Plaintiff, Brownell Combs, II, Administrator C.T.A. of the Estate of Leslie Combs, II, deceased, appeals an order granting Defendant, International Insurance Company, summary judgment against Plaintiff's action in diversity, brought pursuant to 28 U.S.C. § 1332, alleging breach of a directors and officers liability insurance contract, breach of the implied contractual duty of good faith and fair dealing, and bad faith denial of Defendant's duty to defend under the insurance policy. For the reasons set forth below, we AFFIRM the district court.

Decedent founded Spendthrift Farm ("Spendthrift") in 1937 with 120 acres of land near Lexington, Kentucky. By the early 1980s, the farm encompassed 1800 acres and housed forty-three stallions, including the last two Triple Crown winners. Decedent developed the principal method of stallion management used today.1

By 1981, Decedent was over eighty years old and became interested in planning his estate so that Spendthrift would continue after his death. During the early 1980s, Decedent tried several different methods to broaden Spendthrift's ownership. Initial efforts to take the company public failed when the farm could not locate a suitable investment bank. A 1982 effort to distribute forty percent of the ownership interest in Spendthrift failed because of disputes between Plaintiff2 and potential investors.

Eventually, the farm developed and implemented a private stock placement plan described in a Private Placement Memorandum ("PPM"). The PPM made clear that Decedent and Plaintiff, not Spendthrift, retained exclusive control over the private placement:

All sales are subject to the discretion of the Sellers including the right to accept each unit as purchased or none until the entire offering is purchased. Sellers reserve the right, in their absolute discretion, to accept or reject any offer to purchase, and/or to withdraw the offering either partially or in its entirety.

(J.A. at 842.) Furthermore, the lawyers involved in both the private placement (Charles Hembree, Decedent's long-time personal counsel), and the contemplated initial public offering ("IPO") (Frank Wheat of Gibson, Dunn & Crutcher), distinguished between the Combs family and Spendthrift Farm. Wheat took steps "to keep Spendthrift out of this private placement" to "protect Spendthrift against claims that might arise out of the placement." (J.A. at 748.)

Through the PPM, the farm sold blocks of stock to certain investors already involved in the thoroughbred industry, thereby creating a pool of shareholders who could create a board of directors and lead Spendthrift after Decedent's death. Decedent completed the stock sale in 1983. Decedent personally received $17.5 million from the transaction and Plaintiff received another $17.5 million.

The thoroughbred industry, including Spendthrift, prospered during the early 1980s. Later in the decade, however, the industry suffered a downturn from which it did not fully recover until the mid-1990s. Spendthrift's problems in the mid-to-late-eighties upset many of the investors in the private placement.

In January of 1984, Defendant issued excess policy no. 524-029517-1, which provided directors' and officers' liability coverage to Spendthrift for claims made against the farm from November 17, 1983 through November 17, 1986. The policy provided coverage as follows:

If during the policy period any claim or claims are made against the Insured (as hereinafter defined) or any of them for a Wrongful Act (as hereinafter defined) while acting in their individual or collective capacities as Directors or Officers, the Insurer will pay on behalf of the Insureds or any of them, their Executors, Administrators, Assigns 95% of all Loss (as hereinafter defined), which the Insureds or any of them shall become legally obligated to pay in excess of the retentions stated in Item IV(a) and (b) of the Declarations, not exceeding the limit of liability stated in Item III of the Declarations.

(J.A. at 34-35.) The policy defined "Wrongful Act" as "any actual or alleged error or misstatement or breach of duty by the Insureds while acting in their individual or collective capacities, or any matter not excluded by the terms and conditions of this Policy claimed against them solely by reason of their being Directors and Officers of the Company." Id.

In 1986, while the policy was in effect, Fred L. Fredricks sued Decedent, Plaintiff, and other co-defendants "in their individual capacity and as agents and employees of defendant Spendthrift Farm, Inc." (J.A. at 1049.) The Northern District of California consolidated the Fredericks case with seven other cases involving substantially similar claims (hereinafter the consolidated "California Litigation"). Approximately half of the California Litigation plaintiffs sued Spendthrift Farm itself for the alleged misrepresentations of its officers, directors and agents with respect to the private placement. The other half sued only the individual agents of the farm involved in the private placement.

Regardless, the substance of all the claims focused primarily on representations made in the PPM regarding the farm's financial condition and the value of its assets. Specifically, the plaintiffs in the California Litigation made two allegations: (1) that the PPM relied upon financial statements prepared on a current value basis rather than a cost basis; and (2) that the current value presentation misled the plaintiffs because it failed to include any provision for income taxes, thereby overstating the value of Spendthrift's assets by the amount of tax liability that would result from attempting to realize the assets' full value. According to the plaintiffs, the asset valuations were "substantially inflated and based on unrealistic assumptions about the quality, confirmation and other characteristics of the horses." Id. The plaintiffs sought rescission pursuant to Section 12 of the Securities Act of 1933, 15 U.S.C. § 77, along with damages for the allegedly fraudulent misrepresentations.

In a letter dated November 14, 1986, Paul Renne, Plaintiff's counsel in San Francisco, California, notified Defendant that the plaintiffs filed eight complaints against Defendant. Renne's letter sought reimbursement under the policy and Renne requested that Defendant communicate with him about Decedent's coverage demand.

In Defendant's response, Defendant's New York counsel explained, inter alia, that the wrongful conduct alleged against Decedent did not involve acts solely in his capacity as a director or officer of Spendthrift, but rather conduct in his individual capacity as a shareholder selling his shares in Spendthrift for his personal gain (and Plaintiff's personal gain) of $35 million:

The wrongful conduct alleged against the [Plaintiff and Decedent] in the [California Litigation] arises from the sale of their Spendthrift stock. International has observed from a review of the Private Placement Memorandum dated April 1, 1982 ("PPM"), which was distributed along with a supplement thereto dated July 27, 1983 ("PPM Supplement"), in connection with a "private placement" distribution of shares of Spendthrift by the [Plaintiff and Decedent] that the [Plaintiff and Decedent] offered shares in Spendthrift, in units of twenty at $1.75 million per unit, for a total offering of $35 million. All of the proceeds of the offering went to the [Plaintiff and Decedent], and none of the proceeds went to Spendthrift. The PPM states that [Decedent] is selling the portion of stock for the purpose of estate planning, and that [Plaintiff] is selling his portion of stock for estate planning and to diversify his interests....

As indicated above, the directors and officers of Spendthrift are insured only for loss incurred by them solely in their respective capacities as directors and officers. The wrongful conduct alleged against the [Plaintiff and Decedent] in the [California Litigation] does not involve them solely in their capacity as directors or officers but rather in their capacity as individuals who are selling their shares in Spendthrift for their own personal gain. Accordingly, International declines to afford coverage for any loss incurred by the [Plaintiff and Decedent] in connection with the [California Litigation].

(J.A. at 108-09.) Defendant addressed this letter to Renne, in California, as well as to certain defense counsel (located in California, Ohio, and Kentucky) who had an interest in the Policy because they represented other defendants.

Decedent settled some of the claims against him for $2 million in a court-approved settlement. The trial court dismissed numerous other claims. A few issues reached a jury, which returned a verdict in favor of the defendants. In a comprehensive opinion explaining the California Litigation in detail, the Ninth Circuit affirmed. See McGonigle v. Combs, 968 F.2d 810 (9th Cir.1992). When the California Litigation concluded in 1992, Decedent had spent $770,000 in attorney's fees in addition to the $2 million he paid to settle certain claims.

With respect to the court-approved settlement, Decedent submitted an extensive evidentiary record to persuade the district court that he had only a personal role in the disputed transactions:

(1) there is no evidence that [Decedent] wrote any part of the offering memos or that he was even consulted as to their text; (2) there is no evidence that [Decedent] consciously did anything wrong, or that he is guilty of any moral turpitude in this action; and (3) there is no evidence that [Decedent] personally misled anyone or sought to mislead anyone. [Decedent] is in these cases for one reason only: he was a seller, and there is a claim for recission to which he must respond although there is no evidence of scienter on his part.

(J.A. at 379-80.) The evidence Decedent submitted to the district court in the California Litigation included a due diligence memorandum prepared by counsel following a July 26, 1983 meeting with Decedent and Plaintiff. This memorandum distinguishes between the conduct of Decedent and Plaintiff as sellers, and their conduct as officers or directors for Spendthrift:

On Tuesday morning, July 26, we met with [Plaintiff] to discuss with him the timing of the private placement. [Plaintiff] informed us in rather blunt terms that he had no intention of delaying the private placement. He informed us that he was willing to accept the risks of lack of full disclosure and non-compliance with state securities laws. I advised [Plaintiff] that the risks of lack of going forward could be rather significant as it was my view that prudence called for a delay of the private placement of at least two weeks. [Plaintiff] informed the group that he knew each of the investors personally and had conducted a significant amount of business with them in the past. He informed us that he would give back any monies if the investors complained about the adequacy of the disclosure or non-compliance with Blue Sky laws.

(J.A. at 716-20.) Decedent's factual submission to the district court in the California Litigation also established that Decedent "was only interested in [the private placement] being done and having his check delivered to him and [Plaintiff] was supposed to take care of everything." (J.A. at 395.) Decedent quoted Plaintiff as stating that "his father had reached the point in his life where he would not have an attention span sufficient to go through complex legal documents or, for that matter, a conversation that lasted more than five minutes." (J.A. at 413.) Plaintiff did not dispute any part of the factual record submitted to the district court in support of the settlement in the California Litigation.

Decedent died in April of 1990, before the California Litigation concluded. Decedent had not yet brought an action against Defendant for his defense fees or the settlement payments, and Plaintiff still had not done so on Decedent's behalf when the California Litigation terminated in 1992. Decedent's estate was closed four years later, on June 21, 1996. Neither the administrator of the estate at the time, P. Keith Nally, nor the estate's attorney, Charles Hembree, considered commencing a claim against Defendant on Decedent's behalf.

In the Spring of 2000, four years after Decedent's estate closed, Plaintiff reopened the estate and arranged to become the administrator. On June 6, 2000, in the United States District Court for the Eastern District of Kentucky, Plaintiff commenced the action that is the subject of this appeal. Count I of the complaint asserted a breach of contract claim; Count II alleged a breach of the duty of good faith and fair dealing; and Count III asserted a bad faith claim.

Following discovery, Defendant moved for summary judgment on July 2, 2001. Defendant argued (1) that New York's statute of limitations applied and barred Plaintiff's claims; (2) equitable estoppel barred Plaintiff's claims; and (3) the policy did not cover the Decedent's role in the private placement. On September 10, 2001, the district court granted Defendant's motion on the basis that the applicable statute of limitations barred Plaintiff's claims. See Combs v. Int'l Ins. Co., 163 F. Supp. 2d 686 (E.D.Ky.2001). The district court also suggested that the good faith and fair dealing claim and the bad faith claim lacked merit.3 Id. at 695-96.

On September 20, 2001, Plaintiff filed a motion under Fed.R.Civ.P. 59(e) to amend the judgment to certify the statute of limitations issue to the Kentucky Supreme Court. After further briefing, the district court denied Plaintiff's motion on November 16, 2001. On November 29, 2001, Plaintiff timely noticed his appeal.

