FMC Corp. v. Plaisted & Companies (61 Cal. App. 4th 1132, 1998)

Citations

61 Cal. App. 4th 1132

Full opinion text

Opinion

BAMATTRE-MANOUKIAN, J.

— These eight appeals arise from a single superior court action, and address specific aspects of the duty of a general liability insurer to indemnify its insured for the cost of reimbursing government agencies and of complying with orders for investigation and remediation of toxic contamination to soil, surface water, and groundwater beneath the surface of the soil.

The general liability insurance policies, issued by London insurers, are of the type known in American insurance practice as comprehensive or commercial general liability (or CGL) policies. The policies contain indemnification provisions similar to those of standard form CGL policies but omit the explicit duty-to-defend provisions commonly included in CGL policies. The policies’ indemnification provisions require, in pertinent part and subject to many qualifications, that the insurer indemnify the insured for all sums the insured shall be obligated to pay by reason of liability for property damage as defined. In a writ proceeding arising out of this action the Supreme Court has established as general propositions, and insofar as applicable to these policies, that contamination of the environment is property damage and, in essence, that amounts the insured is required to pay to reimburse government agencies and to comply with government orders under statutes such as the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (CERCLA) (42 U.S.C. § 9601 et seq.) and similar statutes, once hazardous wastes have been released, are sums the insured is obligated to pay by reason of liability for property damage. (AIU Ins. Co. v. Superior Court (1990) 51 Cal.3d 807, 824-843 [274 Cal.Rptr. 820, 799 P.2d 1253].) These appeals call upon us to resolve additional coverage issues, as well as procedural and evidentiary questions which arose in the course of trial.

FMC Corporation, a diversified manufacturer of equipment and chemical compounds, acknowledges that over a period of many years its commercial activities, and the activities of others at sites for which FMC was or became responsible under environmental laws, caused toxic contamination to soil and groundwater at and near many sites throughout the country. Beginning as early as 1959, FMC spent substantial sums to investigate the nature and extent of the contamination, to undertake measures of its own to remediate the contamination, to reimburse government agencies for their remediation work, and to comply with related orders of state and federal agencies.

Throughout the relevant period FMC had purchased and maintained extensive general liability insurance. Liberty Mutual Insurance Company provided FMC’s primary coverage to limits of $10 million. FMC also bought umbrella and excess coverage provided by a number of insurers including certain underwriters at Lloyds of London and certain London insurance companies. We shall refer to these London underwriters and insurance companies, including one company that engaged separate counsel in the course of the proceedings, as “the London insurers” or “the London defendants,” and to the umbrella and excess policies they issued as “the London policies.” Many other insurers were involved in the proceedings below, but of all of FMC’s insurers only the London defendants are now before this court. Exhaustion of FMC’s primary coverage (cf. Community Redevelopment Agency v. Aetna Casualty & Surety Co. (1996) 50 Cal.App.4th 329, 337-340 [57 Cal.Rptr.2d 755]) is not an issue on these appeals.

Early in November 1987, FMC for the first time formally demanded that the London defendants investigate, defend and indemnify FMC with respect to toxic contamination at specified sites. Eleven days later FMC brought this action, against Liberty Mutual, the London defendants and many other insurers, for a declaration of rights as against all the insurers and for damages as against Liberty Mutual. Over a period of nearly 10 years, many of the sites for which, and many of the insurers against which, FMC had claimed coverage were removed from the action by pretrial order, voluntary dismissal, or stipulated disposition. In 1988, a few months before he was appointed to this court, the action was assigned to the Honorable Eugene M. Premo for all purposes, and he continued as single-assignment judge in the action until the spring of 1993, shortly before trial began. Justice Premo has taken no part in the consideration or decision of the issues on this appeal.

Beginning in 1993, a total of fifty-seven sites were assigned for trial in eight groups, with eight separate juries empaneled to hear factual issues and render special verdicts at a series of eight separate trial phases to which we shall refer (as the court and counsel did in the trial court) as “trials.” Of the 57 sites 4 did not involve the London defendants, and of the remaining 53 more than half were removed from jury consideration by dismissal, stipulation, or court order in the course of pretrial and trial proceedings. Thus, for example, of the six sites assigned to the first trial, two were dismissed by FMC before trial, one was dismissed by stipulation among the parties, one did not involve the London defendants, and only the remaining two went to special verdicts and judgment with respect to the London defendants. And although all of the six sites assigned to the second trial went to judgment as to the London defendants, as to one of the sites the judgment was based on a summary adjudication order and as to another it was based on the parties’ stipulation. By the time the last trial Was concluded, in early 1997, a total of 35 sites had gone to judgment with respect to the London defendants, 22 of them on the basis of special verdicts. By this time only the London insurers remained as defendants.

Following each of the eight separate trials the trial court entered a separate, detailed recitation of the proceedings had, findings made, and conclusions reached at that trial. Each of these eight recitations was denominated a “judgment.” Neither FMC nor the London defendants were fully satisfied with any of these eight “judgments”: FMC appealed, and the London defendants cross-appealed, from each of them. This court concluded that the interests of justice would best be served by considering the eight appeals together.

Three issues lend themselves to summary disposition:

(1) Appealability. The London defendants initially took the position that none of the individual “judgments,” viewed in isolation, could be deemed a single final judgment disposing of all issues or be brought within an exception to the one final judgment rule, and therefore that none of the individual “judgments” was separately appealable. (Cf. Morehart v. County of Santa Barbara (1994) 7 Cal.4th 725, 736-744 [29 Cal.Rptr.2d 804, 872 P.2d 143].) This court’s decision to consider the appeals from the eight separate “judgments” together effectively moots the appealability issue: We shall treat the eight “judgments” heretofore entered collectively, as a single judgment disposing of all issues, and the eight appeals as a single collective appeal from that single judgment. For convenience we shall continue to use the terms “judgment” and “judgments,” in quotation marks, to refer to the trial court’s eight separate recitations.

