“Who pays” and “how much” continue to be central questions in insurance-recovery litigation by policyholders for asbestos, environmental clean up, pharmaceutical, lead-paint, toxic-tort and other conditions that produce loss over time. Because insurance contracts are governed by state law, the coverage wars apparently will continue until each inch of turf is won or lost. Most recently, the Delaware Supreme Court has weighed in on the question of trigger of coverage (“who pays”) for asbestos- liability claims, the Minnesota Supreme Court has addressed allocation of loss (“how much”) among triggered policies, and the New Hampshire Supreme Court has now been asked to address the allocation question, too.
The Delaware and Minnesota cases suffer from an overly conceptualist approach to looking at the questions presented; instead of applying the contract language, these courts have applied “models” of how the coverage should apply. As these cases show, misframing the question to be answered results in answers that are not entirely satisfactory or coherent.
Starting with Delaware: The question presented in Shook & Fletcher Asbestos Settlement Trust v. Safety National Cas. Corp. (Del. Sept. 26, 2006) was what event must occur during the policy period in order for an occurrence policy to be activated (triggered). An insurer has no obligation to perform unless the claim against the insured triggers (or potentially triggers) its coverage. Shook & Fletcher was decided under Alabama law. As noted by the Delaware court, in prior coverage litigation “after full briefing and argument, the Alabama court held that the ‘exposure coverage theory’ would apply, instead of the ‘continuous trigger’ or ‘triple trigger’ theory.” Slip op. at 3-4.
The Delaware Supreme Court in trying to predict Alabama law sought to resolve the question in part by counting noses among state Supreme Courts and federal appellate courts. As the court explained: “We have analyzed the cases where the parties disagreed upon which trigger theory the court actually adopted. Our analysis shows that the exposure trigger is the majority rule.” Slip op. at 9. The court found that Pennsylvania, New Jersey, Delaware, and the Third and D.C. Circuit’s have adopted a continuous trigger; but the Court’s analysis of the other cases led it to conclude that the following jurisdictions have adopted an exposure trigger for asbestos bodily injury claims: Louisiana, Massachusetts, Maryland, Illinois, Fourth Circuit, Fifth Circuit (Texas), Fifth Circuit (Louisiana), Sixth Circuit (Ohio), Sixth Circuit (Michigan), Eleventh Circuit (Alabama), and the First Circuit.
Holding aside that I quarrel with the court’s table and analysis of cases, the court thus concluded: “We rely on the rationale of those cases and the fact that the exposure trigger is the majority rule.” Slip op at 11.
The Delaware court offers no analysis of the issue on its own but rather tries to predict what the Alabama Supreme Court might do. In this context, it is passing strange that the court does not address on the merits its own prior continuing-injury trigger decisions or come to grips with the rule of construction – applicable in Alabama – that if the policy language admits of more than one reasonable construction the court is to construe the language in favor of coverage. Compare Hercules Inc. v. AIU Ins. Co. (Del. Aug. 15, 2001) (finding plain language compels an “all sums” allocation as do equitable considerations). Is the Delaware court now saying that its own prior decision was unreasonable?
In Shook & Fletcher, no effort was made to look at the policy language. And the language is straightforward: insurance policies like those in Shook & Fletcher are triggered if injury occurs during the policy period. Trigger is the issue of what event must take place during the policy period, and the policy is clear – it is not the occurrence, the exposure, or the accident: trigger is “injury.” While “injury” may follow from “exposure,” the difference is one between cause and effect, and typically liability policies are triggered by the effect. (The number of occurrences is determined by causes, but that is a different legal and contractual issue.)
Policyholders should not argue, as I have sometimes seen them do, that there are as many as eight different trigger theories out there and the job of the court is to pick one. Rather, the policy language is absolutely clear that injury is the trigger, and the only question is whether, in the particular context before the court, can whatever happened during the policy year under scrutiny reasonably be construed to be “injury.”
