replacement of the ACBM “seek these damages as a result of the contamination visited upon the buildings and the contents therein by the product that Wilkin installed. As such, the underlying complaints seek recovery for damage to property other than the product installed by Wilkin . . . .” (Italics in original; see also Dayton Independent School D. v. National Gypsum, supra, 682 F.Supp. at p. 1412 [“the ‘own product’ exclusion does not bar coverage for the damages alleged to have been sustained by the Plaintiffs’ buildings or their contents”]; accord, Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d at p. 1210; Maryland Cas. Co. v. W.R. Grace & Co., supra, 794 F.Supp. at p. 1227; and see Aetna Cas. & Sur. Co. v. PPG Industries, Inc., supra, 554 F.Supp. at p. 294 [damage from foam insulation was damage to the building, not to the insulation itself; hence, exclusion not applicable].)

The reasoning in those preceding asbestos property damage cases conforms to rulings by California courts in analogous settings. (Geddes & Smith, Inc. v. St. Paul Mercury Indem. Co., supra, 63 Cal.2d 602, 606-608 [defective doors caused damage to whole house]; Economy Lumber Co. v. Insurance Co. of North America, supra, 157 Cal.App.3d 641, 649-650 [defective siding caused damage to whole house]; cf. Volf v. Ocean Accident & Guar. Corp. (1958) 50 Cal.2d 373, 375-376 [325 P.2d 987] [defective cement damaged only the wall, and the wall was the insured-contractor’s own product]; Diamond Heights Homeowners Assn. v. National American Ins. Co. (1991) 227 Cal.App.3d 563, 572-573 [277 Cal.Rptr. 906] [various construction defects damaged the house, but the whole house was the insured-contractor’s own product].)

Relying on Volf and Diamond Heights, the insurers argue that because the underlying claimants are seeking the costs of repairing or replacing only the ACBM, and not any other part of the building, only damage to the ACBM is at issue in the building cases. Yet, as explained in part A.l above, we have assumed, for purposes of interpreting the policy language, that the buildings themselves have been injured by the ACBM. We hold only that insofar as Armstrong is held liable for the claimed damage to the buildings, the “own products” exclusion does not bar coverage.

2. Product Recall

The second exclusion advanced by the insurers is the product recall or “sistership” exclusion, which is expressed in the standard policy as follows: “Damages claimed for the withdrawal, inspection, repair, replacement or loss of use of the named insured’s products or work completed by or for the named insured or of any property of which such products or work form a part, if such products, work or property are withdrawn from the market or from use because of any known or suspected defect or deficiency therein.”

The trial court found this exclusion, too, inapplicable to the building cases, and we affirm that ruling.

The term “sistership” stems from the practice in the aircraft industry of recalling planes for repairs when a plane of the same model—a sister ship—had crashed because of a design defect. (Arcos Corporation v. American Mutual Liability Ins. Co. (E.D.Pa. 1972) 350 F.Supp. 380, 384, fn. 2; Annot. (1984) 32 A.L.R.4th 630.) The sistership exclusion “operates to exclude coverage for the cost of ‘preventative or curative action’ when the insured withdraws a product in situations in which a danger is merely apprehended. [Citation.] It does not, however, operate to exclude coverage for actual damage caused by the very product giving rise to such an apprehension.” (Dayton Independent School D. v. National Gypsum, supra, 682 F.Supp. at p. 1412 [asbestos building products], quoting from Todd Shipyards Corp. v. Turbine Service, Inc. (5th Cir. 1982) 674 F.2d 401, 419 cert. den. 459 U.S. 1036 [74 L.Ed.2d 602, 103 S.Ct. 447] [malfunctioning turbine]; accord, Stonewall Ins. Co. v. Asbestos Claims Management, supra, 73 F.3d at p. 1211 [asbestos building products]; Maryland Cas. Co. v. W.R. Grace & Co., supra, 794 F.Supp. at p. 1227 [same]; see also Gulf Mississippi Marine Corp. v. George Engine Co. (5th Cir. 1983) 697 F.2d 668 [defective engines and gears]; Marathon Plastics v. Intern. Ins. Co. (1987) 161 Ill.App.3d 452, [112 Ill.Dec. 816, 514 N.E.2d 479, 487] [defective pipes].)

The insurers argue that because the underlying complaints allege that the mere presence of ACBM poses only a potential health risk, the removal of ACBM from the buildings is based upon an apprehension of danger, not actual damage. The argument is flawed in two respects. First, it ignores the allegations of damage from contamination from asbestos fibers released into the air supply or onto surfaces of the buidings. Second, it ignores the fact that we have assumed, as did the trial court, that actual damage to the buildings has resulted from the mere presence of still-contained ACBM for which Armstrong will be held liable. Insofar as Armstrong is ultimately held liable for such damage, the sistership exclusion does not bar coverage.