(2) Instructions on late-notice defense. As to each of the two sites that reached the jury at the first trial — the Fresno site in California and the Conservation Chemical site in Kansas — the London defendants asserted as a defense that FMC had not given them the notice required by the London policies within a reasonable time. The trial court instructed the jury that, to sustain the defense, the London defendants would be required to prove they had been prejudiced by the delay. In this court FMC contends that in preinstructions the trial court improperly authorized the jury to consider prejudice to the London defendants in defending FMC’s coverage action, arguing that the law permits a late-notice defense only where the insurer can show it was prejudiced in asserting a defense to the underlying claim against the insured. But at the first trial the jury explicitly found that FMC had not been late in giving notice, and the point has not been raised as to any subsequent trial. Because FMC was not aggrieved by the preinstruction of which it complains, we need not reach the issue in order to dispose of FMC’s appeal. We respectfully decline FMC’s request that we nevertheless discuss the issue, essentially in the abstract, against the possibility it will recur in future proceedings.

(3) Standing to raise other issues. In their respondents’ brief on FMC’s first appeal the London defendants asserted that FMC lacked standing, by reason of jury findings favorable to FMC, to raise certain issues in addition to the late-notice preinstruction. It appears that each of these additional issues was subsequently raised, in one or more of the ensuing appeals, in circumstances in which FMC was in the requisite sense aggrieved and therefore had standing to raise the issues. Thus the additional issues are properly before us.

Coverage Issues

The Pollution Exclusions

The London defendants first provided umbrella and excess general liability coverage to FMC as of August 5, 1964, and continued to provide various coverages until 1985, but effective October 1, 1970, the London defendants incorporated into their FMC policies provisions that excluded coverage for liabilities for various kinds of releases of pollutants. For the first few years after 1970 the pollution exclusions were subject to an exception if (as worded in the London policies) “such . . . release ... is sudden and accidental” or if (in another London form) “such . . . pollution . . . is caused by a sudden, unintended and unexpected happening . . . .” Most of FMC’s claims were based on gradual releases of contaminants. Unless FMC could nevertheless invoke the exception, its coverage for environmental contamination liabilities under the London policies would be limited, for purposes relevant to these appeals, to the period from August 5, 1964, to October 1, 1970. To identify the coverage periods with which we must be concerned, we turn first to the construction to be given the exception to the London policies’ pollution exclusions.

To invoke the exception with respect to gradual releases, FMC needed to persuade the London defendants, or the trial court, that the word “sudden,” as used in each form of the exception, did not necessarily mean “abrupt” or otherwise evoke a temporal connotation. To this end FMC argued that the word was defined, was understood, and had been determined by several courts, to mean “unexpected” as an alternative to “abrupt.”

After consideration of FMC’s arguments and extrinsic evidence, the trial court concluded by pretrial ruling that “sudden” meant “abrupt — in a temporal sense — and was not a synonym for unexpected and unintended.” This pretrial ruling was adhered to throughout the trial proceedings.

The trial court’s conclusion was based in part on Shell Oil Co. v. Winterthur Swiss Ins. Co. (1993) 12 Cal.App.4th 715, 752-756 [15 Cal.Rptr.2d 815], and is also supported by ACL Technologies, Inc. v. Northbrook Property & Casualty Ins. Co. (1993) 17 Cal.App.4th 1773, 1779-1794 [22 Cal.Rptr.2d 206]. Although the issue has produced division among courts nationwide, no reported California appellate decision has contradicted the holdings of Shell and ACL that in the relevant context the word “sudden” means, at least, “abrupt,” and thus does not apply to gradual releases of pollutants no matter how unexpected.

We agree with Shell and ACL, and will not expand the literature by attempting to add to their careful analyses of the issue. FMC argues that neither Shell nor ACL “analyzed the drafting and regulatory history that is so critical to our position,” and that in this respect the Shell and ACL courts disregarded the rule of Pacific Gas & E. Co. v. G. W. Thomas Drayage etc. Co. (1968) 69 Cal.2d 33 [69 Cal.Rptr. 561, 442 P.2d 641, 40 A.L.R.3d 1373] (PG&E) that “[t]he test of admissibility of extrinsic evidence to explain the meaning of a written instrument is not whether it appears to the court to be plain and unambiguous on its face, but whether the offered evidence is relevant to prove a meaning to which the language of the instrument is reasonably susceptible.” (69 Cal.2d at p. 37.) ACL considered and rejected similar contentions. Again, we agree with, and will not seek to add to, ACL’s analysis.

From our conclusion that “sudden” means, at least, “abrupt,” it follows that FMC’s coverage under the London policies for any discharge or release of pollutants that could not be shown to be abrupt would be limited to policy periods between August 5, 1964 (when the London companies first provided coverage) and October 1, 1970 (the effective date of the pollution exclusions). For the purpose of the remaining coverage issues we shall focus upon that period.

The 1964-1970 London Policies

Characterization of an insurance policy as a “general liability” policy connotes neither unlimited coverage nor particular coverages. The term is more realistically viewed as a means of distinguishing policies which afford relatively broad coverages from those (such as directors and officers policies) which are explicitly limited to specifically defined areas of liability. (Cf. Fresno Economy Import Used Cars, Inc. v. United States Fid. & Guar. Co. (1977) 76 Cal.App.3d 272, 279-280 [142 Cal.Rptr. 681]; cf. also 11 Couch on Insurance (2d ed. 1982) Risks and Coverages, § 44:264, pp. 410-411.) It is invariably necessary to consult the language of any particular general liability policy to determine what coverages it affords.

Between August 5, 1964, and October 1, 1970, the London companies provided a total of eight general liability policies to FMC, of which two were of particular significance to the coverage issues before us. Policy No. K 79446 was in effect, with periodic modifications, from August 5, 1964, to October 1, 1969, and was in practical effect replaced by policy No. CX2062 for the period from October 1, 1969, to October 1, 1970. These two policies, each captioned as an “umbrella policy,” undertook to provide coverage immediately above and in some respects supplemental to Liberty Mutual’s primary general liability coverage. The remaining six policies provided layers of excess coverage above policy Nos. K 79446 and CX2062, and in relevant respects “followed form” with — which in insurance parlance meant their terms conformed to those of (cf., e.g., Wells Fargo Bank v. California Ins. Guarantee Assn. (1995) 38 Cal.App.4th 936, 940 [45 Cal.Rptr.2d 537]; Coca Cola Bottling Co. v. Columbia Casualty Ins. Co. (1992) 11 Cal.App.4th 1176, 1182 [14 Cal.Rptr.2d 643]) — the umbrella policies.

