AMERICAN HOME PROD. v. LIBERTY MUT. INS. CO.

OPINION AND ORDER

SOFAER, District Judge:

Plaintiff American Home Products Corporation ("AHP"), a diversified company manufacturing drugs, foods, and household products, is the defendant in fifty-four products-liability suits arising from AHP's manufacture and sale of six pharmaceuticals: Ovral and L/Ovral (oral contraceptives), DES (Diethylstilbestrol), Mysoline, Atromid-S, Premarin, and Anacin. Defendant Liberty Mutual Insurance Company ("Liberty"), which provided AHP with insurance from 1944 until 1976, has refused to assume AHP's burden of defense or to indemnify AHP in those lawsuits, because in each case physical harm did not become manifest until after termination of the insurance policies. In this action AHP seeks a judgment declaring that Liberty is obliged to defend and to indemnify AHP in the underlying lawsuits because, regardless of when physical harm became manifest, exposure to the alleged agents of harm occurred during the policy periods, thereby triggering coverage.

Jurisdiction is based on diversity of citizenship, 28 U.S.C. § 1332 (1976), and New York law controls. AHP contends there are no disputed issues of material fact, and has moved for summary judgment awarding the declaration it seeks. Liberty opposes this motion, and has itself moved for partial summary judgment on the basis of a provision in the AHP policies which Liberty argues excludes coverage for all claims involving exposures to allegedly harmful substances after termination of Liberty's coverage on November 1, 1976.

Several courts have recently ruled on the scope of insurance policies covering liability for insidious diseases, which are illnesses that become manifest long after initial exposure to the substances believed to cause them. The policy provisions at issue in these cases were all variants of the Comprehensive General Liability Policy ("CGL"), a standard-form policy for liability coverage drafted during the 1960's by representatives of the insurance industry to deal with the problem of liability for injuries caused over a period of time. Instead of covering only "accidents", a word that connotes an event causing immediate or contemporaneous injury, the CGL was written to cover "occurrences", defined to include "an accident, including injurious exposure to conditions, which results, during the policy period, in bodily injury ... neither expected nor intended from the standpoint of the insured." CGL, Pl.Ex. 22 at 12. This change in terminology made clear the intent of insurers to provide coverage for insidious diseases. But the new language provided no definition of "bodily injury" other than the words themselves, thereby creating a basis for disputes as to the trigger of coverage.

AHP's insurance policies with Liberty were "manuscript" policies written specifically for AHP. Like the 1966 version of the CGL, however, AHP's policies throughout the period relevant to this litigation provided liability coverage for "occurrences" that result in "personal injury, sickness or disease including death resulting therefrom ... sustained by any person." An occurrence is defined by inference from Article IV: "This policy applies only to (1) personal injury, sickness or disease including death resulting therefrom ... which occurs during the policy period." Pl.Ex. 20 at 3. The ultimate question under both the CGL language and the AHP policies at issue here is therefore the same: when does "injury, sickness or disease" occur? Under both policies, coverage exists only for injuries occurring during the policy period. Moreover, both the CGL and AHP's policies require the insurer to defend any suit against the insured that seeks damages for an injury alleged to have occurred under the policy, even if the suit is groundless or fraudulent. See CGL, Pl.Ex. 22 at 1; AHP Policy, Pl.Ex. 20 at 1.

One provision in AHP's policies after 1968 differs from anything in the CGL, however. The provision Liberty relies on in its motion for partial summary judgment states:

Liberty contends that this provision renders these policies inapplicable to twenty-eight specified cases. See Urmston Aff't, Ex. 1.

For the reasons that follow the policies in this case are construed as they are written — to require a showing of actual injury, sickness or disease occurring during the policy period, based upon the facts proved in each particular case. Thus, an occurrence of "personal injury, sickness, or disease" is read to mean any point in time at which a finder of fact determines that the effects of exposure to a drug actually resulted in a diagnosable and compensable injury. Depending upon the facts of each case, the drug involved, the period and intensity of exposure, and the person affected, an injury may occur in this sense upon exposure, at some point in time after exposure but before manifestation of the injury, and at manifestation. This construction is supported by the policy's language and background, the intentions and expectations of the parties, and considerations of practicability and fairness. It provides liberal protection to the insured, without doing violence to the principle — long a part of the law of New York — that insurance policies are contracts under which insureds obtain all the protection for which they may reasonably be said to have paid, but not more.

I. Meaning of The Policy Language

Whether the contract between AHP and Liberty is ambiguous is of central importance to this case, and particularly to the disposition of this motion. AHP argues that the policies supply an ambiguous definition of "occurrence" that is susceptible to at least two plausible constructions. Under one construction coverage would be triggered by every exposure to a harmful substance that could ultimately result in bodily injury. Another construction would trigger coverage only when injury became manifest, which has been defined to mean when the injury was diagnosed or when it produced symptoms that placed or should have placed the injured person on notice. Relying on these alleged ambiguities, AHP invokes the well established rule that resolves ambiguities in insurance contracts in favor of the insured and against the insurer. See, e.g., Breed v. Insurance Co. of North America, 46 N.Y.2d 351, 353, 385 N.E.2d 1280, 1282, 413 N.Y.S.2d 352, 354 (1978). AHP argues that, in this case, the governing rule of construction requires application of the exposure theory, which favors manufacturers, or perhaps an even broader construction providing coverage at any point from exposure to manifestation. An exposure theory makes sense, AHP contends, because it recognizes that manufacturers intend to protect themselves against the long-term risk of claims associated with insidious diseases; "personal injury, sickness or disease" should therefore be read to cover every potentially injurious exposure during the policy period that results at any future time in a claim for injury. If an exposure theory were applied, Liberty would be responsible to pay any liability AHP incurs in any suit against AHP where the plaintiff ingested the harmful product during the policy period, irrespective of how long after termination a diagnosable injury, sickness, or disease developed, or manifested itself.

Several courts have construed the CGL terms to provide coverage upon exposure. The Sixth Circuit, in Insurance Co. of North America v. Forty-Eight Insulations, Inc., 633 F.2d 1212, 1222 (6th Cir.1980), reh'g granted in part and denied in part, 657 F.2d 814 (1981), cert. denied, 454 U.S. 1109, 102 S.Ct. 686, 70 L.Ed.2d 650 (1981), found the terms "bodily injury" and "occurrence" inherently ambiguous as applied in the "progressive disease context." The Court, faced with determining an asbestos manufacturer's coverage under a CGL-derived policy, relied on that ambiguity, on medical testimony of the progressive nature of diseases caused by prolonged exposure to asbestos, and on the presumed intent of the parties, to read the contract as providing coverage for all potentially injurious exposures. See also Porter v. American Optical Corp., 641 F.2d 1128 (5th Cir.), cert. denied, 454 U.S. 1109, 102 S.Ct. 686, 70 L.Ed.2d 650 (1981). In Keene v. Insurance Co. of North America, 667 F.2d 1034 (D.C.Cir.1981), cert. denied, 455 U.S. 1007, 102 S.Ct. 1644, 71 L.Ed.2d 875 (1982), the District of Columbia Circuit also found that exposure to asbestos triggered coverage under the "ambiguous" CGL terms. It concluded that an exposure theory most closely approximated the reasonable expectation of manufacturers, who purchase insurance in order to gain certainty and be free from all risk of liability arising out of products-liability suits. To protect this alleged expectation, the Court held that insurance coverage was triggered not only by every exposure, but by "exposure in residence," when asbestos fibers were present in the body and causing further injury, and by every manifestation of an injury, sickness, or disease. 667 F.2d at 1048.1

Liberty counters with the claim that the relevant policy provisions unambiguously define manifestation as the trigger of insurance coverage. The average person, Liberty argues, would understand the terms "injury" and "disease" to mean an "abnormal condition" that "occurs" when it becomes manifest and is or should be discovered by the claimant. Thus Liberty argues that, under the policy, it incurs no duty to defend unless the claimant seeks recovery for an injury, sickness, or disease that allegedly became manifest while the policy was in effect, and that it has no duty to pay AHP sums for which AHP becomes liable unless the alleged injury became manifest during coverage.

Some courts have indeed rejected the exposure theory and found that, properly construed, the CGL provisions support a manifestation theory. See, e.g., Eagle-Picher Industries, Inc. v. Liberty Mutual Insurance Co., 523 F.Supp. 110 (D.Mass.1981), modified, 682 F.2d 12 (1st Cir.1982); American Motorists Insurance Co. v. E.R. Squibb & Sons, Inc., 95 Misc.2d 222, 406 N.Y.S.2d 658 (N.Y.Sup.Ct.1978). They reason that manufacturers and insurance companies intend to cover only risks that ripen into claims, and that "bodily injury" in the CGL cannot logically be read to mean injurious exposures that are not themselves injuries. Thus, District Judge Zobel, addressing policy coverage for asbestos related diseases, cogently reasoned:

Eagle-Picher Industries, 523 F.Supp. at 114. On appeal, the First Circuit agreed that "[t]he policies clearly distinguish between the event which causes injury — the accident or exposure — and the resulting injury or disease. Yet ... it is the resulting injury, not the exposure, which must take place `during the policy period' in order to trigger coverage ...." Eagle-Picher Industries, 682 F.2d at 17 (emphasis in original). Liberty relies on these cases in seeking denial of AHP's motion for summary judgment; it argues for partial summary judgment, moreover, based on the special limiting clause in its policies after 1968, which it claims incontrovertibly excludes coverage for several of the claims referred to in this action.

