As with all disputed coverage issues, the analysis begins with the language of the policy as interpreted by the applicable state law. A typical definition of occurrence under a CGL policy is: “an accident, including continuous or repeated exposure to substantially the same general harmful conditions.” Courts have taken three main approaches to determining the number of occurrences under this standard definition.
Under the so-called “effects” test, the number of occurrences is “construed from the point of view of the person whose property was injured.” Anchor Cas. Co. v. McCaleb, 178 F.2d 322, 324 (5th Cir. 1949) (holding that an exploding oil well that damaged multiple properties constituted a separate occurrence for each damaged property). Thus, “the policy coverage limits are based on the effect of the accident, extending the insured’s policy limits to each injured party.” American Family Mut. Ins. Co. v. Wilkins, 285 Kan. 1054, 1062 (Kan. 2008).
This approach has been criticized as “unpredictable and limitless, since any one event can cause many injuries to many people.” United Servs. Auto. Ass’n v. Neary, 307 P.3d 907, 915 (Alaska 2013). A single wildfire, for example, can spread over a vast region, injuring thousands of victims. Currently, Tennessee and Louisiana apply a version of the effects test, at least in certain contexts. See Lombard v. Sewerage and Water Bd. of New Orleans, 284 So. 2d 905, 915 (La. 1973) (“The word ‘occurrence’ as used in the policy must be construed from the point of view of the many persons whose property was damaged.”); but see Whetstone v. Dixon, 616 So. 2d 764, 773–74 (La. Ct. App. 1993) (finding only one occurrence where single automobile accident caused two different bodily injuries); see also Am. Mod. Select Ins. Co. v. Humphrey, 2012 WL 529576, at *4 (E.D. Tenn. Feb. 17, 2012) (“The parties agree that Tennessee follows the “effects” theory. This theory states, generally, that limits in a liability policy which limit the insurer's liability to a specified amount “per occurrence” or “per accident” refers to the effect of the occurrence or accident, thus making the entire policy limits available to each injured or damaged party.”).
The vast majority of courts determine the number of occurrences by considering the cause of the covered damages. See, e.g., Donegal Mut. Ins. Co. v. Baumhammers, 959 Pa. 147, 159 (Pa. 2007) (“The prevailing view looks to the ‘cause’ or ‘causes’ of the damage to determine the number of occurrences.”). Application of this “cause” test generally leads to a smaller number of occurrences than the effects test. However, courts differ considerably in their approach to identifying the relevant cause. Some cases “look to the direct, physical cause of the injuries as being the yardstick for measuring whether the claims had a common origin.” In the wildfire context, such a proximate-cause analysis might consider each point of ignition as a separate occurrence. An alternative approach—the “liability-events causal test”—focuses on more remote underlying causes. This approach looks to “the insured’s act or omission that allegedly constitutes a breach of a duty to the claimant rather than at the most ‘direct’ or ‘immediate’ causes of the injury that are outside the insured’s control.” Restatement of the Law of Liability Ins. § 38. This could have the effect of grouping together multiple fires into a single occurrence—if, for example, the plaintiff’s theory of liability is that separate fires resulted from the same alleged negligent conduct.
New York applies its own unique standard, known as the “unfortunate event” test, which purports to look not solely at the cause of the losses but also “the nature of the incident giving rise to damages.” In applying this test, the court of appeals has instructed New York courts to consider: “[1] whether there is a close temporal and spatial relationship between the incidents giving rise to injury or loss, and [2] whether the incidents can be viewed as part of the same causal continuum, without intervening agents or factors.” Appalachian Ins. Co. v. Gen. Elec. Co., 8 N.Y.3d 162, 171-72 (2007). Criticized as “unwieldy in practice and inconsistently applied,” the unfortunate event test has resulted in a relatively narrow interpretation of “occurrence.”
In the seminal case to apply this standard, Arthur A. Johnson Corp. v. Indemnity Ins. Co. of North Am., 7 N.Y.2d 222 (1959), the court of appeals held that the collapse of two independent walls of adjoining buildings constituted separate “accidents,” although damage at each building occurred only 50 minutes apart and resulted from the same rainstorm. While the buildings were adjacent, there was “no suggestion that the collapse of the first wall caused the failure of the second.” Nor did the fact that both collapses ultimately resulted from the same weather event render them a single occurrence. As the court of appeals explained in a subsequent case: “Common causation is pertinent once the incident—the fulcrum of our analysis—is identified, but the cause should not be conflated with the incident.” Appalachian Ins. Co. v. Gen. Elec. Co., 8 N.Y.3d 162, 172 (2007).
At the other end of the spectrum, in Hartford Acc. & Indem. Co. v. Wesolowski, 33 N.Y.2d 169 (1973), the court found a single occurrence in a three-car collision, in which “the insured’s automobile struck one oncoming vehicle, ricocheted off and struck a second.” Unlike the collapse of the two walls in Johnson, the two collisions in Wesolowski “occurred but an instant apart” and “[t]he continuum before the two impacts was unbroken, with no intervening agent or operative factor.”
The New York courts have yet to apply the unfortunate event test to a wildfire liability claim. However, one trial court decision applied the test to a first-party claim involving fire damage at a Manhattan condominium. In Humphries v. Metropolitan Prop. & Cas. Ins. Co., 2020 WL 3104370, at *11 (Sup. Ct. N.Y. Co. May 26, 2020), a fire broke out on the second floor of the condominium, and after having been initially extinguished, was “rekindled” the next day on the third floor. Acknowledging the absence of “cases directly on point,” the court determined that “the two fires are separate occurrences within the meaning of the policy” because “the second fire, although a rekindling, took place a day after the first”; and the fires did not occur in the same space, but on different floors of the building,” such that “there was not the sort of undisrupted continuum the Court of Appeals contemplated” under the “unfortunate event” test.
