Catlin Specialty Ins. Co. v. J.J. White, Inc. , 387 F. Supp. 3d 583 ( 2019 )

Goldberg, J.

This continuing insurance coverage dispute presents complicated questions regarding indemnification obligations stemming from a wrongful death settlement in the Philadelphia Court of Common Pleas. Through an Opinion issued on February 27, 2018, I resolved cross-motions for summary judgment. Defendant in this matter, J.J. White, Inc. ("J.J. White"), who was a settling party in the Common Pleas case, now understandably asks me to clarify an important portion of that Opinion dealing with proof issues on the coverage question. This follow-up Opinion sets forth that clarification.

I. FACTUAL & PROCEDURAL BACKGROUND ¹

The insurance policy at issue, issued to J.J. White by Catlin Specialty Insurance *586Company ("Catlin"), is entitled "Contractor's Protective Professional and Pollution Legal Liability Insurance Policy (hereinafter the "Pollution Policy").² The relevant coverage provision of the Pollution Policy provides as follows:

The Insurer shall pay on behalf of Insured for a Pollution Loss ...which the Insured has or will become legally obligated to pay as a result a [sic, read "as a result of a"] Pollution Claim first made against the Insured and reported to the Insurer, in writing, during the Policy Period ...from Pollution Conditions ..., provided that ... the Pollution Conditions first occurred on or after the Retroactive Date ....

(Ford Decl., Ex. A § IV.Q.)

The italicized text above is referred to as the "Retroactive Date Provision," and, as I concluded in the Previous Opinion, it limits coverage to injuries from toxic chemical exposures that occurred on or after April 30, 2002 (the "Retroactive Date"). (Previous Op. 2-3.)

The underlying wrongful death suit was brought by the estate of a refinery worker, George Gans, who worked for J.J. White at a Sunoco refinery from 1986 until his diagnosis and death from leukemia in 2010. In the wrongful death suit, Gans' estate argued that Gans' leukemia was caused by exposure to certain toxic chemicals (termed "BTEX") at the refinery. Because Gans worked at the refinery both before and after the Retroactive Date, the underlying suit created a risk for J.J. White of both covered and uncovered liability. If Gans' injuries were caused by pre-Retroactive-Date exposures, J.J. White's liability to Gans' estate would not be covered. On the other hand, if Gans' injuries were caused by exposures on or after the Retroactive Date, J.J. White's liability to Gans' estate would be covered.

There was also a third possibility: Gans' leukemia might not have been caused by BTEX exposures at all. Instead, it could have been caused by genetics or some other cause, such as Gans' years of cigarette smoking. In that case, J.J. White would have no liability to Gans' estate, covered or uncovered. But, as noted below, because the underlying case settled before trial, none of these three possibilities was proven. (Previous Op. 4-6.)

In the course of discovery in the underlying lawsuit, Catlin withdrew its defense of J.J. White, contending that there was no coverage. As I determined in my Previous Opinion, because J.J. White faced a risk of covered liability, Catlin was contractually obligated to defend the Gans suit, and breached that duty when it withdrew its defense.

After Catlin withdrew its defense, J.J. White made the decision to settle with Gans' estate, even though J.J. White denied all liability to Gans' estate in the underlying suit. J.J. White then sought indemnity from Catlin for the amount of the settlement. Catlin refused and initiated this declaratory judgment action to establish that the policy did not provide coverage. J.J. White countersued.

In the previously-decided cross-motions for summary judgment, Catlin sought a summary determination that it had neither a duty to defend nor indemnify J.J. White with respect to the Gans suit. J.J. White disagreed, urging that it had a right to a *587defense and indemnity of the settlement. Alternatively, J.J. White sought partial summary judgment that Catlin bore the burden of proving whether the Retroactive Date Provision would bar coverage. Because there was indisputably a possibility that Gans' suit could result in covered liability, I denied Catlin's motion in its entirety, and granted J.J. White's motion insofar as J.J. White argued that Catlin had breached its duty to defend.

I also decided that New York law applied, and that, under Servidone Construction Corp. v. Security Insurance Company, 64 N.Y.2d 419, 488 N.Y.S.2d 139, 477 N.E.2d 441 (1985), the mere fact that Catlin breached its duty to defend did not entitle J.J. White to indemnity for amounts paid in the settlement. I determined that there was a factual dispute as to when, if ever, exposure to BTEX caused Gans' leukemia-and therefore whether the loss was covered. Consequently, I denied both Catlin's and J.J. White's motions for summary judgment on the issue of indemnity, concluding that this issue would have to be resolved by a factfinder. Finally, I decided that J.J. White, as the insured, bore the burden of proving at trial that there was coverage and that the Retroactive Date Provision did not preclude such coverage.

