(back to preceding text)

58.  The solution to the question on these appeals, and the reasons why they should be allowed, seem to me to be found in these steps:

(1)  In order to establish liability against a reinsurer, the reinsured has to establish that the loss is within the risk assumed under the underlying insurance contract; and that the relevant risk has been assumed under the reinsurance contract.

(2)  Whether the relevant risk has been assumed under the reinsurance contract is a question of construction of that contract.

(3)  In principle the relevant terms in a proportional facultative reinsurance - and in particular those relating to the risk - should be construed so as to be consistent with the terms of the insurance contract on the basis that the normal commercial intention is that they should be back-to-back.

(4)  Where the insurance contract and the reinsurance contract are governed by different laws, it remains a question of construction of each contract under its applicable law as to what risk is assumed, and there is no special rule of the conflict of laws which governs the consequences of any inconsistency.

(5)  Both the insurance contract and the reinsurance contract were “losses occurring during” (or “LOD”) policies (or “occurrence policies” as they are known in the United States), which in English law means that an insurer (or reinsurer) is liable to indemnify the insured (or reinsured) in respect of loss or damage which occurs during the policy period.

(6)  There was not in 1977, when the insurance contract and the reinsurance contract were concluded, any identifiable system of law applicable to the insurance contract which could have provided a basis for construing the contract of reinsurance in a manner different from its ordinary meaning in the London insurance market.

(7)  The effect of the decision of the Supreme Court of Washington is to impose liability on Lexington under the contract of insurance for loss and damage which occurred both before and after (as well as during) the policy period in the reinsurance contract.

(8)  It is common ground that under English law an insurer (or reinsurer) would not be liable for losses occurring before and after the policy period.

(9)  Although normally any loss within the coverage of the insurance will be within the coverage of the reinsurance, there is no rule of construction, and no rule of law, that a reinsurer must respond to every valid claim under the insurance irrespective of the terms of the reinsurance.

(10)  The reinsurance contract cannot reasonably be construed to mean that it would respond to any liability which “any court of competent jurisdiction within the United States” (the phrase in the Service of Suit clause) would impose on Lexington irrespective of the period of cover in the reinsurance contract.

59.  It is elementary and obvious that a reinsurer cannot be held liable unless the loss falls within the cover of the underlying insurance contract and within the cover created by the reinsurance: Hill v Mercantile and General Reinsurance Co plc [1996] 1 WLR 1239, at 1251, per Lord Mustill. It is equally elementary that what falls within the cover of a contract of reinsurance is a question of construction of that contract.

60.  In the case of proportional facultative reinsurance the obvious commercial intention is for the original insurer to reinsure part of its own risk and for the reinsurer to accept that part of the risk, and it is therefore equally obvious that the relevant terms in the reinsurance contract should be construed so as to be consistent with the contract of insurance. This is simply commercial common sense. Consequently, in proportional facultative reinsurance the starting point for the construction of the reinsurance policy is that the scope and nature of the cover in the reinsurance is co-extensive with the cover in the insurance. As Staughton LJ said in Youell v Bland Welch & Co Ltd [1992] 2 Lloyd’s Rep 127, 132: “One can … readily assume that a reinsurance contract was intended to cover the same risks on the same conditions as the original contract of insurance, in the absence of some indication to the contrary.”

61.  An early example of this principle is Joyce v Realm Marine Insurance Co (1872) LR 7 QB 580. The insurance covered (inter alia) cargo from ports in West Africa with outward cargo to be considered homeward interest 24 hours after the ship’s arrival at her first port of discharge. The reinsurance was at and from West African ports “to commence from the loading of the goods.” Goods shipped at Liverpool were lost 24 hours after the ship’s arrival at the port of Cabenda. It was held by the Court of Queen’s Bench that “loading” in the reinsurance applied to outward cargo from Liverpool to West Africa which was left on board and considered as homeward cargo under the insurance. The terms in the reinsurance in the light of the insurance showed that “what was meant between the parties was not the actual loading, but a constructive loading, which was what the original underwriters had agreed to treat as a loading on board for the purpose of the homeward voyage": at 586, per Lush J.

