For the past eighteen years, this case, a declaratory judgment action brought by the maker of the drug diethylstilbestrol ("DES") to resolve complex insurance coverage issues, has been slowly making its way through the federal courts. While it is, unfortunately, not unusual for a complicated case such as this one to take years to resolve, eighteen years is a particularly long time and the parties are understandably anxious to reach a conclusion. Thus they were, we think it safe to say, more than slightly miffed when, in our first encounter with the case, we remanded it to the district court for a determination of whether, given the presence, as defendants, of "Certain Underwriters at Lloyd's of London" ("Lloyd's"), federal subject matter jurisdiction existed. See E.R. Squibb & Sons v. Accident & Cas. Ins. Co., 160 F.3d 925 (2d Cir. 1998) ("Squibb I"). At that time, we indicated that such jurisdiction might lie, but only if certain factual premises obtained. We also noted that we, at the appellate level, were not in a position to verify these premises.

On remand, the district court (John S. Martin, Jr., Judge), following our mandate with great care, gathered evidence and concluded that there was subject matter jurisdiction. The case has, as a result, returned to us for review. On this renewed appeal, we are again obligated first to consider the district court's jurisdictional findings and then, if they are correct, to examine the merits. Having done so, we now affirm the district court's finding that diversity jurisdiction exists and also affirm the district court's decision on all but one of the merits issues raised by the defendant insurers. With respect to that single claim -- the question of whether Continental Casualty Company ("CNA") should be obligated to pay defense costs for the period of January 1, 1971 to January 1, 1976 -- we hold that the district court improperly granted summary judgment. Accordingly, we reverse and remand that claim for further proceedings.

We will assume some familiarity with our previous opinion in Squibb I and hence will only sketch briefly the general background of the case. Whatever further factual information is necessary will be provided in the discussion of each issue on appeal.

In the years relevant to this case, 1953 to 1976, E.R. Squibb & Sons, Inc. ("Squibb") insured itself against various risks arising in the course of its business. Its insurance coverage was structured into layers, such that the higher (or "excess") layers would only be reached after the bottom (or "primary") layer had been exhausted. Faced with a tidal wave of litigation arising from the injuries associated with use of its product DES during pregnancy, Squibb turned to both its primary and excess insurers for coverage.

To determine the extent of each product liability insurer's coverage with respect to thousands of settled, pending, and future actions against it in cases involving DES, Squibb in March 1982 brought this declaratory judgment action under 28 U.S.C. 2201. The case was first filed in the district court in the District of Columbia but was transferred to the Southern District of New York. Once transferred, it was consolidated with another suit that Squibb had brought in the same court, and a single unified complaint, filed in 1984, seeking both compensatory and punitive damages, served as the basis for the action.

During the course of the litigation, some of the insurers settled with Squibb. At that point, all of Squibb's disputes with its primary insurers were at an end, and only its excess insurers (the "Excess Insurers") remained as defendants in the action. The Excess Insurers participated in two trials on liability that were held in 1996. The first of these was tried to a jury, which determined (with respect to the damages alleged by the DES claimants) the time at which injury-in-fact occurred. The second trial was to the district court and decided various insurance issues such as when the secondary and tertiary insurers were obligated to begin paying on claims. After the trials, additional hearings took place with respect to the appropriate relief. Later still, a trial on damages and other relief was held, and a final judgment was entered. Most of the parties then appealed to this Court.

Instead of resolving the merits, however, we remanded the case to the district court for a determination of whether it had subject matter jurisdiction over the case, given that Allan Peter Denis Haycock, an underwriter of Lloyd's, was named as a party to this suit, not only in his individual capacity but also as a representative of an undefined number of Lloyd's underwriters. On remand, the district court held an evidentiary hearing, and, on a variety of different grounds, held that diversity jurisdiction existed. See E.R. Squibb & Sons, Inc. v. Accident & Cas. Ins. Co., 1999 U.S. Dist. LEXIS 8333, 1999 WL 350857 (S.D.N.Y. June 2, 1999) ("Squibb II"). One of these grounds required the dismissal of Haycock and the substitution of another party, Stephen Merrett. The lower court, however, did not dismiss or substitute any parties, pending our approval of the specific change. It entered a new final judgment and the parties once again appealed. Because certain insolvent defendants remain before the district court, this appeal was certified under Rule 54(b) of the Federal Rules of Civil Procedure.

In addition to the question of subject matter jurisdiction, we address the following issues going to the merits of this case: (a) whether, given another court's ruling that one of Squibb's primary insurance policies was triggered by the manifestation of DES injury, Squibb is estopped from benefitting in this case from an injury-in-fact trigger of insurance coverage; (b) whether the district court erred in excluding evidence proffered by the Excess Insurers on the question of whether DES caused certain injuries in children of women who ingested DES during pregnancy; (c) whether certain claims by the grandchildren of these women are covered under the insurance policies at issue; (d) whether the district court properly allocated responsibility for Squibb's losses among its various insurers; (e) whether the district court erred in the procedures it established for addressing new post-judgment information and claims; (f) whether the district court erred in its interpretation of when a deductible in certain policies applies; (g) whether certain statements in the final judgment's preamble should be stricken; (h) whether there is a case or controversy concerning the policies of certain Excess Insurers which provide coverage to Squibb only if its liability reaches very high levels; and (i) whether CNA's policies cover Squibb's defense-related costs between January 1, 1971 and January 1, 1976.

As we discussed in Squibb I, Squibb's consolidated complaint named Allan Peter Denis Haycock, a British subject, as a representative of certain underwriters at Lloyd's. See Squibb I, 160 F.3d at 928. The parties later stipulated that Haycock was appearing in the action both "in his individual capacity, and for administrative convenience, as a representative of all Lloyd's Underwriters."1 Id. (internal quotation marks omitted).

When, during the pendency of the last appeal, we asked for additional briefing on the issue of subject matter jurisdiction given the presence of these underwriters at Lloyd's, the parties discovered that Haycock had died. Squibb and Lloyd's therefore moved to add Stephen Merrett, another Lloyd's underwriter and British subject, as a replacement for Haycock. See id. at 928 & n.6. We did not grant this motion, but rather remanded, asking the district court to reassess all the facts supporting diversity jurisdiction. We also asked the district court, after it had all of the information before it, to determine whether the substitution of Merrett would be proper.

In particular (because we had concluded that suing an underwriter in a representative capacity required the court to determine if each and every party that individual represented -- in this case, a large number of unspecified individuals -- met the diversity requirements), we asked the district court to ascertain if the parties represented by Haycock were indispensable or if, instead, one Lloyd's underwriter could be sued solely in his individual capacity without prejudicing the other parties to the litigation. See id. at 939. If this were possible, and if the individual underwriter sued was both diverse in citizenship and satisfied the statutory requirement of a minimum amount in controversy, the jurisdictional problem could be avoided in a fairly straightforward fashion.

On remand, the district court found that, in the instant case, any individual underwriter at Lloyd's (a "Name") could be sued in his individual capacity without creating difficulties. See Squibb II, 1999 WL 350857, at *5. Specifically, it held that dismissing the suit against either Haycock or Merrett in their representative capacity would not result in the dismissal of indispensable parties because the record "clearly establishes that as a matter of contract and the rules of Lloyd's, a judgment [in this country] against either Haycock or Merrett will be honored by the other Names," i.e., by the other underwriters who are bound by contract to share the policy liability. Squibb II, 1999 WL 350857, at *11. As a result, the court concluded that there would be no prejudice to any of the other defendants from dropping the "represented" parties. It expressly found that "the record establishes that the absent Names are, for practical purposes, as bound by a judgment against Haycock or Merrett in their individual capacities as they would be by a judgment against them in their representative capacities." Id. at *13.

The district court also determined that both Haycock and Merrett satisfied the diversity requirements and that, therefore, subject matter jurisdiction could be based on the presence of either one of them, individually. First, as British subjects, they each met the diversity of citizenship standards. Second, for a variety of reasons, most of which we need not consider, the court found that both had satisfied the amount in controversy requirement. See id. at *5-*11. We now ratify the district court's conclusion that subject matter jurisdiction over this suit can be established. But we do so only on the premise that Merrett (in his individual capacity) will be the sole Name present as a defendant. To that end, we order that Merrett be substituted for Haycock as a defendant, and that Merrett's presence in the suit be solely in an individual capacity.

To begin with, we agree with the district court -- substantially for the reasons it gave -- that Merrett, a British subject, may be sued in an individual capacity, and that, contrary to the claims of CNA, no prejudice results if we allow the suit to be recast as a suit against Merrett in that capacity. As a matter of contract, each of the underwriters of the policy will be bound by an individual judgment against Merrett. See Squibb I, 160 F.3d at 929. And given the peculiar structure of Lloyd's, this fact suffices to protect all the parties, substantially to the same degree as if Merrett (or Haycock) were being sued in a representative capacity. Moreover, given Haycock's death, there is no problem in substituting Merrett for Haycock. See Fed. R. App. P. 43(a)(1) (if a party dies and there is no personal representative to act on his behalf, the court may take such action as it deems appropriate); see, e.g., Ward v. Edgeton, 59 F.3d 652 (7th Cir. 1995).

The more difficult question is whether Merrett, himself, satisfies the requirements of diversity, and in particular those concerning the required jurisdictional amount. Merrett's liability -- the district court found -- is $ 23,762.00. As a result, the requirement will be readily satisfied if it is found that the $ 10,000 jurisdictional amount, which was in effect at the time the consolidated complaint was filed, applies, but not if the $ 75,000 amount currently mandated, governs. Which figure controls depends on whether, upon the granting of Squibb's motion to amend its complaint to effect the substitution, the replacement of Haycock in a representative capacity with Merrett in an individual capacity relates back under Rule 15(c)(3) of the Federal Rules of Civil Procedure.