In the Motion for Clarification or Reconsideration of the Previous Opinion currently before me, J.J. White does not challenge any of these conclusions. Rather, J.J. White challenges part of a single sentence in the Previous Opinion that it contends incorrectly sets out the specific facts that it will be required to prove in order to establish coverage. This sentence reads, in relevant part: "J.J. White...bear[s] the burden of proving that the Retroactive Date Provision is inapplicable-i.e. that Gans' leukemia was caused by exposures to BTEX occurring on or after the Retroactive Date [April 30, 2002.]" (Previous Op. 29 (emphasis added.))

J.J. White argues that this sentence imposes an improper burden on it, because it requires it to prove that exposure to BTEX was actually the cause of Gans' leukemia. J.J. White asserts that the previously-decided summary judgment motions did not address this issue and that this sentence is an incorrect expression of the law. J.J. White contends that, under applicable New York law, it is not required to prove actual liability to Gans' estate, but rather potential liability is enough. J.J. White accordingly requests that I either: (1) strike the above sentence, leaving the question as to what exactly it must prove to establish coverage open, or, alternatively, (2) rephrase the sentence to accord with its view of the law.

Catlin opposes either form of relief. In Catlin's view, the issue was presented in the previous summary judgment motions and the Previous Opinion conclusively resolved it in Catlin's favor, by noting that J.J. White cannot show a right to indemnity unless it can prove that it was actually liable to Gans' estate.

For the reasons set out below, I agree with J.J. White that it need not establish that it was, in fact, liable in the underlying suit. Accordingly, I will strike the sentence from my Previous Opinion suggesting otherwise. Going forward, and for the reasons set out below, in order to establish coverage, J.J. White will be required to prove the following: First, at the time it settled with the worker's estate, there was a possibility that the underlying trial verdict could have resulted in a covered loss. This is required because, if there was no possibility that the death was caused by exposure to chemicals during a covered time period, Catlin had no coverage obligations whatsoever. I note that, in my Previous Opinion, I found that there was evidence in *588the record of the underlying suit that fatal exposures could have occurred during a covered time period (i.e. , on or after the Retroactive Date), such that Catlin was required to defend. (See Previous Op. 23 (concluding that a "factfinder could ... have reasonably found [in the underlying suit] that [the worker] was exposed...in substantial amounts after the retroactive date, and that it was this later exposure that caused his [death]-in which case there would be coverage.")) Second, J.J. White must establish that, given the evidence available to it at the time of settlement, it was reasonable to enter into the settlement in the underlying case. And lastly, assuming that exposure to chemicals was the cause of the worker's death, J.J. White must prove that the fatal exposures occurred on or after the Retroactive Date.

II. LEGAL STANDARD

"The general purpose of a motion for clarification is to explain or clarify something ambiguous or vague, not to alter or amend." Resolution Trust Corp. v. KPMG Peat Marwick, No. 92-cv-1373, 1993 WL 211555, at *2 (E.D. Pa. June 8, 1993). By contrast, "[t]he purpose of a motion for reconsideration...is to correct manifest errors of law or fact or to present newly discovered evidence." Max's Seafood Cafe ex rel. Lou-Ann, Inc. v. Quinteros, 176 F.3d 669, 677 (3d Cir. 1999) (quoting Harsco Corp. v. Zlotnicki, 779 F.2d 906, 909 (3d Cir. 1985) ). A motion for reconsideration must rely upon one of the following grounds: "(1) an intervening change in controlling law; (2) the availability of new evidence that was not available when the court [rendered its decision]; or (3) the need to correct a clear error of law or fact or to prevent manifest injustice." Id. (citing N. River Ins. Co. v. CIGNA Reinsurance Co., 52 F.3d 1194, 1218 (3d Cir. 1995) ).

III. CLARIFICATION - WHAT J.J. WHITE MUST PROVE IN ORDER TO OBTAIN INDEMNIFICATION

J.J. White's clarification motion presents the following issue: Under New York law, when an insurer breaches its duty to defend its insured against a claim carrying a risk of both covered and uncovered liability, and the insured has settled with the claimant, what must the insured prove to establish its right to indemnity? This issue can be broken into two subsidiary questions. First, must the insured prove actual liability to the claimant? That is, must the insured prove that, had the underlying suit proceeded to trial, the insured would have been found liable to the claimant? And if the answer to that question is "no," the remaining question is how to determine coverage for a settled claim, where the insured contested liability in the underlying case, and that issue was not resolved by a factfinder.

As noted above, as to the first question, I conclude that J.J. White need not prove that it was actually liable in the underlying action. The answer to the second question is more nuanced and complicated. It requires the factfinder to assume that the deceased refinery worker's death was caused by exposure to harmful chemicals. Under this scenario, in which the factfinder must make that assumption, J.J. White must then prove that the fatal chemical exposures occurred on or after the Retroactive Date of April 30, 2002.

A. Question 1: Under New York Law, Must the Insured Prove Actual Liability?

Although no case of the New York Court of Appeals is directly on point, there is persuasive precedent suggesting that, should the question ever come before *589it,³ that court would conclude that an insured that enters into a reasonable settlement with a claimant for a covered claim is entitled to indemnity from the insurer even if it cannot prove actual liability to the claimant.