We conduct a de novo review of summary judgment. Eastman Kodak Co. v. Image Technical Servs., Inc., 504 U.S. 451, 466 n. 10, 112 S. Ct. 2072, 119 L. Ed. 2d 265 (1992); Buckeye Cmty. Hope Found. v. City of Cuyahoga Falls, 263 F.3d 627, 633 (6th Cir.2001); Johnson v. Econ. Dev. Corp., 241 F.3d 501, 509 (6th Cir.2001). Summary judgment is appropriate when there is no genuine issue of material fact such that the moving party is entitled to a judgment as a matter of law. Kocsis v. Multi-Care Mgmt., Inc., 97 F.3d 876, 882 (6th Cir.1996). In Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 106 S. Ct. 2505, 91 L. Ed. 2d 202 (1986), the Supreme Court explained that

[t]he mere existence of a scintilla of evidence in support of the plaintiff's position will be insufficient; there must be evidence on which the jury could reasonably find for the plaintiff. The judge's inquiry, therefore, unavoidably asks whether reasonable jurors could find by a preponderance of evidence that the plaintiff is entitled to a verdict.

Id. at 252, 106 S. Ct. 2505. The "mere possibility" of a factual dispute does not suffice to create a triable case. Gregg v. Allen-Bradley Co., 801 F.2d 859, 863 (6th Cir.1986). To defeat summary judgment, the plaintiff "must come forward with more persuasive evidence to support [his or her] claim than would otherwise be necessary." Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574, 586, 106 S. Ct. 1348, 89 L. Ed. 2d 538 (1986). If the defendant successfully demonstrates, after a reasonable period of discovery, that the plaintiff cannot produce sufficient evidence beyond the bare allegations of the complaint to support an essential element of his or her case, summary judgment is appropriate. Celotex Corp. v. Catrett, 477 U.S. 317, 325, 106 S. Ct. 2548, 91 L. Ed. 2d 265 (1986). When determining whether to reach this conclusion, we view the evidence and draw all reasonable inferences in the light most favorable to the non-moving party. Adickes v. S.H. Kress & Co., 398 U.S. 144, 157, 90 S. Ct. 1598, 26 L. Ed. 2d 142 (1970); Williams v. Int'l Paper Co., 227 F.3d 706, 710 (6th Cir.2000); Smith v. Thornburg, 136 F.3d 1070, 1074 (6th Cir.1998).

We must decide whether the district court properly ruled that under Ky.Rev.Stat. § 413.320, the Kentucky "borrowing statute," the New York statute of limitations bars Plaintiff's action. This is a highly uncertain area of state law, forcing us to make an educated "Erie guess." Plaintiff contends, alternatively, (1) that the Kentucky judiciary would interpret Kentucky's borrowing statute so as to apply the statute of limitations of the jurisdiction that, considering all the facts, has the most significant connection with the disputed transaction, and (2) even if the locus of accrual is the only relevant consideration, this cause of action accrued in Kentucky, making Kentucky's statute of limitations applicable.

New York's statute of limitations for breach of a written contract is six years. N.Y. CIV. PRAC. LAW & RULES § 213. Kentucky has a fifteen-year statute of limitations for the breach of a written contract. KY.REV.STAT. § 413.090(2). Defendant notified Plaintiff that it would deny coverage on November 14, 1986, and Plaintiff brought suit on June 6, 2000. Thus, if New York's statute of limitations applies, as opposed to Kentucky's, Plaintiff's claim is barred.

In accordance with Erie Railroad Co. v. Tompkins, 304 U.S. 64, 58 S. Ct. 817, 82 L. Ed. 1188 (1938), when evaluating an undecided question of Kentucky law, a federal court sitting in diversity must make the "`the best prediction, even in the absence of direct state precedent, of what the Kentucky Supreme Court would do if it were confronted with [the] question.'" Managed Health Care Assocs., Inc. v. Kethan, 209 F.3d 923, 927 (6th Cir.2000) (quoting Welsh v. United States, 844 F.2d 1239, 1245 (6th Cir.1988)). If the relevant state judiciary has not spoken to the issue, courts sitting in diversity should consider "all relevant data," Kingsley Associates, Inc. v. Moll PlastiCrafters, Inc., 65 F.3d 498, 507 (6th Cir.1995), including jurisprudence from other jurisdictions, Lexington Insurance Co. v. Rugg & Knopp, Inc., 165 F.3d 1087, 1090 (7th Cir.1999).

The Seventh Circuit observed that federal courts "must proceed with caution" when making pronouncements about state law. Lexington, 165 F.3d at 1092. Sitting in diversity, we are "not commissioned to take a position regarding the advisability or fairness of the state rule to be applied, but [must] determine the issue as would the highest court of the state." Kurczi v. Eli Lilly & Co., 113 F.3d 1426, 1429 (6th Cir.1997). This Court's proper reluctance to speculate on any trends of state law applies with special force to a plaintiff in a diversity case, like this one, who has chosen to litigate his state law claim in federal court. Torres v. Goodyear Tire & Rubber, Inc., 867 F.2d 1234, 1238 (9th Cir.1989); Shaw v. Republic Drill Corp., 810 F.2d 149, 150 (7th Cir.1987). Furthermore, "[w]hen given a choice between an interpretation of [state] law which reasonably restricts liability, and one which greatly expands liability, we should choose the narrower and more reasonable path." Todd v. Societe Bic, S.A., 21 F.3d 1402, 1412 (7th Cir.1994) (en banc); see also City of Phila. v. Beretta, U.S.A., Corp., 126 F. Supp. 2d 882, 906 (E.D.Pa.2000) (declining to broaden state law nuisance doctrine in a diversity case).

As the First Circuit explained, federal courts sitting in a diversity case are in "a particularly poor position ... to endorse [a] fundamental policy innovation.... Absent some authoritative signal from the legislature or the courts of [the state], we see no basis for even considering the pros and cons of innovative theories...." Dayton v. Peck, Stow & Wilcox Co. (Pexto), 739 F.2d 690, 694 (1st Cir.1984). Federal courts hearing diversity matters should be extremely cautious about adopting "substantive innovation" in state law. Rhynes v. Branick Mfg. Corp., 629 F.2d 409, 410 (5th Cir. Unit A 1980). Thus, we must handle this issue charily.

As noted, Plaintiff contends that the Kentucky Supreme Court would interpret Kentucky's borrowing statute so that whenever substantive Kentucky law governs a claim involving multiple jurisdictions, Kentucky will apply the statute of limitations of the forum with the most significant relationship to the dispute.

A borrowing statute is a legislative exception from the general rule that the forum always applies its statute of limitation. Miller v. Stauffer Chem. Co., 99 Idaho 299, 581 P.2d 345, 348 (1978) ("Borrowing statutes change the common law rule governing choice of the applicable statute of limitation.") (citation omitted). Borrowing statutes vary somewhat, but all provide that the forum state will apply the statute of limitations from the foreign jurisdiction in which the cause of action accrued. Kentucky's borrowing statute provides:

When a cause of action has arisen in another state or country, and by the laws of this state or country where the cause of action accrued the time for commencement of an action thereon is limited to a shorter period of time than the period of limitation prescribed by the laws of this state for a like cause of action, then said action shall be barred in this state at the expiration of said shorter period.

KY.REV.STAT. § 413.320. Thus, if a cause of action arises in a foreign jurisdiction which has a shorter statute of limitations than Kentucky for the same cause of action, Kentucky courts must "borrow" the foreign jurisdiction's statute of limitations. Almost three-fourths of states have statutes that resemble Kentucky's. See Ibrahim J. Wani, Borrowing Statutes, Statutes of Limitations and Modern Choice of Law, 57 UMKC L.REV. 681, 690 (1989); Donna Mae Endreson, Wisconsin's Borrowing Statute: Did We Shortchange Ourselves?, 70 MARQ. L.REV. 120, 122-27 (1986).

Historically, local statutes of limitations governed. At common law, statutes of limitation were procedural, not substantive, a fact which the Second Circuit rightly termed "an accident of history." Bournias v. Atl. Mar. Co., 220 F.2d 152, 154 (2d Cir.1955). In most civil law countries, the statute of limitations applicable to an action would come from the jurisdiction whose substantive law formed the action's basis. ERNST RABEL, THE CONFLICT OF LAWS: A COMPARATIVE STUDY 511-16 (1964); Edgar H. Ailes, Limitations of Actions and the Conflict of Laws, 31 MICH. L.REV. 474, 478 (1933). Early English courts first confronted with conflict-of-law problems acted to protect English law from continental influence and thus took the opposite route. Ernest H. Lorezen, The Statute of Limitations and the Conflict of Laws, 28 YALE L.J. 492, 496 (1919). Carrying the English tradition to America, in M'Elmoyle v. Cohen, 38 U.S. (13 Pet.) 312, 10 L. Ed. 177 (1839), the Supreme Court held that a state may treat statutes of limitations as procedural and thus may apply its own limitations law even when it must apply another jurisdiction's substantive law. Id. at 328; see also Sun Oil Co. v. Wortman, 486 U.S. 717, 722, 108 S. Ct. 2117, 100 L. Ed. 2d 743 (1988); Wells v. Simonds Abrasive Co., 345 U.S. 514, 516-18, 73 S. Ct. 856, 97 L. Ed. 1211 (1953); Townsend v. Jemison, 50 U.S. (9 How.) 407, 413-20, 13 L. Ed. 194 (1850).

Before borrowing statutes, when a claim arose outside the forum, courts handled the limitations problem in one of two ways. First, if the forum's rules barred the claim, the forum's courts would not hear the claim regardless of its status elsewhere. See, e.g., Panhandle E. Pipe Line Co. v. Parish, 168 F.2d 238, 241 (10th Cir.1948); Corrigan v. Clairol, Inc., 126 F. Supp. 791, 792 (D.Conn.1954). Second, the forum would hear the claim as long as the local limitations period had not expired, even if the limitations period applicable where the claim arose had run. See, e.g., Filson v. Fountain, 197 F.2d 383, 384 (D.C.Cir.1952) (per curiam); Goodwin v. Townsend, 197 F.2d 970 (3d Cir.1952); Jackson v. Cont'l S. Lines, Inc., 172 F. Supp. 809, 812 (W.D.Ark.1959). In essence, lex fori governed the determination of procedural issues, including those associated with statutes of limitation.

Lex fori governs procedural rights, but under lex loci (the law of the place), substantive rights are determined according to the law of the place where the cause of action accrued. Lex loci involves a concept of accrual similar to the idea of vesting, which forms the basis for the vested-rights approach to choice of law. Replacing nineteenth-century comity theory but preceding contemporary choice of law paradigms, the vested rights approach dominated the period from 1900 to 1950. EUGENE F. SCOLES, ET AL., CONFLICT OF LAWS § 2.7, at 20 (3d ed.2000). According to Professor Joseph H. Beale, the main proponent of vested rights analysis, "[a] right having been created by the appropriate law, that recognition of its existence should follow anywhere." JOSEPH H. BEALE, 3 CASES ON THE CONFLICT OF LAWS 517 (1901). As Justice Holmes explained:

[W]hen ... a liability is enforced in a jurisdiction foreign to the place of the wrongful act, obviously that does not mean that the act in any degree is subject to the lex fori, with regard to either its quality or its consequences. On the other hand, it equally little means that the law of the place of the act is operative outside its own territory. The theory of the foreign suit is that, although the act complained of was subject to no law having force in the forum, it gave rise to an obligation, an obligatio, which, like other obligations, follows the person, and may be enforced wherever the person may be found. But as the only source of this obligation is the law of the place of the act, it follows that that law determines not merely the existence of the obligation, but equally determines its extent.

