Appalachian Insurance v. General Electric Co.

OPINION OF THE COURT Graffeo, J. In this declaratory judgment action, we must determine whether, for purposes of exceeding annual “per occurrence” primary insurance policy limits to access excess insurance proceeds, defendant General Electric Company (GE) can group together as a single occurrence numerous personal injury claims arising from the exposure of individuals to asbestos insulation in GE turbines at work sites across the country. We agree with the courts below that, under the terms of the GE primary insurance policies, the claims present multiple occurrences. In this case, GE seeks to obtain excess insurance coverage for asbestos-related personal injury claims brought by individuals who, between 1966 and 1986, were exposed to asbestos-containing insulation used in steam turbines manufactured by GE and installed at more than 22,000 sites throughout the United States. Although GE did not produce the asbestos-related products, for decades it designed, manufactured and, in some cases, installed custom turbines that were insulated with asbestos-containing products manufactured by others. In the typical personal injury case, a plaintiff sued GE on the theory that, with knowledge of the dangers of exposure to asbestos-containing products, it designed, manufactured, sold, installed and/or serviced turbines insulated with asbestos, without warning individu — 167 — als working near its turbines of those dangers. Of all the asbestos exposure claims filed against GE, 95% arose from GE’s turbine business. Ordinarily, GE is only one of many defendants sued in an asbestos exposure case, with the manufacturers of the asbestos-containing insulation products, installation contractors and others also joined on various theories. As a result, GE’s portion of a settlement or verdict in individual cases has been relatively small: as of 2002, over 400,000 asbestos-related claims had been filed against GE, with GE’s share of each judgment averaging $1,500. In the early 1990s, an escalation in the number of asbestos-related personal injury claims filed against GE led to this dispute between GE and its excess insurers over the treatment of asbestos-related personal injury claims under GE’s primary general liability insurance policies in effect between 1966 and 1986. From the 1950s to the 1990s, GE maintained general liability insurance with Electric Mutual Liability Insurance Company (EMLICO), an entity partially owned by GE and its employees. Annual premiums under the policies were calculated through a complicated formula that provided for retrospective payment by GE of a sum that was largely based on prior years’ losses. Thus, the premium structure functioned much like a self-insurance retention or a deductible, with GE reimbursing EMLICO for claims EMLICO paid on GE’s behalf. Before 1966, the EMLICO policies contained a $5 million per-occurrence liability limit and a $10 million aggregate liability limit. From 1966 through 1986, the EMLICO policies retained a $5 million per-occurrence limit (with the exception of the 1986 policy year, which had a $25 million per-occurrence limit) but did not incorporate an aggregate liability limit. Under the retrospective premium formula, GE was generally required to compensate EMLICO for the payout related to each occurrence under $5 million, regardless of the annual number or combined dollar value of the claims. During the 1966-1986 time frame, GE also maintained excess liability insurance that supplied additional layers of coverage beyond the EMLICO limits. Most of GE’s first and second layer general liability excess coverage policies were through Certain Underwriters at Lloyd’s, London or the London Market Insurers (London). The excess insurers involved in this litigation subscribed to manuscript policies — 168 — for the excess coverage over EMLICO and, in most cases, the London policies. In this 20-year period, none of these excess insurance policies covered any claim that was less than the $5 million per-occurrence coverage supplied by EMLICO. All of the excess insurance policies contained the following loss payable provision: “[l]iability under this policy with respect to any occurrence shall not attach unless and until the Insured, or the Insured’s underlying insurer, shall have paid the amount of the underlying limits on account of such occurrence.” As such, excess coverage for a claim was triggered only when the claim exceeded the $5 million per-occurrence limit in the underlying EMLICO primary general liability policy. The EMLICO policies define an occurrence as “an accident, event, happening or continuous or repeated exposure to conditions which unintentionally results in injury or damage during the policy period.” Until 1992, EMLICO treated each asbestos-related personal injury claim brought by an injured individual who alleged exposure during the relevant 20-year time frame as a separate “occurrence.” Because of the retrospective premium structure, the fact that none of the claims exceeded $5 million and the absence of aggregate liability limits, the full cost of each claim was ultimately borne by GE. The number of asbestos-related claims substantially increased in 1991, causing GE to object to EMLICO’s practice of treating each individual claim as a distinct “occurrence.” After negotiation, and without the participation of any of GE’s excess insurers, GE and EMLICO entered into a “Claims Handling Agreement” that required EMLICO to group asbestos-related personal injury claims by product type, meaning that all claims associated with a single product— here, turbines — would constitute a single occurrence with a $5 million annual coverage limit. The agreement therefore allowed GE to combine claims to reach the $5 million per-occurrence general liability policy ceiling, triggering access to GE’s excess insurance coverage. GE’s resolution of its dispute with EMLICO created controversy between GE and its excess insurers because, by reducing the number of occurrences under the primary EMLICO policies, the agreement increased the number of instances when the $5 million per-occurrence threshold would be exceeded, significantly expanding the exposure of GE’s excess carriers. This litigation was commenced in 1996 when Allstate Insurance — 169 — Company sued GE, EMLICO and numerous GE excess insurers, including Appalachian Insurance Company, seeking a declaration of the parties’ rights and responsibilities relating to GE asbestos claims. Although other concerns were raised, the main issue was the propriety of GE’s and EMLICO’s construction of the term “occurrence” in the EMLICO general liability policies. Appalachian answered the complaint, pursuing counterclaims against Allstate and cross claims against GE and EMLICO. Other excess insurers similarly joined the action. When Allstate settled with GE and EMLICO, Supreme Court substituted Appalachian for Allstate as the lead plaintiff. After extensive discovery, Appalachian and other excess insurers moved for summary judgment requesting a declaration that GE’s primary policies had not been exhausted by the payment of the asbestos claims because each claim by an injured plaintiff represented a separate occurrence for purposes of the EMLICO policies and none of the claims came close to reaching the $5 million per-occurrence limit. GE opposed the motion and cross-moved for summary judgment and a judicial declaration that the asbestos litigation arising from GE’s turbine business was a single occurrence under each policy because all the claims could be traced to a single act of negligence — GE’s failure to warn of the dangers of exposure to asbestos insulation in its turbines. Because the excess insurers were not parties to the “Claims Handling Agreement,” GE acknowledged that they were not bound by the retroactive construction of the term “occurrence” in that agreement. The issue therefore distilled to one of contract interpretation: how the term “occurrence” in the 1966-1986 EMLICO policies should be interpreted in the asbestos exposure context. Supreme Court granted the excess insurers’ motion and denied GE’s cross motion. Relying on our decision in Arthur A. Johnson Corp. v Indemnity Ins. Co. of N. Am. (7 NY2d 222 [1959]), the court held that the focus should be on the event for which the insured is being held liable, not a point further back in the causal chain. Applying the “unfortunate event” test to the asbestos exposure claims, the court noted that the turbines were custom-designed based on the specific needs of GE customers, with little or no uniformity in the amounts or types of asbestos insulation incorporated in the design. Moreover, the exposure of the individual plaintiffs varied in duration and intensity and occurred over decades at more than 22,000 work sites throughout the nation. In sum, there was insufficient — 170 — temporal and spatial proximity among the claims to unify them as one occurrence. The court concluded that the exposure of each plaintiff to asbestos-containing insulation was the unfortunate event for which GE was being held liable. It therefore declared that the excess insurers were not obligated to provide excess coverage for individual claims that did not exceed the annual $5 million per-occurrence limit in the EMLICO primary general liability policies. The Appellate Division affirmed, reasoning that the operative occurrence under the definition in the EMLICO policies was the “last link in the causal chain leading to liability, i.e., the exposure of each individual claimant to asbestos contained in the turbines manufactured by the insured, rather than earlier events creating the potential for future injury” (19 AD3d 198 [2005]). This Court granted GE leave to appeal.1 Under the excess insurance policies at issue here, GE is entitled to coverage only if the $5 million per-occurrence policy limit in the underlying EMLICO general liability policies is exceeded. The rights and liabilities of the excess insurers turn on the meaning of the term “occurrence” in the EMLICO policies. Our Court has previously addressed the standard to be applied in resolving whether a set of circumstances amounts to one accident or occurrence, or multiple accidents or occurrences, for purposes of resolving how much coverage is available under a third-party liability insurance policy. In Arthur A. Johnson Corp. v Indemnity Ins. Co. of N. Am. (7 NY2d at 227-229), we described the three approaches used by courts to resolve the question: the sole-proximate-cause approach, which focuses on whether the injuries or losses can be traced to a single, originating cause; the one-accident-per-person approach, which depends on the number of individual claimants seeking recovery; and the unfortunate-event approach, which is based not solely on — 171 — the cause but on the nature of the incident giving rise to damages. In Johnson, the issue was whether the collapse of two independent walls of adjoining buildings during an intense rainfall constituted one or two “accidents” for the purpose of determining whether the insured was entitled to coverage up to the “per-accident” liability limit for the breach of each wall. The Court adopted the unfortunate-event test and