The New York case most often cited for this proposition is Luria Brothers & Company v. Alliance Assurance Company, 780 F.2d 1082 (2d Cir. 1986). In Luria, the insured settled a wrongful death suit regarding the sinking of its shipping vessel and sought indemnity of the settlement amount from its insurer. Id. at 1087. The insurer refused, arguing that it was only obligated to indemnify the insured's actual liability, not the mere potential liability that might motivate an insured to settle. Id. at 1091. The United States Court of Appeals for the Second Circuit disagreed, holding that, to obtain indemnity for the settlement from its insurer, the insured "need not establish actual liability to the party with whom it has settled 'so long as a potential liability on the facts known to the insured is shown to exist, culminating in a settlement in an amount reasonable in view of the size of possible recovery and degree of probability of claimant's success against the insured.' " Id. (alterations omitted) (quoting Damanti v. A/S Inger, 314 F.2d 395, 397 (2d Cir. 1963) ).

Although Luria was governed by New York law, I recognize that it relied on cases applying federal maritime precedent. However, four additional considerations convince me that the New York Court of Appeals would follow Luria: (1) an analogous New York Court of Appeals decision, Cardinal v. State, 304 N.Y. 400, 107 N.E.2d 569 (1952) ; (2) federal and lower state court decisions applying New York law that cite the Luria rule approvingly; (3) persuasive decisions from other jurisdictions applying the Luria rule; and (4) the policy rationale behind the Luria rule.

First , the New York Court of Appeals reached a similar result in Cardinal v. State, 304 N.Y. 400, 107 N.E.2d 569 (1952). In Cardinal, the United States brought a claim against the insured, Cardinal, for contribution on claims brought by Cardinal's injured workers. Id. at 570-71, 107 N.E.2d 569. Cardinal sought a defense from its insurer, which, it was later determined, wrongfully refused to provide a defense. Id. at 572-73, 107 N.E.2d 569. Cardinal then settled the contribution claim with the United States. Id. at 573, 107 N.E.2d 569. In a subsequent suit by Cardinal against its insurer, the insurer argued that later developments in maritime law-developments that occurred after Cardinal entered into a settlement with the United States-undermined the legal theory under which the United States had sought contribution. Id. at 577, 107 N.E.2d 569. Therefore, the insurer argued, the contribution claim was meritless, and thus not covered by the insurance policy. Id.

The New York Court of Appeals rejected the insurer's argument, holding that the viability of the contribution claim must be decided based on what was known to the insured at the time of the settlement-not by later changes in the law.

*590Cardinal, 107 N.E.2d at 578. The sole question, the court concluded, was whether the settlement was "reasonable[ ]." Id. Although Cardinal addressed a different issue than Luria ( Cardinal involved a change in the law, whereas Luria involved a dispute of fact), the case nevertheless instructs that an insured need not prove that it was actually liable in the underlying suit in order to obtain an indemnity for a settlement of that suit.

Second , regarding indemnification issues, state and federal courts applying New York law have consistently followed Luria and cited it approvingly. See J.P. Morgan Sec. Inc. v. Vigilant Ins. Co., 57 Misc.3d 171, 51 N.Y.S.3d 369, 380 (N.Y. Sup. Ct. 2017) ; Uniroyal, Inc. v. Home Ins. Co., 707 F. Supp. 1368, 1379 (E.D.N.Y. 1988) ; Am. States Ins. Co. v. Synod of the Russian Orthodox Church, 183 F. App'x 401, 403 (5th Cir. 2004) ; Md. Cas. Co. v. W.R. Grace & Co., No. 88-cv-2613, 1996 WL 109068, at *6-11 (S.D.N.Y. Mar. 12, 1996) ; Olin Corp. v. Ins. Co. of N. Am., No. 84 Civ. 1968, 2015 WL 1782194 at *4 n.1 (S.D.N.Y. Apr. 15, 2015), vacated in part on other grounds, 864 F.3d 130 (2d Cir. 2017). Other lower courts in New York, though not citing Luria, have found that an insurer who breaches its duty to defend is "conclusively bound by any reasonable good faith settlement the indemnitee may make." Feuer v. Menkes Feuer, Inc., 8 A.D.2d 294, 187 N.Y.S.2d 116, 121 (1959) ; ELRAC, Inc. v. Cruz, 182 Misc.2d 523, 699 N.Y.S.2d 647, 649 (Civ. Ct. 1999). Catlin has not cited, and I have not found, any cases rejecting the Luria rule.