Slater v. Mexican Nat'l R.R. Co., 194 U.S. 120, 126, 24 S. Ct. 581, 48 L. Ed. 900 (1904) (citations omitted). Beale served as the reporter for the Restatement of the Conflict of Laws (1934), which reflected the vested rights approach. Section 384 of the Restatement provides, for instance: "(1) [i]f a cause of action in tort is created at the place of the wrong, a cause of action will be recognized in other states[;] (2)[i]f no cause of action is created at the place of wrong, no recovery in tort can be had in any other state." Kentucky adopted its borrowing statute in 1942, during the heyday of the vested rights approach.

After most states adopted borrowing statutes, problems in the application of borrowing rules and a torrent of scholarly criticism caused some state high courts to consider interpreting their state's borrowing statutes so the forum would "borrow" another state's limitation rule only when the local forum lacked a significant connection to the dispute. Critics of borrowing statutes note that this would be consistent with the Restatement (Second) on Conflict of Laws, which seeks to apply the law of the forum with the most significant relationship with the parties and the dispute. See RESTATEMENT (SECOND) OF CONFLICT OF LAWS § 6 (1969). The "most significant relationship" test that guides the Second Restatement has at least arguable policy advantages over lex fori because the state with the most significant relationship with the parties and the dispute is probably the state with the greatest interest in the action's outcome.4 In fact, one state, Minnesota, repealed its borrowing statute in 1977. See MINN.STAT. § 541.14 (1976), repealed by 1977 MINN. LAWS ch. 187 § 1.

Plaintiff cites a number of law review articles arguing that borrowing statutes should be repealed entirely or interpreted to incorporate a "most significant relationship" test. See, e.g., Ibrahim J. Wani, Borrowing Statutes, Statutes of Limitations, and Modern Choice of Law; 57 UMKC L.REV. 681 (1989); Donna Mae Endreson, Wisconsin's Borrowing Statute: Did We Shortchange Ourselves?, 70 MARQ. L.REV. 120 (1986); Samuel J. Morley, Applying the Significant Relationships Test to Florida's Borrowing Statute, 59 FLA. BAR. J. 17 (1985); Donald R. Rigone, The Impact of Significant Contacts on the Pennsylvania Borrowing Statute, 72 DICK. L.REV. 598 (1968); David H. Vernon, Statutes of Limitation in the Conflict of Laws: Borrowing Statutes, 32 ROCKY MTN. L.REV. 287 (1960). However, none of these law review articles are binding on this Court. For that reason, we will conduct our own inquiry.

Plaintiff highlights five states whose judiciaries have, according to Plaintiff, interpreted borrowing statutes to incorporate a most significant relationship test despite statutory language essentially indistinguishable from Kentucky's borrowing statute.

Beginning with Wyoming, Plaintiff claims that BHP Petroleum v. Texaco Exploration Production, 1 P.3d 1253 (Wyo.2000), "addresses precisely the question before the Court in this case." (Pl.'s Br. at 32) (Plaintiff's emphasis). In BHP, the Wyoming Supreme Court considered a dispute between two oil companies, each with offices located in both Texas and Colorado. Id. at 1255. BHP and Texaco had an ongoing relationship in which they shared royalties from minerals extracted in Wyoming. Id.

BHP allegedly made an offer to Texaco to change the formula by which the companies calculated royalties. Id. BHP sent that letter from Texas to Colorado. Id. Texaco allegedly accepted in a response sent from Colorado back to BHP's office in Texas. Id. Later, Texaco sent another letter from Texas to Colorado that disclaimed any intention to recalculate royalty payments. Id. Wyoming's borrowing statute states that "[i]f by the laws of the state or country where the cause of action arose the action is barred, it is also barred in this state." Wyo. STAT. ANN. § 1-3-117. Colorado had a shorter statute of limitations that would have barred BHP's action; BHP thus argued that Wyoming's longer limitations period should apply. 1 P.3d at 1256. The Wyoming court, applying the borrowing statute, found that "the breach occurred in Colorado where BHP received the letter." Id. at 1258.

We note several points about BHP. First, the BHP court provides scant detail on the nature of the breach. Not only does the court appear skeptical that any breach occurred, the court never defines the dispute with precision. From the facts, it appears the case involved BHP's attempt to modify the terms of an existing transaction — something different than Plaintiff's request that Defendant remit (or pledge to remit) certain monies allegedly in accordance with the terms of a Directors and Officers (hereinafter "D & O") insurance contract.

Second, BHP urged the court to adopt the Wyoming limitations period because, according to BHP, the court should "focus[] upon the location of the subject matter of the contract." Id. at 1256. Thus, the plaintiff in BHP argued in favor of the limitation period associated with the forum of the subject matter (the Wyoming resources), while the defendant argued for the limitation period associated with the forum of the breach. BHP evidently did not disagree that if the breach was the only relevant consideration, the breach occurred in Colorado, as opposed to Texas, where Texaco posted the letter. The latter scenario, however, is much more akin to the issue we presently face. Plaintiff's counsel sent a letter from California to Defendant's New York office, and the New York office replied with a letter sent to lawyers in three states — California, Kentucky and Ohio. Thus, as the origin of the letter that allegedly constituted the breach, New York is the analogue of Texas in BHP, not Colorado. No one argues in the instant case that the California statute of limitations might apply.

Finally, BHP never adopted the "substantial relationship" test. According to the BHP decision:

The thread of the argument presented by BHP is tenuous. The essential premise is that in Stanbury v. Larsen, 803 P.2d 349, 355 (Wyo.1990), this Court adopted Restatement (Second) Conflict of Laws § 188 (1971). BHP then contends that since that provision utilizes the "most significant relationship" test, the cause of action arose in Wyoming because the subject matter of the contract is Wyoming minerals. The elements articulated in the "most significant relationship" test include "(a) the place of contracting, (b) the place of negotiation of the contract, (c) the place of performance, (d) the location of the subject matter of the contract, and (e) the domicil, residence, nationality, place of incorporation and place of business of the parties." RESTATEMENT (SECOND) CONFLICT OF LAWS § 188 at 575.

BHP, 1 P.3d at 1256. After considering its Stanbury decision, the court concluded that "[w]e do not understand that this Court adopted [in Stanbury] the `most significant relationship' test of Restatement (Second) Conflict of Laws § 188.... [W]e invoked the Restatement provision only as containing examples of factors to be considered in making the determination as to where the cause of action arose." Id. at 1257. Thus, BHP does not indicate that the Kentucky Supreme Court would adopt the most substantial relationship test.

Plaintiff next cites New York law and Global Financial Corp. v. Triarc Corp., 93 N.Y.2d 525, 693 N.Y.S.2d 479, 715 N.E.2d 482 (1999). Global Financial involved an action brought in New York for the payment of consulting fees by a corporate plaintiff located in Delaware against a corporate defendant also located in Delaware but with its principal place of business in Pennsylvania. Id. at 484-85. Statutes of limitations in both Delaware and Pennsylvania barred the action, but the suit could proceed under New York's longer limitations period. Id. at 484. New York has a borrowing statute similar to Kentucky's. See id.

The court concluded that the breach occurred in Delaware because, under New York's borrowing statute, "a cause of action accrues at the time and in the place of injury," and "[w]hen an alleged injury is purely economic, the place of injury is usually where the plaintiff resides and sustains the economic impact of the loss." Id. at 485 (citations omitted). The Global Financial court did not adopt a "most substantial relationship" analysis of New York's borrowing statute. In fact, the court rejected the more holistic "center of gravity" test, writing that the goals of New York's borrowing statute are "better served by a rule requiring the single determination of a plaintiff's residence than by a rule dependent on a litany of events relevant to the `center of gravity' of a contract dispute."5 Id. at 486.

Equally important, the notion that the cause of action accrues where the injury is sustained is not particularly helpful in this case because it begs the question of where the plaintiff sustained the injury. Plaintiff's action involves an abstract injury — by allegedly breaching its promise to pay in a letter Defendant mailed from New York to three jurisdictions, Decedent was not reimbursed for litigation expenses accumulated primarily in California but related to a Kentucky enterprise. Asking where Decedent "got hurt" does not help us.6

Citing National Heritage Life Ins. Co. v. Frame, 41 S.W.3d 544 (Mo.Ct.App.2001), Plaintiff claims Missouri interpreted its borrowing statute in a fashion similar to New York. National Heritage involved a suit brought by a Texas plaintiff based on a guaranty allegedly breached by Missouri defendants. Id. at 546-48. Missouri's borrowing statute states that "[w]henever a cause of action has been fully barred by the laws of the state, territory or country in which it originated, said bar shall be a complete defense to any action thereon, brought in any of the courts of this state." MO.REV.STAT. § 516.190. Missouri courts analyze the "origination" of a cause of action just as they analyze its accrual for purposes of beginning the limitations period. Nat'l Heritage, 41 S.W.3d at 552; see also Penalosa Coop. Exch. v. A.S. Polonyi Co., 754 F. Supp. 722, 733 (W.D.Mo.1991); Alvarado v. H & R Block, Inc., 24 S.W.3d 236, 242 (Mo.Ct.App.2000). Since the guaranty required the guarantors to make their payment to the plaintiff in Texas, the National Heritage court found that the cause of action "originated" there. 41 S.W.3d at 553.

Following this rule, however, has not led Missouri courts to always conclude that a cause of action "originates" in the forum where the injury appears to occur. In Finnegan v. Squire Publishers, Inc., 765 S.W.2d 703 (Mo.Ct.App.1989), for instance, the Missouri Court of Appeals applied Missouri's borrowing statute in a libel case. Id. at 704. Finnegan forced the court to determine whether the cause of action for libel "accrue[d] in Kansas where the newspaper containing the defamatory statements was first published, [or] in Missouri where [the plaintiff] is licensed to practice law and where damages to his professional reputation were sustained." Id. Applying Missouri's borrowing statute, the court found that plaintiff's cause of action "originated" where the defendant initially published the defamatory remarks, not where the plaintiff actually suffered the greatest injury. Id. at 706-07. This is in tension with Plaintiff's claim that if we properly interpret Kentucky's borrowing statute, New York's statute of limitations cannot bar his claim because he suffered no injury in New York.

The Finnegan court reached its conclusion in part because the "[p]laintiff's reputation interest is invaded at the time of publication, and arguably that is the time when his damage is sustained." Id. at 706. Thus, at least in Missouri, the location of the wrong is determined by where the plaintiff first sustains any damage, not where he sustains the most damage. Also significant, the Finnegan decision recognized the problems inherent in attempting to determine the location of an intangible injury, like an injury to reputation. Id. at 706-07. The court refused "to adopt the assumption that an attorney's reputation can only be injured in the state where the attorney is licensed to practice law." Id. at 706. Rhetorically, the court wondered, "[w]hat would result if plaintiff is licensed in more than one state or by a federal court or agency? Should a court look to the place of first injury, any injury, or the place of greatest injury?" Id. This discussion shows that Missouri courts seem sensitive to the pitfalls inherent in determining accrual based on where the injury purportedly occurred.