The pertinent basic provisions of policy Nos. K 79446 and CX2062 were closely similar.

The policies provided that the London insurers

K 79446

“agree ... to indemnify the Assured for all sums which the Assured shall be obligated to pay by reason of the liability [imposed by law or assumed by agreement] for damages, direct or consequential and expenses, all as more fully defined by the term ‘ultimate net loss’ on account of:— . . . Property Damage . . . caused by or arising out of each occurrence happening anywhere in the world.”

CX2062

These provisions were made subject to the following “Limit of Liability”:

K 79446

“Underwriters hereon shall only be liable for the ultimate net loss the excess of either

“(a) the limits of the underlying insurances as set out in the attached schedule in respect of each occurrence covered by said underlying insurances,

“or (b) $25,000 ultimate net loss in respect of each occurrence not covered by said underlying insurances,

“(hereinafter called the ‘underlying limits’):

“and then only up to [specified policy limits].”

CX2062

“Underwriters hereon shall only be liable for the ultimate net loss the excess of the greater of “(a) the limits of the underlying insurances as set out in the attached schedule and the amount collectible by the Assured under any other underlying insurance, in respect of each occurrence covered by said underlying insurances,

“o[r] (b) $100,000 ultimate net loss in respect of each occurrence

“(hereinafter called the ‘underlying limits’):

“and then only up to [specified policy limits].”

The policies defined “Property Damage” as

K 79446

“loss of or direct damage to or destruction of tangible property (other than property owned by the Named Assured).”

CX2062

“loss of or direct damage to or destruction of tangible property (other than property owned by the Named Assured).”

Throughout these proceedings the trial court and counsel have referred to property other than that owned by FMC as “third party property,” and to damage to such property as “third party property damage”; occasionally they have referred to FMC’s property as “owned property” and to damage to FMC’s property as “first party property damage.” We shall adopt this terminology. Under California law groundwater — which we broadly define as water beneath the surface of the soil — is publicly owned, and thus California groundwater would be third party property from FMC’s perspective. (AIU Ins. Co. v. Superior Court, supra, 51 Cal.3d at p. 817 fn. 6; cf. Wat. Code, § 102.)

The policies defined “OCCURRENCE” as

K 79446

“an accident or a happening or event or a continuous or repeated exposure to conditions which unexpectedly and unintentionally results in personal injury, property damage or advertising liability during the policy period. All such exposure to substantially the same general conditions existing at or emanating from one premises location shall be deemed one occurrence.”

CX2062

The policies defined “Ultimate Net Loss” as

K 79446

“the total sum which the Assured, or any company as his insurer, or both, become obligated to pay by reason of personal injury, property damage or advertising liability claims, either through adjudication or compromise, and shall also include hospital, medical and funeral charges and all sums paid as salaries, wages, compensation, fees, charges and law costs, premiums on attachment or appeal bonds, interest, expenses for doctors, lawyers, nurses and investigators and other persons, and for litigation, settlement, adjustment and investigation of claims and suits which are paid as a consequence of any occurrence covered hereunder, excluding only the salaries of the Assured’s or of any underlying insurer’s permanent employees.”

CX2062

“the total sum which the Assured, or any company as his insurer, or both, become obligated to pay by reason of personal injury, property damage or advertising liability claims, either through adjudication or compromise, and shall also include hospital, medical and funeral charges and all sums paid as salaries, wages, compensation, fees, charges and law costs, premiums on attachment or appeal bonds, interest, expenses for doctors, lawyers, nurses and ' investigators and other persons, and for litigation, settlement, adjustment and investigation of claims and suits which are paid as a consequence of any occurrence covered hereunder, excluding only the salaries of the Assured’s or of any underlying insurer’s permanent employees.”

Insured Not Involved Until After Policy Period

By definition, coverage under an “occurrence” liability insurance policy such as those before us requires that an event or events relevant to coverage have occurred during the policy period. (Cf. A. C. Label Co. v. Transamerica Ins. Co. (1996) 48 Cal.App.4th 1188, 1192 [56 Cal.Rptr.2d 207].) This concept is encapsulated in the term “trigger of coverage” which the Supreme Court, in Montrose Chemical Corp. v. Admiral Ins. Co. (1995) 10 Cal.4th 645 [42 Cal.Rptr.2d 324, 913 P.2d 878] (Montrose II), defined as “a term of convenience used to describe that which, under the specific terms of [a liability] insurance policy, must happen in the policy period in order for the potential of coverage [relevant to the duty-to-defend issues involved in Montrose 77] to arise. The issue is largely one of timing — what must take place within the policy’s effective dates for the potential of coverage to be ‘triggered’? Whether coverage is ultimately established in any given case may depend on the consideration of many additional factors . . . .” (10 Cal.4th at p. 655, fn. 2; cf. id. at p. 673 et seq.; Armstrong World Industries, Inc. v. Aetna Casualty & Surety Co. (1996) 45 Cal.App.4th 1, 39 [52 Cal.Rptr.2d 690].) But “unless coverage has been triggered under these occurrence policies within the policy period, there is no coverage once the policy period has ended. [Citation.]” (Cooper Companies v. Transcontinental Ins. Co. (1995) 31 Cal.App.4th 1094, 1107 [37 Cal.Rptr.2d 508]; A. C. Label Co. v. Transamerica Ins. Co., supra, 48 Cal.App.4th at p. 1194.)

Many cases have generalized that under an occurrence-based general liability policy it is the injury or harm that triggers coverage. (Cf. Armstrong World Industries, Inc. v. Aetna Casualty & Surety Co., supra, 45 Cal.App.4th at pp. 39-40, and cases and secondary authorities cited.) In Montrose II the Supreme Court identified four varying “trigger theories” applied by courts under such policies (10 Cal.4th at pp. 674-676) and held that, in circumstances analogous to those of this case, liability under such policies for property damage which is “continuous or progressively deteriorating throughout successive policy periods” will be covered by “all policies in effect during those periods,” under a theory which Montrose II labeled “[t]he continuous injury (or multiple) trigger.” (id. at p. 675.)

It appears to be undisputed for purposes of this appeal that the gradual release of pollutants which underlay most of FMC’s coverage claims invoked Montrose 77’s continuous injury trigger, and that coverage under the London policies was triggered by any property damage in any policy period at a site with which FMC had had some involvement relevant to liability under CERCLA or comparable statutes before or during that policy period.