An amendment of a pleading relates back to the date of the original pleading when . . . the amendment changes the party or the naming of the party against whom a claim is asserted if the foregoing provision (2) [requiring that the asserted claims arise out of the occurrence set forth in the original pleading] is satisfied and, within the period provided by Rule 4(m) for service of the summons and complaint, the party to be brought in by amendment (A) has received such notice of the institution of the action that the party will not be prejudiced in maintaining a defense on the merits, and (B) knew or should have known that, but for a mistake concerning the identity of the proper party, the action would have been brought against the party.

Fed. R. Civ. P. 15(c). The district court concluded that all of the rule's requirements for relating back had been satisfied here. It therefore acted as if the substitution had taken place on the date the complaint was filed in 1984 and assessed diversity with respect to Merrett as of that moment.2

We affirm the district court's conclusions (a) that the amendment relates back under Rule 15(c)(3), (b) that, therefore, the $ 10,000 amount in controversy requirement applies, and (c) that it has been met. In reaching this result, we hold that, where it is appropriate to relate back an amendment to a pleading under Rule 15, jurisdiction is assessed as if the amendment had taken place at the time the complaint was first filed. This means that the applicable law as to what amount must be in controversy, and as to how diverse the parties must be, will be that in effect at the time of the filing of the relevant complaint. Similarly, questions of fact, such as how much money is actually at stake and where each party lives, will also be determined with reference to the date on which the relevant complaint was filed.

Although we have never explicitly stated the rule in this fashion before, it clearly follows from the weight of diversity jurisprudence. As a general matter, it is widely accepted that amendments to cure subject matter jurisdiction relate back. See 6A Charles Alan Wright et al., Federal Practice and Procedure 1497, at 80 (2d ed. 1990) ("Amendments curing a defective statement of subject matter jurisdiction . . . will relate back . . . ." (footnotes omitted)); 3 James William Moore, Moore's Federal Practice P 15.15 [3.--2], at 15-154 (1996) ("Even though parties must be dropped to perfect diversity jurisdiction, if they were not indispensable parties the amendment will relate back to allow for entry of judgment on the original verdict where this is not prejudicial to the remaining parties. Similarly, although there may not be complete diversity when the action is brought, an amendment dropping non-indispensable parties to cure the jurisdictional defect will relate back." (footnotes omitted)); see also Carney v. Resolution Trust Corp., 19 F.3d 950, 954 (5th Cir. 1994) (holding that relation back is appropriate "even when the amendment states a new basis for subject matter jurisdiction"); Berkshire Fashions, Inc. v. The M.V. Hakusan II, 954 F.2d 874, 887-88 (3d Cir. 1992) (same, and also holding specifically that when an amendment meets the requirements of Rule 15(c)(3), it is proper to evaluate that amendment based on the law as of the date the complaint was filed when deciding whether the amount in controversy requirement is satisfied).

Moreover, once subject matter jurisdiction is "cured" by an amendment, courts regularly have treated the defect as having been eliminated from the outset of the action. In other words, where a change in parties, necessary to the existence of jurisdiction, is appropriate and is made (even on or after appeal), appellate courts have acted as if the trial court had jurisdiction from the beginning of the litigation. This has permitted such courts to affirm decisions of trial courts on the merits, although the requisite change in the parties did not occur until much later in the action. See, e.g., Newman-Green, Inc. v. Alfonzo-Larrain, 490 U.S. 826, 104 L. Ed. 2d 893, 109 S. Ct. 2218 (1989). By thus retroactively validating the pre-amendment substantive decisions of the lower court, appellate courts -- the Supreme Court has told us -- avoid the "waste of time and resources [that] would be engendered by remanding to the District Court or by forcing the[] parties to begin anew." Id. at 838.

For analogous reasons, federal courts have always assessed jurisdiction and evaluated the factual premises for it, e.g., the amount in controversy and the citizenship of parties, as of the moment the complaint was filed. See 13B Wright et al., Federal Practice and Procedure 3608, at 448-49 (2d ed. 1984 & Supp. 2000). As a result, we have consistently held that federal jurisdiction is not defeated if one party, subsequent to the filing of a complaint, becomes a citizen of the same state as his opponent. See, e.g., In re Agent Orange Prod. Liab. Litig., 818 F.2d 145, 163 (2d Cir. 1987). The same is true with respect to changes in the law. Thus, if a suit is filed and, while the suit is pending, the minimum jurisdictional amount is raised by statute, the court does not lose jurisdiction, even though the amount in controversy in that particular suit does not satisfy the new standard. See, e.g., Central Fiber Corp. v. Site Servs. Ltd., 962 F. Supp. 1426, 1427 (D. Kan. 1997).

Having concluded that the relevant time at which to determine whether Merrett met the jurisdictional amount was the moment the instant suit was filed, and that, as of that time, the jurisdictional requirement was satisfied, the only issue that remains before us is whether Merrett's substitution for Haycock meets Rule 15(c)(3)'s standards. As mentioned earlier, this will be so if (1) there is no prejudice as a result of the substitution, (2) the party to be brought in by the curative amendment had notice of the action and knew or should have known that he would be named as the proper party in order to cure the jurisdictional defect,3 and (3) the claims remain the same as those that were originally asserted.

In this case, as discussed above, (1) no prejudice results from replacing Haycock (in both his individual capacity and as the representative of the other Lloyd's Names) with Merrett in his individual capacity, (2) Merrett had notice that this action was brought and knew or should have known that he was the proper party to be sued had the parties not erred as a jurisdictional matter by going after Haycock in a representative capacity, and (3) the claims asserted against Merrett remained the same as those stated in the original 1984 consolidated complaint.

Because the substitution of Merrett properly relates back under Rule 15(c)(3) and because Merrett satisfies both the diversity of citizenship requirement and the applicable amount in controversy requirement, we find that, with Merrett as an individual defendant, diversity subject matter jurisdiction exists. We therefore dismiss Lloyd's and Haycock from the suit and substitute Merrett in his individual capacity in their place. Since, with this amendment, diversity jurisdiction over the suit has been established, we need not, and do not, express any opinion on any of the other bases for jurisdiction considered by the district court.

Squibb has settled its coverage disputes with all of its solvent primary insurers. The final judgment in this case declares certain coverage obligations of Squibb's Excess Insurers in the years 1953 to 1976 (except for the first-level excess insurer for the years 1968 to 1977, which has also made a separate peace). Because first-level excess coverage in the period 1953-68 was reached by exhaustion of primary limits, the final judgment implements its coverage rulings and imposes liability on underwriters at Lloyd's and various London companies in the amount of $ 36,779,737 for indemnity and defense costs, and in the amount of $ 6,262.82 per day in pre-judgment interest for the period beginning after June 30, 1997.

In general, a liability insurer's "duty to indemnify is 'triggered' by a determination that fortuitous bodily injury or property damage took place during the policy period." Barry R. Ostrager & Thomas R. Newman, Handbook on Insurance Coverage Disputes 9.01, at 408 (9th ed. 1998). The numerous excess policies at issue in this appeal are close analogs of the Comprehensive General Liability Policy ("CGL") form of liability policy drafted by members of the insurance industry in the 1960s. The policy provisions implicated by our analysis are set out below. The parties agree that New York law governs the interpretation of these provisions.

The policies contain similar coverage agreements, which typically provide that the Excess Insurers shall:

Indemnify the Assured for all sums which the Assured shall be obligated to pay by reason of the liability

(a) imposed upon the Assured by law,

or

(b) assumed under contract or agreement by the Named Assured . . . for damages, direct or consequential and expenses, all as more fully defined by the term "ultimate net loss" on account of:-

(i) Personal Injuries, including death at any time resulting therefrom, . . . caused by or arising out of each occurrence happening anywhere in the world.

"Occurrence" is defined (in relevant part) as "an accident or a happening or event or a continuous or repeated exposure to conditions which unexpectedly and unintentionally results in personal injury . . . during the policy period." Accordingly, the parties agree that coverage under the agreements is triggered by "personal injury" during a policy period. "Personal injuries" include "bodily injury, mental injury, mental anguish, shock, sickness, disease [and] disability . . . ."

The policy limits and underlying coverages vary among the policies, but their excess character is expressed by clauses accepting indemnity liability only for a stated ultimate net loss in excess of the stated limits of underlying insurances or stated retentions. The term "ultimate net loss" is defined as:

The total sum which the Assured, or any company as his insurer, or both, become obligated to pay by reason of personal injury . . . either through adjudication or compromise, and shall also include . . . all sums paid . . . for litigation, settlement, adjustment and investigation of claims and suits which are paid as a consequence of any occurrence covered hereunder. . . .

Squibb sought a declaration of coverage for claims brought against it by three generations of DES claimants: 1) women who ingested DES during pregnancy to prevent miscarriage (the "first generation"); 2) the children of those women, who were exposed to DES in utero (the "second generation"); and 3) the grandchildren of those women, who suffered injuries related to premature birth caused by their mothers' anatomical injuries (the "third generation"). At trial in 1996, the jury determined when coverage was triggered for the claims brought by these groups of claimants. The jury found that the claimants' injuries occurred over many years and therefore triggered multiple policy periods. The district court, acting as fact-finder, then determined at which points the Excess Insurers were obliged to begin paying on the claims.

The Excess Insurers appeal on the grounds discussed below.

In 1978, Squibb prevailed in a coverage action in the New York State Supreme Court, New York County. That court ruled that coverage under one of Squibb's primary insurance policies was triggered by manifestation of DES injury. See American Motorists Ins. Co. v. E.R. Squibb & Sons, Inc., 95 Misc. 2d 222, 406 N.Y.S.2d 658 (Sup. Ct. N.Y. Co. 1978). Because the coverage agreement of that policy is identical to the coverage agreements at issue in this federal case, and because the state court adopted the coverage arguments set forth by Squibb at the time, one of the Excess Insurers moved for summary judgment on the ground that Squibb is collaterally and judicially estopped from arguing that coverage under the excess policies is triggered by anything other than manifestation of injury. The district court denied summary judgment on the ground that estoppel would be inappropriate given the unsettled nature of the law concerning the trigger of insurance law for toxic torts. This Court reviews that ruling de novo, considering the record in the light most favorable to Squibb and drawing all reasonable inferences in Squibb's favor. See Levy v. Kosher Overseers Ass'n of Am., 104 F.3d 38, 41 (2d Cir. 1997).