Third , several other jurisdictions follow a related rule announced by the New York Court of Appeals in Servidone Construction Corp. v. Security Insurance Company, 64 N.Y.2d 419, 488 N.Y.S.2d 139, 477 N.E.2d 441 (1985), that an insurer breaching the duty to defend does not waive its coverage defenses. Those courts following Servidone also apply the Luria rule that the insured is not required to prove actual liability to the claimant. See Associated Aviation Underwriters v. Wood, 209 Ariz. 137, 98 P.3d 572, 584 (Ariz. Ct. App. 2004) ; Emp'rs Cas. Co. v. Block, 744 S.W.2d 940, 942-43 (Tex. 1988) ; Elliott v. Hanover Ins. Co., 711 A.2d 1310 (Me. 1998) ; Cambridge Mut. Fire Ins. Co. v. Perry, 197 Me. 94, 692 A.2d 1388 (1997) ; Isaacson v. Cal. Ins. Guarantee Ass'n, 44 Cal.3d 775, 244 Cal.Rptr. 655, 750 P.2d 297, 309 (1988). And Catlin has not identified-and I have not found-any jurisdiction that follows Servidone and yet rejects Luria. This belies Catlin's argument that a rule that does not require an insured to prove actual liability in the underlying lawsuit is in conflict with Servidone.

Fourth , the policy rationale for the rule in Luria comports with common sense. As one judge of the United States District Court for the Eastern District of New York aptly observed, requiring the insured to prove actual liability in the underlying suit would "markedly reduce the advantages to the insured of settling" because, "faced with the choice of defending the tort action vigorously or settling it without hope of insurance reimbursement, insureds would tend to choose the former." Uniroyal, 707 F. Supp. at 1378. Indeed, it appears that every state in the country permits an insured, faced with a suit against it and an insurer that has declined to defend, to make a reasonable settlement with the claimant. See Kenneth S. Abraham, The Liability Insurer's Duty to Settle Uncertain and Mixed Claims, 68 Rutgers U.L. Rev. 337, 358 (2015). Reasonable settlements would make little sense if, in the subsequent coverage lawsuit, the insured would be required to prove it was actually liable in order to obtain indemnification. Uniroyal, 707 F. Supp. at 1378. Because *591public policy favors settlement, a rule that would discourage settlement should be avoided. Id.

Catlin has not raised any persuasive arguments that undermine any of these considerations. Instead, Catlin presses the point that, because an insured must prove coverage, and because no coverage can arise absent liability, an insured must prove actual liability. Catlin argues that this result is supported by Servidone, because that case held that an insured must prove the actual existence of the facts on which coverage turns, and not merely that those facts were possibile at the time the insurer withdrew its defense. 488 N.Y.S.2d 139, 477 N.E.2d at 445. Catlin observes that, while Luria is one of a handful of nonbinding cases, Servidone is binding, and the New York Court of Appeals reaffirmed Servidone as recently as a few years ago. K2 Inv. Grp., LLC v. Am. Guarantee & Liab. Ins. Co., 22 N.Y.3d 578, 983 N.Y.S.2d 761, 6 N.E.3d 1117, 1118 (N.Y. 2014). Therefore, Catlin contends, any conflict between Luria and Servidone must be resolved in favor of Servidone, regardless of the policy implications, or the decisions of federal courts or state courts in other jurisdictions.

Catlin is correct that Servidone is binding as far as it applies to a particular case, and that J.J. White is required to prove that the loss was covered. Catlin is also correct that here, unlike in Luria, coverage for the underlying claim is disputed, and that this may place an additional burden on J.J. White. However, on the narrow question of whether the burden to prove coverage includes a burden to prove actual liability in the underlying action, Catlin is incorrect that Servidone dictates the answer in its favor.

Servidone did not address the issue of whether an insured must prove actual liability in the underlying suit. Indeed, the coverage dispute in Servidone did not turn on the merits of the underlying suit. Rather, in Servidone, the claimant, the United States, brought claims for common law and contractual indemnity, only the first of which would be covered under the policy. 488 N.Y.S.2d 139, 477 N.E.2d at 442. The post-settlement coverage dispute turned on which of those claims-common law or contractual indemnity-was the "actual basis" for the settlement. Id. at 424, 488 N.Y.S.2d 139, 477 N.E.2d 441. Liability was therefore not in dispute-only the basis for liability. Similarly, in K2, the dispute as to coverage turned on whether the insured was acting for a "business enterprise" when he committed the alleged wrong, a fact that could be shown independently of the insured's liability on the underlying claim. 22 N.Y.3d 578, 983 N.Y.S.2d 761, 6 N.E.3d 1117. In short, although Servidone requires the insured to show that the "basis" for its liability would be covered, it did not require the insured to prove that that it was, in fact, liable in the underlying suit.

Moreover, Servidone reaffirmed that "an insurer cannot by virtue of its own breach [of the duty to defend] escape liability for the reasonable settlement of a covered risk." 488 N.Y.S.2d 139, 477 N.E.2d at 444. By requiring only that the settlement be "reasonable," Servidone suggests that actual liability is not required, only a sufficient likelihood of liability to justify the settlement.

