STRYKER CORPORATION; HOWMEDICA OSTEONICS CORP., Plaintiffs-Appellees/Cross-Appellants, v. NATIONAL UNION FIRE INSURANCE COMPANY OF PITTSBURGH, PA., Defendant, TIG INSURANCE COMPANY, Defendant-Appellant/Cross-Appellee.
Stryker Corporation has been engaged in a longstanding row with XL Insurance America, Inc. (its commercial umbrella insurer) and TIG Insurance Company (its excess liability insurer). Fifteen years by our count. See Stryker Corp. v. XL Ins. Am., 576 F. App'x 496 (6th Cir. 2014); Stryker Corp. v. XL Ins. Am., 735 F.3d 349 (6th Cir. 2012); Stryker Corp. v. Nat'l Union Fire Ins. Co., 681 F.3d 819 (6th Cir. 2012). That insurance-coverage dispute, in its current incarnation, requires us to interpret the “consent-to-settle” provision of an excess-liability policy. The district court thought that the insurance contract contained a latent ambiguity, construed the policy against TIG, and entered summary judgment for Stryker. But the contract is not ambiguous, in any sense of the word, so we reverse.
Stryker is a medical technologies firm. In the late 1990s, it purchased a subsidiary of Pfizer, Inc. that made and sold orthopedic products. One of those products, an artificial knee joint called the Duracon Unicompartmental Knee (or “Uni-Knee” for short), turned out to be defective. These medical devices were sterilized using gamma rays, which caused ultra-high-molecular-weight polyethylene in the artificial knees to degrade and, if implanted past their five-year shelf-life, potentially fail. Due to an inventory oversight, a number of expired Uni-Knees were sold to hospitals and implanted in patients. As a result, in the early 2000s, Stryker was subject to over 70 individual product-liability claims and potentially obligated to cover Pfizer's losses as well. See Stryker, 735 F.3d at 352–53.
Stryker turned to its insurers for relief from exposure. Two policies, effective during the year 2000, are relevant here: a “commercial umbrella” policy, issued by XL, and an “excess liability” policy, issued by TIG. The umbrella policy covered any “batch” of losses that Stryker became “legally obligated to pay by reason of liability imposed by law or assumed by the [i]nsured ․ because of [b]odily [i]njury.” That policy was limited to $15 million, after a $2 million self-insured retention. The excess-liability policy followed form, kicked in after the umbrella policy was fully “exhausted,” and extended to Stryker's “ultimate net loss ․ in excess of all underlying insurance” up to $25 million.
The insurance companies balked at Stryker's request for defense and indemnification. In October 2001, XL denied coverage outright, arguing that the Uni-Knee claims were “known or suspected” prior to the inception of the policy, while TIG waited in the wings, hoping that its excess layer would not be implicated at all. Stryker, in turn, filed multiple lawsuits against XL and TIG in the Western District of Michigan. During the pendency of that protracted litigation, Stryker unilaterally settled all of its individual product-liability claims for $7.6 million and was separately adjudicated liable in the Southern District of New York for $17.7 million of Pfizer's losses. See Stryker, 735 F.3d at 353–54; see also Pfizer, Inc. v. Stryker Corp., 348 F. Supp. 2d 131, 159 (S.D.N.Y. 2004).
Stryker's case against XL was resolved over a decade later. In July 2012, we conclusively held that XL was obliged to “provide[ ] coverage for the claims made against Stryker in connection with [the defective] Uni-Knees.” Stryker, 735 F.3d at 356. With concrete figures in hand and its legal obligation apparent, XL decided to cover Stryker's losses. But it did so in non-chronological order: XL paid out the larger Pfizer judgment first, exhausted the limits of its coverage, and left Stryker's individual product-liability claims on the table. See id. at 357 & n.3 (“[T]he general rule is that an insurer may pay claims in any order it chooses.”).
And so, one score was left unsettled. In October 2013, back in the Western District of Michigan, Stryker filed a supplemental complaint against TIG, seeking to recover the remaining $7.6 million paid to settle its direct product-liability claims. TIG disputed its coverage obligation, raising a defense that was “unique to [its] policy.” Stryker, 681 F.3d at 825 & n.4. In TIG's view, the direct Uni-Knee claims did not constitute “ultimate net loss” because Stryker failed to obtain “written consent” at the time the settlements were made. Section III.B of the policy defines “ultimate net loss” as “the amount of the principal sum, award or verdict actually paid or payable in cash in the settlement or satisfaction of claims for which the insured is liable, either by adjudication or compromise with the written consent of [TIG], after making proper deduction for all recoveries and salvages.”