Insofar as AHP and Liberty attack each other's construction, both sides are correct: neither the exposure nor the manifestation theory can be wholly justified by the policy language.

A. Governing Legal Principles

AHP correctly argues that insurance contracts must be liberally construed, with ambiguities in the policy language resolved in favor of the insured. See Pan American World Airways, Inc. v. Aetna Casualty & Surety Co., 505 F.2d 989, 999 (2d Cir.1974); Breed v. Insurance Co. of North America, 46 N.Y.2d 351, 353, 385 N.E.2d 1280, 1282, 413 N.Y.S.2d 352, 354 (1978). This rule of construction, in fact, appears to be the single factor that unifies the discordant opinions applying the CGL and its derivatives to insidious diseases. Whether or not explicitly finding ambiguities, and irrespective of the construction adopted, every court construing these provisions has reached the result that extended coverage to the insured. See, e.g., Eagle-Picher Industries, 682 F.2d at 17 (manifestation); Keene, 667 F.2d at 1041 (manifestation and exposure); Forty-Eight Insulations, 633 F.2d at 1223 (exposure).

Under New York law, however, a court should apply its own construction to a policy's terms only after it has exhausted every effort to derive the meaning of the terms that accurately reflects the intent of the parties. An insurance policy is a contract which, like any other contract, must be construed to effectuate the parties' intent as expressed by their words and purposes. McGrail v. Equitable Life Assurance Society, 292 N.Y. 419, 424, 55 N.E.2d 483, 486 (1944); MetPath Inc. v. Birmingham Fire Insurance Co., 86 A.D.2d 407, 449 N.Y.S.2d 986 (1st Dep't 1982). If the words of the contract are unambiguous, establishing only one meaning when read in the context of the entire policy, then a New York court must enforce the plain meaning of the words and refrain from making or varying "the contract of insurance to accomplish its notions of abstract justice or moral obligation." Breed, 46 N.Y.2d at 355, 385 N.E.2d at 1283, 413 N.Y.S.2d at 355; see Coppotelli v. Insurance Co. of North America, 484 F.Supp. 1327, 1329 (E.D.N.Y.1980).

The principle that requires courts to resolve ambiguities in insurance contracts in favor of the insured, and against the insurer, is not a fundamental departure from established contract principles. Rather, it is one instance, albeit special and emphatic, of the widely recognized rule that resolves ambiguity in a contract against the contract's drafter. Greaves v. Public Service Mutual, Inc., 5 N.Y.2d 120, 155 N.E.2d 390, 181 N.Y.S.2d 489 (1959). Insurance contracts are usually contracts of adhesion in that their terms are generally dictated rather than negotiated. Courts therefore require insurance companies to be clear and unambiguous in creating limitations on coverage. "[T]he insurer has the burden of establishing that `the words and expressions used not only are susceptible of [the] construction [that the insurer advocates], but that it is the only construction that can fairly be placed thereon.'" Bronx Savings Bank v. Weigant, 1 N.Y.2d 545, 552, 136 N.E.2d 848, 851, 154 N.Y.S.2d 878, 883 (1956) (quoting Hartol Products Corp. v. Prudential Insurance Co., 290 N.Y. 44, 49, 47 N.E.2d 687, 690 (1943)). Accord, Vargas v. Insurance Co. of North America, 651 F.2d 838, 839 (2d Cir.1981).

Nevertheless, intent remains controlling. No construction of an ambiguous provision is permitted that is inconsistent with the contract's plain meaning or with the parties' clear intentions. As the Second Circuit stated in Union Insurance Society of Canton, Ltd. v. William Gluckin & Co., 353 F.2d 946 (2d Cir.1965):

353 F.2d at 951 (citation omitted). To disregard express language in an insurance contract because of a claimed ambiguity "would violate the more fundamental rule of construction" that requires a court to construe the contract as a whole and, whenever possible, give effect to all of its parts. Spencer, White & Prentis, Inc. v. Pfizer, Inc., 498 F.2d 358, 363 n. 23 (2d Cir.1974). New York law therefore requires a construction of the CGL language based on the plain meaning of the words employed, read in the context of the policy as a whole, the purposes sought to be accomplished, and the relevant surrounding circumstances. See Murray Oil Production v. Royal Exchange Assurance Co., 21 N.Y.2d 440, 445, 235 N.E.2d 765, 288 N.Y.S.2d 618 (1968); Vanguard Insurance Co. v. Polchlopek, 18 N.Y.2d 376, 222 N.E.2d 383, 275 N.Y.S.2d 515 (1967).

B. Proposed Constructions of the Parties

Neither construction proffered by the parties is supported by the plain meaning of the terms employed.

1. The Exposure Theory. An exposure theory is inconsistent with the policies' plain meaning, and AHP has offered no evidence or explanation to demonstrate how an "occurrence" could logically include every exposure to the substances it manufactures. The policies require that the resulting injury and not the exposure occur "during the policy period." Moreover, the policies were designed to protect against liability from law suits brought because of compensable injuries. To that end they cover "occurrences" wherein both exposure and an injury take place.

The only New York case on point also rejected the exposure theory as unsupported by the language of a policy that covered "[a]n accident or injurious exposure to conditions which results, during the policy period, in bodily injury...." The injured plaintiffs, who had sued the insured, were the daughters of women who had ingested DES while pregnant with the plaintiffs in 1952, 1953, and 1961 respectively; the plaintiffs thereafter developed cervical cancers that were discovered in 1970, 1971, and 1975. Justice Greenfield concentrated on the policy language:

American Motorists Ins. v. E.R. Squibb & Sons, 95 Misc.2d 222, 406 N.Y.S.2d 658, 659-60 (N.Y.Sup.Ct.1978) (emphasis in original).

A limited version of the exposure theory, adopted in some litigations, is at least linguistically respectable. Some courts, relying on medical evidence, have found that, on exposure, asbestos particles enter the body and cause discrete injuries to lung and other tissue, and that these injuries are sufficient to establish coverage under the CGL language, even though they must aggregate over time to cause diseases and sicknesses such as asbestosis, carcinoma, and mesothelioma. Keene, 667 F.2d at 1044; Insurance Co. of North America v. Forty-Eight Insulations, Inc., 451 F.Supp. 1230, 1239 (E.D.Mich.1978), modified, 633 F.2d 1212, 1222-23 (6th Cir.1980), cert. denied, 454 U.S. 1109, 102 S.Ct. 686, 70 L.Ed.2d 650 (1981). But see, e.g., Eagle-Picher, 523 F.Supp. at 115 & 682 F.2d at 17 (medical evidence establishes that (1) not every exposure to asbestos results in injury of any sort, since body's natural mechanisms often remove fibers before they become embedded; and (2) even when discrete injuries are caused, they frequently fail to lead to any compensable injury).

AHP cannot rely on the rationale of cases that have found immediate injury from the ingestion of asbestos fibers, because the drugs at issue in this case differ markedly from asbestos in the manner in which they are alleged to injure humans. AHP has failed to submit any proof with respect to the effects of any of the drugs involved, and has not claimed that any of them injures upon every exposure. The record establishes without material dispute, moreover, that at least two of the drugs at issue (Ovral and DES) do not injure upon every exposure. Flessa Aff't, Def.Ex. 35 (Feb. 14, 1982) (Ovral); Mattingly Aff't, Def.Ex. 36 (Feb. 10, 1982) (DES). As discussed below, a particular plaintiff might be able to establish, despite this evidence, that a particular exposure to one of the drugs at issue constituted an actual and compensable injury. No evidence has been presented, however, that could support a general declaration that every exposure to any of the drugs at issue in this case causes injury, and therefore triggers coverage, under the AHP policies.

Advocates of the exposure theory have suggested that it has some inferential support in the New York Court of Appeals' holdings that an injury occurs upon exposure for purposes of the statute of limitations. See, e.g., Steinhardt v. Johns-Manville Corp., 54 N.Y.2d 1008, 430 N.E.2d 1297, 446 N.Y.S.2d 244 (1981). This rule can be traced back to decisions that appear to have assumed that exposure to harmful chemicals does in fact injure bodily tissue. See, e.g., Thornton v. Roosevelt Hospital, 47 N.Y.2d 780, 781, 391 N.E.2d 1002, 1004, 417 N.Y.S.2d 920, 922 (1979). But these statute-of-limitation decisions do not in fact turn on the untenable proposition that actual injury always occurs upon the ingestion of asbestos fibers, or other bodily invasion. They rest, rather, on the costs and uncertainty that the Court of Appeals long ago concluded would be created by a general rule that only "discovery" triggers a statute of limitation, and on the state legislature's failure to amend the applicable New York statute to extend the time in which plaintiffs may bring suits for the consequences of insidious diseases, other than those caused by "Agent Orange." See Schwartz v. Heyden Newport Chemical Corp., 12 N.Y.2d 212, 218-19, 188 N.E.2d 142, 144-45, 237 N.Y.S.2d 714, 718-19 (1963); N.Y.C.P. L.R. § 214-a (McKinney Supp.1982). These policy considerations have no bearing on the construction of insurance contracts; indeed, every court that has examined CGL policies has refused to give weight to definitions of injury articulated in statute-of-limitation cases. See, e.g., Keene, 667 F.2d at 1043 & n. 17; see generally Wrubel, Comment, Liability Insurance for Insidious Disease: Who Picks Up the Tab?, 48 Fordham L.Rev. 657, 669-70 (1980).