We agree that Squibb is not estopped by the 1978 decision. Under New York law, a party is precluded "from relitigating 'an issue which has previously been decided against him in a proceeding in which he had a fair opportunity to fully litigate the point.'" Kaufman v. Eli Lilly & Co., 65 N.Y.2d 449, 482 N.E.2d 63, 67, 492 N.Y.S.2d 584 (N.Y. 1985) (quoting Gilberg v. Barbieri, 53 N.Y.2d 285, 423 N.E.2d 807, 808, 441 N.Y.S.2d 49 (N.Y. 1981)).

The state court considered whether coverage is triggered under similar policies at the time injury manifested or whether it is triggered at the time of exposure. See American Motorists Ins., 406 N.Y.S.2d at 659-61. In holding that coverage is triggered at manifestation, however, the state court did not hold that manifestation is the exclusive trigger of coverage on which a policyholder may rely, nor did the state court explicitly reject any of the trigger theories now advanced by Squibb in this case. See id.

Squibb's recovery in 1978 on a then available theory satisfactory to its coverage needs does not preclude reliance on a later interpretation of similar coverage agreements.

In American Home Products Corp. v. Liberty Mutual Insurance Co., 748 F.2d 760, 764-65 (2d Cir. 1984), this Court held under New York law that coverage for DES claims is triggered at times of actual injury -- "injury-in-fact." See also Abex Corp. v. Maryland Cas. Co., 252 U.S. App. D.C. 297, 790 F.2d 119, 124-26 (D.C. Cir. 1986) (adopting injury-in-fact trigger set forth in American Home Products in case governed by New York law); Continental Cas. Co. v. Rapid-American Corp., 80 N.Y.2d 640, 609 N.E.2d 506, 511, 593 N.Y.S.2d 966 (N.Y. 1993) (applying injury-in-fact trigger and acknowledging that American Home Products and Abex "concluded that the 'injury-in-fact' rule is most consistent with New York law"). Injury-in-fact often coincides with manifestation; so American Home Products is not at odds with the 1978 decision. Rather, American Home Products provides a fuller interpretation of the relevant policy language, recognizing that "some types of injury to the body occur prior to the appearance of any symptoms; thus, the manifestation of the injury may well occur after the injury itself." American Home Prods., 748 F.2d at 764; see also Continental Cas., 609 N.E.2d at 511 (describing the injury-in-fact test as "resting on when the injury, sickness, disease or disability actually began"). New York trigger law evolved after the 1978 decision, expanding upon and refining the principles that controlled the 1978 ruling.

In arguing that the American Home Products injury-in-fact trigger applies now, Squibb is not changing position. It did not argue against injury-in-fact in 1978, and the 1978 court did not reject it. Injury-in-fact is a lineal development and refinement of the manifestation concept.

The district court held that the injury-in-fact trigger set forth in American Home Products governed the Excess Insurers' obligations. The sole appellate challenge to that legal ruling is the estoppel argument rejected above. The Excess Insurers do, however, appeal the district court's preclusion of evidence bearing on when DES injury-in-fact actually occurs in the second generation. Over the Excess Insurers' objections, the court granted Squibb's pre-trial motion to "preclude the taking of evidence concerning . . . whether there is any causal connection between the drug diethylstilbestrol ("DES") and the injuries alleged in the underlying tort actions," and instructed the jury on its verdict form that, "it is not disputed here that cancer and cancer-related injuries ascribed to DES take place on exposure to DES and continuously during the period beginning five years prior to diagnosis." The district court acknowledged that the Excess Insurers agreed to that stipulation in the verdict form only because the disputed evidentiary ruling dictated that result.

The Excess Insurers now argue: 1) that the district court improperly precluded them from presenting evidence disputing that injury-in-fact occurred during the relevant policy periods for squamous cell cancer and for the ovarian, breast and testicular forms of cancer; and 2) that the verdict form's instruction was error. We review the district court's evidentiary ruling for abuse of discretion, see Barrett v. Orange County Human Rights Comm'n, 194 F.3d 341, 346 (2d Cir. 1999), finding reversible error only if the ruling affects a party's substantial rights, see Schering Corp. v. Pfizer Inc., 189 F.3d 218, 224 (2d Cir. 1999). The instructions for a verdict form are reviewed de novo; we find an instruction erroneous if it "misleads the jury as to the correct legal standard or does not adequately inform the jury on the law," but we reverse on that basis only if a party was prejudiced by the error. Anderson v. Branen, 17 F.3d 552, 556 (2d Cir. 1994).

Squibb argues preliminarily that the Excess Insurers waived these arguments by entering into the stipulation that formed the basis of the verdict form instruction and by failing to object when the verdict form was given to the jury. We disagree. When the Excess Insurers entered into the stipulation, the district court expressly acknowledged their objection to the evidentiary ruling and recited that they had preserved their right to appeal that ruling notwithstanding the stipulation. On the last day of trial the district court received an offer of the proof that the Excess Insurers would have presented but for the evidentiary ruling. When Squibb objected, the district court once again acknowledged that the Excess Insurers had preserved their right to appeal, ruling that the offer of proof would be permitted so "the Court of Appeals [would] know[] all the things that I kept [the Excess Insurers] from proving to [the] jury." The record thus is clear that the Excess Insurers preserved their right to appeal the evidentiary ruling, and that the verdict form instruction was a corollary of that ruling. In light of previous objections, the Excess Insurers' failure to repeat their objection when the verdict form was explained to the jury does not constitute waiver. See City of St. Louis v. Praprotnik, 485 U.S. 112, 119-121, 99 L. Ed. 2d 107, 108 S. Ct. 915 (1998) (finding no waiver where same legal issue was raised by motions and jury instruction, and party repeatedly objected to rulings on motions but did not object to jury charge); 9A Charles A. Wright & Arthur R. Miller, Federal Practice and Procedure 2553, at 411 (2d ed. 1995 & 1997 Supp.) (no need to reiterate objection at time of jury charge if "the party's position previously has been made clear to the trial judge and it is plain that a further objection would be unavailing").

As to the merits, the Excess Insurers do not contest that DES can cause all of the cancers alleged in the underlying bodily injury claims. The Excess Insurers also concede that with respect to most of these cancers, which stem from mutations that began in utero upon exposure to DES, the medical evidence they proffered would have confirmed the district court's assumption that injury occurred at exposure. However, with respect to squamous cell cancer and the ovarian, breast and testicular types of cancer, the Excess Insurers proffered expert testimony that the onset of disease was delayed until after puberty and that "no physical event associated with exposure to DES in utero had any bearing on the subsequent development of the cancer . . . ." The Excess Insurers argue that this evidentiary proffer refutes the district court's instruction that for these cancers injury occurs at exposure and continuously thereafter.

Evaluating their contention requires an understanding of whether Squibb may recover for these specific second-generation claims. Not surprisingly, the Excess Insurers also challenge Squibb's recovery for these claims. They argue that Squibb offered no evidence showing that these injuries occurred at the point of exposure to DES in utero, and instead relied on the district court's allegedly erroneous evidentiary ruling. While this argument could be construed as a challenge to the sufficiency of the evidence, the Excess Insurers did not move for judgment as a matter of law at the close of Squibb's case and thus cannot raise a sufficiency argument on appeal. See Fed. R. Civ. P. 51. We therefore construe this argument to say that the errors the district court allegedly made were not harmless.

Notably, however, the Excess Insurers concede that the second generation claimants' only DES exposure was in utero, and that DES ultimately contributes to the development of these cancers after puberty and (in the case of some cancers) after the initiation of sexual activity. For instance, the appellate briefs of the Excess Insurers state:

--"it was undisputed that cancer occurs through a 'multi-step' process during which cells accumulate a series of genetic mutations in sequence over time, usually many years."

-- "Since DES is eliminated from the body of the mother and the fetus within days, the drug has a direct physiological effect on the fetus only during the short period of exposure to the drug in utero. Thus, to the extent that DES causes or contributes to the development of any form of cancer, including squamous cell cancer, it does so by causing a genetic mutation in utero." (citations omitted).

Despite these concessions, the Excess Insurers argue that their experts, if permitted to testify, would have established that there was no physical effect in utero, and hence no injury-in-fact.

This argument depends on a line drawn between (i) causation and (ii) harmful effect, which (the Excess Insurers argue) is alone sufficient to trigger coverage. This distinction between causation and effect finds support in American Home Products. See American Home Prods., 748 F.2d at 764 ("Since a cause normally precedes its effect, it is plain that an injury could occur during the policy period although the exposure that caused it preceded that period."). However, American Home Products also concluded that "the policies plainly give coverage for injury that occurred during the policy period and was caused by exposure to AHP products [including DES]." Id. at 765 (emphasis added); see also Stonewall Ins. Co. v. Asbestos Claims Mgmt. Corp., 73 F.3d 1178, 1194 (2d Cir. 1995)("Our inquiry as to how New York would answer the triggering inquiry begins with American Home Products . . . where we noted that under New York law, coverage is based upon the occurrence of an injury-in-fact during the policy period.").

Injury-in-fact is not manifestation -- the coverage theory that depends on detectable symptoms. Nothing in American Home Products precludes a finding -- like that made by the jury -- that injury-in-fact encompasses the initial onset of an injury however delayed its symptomatic effects may be. The application of injury-in-fact analysis to progressive diseases "permits triggering at various points when evidence shows injury to have occurred." Stonewall, 73 F.3d at 1195. For asbestosis, for instance, one of these points can be the latency period, during which time the disease develops undetected. See id. at 1198.

We conclude that injury-in-fact can also include, in appropriate circumstances, the inevitable pre-disposition to illness or disability as a result of cell mutation caused by DES. It is undisputed by the parties that second generation claimants, as a result of exposure to DES in utero, develop the cancers at issue as part of their normal human development and experience, and will in fact contract one or more of these cancers if only they survive to puberty and become sexually active.

