Not Your Father s Allocations: Practical Solutions for Modern Long-Tail Insurance Claims 1

Transcription

1 Not Your Father s Allocations: Practical Solutions for Modern Long-Tail Insurance Claims 1 Ginamarie Alvino RiverStone Claims Management, LLC Chicago, IL Suzan F. Charlton Covington & Burling LLP Washington, DC Elizabeth A. Hanke KCIC, LLC Washington, DC Helen K. Michael Howrey LLP Washington, DC Julia Molander Sedgwick, Detert, Moran & Arnold, LLP San Francisco, CA INTRODUCTION You know the difference between all sums and pro rata. You ve argued number of occurrences, drop-down, and whether aggregate limits apply. It s time to crunch the numbers -- now what? After the legal briefs have been written and the motions have been argued, it takes simultaneous mastery of multiple issues to achieve real results in complex, long-tail coverage disputes. This paper, along with the accompanying presentation, will highlight the key legal and practical issues that policyholders, insurers and their lawyers must navigate in order to resolve long-tail claims among multiple layers and periods of insurance coverage. A hypothetical scenario is included in the discussion in order to illustrate concrete principles. 1 The authors views are their own and not that of their firms or their clients. Further, each individual author does not necessarily agree with everything in this paper, which is a joint project and which necessarily contains portions that were authored by other panelists. DC:

2 I. Hypothetical Situation Asbest-R-Us is a subcontractor that installed piping (manufactured by another company) in 10,000 houses, located in 100 subdivisions in 10 different states beginning in 1965 and ending in The subcontractor performed its work under 20 separate contracts with 10 separate contractors. 15 years later, it is discovered that the piping's insulation contains a harmful substance and must be removed at significant expense. $250M has been spent to date and another $300-$1.30B has been estimated to be necessary to address the property damage. Also years later, workers (not Asbest-R-Us employees) who were on the building sites during construction are claiming to have contracted various diseases from exposure to Asbest-R-Us pipe installation operations. 1,000 workers have filed suit so far, and as many as 2,500 more are expected to file. Claimed damages range from $5,000 to $10,000,000 per claim. The peak of the workers exposures occurs in 1971 with no new exposures occurring after The homeowners and workers' lawsuits have bankrupted the pipe manufacturer, so the plaintiffs are now going after the installer. II. Policy Assumptions The coverage includes: available coverage from 1968 to approximately 1986 some primary policies with first dollar coverage and some with self-insured retentions (SIRs) aggregates for products/completed operations, generally no property-damage aggregates one policy with a combined single limit (CSL) or over-all policy aggregate defense costs clearly paid outside of defense at the primary layer; umbrella layer includes policies having drop-down language that is mixed and unclear as to whether they would assume control and responsibility for the defense once their attachment points are reached; excess layers includes policies having are mixed terms as to whether defense is covered at all, paid within limits or in some cases arguable paid in excess of limits. absolute asbestos exclusions in later policies and a few asbestosis exclusions in the last year of coverage various layers of excess insurance (including carriers who issued both primary and excess); some gaps and underinsured periods some insolvents in both primary and excess layers multi-year policies throughout the program 2

3 total coverage limits that will not cover all the damages if each policy only pays once (so there can be a possibility of coming back to primaries for a second pass after pro rata allocation or horizontal exhaustion.. III. Analysis For each issue discussed herein, this paper will address the following considerations: 1. Provide an overview of the relevant issues; 2. Discuss applicable case law, from both sides of the issue, if applicable, highlighting interesting or idiosyncratic cases and/or limiting application of principles by jurisdiction as appropriate; and 3. Discuss practical solutions for the problems, e.g., how to value the issue monetarily, potential ways to compromise certain apportionments; how to address lump-sum or coverage-in-place settlements. A. What is the "Occurrence" and How Many Are There? 2 Summary of Issue How one defines the occurrence in a long-tail claim affects everything from which policies must pay to how many times a particular policy must may pay. Batch clauses and other aggregating clauses further delineate how much coverage may be available. In the above example, if the occurrence is the company s decision to use asbestos-containing materials in the products it installed, there could be just one occurrence. On the other hand, if the occurrence is each injury or incident of damage, then arguably there could be 13,500 separate policy triggers (10,000 damaged houses plus 3,500 injured workers) -- unless the policy also contains language that aggregates injuries or damages that arose out of the same lot of goods, or the same premises, or the same cause. Of course, that raises another layer of questions: what counts as the same lot, premises, or cause? Arguably, there may be some number of occurrences in between one and 3,500: perhaps hundreds (one per construction site), or 50 (one per subdivision), or some other number. Outcomes may vary further to the extent that policies contain either combined single limits or separate limits for bodily injury and property damage. The practical effect of the occurrence issue is that the more an occurrence or its cause can be consolidated, the more likely higher layers of coverage will be reached. Conversely, the more 2 This section is authored by Suzan F. Charlton, Special Counsel in the Washington, D.C. office of Covington & Burling LLP. The opinions expressed in this section are the author's alone and not those of her firm, her firm's clients, or any other author of this joint paper. The opinions expressed in the other sections of this joint paper should not be considered the personal views of this author, her firm, or her firm s clients. 3

