limits applicable to Bridgeport were also aggregate limits.
Accordingly, the LMI policies are entitled to a credit for the full amount of Aetna's Bridgeport per-occurrence coverage for the triggered policy years before they are required to pay any of Chemical Leaman's cleanup costs for the site. It is undisputed that the full Aetna per-occurrence policy limits applicable to Bridgeport total $ 11,055,000, regardless of whether these are also aggregate limits.
Thus, the LMI are liable for only those past and future Bridgeport indemnity costs in excess of $ 11,055,000. Remarkably, the settlement credit matches almost exactly the total past indemnity costs on the record: $ 11,084,226. See Section II.A.2. supra. Accordingly, all of the subject LMI policies' apportioned liability, save for approximately $ 30,000, will be for the projected $ 20 million to $ 50 million in future response costs.
2. POLICIES SUBJECT TO ALLOCATION
The Court next considers the LMI's argument that in accordance with Owens-Illinois, Chemical Leaman should absorb a portion of its past and future indemnity costs related to soil and wetlands contamination, based on the liability-phase determinations that pollution exclusion clauses in the 1971-81 policies bar coverage for soil and wetlands damage for those years.
The Court agrees with Chemical Leaman that the Third Circuit has already ruled against any allocation back to the plaintiff insured on account of the pollution exclusion clauses. As stated supra, the Third Circuit affirmed the Court's judgment of liability as to the LMI policies except as to allocation of damages among the LMI insurers. Whereas the judgment had held the LMI and now-exhausted Aetna policies jointly and severally liable for all of Chemical Leaman's Bridgeport indemnity costs, the Third Circuit remanded for a reallocation of liability "between the [London Market] insurers and among the triggered [London Market] policies in accordance with Owens-Illinois." Chemical Leaman, 89 F.3d at 995.
By "triggered policies," the Third Circuit clearly meant those policies which were found by the judgment to cover response costs for contamination to a given environmental medium: groundwater, soil, or wetlands. The Third Circuit did not intend to include as a "triggered policy" for a given medium an LMI policy where a pollution exclusion clause applied. Notably, the Third Circuit recounted that this Court had ruled that "all of the LMI's policies from 1960 to 1985 had been triggered by the environmental contamination at the Bridgeport site, unless a policy exclusion barred coverage." Id. (emphasis added). An examination of the appellate court's opinion makes clear, therefore, that the Third Circuit intended an allocation of all of Chemical Leaman's groundwater-contamination response costs beyond the Aetna coverage to the LMI policies covering 1960-1981; all soil indemnity costs beyond the Aetna insurance to the 1960-71 LMI policies; and all wetlands costs beyond the primary coverage to the 1961-71 LMI policies.
3. ALLOCATION OF DAMAGES AMONG TRIGGERED POLICIES
The Court finally turns to the question of how Bridgeport indemnity costs are to be allocated among the LMI policies pursuant to Owens-Illinois. As the Third Circuit explained, the New Jersey Supreme Court in Owens-Illinois "held a fair method of allocation appears to be one that is related both to time on the risk and the risk assumed, i.e., proration on the basis of policy limits, multiplied by years of coverage." Chemical Leaman, 89 F.3d at 995 (citations omitted). In so holding, the Owens-Illinois court expressly adopted the risk-based allocation formula developed in Armstrong World Indus., Inc. v. Aetna Cas. and Sur. Co., 20 Cal. App. 4th 296, 26 Cal. Rptr. 2d 35 (Cal. Ct. App. 1993), review granted and opinion superseded sub. nom. In re Asbestos Ins. Coverage Cases, 866 P.2d 1311, 27 Cal. Rptr. 2d 488 (Cal. 1994), review transferred to Court of Appeal sub. nom. In re Asbestos Ins. Coverage Cases, 904 P.2d 370, 46 Cal. Rptr. 2d 174 (1995), decsn. after transfer, 45 Cal. App. 4th 1, 52 Cal. Rptr. 2d 690 (Cal. Ct. App. 1996), review denied (Aug. 21, 1996) ("Armstrong ").