The Excess Insurers argue that there was no injury-in-fact during their policy periods precisely because the second generation claimants were in utero or in swaddling clothes during the relevant policy periods and, depending on other contingencies and life choices, might not live to maturity or become sexually active. We disagree.

This is not a case of a risk or predisposition that is heightened or discounted by other contingencies, choices and influences. The mutation here is a present injury because it triggers cancer without intermediate contingencies other than maturity and sexual activity, events that are not properly characterized as intervening contingencies sufficient to refute the idea of present injury. Puberty is not certain, but it cannot be viewed as an intervening contingency because nowadays and in this country it is only somewhat less certain than ultimate death. Sexual activity entails choice (as the Excess Insurers argue), but a condition that forecloses or limits that choice, let alone a condition that offers the choice between sexual abstinence and a potentially fatal disease, is already an injury-in-fact. For some of these reasons, the Supreme Court has held that reproduction is a major life activity, and that a health impairment that limits that activity constitutes a disability. See Bragdon v. Abbott, 524 U.S. 624, 637-42, 141 L. Ed. 2d 540, 118 S. Ct. 2196 (1998). The mutation at issue is functionally equivalent to the implant of a time bomb with a short and reliable fuse.

Because we interpret injury-in-fact to encompass a genetic mutation in utero that causes a predisposition to cancer unavoidable except by death before puberty or by permanent sexual abstinence, we find that the district court did not err in precluding the Excess Insurers from offering additional evidence of the impact of later contingencies on the development of the cancers at issue. Their proposed proof either threatened to reopen the question of causation, a question they had conceded, or was irrelevant because it raised only the question of whether a further injury-in-fact occurred. The district court therefore did not abuse its discretion in excluding such evidence, and its instruction on the jury verdict form did not relieve Squibb from its burden of proof.

One injury suffered by members of the second DES generation is an abnormality of the cervix that impairs their ability to carry children to term. The resulting premature deliveries cause birth defects of the lung, heart, eye, and brain (including cerebral palsy) in the next generation. These grandchildren of women who ingested DES during pregnancy are the third-generation claimants. See, e.g., Enright v. Eli Lilly & Co., 77 N.Y.2d 377, 570 N.E.2d 198, 199-200, 568 N.Y.S.2d 550 (N.Y. 1991).

The district court granted summary judgment in favor of Squibb on the question of whether the policies in question cover claims presented by third-generation claimants. The court reasoned that the third-generation claimants' injuries are covered because they were causal consequences of "occurrences" -- in utero injuries-in-fact to the reproductive systems of the second DES generation -- that took place during the period of coverage. It explained that "once there is an occurrence within the policy period the insurer can be liable for injuries resulting from that occurrence that take place in subsequent periods." We review de novo the district court's grant of summary judgment, see Commercial Union Assurance Co. v. Oak Park Marina, Inc., 198 F.3d 55, 59 (2d Cir. 1999), and we affirm.

On appeal, there is no dispute as to the three key facts: (1) members of the second DES generation suffered injury-in-fact in utero during the relevant policy periods; (2) the third-generation claimants' injuries are consequences of their mothers' in utero injuries; and (3) the third-generation claimants' injuries did not occur until after the relevant policy periods. The applicable policy language provides coverage for "all sums which the Assured shall be obligated to pay by reason of the liability . . . for damages, direct or consequential . . . [imposed for personal injuries] . . . caused by or arising out of each occurrence happening anywhere in the world."

We agree with the district court that the policy means what it says when it declares that coverage extends to all personal injuries "caused by or arising out of each occurrence." There is no dispute that the third-generation injuries are caused by or arise out of a covered occurrence. The Excess Insurers' protestations that, during the relevant policy periods, there were no injuries-in-fact to the third-generation claimants (who, after all, were years away from conception) are simply beside the point. This is so unless we are to read into the policy the limitation that personal injuries caused by an occurrence are covered only if the same person suffers both the triggering injury-in-fact and the later consequential injury.

We find no basis for imposing such a restriction on these broadly written policies. Cf. Uniroyal, Inc. v. Home Ins. Co., 707 F. Supp. 1368, 1376 (E.D.N.Y. 1988) (noting, in an Agent Orange products liability case, that New York's canon of construing insurance policies against the insurer "would seem to have special vigor when applied to a policy . . . which is by its own terms denominated a 'comprehensive general liability policy.'") (quoting National Screen Serv. Corp. v. United States Fidelity and Guar. Co., 364 F.2d 275, 279-80 (2d Cir. 1966)). Every court to consider similar issues, including a New York court, has agreed that when a policy provides coverage for consequential damages, such coverage extends to claims by persons other than the one whose injury triggered coverage. In County of Chemung v. Hartford Casualty Insurance Co., 130 Misc. 2d 648, 496 N.Y.S.2d 933 (Sup. Ct. 1985), for instance, the New York court considered claims arising from a policy under which coverage was triggered by the rape of a child. The court held that the insurer was required to indemnify the policyholder for the resulting damages recovered by the child's parents, reasoning that these were damages "arising out of the bodily injury to [their daughter]." 496 N.Y.S.2d at 936; see also Commercial Union Ins. Co. v. Gonzalez Rivera, 358 F.2d 480, 483-84 (1st Cir. 1966) (awarding, under a consequential damages theory, indemnity for damages granted to compensate children for their suffering on account of their father's accident, where it was the father's accident that triggered insurance coverage); Danek v. Hommer, 28 N.J. Super. 68, 100 A.2d 198, 203 (N.J. Super. Ct. App. Div. 1953) (requiring indemnity for a husband's damages deriving from his wife's injuries, where it was her injuries that triggered coverage under her employer's policy). Moreover, in the one other decision to address an issue substantially similar to the one presented here, the court ruled, under New York law, that third-generation DES claims were within the scope of a nearly identical policy because "so long as the subsequent injury may be considered a consequence of the first [triggering injury], it may be considered part of the same accident." Burroughs Wellcome Co. v. Commercial Union Ins. Co., 632 F. Supp. 1213, 1221 (S.D.N.Y. 1986). Judge Leisure's opinion in Burroughs Wellcome has received no hint of disapproval from the New York courts.

The Excess Insurers observe that, in addition to affecting persons other than those suffering injury-in-fact during the relevant period, the third-generation injuries are, of necessity, quite remote in time from the triggering injuries-in-fact.5 Their point in both respects, we take it, is essentially that there may be something unfair, and perhaps unmanageable, about requiring coverage for claims that are, in some sense, far removed from the events triggering coverage. This objection has some surface appeal. But that appeal diminishes rapidly when one realizes that the underlying tort law governing consequential claims itself provides limits on insurers' liability. For the policies confine themselves to indemnification for sums paid "by reason of . . . liability," and hence do not apply to any injuries deemed too distant to give rise to tort liability.6 In this case, the Excess Insurers have not argued that Squibb's settlements were not made "by reason of liability."

To the extent that it seeks to deny coverage even when the propriety of payments on account of possible liability was conceded, the Excess Insurers proposed reading has the strong disadvantage that it tends to create unintended exposure for potential tortfeasors. There is no indication in the policy, however, that Squibb meant to bear the risks at issue. Denying coverage here would run contrary to the sound practice of construing insurance policies in a manner "related both to time on the risk and the degree of risk assumed," Stonewall Ins. Co., 73 F.3d at 1203 (quoting Owens-Illinois Inc. v. United Ins. Co., 138 N.J. 437, 650 A.2d 974, 995 (N.J. 1994)). It would impute to Squibb the intent to self-insure7 with respect to one class of risks -- that reproductive injuries would have intergenerational effects -- in the absence of any basis in the policy for such distinctions among the types of injury that might be "caused by or arise[] out of" triggering occurrences. Cf. Stonewall, 73 F.3d at 1203-04 (distinguishing between decisions to assume a risk and the assumption of risk because no insurance is available). This we decline to do.

Given that the terms of the agreement at issue, the case law, and the relevant considerations of policy all point toward coverage of the third-generation claims, we think the resolution of this question is sufficiently clear that its certification to the New York Court of Appeals is inappropriate. In this respect we are, naturally, also influenced by the failure of any party to seek such certification, by the eighteen or so years that this case has been sub judice, and by the parties' joint, and tenacious, desire that their dispute be decided in the federal courts.

Accordingly, we affirm the district court's decision with respect to the third-generation claims.

The Excess Insurers also appeal from the district court's method of allocating Squibb's losses among the various insurers. See E.R. Squibb & Sons, Inc. v. Accident and Cas. Ins. Co., 1997 U.S. Dist. LEXIS 6674, No. 82 Civ. 7327(JSM), 1997 WL 251548 (S.D.N.Y. May 13, 1997). In particular, they argue that the district court's approach resulted in an improper "double recovery" for Squibb by virtue of the fact that its allocation scheme attributed to the Excess Insurers excessive responsibility for certain claims that had been the subject of settlements with the primary insurers. We conclude that, in the particular circumstances of this case, the district court took the best path available through this complex area.

The parties agree that the appropriate starting point for the analysis is the pro rata allocation method elaborated in Stonewall. Under this approach, the insured's loss on any particular claim is attributed in shares to each insurer, with each insurer's responsibility calculated by multiplying the loss "by a fraction that has as its denominator the entire number of years of the claimant's injury, and as its numerator the number of years within that period when the policy was in effect." Stonewall, 73 F.3d 1178, 1202.9 The complications in this case arise from the fact that Squibb carried both primary and excess insurance, and that it has settled with its primary insurers. The question, then, is how the allocation scheme should account for those settlements.

The district court approached this problem by first determining the pro rata shares (taking into account that the Excess Insurers' coverage came into play only after the primary insurers' policy limits were exhausted) as if there had been no settlements. It then simply treated the settling insurers' portions as satisfied by the settlements, regardless of the actual settlement amounts. The result was that, as regards the Excess Insurers, their obligations were exactly as if there had been no settlements.

Appellants contend, however, that they were entitled to an allocation that left them better off on account of the fact that others had settled. The reason, they postulate, is that the settlement agreements, while exhausting in aggregate the primary insurers' policy limits, allocated settlement funds towards particular claims in amounts that, in some cases, exceeded the settling insurers' pro rata share.10 Because, for such claims, the sum of the settlement amounts plus the Excess Insurers' pro rata share is greater than the total value of the claims, appellants argue that Squibb has unfairly recovered twice on these claims.