4 occurrences there are and the less they can be consolidated, the more low-level insurers and policyholders with high SIRs will have to absorb. 3 Legal Discussion The number of occurrences issue arises primarily out of policy language that consolidates damages arising out of or caused by the same thing. But policy language alone rarely answers the question, how many are there? Case law analyzing the number of occurrences provides guidance, but is not consistent from state to state, even though most states have adopted the cause test for determining the number of occurrences. 1. Policy Language Typical occurrence definitions may be phrased as: an accident or event or continuous or repeated exposure to conditions which results in personal injury or property damage during the policy period or an event or series of events or continuous or repeated exposure to conditions which results in legal liability regardless of the number of persons... affected by such act or acts of omission. Under such definitions, occurrences are not limited to individual, one-time accidents, but may include circumstances that are continuous or repeated. Typical occurrence-aggregating clauses (also sometimes referred to as deemer clauses) are usually phrased in terms of the cause of the damage: All bodily injury and property damage resulting from any one accident or from continuous or repeated exposure to substantially the same conditions shall be deemed to result from one occurrence. A variation of this clause adds the phrase existing at or emanating from one premises location to modify the word conditions. Another variation, often found in policies providing coverage for products-completed operations hazards, is a batch clause that allows occurrences to be aggregated if they arise out of one lot of goods or products prepared or acquired by the Named Insured. 2. Case Law Despite the apparent variation in coverage cases deciding the number of occurrences, outcomes are not entirely unpredictable. To determine whether bodily injury or property damage is the result of one occurrence or multiple occurrences, the vast majority of courts look to the cause of the injury or damage. See, e.g., U.S. Gypsum Co. v. Admiral Ins. Co., 643 N.E.2d 1226, 1258 (Ill. App. Ct. 1994); Michigan Chem. Corp. v. Amer. Home Assur. Co., 728 F.2d 374, (6th Cir. 1984) and citations therein; Owens-Illinois, Inc. v. Aetna Cas. & Sur. Co., 597 F. Supp. 1515, 1525 (D.D.C. 1984); see generally Ostrager & Newman, 1 Handbook on Insurance Coverage Disputes 9.02 (12th Ed. 2004). Only a few states still adhere to the so-called effects test. See, e.g., Lombard v. Sewerage & Water Bd., 284 So.2d 905, (La. 1973); Kuhn s of Brownsville v. Bituminous Cas. Co., 270 S.W.2d 358, 360 (Tenn. 1954). Knowing that the test is cause-based, however, does little to help in determining the outcome. Indeed, it is far from clear, based on recent cases, that the cause test 3 This result assumes aggregate limits do not apply. If they do, then primary and low-level policies will be exhausted and upper layer excess policies will be reached sooner, even if there are multiple occurrences. See infra, section D, for further discussion of aggregate limits. 4

5 provides any greater basis for aggregating occurrences than did the now-disfavored effects test. This is because different courts apply the cause test differently. However, recent decisions reveal two basic variations of the cause test: (1) cause of loss, which focuses on the insured loss under the policy and thus an inquiry into the policyholder s actions and resulting liability; 4 and (2) cause of injury, in which the determining factor is the most immediate cause of the injuries, whether or not the insured s actions caused them. Focusing on the immediate cause of injury, without regard to the role of the insured, tends to result in multiple occurrences, often with every injury or claim be found to arise out of a separate occurrence. See, e.g., Koikos v. Travelers Ins. Co., 849 So.2d 263, (Fla. 2003) (multiple shootings were multiple occurrences); Flintkote Co. v. General Accid. Assur. Co. of Canada, No. C MHP, 2006 U.S. Dist. LEXIS 11, at *42 (N.D. Cal. Jan. 3, 2006) (each asbestos exposure that causes and immediately precedes an injury giving rise to liability under the policy was a separate occurrence); Uniroyal Inc. v. American Re-Insurance Co., No A TI (N.J. Super, App. Div. Sept. 13, 2005) ( Uniroyal ), reprinted in Mealey s Litig. Rep.- Ins. 1, 6 (Oct. 18, 2005) (each claimant s individual exposure... to the product that resulted in the injury was a separate occurrence); cf. London Market Insurers v. The Superior Court of Los Angeles County, Truck Ins. Exchange, No. B189000, 2007 Cal. App. LEXIS 21 (Cal. App. 2nd Dist. Jan. 9, 2008) ( Kaiser Cement ) ( occurrence means injurious exposure to asbestos but case remanded for determination of number of injuries arising out of one lot or each premises location ). On the other hand, when courts consider what the policyholder did to cause the injury -- and thus the insured liability -- occurrences tend to be aggregated more often. See, e.g., Washoe County v. Transcontinental Ins. Co., 878 P.2d 306, 308 (Nev. 1994) (county s negligent licensing of day care center where sexual abuse occurred over 3-year period was one occurrence); Owens-Illinois, Inc. v. United Ins. Co., 625 A.2d 1, 23 (N.J. Super Ct. App. Div. 1993), rev d on other grounds, 650 A.2d 974 (N.J. 1994) (manufacture and sale of asbestos-containing product was one occurrence); Donegal Mut. Ins. Co. v. Baumhammers, 2007 WL at *7 (Pa. Dec. 27, 2007) (parents negligent supervision of child who shot multiple victims was one occurrence); Owners Ins. Co. v. Salmonsen, 622 S.E.2d 525, 526 (S.C. 2005) (distribution of defective stucco product was one occurrence); Liberty Mut. Ins. Co. v. Treesdale Inc., 418 F.3d 330, 334 (3d Cir. 2005) (manufacture and sale of asbestos-containing product was one occurrence); Sunoco Inc. v. Illinois Nat l Ins. Co., c.a. No , 2005 U.S. Dist LEXIS 18407, at *22-23 (E.D. Pa. Aug. 29, 2005) (manufacture and sale of MtBE was one occurrence); The Pyrites Co., Inc. v. Century Indemnity Co., 2005 Phila Ct. Com. Pl. LEXIS 393 at *5 4 New York applies a unique variation on the cause test, known as the unfortunate event test. See Arthur A. Johnson Corp. v. Indemnity Ins. Co. of Am., 164 N.E.2d 704, 707 (N.Y. 1959); Appalachian Ins. Co. v. General Electric Co., 863 N.E.2d. 994 (N.Y. 2007) ( GE ). This test essentially starts with an inquiry into the policyholder s liability-inducing acts or omissions, but, at least after GE, it also considers whether the act and injury occur along the same causal continuum, without intervening agents or factors. Because of the complicated way in which the unfortunate event test has been applied by courts in New York and the Second Circuit, a thorough discussion of this particular test is beyond the scope of this paper. 5