determined that there were two accidents, explaining: “[W]e need only point out that it is agreed that, during a heavy rainfall, a protecting wall collapsed under the water pressure and destruction poured into a building. Almost an hour later, another wall gave way and water flooded another building. There is no suggestion that the collapse of the first wall caused the failure of the second ... In addition, the catastrophe was not the rain — that, in itself, did not harm. It was the breach of the wall letting the rain water in. Furthermore, if the walls were located blocks away from each other on different job sites but subject to the same rainfall, no one could contest that there were two accidents” {id. at 230). In Hartford Acc. & Indent. Co. v Wesolowski (33 NY2d 169 [1973]), the Court found no significant difference between the meaning of “accident” in the Johnson policy and the meaning of “occurrence” in resolving the extent of coverage under a policy with a “per occurrence” limit. The issue in Wesolowski was described as follows: “Where the insured’s automobile struck one oncoming vehicle, ricocheted off and struck a second more than 100 feet away, was there more than one ‘occurrence’ within the meaning of the provision fixing limits of liability in an automobile insurance policy?” (Id. at 170.) Applying the Johnson unfortunate-event test, we held that the three-car collision amounted to a single occurrence because “the two collisions here occurred but an instant apart” and “[t]he continuum between the two impacts was unbroken, with no intervening agent or operative factor” (33 NY2d at 174). From our decisions in Johnson and Wesolowski several factors emerge as relevant to distinguishing injuries or losses that arise from a single occurrence as opposed to those that constitute multiple occurrences: whether there is a close temporal and spatial relationship between the incidents giving rise to injury or loss, and whether the incidents can be — 172 — viewed as part of the same causal continuum, without intervening agents or factors. Common causation is pertinent once the incident — the fulcrum of our analysis — is identified, but the cause should not be conflated with the incident.2 In Johnson, we concluded that there had been multiple accidents, noting that 50 minutes elapsed between the wall collapses and there was an insufficient causal nexus between the two incidents since there was “no suggestion that the collapse of the first wall caused the failure of the second” or that the walls shared similar deficiencies traced to a common cause as it was not even clear that the two walls had been built at the same time by the same workers (7 NY2d at 230). But in Wesolowski an analysis of all three factors persuaded us that there had been a single occurrence — the three cars collided seconds apart, in close spatial proximity and without disruption by intervening causes or other factors. The insurance policies we interpreted in Johnson and Wesolowski did not specifically define the terms “accident” and “occurrence” and parties to an insurance contract remain free to include language articulating a different approach. The question we face here is whether GE and EMLICO did so in the EMBICO policies. Those policies defined an occurrence as “an accident, event, happening or continuous or repeated exposure to conditions which unintentionally results in injury or damage during the policy period.” As GE points out, the policies under review were drafted after we decided Johnson, at a time when general liability policies were shifting from “accident” to “occurrence” based. “The insurance industry changed to occurrence-based coverage in 1966 to make clear that gradually occurring losses would be covered so long as they were not intentional” (Continental Cas. Co. v Rapid-American Corp., 80 NY2d 640, 648 [1993]). By encompassing “continuous or repeated exposure to conditions,” as well as — 173 — the sudden events traditionally associated with an accident, the occurrence model more readily embraced a wider range of liabilities — such as liability arising from asbestos exposure or contamination. But as is evident from our decision in Wesolowski, the transition from accident to occurrence coverage had no impact on the process of determining whether injuries or losses could be grouped for purposes of per-occurrence liability limits. With regard to this issue, the term occurrence is synonymous with accident unless the parties include language in the policy indicating otherwise. We see nothing in the definition of occurrence in the EMLICO policies that suggests that GE and EMLICO had any such intent. The Johnson unfortunate-event standard, set forth in 1959, predated the first of the series of policies in this case by more than five years. Certainly, these sophisticated parties could have chosen to define occurrence in a manner that grouped incidents based on the approaches rejected in Johnson (such as the sole-proximate-cause model or the single-occurrence-per-claimant model) or adopted yet another approach not envisioned by the Johnson court.3 But they did not do so. We therefore conclude that the unfortunate-event standard governs the outcome of this appeal. Applying that standard, the asbestos exposure claims GE seeks to join as one occurrence (per policy period) represent multiple occurrences. Using the language adopted by the parties in the EMLICO policies, the incident that gave rise to liability was each individual plaintiffs “continuous or repeated exposure” to asbestos. Before the exposures occurred, there was — 174 — only the potential that some