3. Design Defect

The third exclusion put forth by the insurers was the design defect exclusion. The 1966 standard form CGL excludes: “Property damage resulting from the failure of the named insured’s products or work completed by or for the named insured to perform the function or serve the purpose intended by the named insured, if such failure is due to a mistake or deficiency in any design, formula, plan, specifications, advertising material or printed instructions prepared or developed by any insured; but this exclusion does not apply to . . . property damage resulting from the active malfunctioning of such products or work.”

In American Employers’ Ins. Co. v. Maryland Casualty Co. (1st Cir. 1975) 509 F.2d 128, 130, the court explained the “active malfunctioning” exception to this exclusion: “Design errors resulting in mere ‘passive’ failure to discharge an intended function are regarded as the insured’s normal business risk and are excluded from coverage, while design errors themselves causing some positive or ‘active’ harm deemed extraordinary in the insured’s business are covered. Thus, to recite some of the hypotheticals appearing in commentaries dealing with the clause, the policy is not intended to cover liability resulting from the faulty design of an insecticide which fails to kill insects, a hair tonic which fails to prevent baldness, or a rust inhibitor which fails to inhibit rust. On the other hand, the active malfunctioning exception would apply to provide coverage for liability resulting from an insecticide which harms crops to which it is applied, a hair tonic which causes a scalp rash, or a rust inhibitor which corrodes a radiator to which it is added.” (Fns. omitted.)

In the present case, the trial court found that the “active malfunctioning” exception applies to make the design defect exclusion inapplicable: “In the instant case, the alleged defect in the ACBM is not a ‘passive’ failure to insulate or perform any of the normal functions expected of floor tile, pipe insultation or surfacing material. Rather, as resolved earlier in this phase, the building cases allege that a positive harm results from ACBM. The alleged contamination of the buildings from ACBM is closely analogous to the hypotheticals considered within the scope of the active malfunctioning clause. As in the above examples, the type of resulting harm is a side effect of the product and has nothing to do with the product failing to perform its primary purpose. Therefore, the active malfunctioning limitation applies to the building cases and renders the 1966 CGL design defect exclusion inapplicable.”

The insurers do not challenge this ruling as applied to cases in which the underlying complaints allege asbestos fibers have been released and have contaminated the building. The insurers argue, however, that in cases alleging damages from the mere presence of still-contained ACBM, there is no “active malfunctioning.” Yet, this argument seems to be simply another presentation of the point that the mere presence of still-contained ACBM is not a physical injury. As we have explained in part A.l above, we must assume for purposes of deciding coverage that the presence of ACBM is injurious and will be the basis of Armstrong’s liability to the building owners. And as explained in part A.3, we have concluded that injury from the presence of ACBM qualifies as a physical injury within the meaning of the policies. We therefore hold that insofar as Armstrong is held liable for injuries from the presence of ACBM, the design defect exclusion is inapplicable.

H. Excess Policies: Aggregate or Catastrophe *

In summary, we affirm the trial court’s decision that all claims in the underlying building cases, whether for releases of asbestos fibers or for the mere presence of ACBM, qualify as claims for “physical injury to . . . tangible property” and are covered by the insurance policies. We affirm the trial court’s decision that coverage is triggered if ACBM was installed during a policy period or if releases or reentrainments took place during a policy period, and we affirm the trial court’s ruling that the policies provide coverage “in full,” without the participation of Armstrong, (up to the policy limits, and subject to apportionment based upon “other insurance” clauses). We also affirm the judgment concerning the duty to defend and the duty to indemnify. With respect to the judgment on the INA excess policy (paragraph 54), we remand for findings on the objectively reasonable expectations of the insured.

Disposition

The judgment on phase I denying recovery or relief to GAP Corporation under Continental Casualty Company policy RD 9972548 (paragraph 3) is affirmed.

The judgment on phase III concerning the trigger and scope of coverage for bodily injury claims is affirmed, except that the first sentence of paragraph 8 of the judgment is modified to read that all of a policyholder’s policies subject to this judgment that were in effect from the date of the claimant’s first exposure to the policyholder’s asbestos product until the date of death or claim, whichever occurs first, are triggered on an asbestos-related bodily injury claim, but the claimant is presumed to have been exposed to all defendant-manufacturers’ asbestos products, and the burden is on the insurer to prove that the claimant was not exposed to its policyholder’s product before or during the policy period.

The declaration on the meaning of “expected or intended” language (paragraphs 20 and 21) is reversed, but the judgment that coverage exists under the policy of Commercial Union’s predecessor issued to Armstrong for asbestos bodily injury claims (paragraph 32) is affirmed.

The judgment that the policies issued to GAP prior to May 26, 1967, by Continental Casualty and by predecessors of Commercial Union provide coverage for asbestos-related bodily injury claims related to the Ruberoid Co. (paragraph 30) is reversed.