But the parties sharply disagree as to whether coverage was triggered under the London policies for property damage for which (under CERCLA or comparable state statutes) FMC was liable even though FMC had had no involvement with the site until after the last relevant policy period ended on October 1, 1970.

A peculiarity of CERCLA and comparable statutes is that an entity may become liable (sometimes simply as the current owner of a site) for cleanup and remediation costs for property damage which occurred before, and sometimes long before, the entity had any involvement whatsoever. (Cf. CERCLA § 107(a), 42 U.S.C. § 9607(a); U.S. v. Shell Oil Co. (C.D.Cal. 1993) 841 F.Supp. 962, 968.) FMC found itself in this situation with respect to sites of two general kinds, labeled in motion paperwork as “category A” and “category B.” “Category A” included sites at which FMC had no involvement, either directly or vicariously through a predecessor entity, until after October 1, 1970; “category B” included sites in which FMC had not been involved, but a predecessor entity (subsequently acquired by FMC) had been involved, before October 1, 1970.

No “category A” or “category B” sites were involved in the first trial. Shortly after the first trial the London defendants sought a ruling that they would not be required to provide coverage for injury for which FMC was liable under CERCLA at sites where FMC’s involvement began after October 1, 1970. FMC responded that under the literal language of the policies’ “occurrence” definition it would be entitled to indemnity for its liability for any property damage within a covered policy period, whether or not FMC had caused the property damage or had otherwise been involved during the policy period. The trial court agreed with FMC, concluding that “there is no coverage requirement that the damage be caused by the named insured” and denying the London defendants’ motion. One “category A” site, Andover, in Minnesota, was assigned to the second trial, and the London defendants were found to have coverage, and to be obliged to indemnify FMC for its essentially retroactive CERCLA liability, for certain of their policy periods.

A “category B” site, Bermet, in South Carolina, was assigned to the fourth trial. The London defendants again moved for a ruling that they had no coverage, and this time the trial court, apparently influenced by the intervening appellate opinion in Cooper Companies v. Transcontinental Ins. Co., supra, 31 Cal.App.4th 1094, granted the London defendants’ motion, holding that there was no coverage for the Bermet site.

Three “category A” sites — Ekotek, in Utah, Delaware Sand & Gravel, in Delaware, and Global Landfill, in New Jersey — were assigned to the seventh trial, and again the London defendants made a pretrial motion for a ruling that there was no coverage, citing Cooper Companies and Armstrong World Industries, Inc. v. Aetna Casualty & Surety Co., supra, 45 Cal.App.4th 1. In Armstrong World Industries, an asbestos products liability case, the Court of Appeal among other things construed language from Montrose II “to mean that exposure to hazardous conditions after the policy period will not trigger coverage” (45 Cal.App.4th at p. 61, fn. 24), and concluded that “[a]n insurer has no liability if its policy expired before the claimant was exposed to the policyholder’s product.” (Id. at p. 62.) The London defendants’ position was also supported by this appellate district’s opinion in A. C. Label Co. v. Transamerica Ins. Co., supra, 48 Cal.App.4th 1188, a groundwater contamination case on “category B” facts, which was not yet final at the time the London defendants’ motion was heard.

FMC sought to distinguish Cooper Companies and A. C. Label, but conceded that, although it disagreed with Armstrong World Industries, the trial court would be obliged under Armstrong World Industries to dismiss FMC’s coverage claims as to Ekotek, Delaware Sand & Gravel, and Global Landfill. The trial court granted the London defendants’ motion on the basis both of Armstrong World Industries and of the court’s perception that FMC might not have had an “insurable interest,” as required and defined in Insurance Code section 280 and following sections, in these sites.

FMC made similar concessions, and the trial court ruled similarly, as to four “category A” sites — J.I.S. Landfill, Kin-Buc Landfill, and Higgins Farm, in New Jersey, and Mountaineer Refinery, in Wyoming — assigned to the eighth trial.

On their first cross-appeal the London defendants challenge the trial court’s ruling as to the Andover site. By its fourth, seventh and eighth appeals FMC, having reconsidered the applicability of Armstrong World Industries, challenges the rulings denying coverage as to the “category A” and “category B” sites assigned to those trials.

The crux of FMC’s argument is that the London policies’ definition of an “occurrence” was unambiguous as to what must occur during a policy period, requiring in pertinent part only unexpected or unintended “property damage” without reference to any requirement that FMC have caused, or otherwise have been contemporaneously liable for, the “property damage,” and that it is undisputed that there was in fact “property damage” sufficient to constitute a continuous injury trigger under Montrose II.

The London defendants’ primary position is that this case is controlled by A. C. Label, in which a majority of this court’s panel held, on “category B” facts and under general liability policy language similar in substance to that of the London policies, that “[t]he coverage provided by [the general liability] policy was not triggered during the policy period because plaintiffs had no connection to or nexus with the damage caused by contamination that occurred on the subsequently acquired property during the policy period.” (48 Cal.App.4th at p. 1194.)

FMC respectfully suggests that this court decided A. C. Label incorrectly. In essence FMC agrees with the dissent in A. C. Label that “the clauses are the same standard clauses that were at issue in Montrose [//].... The court therein settled the meaning of the clauses: ‘We find no ambiguity in this language; it clearly and explicitly provides that the occurrence of bodily injury or property damage during the policy period is the operative event that triggers coverage.’ ([10 Cal.4th] at p. 668.) [*[[]... Since there is no question that there was an occurrence during the policy period in this case, the policy affords plaintiff coverage.” (A. C. Label Co. v. Transamerica Ins. Co., supra, 48 Cal.App.4th at pp. 1195-1196 (dis. opn. of Premo, Acting P. J.).)

FMC argues that A. C. Label was flawed by its asserted failure to acknowledge that “[t]he Supreme Court has held that statutorily created, retroactive, strict liability is covered under general liability policies. As long as the necessary triggering event (i.e., injury or damage) takes place during the policy period and the insured is held liable for that injury or damage, coverage exists under occurrence-based CGL policies.”