Finally, the exposure theory is made particularly untenable in this case by the special limitation added to the AHP-Liberty policies since at least 1968, which excludes coverage for injuries "caused by continuous or repeated exposure, any part of which occurred after termination of the policy." If every exposure were an occurrence under the policies, the limitation would have no meaning; it can have meaning only if the occurrences insured against are the injuries, diseases, or sicknesses caused by exposures, and not the exposures themselves.

2. The Manifestation Theory. Liberty contends that the policy words "injury", and "disease" or "sickness" are ordinary words which the average person would understand respectively to mean "damage" and an "abnormal condition," and that the average person would think of these consequences as having occurred when they take place or are discovered. The First Circuit accepted this view in extending coverage to the claims at issue in Eagle-Picher:

682 F.2d at 19 (footnote omitted).

The "manifestation" theory adopted in Eagle-Picher, and now advanced by Liberty, is in part a departure from what has conventionally been understood as the manifestation approach. The meaning of manifestation proposed in prior cases is that an injury, sickness, or disease becomes manifest only when symptoms become noticeable or a diagnosis is made. "The manifestation theorists contend that the date of manifestation is `the date on which the condition became known or should have become known to plaintiff or the date on which plaintiff's condition was medically diagnosed, whichever comes first.'" Wrubel, supra, 48 Fordham L.Rev. at 668 n. 58 (quoting Forty-Eight Insulations, 451 F.Supp. at 1238). Thus stated, the manifestation theory refuses to recognize that any bodily injury may have existed prior to the appearance of symptoms or an actual diagnosis. Under this construction, the manifestation theory would be convenient to apply, but it is inconsistent with the policy language.

The ordinary person may construe an "occurrence" of injury to mean manifestation in the sense of discovery. Discovery of an injury or disease is a truly significant event which makes the victim aware of what had theretofore been only a latent, medical problem without conscious significance. The plain meaning of the policy language is not measured, however, by the understanding of a lay person, but by the understanding of a person engaged in the insured's course of business. See Champion International Corp. v. Continental Casualty Co., 546 F.2d 502, 505-06 (2d Cir. 1976); McGrail v. Equitable Life Assurance Society, 292 N.Y. 419, 424-25, 55 N.E.2d 483, 486 (1944); Loblaw, Inc. v. Employers' Liability Assurance Corp., 85 A.D.2d 880, 446 N.Y.S.2d 743, 745 (4th Dep't 1981). From the point of view of a drug manufacturer, familiar with the potential development of insidious diseases from its products, and seeking to insure against liability for harm, an injury, sickness, or disease would include any compensable, medical condition that is fully developed, even though dormant.

For example, a particular drug may cause a heart attack in some women long after they ingest it. Under the manifestation theory, liability for the heart attack would be covered by a CGL-type policy in effect when the injury occurs. But the manufacturer would also expect that a latent, undiscovered disease caused by the same drug prior to the heart attack would be covered by a policy in effect at the time it arose, regardless of when the disease was discovered or the heart attack or some other manifestation occurred. The CGL policy language covers all injuries, sicknesses, or diseases that occur during coverage, not merely those that become manifest.

The restriction imposed on the policy language by the manifestation approach is, moreover, inconsistent with insurance law principles. If the policy language were deemed ambiguous enough to permit reading in a manifestation requirement, the process of doing so would be unacceptable in New York because that reading is not "the only construction which may fairly be placed on" the policy's words, Filor, Bullard & Smyth v. Insurance Co. of North America, 605 F.2d 598, 602 (2d Cir.1978), cert. denied, 440 U.S. 962, 99 S.Ct. 1506, 59 L.Ed.2d 776 (1979) (emphasis omitted) (citing Lachs v. Fidelity & Casualty Co., 306 N.Y. 357, 365-66, 118 N.E.2d 555, 559 (1954)), and because the construction proposed would be more restrictive than one that treated as "occurrences" all injuries, diseases, and sicknesses, whether manifest or merely discoverable, see Thomas J. Lipton, Inc. v. Liberty Mutual Insurance Co., 34 N.Y.2d 356, 361, 314 N.E.2d 37, 357 N.Y.S.2d 705, 708 (1974); National Screen Service Corp. v. United States Fidelity & Guaranty Co., 364 F.2d 275, 277 (2d Cir.1966); Vargas v. Insurance Co. of North America, 651 F.2d 838, 839-40 (2d Cir.1981). Significantly, courts that have accepted the manifestation theory have invariably done so in extending, rather than restricting, the coverage of an ambiguous policy.

The limitation provision found in AHP's policies after 1968, and relied upon by Liberty to refute the exposure theory, also refutes the manifestation approach. The limitation excludes from coverage injuries "caused by continuous or repeated exposure, any part of which occurred after termination of the policy." If the standard-form, CGL language, standing alone, triggered coverage solely by manifestation or discovery, then this special provision would have no reasonable meaning, since all occurrences to which it could apply would have already become manifest during the policy period. No insured would conceivably agree to give up coverage for liability in cases where every necessary prerequisite for liability has occurred during the policy period, merely because the plaintiff continued thereafter to be exposed to the cause of his already manifest injury. To the contrary, the limitation provision suggests that, absent its application, the policy could arguably apply to injury, sickness or disease, caused in part by exposure during the policy period and in part thereafter. In protecting itself against such claims, Liberty used language that significantly informs the construction of the "occurrence" clause, rendering its manifestation theory as untenable as the same language renders the exposure theory suggested by AHP.

The manifestation theory is also advocated on the basis of decisions construing health and accident insurance policies, and worker compensation statutes. Courts apply the manifestation theory in construing health and accident policies in order to extend coverage to individuals who suffered from latent but unknown injuries or diseases prior to coverage, which only manifested themselves or were discovered during coverage. See, e.g., Silverstein v. Metropolitan Life Insurance Co., 254 N.Y. 81, 84, 171 N.E. 914, 914-15 (1930). This result is reached, moreover, because the parties to such insurance contracts are concerned with the disabling results of injuries and illnesses, and not their medical ascertainability. In a typical health insurance case, Cohen v. North American Life & Casualty Co., 150 Minn. 507, 508, 185 N.W. 939, 939 (1921), for example, the Supreme Court of Minnesota explained: "The insurer intended to give and the insured expected to get indemnity if illness or sickness [occurred] .... They were not concerned with the latent or remote cause of it. The cause of the disease insured against is not mentioned in the policies as a condition of liability." By contrast, the CGL language and AHP's policies are contracts in which the risk insured against is not illness itself, but liability resulting from the manufacture or sale of a product, and in which the parties are expressly concerned with diseases that are caused over time by injurious exposure, and hence often exist in a latent form before discovery:

Wrubel, supra, 48 Fordham L.Rev. at 671 (footnotes omitted).

Resort to the manifestation theory in construing worker compensation laws provides no support for the theory's application in construing the private contracts at issue in this case. In that context, and with characteristic flair, Learned Hand said that "a disease is no disease until it manifests itself." Grain Handling Co. v. Sweeney, 102 F.2d 464, 466 (2d Cir.), cert. denied, 308 U.S. 570, 60 S.Ct. 83, 84 L.Ed. 478 (1939). The claimant's disease (dormant tuberculosis) long antedated his last employment, but Judge Hand treated it as having occurred when it disabled the claimant. As Judge Hand noted, however, the Longshoremen's Act which he was interpreting "is not concerned with pathology, but with industrial disability," id., and the legislature's purpose in passing such legislation was to make the costs of all disabling injuries and illnesses a part of the cost of production, id. at 465. Similarly, in Travelers Insurance Co. v. Cardillo, 225 F.2d 137 (2d Cir.), cert. denied, 350 U.S. 913, 76 S.Ct. 196, 100 L.Ed. 800 (1955), the Second Circuit held that under the same Act a disabled employee's last employer would be entirely liable for the individual's hearing disability, where it was impossible to determine how much hearing loss resulted from any prior employment. 225 F.2d at 144. Accord, General Dynamics Corp. v. Benefits Review Board, 565 F.2d 208 (2d Cir.1977). In this situation, coverage was legislatively mandated; the Court's problem was to determine which employer's fund should be tapped to pay the benefits required. By contrast, coverage under the CGL language depends upon tort liability, not legislative mandate, and is supposed to extend only to those cases upon which the contracting parties agreed, not to a point that is convenient or socially desirable. The practical, legislatively supported rules adopted by the Second Circuit in worker compensation cases are, therefore, no proper basis for resolving the private, contractual disputes presented here; and the language and purpose of the contracts at issue here do not permit a manifestation theory that treats as insignificant the development from injurious exposure of a latent but real illness.

C. Plain Meaning of the Policy Provisions.

1. The Occurrence Clause. The plain meaning of the "occurrence" clause is no secret to the parties. Courts and writers have recognized that "occurrence" is most logically construed to include only those injuries, sicknesses, or diseases that are proved to have existed during coverage. For example, in Forty-Eight Insulations the Sixth Circuit stated:

633 F.2d at 1217 (footnote omitted); see also Schering Corp. v. Home Insurance Co., 544 F.Supp. 613, 616 (E.D.N.Y.1982). See generally Wrubel, supra. Indeed, while Liberty has advanced its position as based on a manifestation theory requiring knowledge or actual diagnosis, Liberty at times defines "manifestation" to include the unknown presence of diagnosable and compensable disease. See, e.g., Portmann Dep., Pl.Ex. 3 at 136, 138; Stevens Dep., Pl.Ex. 8 at 113-17.

