Grant Thornton, LLP v. Federal Deposit Insurance

Affirmed in part and reversed and remanded in part by unpublished opinion. Judge AGEE wrote the opinion, in which Chief Judge TRAXLER and Judge MOTZ concurred. Unpublished opinions are not binding precedent in this circuit. AGEE, Circuit Judge: The Federal Deposit Insurance Corporation (“FDIC”), acting as receiver for the First National Bank of Keystone (“Keystone” or “the Bank”), sued Grant Thornton, LLP (“Grant Thornton”), a national accounting firm, for professional malpractice. Alleging that Grant Thornton negligently performed an audit of Keystone, 1 the FDIC sought to recover damages from the accounting firm after the FDIC closed Keystone as insolvent. After a lengthy bench trial, the district court awarded judgment in the FDIC’s favor in the initial amount of $25,080,777, which was reduced by a settlement credit to $28,737,026.43. On appeal, Grant Thornton does not challenge the district court’s finding that it was negligent in the conduct of the Keystone audit. Instead, Grant Thornton assigns error to the district court’s finding that its negligence was the proximate cause of certain of Keystone’s losses. Grant Thornton also challenges the district court’s refusal to allow some defenses and claims, which required imputing the actions of the Bank’s management to the FDIC. Finally, Grant Thornton claims the court erred in calculating a settlement credit based on the FDIC’s earlier settlement of various claims against Kutak Rock, LLP (“Kutak”), the Bank’s outside legal counsel. The FDIC has filed a cross-appeal, challenging the district court’s denial of an award of prejudgment interest. As explained in more detail below, we affirm the judgment of the district court as to all issues except the district court’s calculation of the settlement credit. As to that issue, we reverse and remand for further proceedings. We underscore, however, that the results in this case are driven by its unique facts, particularly the context of heavy regulatory oversight, known to Grant Thornton, as the sole reason for its engagement. Accordingly, it should be well understood we do not announce any new rule of auditor liability and none should be implied. I. A. Factual Background In Ellis v. Grant Thornton LLP, 530 F.3d 280 (4th Cir.2008), a prior appeal with different parties, we described Keystone’s background and how it came to engage Grant Thornton: Prior to 1992, Keystone was a small community bank providing banking services to clients located primarily in McDowell County, West Virginia. Before its collapse, Keystone was a nation *192 al banking association within the Federal Reserve System, the deposits of which were insured by the FDIC. In 1992, Keystone began to engage in an investment strategy that involved the securitization of high risk mortgage loans.... Keystone would acquire Federal Housing Authority or high loan to value real estate mortgage loans from around the United States, pool a group of these loans, and sell interests in the pool through underwriters to investors. The pooled loans were serviced by third-party loan servicers, including companies like Advanta and Compu-Link. Keystone retained residual interests (residuals) in each loan securitization [but the residuals] would receive payments only after all expenses were paid and all investors in each securitization pool were paid. Thus, Keystone stood to profit from a securitization only after everyone else was paid in full. The residuals were assigned a value that was carried on the books of Keystone as an asset. Over time, the residual valuations came to represent a significant portion of Keystone’s book value. From 1993 until 1998, when the last loan securitization was completed, the size and frequency of these transactions expanded from about $33 million to approximately $565 million for the last one in September 1998. All told, Keystone acquired and securitized over 120,000 loans with a total value in excess of $2.6 billion. [T]he securitization program proved highly unprofitable. Due to the risky nature of many of the underlying mortgage loans, the failure rate was excessive. As a result, the residual interests retained by Keystone proved highly speculative and, in actuality, they did not perform well. Keystone’s valuation of the residuals was greater than their market value. [Some members of Keystone’s management] and others concealed the failure of the securitizations by falsifying Keystone’s books. Bogus entries and documents hid the true financial condition of Keystone from the bank’s directors, shareholders, depositors, and federal regulators. Keystone’s irregular bank records drew the attention of the [Office of the Comptroller of the Currency (“OCC”) ], which began an investigation into Keystone’s banking activities. This investigation revealed major errors in Keystone’s accounting records that financially jeopardized Keystone. In May 1998, the OCC required Keystone to enter into an agreement obligating Keystone to take specific steps to improve its regulatory posture and financial condition. This agreement required Keystone to, among other things, retain a nationally recognized independent accounting firm “to perform an audit of the Bank’s mortgage banking operations and determine the appropriateness of the Bank’s accounting for purchased loans and all securitizations.” In August 1998, Keystone retained Grant Thornton as its outside auditor. Ellis, 530 F.3d at 283-84 (internal citation omitted). 