FMC’s reference is to the following footnote, in the Supreme Court’s opinion in AIU Ins. Co. v. Superior Court, supra, 51 Cal.3d 807, in support of a text generalization that “. . . we generally interpret the coverage clauses of insurance policies broadly, protecting the objectively reasonable expectations of the insured”: “Although our focus is the expectations of the insured at the time the policy is made, this emphasis does not preclude coverage of forms of liability — such as those at issue here — created after the formation of the policy. Because the policies in question here are ‘comprehensive,’ it was within the insured’s reasonable expectation that new types of statutory liability would be covered, as long as they were within the ambit of the language used in the coverage provision. As one court has pointed out, failure to cover new liabilities would create a ‘discordant result, for it would mean that where courts enlarge liability during the effective period of a liability policy, an insured who contracted for complete coverage of a possible risk would be left without coverage because the scope of the risk had been enlarged by decisional law.’ [Citation.] The same is true when legislatures create entirely new forms of liability. The sole relevant inquiry in determining whether such types of liability are covered is whether, in view of the reasonable expectations of the insured, policy language can be interpreted to embrace the liability that may accrue under new statutory schemes.” (51 Cal.3d at p. 822, fn. 8; cf. also Armstrong World Industries, Inc. v. Aetna Casualty & Surety Co., supra, 45 Cal.App.4th at p. 59, fn. 23; Cooper Companies v. Transcontinental Ins. Co., supra, 31 Cal.App.4th at p. 1107, in. 10; Travelers Ins. Co. v. Industrial Indem. Co. (1971) 18 Cal.App.3d 628, 632 [96 Cal.Rptr. 191]; Pacific Indemnity Co. v. Liberty Mutual Ins. Co. (1966) 239 Cal.App.2d 346, 352 [48 Cal.Rptr. 667].)

We understand the Supreme Court to have referred to the well-established rule that liability insurance coverage will extend to liabilities created or enlarged by post-policy period changes in applicable law. We do not understand the parties to disagree with this principle, which is applicable in this case inasmuch as CERCLA (for example) was not enacted until 1980, 10 years after the last London policy period relevant to the issues before us. In no case would, or could, FMC have been declared liable under CERCLA at the time damage occurred within the policy period, because at that time CERCLA did not exist. FMC subsequently became liable, by virtue of CERCLA, for damage which had occurred within policy periods several years before.

But in every instance, other than at Andover, in which the trial court found coverage under the London policies, it was also true that every factual predicate for CERCLA liability, including a nexus between FMC and the damage or the site, could be shown to have existed before or during the policy period.

We consider such a showing of a contemporaneous factual predicate essential to the kind of coverage for statutorily expanded risk to which ALU’s footnote referred. For this reason we shall conclude, as we concluded in A. C. Label, that coverage for liability under subsequently enacted statutes was not triggered in policy periods in which a sufficient factual nexus could not be shown.

Our conclusion that a complete factual predicate for a liability subsequently imposed by law must exist during the policy period is based in our perception that while it is indeed appropriate to require that an insurer that writes, and accepts a substantial premium for, general liability insurance assume the risk of expansion in legal theories of liability applicable to facts which occurred in or before the policy period, it is neither reasonable nor consonant with the terms of the general liability policies before us to require such insurers to cover liabilities based on facts which did not occur until after the policy period. A general liability insurer can realistically be said to be in the business of understanding and taking into account the legislative and judicial dynamics that produce changes in legal theories, but cannot be required to be clairvoyant as to the infinite possible future permutations of facts, fundamental to the very existence of coverage but not in existence during the policy period, once the policy period has expired.

FMC argues that such a conclusion would essentially rewrite the policies in light of judicial concepts of fairness and justice, in violation of the Supreme Court’s recent admonition, in Aerojet-General Corp. v. Transport Indemnity Co. (1997) 17 Cal.4th 38 [70 Cal.Rptr.2d 118, 948 P.2d 909], that the insurance policies before the court in that case “provide what they provide,” and that in agreeing to the policies the parties “established what was ‘fair’ and ‘just’ inter se. We may not rewrite what they themselves wrote. . . . As a general matter at least, we do not add to, take away from, or otherwise modify a contract for ‘public policy considerations.’ ” (17 Cal.4th at p. 75.)

We disagree with FMC’s restrictive reading of the Supreme Court’s language in Aerojet. Aerojet itself furnishes vivid evidence that its admonition that insurance policies “provide what they provide” is applicable only “[a]s a general matter” and by no means inflexibly. In Buss v. Superior Court (1997) 16 Cal.4th 35 [65 Cal.Rptr.2d 366, 939 P.2d 766], the Supreme Court had revalidated the duty of an insurer whose policy contains a duty to defend to undertake the entire defense of a “mixed” action involving both potentially covered claims and claims which were not potentially covered. The Supreme Court explicitly acknowledged that “[w]e cannot justify the insurer’s duty to defend the entire ‘mixed’ action contractually, as an obligation arising out of the policy, and have never even attempted to do so. . . . ffl] That being said, we can, and do, justify the insurer’s duty to defend the entire ‘mixed’ action prophylactically, as an obligation imposed by law in support of the policy.” (16 Cal.4th at pp. 48-49, fn. omitted.) In Aerojet the Supreme Court reaffirmed both this rule and its rationale. Patently the Supreme Court did not regard the duty to defend “mixed” actions as one to which policies should be deemed to “provide what they provide.”

We believe the issue before us is perhaps comparable in significance to that of the duty to defend “mixed” actions. Certainly the problem goes beyond mere generalization as to what might be considered “fair” or “just.” It is fundamental to the concept of liability insurance. We are satisfied that we have reached a valid resolution.

The London defendants also reassert the trial court’s apparent conclusion that, for want of a sufficient nexus, FMC did not have the “insurable interest” required by Insurance Code section 280 and following sections in the sites in “category A” and “category B.” We have concerns about the relevance of property interests to the concept of insurable interest in the context of liability insurance (cf., e.g., Osborne v. Security Ins. Co. (1957) 155 Cal.App.2d 201, 205 [318 P.2d 94]; Davis v. California Highway Indem. Exch. (1931) 118 CaLApp. 403, 406 [5 P.2d 447]), but in light of the conclusion we have stated we need not resolve the question in this action.

Our conclusion requires that the judgment be modified to reflect no coverage under the London policies, for any of their policy periods, at the Andover, Minnesota site.