62.  More than a hundred years later Forsikringsaktieselskapet Vesta v Butcher [1989] AC 852 and Groupama Navigation et Transports v Catatumbo CA Seguros [2000] 2 Lloyd’s Rep. 350 affirmed the continuing significance of the principle. In Vesta v Butcher Lord Griffiths said (at 895):

“In the ordinary course of business reinsurance is referred to as ‘back-to-back’ with the insurance, which means that the reinsurer agrees that if the insurer is liable under the policy the reinsurer will accept liability to pay whatever percentage of the claim he has agreed to reinsure. A reinsurer could, of course, make a special contract with an insurer and agree only to reinsure some of the risks covered by the policy of insurance, leaving the insurer to bear the full cost of the other risks. Such a contract would I believe be wholly exceptional, a departure from the normal understanding of the back-to-back nature of reinsurance and would require to be spelt out in clear terms. I doubt if there is any market for such a reinsurance.”

63.  Where the potential conflict between the insurance contract and the reinsurance arises from the fact that they are governed by different laws, the question whether the conflict can be resolved remains a question of construction. The solution cannot be found in any rules of the conflict of laws.

64.  An early example of such a conflict is St Paul Fire and Marine Insurance Co v Morice (1906) 11 Com Cas 153. St Paul insured under a United States policy a bull shipped from New York to Buenos Aires against (inter alia) “all risks of mortality". The bull was infected with foot and mouth disease and slaughtered on board on arrival in Argentina pursuant to Argentine law and regulations. A Lloyd’s policy of reinsurance (“subject to the same terms … as original policy ...”) insured the bull against all risks “including mortality". The reinsured settled the claim under the United States insurance policy and claimed on the English law reinsurance. The policy was issued by a Minnesota insurance company, and referred to the potential liability of the insurer under “the rules and customs of insurance in Boston or New York.” The reinsured called expert evidence on United States law, rather than the law of Minnesota, New York or Massachusetts, presumably because this case was decided before the Supreme Court ruled in Erie Railroad Co v Tompkins, 304 US 64 (1938) that there was no federal common law, or because there was no difference between the laws of those States. The expert evidence was to the effect that under United States law the reinsured was liable to pay the insured under the words “all risks of mortality". The reinsurers (represented by Mr Scrutton KC) argued that mortality in both the insurance and the reinsurance meant death by such things as accident, but not by intentional killing.

65.  In an unreserved judgment, Kennedy J said that after hearing the expert evidence he did not think that there was any strong reason for supposing that the words did include, as a matter of United States law, slaughter of the kind in question: at 163. But if he were wrong in that, he considered whether, as a matter of construction, the reinsurers were bound to pay. On that point, he decided that the natural construction of the reinsurance policy under English law was the same as the construction he had given to the United States policy, namely that mortality did not include death by the intentional act of the officials at Buenos Aires. If there had been no grounds for rejecting the evidence of United States law (or, as it would now be, the law of the State whose law governed the policy), it is likely that the case would have been decided differently today, and it does not give much support to the appellants’ case.

66.  Vesta v Butcher and Groupama v Catatumbo were cases where the insurance contracts and reinsurance contracts contained, or incorporated, the same or similar language, but were governed by different laws. In those cases the apparent conflict between the insurance and the reinsurance arose, not from a difference in wording between the policies, but from the different effect which identical or similar wording had under the different laws governing the insurance and the reinsurance. They were much easier cases than the present one. In each case the reinsurers were taking the wholly unmeritorious point that they were relieved from liability because the original insured (and not the reinsured) had been guilty of a breach of a warranty. In each case the warranty was held to be a term of both the insurance and the reinsurance contracts. In each case the breach was not, or was assumed not to have been, causative of the loss. In each case the governing law of the insurance contract did not afford a defence where the breach was non-causative.

67.  In Vesta v Butcher the insurance was for loss or damage to a fish farm in Norway. As in the present case, the reinsurance policy was put in place before the original insurance was written. The insurance and the reinsurance were broked as part of a package. London underwriters and brokers had marketed insurance contracts for fish farms across the world. They did not do so directly but made use of a local insurance company to obtain the business. The brokers interested Vesta in the business on the understanding that the brokers would be able to obtain 90% reinsurance of Vesta’s risk in the London market.