Another case, Harris-Laboy v. Blessing Hospital, Inc., 972 S.W.2d 522 (Mo.Ct.App.1998), further exemplifies the manner in which Missouri courts apply the state's borrowing statute. Harris-Laboy involved a Missouri resident who brought a malpractice action against an Illinois hospital and three Illinois physicians after she discovered the doctors left a surgical sponge in her abdomen. Id. at 523. At the time of the surgery, the plaintiff lived in Illinois, but she did not discover the problem until she moved to Missouri several years later. Id. The defendants argued that Missouri should borrow Illinois' statute of limitations, which would bar the plaintiff's action. Id. at 523-24. The court agreed because, under Missouri law, an action originates when and where an injury can be ascertained, and "[d]amage is sustained and capable of ascertainment when it can be discovered or made known, not when the plaintiff actually discovers the injury or wrongful conduct." Id. at 524 (citing Carr v. Anding, 793 S.W.2d 148, 150 (Mo.Ct.App.1990)) (emphasis in original). Consequently, the plaintiff's "damage was sustained and was capable of ascertainment immediately after the sponge was left inside her in Illinois." Id. at 525. Whatever the merits of this rule, it does not appear to help Plaintiff because, applying Harris-Laboy, Plaintiff suffered an injury the moment Defendant drafted its response to Plaintiff's coverage inquiry, as opposed to when Plaintiff actually learned what Defendant would do.7 Viewed overall, Missouri jurisprudence does not necessarily focus on the place of injury, nor has it explicitly endorsed the "most significant contacts" test. In fact, Missouri has twice rejected express invitations to reinterpret Missouri's borrowing statute to exclude situations in which Missouri does have the "most significant contacts." See Dorris v. McClanahan, 725 S.W.2d 870, 872 (Mo.1987) (en banc) ("[T]he Missouri legislature [through its borrowing statute] preempts an analysis of 2d Restatement significant contacts in this case."), overruled on other grounds by Thompson v. Crawford, 833 S.W.2d 868 (Mo.1992); Trzecki v. Gruenewald, 532 S.W.2d 209, 211 (Mo.1976).

Plaintiff next emphasizes Illinois law and Employers Insurance of Wausau v. Ehlco Liquidating Trust, 723 N.E.2d 687 (Ill.Ct.App.1999) (Ehlco II), an Illinois Court of Appeals opinion issued on remand from the Illinois Supreme Court's decision in Employers Insurance Co. v. Ehlco Liquidating Trust, 186 Ill. 2d 127, 237 Ill. Dec. 82, 708 N.E.2d 1122 (1999) (Ehlco I). Elhco II dealt with one portion of a complex insurance coverage dispute. 243 Ill. Dec. 384, 723 N.E.2d at 691. Ehlco was a trust created by order of the Delaware Chancery Court to resolve the contingent liabilities of the Edward Hines Lumber Company ("Hines"), a dissolved Delaware corporation, and various affiliated businesses. Id. at 690. Plaintiff Wausau instituted a declaratory judgment action against Ehlco seeking a declaration, inter alia, that it had no defense or indemnity obligations under certain insurance policies in connection with a lawsuit filed against Ehlco relating to a contaminated industrial site in Wyoming. Id. Elhco filed a counterclaim against Wausau alleging that the insurer breached its duty to defend and indemnify Ehlco in connection with contamination at another site in Arkansas. Id. The Elhco II opinion deals exclusively with the counterclaim. Id. at 692.

In 1982, pursuant to the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 ("CERCLA"), 42 U.S.C. § § 9601-9675, the United States Environmental Protection Agency ("EPA") sent Hines a letter notifying Hines that it faced potential liability for cleanup costs at the Arkansas site. 243 Ill. Dec. 384, 723 N.E.2d at 692. Hines, based in Chicago, notified Wausau that the EPA would assert liability against Hines. Id. The insurer responded, appearing to deny coverage based on various policy exclusions. Id. Hines' counsel then wrote to Wausau twice, once in 1982 and once in March of 1983, each time "formally" requesting that it defend Hines against the EPA's proceedings. Id. Wausau replied several weeks after the March letter with a memorandum that referred to Wausau's first response but also requested "a copy of the [EPA] action or correspondence ... [and] copies of any complaints or summons." Id. No further communication occurred until 1987, when Hines informed Wausau that the company had almost reached a "final agreement" with the EPA. Id. Wausau responded immediately, again recounting its prior reservation of rights and requesting the details of the proceedings. Id. The Illinois court noted that "[t]he submissions of the parties do not show that either Hines or its counsel responded to Wausau's [two] requests for information." Id. at 692-93. In 1988, the EPA filed suit against Hines in Arkansas federal district court and simultaneously proposed a consent decree. Id. at 693. Wausau received news of that filing from a source other than Hines. Id.

Illinois has a borrowing statute similar to Kentucky's. Like Kentucky's, the Illinois statute states that "[w]hen a cause of action has arisen in a state or territory out of this State, or in a foreign country, and, by the laws thereof, an action thereon cannot be maintained by reason of the lapse of time, an action thereon shall not be maintained in this State." 735 ILL. COMP. STAT. 5/13-210. Arkansas' statute of limitations barred Ehlco's action, but Ehlco could proceed under Illinois' longer limitations period. 243 Ill. Dec. 384, 723 N.E.2d at 692. The court relied heavily on the Restatement (Second) of Conflict of Laws and considered numerous connections between Illinois and the dispute including:

the issuance of the insurance policies to Hines in Illinois, the location of Hines's principal place of business in Illinois, and the licensing of Wausau to do business in Illinois. Additional contacts with Illinois include the fact that correspondence relative to the EPA lawsuit was directed to Hines in Illinois; Hines's requests to Wausau for defense originated in Illinois; Wausau's responses to those requests were sent to Hines in Illinois; and Hines's legal counsel that provided Hines's defense with respect to the EPA proceedings was located in Illinois and would have billed its defense fees to Hines in Illinois. Finally, any monies found to be owing Hines by Wausau as a result of the counterclaim would have been due Hines in Illinois.

Ehlco II, 243 Ill. Dec. 384, 723 N.E.2d at 696-97 (citing RESTATEMENT (SECOND) CONFLICT OF LAWS § 188(2) (1971)). The court thus found that Illinois' statute of limitations would apply because Illinois uses the limitation period of the forum "that bears the most significant relationship to the contract dispute," id. at 243 Ill. Dec. 384, 723 N.E.2d at 695, although Illinois law is just one factor among many we must consider when making an Erie guess.

Furthermore, the Ehlco II court was careful to "note that the inquiry as to where the cause of action arose presupposes that the cause of action has in fact arisen." 243 Ill. Dec. 384, 723 N.E.2d at 693. In Ehlco I, the Illinois Supreme Court "held that Wausau's duty to defend was triggered by the filing of the EPA lawsuit in the federal district court in Arkansas seeking entry of the proposed consent decree." Id. (citing Ehlco I, 237 Ill. Dec. 82, 708 N.E.2d at 1130). The Illinois Supreme Court "also found that notwithstanding the triggering event of the filing of the lawsuit, Wausau would not have had a duty to defend Hines in the EPA action unless Hines tendered the lawsuit to Wausau for defense or Wausau had actual notice of the lawsuit." Id. (citing Ehlco I, 237 Ill. Dec. 82, 708 N.E.2d at 1131). The Illinois Supreme Court remanded the notice issue to the trial court with instructions that the parties "be given the opportunity to amend their pleadings to address the actual notice issue." Ehlco I, 237 Ill. Dec. 82, 708 N.E.2d at 1131-32. The Illinois Court of Appeals, however, explained that "notwithstanding the mandated examination of the notice issue by the [trial] court, we should, in the interest of judicial economy and our mandate from the Supreme Court, proceed to consider the questions of whether a cause of action arose and the residency of the parties." Ehlco II, 243 Ill. Dec. 384, 723 N.E.2d at 694. Thus, the entire decision in Ehlco II assumed, without deciding, that the plaintiff had a cause of action.

This assumption, required by the Illinois Supreme Court, probably had the unintended effect of obscuring the potential significance of the forum in which the cause of action accrued. According to the Illinois Supreme Court, no cause of action arose against Wausau unless and until Wausau had actual notice of the "trigger[]," which was the EPA lawsuit. Ehlco I, 237 Ill. Dec. 82, 708 N.E.2d at 1130-31. Therefore, the triggering event was the filing of the EPA's suit in Arkansas, and the cause of action did not arise until Wausau, the defendant, received "actual notice" of the lawsuit. Conceivably, Wausau could have received that notice anywhere-even Arkansas. This casts doubt on Ehlco II's reasoning.8

Additionally, the Ehlco II Court noted that "the location of the subject matter of the contract, such as the location of the risk insured by an insurance policy, is entitled to little weight when the subject matter or risk is located in more than one state." 243 Ill. Dec. 384, 723 N.E.2d at 694 (citing Lapham-Hickey Steel Corp. v. Prot. Mut. Ins. Co., 166 Ill. 2d 520, 211 Ill. Dec. 459, 655 N.E.2d 842, 845 (1995) and RESTATEMENT (SECOND) OF CONFLICT OF LAWS § 193, cmt. b (1971)). As a consequence, the court gave the location of the risk less salience in its "most significant relationship" analysis because the plaintiff's Comprehensive General Liability Policy ("CGL") covered environmental risks in multiple states. Similarly, Decedent's D & O policy covered certain D & O litigation-related expenses wherever they might occur. This counsels against excessively focusing our inquiry on the locus of the policy's "subject matter."

As already alluded to, it is hard to determine where the risks a D & O policy insures against are located. It would be simplistic to conclude that a risk is always located in the insured's state of residence because insurance protects only against financial loss, which is ultimately felt at a person's residence or a corporation's headquarters. A policy can cover risks in multiple states, as the Second Restatement suggests, see RESTATEMENT (SECOND) OF CONFLICT OF LAWS § 193, cmt. b (1971), and the Second Restatement cautions against focusing on the location of the risk when risks are dispersed across forum boundaries, see id., which means the residence of someone insured under a D & O policy like the one Decedent held should make little difference. This is further underscored by the fact that many D & O policies cover multiple directors and officers, each potentially residing in different fora. The limited importance given to risk location in the Ehlco litigation does not support the outcome Plaintiff urges.

Florida is the final jurisdiction whose law Plaintiff asks us to consider. Like Kentucky, Florida's borrowing statute provides that "[w]hen the cause of action arose in another state or territory of the United States, or in a foreign country, and its laws forbid the maintenance of the action because of lapse of time, no action shall be maintained in this state." FLA. STAT. § 95.10. Plaintiff first cites Bates v. Cook, 509 So. 2d 1112 (Fla.1987), for the proposition that Florida applies a "most significant relationship" test to its borrowing statute when the cause of action sounds in tort. Plaintiff is partly correct — Florida adopted the "most significant relationship" test in Bishop v. Florida Specialty Paint Co., 389 So. 2d 999, 1001 (Fla.1980), but that case did not involve Florida's borrowing statute. Rather, Bishop adopted the Second Restatement's approach only with respect to the determination of what substantive law would apply to torts. Id. at 1000. The court said nothing about statutes of limitations, which are procedural. The court also remained silent about contract disputes.9 Id.