This argument is not without force, but it is ultimately unpersuasive in large part because, while Squibb may have gained from the settlements, it undoubtedly took the risk that the size of the settlements would be inadequate to cover the settling insurers' pro rata share. In such a case Squibb would have been left holding the bag. We have little to add to the district court's analysis of the terms of the settlement and insurance agreements, or to its application of relevant policy considerations, and affirm for substantially the reasons given in Judge Martin's opinion. See also Koppers Co. v Aetna Cas. & Sur. Co., 98 F.3d 1440, 1452-54 (3d Cir. 1996) (reducing the excess insurers liability by the settling primary insurers' pro rata shares by relying on the policy limits of the primary policies, not the actual settlement amounts). Under the court's approach, the settling parties are the ones who took the risk of the settlement, and the non-settling parties are left precisely as they would have been had no settlement occurred. That hardly seems unfair.11

Accordingly, we affirm the district court's application of its allocation scheme.

The Excess Insurers argue that the district court's judgment is also flawed because it does not permit reallocation of claims if post-judgment information alters the date of loss, as is expected to happen. However, we conclude that it was within the district court's discretion to enter a final judgment that would not be threatened by inefficient, costly and, in many cases, unnecessary retroactive reallocations. We also conclude that it was within the district court's discretion to establish a protocol for the payment of future claims, namely the requirement that the Excess Insurers are obliged to pay on future claims within 30 days of presentment, because the Excess Insurers' policies contain language to that effect.

Among the Excess Insurers is a subgroup of insurers consisting of various underwriters at Lloyd's and companies in the London insurance market ("London"). In 1967, a $ 100,000 deductible was incorporated into the excess policies of these London insurers for the period January 1, 1967 through January 1, 1970. London argues that the district court erred in ruling that this deductible applies only in the event of non-concurrency between the London excess policy period and the underlying primary policy period. We find this argument to be without merit.

"Under New York law, whether a written contract is ambiguous is a question of law that we review de novo." Bouzo v. Citibank, N.A., 96 F.3d 51, 58 (2d Cir. 1996). Ambiguous language is that which suggests "more than one meaning when viewed objectively by a reasonably intelligent person who has examined the context of the entire integrated agreement and who is cognizant of the customs, practices, usages and terminology as generally understood in the particular trade or business." Lightfoot v. Union Carbide Corp., 110 F.3d 898, 906 (2d Cir. 1997) (internal quotation marks omitted). We review a district court's finding regarding the meaning of ambiguous terms for clear error, bearing in mind that "where there are two permissible views of the evidence, the factfinder's choice between them cannot be clearly erroneous." Haber v. St. Paul Guardian Ins. Co., 137 F.3d 691, 695 (2d Cir. 1998) (quoting Anderson v. Bessemer City, 470 U.S. 564, 574, 84 L. Ed. 2d 518, 105 S. Ct. 1504 (1985)) (alteration in original).

The deductible at issue appears in an endorsement entitled "Exhaustion of Aggregate Endorsement," which reads in full:

1. Whereas, the period of the Primary and/or Underlying policy or policies, including renewal or replacement thereof, with respect to which this policy applies in excess is or may be non-concurrent with the period of this policy.

2. Now, therefore, in consideration of the premium for which this policy is written, in the event of reduction or exhaustion of the aggregate limit or limits contained in such Primary and/or Underlying policy or policies solely by payment of losses in respect to accident or occurrences during the period of such Primary and/or Underlying policy or policies it is hereby understood and agreed that such insurance as is afforded by this policy shall apply, in excess of the reduced underlying limit or, if such limit is exhausted, shall apply as underlying insurance, notwithstanding anything to the contrary in the terms and conditions of this policy.

3. In respect to Products Liability this insurance shall be applicable only after exhaustion of the underlying primary aggregate limit and then only after the application of a $ 100,000 deductible any one occurrence.

While the endorsement's layout and the use of "whereas" suggest that the first paragraph qualifies the third, the third paragraph does not specifically refer to the first or incorporate its terms. The district court thus ruled that the language of the endorsement is ambiguous. We agree with this assessment.12

At trial, the parties introduced considerable evidence to help resolve the ambiguity. Squibb presented the expert testimony of an experienced London insurance professional, who stated that the endorsement is "a widely recognized standard endorsement addressed to the problem of non-concurrency." Squibb's expert also explained that the parties might have wanted to include a deductible that applies only in the event of non-concurrency because in such a situation there is a risk that the primary coverage might exhaust before the excess coverage begins. The district court apparently credited this witness's testimony and therefore concluded that the deductible applies only in periods of non-concurrency.

London contends that in so ruling the district court wrongly disregarded contemporaneous evidence to the contrary. The evidence to which London refers is the London underwriter's handwritten scrawl on the slip, which reads "in regard Products, after exhaustion primary, this policy subject deductible $ 100,000 a.o.occ.," and correspondence from June and July 1968 indicating that London advised Squibb's new brokers that Squibb's existing coverage was subject to a $ 100,000 per occurrence deductible. We do not believe that the district court clearly erred in finding this evidence unpersuasive. Although the underwriter's scrawl does not include an explicit qualification that the deductible applies only during periods of non-concurrency, the scrawl appears adjacent to a line that refers to the "Exhaustion of Aggregate Endorsement," suggesting that its terms would ultimately be placed within the endorsement and be subject to the endorsement's limitations. The persuasiveness of the 1968 letters, meanwhile, is limited by the fact that they were written as part of the process of modifying Squibb's coverage a year after the deductible was incorporated into the 1967 excess policies. Moreover, the evidence in the record from 1968 is mixed at best, as a June 1968 notation on a memo from an insurance executive at Squibb's parent indicates that the deductible applies "only to [claims] where there is no primary insurance." Given this and all of the other evidence presented on this question, the district court did not clearly err in its interpretation of the endorsement.

The Excess Insurers take issue with a portion of the final judgment's preamble that incorporates the rulings set forth in an order issued by the district court on October 10, 1995. The October 10, 1995 order includes, inter alia, a declaration that "the settlements arrived at to date between Plaintiff and the primary-level and other settling insurers were the result of non-collusive, arm's length negotiations" and a declaration that each of the non-settling defendants except Utica Mutual Insurance Company "is obliged to indemnify Squibb for all settlements and judgments and to pay all defense-related expenses on the same bases and criteria as that defendant's underlying carrier has committed to pay." Although the Excess Insurers do not object to the former ruling, they argue that, if the latter ruling is not surplusage in the context of the final judgment, it is based on a misreading of the excess policies and improperly suggests that the excess insurers might be called upon to pay pursuant to the bases and criteria of Squibb's settlements with its primary insurers. In response, Squibb makes only one argument: that the preamble's reference to the October 10, 1995 order must be maintained in order to preserve that order's ruling that the settlements were the result of arm's length negotiations. Because Squibb does not argue that the "bases and criteria" ruling is necessary to the final judgment, we assume that it now concedes that the preamble's incorporation of that ruling adds nothing to the judgment. Given this concession that the "bases and criteria" ruling is surplusage in the final judgment, there is no need for us to consider further the Excess Insurers' argument on this point.

Certain of the Excess Insurers are known as "high level" excess insurers whose coverage begins at amounts well above the primary limits. Five high level excess insurers remain in this action: Insurance Company of the State of Pennsylvania, whose coverage begins at $ 50 million excess of primary limits; American Home Assurance Company, whose coverage begins at $ 30 million excess of primary limits; Northbrook Excess and Surplus Insurance Company, whose coverage begins at $ 10 million excess of primary limits; and Continental Casualty Company ("CNA") and Commercial Union Insurance Company, whose coverage begins at $ 5 million excess of primary limits. Together, these five insurers contend that the district court erred in denying their motions to dismiss and motions for summary judgment based on the lack of an actual case or controversy. We disagree.

Although Squibb initiated this action in 1982 and all five high level excess insurers were named as defendants by 1984, the district court was not asked to rule on this issue until February 1993. At that time, the high level excess insurers jointly moved to dismiss for lack of an actual case or controversy on the ground that their coverage would never be reached. American Home Assurance Company and Insurance Company of the State of Pennsylvania also filed a separate motion to dismiss due to their very high attachment points. In support of the motions, the high level excess insurers noted that Squibb had never submitted a claim to any of them and that, even under Squibb's own data, coverage under the primary policies had not been fully exhausted.

By Memorandum Order dated May 20, 1994, the district court denied the motions, stating:

Statistical evidence which need not be set forth in detail here makes it clear that a significant portion of Squibb's DES exposure remains in the future, and that it is highly uncertain whether primary carriers' policies will cover it.13 Further, the likelihood that primary carriers' policy limits will be exhausted is greatly increased if, as required by the policy language as interpreted here and contrary to the moving carriers' position, policy limits of the primary carriers may be exhausted by settlement as well as payment.14 Under the circumstances, declaratory relief to settle the legal relations between the parties is useful and appropriate.

All excess carrier defendants have joined in the contest of the substantive issues in the case, which is at least obliquely contrary to the argument that they have no interest in the outcome. The complaint was filed in 1982; no motions challenging existence of a genuine controversy were filed until February 15, 1993, eleven years after the filing of this case and almost a year after the April 1992 order. While such delay would not create a genuine controversy were one absent, it is a significant indication that only an afterthought is involved.

E.R. Squibb & Sons, Inc. v. Accident & Cas. Ins. Co., 853 F. Supp. 98, 102 & nn.6-7 (S.D.N.Y. 1994) (citation omitted) (footnotes in original). The district court subsequently reaffirmed its ruling after granting a motion for reconsideration.