6 (Pa. Ct. Com. Pl. Aug. 30, 2005) (environmental contamination at connected sites was all one occurrence). The exception to the aggregation of occurrences under the latter cause test arises where the policyholder acted separately and distinctly in multiple, separate instances. See, e.g., Nicor v. Associated Elec. & Gas Ins. Serv s Ltd., 860 N.E.2d 280, 295 (Ill. 2006) (each gas metering site where there were mercury spills with no common cause was a separate occurrence); Strycker Corp. v. National Union Fire Ins. Co. of Pittsburgh, PA, No. 4:01-CV-157 (W.D. Mich. Aug. 27, 2004) (each sale and implantation of artificial knee was separate occurrence). Practical Solutions Notwithstanding the legal positions advanced by variously situated policyholders and insurance companies, parties should take into account certain practical considerations when determining the effect of the identity and number of occurrences on allocation of losses. First, examine the complete coverage program and the nature of the claims to determine whether an argument about the number of occurrences is even necessary. For example, if aggregate limits clearly apply, the number of occurrences may be irrelevant, because underlying layers of coverage may be exhausted in any event. Second, run allocations in a variety of ways to determine the real impact on all parties. Depending on the coverage program, there may be a breaking point at which a particular number of occurrences has a dramatic impact (e.g., any number of occurrences higher than x will never get out of the primary layer, or any number of occurrences lower than y will hit all insurers in the first three layers, etc.). Third, if a multiple occurrence scenario creates a black hole of coverage for primary insurers but is a realistic risk, parties interested in settlement might consider an approach that simulates aggregate limits equal to some multiple of the primary coverage, with excess insurers participation thereafter, and a preservation of the right to come back to the primary insurers if all coverage is eventually exhausted (see, e.g., Wellington Agreement ( Agreement Concerning Asbestos-Related Claims, signed June 19, 1985). To be attractive to the policyholder, such an arrangement would require agreement by excess insurers as well. To be acceptable to excess insurers, there must also be some risk of a single-occurrence outcome. B. What Trigger of Coverage Theory Applies? 5 Summary of Issue The trigger of coverage inquiry turns upon what must happen during the policy period under the specific policy language in order for the potential for coverage to arise. This issue focuses on the point in time at which an occurrence, as defined by the policy is determined to have happened, and resolution of this timing question serves to establish which policies may have to respond to particular 5 This section is authored by Helen K. Michael, a partner in the Washington, D.C. office of Howrey LLP. The opinions expressed in this section are the author s alone and not those of her firm, her firm s clients, or any other author of this joint paper. The opinions expressed in the other sections of this joint paper should not be considered the personal views of this author, her firm, or her firm s clients. 6

7 claims. Resolution of this timing question is frequently complicated when, as is typical for asbestos and other long-tail injury claims, the negligent act or the conduct complained of is separated in time from numerous resulting injuries, which may be deemed to have occurred over a substantial period of time. The general rule accepted in most jurisdictions is that coverage is potentially triggered not when the wrongful act was committed, but when the damage results during the policy period. The possibility that multiple occurrences may be involved can result in further complications. In the Asbest-R-Us hypothetical, the claims for property damage to the houses in the 100 subdivisions and the numerous bodily injury claims of the construction workers exposed to asbestos-containing material potentially implicate different successive policy periods. The houses were built between 1965 and 1981, and the damage or injury to some of these houses continues to this day. For the construction workers alleging bodily injury claims, the dates of employment will likely serve to demarcate the periods in which the workers were exposed to the asbestos-containing materials, and injury may be deemed to have occurred at any point from exposure through diagnosis of asbestos disease. The trigger theory that is applied can have a significant impact on the policyholder s total recovery because the application of a single point trigger of coverage theory, such as the manifestation trigger of coverage theory, which predicates trigger on the date of discovery of the injury, will put far fewer policies in play than application of the more widely followed continuous trigger theory. Legal Discussion 1. Policy Language The timing question upon which trigger of coverage turns is addressed in different clauses of the policy, depending upon the policy form. Before 1973, the timing requirement was included in the standard ISO definition of occurrence, which was defined as an accident, including injurious exposure to conditions, which results, during the policy period, in bodily injury or property damage neither expected nor intended from the standpoint of the insured. In the 1973 ISO form this timing requirement was addressed in both the occurrence definition defining an occurrence to mean an an accident, including continuous or repeated exposure to conditions, which results in bodily injury or property damage neither expected nor intended from the standpoint of the insured, and the definitions of bodily injury and property damage, which required the specified injury to take place in the policy period. In some policy forms, an additional condition related to the happening of specified conditions within the policy territory and policy period may be included. 2. Case Law The various trigger of coverage theories that may be applied in toxic tort cases have been addressed in scores of decisions, with the same theories ultimately being applied to resolve the timing of the occurrence issue for both bodily injury and property damage claims. A brief summary of each of these theories follows. 7