And finally, every court to have considered Luria and Servidone together has concluded that Luria "is not inconsistent with Servidone." Md. Cas. Co., 1996 WL 109068, at *7 ; Uniroyal, 707 F. Supp. at 1379 (" Servidone...required only that the claim settled by the insured be a 'covered loss' under the policy, and not that the insured independently prove the truth of the underlying claim.");

*592Morgan Stanley Grp., Inc. v. New England Ins. Co., 36 F. Supp. 2d 605, 608 (S.D.N.Y. 1999) (" Servidone and Luria can, and should, be read together to establish the standard of proof for a settling insured."), vacated in part on other grounds, 225 F.3d 270 (2d Cir. 2000).

Accordingly, in order to obtain indemnification, I conclude that J.J. White is not required to prove that it was actually liable to Gans' estate in the underlying litigation. That is, J.J. White need not prove that exposure to BTEX actually caused Gans' leukemia. Because the portion of one sentence of my Previous Opinion cited by J.J. White suggested otherwise, I will grant J.J. White's Motion for Clarification or Reconsideration, and strike that portion of that sentence.

B. Question 2: Because J.J. White Need Not Prove Actual Liability in the Underlying Lawsuit, What Must It Prove to Establish Policy Coverage?

As explained above, J.J. White is not required to prove that chemical exposures caused Gans' death, or actual liability to Gans' estate in the underlying suit. But that still leaves unanswered the question of what J.J. White does have to prove to establish coverage for the settlement with Gans' estate.

Luria, and the cases discussed above that follow it, require only that the settlement be "reasonable in view of the size of possible recovery and degree of probability of claimant's success against the insured." 780 F.2d at 1091. But as J.J. White concedes, that cannot be the only fact that it must prove here, because this case differs from Luria in one important way: In Luria, any finding of liability in the underlying suit would have been covered. See, e.g., id. at 1090-91. The only ground on which the insurer in Luria challenged coverage was that the insured was not actually liable in the underlying case.

Here, by contrast, the underlying suit carried a risk of both covered and uncovered liability. As I explained in my Previous Opinion:

[H]ad the Gans Suit proceeded to trial, the finder of fact may well have found that Gans was exposed to BTEX in substantial amounts before the retroactive date, and that this exposure caused Gans' leukemia, precluding coverage under the...Policy. Conversely, a factfinder could also have reasonably found that Gans was exposed to BTEX in substantial amounts after the retroactive date, and that it was this later exposure that caused his leukemia-in which case there would be coverage under the...Policy.

(Previous Op. 23.)

Because a finding that J.J. White was liable in the underlying suit would not, by itself, resolve the coverage question-i.e. , whether the fatal BTEX exposures occurred before or after the Retroactive Date-it is clear that J.J. White must prove something more than that the settlement was reasonable in order to obtain indemnification.

But what that something is presents a difficult and unsettled question, at least under New York law. The question is difficult because of the tension between two principles of New York law. On the one hand, under Luria, the insured is not required to prove that it was actually liable in the underlying suit, only that there was a possibility of liability sufficient to justify the settlement. Thus, under Luria, J.J. White is not required to prove the opposite of what it argued in the underlying case: that Gans, in fact, died from chemical exposures.

*593On the other hand, however, under Servidone, an insured is required to prove coverage on the basis of the actual facts-and cannot obtain coverage merely by showing a possibility that the factfinder in the underlying case would have found the facts on which coverage turns. Thus, J.J. White faces the dilemma of proving the actual facts on which coverage turns-that is, the timing of the fatal chemical exposures-even though: (1) it has always argued that these exposures were not actually fatal; and (2) it need not prove that the exposures were actually fatal to obtain coverage.

Only J.J. White's briefing directly addresses this dilemma. J.J. White maintains that it can establish an entitlement to indemnity if-in addition to proving that the settlement was reasonable-it proves two things: (1) based on the facts known to it at the time of the settlement, there was a potential covered liability-i.e. , a potential liability for exposures occurring on or after the Retroactive Date; and (2) Gans' leukemia was not caused by chemical exposures occurring before the Retroactive Date. J.J. White explains that this second fact can be shown by demonstrating that Gans' leukemia was not caused by chemical exposure at all, but, rather, by something else-such as cigarette smoking.

J.J. White's position is somewhat counterintuitive, but, as I understand it, J.J. White reasons as follows. J.J. White need not prove that it was actually liable in the underlying case. Rather, for purposes of the coverage question, J.J. White's liability must be assumed. And because liability is assumed, "disproving" the uncovered liability "necessarily proves" the covered liability. (See J.J. White 11/12/2018 Ltr. Br. 2 n.4, 5.) In other words, proving that Gans' death was not caused by pre-Retroactive Date exposures is-given the assumption of liability that the factfinder must make-sufficient to prove that Gans' death was caused by exposures occurring on or after the Retroactive Date. Thus, paradoxically, by proving that Gans' death was caused by something other than chemical exposures, such as cigarette smoking, J.J. White posits that it has proven-in light of the assumption of liability-that the death was caused by post-Retroactive Date exposures. But, as explained in greater detail below, J.J. White cannot both benefit from the assumption of liability, and then prove coverage by denying that very same assumption.