With respect to contract disputes, Plaintiff cites Lumbermens Mutual Casualty Co. v. August, 530 So. 2d 293 (Fla.1988). Lumbermens involved a Massachusetts driver insured by a policy with uninsured motorist benefits. Id. at 294. The plaintiff signed the policy in Massachusetts before having an accident with an uninsured motorist in Florida. Id. She then brought suit against her insurer for benefits after Massachusetts' limitations period expired but before Florida's terminated. Id. The court found Florida's statute of limitations governed the action because "the lex loci contractus rule determines the choice of law for interpretation of provisions of uninsured motorists clauses in automobile insurance policies just as it applies to other issues of automobile coverage." Id. at 295.

lex loci contractus rule dictates that the cause of action arises in the place where the contract is executed. Id. at 296 (citing Colhoun v. Greyhound Lines, Inc., 265 So. 2d 18, 21 (Fla.1972)).

Florida courts continue to apply the "most significant relationship" test to torts only, not contract actions. See, e.g., Allstate Ins. Co. v. Clohessy, 32 F. Supp. 2d 1328, 1333 (M.D.Fla.1998) (refusing to apply the "most significant relationship" test to a contract dispute). Although Plaintiff does not argue that the Kentucky statute of limitations should govern his suit solely because Decedent executed the insurance contract there, one should recognize in passing that Florida's rule is problematic because it offers no guidance in situations where the contract dispute involves the failure to execute a contract or when the parties executed the contract in a forum where the defendant is not amenable to process.

Most important for the purposes of the instant Erie problem, in Pledger v. Burnup & Sims, Inc., 432 So. 2d 1323 (Fla.Dist.Ct.App.1983), the Florida District Court of Appeals declined to construe the Bishop decision that applied the "most significant relationships" to substantive conflict of laws determinations as also covering procedural matters like the statute of limitations determination implicated by Florida's borrowing statute. Id. at 1330. The Florida District Court of Appeals presented the issue this way:

Unquestionably, Florida has the most significant contacts with the issue presented.... Appellant has presented a question on which he has cited no authority, in Florida or elsewhere: that is, in view of the [Florida] Supreme Court's adoption of Restatement (Second) Conflict of Laws § 145, did the cause of action arise in New York, or in Florida, which has the most significant relationship?

Id. The court found New York's limitations period would apply because the Florida legislature, not the judiciary, should make the decision that the "most significant relationship" test should apply to Florida's borrowing statute. Id. The Pledger court explained:

The Legislature has not amended [the borrowing statute] since the [Florida] Supreme Court opinion in Bishop. We have no hint of the legislative will on the question of determining where a cause of action arises. They may have assumed the Bishop conflicts rule would determine where a cause of action arises. They might equally have assumed that Bishop, a case dealing exclusively with the rights and liabilities of the parties, had nothing to do with the procedural borrowing statute.

Id. at 1330-31 (emphasis added). Thus, Florida does not use the test Plaintiff advocates and one Florida court emphasized that modifying a borrowing statute is a legislative prerogative. A federal court forced to make an Erie guess should pause before reinterpreting a state statute if the state law is one that is better modified legislatively. Whether or not the Kentucky courts could reinterpret Kentucky's borrowing statute instead of the Kentucky legislature despite the separation of powers concerns expressed in Pledger, the reinterpretive process raises much more problematic federalism concerns when a federal court construes a state statute.

Although Plaintiff's citations to Wyoming, New York, Missouri, Illinois and Florida law are somewhat illuminating, none of this information makes clear that the Kentucky Supreme Court would interpret Kentucky's borrowing statute as Plaintiff suggests.

As cited above, Plaintiff references various scholarly works that urge courts to construe borrowing statutes in a manner that incorporates the "most significant relationship" test. Despite these academic exhortations, we recognize that borrowing statutes interpreted in accordance with the accrual approach produce several meaningful policy advantages.10

First, borrowing statutes impede forum shopping. As one federal court sitting in diversity explained, the objective of Pennsylvania's borrowing statute is "simply to insure that a plaintiff who sues in Pennsylvania obtains no greater rights than those given in the state where his cause of action arose." Wilt v. Smack, 147 F. Supp. 700, 704 (E.D.Pa.1957); see also Stuart v. Am. Cyanamid Co., 158 F.3d 622, 627 (2d Cir.1998) ("[T]he purpose of the borrowing statute — preventing forum shopping by plaintiffs seeking the holy grail of the longer period — is best served by applying the period of the foreign state, regardless of how it is denominated."); Faigin v. Doubleday Dell Publ'g Group, Inc., 98 F.3d 268, 271 (7th Cir.1996) (observing "the antipathy to forum shopping reflected in Wisconsin's borrowing statute (in all such statutes, really)"); Siegemund v. Shapland, 247 F.Supp.2d. 1, 6 n. 17 (D.Me.2003) ("Borrowing statutes, enacted by states to prevent forum shopping, `enable the forum to borrow and use the statute of limitations of another state in determining the timeliness of an action.'") (quoting Hossler v. Barry, 403 A.2d 762, 765 (Me.1979)); Flowers v. Carville, 112 F. Supp. 2d 1202, 1208 (D.Nev.2000) ("Borrowing statutes are designed to prevent forum shopping.").

An accrual-based approach makes forum shopping impossible because the statute of limitations that governs in the forum where the cause of action accrued will apply no matter where the plaintiff files suit. In contrast, allowing plaintiffs to file suit in any state that has significant contacts with the dispute encourages plaintiffs to shop for the forum with the longest statute of limitations. Although limiting the operation of a borrowing statute to only those instances where a state has the most significant connection with the lawsuit helps make forum shopping somewhat more difficult, many modern commercial transactions transcend state boundaries. If a contract is negotiated in Michigan between an Ohio company and a Kentucky company, with performance scheduled in Tennessee, determining which state has the "most" significant contacts becomes difficult and subjective enough for enterprising attorneys to capitalize on differences between limitations periods. State courts' desire to discourage forum shopping is significant enough that in at least one instance, the New Jersey Supreme Court applied a foreign jurisdiction's limitation period even though New Jersey does not have a borrowing statute. See Heavner v. Uniroyal, Inc. 63 N.J. 130, 305 A.2d 412, 418 (1973) (applying New Hampshire statute of limitation despite lack of borrowing statute in New Jersey, in part to prevent forum shopping).

Second, strictly enforcing borrowing statutes best serves the purpose of statutes of limitation and repose.11 Most states, including Kentucky, have tolling statutes that stall the running of the applicable limitations period if a cause of action accrues against a state resident not present in the jurisdiction when the cause of action accrues. Kentucky law explains that "[i]f, at the time any cause of action mentioned in Ky.Rev.Stat. 413.090 to 413.160 accrues against a resident of this state, he is absent from it, the period limited for the commencement of the action against him shall be computed from the time of his return to this state." KY.REV.STAT. § 413.190(1). Furthermore,

When a cause of action mentioned in KRS 413.090 to 413.160 accrues against a resident of this state, and he by absconding or concealing himself or by any other indirect means obstructs the prosecution of the action, the time of the continuance of the absence from the state or obstruction shall not be computed as any part of the period within which the action shall be commenced.

Id. § 413.190(2). Since tolling statutes eliminate a defendant's ability to flee the jurisdiction and return after the statute of limitations expires, tolling statutes can inadvertently create perpetual liability for a defendant legally residing in Kentucky accused of a tort or breach of contract that occurred in another jurisdiction. If the defendant is not present in Kentucky when the cause of action accrues against him, he will face liability upon his return to the State even if he returns decades later. See id. The borrowing statute minimizes this problem by using the foreign state's statute of limitations when the cause of action accrues in the foreign state, thereby eliminating the problem of perpetual liability in those circumstances. See George v. Douglas Aircraft Co., 332 F.2d 73, 78 (2d Cir.1964) (discussing the problem of perpetual liability); Cvecich v. Giardino, 37 Cal. App. 2d 394, 99 P.2d 573, 574-75 (1940) (same).

Relieving potential defendants of uncertainty is one of the historical purposes of statutes of limitations. As the Supreme Court explained, "[s]tatutes of limitations, like the equitable doctrine of laches, in their conclusive effects are designed to promote justice by preventing surprises through the revival of claims that have been allowed to slumber until evidence is lost, memories have faded, and witnesses have disappeared." Order of R.R. Telegraphers v. Ry. Express Agency, 321 U.S. 342, 348-49, 64 S. Ct. 582, 88 L. Ed. 788 (1944). Earlier, the Supreme Court wrote that:

It is a well-settled principle that a statute of limitations is the law of the forum, and operates upon all who submit themselves to its jurisdiction.... Of late years, the courts, in England and in this country, have considered statutes of limitations more favorably than formerly. They rest upon sound policy, and tend to the peace and welfare of society. The courts do not now, unless compelled by the force of former decisions, give a strained construction to evade the effect of those statutes. By requiring those who complain of injuries to seek redress by action at law, within a reasonable time, a salutary vigilance is imposed, and an end is put to litigation.

McCluny v. Silliman, 28 U.S. (3 Pet.) 270, 276-77, 7 L. Ed. 676 (1830). Statutes of limitations and repose thus serve both the interest in encouraging expeditious dispute resolution and the interest in creating confidence that long bygone incidents will not suddenly reemerge in litigation initiated at some unknown future time. Interpreting Kentucky's borrowing statute in a manner more likely to create instances of lengthy or perpetual liability does a disservice to these objectives.

Third, borrowing statutes reflect respect for state sovereignty. A cause of action is an attempt by the plaintiff to vindicate what he believes is a legal right or obligation owed. Legal rights and obligations vest when the last event necessary to create the cause of action occurs. The law of the state where the final event occurs determines the parties' rights and responsibilities at that point because each state has sovereignty over that which occurs within in its territory. See, e.g., RESTATEMENT OF CONFLICT OF LAWS §§ 377-388. As discussed, the application of the law of the place of the injury is why this system is sometimes called the lex loci approach. Justice Story argued,

[E]very nation possesses an exclusive sovereignty and jurisdiction within its own territory.... The direct consequence of this rule is, that the laws of every state affect, and bind directly all property, whether real or personal, within its territory; and all persons, who are resident within it ...; and also all contracts made, and acts done within it.

JOSEPH STORY, COMMENTARIES ON THE CONFLICT OF LAWS, FOREIGN AND DOMESTIC 18 (2d ed. 1841) (hereinafter "COMMENTARIES"); see also Renfroe v. Eli Lilly & Co., 686 F.2d 642, 647 (8th Cir.1982) ("[T]he cause arises where as well as when the final significant event that is essential to a suable claim occurs.") (citing Mack Trucks, Inc. v. Bendix-Westinghouse Auto. Air Brake Co., 372 F.2d 18, 20 (3d Cir.1966)). The notion of vested rights, therefore, depends on ideas of territoriality and sovereignty.