unidentified claimant would someday be harmed by GE’s alleged failure to warn. To be sure, the exposure to asbestos was not sudden — it was gradual and, for many plaintiffs, continued over a number of years. But that does not make it any less the operative incident or occasion giving rise to liability in this factual context. Having determined that there were numerous exposure incidents, we must analyze the temporal and spatial relationships between the incidents and the extent to which they were part of an undisrupted continuum to determine whether they can, nonetheless, be viewed as a single unfortunate event — a single occurrence. Even if we were to assume that the continuum element was met here because the exposures share a common cause (as GE urges), the scenarios presented would fail the Johnson test because of the lack of any spatial or temporal relationship. On this record, it appears that the incidents share few, if any, commonalities, differing in terms of when and where exposure occurred, whether the exposure was prolonged and for how long, and whether one or more GE turbine sites was involved. Under the circumstances, there were unquestionably multiple occurrences and the excess insurers were entitled to a declaration to that effect.4 In reaching this conclusion we emphasize that the Johnson standard is not, as GE suggests, the equivalent of a one-occurrence-per-injured-party approach. Although there may be instances when application of the rule will result in each individual claimant representing a single occurrence, the standard can lead to the grouping of claims as occurred in Wesolowski. Nor does the rule necessarily bar excess coverage in multi-plaintiff mass tort contexts. Each mass tort scenario must, be examined separately under the Johnson rule. There will undoubtedly be unfortunate occasions, such as a series of explosions, the accidental release of a hazardous substance or some other calamity, that will result in numerous injuries or losses. In those situations, the Johnson test may well allow the grouping of some or all of the claims for purposes of satisfying the per-occurrence limit, thereby triggering excess coverage. — 175 — Accordingly, the order of the Appellate Division should be affirmed, with costs. Chief Judge Kaye and Judges Ciparick, Smith and Pigott concur; Judges Read and Jones taking no part. Order affirmed, with costs. Footnotes [1] Due to constitutional limitations on this Court’s jurisdiction over appeals from nonfinal orders, GE was granted leave to appeal only insofar as it sought relief against nine excess insurers for whom party-finality principles rendered the judgment final: Birmingham Fire Insurance Company of Pennsylvania, China America Insurance Company, Fremont Industrial Indemnity Company, Granite State Insurance Company, Insurance Company of the State of Pennsylvania, North River Insurance Company, Rampart Insurance Company, Riunione Adriatica DiSicurta and Starnet Insurance Company. GE was denied permission to appeal with respect to the claims involving Appalachian and 30 other excess insurers. As the lead plaintiff, Appalachian was permitted to brief and argue the appeal. [2] For example, with respect to the automobile accident at issue in Wesolowski, the cause of the accident (whether it was speeding, reckless driving or the like), though active at the moment of the collisions, is distinct from the incident itself, and it is the incident, not the cause, that is subjected to the Johnson standard (indeed, causation is a factor in the Johnson test). In this case, assuming the cause of plaintiffs’ injuries was a failure to warn as GE contends, that does not transform the cause into the event giving rise to liability for purposes of the per-occurrence limit. If we were to focus only on discerning the common originating cause of multiple events as GE urges, we would not be applying the unfortunate-event test but rather the sole-proximate-cause test that we explicitly rejected in Johnson. [3] There are many ways that parties to an insurance contract can provide for the grouping of claims. Some liability policies, including the excess insurance policies interpreted by the Connecticut Supreme Court in Metropolitan Life Ins. Co. v Aetna Cas. & Sur. Co. (255 Conn 295, 309, 765 A2d 891, 898 [2001]), contain expanded definitions of occurrence that, for example, allow “continuous or repeated exposure to substantially the same general conditions [to] be considered as arising out of one occurrence” (see also, Consolidated Edison Co. of N.Y. v Allstate Ins. Co., 98 NY2d 208 [2002]), indicating an intent that certain types of similar claims be combined. In Travelers Cas. & Sur. Co. v Certain Underwriters at Lloyd’s of London (96 NY2d 583 [2001]), we interpreted language in reinsurance treaties that allowed aggregation of claims in some circumstances, provided temporal and spatial relationships were present. What these cases demonstrate is that the parties to an insurance policy, like the parties to any contract, are free to determine the terms of their arrangement. If they intend to allow grouping of claims, they need only include language expressing that intent. There is no such language in the EMLICO policies. [4] We note that our conclusion is consistent with the results reached by the United States Court of Appeals for the Second Circuit when applying New York law to claims arising from asbestos exposure or contamination (see In re Prudential Lines Inc., 158 F3d 65 [2d Cir 1998]; Stonewall Ins. Co. v Asbestos Claims Mgt. Corp., 73 F3d 1178 [2d Cir 1995], mod on other grounds 85 F3d 49 [1996]).