Accordingly, I will decline to follow J.J. White's approach. Instead, I will follow the approach taken by the Minnesota Supreme Court in an analogous case, which, I conclude, the New York Court of Appeals would follow were it faced with this question. See In re Silicone Implant Insurance Coverage Litigation ("Silicone Implant"), 667 N.W.2d 405 (Minn. 2003). Under that approach, the factfinder must assume that the insured was actually liable in the underlying case, and then determine-based on that assumption-the remaining factual questions necessary to establish coverage. Thus, here, J.J. White will not be permitted to prove coverage by proving that Gans' death was caused by something other than chemical exposures. Rather, it will be required to prove that, assuming Gans' death was caused by chemical exposures, the fatal exposures occurred on or after the Retroactive Date of April 30, 2002. My reasoning for adopting this approach follows.

1. The Dilemma: How to Prove the Timing of Something (the Alleged Chemical Exposure) Without Proving That It Happened?

As noted above, what an insured must prove to establish coverage when it has settled an underlying case *594carrying a risk of both covered and uncovered liability is complicated by two principles of New York insurance law that are somewhat in tension. An insured need not prove actual liability to establish coverage, only a possibility of liability sufficient to render the settlement reasonable. But the insured must still prove the actual facts necessary, such as the timing of the injury, to establish coverage. 488 N.Y.S.2d 139, 477 N.E.2d at 444. These two principles create a dilemma for an insured like J.J. White, which contested liability in the underlying suit: J.J. White must prove the timing of fatal chemical exposures that it has always maintained did not occur, even though J.J. White need not prove that the exposures actually occurred.

An insured was confronted with a similar dilemma in Maryland Casualty Co. v. W.R. Grace & Co., No. 88-cv-2613, 1996 WL 109068, at *6-11 (S.D.N.Y. Mar. 12, 1996), which, while not binding here, applied New York law. There, the insured, Grace, sought coverage for claims that its asbestos-containing products had caused "property damage" by creating asbestos hazards on its customers' properties. Id. Grace had multiple property damage insurance policies covering different time periods, and coverage under these policies turned on the date of installation of the asbestos-containing products. Id. at *6. In the underlying suit, Grace denied that it had installed the asbestos-containing products at issue. But it nevertheless settled with some of the claimants and then sought indemnity for the settlement from its insurers. Id.

In determining how Grace could prove coverage for the settled claims, the court recognized that New York law placed Grace in a difficult position: Although Grace was not required, under Luria, to prove actual liability in the underlying suit-by proving installations that it had denied occurred in the underlying suit-it was required, under Servidone, to prove coverage by proving the dates on which the asbestos-containing products were installed. Md. Cas. Co., 1996 WL 109068, at *7. Grace was therefore seemingly required to prove "the date of non-installation." Id. Recognizing this dilemma, the court confronted the question of how Grace could establish the timing of the installation of asbestos-containing products that it had always maintained had not actually been installed.

In resolving this dilemma, the Maryland Casualty court held that Grace could obtain coverage if it could show a "reasonable basis" to conclude that the liability from the underlying settlement arose within a given time period. Md. Cas. Co., 1996 WL 109068, at *9. The court explained that Grace could do this by offering statistical evidence of known dates on which it had, in fact, installed other asbestos-containing products that were not at issue in the underlying claims, and by assuming that the dates of the settled claims followed the same distribution of dates as the other claims. Id. The court noted that this method would be efficient because it would not require a difficult factfinding as to when the products at issue had been installed when it had always maintained that those products had not been installed. Id. at *10.

Maryland Casualty illustrates the dilemma J.J. White faces, but, unfortunately, its use of statistical evidence from similar claims does not provide a workable solution for the case before me. Here, there is only one underlying claim-a single death from leukemia-and there are no similar injury dates outside of this case to use as a statistical model to prove timing. Moreover, Maryland Casualty's way of generally framing its rule-that an insured can prove coverage by demonstrating a "reasonable basis" to attribute the settled *595claim to a given time period-does not provide a concrete test that can be applied to this case. Thus, while Maryland Casualty is useful in illustrating the dilemma that J.J. White now faces-proving the date of fatal chemical exposures that it has always maintained did not occur- Maryland Casualty is not instructive as to what, exactly, J.J. White must prove in this case to provide coverage.

2. The Silicone Implant Approach: Assume Liability and Decide Other Facts Necessary to Establish Coverage

Fortunately, another case presenting the same issues-but decided under Minnesota law-does present a workable approach. See In re Silicone Implant Insurance Coverage Litigation ("Silicone Implant"), 667 N.W.2d 405 (Minn. 2003). In Silicone Implant, the insured-a manufacturer of silicone breast implants-sought coverage from its insurer after it settled several products liability suits over the alleged harmful effects of its implants. Id. at 408-409. The insured manufacturer had defended the underlying suits on the theory that its implants did not cause injury. Id. at 410. Under the pertinent policy, coverage turned on the date of "actual injury" caused by the implants. Id. at 415.