So, the right way for policyholders to argue is not to serve up the question as “which trigger applies: exposure, manifestation, continuous”? The right question is: “Because the trigger is injury, did injury occur during this policy period?” For policies in effect during the period of exposure, one argues that following exposure the body suffers injury. For policies in the period after exposure but while the asbestos fibers persist in the body, one argues that the continued presence of asbestos in the body produces a reaction such that that further effect constitutes injury that year. For policies in effect during the period that an asbestos-related disease is diagnosed, one argues that under the definition of bodily injury – which includes “bodily injury [and] disease” – there is disease and thus injury during the policy period.
It doesn’t matter whether we are arguing about asbestos, or pharmaceutical, or lead paint, or toxic torts: the question of trigger under “occurrence” CGL policies is injury.
A policy’s being triggered does not answer the question of how much the triggered policy is required to pay for the insured’s loss. The “how much” question is known variously as “allocation” or “scope of coverage.” The Minnesota Supreme Court recently weighed in on allocation of coverage, but in doing so it had to deal with prior authority holding that a continuing-injury scenario triggered multiple policies across time and that presumptively each insurer’s obligation to perform was dependent on the quantum of injury that occurred in its policy year. So for illustration, if damage occurred in a steady-stream way for 10 years and caused $10 million in damages, then $1 million would be allocated per year, no matter that in some years the insured bought much more coverage or that in other years the primary layer was, e.g., $200,000 and in other years it was $50,000.
In Wooddale Builders, Inc. v. Maryland Cas. Co. (Minn. Oct. 8, 2006), the court confronted a number of questions concerning the implementation of this kind of scheme: if the damage occurs during only part of a year, does the carrier have to pay its full limit? What happens for years in which there is no coverage? Can the insured buy more coverage once it has knowledge of the problem?
The Court framed the basic issue as follows:
The relationship [among the policies] can be expressed – and perhaps best understood – in terms of a simple equation: A/B x C = D. In this equation, A is each insurer’s time on the risk, B is the total period over which liability is allocated, C is the total damages to be allocated, and D is the damages allocated to each individual insurer.
The court first addressed whether the coverage period is cut off at some moment due to the insured’s knowledge of the risk of liability. In Wooddale, the case involved damage to a number of homes that had been built, and the court ruled that once the insured had knowledge that a home had been damaged no further insurance available. The court ruled – in error – that damage was expected or intended (and thus excluded) once the insured had knowledge that some homes had been damage: “The practical effect of the policy language excluding expected damage and the rationale behind the known loss/loss in progress doctrine is that no additional insurance policies are triggered by continuing damage to homes for which claims had been made before those policies took effect. . . . We therefore hold that only insurers that provided coverage to Wooddale between the closing date of a particular home and Wooddale’s receipt of notice of claim with respect to that property are on the risk for that claim.” (fn. omitted).
(The court’s cutoff based on these theories is wrong because expected/intended damage is meant to stop the insured from acting or setting in motion the chain of events leading to damage rather than policing the insured’s knowledge of a ticking time bomb – the right mechanism to protect insurers on this point is nondisclosure; and known loss is not properly applied here either because from an insurance perspective there is insurable risk, and thus no known loss, so long as there is uncertainty as to whether, when or how much liability will be found – again, all subject to non disclosure rules. See Transamerica Ins. Group v. Meere, 694 P.2d 181, 186 (Ariz. 1984) (“”The exclusion ‘is designed to prevent an insured from acting wrongfully with the security of knowing that his insurance company will “pay the piper” for the damages.”’); Bob Hedges, Back-Dated Coverage as Insurance, 34 CPCU J. 181, 181 (1981). And the court does not address equities such as whether the policyholder paid premium in effect for this risk exposure that the carrier gets to pocket under the court’s analysis.)
The court further examined the consequence on an insurer’s obligation to pay from the trigger cutting off during its policy period, or put differently is the insurer that provides coverage for only part of a policy year required to pay its full limit? Here, the court got the answer right, but because its rationale is embedded within the pro-rata-by-quantum-of-damage methodology its reasoning is a bit unsatisfactory. The court offered only the rationale that it was aware of no basis for concluding that a full limit does not apply. “Instead, each of the policies provides coverage for ‘property damage’ that ‘occurs during the policy period,’ indicating that the insurer has agreed to indemnify the insured for damages that occur during the entire policy period, including the part of the policy period that runs after notice of the claim.”