On April 19, 1996, the high level excess insurers filed a joint motion for summary judgment based on the absence of an actual case or controversy. American Home Assurance Company and Insurance Company of the State of Pennsylvania again filed an additional separate motion due to their very high attachment points. At a conference held May 17, 1996, the district court stated:

I am going to deny the motions to dismiss. It seems to me that there are many issues still to be resolved in this lawsuit, and the ultimate resolution of those issues can very well affect the liability of each of the excess carriers. Clearly, the statistical analysis submitted by Squibb with respect to the following of settlements is one that may ultimately not apply in this case. There are issues to be tried that would determine when injury in fact occurs. It is very conceivable that that could end up placing in a single policy year a substantial number of claims that might very well reach the levels of each of the carriers in this litigation, and there is a substantial enough controversy that the matter should be fully resolved in this case and not left to future litigation when the claims actually are made against the excess policy.

The district court thus repeatedly ruled that Squibb's declaratory judgment action against the high level excess insurers presented an actual case or controversy. This Court recently questioned in dicta whether we review such rulings de novo or for abuse of discretion in light of the Supreme Court's decision in Wilton v. Seven Falls Co., 515 U.S. 277, 132 L. Ed. 2d 214, 115 S. Ct. 2137 (1995), which held that district courts have a "unique breadth of . . . discretion to decline to enter a declaratory judgment," id. at 287. See Certain Underwriters at Lloyd's, London v. St. Joe Minerals Corp., 90 F.3d 671, 675 (2d Cir. 1996) (citing a line of cases applying de novo review to a district court's conclusion that a declaratory judgment action presents an actual case or controversy, but suggesting that Wilton casts doubt on the vitality of these cases). The fact that the high level excess insurers challenge the district court's subject matter jurisdiction to hear this action rather than its discretionary decision to exercise jurisdiction under the Declaratory Judgment Act would seem to suggest that de novo review is appropriate. See American States Ins. Co. v. Bailey, 133 F.3d 363, 368 (5th Cir. 1998) (applying de novo review to a district court's determination that a declaratory judgment action presented an actual case or controversy, while reviewing the district court's decision to exercise its declaratory judgment jurisdiction for abuse of discretion under Wilton). However, we need not resolve this question because, for the reasons that follow, we would affirm the district court's rulings under either standard.

The legal principles governing the high level excess insurers' challenge are well-established. A party seeking a declaratory judgment bears the burden of proving that the district court has jurisdiction. See Cardinal Chem. Co. v. Morton Int'l, Inc., 508 U.S. 83, 95, 124 L. Ed. 2d 1, 113 S. Ct. 1967 (1993). Jurisdiction exists only if there is an "actual controversy," 28 U.S.C. 2201(a); see also Jefferson v. Abrams, 747 F.2d 94, 96 (2d Cir. 1984) ("Article III, Section 2 of the United States Constitution limits federal court jurisdiction to actual cases and controversies."), which has been defined as one that is "'real and substantial . . . admitting of specific relief through a decree of a conclusive character, as distinguished from an opinion advising what the law would be upon a hypothetical state of facts.'" Olin Corp. v. Consolidated Aluminum Corp., 5 F.3d 10, 17 (2d Cir. 1993) (quoting Aetna Life Ins. Co. v. Haworth, 300 U.S. 227, 241, 81 L. Ed. 617, 57 S. Ct. 461 (1937)). As we stated in Associated Indemnity Corp. v. Fairchild Industries, Inc., 961 F.2d 32 (2d Cir. 1992):

Like any other action brought in federal court, a declaratory judgment is available to resolve a real question of conflicting legal interests. That the liability may be contingent does not necessarily defeat jurisdiction of a declaratory judgment action. Rather, courts should focus on the practical likelihood that the contingencies will occur[]. Indeed, litigation over insurance coverage has become the paradigm for asserting jurisdiction despite future contingencies that will determine whether a controversy ever actually becomes real.

Id. at 35 (citations and internal quotation marks omitted).

Given these standards, the parties agree that the central question is whether the district court erred in concluding that there was a "practical likelihood" that the high level excess carriers' policies would be reached. Given the record before us, we find no error. As of 1993, Squibb had incurred approximately $ 100 million in damages. Further, an estimated 900 DES claims were still pending against Squibb, and Squibb's expert averred that it was reasonable to predict that Squibb would incur $ 100 million in additional DES-related indemnity and defense expenses in the future. As Squibb notes, there was also a substantial possibility that its expenses would exceed this projection because it was impossible to predict fully how many cases would be filed, how substantial their verdicts would be, and whether changes in state law would revive time-barred claims. Moreover, a number of issues remained to be decided at the time the motions were filed that could have affected the high level excess insurers' liability, including the jury's trigger determinations. Accordingly, the district court was correct to rule that Squibb's claims against the high level excess insurers presented an actual case or controversy.

CNA argues that the district court erred by ruling in its October 10, 1995 order that three CNA policies covering the period effective January 1, 1971 through January 1, 1976 obligate CNA to pay Squibb's defense-related expenses based upon the terms of the underlying carriers' policies. Each of the three CNA policies includes an endorsement with the following language:

In the event of reduction or exhaustion of the aggregate limit or limits designated in the underlying policy or policies solely by payment of losses in respect to accidents or occurrences during the period of such underlying policy or policies, it is hereby understood and agreed that such insurance as is afforded by this policy shall apply in excess of the reduced underlying limit or, if such limit is exhausted, shall apply as underlying insurance, notwithstanding anything to the contrary in the term and conditions of this policy.

With respect to similar language in policies issued by Aetna Casualty and Surety Company, the district court had granted partial summary judgment in 1992, after concluding that:

where a clause in an Aetna policy provides that coverage 'shall apply as underlying insurance, notwithstanding anything to the contrary in the terms and conditions of this policy,' no prohibition on liability for legal defense costs in that policy shall be applicable where the conditions for applicability of the quoted clause are met.

Based on this ruling and the undisputed fact that the policies underlying the three CNA policies at issue pay defense costs, the district court concluded that CNA is obligated to pay defense costs once its coverage is reached.

CNA argues that, contrary to the district court's ruling, the endorsement plainly provides no coverage for defense costs. CNA first asserts that the endorsement applies only when a specified contingency is met: that is, when one or more underlying policies have been reduced or exhausted by payment of claims arising from occurrences that fall within their policy periods but outside of CNA's policy period. CNA then notes that the district court did not dispute that, aside from the endorsement, the three CNA policies only cover "loss," which is defined not to include defense costs. CNA also points out that each of the three CNA policies includes a "follow form" provision that incorporates the provisions of the immediate underlying policies except for any obligation to pay defense costs. In light of this policy language defining coverage not to include defense costs, CNA argues that the endorsement cannot embrace a defense obligation because, when triggered, it only requires CNA to pay "such insurance as is afforded by this policy." CNA also contends that there can be no defense obligation because the endorsement does not address "defense costs" but instead speaks only of "loss."

In response, Squibb maintains that the district court's ruling was correct because the endorsement's language unambiguously provides coverage for defense costs. Without addressing CNA's argument as to when the endorsement is triggered, Squibb contends that "notwithstanding anything to the contrary in the terms and conditions of this policy" means that CNA must pay defense costs if the underlying insurance does so, even if defense costs are not otherwise covered by its policies. Squibb also argues that the reference to "loss" in the endorsement merely identifies the point of exhaustion of the underlying insurance and does not in any way limit the type of insurance to be provided under the endorsement. Finally, Squibb asserts that "such insurance as is afforded by this policy" is just a reference to the monetary limits of CNA's insurance. In short, both Squibb and CNA contend that the endorsement is unambiguous, but interpret it differently.

We reject both parties' positions. As noted above, we review de novo the district court's determination that a contract is ambiguous. See Bouzo v. Citibank, N.A., 96 F.3d 51, 58 (2d Cir. 1996). "Only where the court finds that the terms are unambiguous, or where no extrinsic evidence exists, may it properly grant summary judgment to one of the parties." Alexander & Alexander Servs., Inc. v. These Certain Underwriters at Lloyd's, London, 136 F.3d 82, 86 (2d Cir. 1998).

Analyzing the terms of the endorsement, we conclude that the district court erred in granting Squibb summary judgment on this issue. As an initial matter, the district court does not appear to have considered the circumstances under which the endorsement is triggered, a matter to which we direct its attention on remand. More fundamentally, we find the endorsement's terms to be ambiguous in at least two respects. First, it is unclear whether the reference to "such insurance as is afforded by this policy" is meant to limit coverage so as not to include defense costs or whether it is intended only to invoke the monetary limits of the CNA policies. Second, it is not clear to what extent "notwithstanding anything to the contrary in the terms and conditions of this policy" trumps the language that appears elsewhere in the CNA policies defining coverage not to include defense costs. In light of this ambiguity, any available extrinsic evidence should have been used in determining the meaning of the endorsement. Because we find no indication in the record that extrinsic evidence is unavailable on this issue, we conclude that the district court erred in resolving this question on summary judgment. We therefore reverse the district court's ruling and remand the case to the district court for further proceedings related to this issue.

For the reasons given above, we affirm the decision of the district court in part and reverse in part. Specifically, we (1) affirm the district court's finding that diversity jurisdiction exists; (2) affirm the district court's holding that Squibb is not estopped by its position in a 1978 state court action; (3) hold that the district court did not abuse its discretion in excluding evidence bearing on when DES injury-in-fact occurs in second-generation claimants nor did its instruction on the jury verdict form relieve Squibb from its burden of proof on this issue; (4) affirm the district court's grant of summary judgment in favor of Squibb as to the third-generation claims; (5) affirm the district court's application of its scheme to allocate losses among the insurers; (6) affirm the district court's decision as to the payments of future claims; (7) affirm the district court's ruling regarding the 1967 deductible; (8) find that the "bases and criteria" portion of the preamble is surplusage in the context of the final judgment; (9) affirm the district court's denial of the high level excess insurers' motions for summary judgment based on the alleged lack of an actual case or controversy; but (10) reverse the district court's grant of summary judgment in favor of Squibb regarding the CNA policies' endorsement and therefore remand the case to the district court for further proceedings related to this issue. Each party shall bear its own appellate costs.

I respectfully dissent (A) from part II.C of the per curiam opinion, which requires coverage for the third-generation claims, and (B) from part II.D, which requires that Excess Insurers pay the policyholder in respect of claims already paid and satisfied by the underlying primary coverages.