8 (a) Exposure Trigger Theory The exposure trigger activates coverage under every policy in effect during the periods in which the claimant was exposed to the toxic substance. This theory relies on medical evidence showing that tissue is injured with every inhalation of asbestos, or exposure to other toxic substances. The seminal case adopting the exposure rule is INA v. Forty-Eight Insulations, 633 F.2d 1212 (6th Cir. 1980), aff d on reh g, 657 F. 2d 814 (6th Cir. 1981). In the asbestos context, many courts continue to use the exposure trigger to determine the policy period(s) during which asbestos bodily injury will be found to have occurred. 6 (b) Manifestation Trigger Theory This theory posits that the occurrence transpires when the injury is detected or manifests itself. The First Circuit s decision in Eagle-Picher Industries, Inc. v. Liberty Mutual Insurance Co., 682 F.2d 12 (1st Cir. 1982), is one of the few approving the time of manifestation as a single point trigger in the context of an asbestos claim. The manifestation trigger has, however, has been applied in coverage litigation for property damage claims due to mold and defective construction. See, e.g., Bayou Bend Homes, Inc. v. Scottsdale Ins. Co., No. H , 2006 WL , at *7 (S.D. Tex. July 18, 2006); Audubon Trace Condominium Ass'n, Inc. v. Brignac-Derbes, Inc., 924 So.2d 1131, 1132 (5th Cir. 2006); DeLuca Enterprises v. Assurance Co., 911 A.2d 190 (Pa. Super. 2006). (c) Injury-in-Fact Trigger Theory This theory, which is the brainchild of Judge Sofaer, requires proof of the point or points in time at which each individual claimant actually suffered injury. See American Home Products Corp. v. Liberty Mutual Insurance Co., 565 F. Supp (S.D.N.Y. 1983), aff d, 748 F. 2d 760 (2d Cir. 1984). Although less commonly applied than the continuous trigger methodologies, the injury-infact trigger continues to be utilized by some courts. See, e.g., Western Nat. Mut. Ins. Co. v. Barbes, No. A , 2006 WL , at *3 (Minn. Ct. App. June 20, 2006). (d) Multiple or Continuous Trigger Theory The multiple trigger or continuous trigger theory has emerged in recent years as the most widelyaccepted trigger theory in toxic tort and defective construction cases. 7 This theory holds that every 6 See, e.g., American Guarantee & Liability Ins. Co. v. Anco Insulation, Inc., No. Civ. A A1, 2005 WL , at *4-*5 (M.D. La., July 29, 2005) ( Louisiana law has expressly rejected the application of the triple-trigger theory and consistently applied the exposure theory when determining the date of accrual of a claim resulting from long-term exposure to latent, disease-producing substances such as asbestos ); Shook & Fletcher Asbestos Settlement Trust v. Safety Nat. Cas. Corp., 909 A.2d 125, 128 (Del. 2006) (applying exposure trigger); A.W. Chesterton Co. v. Massachusetts Insurers Insolvency Fund, 838 N.E.2d 1237 (Mass. 2005) (applying exposure trigger and rejecting refusing to apply a continuous trigger of coverage); Cole v. Celotex Corp., 599 So.2d 1058 (1992)(Sup. Ct. of Louisiana adopted an exposure trigger in the context of bodily injury asbestos claims). 7 See Standard Fire Ins. Co. v. Spectrum Community Ass n., 46 Cal. Rptr. 3d 804 (Cal. App. 2006) (applying continuous trigger to defective construction litigation); Essex Ins. Co. v. Massey Land & Timber, LLC, No. 5:04 CV 102 DCB JCS, 2006 WL (S.D. Miss., May 24, 2006) (same); Plum v. W. Am. Ins. Co., No. C , 2006 WL (Ohio Ct. App., February 3, 2006) (same); Maryland Cas. Co. v. Hanson, 902 A.2d 152 (Md. App. 2006) (continuous injury is the appropriate trigger of coverage in lead-poisoning cases); Riley v. United Services Automobile Ass'n, 871 A.2d 599 (Md. Ct. Spec. App. 2005) (same); Moore v. Aetna, No. A T2F, 2003 WL , at *3 (N.J. Super. App. Div. Jan. 14, 2003) (same). 8

9 insurer on the risk from the beginning of the exposure period and continuing through the progression of the injurious process, until manifestation, is triggered. This theory has received widespread acceptance since Keene v. INA, 667 F.2d 1034 (D.C. Cir. 1981). Practical Solutions In cases involving asbestos bodily injury claims, such as those against Asbest-R-Us in our hypothetical, some variant of a continuous trigger of coverage or an injury-in-fact theory most likely would be applied. Assuming that all of the property damage and bodily injury claims are deemed to arise from a single occurrence, determining what group of policies is triggered with respect to any given claim should be relatively straightforward. Coverage for the property damage claims will be triggered under all policies in effect at the time the Asbest-R-Us piping was installed in each particular house until the injury or damage ends or the claim is paid. Completion and other building certificates should provide information regarding when the Asbest-R-Us piping was installed. Coverage for the bodily injury claims, in turn, will be triggered under all policies in effect from the time an individual construction worker was exposed to the Asbest-R-Us piping through diagnosis of asbestos disease. If a pro rata allocation approach is to be applied, the possibility that losses need not be spread to periods in which insurance was not available should be considered. For settlement and coverage valuation purposes, it also may be necessary to consider alternative scenarios involving application of a continuous trigger theory to possible multiple occurrence scenarios. Among one of a number of possible scenarios presented by the Asbest-R-Us hypothetical is that the property damage claims may all represent a single occurrence, while each bodily injury claim represents a separate occurrence. Assessment of which policies may be triggered under such alternative scenarios likely will prove to be complex because, among other things, policies in particular policy periods and layers are likely to exhaust unevenly. 8 In the event that an injury-in-fact trigger may apply, the policyholder may also need to develop both factual evidence and expert testimony to establish when a particular claimant suffered bodily injury or property damage occurred to a particular structure. As noted above, some courts have shifted the burden to insurers contesting coverage to prove when a particular injury occurred, but others have declined to relieve the policyholder of its usual burden of proving coverage. In jurisdictions in which policyholders bear the burden of proving when injury occurred, insurers typically will challenge both the sufficiency of the policyholder s evidence regarding commencement of the injury and the policyholder s ability to establish whether injury occurred in any particular policy period. In the context of asbestos bodily injury claims, for example, there are studies indicating that the onset of asbestos-related disease does not commence until several years after first exposure. There also is evidence that the pathogenic process is dose dependent and turns on the amount of exposure and other 8 Under the approach adopted in In Re Wallace & Gale Co., 385 F.3d 820, 834 (4th Cir. 2004) ( Wallace & Gale ), an added complexity may arise for policyholders that are contractors seeking coverage under policies affording completed operations coverage. As discussed in detail in Section D below, the Wallace & Gale approach fragments an injury continuing over multiple successive policies into separate injuries based on when the policyholder ceased its contracting operations. As also discussed in further detail below, this approach effectively requires application of an exposure trigger of coverage in cases in which a continuous trigger ordinarily would apply, while precluding the policyholder from obtaining coverage from the policies in effect at the time of the first exposures, before the contracting operations were completed. 9

