This is the third and final article about the decision of a New York appellate court in National Union Fire Ins. Co. of Pittsburgh v. TransCanada Energy USA, Inc., 119 A.D.3d 492 (1st Dep’t 2014).  The first article discussed the legal background against which the case was decided.  The second article discussed the various opinions that have been issued in the case.  This article will discuss the implications of those opinions for insurers in New York.

The Work Product Doctrine

One troubling aspect of the TransCanada case is an apparent disconnect between the facts and the law.  As discussed in the second part of this series, the policies at issue insured an enormous power plant, and the first evidence that “Big Allis” had a damaged rotor surfaced just three days after the policies had incepted.  TransCanada provided notice of the claim about two weeks later.  As soon as it did so, its insurers had reason to believe both that the claim under the policies would be enormous, and that the insurers had a possible defense to the claim, depending on when the crack in the rotor had first appeared.  In June 2009, after TransCanada valued its claim at over $70 million, the insurers articulated their possible defense, and TransCanada expressly disputed it.  The market insurers did not deny the claim until almost a year later, after a lengthy technical examination had been completed, and after a second expert had reviewed its results.

In other contexts, New York law provides that “[w]hether a particular document is … protected by the work product doctrine … is necessarily a fact-specific determination.”  Charter One Bank, F.S.B. v. Midtown Rochester, L.L.C.,738 N.Y.S.2d 179 (N.Y. Sup. 2002), quoting Spectrum Systems Intl. Corp. v. Chemical Bank, 575 N.Y.S.2d 809 (N.Y. 1991).  For TransCanada, the facts simply do not support the conclusion that none of the communications the insurers conducted with their attorneys after June 2009 was “prepared or obtained because of the prospect of litigation.”  See Royal Indemn. Co. v. Salomon Smith Barney, Inc., 4 Misc.3d 10006(A) (N.Y. Sup. Ct. 2004).

Yet both the trial court and the Appellate Division followed the rule—established in such cases as Bertalo’s Rest. Inc.v. Exchange Ins. Co., 658 N.Y.S.2d 656 (2d Dep’t 1997), and Bonbard v. Amica Mut. Ins. Co., 783 N.Y.S.2d 85 (2d Dep’t 2004)—that pre-denial communications, as such, are never made in “anticipation of litigation” for purposes of the work product doctrine.  Justice Jaffe suggested that the market insurers might have helped their case by “submit[ing] evidence stating that a firm decision to deny coverage was made earlier.”  But she did not entertain even the possibility that an insurer could prepare for litigation, based on a decision that was only contingent.

A second consideration that makes this mechanistic approach to rejecting work product protection especially problematic is the broader context in which it was confirmed.  As discussed in the first part of this series, large and complex claims in New York are now likely to generate not only disputes over coverage, but also suits seeking extracontractual damages for alleged breach of the insurer’s covenant of good faith and fair dealing.  See Kings Infiniti Inc. v. Zurich American Ins. Co., 43 Misc.3d 1207(A) (N.Y. Sup. Ct. 2014) (“although a ‘plaintiff’s cause of action alleging bad faith conduct on the part of the insurer cannot stand as a distinct tort cause of action, … its allegations may be employed to interpose a claim for consequential damages beyond the limits of the policy for the claimed [breach of covenant]“) (citation omitted); Bi-Economy Market, Inc. v. Harleysville Ins. Co. of N.Y., 10 N.Y.3d 187 (2008).

Claims of that type, in which the insurer’s exposure is potentially unlimited, can turn on conclusions about the insurer’s state of mind.  In TransCanada, it would have been a reasonable precaution for the insurers’ attorneys to begin marshalling arguments and materials they would need to litigate the claim while they were still awaiting the results of the Siemens analysis—a year or more before the claim was actually denied.  In future cases, policyholders are likely to argue that the products of such activity are evidence that the insurers pre-judged their claims.  And under the analysis used in TransCanada, the fact that the insurers had sound reasons to prepare for litigation would beirrelevant to deciding whether the results of that preparation qualify as work product.

The Common Interest Privilege

TransCanada’s discussion of the common interest privilege represents a significant problem for insurers.  InTransCanada, there was good reason for the market insurers to share a single coverage counsel, because they were defending claims under identical policies that were all based on the same occurrence.  Groups of insurers aren’t always in that position; the situation is most likely to arise in connection with very large exposures.  It is just such exposures, however, that often generate difficult and complex claims.  Insurers have a lot to gain from an open exchange of information and opinions about their legal obligations with respect to such claims; they also have a lot to lose, if elements of that exchange can be used as evidence of bad faith in a claim for extracontractual damages.

Extending a litigation requirement that only one New York appellate court had previously recognized, the trial court ruled (and the Appellate Division originally agreed) that insurers in this position may never exchange privileged communications before the claim has been denied without waiving the privilege in its entirety—and potentially waiving it as to other communications, as well.  National Union Fire Ins. Co. of Pittsburgh v. TransCanada Energy USA, Inc., 2013 WL 4446917, at *4-*5 (N.Y. Sup. Ct. Aug. 15, 2013).  See, e.g.AMBAC Indemn. Corp. v. Bankers Trust Co., 573 N.Y.S.2d 204, 208 (N.Y. Sup. 1991) (“voluntary disclosure of the content of a privileged communication constitutes a waiver of the privilege as to all other communications on the same subject”).  But see Charter One Bank, F.S.B. v. Midtown Rochester, L.L.C., 738 N.Y.S.2d 179, 187-88 (N.Y. Sup. 2002) (declining to find waiver as to undisclosed documents).