The district court granted summary judgment in favor of Squibb on coverage for claims presented by the grandchildren of women who ingested DES ("third-generation" claimants). The majority affirms on the ground that the third-generation claims stem from injuries caused by or arising out of a second-generation injury. See [Maj. Op. at 40]. I do not subscribe to that view.

* * *

Under American Home Products, the only applicable trigger of coverage is injury-in-fact during the policy period. See American Home Prods. Corp. v. Liberty Mutual Ins. Co., 748 F.2d 760, 764-65 (2d Cir. 1984). The third-generation claimants did not exist during the relevant policy periods -- even in utero -- and having no body, could suffer no bodily injury, or become unavoidably predisposed to injury, during those periods. The consequential damages clauses do not alter or affect the American Home Products injury-in-fact analysis.

The majority cites with seeming approval the observation in Uniroyal, Inc. v. Home Ins. Co., 707 F. Supp. 1368, 1376 (E.D.N.Y. 1988), that coverage may be augmented when a contract "is by its own terms denominated a 'comprehensive general liability policy.'" But that denomination does not expand the policy's protections; for example, policyholders should not expect their umbrella policies to keep them dry in the rain. I therefore begin with the relevant policy language, which reads: "Underwriters hereby agree . . . to indemnify the Assured for all sums which the Assured shall be obligated to pay by reason of the liability . . . for damages, direct or consequential and expenses, all as more fully defined by the term 'ultimate net loss' on account of . . . Personal Injuries . . . caused by or arising out of each occurrence happening anywhere in the world." Ultimate net loss is defined as: "The total sum which the Assured . . . becomes obligated to pay by reason of personal injury, property damage, or advertising liability claims, either through adjudication or compromise . . . ."; personal injury is defined as: "Bodily injury, mental injury, mental anguish, shock, sickness, disease, disability . . . ."

The policy wording thus affords coverage for loss on account of personal injury, caused by an occurrence, and payable as damages, direct or consequential. Under this wording, read in light of American Home Products, a covered personal injury must occur in fact during the policy period even though it may be caused by an occurrence that happened during that period or before it, and even though it entails payment of damages suffered during or after the policy period (or even, theoretically, before it).

The causal consideration that bears on coverage is whether the personal injury was caused by an occurrence. This point is not in dispute, as the third-generation claims are caused by the same occurrence as the first and second-generation claims: the ingestion of DES. Notwithstanding that common cause, however, the damages that are insured by the policy (direct or consequential) must be on account of the personal injury that occurs in fact during the policy period. Thus the claims of the first and second-generation claimants themselves, as well as any hypothetical personal injury claims for loss of consortium (by their spouses) or for loss of parental guidance (by their children) are on account of bodily injury in fact suffered in the policy period. By contrast, the damages suffered by the third-generation claims are in respect of and on account of those third-generation claimants' own personal injury. Their injury is caused by the same occurrence, but their injury did not in fact occur during the policy period. Nor are their damages "on account of" an injury that did.

* * *

Consequential damages attributable to sequellae, or progressive or latent injuries, are covered (subject to any allocation or other insurance), but it is very difficult to see how the initial (coverage-triggering) bodily injury can be sustained prior to conception, and impossible to square such a coverage ruling with the settled requirement under New York law that injury-in-fact must occur during a policy period to trigger coverage. In short, the consequential damage clause does not affect the threshold issue of when an injury in fact occurred. See Stonewall, 73 F.3d 1178, 1203 (2d Cir. 1995) (holding that insurers have an "obligation to indemnify for subsequent damages attributable to an injury occurring during the relevant policy period.")

Under our reasoning on the first and second-generation claims, those claimants suffer personal injury on account of their own injuries. Depending on tort law in a given jurisdiction, consequential damages on account of those injuries may reflect third-generation injury to the extent that the inability of second-generation claimants to bear normal children is part of their own harm, and that harm may be compensable by consequential damages that would be on account of the injury that second-generation claimant suffered in the policy period, and therefore covered. But the third-generation claimant suffers and claims damages on account of his or her own injury, and not on account of injury suffered by anyone else. Thus (as the majority points out) a covered claim can be asserted by someone who did not incur the injury-in-fact in the policy period, such as claims for loss of consortium, parental guidance, and so on. But such losses are on account of the bodily injury that occurred, in fact, in the policy period.

When the later victim is the child of the earlier victim, it is easy to succumb to the fallacy of casting the later injury in terms of consequential damages, as the majority has done. That is because an injured parent can theoretically collect as consequential damages for the parent's impaired capacity to bear normal children, or even for the psychological and financial harm of bearing a child with birth defects. But there is a real distinction between (i) consequential damages and (ii) a new injury. That distinction can be illustrated by an example that involves no inter-generational link. If in one period of a hospital's liability policy a patient is injured in fact by the hospital's negligence in allowing the patient to contract a chronic and communicable disease, it cannot be expected that the same policy period is implicated if years later the discharged patient coughs on a co-worker who contracts the disease and sues the hospital. Putting aside the remoteness of causation in tort (which is no more remote than in the third-generation DES claims), I think it is quite clear that the co-worker's injury was not suffered in fact in the original policy period, and that the injury (though a consequence of the hospital's negligence) cannot be deemed consequential damages on account of the patient's injury.

* * *

The majority opinion guesses that New York law will construe the consequential damages clause to provide that when there is injury in fact in a policy period, liability insurance covers subsequent physical injuries to other persons in other times, and that the Court of Appeals will be satisfied to impose no limit on that otherwise limitless chain except such limitation as is afforded by the reach of tort law itself. See [Maj. Op. at 42-43, 45]. I doubt, however, that the scope of tort law is a useful limitation on the contractual meaning of the insurance policies. For one thing, New York does not recognize claims arising from pre-conception torts. See Enright v. Eli Lilly & Co., 77 N.Y.2d 377, 570 N.E.2d 198, 201, 568 N.Y.S.2d 550 (N.Y. 1991). Therefore the majority's own analysis would not support the coverage decision it makes under New York law. It is true (i) that this argument may not bear on claims paid prior to New York's adoption of the bar on recovery for pre-conception torts; and (ii) that the insurance contracts at issue, though construed under New York law, pay loss on claims asserted under the law of many jurisdictions. These provisos only confirm, however, that the shape of New York tort law (or the tort law of any other jurisdiction) is not a useful limiting principle for the scope of products liability insurance under New York law. Second, tort law tends to expand in order to use up any insurance coverages that are arguably available, so that (under the majority opinion) insurance coverage expands to cover widening principles of tort law while tort law itself expands by feeding off insurance coverage.

Finally, the majority opinion has unpredictable ramifications for wrongful birth claims and claims based on genetic mutation. I concur entirely with my colleagues' view that the age and procedural history of the present case militate against certification; but the Court of Appeals presumably will be able to visit these questions before the majority opinion influences the shape of tort law and insurance coverage in New York.

The Excess Insurers argue that the judgment improperly affords a recovery to Squibb for more than 100% of its loss in 191 claims, and allows Squibb to recover more than 100% of its defense costs on 1,999 claims, leading to a total "double recovery" of at least $ 12.38 million. I conclude that a double recovery occurred, and that it was error to allow it.

At trial, the jury determined when coverage was triggered for various DES-related injuries, finding that injury-in-fact occurred upon exposure and several years thereafter. The district court then undertook the heavy labor of deciding, as a matter of law, "the issue of how the amounts for which Squibb has been held liable, or has paid in settlements on the underlying claims, [were] to be allocated among Squibb's insurers." E.R. Squibb & Sons, Inc. v. Accident & Cas. Ins. Co., 1997 U.S. Dist. LEXIS 6674, No. 82 Civ. 7327 (JSM), 1997 WL 251548, at *1 (S.D.N.Y. May 13, 1997). That task is particularly difficult in a case (like this one) in which any given personal injury claim "implicates multiple policies in effect during the multi-year period of the injury process." Stonewall, 73 F.3d at 1201 (2d Cir. 1995). Unlike many such cases, however, the district court did not have to allocate Squibb's loss among the issuers of Squibb's primary insurance policies, because the parties agree that settlement payments by the primary insurers exhausted that underlying insurance. The district court's task was limited therefore to allocating Squibb's remaining loss among the Excess Insurers.

* * *

As the district court held in an earlier stage of this litigation: "For an excess insurer to be liable to Squibb, the requirements of the insurance policy issued to Squibb by that carrier must be satisfied." E.R. Squibb & Sons, Inc. v. Accident & Cas. Ins. Co., 860 F. Supp. 124, 126 (S.D.N.Y. 1994). The relevant requirements are: 1) that covered injury in fact occur during the period covered by the excess policy; and 2) that the insured's underlying insurances -- to which each excess policy is "excess" -- be exhausted. See id. It is undisputed on appeal that these two "separate and independent" requirements (id.) were satisfied, i.e., the policies underlying the excess policies here at issue were exhausted by settlement, and the jury found that each type of injury occurred over multiple policy periods. These requirements also affect the allocation of loss among the excess policies because not every loss is allocated to every policy period, and because the underlying coverages are therefore exhausted at different times. At issue is the district court's determination as to when exhaustion occurred and the district court's subsequent allocation of loss based upon those rulings as to exhaustion.

Since the parties agree that the exhaustion of primary coverage occurred via settlement, the mechanical way to fix the date on which the underlying coverages were exhausted is to add up the settlement amounts properly paid by the underlying policies in respect of each policy period. Years before trial, the district court appeared to have adopted this method. See id. ("Exhaustion may occur by payment or settlement, provided the settlement is noncollusive and at arm's length."); E.R. Squibb & Sons, Inc. v. Accident & Cas. Ins. Co., 853 F. Supp. 98, 99 (S.D.N.Y. 1994) ("The excess carrier is liable when the underlying policy limits of the primary carrier or carriers covering the period of time involved is exhausted, whether by payment or by arm's length settlement with the insured.").