The judgment on the effects of the Wellington Agreement (paragraphs 17 through 19) is affirmed.

The judgment on phase V concerning defense and indemnity obligations of the insurers on asbestos-related property damage claims against Armstrong, concerning policy exclusions, and concerning the trigger and scope of coverage is affirmed, except that the judgment on the INA excess policy (paragraph 54) is reversed and the matter is remanded for findings on the objectively reasonable expectations of the insured.

Newsom, J., * and Stein, Acting P. J., concurred.

The petitions of all appellants for review by the Supreme Court were denied August 21, 1996.

1 Phase I involved the existence and terms of missing insurance policies. Phase II concerned the application of exclusions for “asbestosis.” Phase III involved the trigger and scope of coverage for bodily injury claims, the meaning of the “neither expected nor intended” language contained in some of the policies, and the defense obligations of various insurers under their policies. Phase IV involved various coverage issues not resolved in phase III. Phase V concerned coverage for property damage claims, and phase VI involved issues of damages, bad faith, and contribution claims.

No issues have been raised on appeal concerning phase II. The appeals pertaining to phase VI have been dismissed.

* See footnote, ante, page 1.

8 The trial court found that for purposes of this trigger issue the language contained in the various policies at issue is functionally identical. The trial court also found the meaning of the policy language to be plain and unambiguous in requiring indemnification and defense when any one of three distinct conditions—bodily injury, sickness or disease—is present during the policy period.

9 This general rule has not been followed where the policy language was substantially distinguishable. (Insurance Co. of North America v. Sam Harris Constr. Co. (1978) 22 Cal.3d 409, 412 [149 Cal.Rptr. 292, 583 P.2d 1335] [“occurrence” not defined in policy; coverage held triggered by act of negligence committed within policy period]; Harbor Ins. Co. v. Central National Ins. Co. (1985) 165 Cal.App.3d 1029, 1035 [211 Cal.Rptr. 902] [policy covered offenses committed during policy period; held, no coverage when offense committed before policy took effect].)

10 “Throughout the briefs in this appeal—in Issue Group II no less than in other issue groups—certain parties have cited and relied upon unpublished opinions, in violation of rule 977 of the California Rules of Court. We emphasize that such citations are inappropriate, and we have paid them no heed. (Casella v. City of Morgan Hill (1991) 230 Cal.App.3d 43, 58 [280 Cal.Rptr. 876], cert. den. 503 U.S. 983 [118 L.Ed.2d 387, 112 S.Ct. 1665].)

11 Yet another approach—a theory of “double triggers”—was taken by the Illinois Supreme Court in Zurich Ins. Co. v. Raymark Industries (1987) 118 Ill.2d 23 [112 Ill.Dec. 684, 514 N.E.2d 150]. There, the trial court found that “injury” occurs when asbestos fibers are inhaled and retained in the lung, i.e., when the claimant is exposed to asbestos; “disease” occurs when an asbestos-related disease has progressed to the point that it significantly impairs the lungs’ function and is thereby capable of clinical detection and diagnosis, but between those two points, after exposure to asbestos ceases and before an asbestos-related disease becomes diagnosable, there is no continuous injury. The court based this conclusion on the expert testimony that asbestos-related disease may or may not progress during periods of nonexposure. (514 N.E.2d at pp. 160-161.) In addition, the court found that “sickness” may take place in the period before clinical manifestation of the disease; whether and when a claimant suffered sickness must be determined on a case-by-case basis. The Illinois Supreme Court upheld this approach as consistent with the plain and unambiguous language of the policies.

12 The trial court further concluded that once a claim is filed by a living claimant the claimant’s bodily injury is no longer an unknown event and, accordingly, under the loss-in-progress rule (Ins. Code, §§ 22, 250) policies beginning after the claim is filed are not triggered. This aspect of the trial court’s decision is erroneous, as the Supreme Court has now clarified that as long as the insured’s liability remains uncertain the loss-in-progress rule will not bar coverage. (Montrose Chemical Corp. v. Admiral Ins. Co., supra, 10 Cal.4th at pp. 689-693.) However, because the policyholders have not challenged the trial court’s trigger decision, the error is waived.

13 The trial court found as follows: “The Court determines that regardless of which terms the medical experts use to describe the physiological processes associated with the inhalation of asbestos, the medical evidence establishes that these processes impair the gas exchange function of the lung cells and tissue. The Court finds that these processes begin almost immediately upon the inhalation and deposition of asbestos fibers into the lung, and slowly and continuously impair new portions of lung tissue throughout one’s life, even after exposure to asbestos ceases. This continuing process could properly be termed ‘progression,’ with the understanding that this refers to progressive involvement of new cells and tissue and constitutes new injury to the cells, tissue and body.