Owens-Illinois, 650 A.2d at 993.
The New Jersey Supreme Court explained the Armstrong method using a hypothetical in which the continuous trigger theory activated office-building owners' CGL policies spanning nine years for coverage of asbestos-related injury claims filed by building workers. The policyholders' coverage profile in one of the scenarios posited by Owens-Illinois under this example was as follows: During each of the first three years, the insureds were covered under CGL policies for $ 2 million per occurrence per year. During each of the middle three years, policies were in effect at $ 3 million per year. During the last three years, the building owners were uninsured; for allocation purposes, the Owens-Illinois court said the risk assumed by the owners for these final three years was $ 4 million per year. Under the Armstrong method as interpreted by the New Jersey Supreme Court, "carriers during the first three years would bear roughly twenty-two percent (6/27ths) [of the covered losses], carriers covering the middle three years would bear thirty-three percent (9/27ths); and the building owners would bear forty-four percent of the risk (12/27ths). Of course, policy limits and exclusions must be taken into account." Owens-Illinois, 650 A.2d at 979-980, 994.
a. Allocation Among Layers of LMI Policies
An important Owens-Illinois issue about which the parties disagree is how indemnity costs should be allocated among various layers of insurance. The LMI argue that each layer of liable coverage must "horizontally exhaust," i.e. that the underlying layer must be exhausted across all of the triggered policy years before the next layer would begin paying. Chemical Leaman, by contrast, argues that the Owens-Illinois method divides damages among triggered policy years, and that the attachment point for a given excess policy will depend on the amount of damages allocated to that year and the limits of the underlying policy.
The parties' arguments on this point focus on the issue of exhaustion of the Aetna primary policies in the Owens-Illinois proration scheme. This question has been mooted by the Court's treatment of the settlement supra. Moreover, Owens-Illinois mandates that neither the Aetna policies nor their limits should be included in the proration because they will not be called on for future contribution in the wake of the settlement. See Owens-Illinois, 650 A.2d at 994 ("We are informed that Aetna has paid its policy limits for the years 1963 to 1977, so that Aetna's policy proceeds will not be called upon for future contribution. If that be so, rather than go back to revisit Aetna's contributions, we shall start forward . . . .").
The parties' same arguments, however, apply to the important issue of exhaustion of each layer of LMI excess insurance. That is, the Court must determine whether each layer of liable excess coverage must horizontally exhaust before the next layer would respond. This is a central question in this case, because the triggered LMI policies for purposes of liability allocation are arrayed across a number of layers. In policy year 1970-71, for example, there are three layers of LMI excess policies overlying the exhausted Aetna primary policy for that year.
The Owens-Illinois court did not directly resolve this issue, and acknowledged as much. The New Jersey Supreme Court stated:
We realize that many complexities encumber the solution that we suggest involving, as it does, proration by time and degree of risk assumed -- for example, determining how primary and excess coverage is to be taken into account or the order in which policies are triggered. See Hickman & DeYoung, supra, at 310-11; Roger Westendorf & Ronald R. Robinson, Insurance Coverage for Environmental Claims Under the Comprehensive General Liability Policy, in Pollution Liability: Managing the Challenges of Coverage and Defense in 1991, at 57 (ALI-ABA Video Law Review Study, Q205, 1991), available in WESTLAW, ALI-ABA Database, *54-57 (each discussing various theories of horizontal and vertical stacking and relationship to excess coverage issues). The parties did not focus on these issues. Still, we do not believe that the issues are unmanageable.
Owens-Illinois, 650 A.2d at 994.
The state supreme court did, however, provide some indications as to the appropriateness of horizontal exhaustion. It is notable that in the example it provided of how its method would operate, it allocated damages to "the carriers on the risk" for a given three-year period according to a ratio of the amount of the total of the per-occurrence limits in each of the three years to the total per- occurrence limits for all of the triggered years.
Chemical Leaman is correct, therefore, that the Owens-Illinois court called for a division of damages among the triggered policy years. 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. A year-by-year increase in policy limits, the New Jersey Supreme Court suggests, may well reflect an increasing awareness of the escalating nature of the risks sought to be transferred. Owens-Illinois, 650 A.2d at 993.