This straight-forward interpretation has been rejected, however, as impractical and unfair. According to critics, to require the courts and the parties to CGL-type contracts to determine coverage in the many thousands of pending and anticipated latent-injury claims on a case-by-case basis would be impossible, e.g., Forty-Eight Insulations, 633 F.2d at 1218, and also inconsistent with the reasonable expectations of the insured, and with developing tort-law doctrine in the latent-disease area, e.g., Keene, 667 F.2d at 1044 n. 20; Forty-Eight Insulations, 633 F.2d at 1219. These practical and ethical concerns are shown below to be unpersuasive. Before addressing them, however, a full appraisal of the policy's plain meaning is in order to demonstrate why, linguistically at least, the injury-in-fact approach is plainly the proper construction.

The most basic demand of the policy language is that to establish Liberty's liability the insured must prove that an "occurrence" — injury, sickness, or disease — arose during the policy period. The plain language demands that the insured prove the cause of the occurrence (accident or exposure), the result (injury, sickness, or disease), and that the result occurred during the policy period. An exposure that does not result in injury during coverage would not satisfy the policy's terms. On the other hand, a real but undiscovered injury, proved in retrospect to have existed at the relevant time, would establish coverage, irrespective of the time the injury became manifest.

This approach is faithful to the policy language because it gives separate meaning to the three concepts of exposure, injury, and discovery. The policy expressly gives separate significance to the first two concepts, and its provisions strongly suggest that an occurrence means something different — and more expansive — than manifestation or discovery. On the other hand, nothing in the policy language precludes a finding that a single exposure, or a single period of exposures, immediately injured a person to a compensable extent; similarly the policy language also permits courts or juries to find that injury, sickness, or disease, occurred in a particular case at the time it became manifest. So long as the insured is held liable for an identifiable and compensable injury, sickness, or disease that is shown to have existed during coverage, that liability will be insured against whether or not the injury coincides with exposure or manifestation.

To state this result in the abstract is, of course, much easier than to apply it in particular contexts, given the variables involved. Injuries that are diagnosable and compensable may be established in particular cases at almost any point in the period between exposure and manifestation, depending among other things upon the nature of the substance involved, the intensity (amount or duration) of the exposure, the capacity of the individual exposed to withstand the danger, and chance. For example, although exposure to asbestos does not usually injure seriously enough to constitute an "occurrence" in the context of a liability insurance policy, a finder of fact might, nevertheless, find that a particular exposure or period of exposure contemporaneously caused a compensable injury, sickness, or disease during coverage. If an expert testified that, in his opinion, formed to a degree of reasonable medical certainty, a compensable injury resulted from the particular, alleged exposure, a finding could properly be reached that compensable injury was caused by that exposure, even if it later leads to other, compensable injuries that occur after coverage has ended. Similarly, the manifestation of a long-latent injury or illness may itself constitute an injury, sickness, or disease. For example, the manifestation of asbestosis or mesothelioma, although previously developed but undiscovered, may also constitute an occurrence if accompanied by some additional injury, such as the coughing of blood, or some other identifiable and compensable complication. In sum, the insured would be permitted under a plain reading of the policy to establish that occurrences have taken place at any or all points from exposure to manifestation, on the ground that identifiable injuries have been proved to have occurred at each point to a reasonable degree of medical certainty.

The language remains clear even though its application is complicated by the many, strong variables at work in the insidious disease process. The general rule applied in health and property damage insurance cases provides a strong analogy. In those areas, courts construe language such as "injury", "damage", and "accident" to occur at any point in time that a compensable event could have been or was ascertained. The indicated presence of silicosis was held to be "bodily injury" in Tomnitz v. Employers' Liability Assurance Corp., 343 Mo. 321, 121 S.W.2d 745 (1938), and therefore compensable under a liability policy in effect at that time, even though the disease manifested itself through symptoms only after the policy had expired. Similarly, casualty insurance coverage is extended to a home "damaged" by slow-moving causes, such as a shifting foundation or an attack of termites, as soon as the effect of those causes is shown to have inflicted measurable, compensable, albeit unascertained, damage on the policyholder during the policy period. See, e.g., Harman v. American Casualty Co., 155 F.Supp. 612 (S.D.Cal.1957); Snapp v. State Farm Fire & Casualty Co., 206 Cal.App.2d 827, 24 Cal.Rptr. 44 (1962). Here, again, as in the private contracts in this case, compensable injury may be shown to occur in particular cases at any time from initial exposure until manifestation.

2. The Limitation Provision. The special limitation, appearing first in the AHP-Liberty policy for 1964 and made effective in 1968, provides strong support for this reading of the plain language.2 Liberty argues correctly that the provision is unambiguous, but its meaning is not the one Liberty suggests. Rather, on its face the provision removes from coverage only injury, death, or destruction that is caused by continuous or repeated exposure occurring in part after November 1, 1976, the date AHP terminated the policy. It does not remove from coverage harm caused wholly by exposure occurring prior to November 1, 1976. Nothing in the language of the provision takes such harm out of the policy simply because exposure to the same products continued after November 1, 1976, or indeed because further harm may have occurred from such continued exposure. If, in a particular case, pre-November 1, 1976 exposure is found to have caused a particular harm, no exposure after November 1, 1976 would be a cause of that particular harm; such harm is not "caused by ... exposure any part of which" occurred after November 1, 1976, because all of that harm-causing exposure occurred before that date.

Plaintiff and defendant agree that there is no evidence about who wrote the limitation provision, when it was written, or when it first appeared in AHP policies. The 1952 policy does not contain the provision, the policies from 1964 to 1976 do contain it, and the parties have been unable to locate the policies from 1953 to 1963. Moreover, no reliable evidence exists about the original purpose of the provision or about the parties' understanding of its meaning, prior to November 1, 1976. Nor has any possibly relevant evidence been submitted about whether the addition of the contested provision was reflected in premium amounts. Cf. Pan American World Airways, Inc. v. Aetna Casualty & Surety Co., 505 F.2d 989, 1003 (2d Cir.1974) (recognizing that amount of premium varies with scope of an insurance policy's coverage). Some evidence does exist, however, in the depositions, and as discussed below it supports the interpretation of the limiting provision suggested here, as well as the proposition that the parties intended and understood basic policy coverage to apply to injuries shown in fact to have occurred during the policy period.

3. Obligation to Defend. The plain language of the policies at issue makes clear that Liberty is obliged to defend any suit, however meritless, seeking to impose a liability upon AHP that is covered by the policy. The policies provide:

E.g., Pl.Ex. 20, Sec. II. This expansive language must, under New York law, be construed broadly to provide coverage. See, e.g., Chertok v. Hotel Salisbury, Inc., 516 F.Supp. 766, 767 (S.D.N.Y.1981); Goldberg v. Lumber Mutual Casualty Insurance Co., 297 N.Y. 148, 77 N.E.2d 131 (1948); 31 N.Y.Juris. § 1324 (1963).

Liberty concedes the principle that the duty to defend is broader than the duty to indemnify, in that it must defend meritless suits, so long as they could conceivably result in a liability covered by the policy. But Liberty relies upon the equally valid principle that it has no duty to defend claims which may have merit, but for which liability coverage does not exist because injury occurred outside the policy period. See Appleman, Insurance Law Practice § 4684 at 72 (1979). New York's Court of Appeals has established that "if we can determine that no basis for recovery within the coverage of the policy is stated in the complaint, we may sustain defendant's refusal to defend." Lionel Freedman, Inc. v. Glens Falls Insurance Co., 27 N.Y.2d 364, 368, 267 N.E.2d 93, 95, 318 N.Y.S.2d 303, 305 (1971). Furthermore, even if a complaint states a basis for possible recovery, once an insurer is able to "confine the claim" — exclude the possibility of a recovery for which it has provided insurance — the insurer has no further duty to defend. See Lee v. Aetna Casualty & Surety Co., 178 F.2d 750, 752-53 (2d Cir.1949) (L. Hand, J.).

If manifestation in the form of discovery were required to trigger coverage under AHP's policy, then its motion for summary judgment with respect to the obligation to defend would have to be denied. But the injury-in-fact standard adopted here permits a declaration in AHP's favor requiring Liberty to defend until it has "confined the claim" in any underlying case by demonstrating that coverage will be denied even if the case succeeds. As Liberty itself argued in another suit in this District, an insured is entitled to a defense if the underlying complaint "permits proof" of facts establishing coverage, or, as in this case, if the complaint does not exclude the possibility that injury, sickness, or disease in fact occurred during the policy period. See Pl.Ex. 42 at 16-17; Emons Industries, Inc. v. Liberty Mutual Fire Insurance Co., 481 F.Supp. 1022 (S.D.N.Y.1979). This showing must be made in each underlying litigation, and is an inappropriate issue to resolve in a declaratory judgment action, where the underlying cases are not before the Court. So long as the allegations in a case could conceivably result in liability covered by a policy at issue, Liberty must defend. Tenneco Chemical, Inc. v. Employers Mutual Liability Insurance Co., 1968 Fire & Casualty Cas. (CCH) 984, 985 (S.D.N. Y.1968) (Tyler, J.).