“The issue concerning progressive involvement of new tissue during the latency period was one of the most vigorously contested at trial. Admittedly, it was not conclusively established that every person exposed to asbestos experiences such progression. However, the case before this Court concerns coverage for only those individuals who present asbestos-related claims against the manufacturers for which the manufacturers seek indemnity from their insurers. Even within this group of individuals, it is impossible to say with absolute certainty whether all or any particular percentage of these individuals continued to develop asbestos-related injury in each and every policy period following cessation of exposure and preceding manifestation.

“Such certainty and universality are not required, however, for the Court to make the foregoing determinations. More than sufficient evidence was presented to enable the Court to determine that among individuals who present claims for asbestos-related injury and/or disease, the injurious physiological processes associated with the inhalation of asbestos continue to occur from initial exposure, after cessation of exposure, and throughout those individuals’ lives.

“This determination is amply and convincingly supported by the record. The non-biodegradable nature of the fibers and their continued retention in the lungs elicit a sustained response by the body’s cellular defense mechanisms. Animal studies demonstrate such progression in more detail than is possible by medically ethical studies of humans and can be extrapolated to the human experience. Clinical diagnostic tools such as serial x-rays and lung function tests, though relatively crude in their ability to measure progression, have also confirmed such progression.”

14 In the trial court, the insurers took various positions on the trigger of coverage: some advocated the exposure theory; others the manifestation theory, and still others took no position. On appeal, the joint briefs of the insurers which remain in the action do not clearly articulate what events should trigger coverage, but their arguments imply an advocacy for the manifestation theory. We were advised at oral argument, however, that the insurers accept the injury-in-fact approach.

15 In phase IV, the trial court qualified its “in full” ruling by concluding that only one policy’s limits can apply to each claim, and the policyholder may select the policy under which it is to be indemnified. That decision, too, is supported by Keene: “The principle of indemnity implicit in the policies requires that successive policies cover single asbestos-related injuries. That principle, however, does not require that Keene be entitled to ‘stack’ applicable policies’ limits of liability. . . . Therefore, we hold that only one policy’s limits can apply to each injury. Keene may select the policy under which it is to be indemnified.” (Keene Corp. v. Ins. Co. of North America, supra, 661 F.2d at pp. 1049-1050; see also Owens-Illinois, Inc. v. Aetna Cas. and Sur. Co. (D.D.C. 1984) 597 F.Supp. 1515, 1524; contra, Cole v. Celotex Corp., supra, 599 So.2d 1058, 1074-1080; J.H. France Refractories v. Allstate, supra, 626 A.2d 502, 510.) The policyholders have not challenged this ruling.

16 Other approaches include apportionment (1) based on the premiums paid (Insurance Co. of Tex. v. Employers Liability Assur. Corp. (S.D.Cal. 1958) 163 F.Supp. 143, 147, 151); (2) in equal shares (Reliance Ins. Co. v. St. Paul Surplus Lines Ins. (4th Cir. 1985) 753 F.2d 1288, 1292); and (3) using a “maximum loss” method (Mission Ins. Co. v. Allendale Mut. Ins. Co. (1981) 95 Wn.2d 464 [626 P.2d 505]; and see Uniroyal, Inc. v. Home Ins. Co. (E.D.N.Y. 1988) 707 F.Supp. 1368, 1392-1393 [agent orange]). (See generally, Ostrager & Newman, Handbook on Insurance Coverage Disputes (4th ed. 1991) § 9.04, p. 338.)

17 Despite the phase III decision obligating the insurers to respond “in full,” the effect of the trial court’s apportionment formula is to make the liability of the insurers proportionate. In their briefs on phase III, the insurers make little mention of the phase IV decision and imply that the trial court did not prorate the losses or take into account the amount of time that insurance coverage was provided to the policyholder. Yet when the phase III and phase IV decisions are read together, the insurance carriers got what they want—pro rata allocation of liability.

As a practical matter, the point is academic for most insurers. The trial court noted that “all primary policies have been or will be exhausted by asbestos-related claims. The method of allocation affects only the timing of payments.” With respect to Fibreboard, however, two of its insurers (Pacific Indemnity and Continental) issued policies without aggregate limits. Thus, as to presently unpaid claims against Fibreboard, the only allocation will be between those two unexhausted policies. Contrary to the assertions of the insurers, this fact is not the result of any flaw in the trial court’s decision on the scope of coverage. Whether the policies provide unlimited coverage is before us on other issues which Pacific Indemnity and Continental have asked us to defer.

18 The insurers have relied upon the appellate court’s decision in J.H. France Refractories v. Allstate (1990) 396 Pa.Super. 185 [578 A.2d 468], which applied a continuous trigger and also allocated liability pro rata among insurers based upon the time each insurer was on the risk. But during the pendency of this appeal, the Pennsylvania Supreme Court reversed that decision and instead followed Keene to hold, as we do here, that each insurer must bear potential liability for the entire claim, subject to allocation based on the “other insurance” provisions of the policies. (J.H. France Refractories v. Allstate, supra, 626 A.2d 502.)