“Unexpectedly”

Under the London policies’ definition of “occurrence,” FMC could establish coverage only if the property damage for which it was liable had occurred “unexpectedly and unintentionally.” The parties appear to have agreed, and early in the proceedings the trial court ruled, that “unintentionally” was not in issue. But the question whether environmental harm had in any particular instance been unexpected from FMC’s standpoint pervaded the proceedings. Shortly before the first trial, the Court of Appeal, First Appellate District held that “[t]he appropriate test for ‘expected’ damage is whether the insured knew or believed its conduct was substantially certain or highly likely to result in that kind of damage.” (Shell Oil Co. v. Winterthur Swiss Ins. Co., supra, 12 Cal.App.4th at p. 748.) The trial court applied this subjective test throughout these proceedings.

1. Burden of Proof as to “Unexpectedly”

Would FMC be required to prove that the property damage had been unexpected? Or would the insurers be required to prove that the property damage had not been unexpected? The trial court concluded that the burden of proving the property damage was unexpected should be borne by FMC, as the insured, and it instructed the juries accordingly at each of the eight trials.

The juries reached the “unexpectedly” issue with respect to a total of 19 sites, and concluded that the property damage had not been unexpected (and thus, in necessary effect, that FMC had not borne the burden assigned to it by the trial court) at 7 of these sites. On the basis of the juries’ findings, the trial court denied coverage at each of these seven sites. On appeal, FMC argues that it should not have been required to bear the burden of proof as to this issue, and that the trial court’s error manifestly prejudiced it with respect to five of the seven sites. FMC does not raise the issue as to the Julian Street site involved in the third trial or the Mouat site in the fifth trial.

FMC argues, essentially in the alternative, (a) that the “unexpectedly and unintentionally” language of the policies’ “occurrence” definition was in essence an exclusion from coverage as to which, under the general rule, the insurer should bear the burden of proof, and (b) that to place the burden on FMC would be to require FMC to implicate itself, in violation of Evidence Code section 520, in the commission of “wrongdoing.” We disagree with both branches of FMC’s argument and shall validate the trial court’s ruling and instructions.

a. Exclusion From Coverage

Each of the London umbrella policies, like standard form CGL policies, contained an insuring agreement followed by separately stated exclusions. The Supreme Court has recently restated the approach courts must take to coverage questions under such policies: “[T]he insuring agreement . . . states the risk or risks covered by the policy, and the exclusion clauses . . . remove coverage for risks that would otherwise fall within the insuring clause. [Citation.] Before ‘even considering exclusions, a court must examine the coverage provisions to determine whether a claim falls within [the policy terms].’ [Citation.] ‘This is significant for two reasons. First, “. . . when an occurrence is clearly not included within the coverage afforded by the insuring clause, it need not also be specifically excluded.” ’ [Citation.]

“ ‘Second, although exclusions are construed narrowly and must be proven by the insurer, the burden is on the insured to bring the claim within the basic scope of coverage, and (unlike exclusions) courts will not indulge in a forced construction of the policy’s insuring clause to bring a claim within the policy’s coverage.’ [Citation.] Accordingly, the insured has the burden of showing that there has been an ‘occurrence’ within the terms of the policy.” (Waller v. Truck Ins. Exchange, Inc. (1995) 11 Cal.4th 1, 16 [44 Cal.Rptr.2d 370, 900 P.2d 619]; cf. also Collin v. American Empire Ins. Co. (1994) 21 Cal.App.4th 787, 802-803 [26 Cal.Rptr.2d 391]; Hallmark Ins. Co. v. Superior Court (1988) 201 Cal.App.3d 1014, 1017 [247 Cal.Rptr. 638].)

Once the insured has borne the burden of bringing the claim within the basic scope of coverage, “[t]he burden then shifts to the insurer to prove the claim falls within an exclusion.” (Merced Mutual Ins. Co. v. Mendez (1989) 213 Cal.App.3d 41, 47 [261 Cal.Rptr. 273]; cf. also Royal Globe Ins. Co. v. Whitaker (1986) 181 Cal.App.3d 532, 537 [226 Cal.Rptr. 435]; Executive Aviation, Inc. v. National Ins. Underwriters (1971) 16 Cal.App.3d 799, 806 [94 Cal.Rptr. 347].)

The London policies’ “occurrence” definition is manifestly an inseparable element of the policies’ insuring agreements, which state the London defendants’ obligations in terms (among others) of “occurrence.” On the face of the policies, and under the stated rules, the burden would be upon FMC initially to prove that its claim was based on facts which came within the policy definition of “occurrence,” including the definition’s requirement that property damage resulted “unexpectedly.”

To sustain its argument, to the contrary, that the “unexpectedly . . . results in . . . property damage” clause of the “occurrence” definition was the functional equivalent of a policy exclusion as to which the London defendants should bear the burden of proof, FMC relies primarily on United Pacific Ins. Co. v. McGuire Co. (1991) 229 Cal.App.3d 1560 [281 Cal.Rptr. 375]. United Pacific did not involve allocation of the burden of proof. The issue, raised on appeal from a summary judgment, was whether a CGL insurer owed a duty to defend a wrongful termination action under an “ ‘Extended Definition of Occurrence’ ” as “ ‘an accident, an event or a continuous or repeated exposure to conditions which results, during the policy period, in bodily injury or property damage neither expected nor intended by the insured.’ ” (229 Cal.App.3d at p. 1563.) Earlier cases had held, in essence, that wrongful termination could not be regarded as an “accident,” but the United Pacific court perceived that the word “event” “obviously embraces intentional conduct” (id. at p. 1565); the dispositive issue was whether the phrase “neither expected nor intended” should be construed to qualify the word “event” (so as to limit coverage to events that were neither expected nor intended) or, instead, the words “bodily injury or property damage” (so as to permit coverage of even intentional conduct so long as particular resultant harm was not expected or intended). Focusing on construction of the phrase “neither expected nor intended,” the United Pacific court said: “Since the word ‘event’ is not limited to fortuitous happenings, the phrase ‘not expected or intended’ [az'c] cannot be read as language confirming the meaning of the term; in the context of the extended definition, the phrase must be regarded as language of limitation, narrowing the coverage otherwise provided by the word ‘event.’ As a provision limiting coverage, the phrase performs precisely the same function as the common exclusion for intentional conduct. It does not matter that the phrase appears in the ‘definitions’ section of the policy rather than the ‘exclusions’ section; in either case it performs the function of an exclusion. Consequently . . .the phrase is subject to the ‘well settled principle that such exclusionary clauses should be interpreted as narrowly as possible.’ ” (229 Cal.App.3d at p. 1565.)