Discovery ensued and ultimately, in July 2014, the parties filed cross-motions for summary judgment. Stryker claimed that the policy, as applied to the idiosyncratic facts of this case, is latently ambiguous: because XL satisfied the Pfizer judgment first (and exhausted its policy), Stryker was forced to present its direct settlements to TIG years after they were made. Relying on the testimony of TIG's former claims adjusters and underwriters, Stryker argued that the excess-liability policy did not actually require “consent to the Uni-Knee settlements when they were made.” In response, TIG urged the court to apply the plain language of the policy, and disputed Stryker's characterization of its former employees' testimony.
In October 2014, the district court granted summary judgment to Stryker, concluding that the contract indeed contained a latent ambiguity. Stryker Corp. v. XL Ins. Co., 57 F. Supp. 3d 823 (W.D. Mich. 2014). “On the one hand,” the court said, “the policy, as interpreted and applied by TIG's own employees, does not require TIG's consent to settlements entered into below the TIG layer.” Id. at 831. “[O]n the other hand,” it recognized, the policy's plain language “does require consent if those settlements are [ultimately] offered to TIG for payment.” Id. Thus, in the court's view, the term “claims” in the definition of ultimate net loss was “susceptible to more than one interpretation.” Id. Having found the term ambiguous, the court then construed the contract against TIG based on various policy rationales. Id. at 831–35. All in all, the district court found that Stryker's claims were covered under the policy, TIG could muster no defenses to enforcement, and Stryker was entitled to a grand total of $8.6 million in damages and interest.
This appeal and cross-appeal followed.
We review the district court's grant of summary judgment de novo. Hobart Corp. v. Waste Mgmt. of Ohio, Inc., 758 F.3d 757, 765 (6th Cir. 2014). Summary judgment is proper if “there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law.” Fed. R. Civ. P. 56(a). At this stage, we view the facts and draw all reasonable inferences “in the light most favorable to the party opposing the [summary judgment] motion.” Scott v. Harris, 550 U.S. 372, 378 (2007) (quoting United States v. Diebold, Inc., 369 U.S. 654, 655 (1962) (per curiam)).
Michigan law governs this diversity suit, see Gasperini v. Ctr. for Humanities, Inc., 518 U.S. 415, 427 (1996), and an age-old maxim of contract law controls. When a written instrument is patently unambiguous, courts must give effect to that objective expression of contractual intent. See Rory v. Cont'l Ins. Co., 703 N.W.2d 23, 30 (Mich. 2005); Wilkie v. Auto-Owners Ins. Co., 664 N.W.2d 776, 782 (Mich. 2003); Rogers v. Great N. Life Ins. Co., 279 N.W. 906, 908 (Mich. 1938); 11 Williston on Contracts § 30:6 (4th ed. 2012). Extrinsic evidence cannot be summoned to aid interpretation. See Shay v. Aldrich, 790 N.W.2d 629, 641 (Mich. 2010); Restatement (Second) of Contracts § 213 (1981); James B. Thayer, The “Parol Evidence” Rule, 6 Harv. L. Rev. 417, 418 (1893) (“[Y]ou cannot, in construing a written contract, deed, will, or other like writing, use the direct extrinsic expressions of the writer's meaning evidentially.”).
True linguistic ambiguities are rather “rare in contract cases.” See E. Allen Farnsworth, “Meaning” in the Law of Contracts, 76 Yale L.J. 939, 954 (1967). All agree that this one is no exception. The excess-liability policy covers Stryker's “ultimate net loss,” defined in relevant part, as the amount paid “in the settlement or satisfaction of claims” for which the insured's liability is established by “adjudication” or “compromise with the written consent of [TIG].” In determining whether a contract is ambiguous, a district court must “give the contract language its ordinary and natural meaning.” Davis v. Sodexho, 157 F.3d 460, 585 (6th Cir. 1998) (citing Comerica Bank v. Lexington Ins. Co., 3 F.3d 939, 942 (6th Cir. 1993)). A reasonable person, with an ordinary understanding of the English language, would know what those italicized words mean: the policy requires TIG's “written consent” for any and all settlements. Nothing within the four corners of the contract suggests some otherwise hidden meaning.