19 Despite the California Union court’s appellation of the liability as “joint and several,” the court went on to apportion the damages pro rata between the two insurers based on the policy limits. (145 Cal.App.3d at p. 478.)

20 The trial court expressly disavowed joint and several liability: “The Court recognizes that ‘joint and several liability’ is a doctrine of tort liability, not contract law. Furthermore, the Court is aware that liability for many of the underlying injury claims which are the subject of this coverage dispute is not joint and several among the policyholders, but rather is pursuant to settlement agreements which specify distinct damage amounts attributed to each policyholder. [1 This Court emphasizes that its determination of the scope of coverage for asbestos-related bodily injury claims is not predicated on the joint and several liability which is sometimes but not always imposed on the policyholders in the underlying lawsuits.”

21 In phase IV, where the issue was the method of allocating indemnity and defense costs among multiple insurers pursuant to “other insurance” clauses, no insurer argued that the “other insurance” clauses apply to periods when the policyholder is self-insured. Indeed, most courts hold that “other insurance” refers only to another policy, not to self-insurance. (E.g., Metro U.S. Services v. City of Los Angeles (1979) 96 Cal.App.3d 678, 683 [158 Cal.Rptr. 207]; Universal Underwrit. Ins. Co. v. Marriott Homes, Inc. (1970) 286 Ala. 231 [238 So.2d 730, 732]; American Nurses Ass’n v. Passaic Gen. Hosp. (1984) 98 NJ. 82 [484 A.2d 670, 674]; contra, Aetna Casualty & Surety Co. v. Market Insurance Co. (Fla.Dist.Ct.App. 1974) 296 So.2d 555, 558; see generally, Annot. (1986) 46 A.L.R.4th 707.) Thus, we construe the insurers’ argument as one directed to the insurers’ obligations to its insured and not as a challenge to the allocation formula governing the insurers’ obligations to other insurers.

In any event, the trial court employed a hybrid method of allocation, taking into account both the time on the risk and the policy limits. Because the policy limits are an essential part of the trial court’s apportionment formula, it would be virtually impossible to compute the policyholder’s share of the losses for periods of no insurance. The Keene court so reasoned: “We have no authority upon which to pretend that Keene also has a ‘ self-insurance’ policy that is triggered for periods in which no other policy was purchased. Even if we had the authority, what would we pretend that the policy provides? What would its limits be?” (667 F.2d at pp. 1048-1049; accord, J.H. France Refractories v. Allstate, supra, 626 A.2d 502, 508; contra, Owens-Illinois, Inc. v. United Ins. Co., supra, 650 A.2d 974, 995.)

22 In California, as in many states, statutory modifications have been made to the doctrine of joint and several liability. Liability among concurrent tortfeasors for noneconomic damages is now several—proportionate to fault, not joint. (Civ. Code, § 1431.2, subd. (a).) The purpose of this change was to eliminate the unfairness of requiring a tortfeasor who was minimally culpable to bear all of the plaintiff’s damages when the more culpable tortfeasor became insolvent. (Evangelatos v. Superior Court (1988) 44 Cal.3d 1188, 1198 [246 Cal.Rptr. 629, 753 P.2d 585].)

23 For purposes of our discussion, we assume that the claimant was exposed to a policyholder’s product at some point. In most jurisdictions, an asbestos manufacturer will not be held liable unless the plaintiff proves, directly or circumstantially, actual exposure to the defendant’s product. (See Lineaweaver v. Plant Insulation Co. (1995) 31 Cal.App.4th 1409, 1415-1419 [37 Cal.Rptr.2d 902]; Mullen v. Armstrong World Industries, Inc. (1988) 200 Cal.App.3d 250, 257-258 [246 Cal.Rptr. 32]; In re Hawaii Federal Asbestos Cases (9th Cir. 1992) 960 F.2d 806, 816-818; Bauer v. Raymark Industries, Inc. (2d Cir. 1988) 849 F.2d 790, 792-793; Roehling v. Nat. Gypsum Co. Gold Bond Bldg. (4th Cir. 1986) 786 F.2d 1225, 1228, fn. 5; Blackston v. Shook & Fletcher Insulation Co. (11th Cir. 1985) 764 F.2d 1480, 1482, 1485; Gideon v. Johns-Manville Sales Corp. (5th Cir. 1985) 761 F.2d 1129, 1144-1145.) In other jurisdictions, under the market share doctrine, the plaintiff need not identify the manufacturer of the asbestos product; the burden of proof is shifted to the defendant-manufacturer to prove its product did not cause the plaintiff’s injury. (Wheeler v. RaybestosManhattan, supra, 8 Cal.App.4th 1152; Menne v. Celotex Corp. (10th Cir. 1988) 861 F.2d 1453, 1469; see Sindell v. Abbott Laboratories, supra, 26 Cal.3d 588.) Nevertheless, if the plaintiff was not exposed to the manufacturer’s product, the manufacturer will not be liable.