Construing the phrase narrowly, the United Pacific court concluded that the phrase limited only the scope of covered harms, and that (inferably because there was potential coverage for the arguably unexpected and unintended mental and emotional distress damages the discharged employee sought) the insurer had owed a duty to defend.

In our view United Pacific’s construction of the phrase in the policy before it, for purposes of invoking a rule of policy interpretation, does not resolve the burden of proof issue with respect to the syntactically distinguishable “occurrence” definition before us. Nor are we bound by the findings and conclusions of a federal trial judge, apparently based in part on the testimony of expert witnesses, that the phrase “neither expected no[r] intended” in a standard form CGL policy’s “occurrence” definition was an “exclusion of coverage” as to which the insurer would bear the burden under California law. (Clemco Industries v. Commercial Union Ins. Co. (N.D.Cal. 1987) 665 F.Supp. 816, 820-821.) California courts that have chosen to address the burden of proof issue, in cases in which the issue appears not to have been directly raised, have not reached consistent results. (Compare, e.g., Waller v. Truck Ins. Exchange, Inc., supra, 11 Cal.4th at p. 20 [regarding duty to defend, “insured must show . . . neither expected nor intended”] with Armstrong World Industries, Inc. v. Aetna Casualty & Surety Co., supra, 45 Cal.App.4th at p. 77 [“insurer’s burden was to prove . . . that [insured] actually did expect . . .”].) Cases from other jurisdictions are similarly divided. (Compare Clemco Industries v. Commercial Union Ins. Co., with Queen City Farms v. Central Nat. Ins. Co. (1994) 126 Wn.2d 50 [882 P.2d 703] [burden is on insured].)

After consideration of these authorities, we adhere to our conclusion that the phrase “unexpectedly and unintentionally” is an integral part of the “occurrence” definition before us in this case, and thus is subject to the rule assigning the burden of proof of coverage to the insured. Our conclusion is reinforced by the persuasive insights of the Washington Supreme Court in Queen City Farms v. Central Nat. Ins. Co., on this issue: “[T]he argument that the ‘unexpected or unintended’ language is exclusionary is not a particularly strong argument when deciding who has the burden of proof on this issue, because ‘virtually all the language in the Insuring Agreement of CGL policies after the insurer’s promise to “pay all sums the insured shall become legally obligated to pay . . .” qualifies or limits the scope of this promise in one way or another.’ K. Abraham, Environmental Liability Insurance Law 140-41 (1991). [H] . . . RD Professor Abraham ■ suggests other considerations, including the notion that the burden of proof should be on the insured because the insured is ‘likely to be in possession of or have greater access to whatever information exists about its expectations or intentions . . . .’ K. Abraham, at 140. This reasoning is compelling in the case where a subjective standard is applied.” (126 Wn.2d at pp. 556-557 [882 P.2d at pp. 715-716].) These observations are germane under California law (cf. 1 Witkin, Cal. Evidence (3d ed. 1986) Burden of Proof and Presumptions, § 136, p. 119 et seq.) and to the analogous language of the London defendants’ general liability insurance policies.

b. Evidence Code Section 520

Evidence Code section 520 provides that “[t]he party claiming that a person is guilty of crime or wrongdoing has the burden of proof on that issue.”

FMC did not invoke Evidence Code section 520 in the trial court, where the section’s application to particular circumstances might have been assessed in light of those circumstances. In this court, to bring the “unexpectedly” element of the London policies’ “occurrence” definition within section 520, FMC is obliged to establish, essentially as an abstract generalization, that proof that property damage caused by its releases of toxic contaminants was not unexpected from its standpoint would be tantamount to proof that it was “guilty of crime or wrongdoing.”

For this purpose FMC relies on the subjective test for “expected” damage enunciated in Shell and applied for other purposes at the trials of this action: “whether the insured knew or believed its conduct was substantially certain or highly likely to result in that kind of damage.” (Shell Oil Co. v. Winterthur Swiss Ins. Co., supra, 12 Cal.App.4th at p. 748.) FMC asks us to conclude that conduct that met the Shell test, in the factual context of this action, would necessarily amount at least to “wrongdoing” within the meaning of Evidence Code section 520.

We cannot so conclude. The wording, history and prior judicial application of Evidence Code section 520 all support a finding that “wrongdoing” connotes an element of moral disapprobation substantially greater than that which might be implied by the Shell test. Thus (to take as examples two cases FMC has cited) in Clemmer v. Hartford Insurance Co. (1978) 22 Cal.3d 865, 880 [151 Cal.Rptr. 285, 587 P.2d 1098], it was deemed “consistent with . . . section 520” to require an insurer to show that its insured had acted wilfully (within the meaning of Insurance Code section 533) in shooting his employer to death, and in Fisher v. Superior Court (1980) 103 Cal.App.3d 434, 448-449 [163 Cal.Rptr. 47], the Court of Appeal referred to section 520 in apparent support of its conclusion that the burden of proof as to the “good faith” of a settlement should be upon the party who asserted the settlement was not made in good faith, i.e., was made in bad faith. We find no comparable connotation compelled in the context of liability, under essentially regulatory statutes, for toxic contamination by reason of acts in many instances done long before the statutes were enacted.

2. Separate Sources of Damage

FMC occasionally perceived that it might be unable to prove that certain of the third party property damage at or attributable to a site had been unexpected. One tactical response to such a perception was to argue that such damage could be isolated from other third party property damage at or attributable to the site, and that FMC should be permitted to prove that the other damage, viewed in isolation, had been unexpected and thus to obtain coverage at least for its liability for the other damage. Under the policy language such an argument was necessarily addressed to, and phrased in terms of, the policy definition of “occurrence.”