Accordingly, the Court rejects the LMI's argument that horizontal exhaustion of liable policies is required.
Instead, the Court directs apportionment of damages among policy years without reference to the layering of policies in the triggered years. Given that there are different sets of triggered years for the three different media, costs classified as groundwater, soil and wetlands will be allocated according to medium. Thus groundwater costs will allocated among the 1960-81 years; soil costs will allocated among the 1960-71 years; and wetlands costs will be allocated among the 1961-71 years.
Groundwater-cost proration among triggered years, for example, will work as follows: The Court having concluded that the total LMI coverage for Bridgeport between 1960-81 is $ 415,225,000,
and that LMI policies are in effect for 1960-61 at limits of $ 1,225,000, the LMI policies for 1960-61 will bear .29% of the groundwater indemnity costs above the LMI's $ 11,055,000 settlement credit against all Bridgeport indemnity. Since LMI policies were in effect for 1980-81 at $ 59,000,000, by contrast, these policies will bear 14.21% of the groundwater costs allocated to the LMI policies. Apportionment to the other policy years will proceed according to this model.
Though it rejects horizontal exhaustion, the Court will take into account the layering of the LMI policies in any given policy year by directing that any indemnity costs allocated to that year be applied against the lowest-layer LMI policy not yet exhausted in that year. Stated another way, each layer of excess coverage in a given year must exhaust before the next layer will be required to begin paying indemnity costs to Chemical Leaman. For example, suppose $ 1.4 million in groundwater contamination indemnity costs are ultimately allocated to the LMI policies in the 1970-71 policy year, which will bear 4.70% of the groundwater costs allocable to the LMI policies.
For reasons the Court will discuss in subsection II.B.3.b. infra, the first-layer LMI excess policy for 1970-71 has a $ 1 million operative Bridgeport limit for this policy year, the second-layer policy has an operative limit of $ 8.5 million for the year, and the third-layer policy has a limit of $ 10 million. (LMI Ex. D1; LMI Ex. D2, Tab 3; Pl.'s Ex. P1). Therefore, the first-layer policy would be allocated $ 1 million and be exhausted, the second-layer policy would attach and incur a payment obligation of $ 400,000, and the third-layer policy would not attach for Bridgeport unless the 1970-71 policy year was allocated a substantial amount of soil and wetlands response costs.
b. Annualization of Per-Occurrence Limits
Another dispute between the parties concerns the application of the LMI policies' per-occurrence limits to the Owens-Illinois allocation. Chemical Leaman does not dispute the LMI's argument that a single per-occurrence limit exists in each LMI policy, even those policies with stated policy terms greater than one year. But plaintiff contends it is entitled to coverage up to the per-occurrence limit in each triggered year of every penetrated LMI policy. Plaintiff argues that the Third Circuit's decision in this case mandates treatment of continuous and indivisible property damage such as at Bridgeport as separate and distinct occurrences in each of the years of a liable CGL policy. The LMI responds that the Third Circuit directs treatment of the Bridgeport contamination as a single, continuous occurrence spanning triggered policy years.
Chemical Leaman notes that at the beginning of its discussion of the continuous trigger theory, the Third Circuit stated the following:
The New Jersey Supreme Court adopted the "continuous trigger" theory to identify the time of an "occurrence" in Owens-Illinois, Inc. v. United Ins. Co. The continuous trigger theory recognizes that "when progressive indivisible injury or damage results from exposure to injurious conditions for which civil liability may be imposed, courts may reasonably treat the progressive injury or damage as an occurrence within each of the years of a CGL policy."
Chemical Leaman, 89 F.3d at 994-095 (citations omitted) (quoting Owens-Illinois, 650 A.2d at 974, 995). Through this statement, Chemical Leaman asserts, the Third Circuit expressly adopted a determination in Owens-Illinois that progressive property damage is to be treated as separate occurrences in each of the triggered years of a triggered insurance policy, regardless of whether that policy covers one year or three years. Plaintiff contends that the Third Circuit's quotation of Owens-Illinois should be held to resolve the number-of-occurrences issue in part "because at the time of the Third Circuit's decision, the only policies left before the Court were the LMI policies [since] Aetna settled, and the Third Circuit was very well aware that these were multiyear policies." (Tr. at 105).