Some CGL policies include “grouping” provisions under which incidents that might constitute separate occurrences under the default rules are treated as a single occurrence. For example, in one standard formulation, the policy’s definition of occurrence states: “All exposure to substantially the same general harmful conditions shall be deemed to arise out of one Occurrence.”
Citing with approval an earlier decision of the Connecticut Supreme Court, the New York Court of Appeals characterized such provisions as “expanded definitions of occurrence . . ., indicating an intent that certain types of similar claims be combined.” Appalachian, 8 N.Y.3d at 175 (citing Metro. Life Ins. Co. v. Aetna Cas. & Sur. Co., 255 Conn. 295, 305 (2001) (holding that similar grouping language “serves to combine claims arising from exposure to asbestos at the same place at roughly the same time into one occurrence”)).
Other New York state and federal “Courts have interpreted identical or similar grouping provisions as combining into a single occurrence exposures emanating from the same location at a substantially similar time.” Mt. McKinley Ins. Co. v. Corning Inc., 96 A.D.3d 451 (1st Dep’t 2012) (holding that “any group of claims arising from exposure to an asbestos condition at a common location, at approximately the same time (for example, at the same steel mill or factory), may be found to have arisen from the same occurrence”); Ramirez v. Allstate Ins. Co., 26 A.D.3d 266 (1st Dep’t 2006) (claims for bodily injury to two infant plaintiffs as the result of exposure to lead paint in the same apartment at different times constituted a single “occurrence”); but see Bausch & Lomb, Inc. v. Lexington Ins. Co., 414 Fed. App’x 366, 369 (2d Cir. 2011) (declining to aggregate claims that “involve differing times, locations, and circumstances”).
What degree of proximity is required to group events together under such provisions? One New York appellate decision (later affirmed by the Court of Appeals) clarifies that to qualify as a single occurrence under a policy provision aggregating claims arising from “exposure to the same general conditions,” the events underlying the claims need only be “close enough that there were no intervening changes in the injury causing conditions.” Nesmith v. Allstate Ins. Co., 103 A.D.3d 190, 194 (4th Dep’t 2013), aff’d 24 N.Y.3d 520 (2014). Nesmith involved lead paint exposure in an apartment building. Although the injuries were separated in time by months, they were nevertheless treated as a single occurrence. As the court of appeals observed in its decision affirming the appellate division, the policy’s grouping language did not require exposure to “exactly the same conditions” but rather the same “general conditions,” and “to say that the ‘general conditions’ were not the same would deprive the word ‘general’ of all meaning.” Nesmith, 24 N.Y.3d at 525.
The number-of-occurrences issue has not resulted in many published decisions in the wildfire context. The Wisconsin Supreme Court addressed this question in Secura Insurance
v. Lyme St. Croix Forest Company, LLC, 384 Wis. 2d 282 (Wis. 2018), holding that a wildfire that burned for three days over 7,422 acres constituted a single occurrence under a standard CGL policy.
The fire at issue in Lyme St. Croix was caused by a logging company’s equipment. The company’s CGL policy had a $2 million general aggregate policy limit, and a per-occurrence sub-limit of $500,000 for damage from fire. The insurer (Secura) argued that the fire was a single occurrence and coverage was therefore limited to $500,000. The trial court disagreed, holding that “although there was one uninterrupted cause of the fire, each ‘seepage’ of fire onto another’s property constitutes a separate occurrence for purposes of the policy.” The intermediate appellate court affirmed, finding a new occurrence “each time the fire—fuel and expanded by the consumption of new materials—spread to a new piece of real property and caused damage.”
The Wisconsin Supreme Court reversed, holding that the fire was a single occurrence. The court explained that under Wisconsin’s cause test, “where a single, uninterrupted cause results in all of the injuries and damages, there is but on ‘accident’ or ‘occurrence.’” Further, “if a cause and result are so simultaneous or so closely linked in time and space as to be considered by the average person as one event, then only a single occurrence has taken place.”
Applying these principles, the court found that the lower courts erred “by focusing not on the cause of the damage, but on the effect on individual property owners,” which amounts to “an application of the effects test rejected by this court.” Further, a finding of multiple occurrences would have “arbitrary and unreasonable consequences,” since “under the court of appeals’ analysis, the fire could have burned exactly the same amount of land over exactly the same amount of time, but if all the land were owned by one person instead of several, the fire would constitute one occurrence.”
Finally, the supreme court distinguished an earlier precedent, Wilson Mut. Ins. Co. v. Falk, 360 Wis. 2d 67 (2014), in which manure spread as fertilizer on a farm seeped into the wells of other landowners. Although the decision in Falk also applied the cause test, the court found multiple occurrences, reasoning that “[b]ecause the occurrence under the policy is well contamination, not manure application, there was an occurrence each time manure seeped into a unique well.” The Wisconsin Supreme Court observed that unlike Falk, where the insured’s “manure seeped over the course of an unspecified period of time,” “a three-day fire in a discrete area caused by a single precipitating event would reasonably be considered by the average person as one event.”
As the Lyme St. Croix case demonstrates, determining the number of occurrences at issue in a wildfire claim can have a significant impact on the amount of coverage. The analysis is highly fact-intensive and subject to differing standards depending on the applicable state law. Utility companies and other businesses facing such exposure are well advised to consult with coverage counsel before a claim arises to review and assess their coverage program.