We note, however, that the New York Court of Appeals has modified tort law principles of causation and held that a manufacturer may be held severally liable in proportion to its market share even if it can prove that its product did not contribute to the plaintiff’s injury; only those defendants who can prove that they never participated in the marketing of the product for use by the class of injured victims are exculpated. (Hymowitz v. Eli Lilly and Co. (1989) 73 N.Y. 487 [541 N.Y.S.2d 941, 950, 539 N.Ed.2d 1069], cert. den. (1989) 493 U.S. 944 [107 L.Ed.2d 338, 110 S.Ct. 350] [DES case].)

Of course, new forms of liability created by statute or by judicial pronouncement come within the scope of liability covered by a CGL policy. (AIU Ins. Co. v. Superior Court (1990) 51 Cal.3d 807, 822, fn. 8 [274 Cal.Rptr. 820, 799 P.2d 1253]; Travelers Ins. Co. v. Industrial Indem. Co. (1971) 18 Cal.App.3d 628, 632 [96 Cal.Rptr. 191].) The question before us, however, is not whether the manufacturers’ liability is covered but when coverage begins.

24 Language in Montrose supports our conclusion. In describing the continuous trigger, the Montrose court observed that the timing of the event or conditions causing the injury or damage is largely immaterial: “it can occur before or during the policy period.” (10 Cal.4th at p. 675, italics added.) We construe this language to mean that exposure to hazardous conditions after the policy period will not trigger coverage.

25 Throughout our discussion we have referred to the claimant’s “contact with” or “exposure to” the policyholder’s products, for in this case the policyholders’ wrongful conduct, upon which their underlying liability is based, consists of exposing the asbestos victims to hazardous products. (Fn. 23, ante.) The parties have neither raised nor briefed the question of trigger of coverage when the underlying liability is established in a jurisdiction such as New York, where a manufacturer may be held liable in proportion to its market share despite proof that the claimant was not exposed to its products. We express no opinion on that question. We hold only that coverage is not triggered if the policyholder’s liability-producing conduct took place after the policy expired.

26 Because most claims filed against Armstrong for asbestos-related injuries involve insulation workers, the focus of the trial was on Armstrong’s insulation products. In addition, Armstrong also manufactured floor tiles, which until 1983 contained asbestos. Again, the asbestos was not friable; it was bonded into the tiles. And until 1984, Armstrong manufactured an asbestos-containing gasket used for automotive purposes. The phase IV record does not indicate whether any bodily injury claims have arisen from these products. (Note that the floor tiles are involved in the phase V dispute over property damage.)

27 In phase III, the trial court determined that because the effect of the “neither expected nor intended” clause was to limit coverage, the burden was on the insurer to prove the injury was expected or intended. In phase IV, the trial court explained that it would have found that Armstrong did not act for the purpose of causing injury even if the burden of proof had been upon the insured.

28 “Expect” is given the following dictionary definitions: “to look forward: look with anticipation . . . : to look forward to; specif.: to anticipate the occurrence of . . . : to consider probable or certain . . . .” (Webster’s New Intemat. Diet. (3d ed. 1965) p. 799); “to look forward to (an event), regard (it) as about to happen; to anticipate the occurrence of (something whether good or evil)” (5 Oxford English Diet. (2d ed. 1989) p. 556); “1. To look forward to the probable occurrence or appearance of. 2. To consider likely or certain” (American Heritage Diet. (2d college ed. 1982) p. 476).

29 Insurance Code section 533 provides: “An insurer is not liable for a loss caused by the wilful act of the insured; but [the insurer] is not exonerated by the negligence of the insured, or of the insured’s agents or others.”

30 The laws of Delaware and New Jersey, the respective states of GAF and Ruberoid, are in accord with the general rule that upon a merger a surviving corporation is answerable for the debts and liabilities of the acquired corporation. (N.J. Rev. Stat. § 14A:10-6(e); 8 Del. Code § 259(a).)