As in Maryland Casualty, the insured was placed in the difficult position of proving the date of an injury that it had always denied occurred. Indeed, the situation was more difficult for the insured in Silicone Implant, because "solid conventional science establishe[d] no causal connection whatsoever between silicone gel breast implants and systemic disease." Silicone Implant, 667 N.W.2d at 413 (emphasis added). Thus, unlike in Maryland Casualty, the evidence in Silicone Implant showed that the underlying claims were actually meritless, leaving the insured to prove-in order to obtain coverage-the date of injuries that the actual facts showed did not occur.

The court in Silicone Implant resolved this dilemma by requiring the factfinder to assume, for the limited purpose of determining coverage, that the science was the opposite of what it actually was-that is, to assume that the silicone implants did cause injury. 667 N.W.2d at 413. The court then required the factfinder to determine the timing of the injury on the basis of that assumption, instructing the expert witnesses to take that assumption into account in their opinions as to timing. E.g., id. at 413-14 (quoting one expert's testimony acknowledging, " 'if I have to work on the assumption that silicone is a causal factor, then actual injury likely precedes first manifestations of disease symptoms by a short period of time.' " (alterations omitted)). Based on this assumption, that the silicone implants caused injury, the factfinder determined that, for purposes of coverage, the injury occurred shortly after implantation. Id. at 414.

The insurers had objected to this approach on the ground that a "fictional" injury could not be covered. Silicone Implant, 667 N.W.2d at 415-16. In their view, the plain language of the policy covered only "bodily injury or property damage," and this did not include "nonexistent" injuries of the type that the court required the factfinder to assume had occurred. Id. The Minnesota Supreme Court rejected this argument, holding that, when deciding coverage for a settlement "premised on the notion that silicone gel breast implants caused plaintiffs' injuries," the factfinder was obligated to assume that the injuries were real for the purpose of determining their timing. Id. at 416.

The Silicone Implant approach has advantages and disadvantages. Its primary advantage is that it allows the parties to *596offer evidence as to the timing of the injury at issue, which evidence understandably might not have been developed in the underlying suit. Here, for example, the underlying Gans suit did not focus on the timing of Gans' exposures to BTEX, but rather focused on whether those exposures actually caused Gans' death. Given that the record in the underlying suit will often be underdeveloped on the critical coverage question of timing, the Silicone Implant approach allows for factfinding on that critical question, without requiring the insured to prove the opposite of what it had argued in the underlying case-that it was actually liable.

Perhaps more importantly, this approach also permits an insurer to avoid providing coverage for injuries that it never contracted to cover. As I explain below in rejecting J.J. White's proposed approach, where the underlying suit carried a possibility both of covered and uncovered liability, a rule entitling the insured to full indemnity of a settlement, merely by proving non-liability, would be an unjust windfall to the insured. Put in real terms, J.J. White's approach may lead to coverage if the factfinder determines that Gans died from cigarette smoking. This, of course, would be an absurd outcome. The Silicone Implant approach is more sensible, and requires the insured to offer evidence showing that the settlement corresponds to a loss covered by the policy, without requiring it to prove its actual liability in the underlying case, which liability it has always contested.

Finally, the Silicone Implant approach most closely reconciles Servidone's requirement that the insured establish the "actual facts" necessary to demonstrate coverage, with Luria's rule that an insured need not prove its own liability in the underlying case. As the court in Silicone Implant observed, perfect adherence to the "actual facts" requirement is impossible where the injury at issue did not in fact occur. 667 N.W.2d at 413. Under Luria an insured cannot be required to prove that it was actually liable in the underlying case. The Silicone Implant approach accommodates the Luria rule, while still requiring an insured to prove the other facts necessary to the coverage determination-here, the timing of the injury.

The primary disadvantage of the Silicone Implant approach is that it requires what can be a counterintuitive factfinding. It asks the factfinder to assume the factual basis for liability-even if those facts could actually be false-while determining other facts that may be intertwined with the assumed (but potentially false) facts. See, e.g., Silicone Implant, 667 N.W. 2d at 413-14 (factfinder was required to determine the timing of injuries on the assumption that the injuries were caused by silicone implants, even though, in fact, the injuries were not caused by silicone implants ). Similarly here, the factfinder will be required to assume that Gans' death was caused by BTEX exposures, and then determine the timing of the fatal exposures, even though the exposures may not, in fact, have been fatal.