The court does not address the case law on “stub policies,” that is, the limits available in a part year policy by analogy (as opposed to a full year policy with only part damage). Stub policy case law uniformly recognizes that the carrier’s policy limits are at risk every day of the policy, and that the premium is paid for the availability of the full limits each day. See United States Mineral Products Co. v. American Ins. Co., 792 A.2d 500, 502 (N.J. Super. App. Div. 2002); Stonewall Ins. Co. v. ACMC, 73 F.2d 1178, 1216-17 (2d Cir. 1995); Cadet Mfg. Co. v. Am. Ins. Co., 391 F. Supp. 2d 884, 890 (W.D. Wash. 2005). More importantly, the allocation method that the court has previously embraced does suggest only a partial limit for part-year damage, and the court does little to explain other than offer its ipse dixit.
The court reached the parallel conclusion for policies in effect during the year in which damage first occurred, likewise holding that the full limits were available in the allocation formula.
Finally, the court considered whether periods of no insurance counted in the allocation formula, and the court divided periods of no insurance into (i) periods where the insured elected not to purchase insurance voluntarily (self-insurance) and (ii) periods where due to market forces there was no insurance to be purchased. The court concluded that allocating to periods of self-insurance was fair but that allocating to periods of no insurance was not, and so periods where there was no insurance get removed from the denominator of the allocation formula.
This has always been an unsatisfying approach, in part because it makes the contractual obligations of a party with a closed period dependent upon failures in insurance markets years later perhaps (because if a later period of no insurance is not included in the allocation formula each insurer’s relative share of the pie goes up). It accomplishes the objective of avoiding penalizing the insured for no “fault” of its own, but it also is somewhat naïve in its assessment of insurance markets and whether some underwriter somewhere might be willing to insure some risk for some ridiculous price (as the London market saying goes, there is no bad risk, just a bad price). So even though there was a worldwide conspiracy among insurers and reinsurers to cut off asbestos coverage in the mid 1980s, if an underwriter wanted to sell asbestos insurance doing so would not violate any positive law or prohibition against meretricious contracts. These are, to my way of thinking, too unstable of bases to support a rationale for how to apply an insurance contract that may have been written decades before.
Overall the Minnesota court made the following simplifying assumptions in its mathematical formula:
a. Each policy in effect during when any damage occurred is exposed for its whole limit.
b. Policies incepting after the insured expected claims have no obligation to perform.
c. A policyholder should be treated as if it were an insurance company if it voluntarily elected not to purchase insurance in the commercial markets.
d. A policyholder should not be treated as an insurer if the reason it did not purchase insurance is through no fault of its own.
Based on these elaborations, the court held that as to the duty to indemnify each carrier’s obligation to pay can be derived. (The court however embraced a per-capita allocation for defense.)
The Minnesota court recognized that its elaborations meant that its formulaic approach was becoming unmoored from the principles from which it derived. (No matter; such is the privilege of being the last word.) As the court said:
We are aware it is possible under our construction of factor B that an insurer is liable for more than those damages than would otherwise be deemed to have occurred during its policy period. This is possible because our definition of factor B, the period over which damages are to be allocated, excludes periods during which the insured lacks coverage because no such coverage was available. We are also aware that this result may be contrary to the broad language we used in [prior authority] to describe the ‘actual injury’ [trigger] rule, namely, that under this rule each insurer is liable ‘only for those damages which occurred during its policy period.’ [citation omitted] However, as we [there] indicated . . . ., the allocation of liability between insurers requires a flexible approach. [citation] Further, as we noted [previously], it is inaccurate to conclude that a CGL insurer can never be liable for damages occurring outside of a policy period. [citation] We deem the facts of this case to justify a departure from the typical ‘actual injury’ approach.
So is there a principled way to come to grips with the language of the insurance policy and figure out a principled manner to determine the obligation of each carrier where a single loss situation produces damage and injury both during and outside of its policy period? This is now the question that the New Hampshire Supreme Court has been asked in an environmental-coverage case. See EnergyNorth Natural Gas Inc. v. Certain Underwriters at Lloyd’s Underwriters, 2006 U.S. Dist. LEXIS 73468 (D.N.H.)