The district court thought so as well. In June 2013, the court denied Stryker's prior motion for summary judgment, finding the contract clear on its face. “The TIG policy clearly provides only for payment of ․ settlements which have been entered into with TIG's written consent,” the court said, and “does not exclude the consent requirement for settlements within the underlying policy limits.” Stryker Corp. v. XL Ins. Am., Inc., No. 1:05-CV-51, 2013 WL 3276408, at *6 (W.D. Mich. June 27, 2013). We agree. Because Stryker did not satisfy the consent requirement, its direct settlements cannot constitute ultimate net loss, and there is no coverage under the policy. TIG rests its case there.
But Stryker seeks shelter in uncertainty of a different kind. It urges us to consider “the unusual facts of this case,” and conclude, as the district court did below, that the insurance contract contains a “latent ambiguity.” A written instrument that is clear on its face may, upon consideration of extrinsic evidence or some collateral matter, be rendered latently ambiguous when applied in the real world. See City of Grosse Pointe Park v. Mich. Mun. Liab. & Prop. Pool, 702 N.W.2d 106, 113 (Mich. 2005); In re Kremlick Estate, 331 N.W.2d 228, 230 (Mich. 1983) (per curiam); Mich. Chandelier Co. v. Morse, 297 N.W. 64, 66–67 (Mich. 1941); 11 Williston § 33:43.
The classic example is Raffles v. Wichelhaus, 2 H. & C. 906, 159 Eng. Rep. 375 (Ex. 1864). There, two parties contracted for a shipment of cotton “to arrive ex ‘Peerless' from Bombay.” Id. The terms were perfectly clear on the surface; but an ambiguity lurked just below. For there were in fact two cotton-bearing ships named Peerless setting sail from India, and the parties, of course, disputed which ship the contract referred to. The ambiguity in that contract arose, not from the words, but from their connection to the real world: “[T]he moment it appears that two ships called the ‘Peerless' were about to sail from Bombay there is a latent ambiguity, and parol evidence may be given for the purpose of shewing that the defendant meant one ‘Peerless,’ and the plaintiff another.” Id. at 376; see also A. W. Brian Simpson, Contracts for Cotton to Arrive: The Case of the Two Ships Peerless, 11 Cardozo L. Rev. 287, 291–92 (1989).
That rule holds true today. The Michigan courts admit parol evidence, in the first instance, to reveal the presence of a latent ambiguity and, at a later time, to allow the trier of fact to elucidate the meaning of that ambiguous contract. See Shay, 790 N.W.2d at 641; see also Sault Ste. Marie Tribe of Chippewa Indians v. Granholm, 475 F.3d 805, 816 (6th Cir. 2007); cf. 11 Williston § 33:43 (“[A]fter the court is satisfied that a latent ambiguity exists,” the question of “what the parties intended by language used in the contract—taking into consideration the extrinsic facts and circumstances—[is] an issue to be submitted to the jury.” (internal quotation marks and emphasis omitted)).
The doctrine has long been applied within certain “well defined limits.” See Ives v. Kimball, 1 Mich. 308, 313 (1849). The usual case of latent ambiguity will concern, for example, evidence of industry custom or trade usage, e.g., Sault Ste. Marie, 475 F.3d at 813 (unique meaning of the term “wager” in the gaming industry), course of dealing, e.g., Grosse Pointe Park, 702 N.W.2d at 115 (insurer's practice of covering certain claims without invoking an exclusion clause), or a misidentification, e.g., Kremlick, 331 N.W.2d at 230 (identity of a will's beneficiary). Breaking from the parol evidence rule, on such an occasion, is justified because it “enabl[es] courts to ascertain and carry into effect the intention of contracting parties.” See Ives, 1 Mich. at 313. But in balancing those “seemingly conflicting principles of contract law,” courts must be wary of altering a written instrument “under the guise of contract interpretation.” Grosse Pointe Park, 702 N.W.2d at 123 (Young, J., concurring in judgment). The courts may not admit extrinsic evidence to “create ambiguity where the terms of the contract are clear.” Frankenmuth Mut. Ins. Co. v. Masters, 595 N.W.2d 832, 837 (Mich. 1999).