Yet there does not appear to be any strong public policy of New York that would be advanced by preventing insurers from sharing privileged communications about their legal rights and obligations with respect to complex claims, and neither court tried to articulate one.  In fact, such sharing would almost certainly be permitted in other jurisdictions. See, e.g., Bank of America, N.A. v. Terra Nova Ins. Co. Ltd., 211 F.Supp.2d 493, 498 (S.D.N.Y. 2002); O’Boyle v. Borough of Longport, 94 A.3d 299, 310 (N.J. 2014).

Furthermore, as discussed in the second part of this series, the courts need never have addressed the common interest doctrine at all: the doctrine commonly applies to communications among parties that are separately represented.  E.g.Bank of America, N.A. v. Terra Nova Ins. Co. Ltd., 211 F.Supp.2d 493, 496 (S.D.N.Y. 2002).  Where, as in TransCanada, there is a joint representation, communications involving the jointly-represented clients usually remain privileged as against third parties.  E.g.Arkin Kaplan Rice LLP v. Kaplan, 988 N.Y.S.2d 22 (1st Dep’t 2014).

One reason the TransCanada court did not consider this possibility might be that the underlying facts were not clearly established: the trial court noted that the market insurers had “produced no joint defense agreement, or similar evidence, that explains the terms and conditions of this joint representation.”  This language suggests that joint representation might still be a viable option in future cases—provided that the terms of the representation are made clear in a written retainer agreement.  But insurers who try that approach from now on can expect policyholders to argue that TransCanada stands for the proposition that the joint representation constitutes a waiver. Consequently, even jointly-represented insurers should avoid sharing privileged communications; that is, counsel should communicate with each client separately.

On the other hand, sharing privileged communications among separately-represented insurers at any time before the claim is denied is probably even riskier.  If insurers are not jointly represented, they should share information only through their attorneys, and they should not share privileged communications at all.  A written agreement governing the sharing of information would still be advisable, but it must be drafted in a way that does not suggest the insurers have reached any conclusion about the merits of the underlying claim.

Attorney-Client Privilege

The good news from TransCanada relates to the attorney-client privilege.  The fact that Judge Jaffe upheld the privilege for some of Factory Mutual’s documents, even though the court found that the lawyers were involved in investigating the claim and deciding whether to pay, shows that privilege determinations must depend on the nature and contents of communications, rather than the time at which they occurred.  Although the First Department did not expand on this point, it did clearly acknowledge that the trial court had protected some of the documents at issue. TransCanada, therefore, represents an improvement over the sweeping language that appears in some Second Department decisions.  See, e.g.Melworm v. Encompass Indemn. Co., 977 N.Y.S.2d 321 (2d Dep’t 2013) (“Reports prepared by … attorneys before the decision is made to pay or reject a claim are … not privileged and are discoverable”).


Evidence that can be used to suggest bad faith in the handling of a New York claim creates a greater risk than ever of extracontractual liability.  The opinions in TransCanada illustrate some of the ways in which pre-denial communications with attorneys might end up being applied to that purpose.  Here are some ways to reduce that risk:

  • For large, complex claims, it is best to establish separate investigative and legal teams.  This practice did not end up preventing the disclosure of many documents in TransCanada, but it is still advisable.Pre-denial communications with attorneys about the progress of the investigation should be carefully managed in a way that documents their “legal character.”  TransCanada preserves the privilege for some pre-denial communications, but only to the extent the insurer can persuade the court that they are not part of its regular business.  Communications should clearly reflect that they are being made for the purpose of advising the insurer about its legal rights and obligations under the policy—for example, about how courts have applied or interpreted a particular exclusion.
  • In TransCanada, the insurers had reason to anticipate litigation before they had a conclusive basis for denying the claim, because they knew Siemens might conclude that Big Allis had already been damaged  before the policies incepted.  That situation is likely to come up again—whether because a technical investigation is likely to establish a defense to coverage, or because an insured is asserting an extreme or intransigent position. Under these circumstances, counsel cannot always afford to wait before marshalling the evidence, information and arguments it will need in the anticipated coverage suit.  Counsel must be aware, however, of the likelihood that written materials prepared in this situation will not be protected as work product.
  • One problem this situation creates has to do with legal strategy.  It will be best to document clearly the nature of the litigation risk these materials are designed to address—for example, by stating (in the situation presented by TransCanada) that the parties have already taken conflicting positions about the significance of pending test results; that litigation is expected to ensue if the results cause the insurer to deny the claim; and that the attorneys’ work is being undertaken solely to prepare for that prospect.  With materials that establish these facts, it might be possible to persuade a court that the current, mechanistic approach to disclosing work product in coverage disputes should be modified.
  • More fundamentally, this situation creates an evidentiary problem.  It is a problem where an insurer is actually guilty of bad faith, of course, but it is also, and more broadly, a problem where the record of a good faith investigation has been created carelessly. This leads to the next bullet point.
  • All pre-denial materials should be prepared under the assumption that they will be reviewed by a court (if only in the course of an in camera inspection to resolve a discovery dispute) and, probably, by a jury, as well. That is, they must clearly reflect the insurer’s intention to comply fully with its obligations, and to explore potential coverage defenses only where there is a valid basis to believe they might apply. Failure to document the reasons that a lawyer is investigating of a particular issue can turn out to be worse than a failure to investigate the issue at all, because it can suggest that the insurer is looking for a pretext to deny the claim.
  • Where, as in TransCanada, several insurers have an interest in a single large and complex claim, they must enter into a written agreement setting forth the terms of their co-operation.  They should communicate only through counsel, and they should not share privileged communications at all.  Even if they are jointly represented, counsel should communicate with each client separately.

New Yorker’s have always been tough; now New York insurers need to do something about it.