68.  The contract of insurance contained the terms: “It is warranted that a 24 hour watch be kept over the site … Failure to comply with any of the warranties will render this policy null and void.” The reinsurance policy form was Form J1, and the slip annexed the original insurance terms. The litigation was conducted on the basis that the warranty in the insurance contract was also a term of the reinsurance. Hobhouse J refused the brokers (who were being sued by Vesta for failure to obtain an effective reinsurance) leave to amend so as to plead that the 24 hour watch clause was not a term of the reinsurance: [1986] 2 All ER 488, 496-497. Lord Templeman (with whom Lords Bridge and Ackner agreed) treated Form J1 as emphasising that the two policies were on the same terms ([1989] AC at 891), and Lord Lowry (with whom Lords Bridge and Ackner also agreed) approved Hobhouse J’s statement to the same effect (at 901). Lord Griffiths expressed doubts (which I have to say have considerable force) about whether the effect of Form J1 was to incorporate the warranty in the reinsurance (at 896). The insurance contract was governed by Norwegian law, and the reinsurance contract was held by the Court of Appeal to be governed by English law (and there was no appeal on that point to this House).

69.  In Groupama v Catatumbo the insurance gave hull and machinery cover for a fleet of vessels. There was a warranty as to maintenance of existing class in the insurance contract (“guarantee of maintenance of existing class”) which had been incorporated in the reinsurance contract in similar but not identical terms (“Warranted existing class maintained”). The insurance policy had been issued in Spanish by a Venezuelan insurance company to a Venezuelan insured providing for jurisdiction of a Venezuelan court if the parties did not agree to arbitration. It was accepted that it was governed by Venezuelan law. It was common ground that the reinsurance contract was governed by English law.

70.  In Vesta v Butcher the express term that failure to comply with the warranties rendered the policy null and void was ineffective under Norwegian law if the breach was non-causative, whereas a similar term in the reinsurance would be valid under English law. In Groupama v Catatumbo breach of warranty affected the insurance cover under Venezuelan law only if it were causative, while English law (Marine Insurance Act 1906, sections 33(3), 34(2)) discharged an insurer from the date of the breach irrespective of whether it had been remedied before the loss.

71.  In both cases the reinsurers failed because the reinsurance was held to have the same effect as the insurance. In Vesta v Butcher the speeches of both Lord Templeman and Lord Lowry commanded a majority. They were agreed that the question was one of construction of the reinsurance contract. Lord Templeman’s conclusion was founded on his view that “the effect of a warranty in the reinsurance policy is governed by the effect of the warranty in the insurance policy because the reinsurance policy is a contract by the underwriters to indemnify Vesta against liability under the insurance policy” ([1989] AC at 892). For Lord Lowry, the main point was that the relevant words in the reinsurance contract (“failure to comply”) had the same meaning and effect as they had in the Norwegian insurance contract.

72.  In Groupama v Catatumbo it was held that the parties to the reinsurance contract must be taken to have intended that the incorporation in the reinsurance contract of terms in the original insurance retained the same significance which they had in the original insurance. It was a question of construction, against the background that “reinsurers conducting international business must be taken to have intended that the warranties in the two contracts will have the same effect” (at [20] per Tuckey LJ) and that the “reinsurance is … a contract which in terms relates to and must be read in conjunction with the terms of the original insurance” (at [26] per Mance LJ).

73.  Tuckey LJ rightly emphasised at [20]: “… reinsurers conducting international business must be taken to have intended that the warranties in the two contracts will have the same effect. They will be aware that the laws of some countries give more restrictive effect to warranties than English law, but that is a risk they must be taken to have assumed by writing international business. They will be protected to the same extent as the insurer.” Mance LJ warned (at [30]) against a narrow English law-centred approach: “The … submission that the warranty of existing class maintained in the reinsurance retains a stubbornly domestic English significance, trumping any limited significance of such a warranty included in the original and also incorporated by reference into the reinsurance is, to my mind, both commercially and legally unattractive.”