Turning back to the case at hand, we ask whether the excess-liability policy is latently ambiguous. Is there an alternate real-world meaning for the plain language? Does it mean something other than what it appears to mean? See Shay, 790 N.W.2d at 641. We think not.
For a start, Stryker's understanding of the provision at issue cannot be squared with the contract as a whole. Stryker insists that the term “claims” actually means liability for settlements made without consent, so long as the compromise originally occurred below TIG's coverage layer. But the policy itself defines the universe of possible “claims” in two, and only two, ways. Liability must be established either by “adjudication” or by “compromise with the written consent of [TIG].” Stryker's reading thus goes too far: it seeks to fill a gap that does not exist, and contradicts the fundamental principle that “[p]arol evidence under the guise of a claimed latent ambiguity is not permissible to vary, add to, or contradict” any other “plainly expressed terms of [the] writing.” See Morse, 297 N.W. at 66.
But put that difficulty to one side. Even then Stryker's parol evidence is insufficient to reveal a latent ambiguity. See Sault Ste. Marie, 475 F.3d at 812 (“[T]he burden is on the party alleging the ambiguity.”). Stryker has not directed us, for instance, to any evidence that the term “claims” has a technical meaning in the excess insurance industry; it has not shown, by course-of-dealing evidence, that TIG habitually excludes the consent requirement in situations like this one; nor has Stryker pointed out the sort of real-world peculiarity at issue in the case of the two ships Peerless.
Stryker instead marshals the contested opinion testimony of TIG's former claims adjusters and underwriters. As each of these employees read the contract, Stryker informs us, consent was not required for settlements that originated “beneath TIG's policy layer.” But in the ordinary course, a latent ambiguity must be revealed by objective means—for instance, an admission, uncontested evidence, or the testimony of a disinterested third party. As we have said, interpreting Michigan law, one party's “subjective understanding” of what the contract meant is plainly “insufficient to create a latent ambiguity.” See Universal Settlements Int'l, Inc. v. Nat'l Viatical, Inc., 568 F. App'x 398, 402 (6th Cir. 2014). Absent that limitation, an ambiguity could be manufactured from whole cloth, and contracts could all too easily be rewritten according to one party's interests. Cf. AM Int'l, Inc. v. Graphic Mgmt. Assocs., Inc., 44 F.3d 572, 575 (7th Cir. 1995) (noting that, under Illinois law, only “ ‘[o]bjective’ evidence is admissible to demonstrate that apparently clear contract language means something different from what it seems to mean” because “[t]he ability of one of the contracting parties to ‘fake’ such evidence, and fool a judge or jury, is limited”).
This case proves the point. Stryker maintains, with some self-serving spin, that “all of TIG's claims handlers” believed it “was not required to obtain ․ consent when settling claims within XL's policy layer.” But the testimony, read in the light most favorable to TIG, see Scott, 550 U.S. at 378, demonstrates precisely the opposite. It shows, consistent with the plain language of the policy, that Stryker was required to obtain consent for any settlements that were ultimately presented to TIG for payment:
■ Michael Mulkey, a claims adjuster, stated that consent was not required for any sums within the “self-insured retention” and the “XL layer,” but that TIG's interest would be piqued if Stryker was “asking [it] to spend money.”
■ Valerie Lameka, another claims adjuster, said that she “could[n't] care less” about consent, unless TIG was “going to write the check.”
■ Norton Geller, a fellow claims adjuster, testified that “Stryker was free to settle the lawsuits when XL denied coverage, as long as it did not impact the TIG policy.”
■ Peter Salvin, yet another claims adjuster, similarly testified that consent was not required, so long as Stryker was “not spending [TIG's] money.”
■ Donald Bendure, an underwriter, said that consent was only required under the excess-liability policy if Stryker “call[ed] upon TIG to [pay].”
■ And Marilyn Schultz, another underwriter, attested that the consent-to-settle provision was not implicated “[a]s long as the settlements were paid by XL.”
Whatever its import, the inherent subjectivity of this testimony makes it insufficient for the purpose of detecting a latent ambiguity. See Universal Settlements Int'l, 568 F. App'x at 402; cf. AM Int'l, 44 F.3d at 575 (“[T]he testimony of the parties themselves as to what they believe the contract means ․ is invariably self-serving ․ and is inherently difficult to verify.”). We therefore conclude that Stryker has failed to reveal an ambiguity, and the plain language of the written instrument controls. See Sault Ste. Marie, 475 F.3d at 812.