Borrowing statutes embody the idea that the foreign jurisdiction's law should control out of respect for that jurisdiction's territorial sovereignty. See, e.g., Devine v. Rook, 314 S.W.2d 932, 935 (Mo.Ct.App.1958) ("It is fundamental that the law of the place where the cause of action first came to life controls the substantive law of the cause, since the cause owes its existence, and the character of its existence, to that place."). States, in turn, should respect each others' sovereign rights. In his classic explanation of the comity argument, Justice Story wrote:

The true foundation of which the administration of international law must rest is that the rules which are to govern are those which arise from mutual interest and utility, from a sense of the inconveniences which would result from a contrary doctrine, and from a spirit of moral necessity to do justice, in order that justice may be done to us in return.

COMMENTARIES, at 34. If Kentucky fails to respect that a cause of action accrues in a foreign jurisdiction, like New York, although the final event necessary for the cause of action occurred in New York, Kentucky shows disrespect for New York's territoriality in derogation of comity principles that the Kentucky Supreme Court may value.

Without a borrowing statute, New York could establish a cause of action to remedy a particular wrong, but reach a legislative judgment about the significance of the problem and the importance of finality, and choose to attach a short statute of limitations to the new cause of action. By failing to recognize New York's decision, Kentucky would effectively undermine a quasi-substantive component of New York law — not something we should lightly assume the Kentucky Supreme Court would choose to do.

Thus, despite Plaintiff's attempt to capitalize on the scholarly pressure to interpret Kentucky's borrowing statute as ineffective when Kentucky does not have the "most significant relationship" with the dispute, adopting Plaintiff's approach might encourage forum shopping, work against the purposes of statutes of limitations and repose, and demonstrate an inappropriate disregard for another state's sovereignty.

In a diversity case, the Kentucky statute of limitations will be applied as interpreted by this Court. Atkins v. Schmutz Mfg. Co., 372 F.2d 762, 763 (6th Cir.1967). Kentucky would not apply a "most significant relationship" analysis when applying Kentucky's borrowing statute to the instant case.

The statutory language supports Defendant's position. The "fundamental rule of statutory construction" in Kentucky "is to determine the intent of the legislature, considering the evil the law was intended to remedy." Beach v. Commonwealth, 927 S.W.2d 826, 828 (Ky.1996); see also Kelly v. Marr, 299 Ky. 447, 185 S.W.2d 945, 949 (1945). To ascertain legislative intent, Kentucky courts must refer to "the words used in enacting the statute rather than surmising what may have been intended but was not expressed." Flying J Travel Plaza v. Commonwealth, 928 S.W.2d 344, 347 (1996) (citing Ky. Assoc. of Chiropractors, Inc. v. Jefferson County Med. Soc., 549 S.W.2d 817 (Ky.1977)). To reiterate, Kentucky's borrowing statute states:

When a cause of action has arisen in another state or country, and by the laws of this state or country where the cause of action accrued the time for the commencement of an action thereon is limited to a shorter period of time than the period of limitation prescribed by the laws of this state for a like cause of action, then said action shall be barred in this state at the expiration of said shorter period.

KY.REV.STAT. § 413.320. Nowhere in the statutory language does the "most significant relationship" test appear. We are "not at liberty to add or subtract from the legislative enactment nor discover meaning not reasonably ascertainable from the language used." Commonwealth v. Gaitherwright, 70 S.W.3d 411, 413 (Ky.2002) (citing Commonwealth v. Frodge, 962 S.W.2d 864, 866 (Ky.1998)). Nevertheless, we certainly acknowledge Plaintiff's argument that other state supreme courts have read elements of a "most significant relationship" test into borrowing statutes that lacked such express language.

Kentucky last amended its borrowing statute in 1942, at a time when Kentucky still followed the vested rights approach to substantive choice of law questions. See KY.REV.STAT. § 413.320. Prior to 1967, all Kentucky case law applied lex loci. See, e.g., Ansback v. Greenberg, 256 S.W.2d 1 (Ky.1952) (applying lex loci to an automobile accident case). In a series of cases in the late 1960s, however, Kentucky began to depart from the lex loci approach to torts. See Foster v. Leggett, 484 S.W.2d 827, 828-29 (Ky.1972); Arnett v. Thompson, 433 S.W.2d 109, 112 (Ky.1968); Story v. Burgess, 420 S.W.2d 548, 548 (Ky.1967); Wessling v. Paris, 417 S.W.2d 259, 260-61 (Ky.1967). The Kentucky Supreme Court, however, did not make the same change with respect to contract cases.

In Lewis v. American Family Ins. Group, 555 S.W.2d 579 (Ky.1977), an Indiana resident driver insured in Indiana was involved in an accident in Kentucky with an uninsured Kentucky driver. Id. at 580. The Indiana driver filed suit in Kentucky against his insurer for uninsured motorist coverage and the court applied Indiana substantive law. Id. at 581. Under Lewis, the rights vested in the place of insurance, which was Indiana, not Kentucky, where the accident occurred. More recently, in State Farm Mutual Automobile Insurance Co. v. Tennessee Farmers Mutual Insurance Co., 785 S.W.2d 520 (Ky.Ct.App.1990), the Kentucky Court of Appeals refused to apply Kentucky substantive law to a Tennessee insurance contract following an automobile accident in Kentucky between a Kentucky driver and a Tennessee driver. Id. at 521. As long as Kentucky courts continue to apply the vested rights approach to contracts, and insurance contracts in particular, this Court, sitting in diversity, lacks a basis to predict that the Kentucky Supreme Court would take a different route.

Willits v. Peabody Coal Co., 188 F.3d 510, 1999 WL 701916 (6th Cir. Sept.1, 1999) (unpublished), is the only case to deal with questions like those Plaintiff presents in the context of Kentucky's borrowing statute. Although Willits is an unpublished opinion, its reasoning is instructive.

In Willits, the plaintiffs asserted a breach of contract claim alleging, inter alia, improper calculation of royalties under coal mining agreements. 1999 WL 701916, at *11. The parties originally executed the contract in Kansas. Id. at *13. The defendants calculated royalty payments in Missouri and then mailed them to various states, including Kentucky. Id. Missouri's shorter limitations period would have barred the plaintiffs' claims. Id. at *12. Both parties agreed that Kentucky's borrowing statute controlled the case. Id. at *11. The Willits Court explained how it would apply the borrowing statute:

The Kentucky borrowing statute requires a three step analysis: (1) we must determine whether the cause of action accrued in another state; (2) if the cause of action did accrue in another state, we must determine whether that state's statute of limitations for the particular cause of action is shorter than Kentucky's; (3) if the accrual state's statute of limitations is shorter than Kentucky's, we apply the statute of limitations of the accrual state; however, if the statute of limitations for the cause of action in that state is longer than Kentucky's, we apply Kentucky's shorter statute.

Id. at *12. The second and third parts of this process are fairly simple: if the cause of action accrued in a foreign state, then determine that state's statute of limitations, and if the foreign state has a shorter limitations period than Kentucky, borrow the foreign state's shorter rule. The trouble occurs in reaching the second and third components of the analysis because, as Willits correctly observed, "there is little or no law in Kentucky concerning where a breach of contract action accrues." Id. (emphasis in original).

To help ascertain how the Kentucky judiciary might determine where a cause of action accrues, Willits noted that Kentucky's statute of limitations for contracts of sale under the Uniform Commercial Code (UCC) provides: "`A cause of action accrues when the breach occurs, regardless of the aggrieved party's lack of knowledge of the breach.'" Willits, 1999 WL 701916, at *13 (quoting KY.REV.STAT. § 355.2-725(2)). This supports the conclusion, relevant to Plaintiff's lawsuit, "that a breach can occur before the aggrieved party actually knows of it and is damaged by it." Id. (emphasis added). This supports Willits decision that the breach occurred in Missouri, "where [the defendant] improperly calculated the royalties and from whence [the defendant] mailed out the royalty payments to the Plaintiffs in their various states of residence."12 Id. at *14.

Plaintiff offers three challenges to the reasoning offered in Willits and to the application of Willits to the matter presently before us. First, Plaintiff notes that the district court applied Willits to the instant action in part because of Wisconsin jurisprudence the district court found useful. Wisconsin has a borrowing statute similar to Kentucky's, but Plaintiff complains that the Wisconsin cases cited by the district court do not clearly establish what the Kentucky Supreme Court would do. The district court initially cites Abraham v. General Cas. Co. of Wisconsin, 217 Wis. 2d 294, 576 N.W.2d 46 (1998). In Abraham, the Wisconsin Supreme Court held that a cause of action is "foreign" when "the final significant event giving rise to a suable claim occurs outside the state of Wisconsin." Id. at 53-54. Plaintiff disputes Abraham's persuasive value because Abraham involved a set of facts in which all significant events occurred in Wisconsin. See id. at 54-55 (Bradley, J., concurring). Yet even assuming the Abraham decision depended on significantly different facts than the case presented here, Abraham still reinforces the vested rights theory that a cause of action accrues wherever the "last act" necessary to create the claim occurs.

The district court also relied on Ristow v. Threadneedle Ins. Co., 583 N.W.2d 452 (Wis.Ct.App.1998). In Ristow, a Wisconsin resident sued a South Carolina insurer because he never received a settlement check after suffering an injury in South Carolina. Id. at 453. Significantly, Ristow found that the final event creating the cause of action was the defendant's failure to issue the check from its South Carolina office, not the plaintiff's failure to receive the check in Wisconsin. Id. at 455. Citing Western Union Tel. Co. v. Lacer, 122 Ky. 839, 93 S.W. 34 (1906), Plaintiff claims the district court should not have relied on Ristow because Kentucky courts would have reached a different conclusion if faced with the Ristow facts. The Lacer court had to decide whether a cause of action against Western Union accrued in Indiana, where the company affixed the wrong address to a telegraph, or in Kentucky, where the company failed to promptly deliver the message. Id. at 35. Although Lacer found that the cause of action accrued in Kentucky, that conclusion does not aide Plaintiff much because the Lacer plaintiff expressly contracted for Western Union to "expeditiously deliver the correct message to the addressee at the point addressed." Id. at 35. Neither Ristow nor Plaintiff's case involves a contract that specified a delivery point. More important, Lacer requires a different "last act" analysis than Ristow because the Lacer Plaintiff could not have sued Western Union until the company failed to promptly deliver the message, whereas the Ristow plaintiff could have sued the defendant insurer whenever (and however) he discovered that the defendant did not mail him his settlement check.

Plaintiff's second direct challenge to the application of Willits to the instant problem depends upon Kentucky venue jurisprudence. Plaintiff cites a series of cases interpreting Kentucky's venue statute, Ky.Rev.Stat. Ann. § 452.450. See, e.g., T.C. Young Constr. Co. v. Hartford Accident & Indem. Co., 441 S.W.2d 781 (Ky.1969); Prudential Ins. Co. of Am. v. Terry, 265 Ky. 91, 95 S.W.2d 1109 (1936); Ins. Co. of N. Am. v. Hopper, 253 Ky. 402, 69 S.W.2d 728 (1934); Torrent Lodge v. Nat'l Sur. Co., 231 Ky. 302, 21 S.W.2d 439 (1929). These cases, however, merely address the question of venue for a suit based on an insurance policy, not where a cause of action accrues under Kentucky's borrowing statute. Under Kentucky's venue statute, plaintiffs must bring lawsuits for breach of contract against a corporation in the county where either (1) the corporation is present; (2) its agent resides; (3) the contract was made; or (4) the parties were to perform the contract. KY.REV.STAT. § 452-450. The planned place of performance is a possibly appropriate venue, but that says little about where a cause of action accrues pursuant to the borrowing statute.13

Third, in an attempt to minimize the importance of Willits as applied to the facts of his case, Plaintiff presents us with this hypothetical:

Buyer buys a used car from Seller using a contract that calls for a full-sized spare and shortens the statute of limitations to one year (as allowed by [Ky.Rev.Stat.] 355.2-725). The contract also provides a one year warranty for all electrical components, which requires Buyer to submit a claim form for any repair. Eleven months later, the alternator fails, and Buyer submits a warranty claim form. Two months after that, Buyer has a flat and discovers that there is no spare at all. Three months after the flat, the warranty company delivers a letter claiming that the alternator is not an electrical component.