Notwithstanding the impact of this holding on the LMI's multi-year policies, plaintiff states, defendant insurers did not seek rehearing on this aspect of the Third Circuit's decision. As a result, the insured concludes, the appellate court's determination is the law of the case and the LMI are now barred from disputing that Chemical Leaman is entitled to recover up to the per-occurrence limit of each attached LMI policy for each triggered policy year.
The LMI argue for a much different reading of the Third Circuit's decision. They cite the following portion of the appellate opinion, later in the Third Circuit's discussion of the continuous trigger theory and Owens-Illinois allocation:
Although considering the issue before the New Jersey Supreme Court's decision in Owens-Illinois, the district court applied the continuous trigger theory, ruling all of the LMI's policies from 1960 through 1985 had been triggered by the environmental contamination at the Bridgeport site, unless a policy exclusion barred coverage. The district court also held that all insurance policies activated by a continuing occurrence are jointly and severally liable to policy limits for property damage resulting from the occurrence.
Chemical Leaman, 89 F.3d at 995 (citation omitted). Relying on this passage, defendant insurers contend that the Court of Appeals accepted a premise in Owens-Illinois that property damage like that at issue in this case should be treated as one single, continuous occurrence.
This Court has determined that the Third Circuit has not definitively ruled, either expressly or implicitly, on whether the Bridgeport contamination should be treated as one occurrence or as a separate occurrence in each year for purposes of applying the LMI policies' per-occurrence limits. Accordingly, the Court will consider the issue as a matter of first impression. Casey, 14 F.3d at 857.
As recounted above, the New Jersey Supreme Court in Owens-Illinois held that under the continuous trigger theory it was propounding, progressive indivisible property damage should be treated "as an occurrence within each of the years of a CGL policy." Owens-Illinois, 650 A.2d at 995. On its face, this language appears to direct treatment of progressive property damage as distinct occurrences triggering per-occurrence limits in each year of a policy. Furthermore, a consideration of the context of this statement leads the Court to conclude that this is in fact the proper construction of Owens-Illinois. The Court considers both the continuous-trigger discussion and the allocation discussion in Owens-Illinois, because the New Jersey Supreme Court determined that the two issues were inextricably linked; "We believe," the Owens-Illinois court explains, "that common-law resolution of the trigger-of-coverage issue requires that we consider, at the same time, the issue of scope of coverage if a policy is triggered." Owens-Illinois, 650 A.2d at 985.
The Owens-Illinois court relies very heavily throughout its opinion on Armstrong ; indeed, as related supra, the New Jersey Supreme Court expressly adopted in full the Armstrong allocation formula. The Armstrong trial court's elaboration of the formula, in the context of asbestos-related bodily injury claims, clearly reflects and responds to an underlying premise that injury should be treated as distinct occurrences activating per-occurrence limits in every triggered policy year. For the sake of context, the Armstrong trial court's statement of the allocation method embraced by Owens-Illinois will be quoted at length:
This Court finds that the most equitable method of allocation is proration on the basis of policy limits, multiplied by years of coverage. This method is consistent with the policy language in that it takes policy limits into consideration . . . This method also reflects the fact that higher premiums are generally paid for higher 'per person' or 'per occurrence' limits. Since some policies are in effect for more than one year, and injury occurs during every year from first exposure to asbestos until death (Phase III decision at p. 42). Multiplying the policy limits by years of coverage results in a more equitable allocation than proration based on policy limits alone. Thus, when a particular claim triggers more than one policy, each insurer's share of liability shall be determined by the proportion that each policy's applicable 'per occurrence' limits multiplied by years the policy was in effect bears to the sum total of the applicable 'per occurrence' limits of all triggered policies multiplied by the years each policy was in effect. When a policy does not contain a 'per occurrence' limit, the 'per person' limit shall be used in this calculation."