Stryker asks us to affirm the district court's judgment anyway, raising a few defenses to contract enforcement. For example, Stryker contends that it sought “retroactive” consent, but was met with “unreasonable” recalcitrance on TIG's part. Stryker also tells us that any prior efforts to seek consent would have been “meaningless” and that TIG “waived” enforcement of the consent-to-settle provision. The burden of avoiding contractual arrangements “rests with those who would avoid them.” Morris v. Metriyakool, 344 N.W.2d 736, 742 (Mich. 1984). We conclude that Stryker has failed to carry that burden.
Stryker first claims that TIG has violated the implied covenant of good faith and fair dealing. Citing California law, Stryker argues that an “excess insurer does not have an absolute right to veto arbitrarily a reasonable settlement.” See Diamond Heights Homeowners Ass'n v. Nat'l Am. Ins. Co., 277 Cal. Rptr. 906, 916 (Cal. Ct. App. 1991). Maybe so, TIG responds, but the Michigan courts only impose that covenant “where one party to the contract makes its performance a matter of its own discretion.” Stephenson v. Allstate Ins. Co., 328 F.3d 822, 826 (6th Cir. 2003). The consent-to-settle provision, in TIG's view, is not such a discretionary term.
That debate is beside the point. For the first time, in 2013, Stryker sought TIG's consent to the direct product-liability settlements. TIG understandably denied the request as “untimely.” But Stryker now maintains, as the district court suggested, that “timeliness is not a valid basis for refusing consent” under the policy. See Stryker Corp., 57 F. Supp. 3d at 834. The contract cannot be read to suggest anything of the sort. The plain language of the policy mandates “compromise with the written consent of [TIG].” Not “separate and apart from,” not “retroactively,” and certainly not over a decade “after the fact.” Stryker's retroactivity argument, we thus conclude, improperly employs the obligation of good faith and fair dealing to “override express contract terms.” See Cook v. Little Caesar Enters., Inc., 210 F.3d 653, 657 (6th Cir. 2000). What is more, Stryker's reading of the contract runs up against the obvious purpose of the consent-to-settle provision—to prospectively “give the insurer the opportunity to contest liability, to participate in settlement negotiations and to have input as to the value of the claim.” See Alyas v. Gillard, 446 N.W.2d 610, 613 (Mich. Ct. App. 1989).
Stryker also argues that the consent-to-settle provision is an “immaterial condition,” one that TIG has, in any event, “waived” its right to invoke. We are not persuaded on either score. First, relying on Ranck v. Springer, 53 N.W.2d 678, 680 (Mich. 1952), Stryker asserts that its nonperformance was immaterial because TIG surely would have withheld consent. Even assuming the doctrine of futility applies here, Ranck is an inapt analogy. There, the Michigan Supreme Court simply recognized that, in a contract for the sale of land, a formal offer of tender is not required where one party attempts to provide payment, and the other party manifests its intent to refuse. Id. Speculation aside, no manifestation of futility is apparent on this record. See Weinburgh v. Saier, 6 N.W.2d 921, 923 (Mich. 1942). Stryker did not seek consent at the time of its settlements, so TIG obviously had no occasion to refuse it.
Second, citing Alyas, 446 N.W.2d at 613, Stryker maintains that it was “released from any agreement not to settle without the insurer's consent” because TIG “denied liability and wrongfully refused to defend.” But that contention rests on the false premise that XL's denial of coverage should be imputed to TIG, simply because the excess-liability policy “followed form.” As we have already explained, the latter contract contains provisions “that are unique to the TIG policy.” Stryker Corp., 681 F.3d at 825 n.4. And, as the district court previously reiterated, “XL's denial of coverage does not automatically release Stryker from the ‘consent to settle’ requirement.” Stryker Corp., 2013 WL 3276408, at *8. We decline to impute, so lightly, a waiver of the contractual right to “prevent collusion” and exercise “control over settlement negotiations.” See Coil Anodizers, Inc. v. Wolverine Ins. Co., 327 N.W.2d 416, 418 (Mich. Ct. App. 1982). Stryker did not seek excess coverage until long after entering into its settlements, so it cannot be said that TIG wrongly denied liability or refused to provide a defense at that time.
The district court's judgment is reversed, and the case is remanded with instructions to enter judgment for TIG.
COLE, Chief Judge.