This disadvantage, however, is not as significant as it may seem. The paucity of cases presenting the dilemma discussed above suggests that disputes like those at issue in Maryland Casualty and Silicone Implant are relatively rare. If the underlying claims are entirely groundless, the insured will be hard pressed to demonstrate that it was reasonable to pay to settle them. Most disputes will involve claims falling somewhere between the clearly meritorious and clearly meritless. In this case, for example, the evidence does not suggest that Gans' claim was completely meritless, as the estate did offer expert testimony as to the propensity of BTEX to *597cause leukemia. Because there was competing expert testimony as to whether BTEX actually caused Gans' leukemia, it will not be particularly difficult for a factfinder to assume that BTEX exposures were the cause of the death-and then determine when the fatal exposures occurred.

Because Silicone Implant presents a workable rule, I conclude that the New York Court of Appeals would adopt it, and I will accordingly apply it to this case.

3. J.J. White's Suggested Approach: The Insured Need Only Prove Potential Covered Liability and an Absence of Uncovered Liability

As noted above, J.J. White urges a different rule: that an insured need only prove that: (1) at the time of settlement, a covered liability was a possibility; and (2) the events that would give rise to an uncovered liability did not, in fact, occur. Thus, to put it in terms of the facts of this case, J.J. White suggests it could establish coverage by demonstrating: (1) at the time of settlement, it was possible that the factfinder in the underlying case could have found that Gans' leukemia was caused by BTEX exposures on or after the Retroactive Date; and (2) Gans' leukemia was not, in fact, caused by pre-Retroactive Date exposures to BTEX, but rather by something else, e.g. smoking. Thus, in effect, J.J. White contends that it can establish coverage by proving that it was not liable in the underlying case. For the reasons stated previously and set out below, I will reject this approach.

First, I note that this standard substantially relaxes the insured's burden. It allows the insured to obtain complete indemnity for a settled claim by proving that there was some chance, however slight, that the claim might be covered. Neither Maryland Casualty nor Silicone Implant considered relaxing the insured's burden in this way. And it makes sense that this approach was not considered, because both cases started by assuming-consistent with Luria-that the insurer was barred from disputing the insured's actual liability. It cannot be the case that an insurer is barred from disputing coverage on the basis that its insured was not actually liable, but that the insured can obtain coverage by proving the same thing. To allow the insured to obtain coverage merely by proving that it was not liable in the underlying case, and that there was a possibility of coverage, would neglect Servidone's instruction that the insured prove the "actual facts" establishing coverage.

An example helps illustrate the significant windfall to an insured that J.J. White's proposed rule could create. Assume that, at the time that J.J. White settled with Gans' estate, there was a moderate chance (say a 40% chance) that it would have been found liable-that is, a 40% chance that a factfinder would have found that Gans' leukemia was caused by exposure to BTEX. However, assume that there was only a 1% chance that Gans' leukemia was caused by exposures that occurred on or after the Retroactive Date. Under J.J. White's proposed rule, it could obtain complete indemnification from Catlin merely by showing that it was not in fact liable in the Gans suit, thereby recouping the entirety of what it paid out to settle a risk for which there was only a 1% chance of coverage. This would "enlarge[ ] the bargained-for coverage" something Servidone instructs courts not to do. 488 N.Y.S.2d 139, 477 N.E.2d at 444. For this reason, I am convinced that the New York Court of Appeals would reject J.J. White's proposed rule.

IV. CONCLUSION

It is clear that under New York law, an insured need not prove that it was actually *598liable in the underlying litigation in order to obtain indemnification of a reasonable settlement of that litigation. However, what an insured does have to prove is less clear. For the reasons set out above, I conclude that the Silicone Implant approach-which requires the factfinder to assume liability in the underlying case and determine the remaining facts necessary to evaluate coverage on the basis of that assumption-adheres closest to New York law as set out in Luria and Servidone. The rule from that case will therefore be applied in this matter.

Accordingly, J.J. White's Motion for Clarification or Reconsideration will be granted, and my Previous Opinion will be amended to reflect the conclusions set out above.

An appropriate Order follows.

1 The facts of this case are set out in greater detail in my Previous Opinion on the parties' cross-motions for summary judgment. (See Doc. No. 114.)

2 Counterplaintiffs Sunoco, Inc. and Sunoco, Inc. (R & M) also seek coverage as insureds under the Catlin policy, and join in the instant motion. For simplicity, I will refer to these entities and J.J. White collectively as "J.J. White."

3 A "federal court sitting in diversity...[is] required to apply the substantive law of the state whose law governs the action." Spence v. ESAB Group, Inc., 623 F.3d 212, 216 (3d Cir. 2010). In so doing, absent a controlling decision of the state's highest court, a federal court must "predict how [the state's highest] court would rule if faced with the issue," looking to "decisions of state intermediate appellate courts, of federal courts interpreting that state's law, and of other state supreme courts that have addressed the issue, as well as to analogous decisions, considered dicta, scholarly works, and any other reliable data tending convincingly to show how the highest court in the state would decide the issue at hand." Id. at 216-17 (internal quotation marks omitted).