I believe there is an approach to the contract language that takes the language seriously yet produces results that are (I submit) intellectually consistent.
The first principle is that insurance policies create rights in the insured, and it is the insured’s right to exercise. Insurance policies are not issued in dependence on the existence of coverage in other years; no credit is given on the premium from having more coverage in the past year than not or coverage from carriers with better credit ratings. Policyholders don’t promise carriers to purchase insurance policies in the future. And insurers are not understood to be third-party beneficiaries of each other’s contract, even though those contracts are with the same party, the policyholder. See Signal Co., Inc. v. Harbor Ins. Co., 27 Cal. 3d 359, 369 (1980); L.E. Myers Co. v. Harbor Ins. Co., 394 N.E. 2d 1200 (Ill. 1979).
The second, and more important, principle is that insurance policies insure against the risk that the insured will be held liable for injury or damage during the policy period; policies do not insure against the risk of injury or damage occurring per se. As one leading treatise explains, “[t]he hazard insured against under the liability feature is not injury or loss . . . but liability or responsibility of the insured for loss or injury.” 6B J. Appleman & J.A. Appleman, Insurance Law and Practice §4254 at 26-27 (Rev. ed. 1979). The policy language makes this plain. CGL policies contain a broad promise to pay “all sums” that “the insured shall become legally obligated to pay as damages because of bodily injury . . . to which this insurance applied caused by an occurrence.” In other words, subject to the applicable limit of liability, the policy covers the totality of damages incurred by an insured “because of,” i.e. “by reason of” or “on account of,” bodily injury within the policy period. The question then is, what is the insured’s liability because of the injury during the policy period. This is key and shows both where Minnesota law goes wrong and creates the hydraulic pressure on the court to backfill we see so clearly in Wooddale.
These two principles illuminate the right way to argue the point for policyholders (and I submit the right way for courts to resolve the question). Instead of the Minnesota court’s mathematical formula, one takes a single policy and asks the following question: Is there bodily injury or property damage that occurred during the policy period, and if so what is the insured’s liability for the bodily injury or property damage that happened? (Contrary to what some insurers have argued, the “Policy Period; Territory” provision does not alter the question, because that provision ensures that the trigger is damage, not the negligent act, see State v. Glens Falls Ins. Co., 609 P.2d 598, 600-01 (Ariz. App. 1980), and while the language in the insuring agreement referring to property damage “to which this insurance applies” refers implicitly to the “policy period; territory” provision, that does nothing to the question on which I focus, viz., what is the insured’s liability “because of” the property damage to which the insurance applies.) I don’t need to reflexively or exclusively rely on the insurance policy’s promise to pay “all sums” or “those sums” – that is handy language but is not the key issue. The key is always and only, what is the insured’s liability because of bodily injury/property damage during the policy period.
As a matter of tort (not insurance) law, in most of the situations we deal with the answer to that question is this: for bodily injury or property damage that occurred during the particular carrier’s policy period, the insured’s liability is equal to the entirety of the plaintiff’s claim. This is the result of the rule of tort law that imposes joint and several liability upon tort defendants. John Crane v. Scribner, 800 A.2d 727, 741 (Md. 2002) (“it is as impossible to ascertain which fiber ultimately caused which cell, over time, to escape the body’s defenses and turn cancerous, as it is to determine when that occurred”).
Note that I am not saying that insurers are jointly and severally liable under their policies; that would be to import a tort concept to the contract context, and I don’t think that is quite precise enough. But what is central is that the insured is jointly and severally liable, and such liability is imposed where the injury or damage at issue is indivisible from injury or damage that occurred in other policy periods (or from other tortfeasors). See United States v. Alcan Aluminum Corp., 964 F.2d 252, 268-69 (3d Cir. 1992); Matter of Bell Petroleum Services, Inc., 3 F. 3d 889-896 (5th Cir. 1993); O’Neil v. Piccillo, 883 F.2d 176 (1st Cir. 1989) (interesting discussions all re indivisible damage/injury and the imposition of joint-and-several liability under tort law).