10 factors existing during particular time periods. To the extent that evidence can be developed that the injury at issue commenced at the time of exposure and continued thereafter, however, application of an injury-in-fact trigger of coverage should have essentially the same impact of maximizing the number of policies triggered as would a continuous trigger of coverage theory. C. What Allocation Theory Applies? 9 Summary of Issue The allocation methodology that is applied in cases involving continuous or progressive injuries spanning successive policy periods, such as the claims for asbestos bodily injury and property damage involved in the hypothetical presented at the beginning of this paper, can have a profound impact on the policyholder s total recovery and the amount that particular insurers may be obligated to pay in particular cases. Application of the so-called all sums approach, which also has been described as joint and several, pick and choose, or vertical allocation approach will favor the policyholder, while the competing pro rata approach typically will favor the insurers. There also are variants of both of these theories. Some that may serve materially to increase the policyholders recovery in certain cases, while others may be more acceptable to both sides for purposes of reaching a lump-sum or coverage-in-place settlement. Before any allocation methodology is applied, an analysis must be undertaken to determine the time period over which various losses are to be spread. This analysis must include identifying 1) the total time period implicated by given claims, including the periods before coverage may have been acquired and after coverage may no longer have been available for the particular risk, 2) under-insured periods, and 3) gaps in coverage created by insurer insolvencies, less-than-policy-limits settlements, and policy provisions. Legal Discussion 1. Policy Language Courts employing the all sums allocation approach and those employing the pro rata allocation approach both have relied on the same principal terms in standard-form CGL policies as justifications for the allocation methodology selected, but have reached diametrically opposed conclusions about the effect of this policy language. These key provisions include: The standard form insuring agreement obligating the insurer to pay all sums which the insured shall become legally obligated to pay as damages because of bodily injury or property damage to which th[e] insurance applies, caused by an occurrence... The standard-form provision defining an occurrence to mean an accident, including injurious exposure to conditions, which results, during the policy period, in bodily injury or property damage... 9 This section is authored by Helen K. Michael, a partner in the Washington, D.C. office of Howrey LLP. The opinions expressed in this section are the author's alone and not those of her firm, her firm's clients, or any other author of this joint paper. The opinions expressed in the other sections of this joint paper should not be considered the personal views of this author, her firm, or her firm s clients. 10

11 2. Case Law Not surprisingly, there are conflicting decisions regarding appropriate allocation methodology. (a) The All Sums Approach General Principles The all sums approach is one of the two principal allocation approaches. 10 The basic rationale for the all sums approach is that the language in each policy, often purchased by the policyholder for substantial premiums over the years, expressly obligates the insurer to pay all sums incurred by the policyholder. Courts adopting this approach have interpreted this same language to mean that, when multiple policies are triggered to cover the same loss, each policy provides indemnity for the insured s entire liability, and each insurer is jointly and severally liable for the entire claim. Rubenstein, 694 N.E.2d at 388. Courts adopting this approach also have emphasized the absence of policy language expressly calling for any proportional allocation once the policy is triggered. As the U.S. Court of Appeals for the District of Columbia explained in its seminal decision first recognizing the all sums approach: [E]ach policy has a built-in trigger of coverage. Once triggered, each policy covers [the policyholder s] liability. There is nothing in the policies that provides for a reduction of the insurer s liability if an injury occurs only in part during a policy period. As we interpret the policies, they cover [the policyholder s] entire liability once they are triggered. That interpretation is based on the terms of the policies themselves. Keene, 667 F.2d at (emphasis added). See, e.g., Armstrong World Indus. v. Aetna Cas. & Sur. Co., 45 Cal. App. 4th 1, 105 (1995) (same). Accord Allstate, 759 N.E. 2d at ( there is no language in the coverage grant... that limits Allstate s responsibility to indemnification for liability derived solely for that portion of damages taking place within the policy period. ); B&L Trucking, 951 P. 2d at 256 ( If the insurer wished to limit its liability through a pro rata allocation of damages once a policy is triggered, the insurer could have included that language in the policy. ). A number of courts adopting the all sums approach also have reasoned that the definition of occurrence in CGL policies also shows that each insurer assumes a separate and independent obligation to pay all resulting damages up to policy limits: 10 See, e.g., Keene Corp. v. Ins. Co. of N. Am., 667 F.2d 1034 (D.C. Cir. 1981), cert. denied, 455 U.S (1982), reh g denied, 456 U.S. 951 (1982) (multiple states); Aerojet-General Corp. v. Transp. Indem. Co., 948 P.2d 909, n.10 (Cal. 1997); Armstrong World Ind. v. Aetna Cas. & Sur. Co., 52 Cal. Rptr.2d 690, , (Cal. Ct. App. 1996); Hercules, Inc. v. AIU Ins. Co., 784 A.2d 481, 492 (Del. 2001); Caterpillar, Inc. v. Century Indem. Co., 2007 Ill. App. 1420, at *25 (Ill. App. Feb. 2, 2007); Allstate Ins. Co. v. Dana Corp., 759 N.E.2d 1049, (Ind. 2001); Monsanto Co. v. C.E. Heath Comp. & Liab. Ins. Co., 652 A.2d 30, 35 (Del. 1994) (applying Missouri law); Rubenstein v. Royal Ins. Co., 694 N.E.2d 381, 388 (Mass. App. Ct.), aff d in part on other grounds, 707 N.E.2d 367 (Mass. 1998); Goodyear Tire & Rubber Co. v. Aetna Cas. & Sur. Co., 769 N.E.2d 835, 841 (Ohio 2002); J.H. France Refractories Co. v. Allstate Ins. Co., 626 A.2d 502, (Pa. 1993); Tex. Prop. & Cas. Ins. Guar. Ass n v. Southwest Aggregates, Inc., 982 S.W.2d 600, 607 (Tex. App. 1998); Am. Nat l Fire Ins. Co. v. B&L Trucking & Const. Co., 951 P.2d 250, 251 (Wash. 1998). But see Boston Gas Co. v. Century Indemn. Co., 529 F.3d 8, 13 (3d Cir. 2008) (certifying question at appropriate allocation methodology to the issue to the Massachusetts Supreme Court). 11