However, the district court rejected that method of ascertaining the exhaustion status of the underlying policies that were issued by one of Squibb's primary insurers, Kemper/Amico. Squibb's settlement with Kemper/Amico specified which sums were paid in respect of which claims, but the district court did not adopt that matching of dollars to claims, and ruled instead that settlement amounts (as distinct from claims paid by Kemper/Amico) should not be assigned to the underlying policies except to the extent that "the law applicable to those policies, including the triggers of coverage which have been established in this case, would obligate those insurers to pay." Squibb, 1997 WL 251548, at *1, *3. This re-allocation was based on the district court's view that "the settlement agreements are not contracts of insurance" and "[do] not reflect the insurer's legal obligation under its contract." Id.

The district court therefore ascertained the point of exhaustion not on the basis of which claims remained to be indemnified after payments were made by the underlying policies, but on "the assumption that the trigger of coverage established in the trial of this case applied to each of the insurers who provided coverage during the relevant years." Id. Employing the allocation scheme of Stonewall, 73 F.3d at 1201, the district court then pro rated Squibb's loss among the primary policies triggered for given claims until the limits of those primary policies were reached, and then pro rated the remainder of Squibb's loss among the Excess Insurers. In so doing, the district court assumed that the exhaustion points of the excess policies were the dollar limits of the primary policies, not the greater amounts actually paid to Squibb in satisfaction of additional claims and in settlement of the primary insurers' obligations under those policies.

* * *

The district court's pro rata allocation would have been a sound way to allocate Squibb's loss among the Excess Insurers if the settlements did not exceed the limits of the primary policies. This Court has recently held that allocation among multiple triggered policies is appropriate under New York law because it is consistent with policy concerns (similar to policy issues raised in this appeal) and serves the cause of efficiency. See Olin Corp. v. Insurance Co. of N. Am., 221 F.3d 307 (2d Cir. 2000). We have also upheld proration that allocated loss to the insured during periods when it chose to be uninsured. See Stonewall, 73 F.3d at 1203.

The district court's pro rata allocation here, however, ignored the fact that the policies of one of the major primary insurers underlying excess policies at issue here were exhausted by a pre-trial insurance settlement that exceeded the limits of those underlying policies.1 The district court's allocation required the Excess Insurers to pay certain losses and costs even though Squibb had already been reimbursed for those losses and costs. The district court thereby permitted Squibb to recover twice on the same loss.

I do not think that an insured may recover more than 100% for any individual claim covered by its indemnity liability policies, even if that claim falls within multiple policy periods.

The majority affirms for substantially the reasons given in the district court's opinion, which justified the allocation as follows: 1) "it is far from clear that, at the end of the day, Squibb will recover more than it pays out. . . . It is doubtful that the ultimate total liability of Squibb will be determined for years and no one can today predict whether Squibb will have paid more or less than the insurance proceeds that it recovers;" and 2) "even if there were to be a windfall, the question is whether that windfall should go to Squibb or the non-settling insurers. . . . Considerations of public policy suggest that, if there is to be a windfall, it should go to Squibb and not the non-settling insurers." Squibb, 1997 WL 251548, at *2.

These rationales cannot be squared with New York insurance law. As discussed in American Home Products Corp. v. Liberty Mutual Insurance Co., 748 F.2d 760, 764-65 (2d Cir. 1984), New York law mandates that coverage be determined on a claim-by-claim basis, according to an injury-in-fact analysis. And it is undisputed that the policies at issue, in accord with traditional principles of indemnity, limit the maximum indemnity per claim at 100% of the value of the claim. New York law does not allow Squibb to make up for under-recovery on certain claims by compelling insurance payments that exceed a 100% recovery on others. New York law also does not allow Squibb to fund anticipated future claims by super-recoveries on other claims.

Indemnity liability insurance reimburses policyholders for covered losses sustained during a given policy period; it is designed "to place the insured as nearly as possible [in the position the insured was in] when the risk began, the object being to make the insured whole." 12 John A. Appleman & Jean Appleman, Insurance Law & Practice 7001, at 11 (1983 ed.). "The policy cannot be made the subject of profit by the insured, and [the insured] may only recover such loss as [the insured] has actually sustained." Id. This applies even when more than one insurer is on the risk. See Goodman v. Allstate Ins. Co., 137 Misc. 2d 963, 523 N.Y.S.2d 391, 394 (N.Y. Sup. Ct. 1987) ("Of course, even though both policies apply, the insured is entitled to only one satisfaction for the loss actually incurred by reason of his liability for damages plaintiff sustained."); 8A Appleman 4907.35, at 354 ("The total amount which the insured can recover from all insurers of the same risk cannot exceed the amount of [the insured's] loss."). It is "a generally accepted fundamental tenet of insurance law that opportunities for net gain to an insured through the receipt of insurance proceeds exceeding a loss should be regarded as inimical to the public interest." Robert E. Keeton & Alan I. Widiss, Insurance Law 3.1(a), at 135 (1988); see also E.R. Squibb & Sons, Inc. v. Accident & Cas. Ins. Co., 860 F. Supp. 124, 126 (S.D.N.Y. 1994) ("Squibb cannot of course recover twice for the same expenses incurred by it."); Silinsky v. State-Wide Ins. Co., 30 A.D.2d 1, 289 N.Y.S.2d 541, 547 (N.Y. App. Div. 1968) ("It is the policy of this court to prevent double recoveries and avoid unjust enrichment by an injured person."); Kahane v. American Motorists Ins. Co., 65 Misc. 2d 1065, 319 N.Y.S.2d 882, 884 (N.Y. Civ. Ct. 1971) ("It need hardly be pointed out that the insured may not under any circumstances recover twice for the same loss or recover more than the value of the loss."). In short, liability indemnity insurance is not a profit center.

* * *

Squibb does not deny that the judgment permits more than 100% recovery on numerous individual claims, and its arguments do not justify such a super-payment in this case. For example, Squibb argues that the Excess Insurers are already receiving a reduction in their liability because the district court, in accordance with Stonewall, pro-rated their liability among the various triggered policies when Squibb was "within its contract rights to designate one policy to pay all of any claim covered by that policy subject of course to exhaustion of the policy limits." Prudential Lines Inc. v. American Steamship Owners Mut. Protectional & Indem. Ass'n, 158 F.3d 65, 83 (2d Cir. 1998) (construing marine policies issued by a mutual insurance association of shipowners). Squibb did not appeal the district court's pro-rata allocation, however. If it had, our recent decision in Olin strongly suggests that the allocation was correct. In any event, the designation of a single excess policy would not have necessarily eliminated the need to determine the exhaustion point of the primary policy underlying each designated excess policy based on the claims paid under that policy.

Squibb assumes the validity of the district court's pro-rata allocation, and argues that the settlements cannot be used to determine the exhaustion points of the underlying policies because the settlements were not claim-specific. And the Excess Insurers agree that where the settlements with the primary insurers were not claim-specific, the exhaustion of those primary policies should be determined based on the policy limits rather than on the settlement amounts. Accordingly, I would remand this case. However, I would not ask the district court on remand to allocate lump-sum settlements to individual claims; it is enough to rely on policy limits to determine the exhaustion points of those primary policies exhausted by the lump-sum settlements. But for the policies exhausted by claim-specific settlements, I would direct the district court to calculate exhaustion using the settlement amounts, or to cap Squibb's recovery for each individual claim at 100%.

Similarly, Squibb argues that the Excess Insurers failed to shoulder their burden of proving that the settlements are valid and collectible insurance under the primary policies. Squibb provides no authority indicating that this is the Excess Insurers' burden to bear. In any event, Squibb's argument proves too much. If the settlements did not operate to exhaust the underlying policies, Squibb would have been unable to reach the excess policies.

Squibb also argues that it reaps no windfall under the district court's judgment because: 1) it has been under-compensated on many claims; and 2) Squibb's damages expert discredited the methodology of the Excess Insurers' damages expert. As discussed above, this first argument impermissibly aggregates the claims for which Squibb is liable, instead of considering them claim-by-claim in accordance with New York law. Moreover, any aggregate under-compensation appears to be the result of deductible clauses in the primary policies and of settlement terms to which Squibb agreed, rather than any insufficiency of payment by the Excess Insurers. With respect to the second argument, the district court made no findings regarding the analysis of either expert, instead excluding such evidence on the ground that it was irrelevant. Squibb's arguments about the reliability and credibility of the Excess Insurers' allocation evidence are therefore beside the point.

Finally, apparently in the alternative, Squibb argues that it is entitled to a windfall under New York's collateral source doctrine, which prohibits tortfeasors from reaping the benefits of settlements that they had no part in producing. However, the Excess Insurers are not tortfeasors, and New York law has refused to equate insurers with tortfeasors. See Silinsky, 289 N.Y.S.2d at 547 ("In our opinion, the insurer and the wrongdoer stand on a very different footing . . . ."). In any event, the Excess Insurers were not in the same position to settle as the primary insurers because, given the nature of excess insurance, the Excess Insurers did not become liable to Squibb until the underlying primary policies were exhausted.

Squibb offers no convincing argument that would permit it to recover indemnity greater than its loss on any given claim.

* * *

I emphasize that the pro rata method of distributing coverage over policy periods is not in question or at issue here. But implicit in pro rata distribution is that the coverage is distributed pro rata for the payment of claims. The radical defect in the majority's position is that it requires the allocation of insurance coverage under indemnity policies for the payment of claims that have already been paid and satisfied, without reopening the accounts on those cases to reallocate the loss in accordance with coverage obligations. The majority justifies this step on the basis in part that the Excess Insurers should not be "better off" because other insurers (whose coverage was first reached) settled first. But I am unwilling to violate the essential character of an indemnity contract in order to avoid an irony. The majority and the district court point out that very likely there will be more than enough loss to go around, and that the double recovery on certain claims, paid pursuant to this judgment, will eventually be paid by Squibb in respect of claims that otherwise would not be covered by insurance. See [Maj. Op. at 49-50]. But that is exactly my objection to this result: indemnity coverage is payable in respect of outstanding covered claims and losses, not in respect of claims that have been already indemnified or (for other reasons) are not payable under the policy wording.