A site might (for example) have had two sources of pollutants, each of which had contributed to groundwater contamination, and while FMC could hope to prove that it had not expected one of the sources to cause the damage it could not reasonably expect to prove that the damage caused by the other source was unexpected. If the totality of damage attributable to the two sources were considered a single occurrence, then FMC’s inability to establish the “unexpectedly” element as to one of the sources would jeopardize its prospects for coverage as to either source. If, on the other hand, FMC could establish that each source should be treated as the basis of a separate occurrence, then it might be able to show that damage was unexpected as to one source, and thus preserve coverage for its liability for damage attributable to that source, even though it could not make a similar showing as to the other source.

Under California case law, for the purpose of determining the number of occurrences under a liability insurance policy (usually as a means of calculating policy limits) “occurrence has generally been held to mean the underlying cause of the injury, rather than the injury or claim itself . . . .” (Whittaker Corp. v. Allianz Underwriters, Inc. (1992) 11 Cal.App.4th 1236, 1242 [14 Cal.Rptr.2d 659].)

The London umbrella policies addressed the determination of the number of occurrences only in the provision, appended to the “occurrence” definition, that “[a]ll such exposure to substantially the same general conditions existing at or emanating from one premises location shall be deemed one occurrence.” In standard form policies this provision has been referred to as a “one occurrence clause.” (Cf. Endicott Johnson Corp. v. Liberty Mut. Ins. Co. (N.D.N.Y. 1996) 928 F.Supp. 176, 179.)

Before the first trial in this action the trial court filed a written ruling that in essence made both the case law rule and the one occurrence clause applicable to these trials without undertaking to synthesize the two. The appropriate synthesis is apparent: The one occurrence clause would qualify the case law rule in any case to which the one occurrence clause would be applicable, so that more than one occurrence could be found on the basis of separately identifiable causes of property damage unless the property damage had been caused by an exposure to “substantially the same general conditions existing at or emanating from one premises location.”

In the proceedings that led to this ruling, counsel for FMC acknowledged “the general proposition that there is one occurrence per site, unless there [are] peculiar facts under this language where there would be more than one.”

One of the sites assigned to the third trial, the Green River site in Wyoming, potentially raised a problem similar to that illustrated by our example: There were at least two separately identifiable sources of contamination on the site, and the likelihood of contamination arguably was more apparent at one source than at another. The third trial jury was given a definition of “occurrence” but was neither instructed as to multiple occurrences nor given the option on the special verdict form of making separate findings as to the separate sources on the Green River site. By their questions to the court during their deliberation, in November 1995, the jurors made clear that they had focused on and were concerned by the distinctions between the two sources. The trial court responded by referring the jury to the instructions already given, but then added the one occurrence clause: “As to property damage resulting from a continuous or repeated exposure to conditions, you are now further instructed that all such exposure to substantially the same general conditions existing or emanating from the Green River site shall be deemed one occurrence.” The court did not alter the special verdict form. The jury ultimately found that there had been third party property damage at the Green River site and that the damage had been unexpected, and the trial court found coverage under the London policies for the Green River site as a whole.

The multiple-source problem arose again at the Carteret site in New Jersey (assigned to the fourth trial) and at the Pocatello site in Idaho (assigned to the fifth trial), but in these instances FMC did not obtain an adjudication of coverage. On appeal FMC complains of trial court rulings which effectively denied FMC an opportunity to argue, as to the Carteret and Pocatello sites, that the jury could consider the “unexpectedly” issue separately as to sources where FMC believed it had better prospects for coverage.

a. Carteret

At Carteret, groundwater contamination was traced to two sources on the site: a sand-lined “dewatering pit” which was identified as a source of arsenic contamination, and a concrete storage tank known as the “P-4 pit” which was identified as a source of contamination by phosphorus and phosphates. It was apparent from the evidence at the fourth trial that the jury could have concluded that FMC was more likely to have expected contamination from the dewatering pit than from the P-4 pit.

Proposed jury instructions and special verdict forms for the fourth trial had been due, under the trial court’s elaborate case management orders, on January 2, 1996, shortly after the jury questions and verdicts at the third trial. Possibly sensing how close it had come to an all-or-nothing disaster as to the Green River site, FMC submitted a proposed special verdict form that would have called for separate findings as to the dewatering pit and the P-4 pit at the Carteret site. But FMC did not at that time request any instruction on multiple occurrences, or otherwise in support of its proposal for separate verdicts for the two Carteret sources; FMC did request a general definition of “occurrence” which tracked the London policies’ definition (but without the one occurrence clause).

On February 20, 1996, after approximately four weeks of FMC’s five-week case-in-chief at the fourth trial, FMC submitted “proposed additional and revised jury instructions” which included two new requested instructions germane to the issue before us:

“Instruction No. 33” would have added the one occurrence clause to FMC’s previously requested “occurrence” definition.

“Instruction No. 33A” would have told the jury that “[mjore than one ‘occurrence’ may exist at a site for coverage purposes. The number of occurrences is determined by the number of cause(s) of the property damage. You must decide whether particular acts, practices, or activities at a particular site location are so unrelated as to support a finding of more than one occurrence at a site.”

At a hearing three days later, on February 23, the trial court questioned counsel for FMC concerning the proposed instructions: “Now we are talking about multiple occurrences, or multiple something? [5D [Counsel for FMC]: Sources, ffl]. • • • ffl] Or causes for the same injury.” Counsel told the court she was requesting the instruction “because the burden of proof being on the plaintiff, we think we need to address it on a source-by-source basis.”

It was not until March 13, after both sides had rested and shortly before summations, that counsel for FMC explicitly stated that “we believe that the exposure from the P-4 pit could be considered a different issue, and it is, in fact, a different occurrence from the exposure relating to the dewatering pit.” Counsel rationalized FMC’s position in terms of perceived “fundamental differences” between the dewatering pit and the P-4 pit.

After further argument on March 13, the trial court refused both “Instruction No. 33” and “Instruction No. 33A.” Apparently the court’s only reservation as to “Instruction No. 33” was that it had been submitted solely to lend context to “Instruction No. 33A,” but the court was not prepared to give “Instruction No. 33A” in any event: “From the plaintiff’s argument, it sounds arguable that they could have made a case for two occurrences, but I find that they did not. Cannot try a case in a vacuum, it has not been noticed, it comes too late, so the issue has been deemed waived. HQ . . . HQ . . . The instruction that plaintiff seeks ... is that there are two occurrences, or can be two occurrences at Carteret. There is no such instruction, it comes too late, comes about 8 weeks too lat