12 An occurrence includes continuous or repeated exposure to conditions which result in bodily injury. The insurers which drafted the definition obviously contemplated the possibility of injury resulting from continuous or repeated exposure to conditions, and specified that the process of exposure was to constitute one occurrence. If prolonged exposure, constituting one occurrence, resulted in injury, and if the injury occurred during the time a given policy was in effect, then the injury is an insurable risk under the terms of that policy. Being defined as one occurrence, the entire injury, and all damages resulting therefrom, fall within the indemnification obligation of the insurer. In other words, once the liability of a given insurer is triggered, it is irrelevant that additional exposure or injury occurred at times other than when the insurer was on the risk. The insurer in question must bear potential liability for the entire claim. J.H. France Refractories Co. v. Allstate Ins. Co., 626 A.2d 502, 508 (Pa. 1993). See Aetna Cas. & Sur. Co. v. Pintlar Corp., 948 F.2d 1507, 1516 (9th Cir. 1991) ( coverage is limited to damages occurring during the policy period; it also includes damages occurring after the period, as long as those damages are caused by an occurrence that is a release which results in injury during the policy ) (applying Idaho law). Other courts have concluded that the standard CGL policy language is at least ambiguous with respect to how losses should be allocated once a standard-from CGL policy is triggered. Accordingly, such courts have applied the rule of contra proferentem, which requires conflicting policy interpretations to be construed against the insurer and in favor of coverage. See, e.g., B&L Trucking, 951 P.2d at 256 (citations and footnote omitted); Hercules, 784 A.2d at & 492 n.30. Alternative Approach: Hopscotching Around the Coverage Chart Although it has yet to be considered in any reported appellate decision, at least two trial court decisions in California have approved a variant of the all sums allocation methodology based on standard-form language frequently contained in CGL policies. This methodology enables policyholders confronting mass tort and other long-tail liabilities to avoid having to absorb gaps in coverage created by insolvencies or less-than-limits settlements as a precondition for accessing the overlying excess coverage. See International Paper Co., et al. v. Affiliated FM Ins. Co., et al., No , Ruling re Issue H Allocation of Insolvent and Less Than Policy Limit Settlements ( International Paper ), at 1 (Cal. Super. Ct., San Francisco, Mar. 17, 2005); Kaiser Aluminum & Chemical Corp. v. Certain Underwriters at Lloyd, No , Decision on Group IIA Trial Issues ( Kaiser I ), at 9 (Cal. Super. Ct., San Francisco, June 13, 2003); see Kaiser Aluminum & Chemical Corp. v. Certain Underwriters at Lloyd, No , Decision on Group IIB Trial Issues at 5 ( Kaiser II ) (Cal. Super. Ct., San Francisco, Feb. 20, 2004). This methodology, which has been described as hopscotching around the coverage chart (International Paper at 1; Kaiser I at 5), permits a policyholder in specified circumstances to move outside a vertical line of underlying insurance in one policy year and to tap into a horizontally located policy in a different policy period in order to fill gaps in coverage. 12