This standard also applies to those suits that Liberty claims are controlled by the special limitation provision. So long as a complaint permits proof that an occurrence took place within the coverage period, Liberty must defend the suit until and unless it is able to establish, not only that post-1976 exposures took place, but also that no compensable injury was diagnosable before those exposures commenced.

II. Background of the CGL and Intention of the Parties

Both AHP and Liberty claim that the background of the CGL — its history and purpose — lends support to their respective theories. AHP argues that the CGL's drafters indisputably rejected the manifestation theory, and thereby signalled an intent consistent with an exposure approach. Liberty replies that the "legislative history" of the CGL is irrelevant here, and that it in any event involves issues of fact that cannot be resolved without an evidentiary hearing. Furthermore, Liberty contends, the evidence relied upon by AHP does not support an exposure theory, and the relationship between the parties — especially AHP's acceptance of the provision limiting coverage to injuries resulting only from exposure during coverage, and the swine-flu policy adopted by the parties in 1976 — refutes the notion that the parties ever intended to adopt an exposure theory.

The background of the CGL, and the relationship that existed between these parties, are relevant in supplementing the policy's plain meaning.3 Union Insurance Co. of Canton, Ltd. v. William Gluckin & Co., 353 F.2d 946, 951 (2d Cir.1965). Where the policy language is unambiguous, a court in New York may not look to circumstances external to the contract to alter its meaning. See, e.g., Breed, 46 N.Y.2d at 355, 385 N.E.2d at 1283, 413 N.Y.S.2d at 355; Teal v. Place, 85 A.D.2d 788, 445 N.Y.S.2d 309, 312 (3d Dep't 1981). Parol evidence can provide strong corroborative proof, however, of the parties' intent as suggested in their words. See Theatre Guild Productions, Inc. v. Insurance Corp. of Ireland, 25 A.D.2d 109, 267 N.Y.S.2d 297, 299 (1st Dep't 1966), aff'd, 19 N.Y.2d 656, 225 N.E.2d 216, 278 N.Y.S.2d 625 (1967). The evidence here shows, as the parties respectively contend, that neither the exposure nor the manifestation theory was intended. Rather, the evidence establishes that the CGL draftsmen, as well as the parties, contemplated coverage only for injuries shown in fact to have occurred during periods of coverage.

A. History and Purpose of the CGL.

The CGL evolved out of the difficulties faced by courts and parties in dealing with personal injuries and property damage sustained as a result of gradual processes. Prior to 1966, general liability policies covered liability "because of bodily injury, sickness or disease, including death at any time resulting therefrom, sustained by any person, caused by accident and arising out of the hazards hereinafter defined." 1 R. Long, Law of Liability Insurance § 11.01 at 11-4 (1979) (emphasis added). The word "accident" suggested an intent to cover only sudden, unexpected, but identifiable events. The courts were left in doubt as to whether, and to what extent, the standard policy was meant to cover liability for injuries that resulted from gradual processes, rather than from sudden events.

The insurance industry responded to the uncertainty created by the "accident" orientation of pre-1966 liability policies by establishing a task force to draft what eventually became the CGL. In 1960, the Mutual Insurance Rating Bureau ("Mutual") and the National Bureau of Casualty Underwriters ("National") established several committees, which engaged in the revision process. Liberty, a member of Mutual, became actively involved; Richard Schmalz, then an attorney for Liberty, was a member of the Joint Drafting Committee, and Gilbert Bean, a Liberty officer, became a member of the Joint Rating Committee, which ultimately approved the form policy. Schmalz Dep., Pl.Ex. 23 at 16. The task force substituted the "occurrence" approach now in the CGL and in AHP's policies for the "accident" approach, and it expressly provided that an occurrence included any injury or damage that resulted, not only from an accident, but also from injurious exposure over an extended period. This change adopted the result reached by courts that construed "accident" to include injuries resulting from long exposures.

The new CGL failed, however, to resolve definitively the time injury should be found to occur in insidious disease cases. The CGL, like its predecessor, requires only that an injury take place during the coverage period. But, since the injuries expressly covered under the CGL include those resulting from long-term injurious exposure, the difficulty of determining time of injury was certain to be even greater under the CGL than it had been under predecessor policies. The insurance industry task force recognized this problem from the very outset of its labors, just as courts had recognized by then that an injury could occur in scientific fact long before it became manifest. See, e.g., Tomnitz v. Employers' Liability Assurance Corp., 343 Mo. 321, 121 S.W.2d 745 (1938). Nevertheless, the task force refused to alter the standard policy language to define precisely the time of injury; in the process, moreover, they refused to accept language that would have incorporated into the CGL either the manifestation or the exposure theory. Schmalz Dep., Pl.Ex. 23 at 65-69.

Substantial evidence supports the view that the CGL draftsmen rejected the manifestation theory as a limitation on coverage. One of the first drafts, prepared by the Joint Drafting Committee in 1960, "would deem all of the injury to have occurred" at the first manifestation of injury or damage. That draft was sent back to the Drafting Committee by the higher level Joint Rating Committee, and Mr. Schmalz has testified without contradiction that this remand was a rejection by the task force leadership of the manifestation approach. Id. at 42, 64-65. The various committees rejected all other drafts that included a manifestation concept, see id. at 42; they found the concept unacceptable because, among other reasons, "it would permit a carrier to cancel following the first notice of injury and leave the insured without coverage for other injuries emerging from the same exposure," and because in many cases injuries sufficiently serious to trigger coverage could occur prior to any form of manifestation. Memorandum: "Action of the Liability Rating Committee of the Mutual Bureau Meeting" (Sept. 12-14, 1961), Pl.Ex. 25 at 1; see also Pl.Ex. 26 (Minutes of Meeting of Joint Scope of Coverage Subcommittee (Sept. 19-20, 1961)) at 3; Wrubel, supra, at 684-85.

Manifestation-theory advocates argued that a nonmanifestation policy would permit insureds to pyramid the limits on policies that cover injuries caused by long-term exposure. See Schmalz Dep., Pl.Ex. 23 at 78-80. Rejection of the manifestation concept in the CGL was thus implicit in its drafters' widespread recognition that the CGL would permit the insured to pyramid the limits of multiple policies when injuries occurred in more than one policy period. See, e.g., Elliott, The New Comprehensive General Liability Policy, in Liability Insurance Disputes 12-5 (S. Schreiber ed. 1968) ("In some exposure types of cases involving cumulative injuries, it is possible that more than one policy will afford coverage."). Gilbert Bean, a Liberty officer who served on the Joint Rating Committee, distributed a memorandum to Liberty underwriters in July 1966 that explained how injuries could occur at different points under the CGL:

Pl.Ex. 29 at 3.

Liberty contends that the history of the CGL cannot be relied upon by AHP, because material issues of fact exist as to what the task force intended. But Liberty has failed to raise a material issue of fact by offering evidence to contradict the view that the CGL drafters rejected the manifestation approach. On the other hand, Liberty correctly argues that the CGL's history does not permit AHP to succeed on its claim that the task force adopted the exposure theory. Indeed, the evidence from the CGL drafting process is almost as clearly against an exposure theory as it is against manifestation.

Once again, Mr. Schmalz casts important light on what occurred. Soon after the Joint Rating Committee remanded the draft incorporating a manifestation limitation to the Joint Drafting Committee, Schmalz circulated a memorandum to his colleagues reviewing alternatives to manifestation. One of his suggestions for dealing with continuous or repeated exposure situations was what he called the "contact with the means of injury" approach. Recognizing that the approach was "novel" for insurance coverage, he suggested specific language to implement the proposal, which would have provided insurance for any "physical contact with the means of injury" during a period of coverage, even if the actual injury did not result until much later. Def.Ex. 37 (Schmalz Memorandum); see Schmalz Dep., Def.Ex. 22 at 118-19. The Committee rejected this language, however, adopting instead the CGL's coverage of any injury that results during the policy period, irrespective of when "the contact with the means of injury" — the exposure — took place.

Mr. Schmalz' own understanding of how the CGL was to operate lends strong support to the injury-in-fact approach suggested by the plain language. Schmalz believes that use of the insured's product during the policy period should trigger coverage under the CGL, but only if the policyholder is able to show that injurious biological effects occurred each time the product was used. Consistent with the clear language of the CGL, Schmalz construes it to require, not only an exposure, but an exposure that injures; and whether an injury has occurred in a given case is a "medical fact question" to be resolved on an ad hoc basis for each product and malady involved. Schmalz Dep., Def.Ex. 22 at 84. Thus he acknowledges that, while in some cases exposure and CGL "injury" may coincide, in many cases they will not. See id. at 163-65. The insurance industry task force intended, he says, that the CGL policyholder carry the burden of proving that the injury at issue occurred during a period of coverage, id. at 110, and he repeatedly testified that the issue of when an injury occurs "is a matter to be determined on the basis of facts relating to medical issues as to when injury occurred and factual issues as to when exposure took place," id. at 136-37, 151.