31 The insurers rely upon Aetna Life & Cas. v. United Pac. Rel. Ins. (Utah 1978) 580 P.2d 230, but that case holds that insurance coverage survives a corporation’s merger and passes to the surviving corporation along with the liabilities. (See also Oklahoma Morris Plan Co. v. Security Mutual Cas. Co. (8th Cir. 1972) 455 F.2d 1209; Maryland Cas. Co. v. W.R. Grace & Co., supra, 794 F.Supp. at pp. 1233-1236.) Application of that principle here means that upon GAF’s succession to Ruberoid’s liabilities, GAF became entitled to insurance coverage by Ruberoid’s insurers. This principle is of no relevance to the issue before us. As a general rule, insurance policies should be interpreted as if no other insurance is available. (Pacific Indemnity Co. v. Imperial Casualty & Indemnity Co. (1976) 176 Cal.App.3d 622, 627 [222 Cal.Rptr. 115]; Chamberlin v. Smith, supra, 72 Cal.App.3d at p. 844.) Therefore, in construing GAF’s premerger policies we do not consider the availability of coverage under Ruberoid’s policies.

32 The ELAC policy provides in pertinent part: “It is agreed that the Named Insured shall read as follows: [GAP] and any other business organization while the foregoing named insured owns an interest therein of more than fifty percent (50%) during the policy period.”

33 Similarly, in light of the phrase “as now existing,” coverage would extend to a subsidiary severed after the policy period began.

34 Although we rely upon a plain reading of the language of the policy, we reach the same result as was reached in Cooper Companies v. Transcontinental Ins. Co. (1995) 31 Cal.App.4th 1094 [37 Cal.Rptr.2d 508], in which the court found the language ambiguous and relied upon the reasonable expectations of the insured. We find the policy language, when construed in the context of the policy as a whole, capable of only one plausible construction; hence, we discern no ambiguity. (See generally Bay Cities Paving Grading, Inc. v. Lawyers’ Mutual Ins. Co. (1993) 5 Cal.4th 854, 867 [21 Cal.Rptr.2d 691, 855 P.2d 1263], defining ambiguity.)

35 Continental and Commercial Union contend an acquired company qualifies as a named insured only if it was already acquired at the time of the occurrence—at the time insurance coverage was triggered. This contention suggests that insurance coverage would be excluded if an occurrence arising from the acquired company’s conduct took place before the merger (and during the policy period), even though the tortfeasor-company was acquired during the policy period. We need not decide this point, i.e., whether the phrase “or hereafter constituted” would qualify the acquired company as a named insured during the entire policy period and extend insurance coverage to occurrences that took place before the merger as well as occurrences that took place afterward. We decide only that there is no coverage when the merger took place after the policy period had expired.

36 Although the decision in Oliver Machinery Co. v. United States Fid. & Guar. Co. (1986) 187 Cal.App.3d 1510 [232 Cal.Rptr. 691] (product of predecessor company was not one of named insured’s products) supports the decision here, we do not rely upon it, as it involved very different facts and issues. The question before the Oliver court was whether the named insured’s distributor qualified as an additional insured. The meaning of “hereafter constituted” was not in issue.

Two other decisions reached the same conclusion reached here, that the premerger insurer of the acquiring company does not provide coverage for liabilities of the acquired company: State of Idaho v. Bunker Hill Co. (D.Idaho 1986) 647 F.Supp. 1064, 1077, and Maryland Cas. Co. v. W.R. Grace & Co., supra, 794 F.Supp. 1206, 1230-1232. However, the courts in those cases employed a somewhat different analysis, reasoning that during the premerger policy period the insured was not responsible for the liabilities of the later-acquired company. Although we focus instead on the fact that Ruberoid was not a named insured under the premerger policies, those cases support our view that the named insured must qualify as such during the policy period.

37 Language in Isaacson suggests that the “reasonableness” of a settlement includes a showing that the underlying claim was covered, although the insured “need not prove his actual liability on the underlying claim, and establishing a breach [of the insurer’s duty to settle for a reasonable amount] does not require a trial of the underlying action.” (44 Cal.3d at p. 793.) Thus, the trial court’s finding that the Wellington Agreement was a reasonable settlement disposes of the objection of the CCC Companies that the Wellington payments go to claims for which the manufacturers are not legally liable. The trial court found the Wellington Agreement to be a reasonable means to resolve the claims among the producers: “Although a producer pays on all claims, whether or not the producer is named in a particular claim, it is equally true that other producers pay on claims in which they are not named and thus pay a proportionate share of the producer’s liability.” The CCC Companies have not challenged this finding on appeal; their appellate arguments pertain to the application of the presumption of liability.

38 While this appeal was pending, we received notices of injunctions and stays issued in connection with receivership and liquidation proceedings involving the following subscribers to certain Armstrong policies: Kingscroft Insurance Company Ltd., El Paso Insurance Company Ltd., Lime Street Insurance Company Ltd., Mutual Reinsurance Company Ltd., and Walbrook Insurance Company Ltd. Nothing in this opinion should be construed as being inconsistent with the orders staying proceedings against those parties.

39 We discuss separately, in part C below, the early policies covering property damage “caused by accident.”

40 The insurers on this brief are United States Fire Insurance Co., Central National Insurance Co. of Omaha, Puritan Insurance Co., and Interstate Fire & Casualty Company.