13 Building on the all sums approach, this methodology also relies on the standard-form attachment point provisions typically found in CGL policies, which obligate carriers to provide coverage once a specified amount of underlying liability has been incurred. The methodology follows the precedent reading such standard-form attachment point provisions 11 as requiring a policy to pay so long as the policyholder s liability has exceeded a specified amount and regardless of whether the underlying policy actually has paid its specified limit. See, e.g., Rummel v. Lexington Ins. Co., 945 P.2d 970, 981 (N.M. 1997) ( the excess insurer has no rational interest in whether the [underlying] policies are collected in full, as long as it is only required to pay the loss for which it would otherwise have been liable under the terms of its contract ); Denny s, Inc. v. Chicago Ins. Co., 234 Cal.App.3d 1786, 1794 (1991) ( the insured s liability [must] exceed the underlying insurance limits before any liability attaches to the excess insurer, whether or not the underlying insurer has actually paid out the policy limit ); Gasquet v. Commercial Union Ins. Co., 391 So.2d 466, 471 (La. 1980) (settlement with underlying insurance company for less than policy limits does not relieve excess insurance company from its obligations). But see Qualcomm, Inc. v. Certain Underwriters at Lloyd's, London, 161 Cal. App. 4th 184, (Cal. App. 4th Dist. 2008) (less-than-policy-limits settlement relieved excess carrier of any coverage obligations because the policy at issue prescribed that no liability would attach until the underlying insurer had paid or been held liable to pay its full policy limit) 12 Applying the all sums and attachment point provisions together, the methodology prescribes that each triggered policy is responsible for its appropriate level of all sums liability even when [the policy] is filling an insolvency [or settlement] gap from a different position in the coverage chart, and that the excess policy overlying the gap in coverage must provide indemnification once the policyholder has paid the specified amount of underlying liability. Kaiser I at 10. It is irrelevant under this methodology whether the underlying policy actually pays its specified limit of liability. Once the policyholder has paid an amount exceeding the specified attachment point of the overlying excess policy, this excess policy may be chosen to provide coverage because the amount of liability incurred by the policyholder, and not whether the underlying insurer actually has provided indemnification for that liability, controls the point at which each policy must assume its coverage obligations. (b) The Pro Rata Approach The pro rata allocation approach is the second principal allocation approach. 13 This approach divides the policyholder s loss proportionally among all liable insurance companies based on a portion 11 Typical attachment point provisions include those providing that liability shall attach to the Company only after the underlying Umbrella Insurers have paid or have been held liable to pay the full amount of their respective underlying liability, or that [l]iability of the Company under this policy shall not attach unless and until the Insured or the Insured s Underlying Insurance has paid or has been held liable to pay the total applicable underlying limits. Others may contain terms stating that the excess policy attaches after all underlying insurance has been exhausted by the payment of the limits of such insurance. 12 The Qualcomm court s reading of this attachment point language is inconsistent with the Denny s court s interpretation, and also conflicts other decisions declining to read such attachment point language as requiring full payment of the specified policy limits by the underlying insurers. See, e.g., Rummel, 945 P.2d at 981 (interpreting functionally identical policy language). It remains to be seen whether the Qualcomm decision will spark more conflicting decisions. This is an issue that should be monitored by both policyholders and insurers counsel. 13 See, e.g., Jones v. Liberty Mut. Ins. Co. (In re Wallace & Gale Co.), 385 F.3d 820, 835 (4th Cir. 2004) (applying Maryland law, time-on-the-risk approach), reh g denied and corrected by 2004 U.S. App. LEXIS (4th Cir. Nov. 15, 13

14 of the loss that occurred during the time each insurer s policy was in place. Courts adopting this approach often focus on policy provisions limiting coverage to occurrences taking place during the policy period and find that the all sums language in the policy is not to the contrary. See, e.g., Public Serv. Co., 986 P.2d at 939. There are two principle variations on this approach, the time-on-the risk approach and percentage-ofthe-limits approach. Insurers prefer the pro rata approach because it often reduces their overall liability by requiring policyholders to contribute for periods of missing coverage in which they did not purchase coverage, or have lost their policies or have other gaps in coverage. This approach also places the burden of transaction costs principally on policyholders because they must pursue each individual insurer on the risk for coverage. Time-on-the-Risk Approach Most courts adopting the pro rata approach allocate the loss among liable all triggered periods based on the proportional time that each insurer covered the risk. Under this approach, the court should make a reasonable estimate of the portion of the occurrence that is fairly attributable to each year by dividing the total amount of liability by the number of years at issue, and should then allocate liability accordingly to each policy-year, taking into account primary and excess coverage, SIRs, policy limits, and other insurance on the risk.... Public Service, 986 P.2d at 940. Courts adopting this approach often focus on policy provisions limiting coverage to occurrences taking place during the policy period and find that the all sums language in the policy is not to the contrary. See, e.g., id.; Porter, 641 F.2d at 1145; Public Serv., 986 P.2d at 939. The New York Court of Appeals reasoning is illustrative: Although more than one policy may be implicated by a gradual harm, joint and several allocation is not consistent with the language of the policies providing indemnification for all sums of liability that resulted from an accident or 2004); Stonewall Ins. Co. v. Asbestos Claims Management Corp., 73 F.3d 1178, (2d Cir. 1995) (applying New York and Texas law, time-on-the-risk approach); Ins. Co. of N. Am. v. Forty-Eight Insulations, Inc., 633 F.2d 1212, (6th Cir. 1980), clarified by 657 F.2d 814 (6th Cir. 1981), cert. denied 455 U.S (1982) (applying Illinois and New Jersey law, time-on-the-risk); Commercial Union Ins. Co. v. Sepco Corp., 918 F.2d 920, (11th Cir. 1990) (holding Alabama Insurance Guaranty Association liable for insolvent insurer s proportionate share based on approval of district court s time-on-the-risk allocation); Southern Silica of Louisiana, Inc. v. Louisiana Insurance Guaranty Ass'n, No. 07-CA-1680, 2008 WL (La. Apr. 8, 2008) (time-on-the risk approach); EnergyNorth Natural Gas, Inc. v. Certain Underwriters at Lloyd s, 934 A.2d 517, 526 (N.H. 2007) (adopting pro rata approach, but not deciding which pro rata scheme applies); Atchison, Topeka & Santa Fe Ry. Co. v. Stonewall Ins. Co., 71 P.3d 1097, (Kan. 2003) (time-on-the risk approach); Sec. Ins. Co. v. Lumbermens Mut. Cas. Co., 826 A.2d 107, 116, 121 (Conn. 2003) (time-on-the risk approach); Pub. Serv. Co. v. Wallis & Cos., 986 P.2d 924, (Colo. 1999) (time-onthe-risk approach); Domtar, Inc. v. Niagra Falls Ins. Co., 563 N.W.2d 724, 732 (Minn. 1997) (time-on-the-risk approach); Carter-Wallace, Inc. v. Admiral Ins. Co., 712 A.2d 1116, 1124 (N.J. 1998) (mixed time-on-the-risk and percentage of coverage); Owens-Illinois, Inc. v. United Ins. Co., 650 A.2d 974, 993 (N.J. 1194) (mixed time-on-the-risk and percentageof-coverage approach); Consol. Edison Co. v. Allstate Ins. Co., 774 N.E.2d 687, 695 (N.Y. 2002) (adopting pro rata approach, but not deciding which scheme applies); Sharon Steel Corp. v. Aetna Cas. & Sur. Co., 931 P.2d 127, (Utah 1997) (time-on-the-risk and percentage of coverage). 14