Buyer's breach action for the lack of a spare fails because the breach occurred at delivery, as Seller could have put the spare in at any time until then. Buyer could have discovered the spare within the year, but did not and is barred. By not requiring actual knowledge, the statute [of limitations] places the burden on Buyer to examine the goods on delivery, and limitations begin at the point he should have known of the breach. This concept of accrual when the plaintiff should have known of the injury is found both in other limitations statutes and case law. See [KY.REV.STAT. ANN.] 413.140(2); Ky. Title Trust Co. v. Weil, 281 Ky. 763, 136 S.W.2d 1097 (1939).

Buyer's action for breach of the warranty, however, does not accrue under the statute until he receives the letter denying his claim, sixteen months after the purchase, and five months after the alternator failed. Buyer could not sue the warranty company when the alternator failed because he had not complied with the condition precedent of submitting the claim form at that point. The breach is only discovered when the Buyer receives the letter denying the claim and, under the statute [of limitations], the cause accrues at that point in time and (under the analysis in Willits) space.

(Pl.'s Br. at 18-19) (emphasis in original). This hypothetical is somewhat deceptive.

Plaintiff implies that the warranty in this hypothetical is like the D & O policy. Presumably, Plaintiff intends the hypothetical to demonstrate the difference between the breach of contract action based on the spare tire (which the hypothetical concedes the statute of limitations would bar) and the breach of warranty action based on the failed alternator. Plaintiff's analysis of the spare tire issue is correct because Buyer could have discovered the breach at any time after the sale and brought suit accordingly. Of course, Plaintiff could (but does not) explain it this way: as soon as Seller sold the car without the spare tire, the breach of contract claim "accrued" because selling the car without the spare was the final act necessary to create the cause of action.

Plaintiff wants us to confuse the "final act" notion of accrual with a discovery concept more favorable to his position in this case. In the hypothetical, Plaintiff assumes that Buyer had no warranty action until he received the letter from the warranty company declaring that an alternator does not qualify as an electrical component. Only then, argues Plaintiff, did Buyer discover his cause of action. Yet the final act necessary to create the cause of action was the warranty company's decision to deny the claim — a judgment that most likely transpired where the company posted the letter, not where Buyer received it. Quoting one of this Court's earlier decisions, Willits notes that "[t]he final act which transforms the liability into a cause of action necessarily has aspects of time and place. It occurs at a certain time and in a certain geographical spot." 1999 WL 701916, at *12 (emphasis added) (quoting Helmers v. Anderson, 156 F.2d 47, 51 (6th Cir.1946), aff'd sub nom. Cope v. Anderson, 331 U.S. 461, 67 S. Ct. 1340, 91 L. Ed. 1602 (1947)). As discussed above, Willits concluded "that a breach can occur before the aggrieved party actually knows of it and is damaged by it." 1999 WL 701916, at *13.

Although Willits is an unpublished opinion that constitutes only persuasive authority, Willits supports Defendant, not Plaintiff.

As already discussed, the judicial activity in other states does not constitute particularly persuasive evidence that the Kentucky Supreme Court would reinterpret Kentucky's borrowing statute to limit its function when Kentucky has the "most significant relationship" to the disputed transaction. Although a few state judiciaries have interpreted their borrowing statutes somewhat progressively, Plaintiff's references to the law in other jurisdictions are too equivocal to form a sound basis for an Erie guess.

To summarize, Wyoming has not adopted the Second Restatement's approach. See BHP, 1 P.3d at 1256. New York's "place of injury" test is different from the "most significant relationship" test and not particularly helpful when the injury is intangible, as in this case. See Global Financial, 693 N.Y.S.2d 479, 715 N.E.2d at 486. Plaintiff also points to Missouri, but Missouri has twice rejected invitations to apply the Second Restatement's approach to its borrowing statute. See Dorris, 725 S.W.2d at 872; Trzecki, 532 S.W.2d at 211. And despite Plaintiff's attempt to use Florida law to his advantage, the Florida Court of Appeals declined to adopt the Second Restatement approach without a clear legislative instruction. See Pledger, 432 So.2d at 1330-31. That leaves Illinois, which applied the "most significant relationship" test in Ehlco II, although that case's unusual procedural posture lessens its precedential value.

There is also relevant information from jurisdictions Plaintiff neglects to mention. Minnesota repealed its borrowing statute, but a federal court sitting in diversity dutifully applied Minnesota's borrowing statute before the Minnesota legislature revoked it. See Devine v. Rayette-Faberge, Inc., 285 F. Supp. 1006, 1008 (D.Minn.1968) (finding plaintiff's action time-barred by using Minnesota's borrowing statute to apply New York's more restrictive statute of limitations). More important, in Mack Trucks, Inc. v. Bendix-Westinghouse Auto. Air Brake Co., 372 F.2d 18 (3d Cir.1966), the Third Circuit expressly declined a plaintiff's invitation to interpret Pennsylvania's borrowing statute so as to apply "the law of the place having the most significant contacts with the relevant transactions and with the parties rather than by the law of the place of the wrong."14 Id. at 20-21.

The opinions from other jurisdictions that Plaintiff cites do not make a compelling case upon which we can make a judicious Erie guess. If anything, developments elsewhere tend to affirmatively indicate that the Kentucky Supreme Court would not limit the application of its borrowing statute to those situations in which Kentucky has the "most significant relationship" with the dispute.

Although the Kentucky Supreme Court might heed the scholarly criticism of borrowing statutes interpreted in accordance with the accrual theory, the Kentucky Supreme Court might just as easily choose to respect the strong policy interests borrowing statutes serve: they impede forum shopping, effectuate the goals of statutes of limitations and repose, and demonstrate respect for other states' sovereignty.

This Court has a circumspect role when sitting in diversity and should not unduly speculate about what a state court might conclude, nor should we impose whatever rule we think best. Given (1) the borrowing statute's language; (2) Kentucky's focus on vested rights in interstate contract litigation; (3) our unpublished Willits decision; (4) Plaintiff's inability to construct a strong argument based on the law of other jurisdictions; and (5) the policy interests supported by leaving Kentucky's borrowing statute unadulterated by the Second Restatement, we conclude that the Kentucky Supreme Court would apply the borrowing statute by focusing only on where the cause of action accrued, not on which state has the greatest interest in the dispute.

Having decided that the Kentucky Supreme Court would determine the applicable statute of limitations by looking only to where the cause of action accrued, we must now attempt to apply that rule to the facts of the case before us.

Where an insurance policy is breached depends in part on whether the contract is a liability policy or an indemnity policy. "[U]nder a liability policy a cause of action accrues when liability attaches, whereas under an indemnification policy there is no cause of action until the liability has been discharged, as by payment of the judgment by the insured." Quinlan v. Liberty Bank & Trust Co., 575 So. 2d 336, 355 (La.1990).

Plaintiff's D & O policy defines "loss" as "any amount which Insureds are legally obligated to pay for a claim or claims against them for Wrongful Acts." (J.A. at 16.) As the Eighth Circuit described a similarly-worded D & O policy issued by American Casualty Company:

The definition of "Loss" unequivocally includes defense costs, and seems to require American Casualty to pay these costs when the insureds are "legally obligated to pay" them, i.e., when the costs are incurred, and not, as American Casualty appears to suggest, when the lawsuit is finally disposed of and American Casualty has determined that the claims were covered. This is clearly a liability policy, despite American Casualty's attempt to treat it as an indemnity policy. The significance of this distinction is that under a liability policy, whenever a covered "loss" occurs (i.e., the insureds are "legally obligated to pay"), American Casualty must pay that amount-payment by the insurer is not conditioned upon a previous payment by the insured, as in an indemnity policy.

McCuen v. Am. Cas. Co., 946 F.2d 1401, 1406-07 (8th Cir.1991); see also Little v. MGIC Indem. Corp., 836 F.2d 789, 793 (3d Cir.1987) ("A `loss' is defined as an amount that the insured is `legally obligated to pay.' Although this section [of the policy] does not explicitly speak to the timing of the insurer's duty to pay, the only reasonable interpretation is that the duty arises at the time the insured becomes `legally obligated to pay.'"); Okada v. MGIC Indem. Corp., 823 F.2d 276, 280 (9th Cir.1986) ("In a liability contract, the insurer agrees to cover liability for damages. If the insured is liable, the insurance company must pay the damages. In an indemnity contract, by contrast, the insurer agrees to reimburse expenses to the insurer that the insured is liable to pay and has paid.") (quoting Cont'l Oil Co. v. Bonanza Corp., 677 F.2d 455, 459 (5th Cir.1982)). Plaintiff, therefore, held a liability policy, not an indemnity policy.

Assuming Defendant had an obligation to pay under the policy's language, Defendant had no responsibility before Decedent himself became "legally obligated to pay" attorney's fees, settlements, or other related sums. The two-page letter sent from Plaintiff's California counsel to Defendant is simply a list of the eight different lawsuits filed in the California Litigation. Renne, Plaintiff's California counsel, made clear in the text that "this letter constitutes a claim by [the officers and directors of Spendthrift] under the terms of the above-captioned policy and they will look to you [Defendant] for reimbursement of all costs and damages which may be assessed against them."15 (J.A. at 102.) Thus, at the moment Defendant received Renne's letter, Defendant had no obligation to reimburse Decedent — the letter did not include any demand for the reimbursement of a specific amount of money. Rather, it notified Defendant that the Spendthrift directors would attempt to recoup from Defendant "all costs and damages which may be assessed against them." Id.

Since Defendant renounced the D & O policy (with respect to the California Litigation) before Defendant had any obligation to make payments pursuant to the policy, Defendant's behavior constituted an anticipatory repudiation.16 As the Supreme Court explained, "[i]t has always been the law that where a party deliberately incapacitates himself or renders performance of his contract impossible, his act amounts to an injury to the other party, which gives the other party a cause of action for breach of contract." Roehm v. Horst, 178 U.S. 1, 18, 20 S. Ct. 780, 44 L. Ed. 953 (1900); see also Hochster v. De la Tour, 2 El. & Bl. 678 (1853). "The disclaimer of a contractual duty is a breach of contract even if the time specified in the contract for performing the duty has not yet arrived. It is what is called anticipatory breach." Wis. Power & Light Co. v. Century Indem. Co., 130 F.3d 787, 793 (7th Cir.1997); see also Quivirian Dev. Co. v. Poteet, 268 F.2d 433, 439 (8th Cir.1959) ("An anticipatory repudiation is declared to exist when a promisor, not otherwise in default, engages in making, without justification, a positive statement to the promisee, or other person having a right under the contract, indicating that the promisor will not or cannot substantially perform his contract duties."). Kentucky has recognized the doctrine of anticipatory breach. See, e.g., Jordon v. Nickell, 253 S.W.2d 237, 239 (1952); Fid. & Deposit Co. of Md. v. Brown, 230 Ky. 534, 20 S.W.2d 284, 286 (1929); Paducah Cooperage Co. v. Ark. Stave Co., 193 Ky. 774, 237 S.W. 412, 413 (1922).