Nothing in the policies reduces the carrier’s obligation to pay simply because injury may also have occurred outside the policy period; the sole determinant of the extent of coverage is what is the insured’s liability because of injury during the policy period. The policy language addressing the application of the policy limits for a covered claim nowhere confines the carrier’s obligation to pay to some aliquot share based on the quantum of injury occurring in its policy years. E.g., ACandS v. Aetna Cas. & Sur. Co., 764 F.2d 968, 974 (3d Cir. 1985) (Even where sums paid by the insured party are partly attributable to injury that occurred in another policy period, “the language of the policy makes [that fact] irrelevant.”).
This is made particularly clear in the “Limits” section of standard liability insurance policies. The “Limits of Liability” section of policies typically provides that, for a covered bodily-injury (or property damage) “occurrence,” the policy will pay for “all damages . . . because of bodily injury sustained by one or more persons as the result of any one occurrence.” The sole limitation on each policy’s obligation to pay is that the payout on the claim “shall not exceed the limit of bodily injury liability stated in the declarations as applicable to ‘each occurrence.’” The Limits provision further elaborates by stating:
For the purpose of determining the limit of the company’s liability [under the policy], all bodily injury . . . arising out of continuous or repeated exposures to substantially the same general conditions shall be considered as arising out of one occurrence.
Here the policy language unambiguously favors coverage: The amounts any given CGL policy pays are expressed in dollars “per occurrence.” The insuring agreement and the limits-of-liability provisions together teach that, where there is covered injury in the policy period, the policy pays “all sums” for “all damages sustained by one . . . person . . . as the result of any one occurrence,” subject only to the per-occurrence limit.
In fact, even if the policies were uncertain as to the proper method of allocation, a reasonable interpretation of the policies requires an “all sums” approach. And of course insurers have been in the position to protect themselves by drafting clear and explicit language that addresses how the obligation to perform should be measured where there is indivisible damage triggering their coverage. Compare Rochester German Ins. Co. v. Schmidt, 175 F. 720, 725-726 (4th Cir. 1909). Principles of insurance law, therefore, require that the interpretation of uncertain or ambiguous policy provisions favoring the insured must govern.
This contract-based approach to allocation also “solves” the question of horizontal versus vertical exhaustion: that is, where policies are triggered across time, must the insured collect first from all the primaries before tapping any excess coverage (and thus absorb cumulated deductibles and insolvent primary layers) or may the insured select a single year of coverage and access the triggered primary and overlying excess (either in a single year or in more than one). General-liability policies do not speak to this issue, and so we return to the principle that the contract rights belong to the insured; consequently, the insured has the option to select or target its triggered policies however it sees fit. In each instance, one asks the same questions: is the policy triggered, and if so how much does the policy pay per occurrence for the insured’s liability because of the (indivisible) damage/injury that year? Any targeted policy may seek to pursue equitable contribution against other carriers that could have been targeted but were not. Indeed, I submit that part of the peace of mind offered by broad CGL insurance is precisely the insured’s ability to collect its full per-occurrence limits and then go home – leaving further redistribution to the carriers to sort out.
This approach also gives the framework to address stacking or cumulation of policy limits, for looking at each contract the stacking or cumulation of policy limits again is straightforward. See United Services Automobile Ass’n v. Riley (Md. June 1, 2006). Until the insured is fully indemnified for its damages because of injury/damage during the policy year, the insured is entitled to collect on its coverage. In other words, stacking is only an issue if an insurer has clear anti-stacking language in its policy.
The bottom line of all of this is that policyholders should look hard at the contract without preconceived notions and see how each individual insurance contract’s obligation to perform is measured. Bearing in mind the crucial question of divisible versus indivisible harm and damage – and its impact on answering what is the insured’s liability for injury or damage during the policy period – insurance policies require insurers to perform if any “injury” occurred during the policy period and to pay their entire per-occurrence policy limit until such time as the insured has been fully indemnified or the insurance policy’s limits exhausted.
We confront the particular language of particular policies every time we settle a claim or file a complaint. Each contract should be analyzed using the toolkit I’ve sketched out here, and then you should move on to the next contract. Shampoo. Rinse. Repeat.