15 occurrence during the policy period.... Most fundamentally, the policies provide indemnification for liability incurred as a result of an accident or occurrence during the policy period, not outside that period. [The policyholder s] singular focus on all sums would read this important qualification out of the policy. Consolidated Edison, 774 N.E.2d at 695 (citations omitted). Other courts have concluded that equity demands allocation among insurers based on the time periods covered by their policies. See, e.g., Public Serv., 986 P.2d at 941 ( In our view, the most equitable method of allocation is a system of time-on-the-risk that also takes into account the degree of the risk assumed. ). In considering the equitable justifications for adopting a pro rata allocation approach, some courts have emphasized that a critical inquiry with respect to the fairness of requiring the policyholder to accept a proportionate share of the risk is whether the policyholder elected to assume the risk, either by declining to purchase available insurance or by purchasing what turned out to be an insufficient amount of insurance. Stonewall, 73 F.3d at This inquiry dictates that the policyholder should be required to absorb only losses allocable to policy periods in which the policyholder failed to purchase available insurance, and not those in which insurance coverage for the particular risk ceased to be available. See id. (ruling that proration-to-the-insured should not have been applied to years after 1985 when asbestos liability insurance was no longer available ). Percentage of Limits Approach A few courts that have adopted the pro rata allocation approach have developed an alternative apportionment theory. Instead of allocating solely based on each insurer s time on the risk, these courts divide the liability into shares across the triggered period with heavier weighting for years in which the policyholder purchased more coverage. This version of pro rata allocation was first adopted in Owens-Illinois, Inc. v. United Ins. Co., 650 A.2d 974 (N.J. 1994). Owens-Illinois sought to strike a middle ground between the extremes imposed by both the all sums and pro rata approaches. Id. at In this case, the New Jersey Supreme Court reasoned that, because [t]he language of the policies does not itself yield either result and the usual rules of [contract] interpretation are less helpful in this context, it would be guided by public interest factors bearing on the fairness of requiring particular insurers or policyholders to bear a portion of the risk. Id. at 992. Applying these public interest factors, the court concluded allocation of coverage for a long-tail loss should be spread horizontally among triggered insurers in proportion to the amount of coverage available in each policy period. Id. at The New Jersey Supreme Court has subsequently clarified that the Owens-Illinois method intentionally assigns a greater portion of indemnity costs to years in which greater amounts of insurance were purchased, based on the view that this measure of allocation is more consistent with the economic realities of risk retention and risk transfer. Carter-Wallace, 712 A.2d at 1123 (quoting Chem. Leaman Tank Lines, Inc. v. Aetna Cas. & Sur. Co., 978 F. Supp. 589, 605 (D.N.J. 1997)). This method also dictates that the policyholder should be required to bear a portion of the loss only with respect to periods in which it chose to purchase no or insufficient insurance. Owens-Illinois, 650A.2d 15

16 at 995 ( When periods of no insurance reflect a decision by an actor to assume or retain a risk, as opposed to periods when coverage for a risk is not available, to expect the risk-bearer to share in the allocation is reasonable. ). Practical Solutions 1. All Sums Approach -- Basic The all sums approach most favors policyholders because it enables them to choose which among multiple triggered policies will pay particular losses, and thereby to avoid periods in which they may have been uninsured or self-insured, or otherwise have gaps in their insurance coverage. This approach also reduces the policyholder s transaction costs by shifting to the selected insurers the costs of obtaining contribution from non-selected insurers providing coverage in other policy years triggered by the loss, after the policyholder has obtained reimbursement from the insurance policies of its choosing. The allocation depicted in Table C.1 below provides an illustration. This illustration assumes that, for the period between 1968 and 1972, Asbest-R-Us has primary coverage of $500,000 in each year, but did not purchase excess coverage, and that, in the later period between 1973 and 1978, Asbest-R-Us has 5 years with $1 million primary coverage, as well as has $1 million in excess coverage for each of these years. This illustration also assumes that the claims for asbestos property damage in one of the subdivisions in which insulation was installed by our hypothetical policyholder Asbest-R-Us triggers the 10-year period, that Asbest-R-Us liability is $10 million, and that Abest-R-Us may stack policy limits to obtain full recovery for particular claims. 14 TABLE C.1 $2 M $1 M $ 500K Years PRIMARY A PRIMARY B EXCESS 14 In jurisdictions that have adopted the all sums approach and that apply a continuous or multiple point trigger of coverage theory for progressive injuries, such as those caused by asbestos and environmental contamination, the question of whether stacking of policy limits is permissible will arise when the policies in a given policy period do not afford coverage sufficient to pay for a particular claim or loss. Resolution of this issue can have a dramatic impact on the extent of the policyholder s recovery because an anti-stacking approach may limit the policyholder to only a single or a few towers of coverage, depending on the number and nature of the claims. Courts in all sums jurisdictions have split on this issue. Compare Society Ins. v. Town of Franklin, 607 N.W.2d 342, (Wis. Ct. App. 2000) and Cole v. Celotex Corp., 599 So. 2d 1058, (La. 1992) (allowing stacking) with American Physicians Ins. Exchange v. Garcia,876 S.W.2d 842, (Tex. 1994) and Keene, 667 F.2d at and FMC Corp. v. Plaisted & Companies, 61 Cal.App.4th 1132, 1189 (1998) (refusing to allow stacking). In a recent decision, the California Court of Appeal for the Fourth Appellate District has rejected the reasoning of its sister appellate court for the the Sixth Appellate District in FMC, and approved the reasoning of the courts holding that, absent a policy provision containing an express anti-stacking provision, a policyholder is entitled to stack policy limits. See State of California v. Continental Ins. Co., 2009 Cal. App. LEXIS 1, ** 46-53(Cal. App. Ct. Jan. 5, 2009). 16