Mr. Schmalz is not the only person involved in the process of writing the CGL who viewed coverage as depending upon injury-in-fact, rather than on either manifestation or exposure. For example, Richard Elliott, then Secretary of the National Bureau of Casualty Underwriters, in presenting what he described as "the underwriting intent" of the CGL's revisers, explained that "the definition of occurrence serves to identify the time of loss for the purpose of applying coverage — the injury must take place during the policy period." He recognized that long exposure could involve "cumulative injuries," in which "more than one policy afford[s] coverage," and "each policy will afford coverage to the bodily injury or property damage which occurs during the policy period." Elliott, The New Comprehensive General Liability Policy, supra. Other contemporary explanations and analyses likewise state or assume that coverage under the CGL would require proof of injury during that period, not merely exposure and not necessarily manifestation. Thus, one commentator stated: "The policy will not depend upon the causative event of occurrence but will be based upon injuries or damages which result from such an event and which happened during the policy period. It will not be material whether the causative event happened during or before the policy period." Obrist, The New Comprehensive General Liability Insurance Policy — A Coverage Analysis 6 (Defense Research Inst. Monograph 1966). The same commentator defined "occurrence" "not only as an accident, but also exposure to conditions which may continue for days, weeks, months or even years. There is no `occurrence' until an injury results from a happening. Injury or damage must occur during the policy period to be covered...." Id. at 16 (emphasis omitted); see generally Wrubel, supra, 48 Fordham L.Rev. at 667 n. 50, 684 n. 152.

Thus, the background and purposes of the CGL terms, which are equivalent to the provisions of the AHP-Liberty policies, support the policies' plain meaning and make clear that the language used by the CGL drafters was not intended to make either exposure or manifestation serve as the trigger for coverage.

B. Practice and Course of Dealing.

Conduct of the parties provides another important source for deriving their intent as to the meaning of the insurance contracts at issue. If a court is genuinely interested in what the parties to a contract meant, "[t]here is no surer way to find out ... than to see what they have done." Brooklyn Life Insurance Co. v. Dutcher, 95 U.S. 269, 273, 24 L.Ed. 410 (1877); Webster's Red Seal Publications, Inc. v. Gilbertson World-Wide Publications, Inc., 67 A.D.2d 339, 415 N.Y.S.2d 229, 230 (1st Dep't 1979), aff'd, 53 N.Y.2d 643, 421 N.E.2d 118, 438 N.Y.S.2d 998 (1981). The New York courts have frequently resorted to the conduct of the parties in determining a contract's meaning, and insurance contracts are no exception to this practice. See Green v. Travelers Insurance Co., 286 N.Y. 358, 363, 36 N.E.2d 620, 622 (1941); American Stevedores, Inc. v. American Policyholders' Insurance Co., 138 N.Y.S.2d 513, 517 (N.Y.Sup.Ct.1955). In the present case, the available evidence establishes that the parties behaved consistently with the plain meaning and purposes of the policies, treating actual injury as the trigger for coverage, and inconsistently with either the exposure or manifestation theory.

Evidence on course of dealing can be ambiguous and misleading, especially in construing insurance contracts covering both liability for, and the duty to defend suits concerning, insidious diseases. Insured parties and insurers alike can be expected to behave according to their perceived self-interest by asserting in any particular litigation the theory that best suits their purposes. For example, where an insured had no coverage during the periods consumers were exposed to its products, but thereafter acquired coverage, it can be expected to advocate the manifestation theory; conversely, the insurance company in such a situation would advocate exposure as the policy trigger. Furthermore, an insurance company's willingness to provide coverage in a particular case may simply be based on its desire to avoid litigation and may provide only spurious support for the theory its behavior reflects. The vast majority of all products liability claims are settled without a suit being filed, and only about four percent of all claims ultimately go to verdict. Wrubel, supra, 48 Fordham L.Rev. at 685 n. 156 (reciting statistics from Insurance Services Office Products Liability Closed Claims Survey: A Technical Analysis of the Survey Results (1977)). Finally, the parameters of coverage under these liability policies are much broader with respect to the duty to defend than with respect to the obligation to compensate the insured for its ultimate liability. Consequently, given the breadth of the insurance company's duty to defend, its willingness to defend or settle virtually all claims based on injuries having any arguable relationship to the coverage period might only establish that it was acting on the basis of practical considerations.

In this case, however, the parties engaged in at least two negotiations apart from any ongoing litigation, that provide relatively reliable evidence that their policies incorporated neither the manifestation nor the exposure theory. In addition, the year-to-year process by which the parties established the rates to be paid under the policies at issue is highly probative of what they expected as businessmen when they extended the policies from year to year, and when they terminated coverage in 1976.

1. Revision applicable after 1968. AHP and Liberty agreed, some time before 1968, to a revision that limited recovery under all future policies to injuries that occurred only as a result of exposures that took place prior to termination of the policies.4 The significance of this change with respect to the plain meaning of the policies is discussed above. The change suggests that the parties proceeded during their relationship on a theory that anticipated an injury-in-fact during the coverage period to entitle the insured to relief; and the change made clear that no injury would be compensable if it resulted even in part from exposures outside the policy period.

The only available proof of AHP's understanding of the limitation comes from Mr. Pizl, who negotiated and coordinated AHP's insurance coverage. Mr. Hogarty, an independent party evaluating AHP's insurance needs, drew Mr. Pizl's attention to the limitation provision and expressed his own concern as to its possible meaning. He noted thereafter, however: "The insured does not feel that the exclusion relating to injury caused by continuous or repeated exposure any part of which occurs after the termination of the policy restricts the coverage that they intend to be provided under the Liberty policy." Hogarty Dep., Def.Ex. 3 at 71. AHP apparently understood the clause to be insignificant, because its coverage in any event extended only to those injuries that occurred under the policy. Similarly, Mr. Portmann, Special Claims Agent for Liberty, construed the provision as meaningless:

Portmann Dep., Pl.Ex. 3 at 99. Although Portmann used the word "manifestation", he defined it to mean "when you could have medical diagnosis, when it is first medically diagnosable, if you will." Id. at 136. He continued in the following colloquy:

Id. at 138. Mr. Portmann acknowledged that Liberty made no record of its intent when it first drafted this provision. But his disregard of the provision reflects an understanding that injury-in-fact establishes coverage.5

2. The Flu Vaccine Program of 1976. When the federal government commenced its flu vaccine program in 1976, AHP, which was one of four manufacturers involved, sought special coverage from Liberty and other insurers. Rather than have protection depend upon an injury during the coverage period, the parties agreed to make coverage depend solely upon use of the vaccine ("exposure") during the policy period: "This policy applies to personal injury or death whenever occurring arising out of the administration of swine flue vaccine during the Policy Period." Def.Ex. 16. Mr. Hogarty, who negotiated the policy on AHP's behalf, testified that the brokers who participated in this underwriting generally assumed that a change from the usual CGL format was necessary:

Def.Ex. 17 at 5. This successful effort to obtain the very sort of coverage AHP now seeks belies its theory that the policies at issue here were meant to protect AHP on an exposure basis.

3. Implications of rate computation. The federal courts that have heretofore ruled on the meaning of the CGL language have ignored the manner in which the parties to such contracts proceeded from year to year to set rates for coverage. Evidence of the premiums agreed to by the parties and their relation to the claims for which the insurer provided coverage or assumed the burden of defense, is probative of how the parties construed the contracts at issue. See, e.g., Pan American World Airways, Ltd. v. Aetna Casualty & Surety Co., 505 F.2d 989, 1003 (2d Cir.1974). An analysis of the practice of parties in negotiating rates must be sensitive to the reasons of convenience or practicability that could explain their willingness to accord special significance to dates of exposure or manifestation in the day-to-day administration of the policies. Nevertheless, the practices by which the parties allocated specific claims to the policies and computed premiums provide strong proof in this case of the intent embodied in the occurrence clause.

The parties employed a methodology apparently typical for such insurance arrangements, that enabled both sides to avoid excessive profits or losses through an annual appraisal based on actual experience. Each insurance policy issued to AHP by Liberty was for a three-year term. However, premiums due Liberty, and the reserves necessary to cover claims that remained open at each year's end, were computed annually by Liberty and then negotiated by the parties. The annual premium agreed to was initially only an estimate. For example, the 1964 policy states:

Pl.Ex. 17 at 4.

The process of negotiating the premium for a particular year would begin with Liberty itemizing the losses assigned to the policies for the previous three years. "In listing these losses, each claim was identified by name and was assigned to one policy period. Claims were always assigned to the first policy period in which the injury or illness was manifested." Conner Aff't, Def.Ex. 15, ¶ 4. (Feb. 3, 1982). The concept of manifestation utilized by the parties was not the narrow, discovery theory, but an estimate of the presence-in-fact of compensable injury. According to William E. Urmston, Supervising Examiner in the Special Claims Department which handled unusually large or difficult AHP claims, Urmston Aff't, Def.Ex. 14, ¶ 1 (Feb. 8, 1982), an injury was considered manifest when it was "capable of detection," id. ¶ 3 — i.e. when it existed, albeit undetected. See also Stevens Dep., Pl.Ex. 8, at 113 ("Q: Under Liberty's manifestation theory, when is an injury manifest? A: When it's diagnosed or capable of being diagnosed."). Liberty would then compute a proposed annual premium based on the three-year loss experience thus compiled. In addition, Liberty would itemize claims that were still open and compute the additional reserves required to cover those claims. It would estimate this reserve either on a case-by-case basis or, when a number of suits alleged injuries resulting from ingestion of the same drug, Liberty would estimate reserves for the group as a whole, taking into consideration the number of claims and the likelihood that payment would become necessary on a certain percentage of these claims. See id. at 113.