74.  In English law, where an insurance or reinsurance contract provides cover for loss or damage to property on an occurrence basis, the insurer (or reinsurer) is liable to indemnify the insured (or reinsured) in respect of loss and damage which occurs within the period of cover but will not be liable to indemnify the insured (or reinsured) in respect of loss and damage which occurs either before inception or after expiry of the risk. As Lord Campbell CJ said in Knight v Faith (1850) 15 QB 649, at 667, “the principle of insurance law [is] that the insurer is liable for a loss actually sustained from a peril insured against during the continuance of the risk.” An early example of a “losses occurring during” insurance policy is Re London Marine Insurance Association (1869) LR 8 Eq 176 (Sir William James V-C). I accept that there may be scope for considerable argument as to what would constitute loss or damage within the policy period: cf Bolton Metropolitan Borough Council v Municipal Mutual Insurance Ltd [2006] EWCA Civ 50, [2006] 1 WLR 1492 (mesothelioma in the context of “loss or damage [which] occurs during the currency of the policy”).

75.  In the present case the contract of insurance is described as a “Special Floater Policy” and is expressed to have been issued by Lexington to Alcoa on August 22, 1977. The printed section (or “policy jacket”) has a section for “From the … day of … 19.. To the … day of 19… beginning and ending at noon (Standard Time at the place of issuance of this Policy)” and the dates July 1, 1977 to July 1, 1980 have been added. The rest of the jacket contains standard conditions. The Supreme Court of Washington set out the history of the cover: Aluminum Co of America v Aetna Casualty & Surety Co, 998 P 2d 856, at 863 (Wash 2000). The lengthy tailor-made terms were prepared by Alcoa’s internal insurance department and its brokers. Large firms of brokers shopped the terms to various insurers. The insurers responded with price quotations, and upon placement of coverage, the insurers sent the policy jackets with standard policy language to the brokers for inclusion in the policies to be added.

76.  The reinsurance contract (which was put in place while the insurance policy was being marketed) covered “All Risks of Physical Loss or damage” and provided cover in respect of loss and damage occurring between 1 July 1977 and 1 July 1980 (“PERIOD: 36 months at 1.7.77 ..”). Consequently this was reinsurance on the “loss occurring” basis, under which a reinsurer is obliged to pay its share of the loss suffered by the reinsured, if it occurred during the period when the reinsurance contract was in force: Balfour v Beaumont [1984] 1 Lloyd’s Rep 272, at 274, per Donaldson LJ; Youell v Bland Welch & Co Ltd [1992] 2 Lloyd’s Rep 127, at 131, per Staughton LJ.

77.  A case in which there was a mismatch between the periods of cover in the insurance contracts and the reinsurance contracts was Municipal Mutual Insurance Ltd v Sea Insurance Co Ltd [1998] Lloyd’s Rep IR 421, where it was held that the reinsurance did not have to respond to the insurance because the vandalism for which the plaintiff insurers had had to indemnify the Port of Sunderland had occurred outside the policy period in the reinsurance. The importance of the period of cover was rightly emphasised by Hobhouse LJ, at 435-6:

“… It is wrong in principle to distort or disregard the terms of the reinsurance contracts in order to make them fit in with what may be a different position under the original cover…

…When the relevant cover is placed on a time basis, the stated period of time is fundamental and must be given effect to. It is for that period of risk that the premium payable is assessed. This is so whether the cover is defined as in the present case by reference to when the physical loss or damage occurred, or by reference to when a liability was incurred or a claim made. Contracts of insurance (including reinsurance) are or can be sophisticated instruments containing a wide variety of provisions, but the definition of the period of cover is basic and clear. It provides a temporal limit to the cover and does not provide cover outside that period; the insurer is not then ‘on risk'…”

78.  In summary, what was decided by the Supreme Court of Washington (Aluminum Co of America v Aetna Casualty & Surety Co, 998 P 2d 856 (Wash 2000)) was that under Pennsylvania law (which Judge Learned had found applicable) all insurers were jointly and severally liable for all losses which flowed from the property damage even if the damage occurred before (or after) inception, because the policies were not limited as to time. The decision of the Supreme Court of Washington has to be read in the context of the development of the law in the United States on the liability of successive insurers on policies covering liability for asbestos-related claims and for environmental claims.