41 As a practical matter, the insurers’ argument seems academic. We find it difficult to envision a scenario in which a single released fiber would be detected, much less serve as the sole basis for triggering coverage.

42 The trial court concluded that “the duty to indemnify is triggered when Armstrong proves an occurrence during the policy period to [j/c] which it is held liable.” We infer from this ruling that if Armstrong is held liable not for contamination of a building from released asbestos fibers but for the mere presence of ACBM (and the potential for releases), coverage would be triggered only at installation. This is so because any incidental releases that may have occurred during subsequent policy periods would not constitute damage for which Armstrong was held liable. We agree. Releases or reentrainments of asbestos fibers during a policy period will trigger coverage only if the basis of Armstrong’s liability is contamination from released asbestos fibers.

43 The courts in the cited cases found the damage both immediate and continuous and employed a continuous trigger. As we have discussed in part B.l above, the continuous trigger was not applied here, as the trial court found asbestos property damage is not continuous. But the fact that property damage is not continuous does not preclude the damage from being immediate upon exposure to the hazardous material.

44 There is no dispute that in order for the policy to provide coverage it is the act or event that must be accidental, not the consequences. (Shell Oil Co. v. Winterthur Swiss Ins. Co., supra, 12 Cal.App.4th 715, 749-750; Commercial Union Ins. Co. v. Superior Court (1987) 196 Cal.App.3d 1205, 1208 [242 Cal.Rptr. 454].)

45 Reliance also argues that installation of ACBM during a policy period cannot trigger coverage because installation poses only the potential for release of asbestos fibers. As we have explained in part B.3 above, however, we conclude that actual injury was inflicted at the moment of installation.

Neither Reliance nor any other party has argued that the installation of ACBM fails to meet the suddenness requirement. Nor has any party argued that the installation of ACBM fails to qualify as an unexpected or unintended event so as to meet the other defining element of “accident.” Accordingly, we do not decide whether installation of ACBM constitutes an “accident” or an “occurrence” under policies defining “occurrence” as an “accident, including continuous or repeated exposure to conditions.”

46 In contrast to the insurers’ position on the bodily injury claims (Issue Group II, part A, ante), the insurers have not raised the argument with respect to the property damage claims that liability should be allocated proportionately among the insurers based upon the time each insurer was on the risk. Their argument is confined to the assertion that Armstrong must contribute for periods during which it had no insurance.

47 The trial court was not asked to decide the number of occurrences in the building cases, and we express no opinion on that issue.

48 Obviously this question does not arise in connection with the asbestos-related bodily injury claims, as there is no dispute that the claimants have suffered bodily injuries; the key question is when the injuries occurred for purposes of triggering coverage. (Issue Group II, part A, ante.)

49 We express no opinion on whether declaratory relief would be appropriate if the insured objected to a determination of coverage questions in advance of a determination of liability.

50 In an analogous context the same rule applies: a direct action by a third party claimant against an insurer must await a final determination of the insured’s underlying liability. (McKee v. National Union Fire Ins. Co. (1993) 15 Cal.App.4th 282 [19 Cal.Rptr.2d 286]; Nationwide Ins. Co. v. Superior Court (1982) 128 Cal.App.3d 711 [180 Cal.Rptr. 464]; Laguna Pub. Co. v. Employers Reinsurance Corp. (C.D.Cal. 1985) 617 F.Supp. 271.)

51 In its phase V decision on property damage, the trial court referred back to its decision in phase III (on bodily injury claims) in which the trial court concluded that although as a legal matter the duty to defend is broader than the duty to indemnify, when applied here the duty to defend coincides with the duty to indemnify: “the trigger and scope of defense follows [sic] the trigger and scope of indemnity.” That is, with respect to bodily injury claims, “since any policy triggered by a claim has an independent obligation to indemnify, each triggered policy which contains a defense obligation to the insured has an independent obligation to defend or reimburse defense costs, depending on the policy language.”

52 The duty to defend is not dependent solely on the allegations in the underlying complaint; the duty to defend may be triggered if the insurer learns of facts from any source that create the potential for liability. (Gray v. Zurich Insurance Co., supra, 65 Cal.2d at pp. 276-277; CNA Casualty of California v. Seaboard Surety Co. (1986) 176 Cal.App.3d 598, 606 [222 Cal.Rptr. 276].) Likewise, the insurer may produce undisputed extrinsic evidence that eliminates the possibility of coverage. (See Montrose Chemical Corp. v. Superior Court (1993) 6 Cal.4th 287, 304 [24 Cal.Rptr.2d 467, 861 P.2d 1153].)

*

* Retired Associate Justice of the Court of Appeal, First District, sitting under assignment by the Chairperson of the Judicial Council.