Liberty would then send this information to AHP's insurance manager and, after he had reviewed the data the parties would negotiate any points of disagreement and arrive at appropriate figures for premiums and reserves. Conner Aff't ¶¶ 5-7. During these negotiations AHP was often able to gain concessions on the reserves as well as on the premiums charged. Pizl Dep., Def.Ex. 1 at 123-25. After the parties agreed to a premium estimate, the policy would run its course. Finally, at the end of the year, the parties would resolve any differences between the estimated and the actually earned premium, and begin negotiating an estimate of the next year's premium.

Thus, according to the undisputed testimony of the parties' officers, AHP's premiums were computed on a loss record that used injury-in-fact as the basis for assigning claims to the various policies. As Urmston testified: "How reserves and payments of costs, settlements, and judgments, are charged to policies is of great importance to the policyholder, because it affects the premiums which Liberty will charge upon the renewal of the policy." Urmston Aff't ¶ 4. Yet, AHP raised no objection to the injury-in-fact basis utilized by the parties during their annual negotiations. Pizl Dep., Def.Ex. 1 at 128-30.

The rate-setting process actually followed by the parties puts to rest simplistic notions of what an insured such as AHP is paying for when it obtains coverage. Although courts have correctly noted that companies such as AHP pay insurers to absorb risks, this proposition lacks meaning unless placed in the context of the rational process that goes into rate formulation. Liability insurance is not in this context a process by which insured and insurer gamble on the amount of coverage that will be required in a given year. The payments made are annually estimated to reflect the anticipated liability for the period of coverage, plus a reserve for estimated future losses, and a commercially negotiated margin of profit. The premiums and reserves are then recalculated and renegotiated at each year's end on the basis of actual experience.

Thus, when major corporate insureds such as AHP purchase liability insurance, they cannot reasonably be said to expect the insurer to have assumed their entire risk of liability, compare Keene, 667 F.2d at 1047, but rather are properly seen as using the insurance contract as a device for spreading the risks they face in a rational manner. See Wrubel, supra, 48 Fordham L.Rev. at 683 n. 149. Typically, the premium agreed upon is based on the prior loss record, which reflects anticipated experience. When actual experience varies from this anticipated experience, and the insurer or the insured might be able to reap unanticipated profits or be required to suffer unanticipated losses, the parties either terminate their relationship or continue it by renegotiating the rates and reserves to negate the noted disparity. In this way, no one side in an ongoing relationship is able to benefit in an unexpected manner, because of unanticipated risks that mature. The mechanism is more a device to spread than to shift risk. The arguments asserted here by AHP and Liberty therefore have no basis in their prior dealings, but arise out of self-interested positions formulated after the rational process for risk spreading in which they were engaged ended with the termination of their relationship.

AHP cancelled its policy with Liberty simply because Liberty sought much higher premiums as AHP's liability from drug litigation steadily escalated. In June 1976, Liberty submitted to Rudolph Pizl, AHP's insurance manager, a renewal quotation for the period of July 1, 1976 to July 1, 1977, in excess of $6,000,000. This compared to deposit premiums of $2,809,821 for the prior year, and $1,858,530 for the year preceding that. These substantial increases, based on actual and projected experience, are what prompted Mr. Pizl to inquire into the cost of self-insurance. He concluded, based on an in-depth study, that AHP's interests would best be served by self-insurance. See Pizl Memorandum, Pl.Ex. 32 (Sept. 1, 1976). Accordingly, on Pizl's recommendation, AHP cancelled its insurance with Liberty effective November 1, 1976. After termination, AHP continued to seek policy coverage for claims asserted against it after 1976. But only then, for the first time in the parties' relationship, did AHP assert claims, not only for liability on account of injuries that actually occurred during pre-1976 coverage, but also for liability on account of exposures occurring during all prior periods of coverage.

Whatever other basis AHP may have to commend its present position, nothing in the course of dealing and commercial expectations of the parties supports the exposure theory. Upon termination of coverage after 1976, the proper degree of risk accepted by Liberty was no less and no more than the risk anticipated by the parties in setting the rate for the last year of coverage. Given the history of the AHP-Liberty relationship, that rate did not anticipate reimbursement for all injuries whenever they might occur as a result of exposures during coverage; and neither did it anticipate reimbursement only for injuries that had become manifest before coverage ended. Rather, consistent with the clear language of the policy, and with the background and purposes of the parties, coverage was anticipated to continue to extend to injuries that had in fact occurred while prior policies were in effect, irrespective of when exposure took place or when the injuries became manifest.

III. Considerations of Practicability and Fairness

The foregoing analysis demonstrates that the parties intended by their language to provide coverage for injuries that in fact occurred while the policies were in effect. The demonstration is based on the actual language and experience of the parties, and could have been made by any of the several federal courts that have passed on this question. But these courts, and the parties before them, have deemed the injury-in-fact approach embodied in the CGL's language to be unacceptable for a variety of reasons. They contend, first, that an injury-in-fact rule is impractical — that it would require thousands of mini-trials in individual cases to determine when each injury occurred. They also claim that these hearings will lead to unreliable findings, since medical science cannot estimate with accuracy when an injury, sickness, or disease first arose in a particular person. Although these reasons for rejecting injury-in-fact are present in virtually every discussion of how the CGL language should be construed, they are most clearly expressed by the Sixth Circuit in Forty-Eight Insulations:

633 F.2d at 1218.

In addition to these practical considerations, the injury-in-fact approach has been rejected, expressly or implicitly, by at least two federal courts for its failure to correlate insurance coverage with underlying tort doctrine, and particularly to take account of recent developments that expand manufacturer liability through concepts such as enterprise liability. See, e.g., Keene, 667 F.2d at 1051; Forty-Eight Insulations, 633 F.2d at 1215, 1225. The injury-in-fact approach has also been deemed inconsistent with the "premise" that coverage must exist in order to satisfy the insured's legitimate expectations. Keene, 667 F.2d at 1044-45. Finally, the injury-in-fact approach would not guarantee the expanded coverage for liability to the underlying tort claimants that courts addressing the issue uniformly have provided. Although an injury-in-fact test would provide recovery in many suits, and would require the insurance companies involved to defend all claims that merely permit proof of the occurrence of injury during the policy period, it would deny the right to a defense where allegations do not permit proof of injury during coverage, and would deny recovery for liability where the insured could not prove that liability was based on an injury, sickness, or disease that occurred during coverage. These various reasons for rejecting injury-in-fact lack empirical support, and are in any event inappropriate bases for federal courts in diversity proceedings to disregard the plain language and bargained-for exchanges embodied in private insurance contracts arrived at through arms-length negotiations.

The courts are wrong to conclude that medical experts and finders of fact will be unable to determine the time of actual injury, sickness, or disease with sufficient accuracy. In many cases, the onset of injury, sickness, or disease will be possible to pinpoint with great certainty, since the time of such an occurrence will be simultaneous with exposure or manifestation. Even where the injury, sickness, or disease cannot be pinpointed, however, experts will be able to place before the parties for purposes of negotiation, and before the finders of fact at trial, estimates based on a reasonable degree of medical certainty as to when the occurrence became diagnosable. This is precisely the manner in which pre-CGL cases were handled, and the manner in which many types of tort and insurance claims are handled today. Of course experts will be unable in many cases to identify the exact day on which an injury, sickness, or disease occurred. But such precision is not required; all that is necessary is reasonably reliable evidence that the injury, sickness, or disease more likely than not occurred during a period of coverage, which usually is at least one year long, and in the policies at issue extends over three years. Finally, the notion that an exposure or manifestation approach will provide more reliable findings than the usual process for estimating the time of injury is largely illusory. A plaintiff's testimony as to the exact time of exposure or manifestation will likely be no more reliable or precise than an expert's estimate based on the injury's objectively verifiable effects and the entire history of the case. Defendants frequently challenge allegations as to the intensity and duration of exposures and as to the time particular plaintiffs knew or should have known of an injury, sickness, or disease.

Collateral estoppel on medical issues might also prove useful in proving when injury occurred. Unlike the variable manner in which injuries are caused by asbestos fibres, other products may produce specific consequences at particular times. When a manufacturer's product has been established to injure in a particular manner, or at an identifiable time, courts may be free to rule in all similar cases that the product caused the injury and/or that it did so at the time identified. See generally Hardy v. Johns-Manville Sales Corp., 681 F.2d 334 (5th Cir.1982); Wetherill v. University of Chicago, 548 F.Supp. 66 (N.D.Ill. 1982). In all such situations, the problem of estimating the time of injury will be greatly alleviated. See also Williams, Mass Tort Class Actions; Going, Going, Gone? 22 Judges' J. 8 (1983) (suggesting the use of class actions to resolve issues of fact in mass tort litigation).

A practical appreciation of the likelihood of mini-trials in "thousands" of insidious disease cases to establish insurance coverage must begin with two crucial considerations: (1) that the insurance company will be required to defend all cases in which liability is conceivable; and (2) that (in part because of this duty to defend) the vast majority (about 95%) of all underlying tort cases will settle. Once an insurance company is required to defend a case, it rather than the insured handles the litigation for all purposes including settlement. Furthermore, when cases are settled, the parties will usually have worked out not only what the plaintiff is to be paid, but also to which policy or policies the settlement will be attributed, and to what extent.