79.  The central decision in the development of the law in the United States is Keene Corp v Insurance Co of North America, 667 F 2d 1034 (DC Cir 1981), cert den 455 US 1007 (1982). The Court of Appeals for the District of Columbia Circuit decided that, in asbestos-related claims, coverage under insurance policies was triggered by any one of: manifestation of disease, inhalation exposure, and exposure in residence (i.e. the subsequent development of the disease). The Court of Appeals then went on to consider the extent of coverage, and held that each insurer was liable to indemnify Keene in full (and not merely pro rata) for the whole of the damages for which it was liable to the plaintiffs in the underlying actions (more than 6000 actions were pending). The policies typically provided that the insurer would “pay on behalf of the insured all sums which the insured shall become legally obligated to pay as damages because of bodily injury … to which this insurance applies, caused by an occurrence ..” “Occurrence” was defined as “an accident, including injurious exposure to conditions, which results, during the policy period, in bodily injury …”

80.  The Court of Appeals took the view that Keene did not expect, nor should it have expected, that its security was undermined by the existence of prior periods in which it was uninsured, and in which no known or knowable injury occurred. If an insurer were obliged to pay only on a pro rata basis, those reasonable expectations would be violated. There was nothing in the policies which provided for a reduction of the insurer’s liability if an injury occurred only in part during a policy period. The court interpreted the policies to cover Keene’s entire liability if an injury occurred only in part during a policy period. For an insurer to be only partially liable for an injury which occurred, in part, during its policy period would deprive Keene of insurance coverage for which it paid.

81.  Shortly before the decision in Keene Corp v Insurance Co of North America the Court of Appeals for the Sixth Circuit had held that the insurers were liable pro rata to the periods of coverage (with the insured being treated as a self-insurer for years when it was not covered): Insurance Co of North America v Forty-Eight Insulations, Inc, 633 F 2d 1212 (6th Cir 1980), cert den 454 US 1109 (1981).

82.  These two approaches have spawned an enormous number of decisions in asbestos-related claims and in environmental claims, many of which are discussed or referred to in Holmes (ed), Appleman on Insurance 2d, 2003, chap 145, and in Ostrager and Newman, Handbook on Insurance Coverage Disputes, 13th ed 2006, chap 9 (who point out, at 618, that the joint and several liability approach has been used more frequently in personal injury cases than in property damage).

83.  The position as it was in 2008 was reviewed by the Court of Appeals for the First Circuit in Boston Gas Co v Century Indemnity Co, 529 F 3d 8 (1st Cir 2008). In cases such as the present a federal court must apply the law of the State in which it sits. The Massachusetts courts had not yet resolved the allocation question as a matter of law, at the highest level, although the joint and several liability approach had been adopted by lower courts: see Rubenstein v Royal Insurance Co of America, 694 NE 2d 381 (Mass. 1998), affd on other grounds, 708 NE 2d 639 (Mass. 1999); Peabody Essex Museum, Inc v United States Fire Insurance Co, 2009 WL 901869 (D.Mass.2009). Consequently, the Circuit Court of Appeals certified the question for decision by the Massachusetts Supreme Judicial Court. In so doing the Circuit Court of Appeals noted (at 13-14) that a “growing plurality have adopted some form of pro rata allocation but a significant number of courts impose joint and several allocation.” The Court of Appeals said (at 14):

“Nor do policy arguments line up solely behind one solution. At first blush it may seem illogical to hold a single insurer, who may have only covered the insured for a single year, fully liable for the costs of environmental damage that may have accrued over the course of a century. But that insurer can seek contribution from other insurers ‘on the risk’ during the contamination period. … And the alternative may force the insured to sue numerous companies in one suit, if this is possible at all, to avoid inconsistencies.

Either method forces courts to indulge in a probable fiction as to when the event triggering coverage occurred. The pro rata method assumes an ongoing occurrence causing stable amounts of damage over time; the joint and several method pretends, even less plausibly, that a single occurrence caused all the damage, and allows the insured effectively to choose the year in which that happened. Both are crude approximations made under conditions of uncertainty.”