The repudiation must be absolute or unequivocal. Thunder Basin Coal Co. v. Southwestern Pub. Serv. Co., 104 F.3d 1205, 1213 (10th Cir.1997); Sys. Council EM-3, Int'l Bhd. of Elec. Workers, AFL-CIO v. AT & T, 972 F. Supp. 21 (D.D.C.1997); Brownsboro Rd. Rest., Inc. v. Jerrico, Inc., 674 S.W.2d 40, 42 (Ky.Ct.App.1984) (citing 17 AM.JUR.2d Contracts § 448 (1964)). Courts determine whether anticipatory repudiation has occurred on a case-by-case basis, depending on the particular contract language involved. Truman L. Flatt & Sons Co., Inc. v. Schupf, 271 Ill.App.3d 983, 208 Ill. Dec. 630, 649 N.E.2d 990, 994 (1995). No precise set of words is necessary to constitute an unequivocal repudiation. Ewing v. Von Nieda, 76 F.2d 177, 182 (8th Cir.1935). Nevertheless, Defendant's reply to the Renne inquiry explained in detail why Defendant would not reimburse Plaintiff for expenses related to the California Litigation. The five-page letter made completely clear that Defendant "declines to afford coverage for any loss incurred by the [Plaintiff and Decedent] in connection with the [California Litigation]." (J.A. at 109.)

There is some disagreement as to whether the doctrine of anticipatory repudiation applies only to bilateral contracts, see, e.g., Saewitz v. Epstein, 6 F. Supp. 2d 151 (N.D.N.Y.1998) (citing 4 ARTHUR L. CORBIN, CORBIN ON CONTRACTS § 962 (1951)), or to both bilateral and unilateral contracts, see, e.g., Placid Oil Co. v. Humphrey, 244 F.2d 184 (5th Cir.1957). Insurance policies are generally unilateral contracts.17 A D & O insurance policy is a unilateral contract-the insured has already performed by paying the premium in exchange for the insurance company's promise to provide insurance.18 Nevertheless, there is ample authority for the proposition that the promising party can anticipatorily repudiate a unilateral contract. See, e.g., Dingley v. Oler, 117 U.S. 490, 502-04, 6 S. Ct. 850, 29 L. Ed. 984 (1886) (evaluating a claim that defendant had anticipatorily repudiated a unilateral contract without mentioning the contention that anticipatory repudiation might not apply to unilateral contracts); Xebec Dev. Partners, Ltd. v. Nat'l Union Fire Ins. Co., 12 Cal. App. 4th 501, 15 Cal. Rptr. 2d 726, 740 (1993) (finding that case law does not "establish that there can never be an anticipatory repudiation of an insurance contract"); Cameron v. Eynon, 332 Pa. 529, 3 A.2d 423, 423-25 (1939) (recognizing an anticipatory breach of a unilateral contract).

To limit anticipatory repudiation to bilateral contracts makes little sense. The theory, presumably, is that if the first party has already performed when the second party announces his intention to breach, the first party (the injured party) has nothing to gain through an immediate action for damages based on anticipatory breach. Yet "[t]he harm caused to the plaintiff is equally great [whether or not he has already performed]; and it seems strange to deny to a plaintiff a remedy [for anticipatory breach] merely on the ground that he has already fully performed as his contract required." 9 ARTHUR L. CORBIN, CORBIN ON CONTRACTS § 962, at 767 (interim ed.1979).

Furthermore, courts have frequently recognized anticipatory repudiation in insurance coverage disputes, and insurance policies are unilateral agreements. See, e.g., N.Y. Life Ins. Co. v. Viglas, 297 U.S. 672, 56 S. Ct. 615, 80 L. Ed. 971 (1936); Lovell v. St. Louis Mut. Life Ins. Co., 111 U.S. 264, 4 S. Ct. 390, 28 L. Ed. 423 (1884); Caminetti v. Manierre, 23 Cal. 2d 94, 142 P.2d 741 (1943); Scott v. Life & Cas. Ins. Co., 34 Ga.App. 479, 129 S.E. 903 (Ga.1925); Van Werden v. Equitable Life Assurance Soc'y of the United States, 99 Iowa 621, 68 N.W. 892 (Iowa 1896); Ebert v. Mut. Res. Fund L. Ass'n, 81 Minn. 116, 83 N.W. 506 (1900); McKee v. Phoenix Ins. Co., 28 Mo. 383 (1859); Garland v. Jefferson Standard Life Ins. Co., 179 N.C. 67, 101 S.E. 616 (1919); Fischer v. Hope Mut. Life Ins. Co., 69 N.Y. 161 (1877); Am. Ins. Union v. Woodard, 118 Okla. 248, 247 P. 398 (1926); Gaskill v. Pittsburgh Life & Trust Co., 261 Pa. 546, 104 A. 775 (1918); Mut. Reserve Fund Life Ass'n v. Taylor, 99 Va. 208, 37 S.E. 854 (1901); Merrick v. Northwestern Nat'l Life Ins. Co., 124 Wis. 221, 102 N.W. 593 (1905). The long list of cases involving the alleged anticipatory repudiation of an insurance contract includes disputes, like this one, that involve the alleged renunciation of a D & O policy. See, e.g., Fed. Sav. & Loan Ins. Corp. v. Oldenburg, 671 F. Supp. 720, 724 (D.Utah 1987) (finding insurer had anticipitorily repudiated a D & O insurance contract through letter that "was a definitive statement of an intent not to perform sent to the insured while there was a claim pending"); Xebec, 15 Cal. Rptr.2d at 740 (finding, in litigation over a directors and officers insurance policy, that case law does not "establish that there can never be an anticipatory repudiation of an insurance contract"). If a breach occurred, Defendant's response to the Renne letter breached the insurance contract by anticipatorily repudiating Defendant's alleged obligations. We must next ask where the breach (or the repudiation) occurred.

We acknowledge that the question of where an anticipatory repudiation occurs "ha[s] not yet been clearly settled." 9 ARTHUR L. CORBIN, CORBIN ON CONTRACTS § 988, at 851 (interim ed.1979). Nevertheless, there are a few cases involving disputes over where a repudiation occurred when the repudiating party sent the renunciation from one jurisdiction to the non-repudiating party in a second jurisdiction. The limited available jurisprudence strongly suggests that a repudiation occurs in the place where the repudiator posts the renunciation, not where the other party received notice. See Auglaize Box Bd. Co. v. Kansas City Fibre Box Co., 35 F.2d 822, 823 (6th Cir.1929) (holding action to have arisen in jurisdiction where repudiation occurred, not place of performance); Wester v. Casein Co. of Am., 206 N.Y. 506, 513-14, 100 N.E. 488 (1912) (concluding, when anticipatory repudiation was cabled from New York, addressed to and later received by the other party in South America, that the breach occurred in New York, because "[t]he breach of the contract occurred in this case upon delivery of the cablegram to the telegraph company, and constituted a breach, not only at that time, but also at that place."); Kumar v. Embassy Kosher Tours, Inc., 696 So. 2d 393, 394 (Fla.Dist. Ct.App.1997) ("[A]n anticipatory breach is deemed to `accrue,' if at all, in the county... from which the `breachor' transmits the repudiation, rather than where the `breachee' receives it."); Karson Indus., Inc. v. Superior Court, 273 Cal. App. 2d 7, 77 Cal. Rptr. 714, 716 (1969) (deciding, in a dispute where a letter cancelling a contract for mobile homes was posted in one county but received in another, "that the weight of authority and the better reasoned cases support the rule that repudiation by letter is a breach at the time when and the place where the letter is dispatched").

These cases ultimately derive their holdings from a series of nineteenth-century British and Irish decisions that reached similar conclusions. In Cherry v. Thompson, 7 Q.B. 573 (1872), for instance, the court found that the anticipatory breach of a contract to marry occurred in Germany when the defendant mailed a letter from Germany withdrawing from the engagement. As Judge Blackburn explained:

In this case the receipt of the letter by the plaintiff in England furnished him with evidence that the defendant had in Germany renounced [their] relation. Had his letters followed him to Ireland he would have received the evidence in Ireland. But the act which he had the option to treat as a breach took place in Germany, and in Germany alone.

Id. at 579 (emphasis added). Similarly, in Hamilton v. Barr, [1886] 18 L.R. Ir. 297, the plaintiff made a contract in Scotland with the defendant to act as the defendant's agent in Ireland. Later, the defendant mailed a letter to his agent in Scotland abrogating their agreement in advance. As Chancellor Naish explained, "I consider that the dismissal took place in Glasgow by the posting of the letter in Glasgow. The dropping of the letter in to the post office there was, I think, the act of dismissal." Id. at 301. Chancellor Porter agreed that the contract ended "the moment the notice of dismissal had left the hands and control of the defendant, which it did in Glasgow." Id. at 302.

In Holland v. Bennett, 1 K.B. 867 (1902), the proprietor of a periodical who resided in France employed a London correspondent. The proprietor discharged the plaintiff in a letter mailed from Italy. Justice Williams concluded that "[t]he effect is that there was a complete breach of the contract when the letter giving notice of dismissal was posted abroad." Id. at 869. Finally, Matthews v. Alexander, [1873] Ir. R.-C.L. 573, involved a wholesale tea dealer residing in London who hired a salesman and sent him to Ireland. Id. at 575. The salesman received a letter of dismissal before he could begin his work. Id. at 576. The message arrived in Dublin, but the letter was posted in London. Id. According to Judge Whiteside, "[t]he act in this case was done in England. The letter informs the Plaintiff that the Defendant in England had renounced the relation that had subsisted between them; and if that was a breach of contract, it took place in England." Id. at 579. This view is shared by the Restatement of Contracts, which explains that the repudiation occurs "as of the time when and the place where the letter or telegram is dispatched." RESTATEMENT OF CONTRACTS § 321. Williston also felt that repudiation by letter is a breach "as of the time when and the place where the letter or telegram is dispatched." 11 WILLISTON ON CONTRACTS § 1332, at 174 (3d ed.1968).

Thus, although the jurisprudence is fairly sparse, the available information strongly indicates that an anticipatory breach occurs where the breaching party posts its letter of renunciation.19 Since Defendant posted the message in New York, the breach occurred in New York. Under Kentucky's borrowing statute, Ky.Rev.Stat. § 413.320, the Kentucky Supreme Court would apply New York's statute of limitations, rendering Plaintiff's claim eight years late. Consequently, the district court properly ruled that under Ky.Rev. Stat. § 413.320, the New York statute of limitations bars Plaintiff's action.

For the aforementioned reasons, we AFFIRM the district court's decision.