Those tort litigations that proceed to trial will not all result, moreover, in a finding of liability. Insurance companies are experienced at evaluating the merits of claims; consequently, of the tort suits that go to trial, a large proportion will result in verdicts for the defendant. Furthermore, when a suit results in liability, the trial will often also result in a finding of when the compensable injury occurred. Despite suggestions to the contrary in recent opinions construing the CGL language, tort suits in most American jurisdictions generally require, for one reason or another, proof of causation and compensable injury; and the process of establishing these necessary prerequisites to liability often will result in proof of how and when the injury took place. Trial courts could make this process even more productive by requiring findings of the time of injury, to the extent such findings are not apparent or implicit in the evidence and general verdict. Finally, once a trial of the underlying tort suit has been held, settlements seem likely to be common on the issue of coverage in those cases in which the issue remains unresolved at the end of the litigation process. Thus, the threat of thousands of mini-hearings on coverage seems grounded on assumptions having no basis in the practical realities of tort litigation.

Courts intent on particular results, and unsatisfied with enforcing bargained-for exchanges, quite naturally have purported to create rules for insurance contracts that match the rules in underlying tort suits. But in fact courts have not and cannot achieve this symmetry. No simple set of rules exists in tort litigation; the state courts have adopted a variety of approaches in cases involving insidious diseases, and different rules sometimes exist within a single state for different types of diseases. The rules are continuously in flux, moreover, and if insurance rules were ever to correlate with tort rules, their symmetry would be short-lived, as particular states experimented with new ones. Finally, if this notion of symmetry between insurance and tort rules were to be taken seriously as an indicator of the CGL's or the parties' intentions, then the relevant inquiry would arguably turn to the tort liability rules most widely in effect at the time the CGL was adopted, under which liability would be imposed only upon proof of actual causation, and only to the extent of the individual company's proven responsibility. The expanded notions of liability to which courts have pointed in construing the policy language are entirely the product of conceptions formulated and rulings made years after the language at issue had been promulgated.

The true rationale for recent decisions construing the CGL language involved in this case seems most clearly related to the view that insureds must be deemed to have negotiated for the purpose of obtaining coverage for all liability. As Judge Bazelon said in Keene:

667 F.2d at 1041 (citations omitted). Most other decisions purport to be less result-oriented than Judge Bazelon's forthright exposition. But as AHP argues in this case, the one consistent way to read these cases is to examine the results they reached in terms of coverage or noncoverage, rather than to study their holdings. For, while each case's holding varies from that of the others, all the cases find coverage for the claim or claims specifically presented. This result is particularly attractive in connection with insidious diseases, since liability for such diseases may be so extensive with respect to particular products that even the largest corporate defendants might be forced out of business unless access to insurance funds is assured.

Assuming for now the legitimacy of relying upon social considerations that so obviously favor insureds, the fact is that courts cannot safely predict the effects of the rules they have adopted on all categories of tort litigants. The manifestation theory may up to now have been used only to extend coverage, but it in theory certainly points to the rejection of coverage in many cases. No doubt the exposure theory will potentially extend coverage to greater numbers of claimants' recoveries than requiring proof of injury-in-fact, especially if all periods after exposure are treated as "exposures in residence." But the practical effects of these rules cannot readily be known and may prove unjust. Insurance contracts have absolute monetary limits, and the decision to expand coverage to a greater number of claims than the policy language indicates will impose a redistribution of the limited proceeds available. Some claims that would have been entirely covered by a narrower construction of a policy will be covered only in part; others will be entirely deprived of coverage when a policy limit is exhausted in satisfying claims based on a broader construction of the insurance contract. Thus, a policy obtained by a manufacturer to cover workers or consumers actually injured during coverage may have its purpose undermined significantly if it is expanded to cover the claims of workers or consumers who are injured before or after that period. A worker exposed at an employer's plant during a particular year would be able to obtain insurance proceeds allocated by that employer to that year, even though the worker thereafter went on to another employer's plant and suffered real injury many years thereafter.

Courts also must consider the fact that they cannot force insurance companies to renew contracts for liability coverage. By imposing greater costs upon insurers, courts make it necessary for them to increase their rates, not merely to compensate for their increased liability, but also to anticipate other, judicially created liabilities that they fail to exclude with sufficient clarity. Some insurers will simply withdraw from the business of liability coverage, leaving the field with less competition and higher rates. See, e.g., T. Lewin, "Insurers Balking on Company Losses," N.Y. Times, Mar. 30, 1983, at 1, col. 1. While insured companies may be willing at this time to ignore these long-term risks because of the vast liabilities they face, the persons whom they now cover may well be grievously hurt in future years by the lower coverage that results, or by the bankruptcies caused by companies becoming self-insured in an effort to avoid the higher rates required to pay for broader theories of coverage. In short, broader coverage has a price that must ultimately be paid. And in the long run insurance companies must pass on that price or must themselves suffer hardship or failure.

Relying on social policy to justify imputing an expectation of complete coverage to the insured is, in any event, legally unsupportable. The only doctrinal basis for enforcing expectations in the face of contrary policy provisions — the so-called reasonable-expectations doctrine relied on in Keene, 667 F.2d at 1042 n. 12 — is reserved for situations in which the expectations enforced are strongly demonstrated, and the policy involved is a contract of adhesion. See Keeton, Insurance Law Rights at Variance with Policy Provisions, 83 Harv.L.Rev. 961, 967 (1970). Here, the parties' expectations were entirely consistent with the policy language, and one cannot equate AHP's manuscript policies with the usual adhesion contract. Furthermore, New Hampshire appears to be the only state that adheres to the reasonable-expectations doctrine; no state law arguably applicable in this case would permit reliance upon the doctrine in connection with unambiguous passages. See generally, Note, A Common Law Alternative to the Doctrine of Reasonable Expectations in the Construction of Insurance Contracts, 57 N.Y.U.L.Rev. 1175 (1982).

Finally, the judiciary's refusal to adhere to the plain meaning and intent of the CGL language creates uncertainty for tort claimants as to the scope of coverage. Courts have assumed that they should defer to the demand of insurers and insureds that the injury-in-fact approach must be rejected as impractical. Yet, the evidence seems now to establish what human experience should have suggested: once courts deem themselves free to ignore the language and intent of a negotiated contract, they are left with little or no basis upon which to arrive at consistent results in deciding how the contract should be read. Judges in such situations are left to act essentially as legislators and, along with the flexibility they obtain to choose among possible results, they also reap the uncertainty that stems from having to rely heavily on personal values and inclinations, as well as upon empirically unsound and potentially disruptive perceptions of fairness. The result for tort litigants has in fact been a great slowdown in the processing of their claims, as one federal court after another devises its own prescription for the problem of insidious-disease coverage.6

That litigation has now continued for years over which of the theories — exposure or manifestation — is proper, belies the notion that a choice between these theories will advance to any material extent a settlement of the overall theoretical dispute, let alone the thousands of lawsuits that are being delayed while manufacturers and insurers are purportedly attempting to simplify the process of evaluating and processing these claims. Nor has the extensive delay resulted in the development of any accepted rule of law or procedure by which insurers have agreed to proceed. Four Circuit Courts of Appeals have reached four different results on the questions at issue, some diametrically opposed to others, and the Supreme Court has refused to review a single one of those cases, leaving the law in conceptual chaos. The Supreme Court seems unlikely to pass on these questions, moreover, because of its long-standing policy against reviewing Circuit Court rulings that purport to be based upon state law. See Ruhlin v. New York Life Insurance Co., 304 U.S. 202, 206, 58 S.Ct. 860, 861, 82 L.Ed. 1290 (1938); R. Stern & E. Gressman, Supreme Court Practice 265-66 (5th ed. 1978). Even if a rule of law is established within individual circuits, incongruous differences in coverage will result for insureds in different geographic areas, under the same insurance contracts. Nor will the applicable law remain clear even within individual federal circuits. Thus far, parties to these insurance contracts have been avoiding the state courts. In time, however, the state courts will rule on the meaning of the CGL language and its variants, and when they do, they seem likely to differ with federal courts that have chosen to disregard the plain language. These rulings will create disparities among the states within individual circuits, adding to the uncertainty that already exists.7

The only judicial solution to the problem of obtaining an authoritative construction of the CGL language is review by the Supreme Court. Review by the High Court would in fact be appropriate on this issue, despite the general wisdom of the Court's policy of avoiding state-law questions. The disparities of construction among the federal courts is a matter of national concern, because many of the courts involved are not even attempting to apply state law in reaching their results, and because these rulings are having significant effects upon the victims of insidious diseases, upon manufacturers, and upon the insurance industry, throughout the nation. Meanwhile, however, the proper course in this case is to follow New York law, which requires adherence to the plain language of the insurance contracts at issue as supported by the parties' proven intent.

IV. Conclusion

The policies at issue in this case explicitly cover liability for injuries, sicknesses, or diseases that occur during each period of coverage. They were intended to cover no more and no less. Coverage is triggered by neither exposure nor manifestation, except when those events constitute in themselves an injury, sickness, or disease for which an insured may be held liable. The policies also expressly require Liberty to bear the costs of defending AHP in all suits, however meritless, that can be read to permit proof that an injury, sickness, or disease occurred during a period of coverage. No considerations of practicability or fairness justify ignoring the plain meaning and purposes of these policies.

The parties will submit within twenty days a declaratory order encompassing these rulings.

SO ORDERED.