2 B. Grant Thornton’s Audit of Keystone Since Grant Thornton does not challenge the district court’s finding that it was negligent in performing the Keystone audit, it is not necessary to fully discuss all the negligent acts found by the court. We *193 simply note the record fully supports the numerous factual findings by the district court regarding Grant Thornton’s negligence. In particular, the district court 'concluded two employees of Grant Thornton with primary responsibility for Keystone’s audit, Stan Quay and Susan Buenger, committed various violations of the Generally Accepted Auditing Standards (“GAAS”). See also Grant Thornton, LLP v. Office of the Comptroller of the Currency, 514 F.3d 1328, 1340-41 (D.C.Cir.2008) (Henderson, J., concurring) (discussing facts of instant case and describing Quay and Buenger’s conduct as “strikingly incompetent”). A crucial error was Buenger’s failure to obtain written confirmation of a purported oral representation from Advanta, one of Keystone’s loan servicers, that a substantial number of mortgages were properly documented on Keystone’s books of account. Buenger testified that she had a telephone conversation with Patricia Ramirez, who worked for Advanta, in which Ramirez told Buenger that she had located a pool of mortgages owned by Keystone worth approximately $236 million. But an e-mail from Ramirez minutes later, as well as an earlier written confirmation, showed that the loans were not owned by Keystone, but by “Investor Number 406,” identified as “United National Bank,” a separate banking entity. (J.A. at 1151.) While the district court expressly concluded that the oral statements Buenger attributed to Ramirez were not in fact made, it held that even if they had been, Buenger had an obligation under GAAS to obtain all “significant” confirmations of financial data in writing. Since the $236 million mortgage portfolio at issue constituted about one-fourth of the Bank’s claimed assets, it was clearly significant. Yet Buenger did not utilize the written statements from Advanta and instead chose to rely on the alleged oral misrepresentation, despite the fact that it conflicted with the written evidence and doing so was contrary to GAAS accounting procedure. Similarly, “Quay violated GAAS by failing to supervise or participate in the evaluation of the Advanta confirmation responses.” (J.A. at 798.) Because of this and other negligent acts, Buenger and Quay failed to find numerous problems with the Bank’s financial statements. Instead, Grant Thornton issued a clean audit opinion on April 19, 1999, stating that the audit had been conducted “in accordance with generally accepted auditing standards” and that the Bank’s financial statements were “free of material misstatement[s].” (J.A. at 1155.) In point of fact, however, those financial statements overstated Keystone’s assets by $515 million, making the Bank grossly insolvent. During an annual examination of Keystone a few months later, in August 1999, the OCC discovered the discrepancies and closed the Bank on September 1, 1999, appointing the FDIC as receiver. Of particular importance, the district court concluded that “[i]f Grant Thornton had exercised due professional care in connection with its audit, the fraud would have been discovered” and that “[i]f Grant Thornton had disclosed to Keystone’s board or the OCC the fact that Keystone was carrying over $400 million in loans on its books that were not owned by Keystone, the Bank would have been closed by April 21, 1999.” (J.A. at 800.) The court thus concluded that Grant Thornton’s negligence proximately caused damages in the amount of Keystone’s net operating expenses from April 21, 1999, two days after the audit report was released, until September 1, 1999, when the Bank was closed. In the context of the analysis of each argument raised on appeal, we discuss pertinent findings of fact that place each issue in perspective. Grant Thornton and the *194 FDIC have timely filed their respective appeal and cross-appeal. We have jurisdiction pursuant to 28 U.S.C. § 1291. II. Proximate Causation Grant Thornton makes three challenges to the district court’s finding that its negligence proximately caused the Bank’s post-audit operating losses. First, it argues that the district court applied the incorrect legal standard. Second, it contends that the district court’s finding of proximate cause was clearly erroneous. Third, it argues that the district court erred in refusing to consider that actions and knowledge of Keystone’s management subsequent to Grant Thornton’s audit were a superseding and intervening cause that cut off any of Grant Thornton’s liability for post-audit damages. We employ a “mixed standard of review” when judgment results from a bench trial. Universal Furniture Int’l, Inc. v. Collezione Europa USA, Inc., 618 F.3d 417, 427 (4th Cir.2010) (citation and quotation marks omitted). Specifically, we review the district court’s legal rulings de novo. Id.; see also Murray v. United States, 215 F.3d 460, 463 (4th Cir.2000) (legal conclusions regarding the correct standard of proof for proximate cause are reviewed de novo). We review the district court’s factual determinations for clear error. Universal Furniture Int’l, Inc., 618 F.3d at 427. Under clear error review, this Court must affirm factual findings if they are “plausible in light of the [entire] record,” “even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently.” Walton v. Johnson, 440 F.3d 160, 173 (4th Cir.2006). A. Legal Standard for Proximate Cause We find no merit in the contention that the district court failed to apply the proper legal standard for proximate cause. As the district court concluded and the parties agree, West Virginia law governs the common law claims of the FDIC here. See O’Melveny & Myers v. Fed. Deposit Ins. Corp., 512 U.S. 79, 89, 114 S.Ct. 2048, 129 L.Ed.2d 67 (1994) (“[w]hat sort of tort liability to impose on lawyers and accountants in general, and on lawyers and accountants who provide services to federally insured financial institutions in particular” did not warrant the “judicial creation of a federal rule of decision....”). State law thus governs the claims here. See Resolution Trust Corp. v. Everhart, 37 F.3d 151, 154-55 (4th Cir.1994) (relying on O’Melveny to hold that whether a federal receiver timely filed a claim against the former directors and officers of a failed financial institution was an issue controlled by state law). West Virginia law defines a proximate cause of an injury as one “which, in natural and continuous sequence, produces foreseeable injury and without which the injury would not have occurred.” Hudnall v. Mate Creek Trucking, Inc., 200 W.Va. 454, 490 S.E.2d 56, 61 (1997). Thus, the test requires both (1) “foreseeable injury”; and (2) but-for causation. Id. Grant Thornton contends that the district court relied solely on its negligence as a but-for cause of the damages awarded, but failed to consider whether the resulting injury to the Bank was in fact foreseeable. The record plainly rebuts this contention as the district court applied the Hudnall two-part test, and addressed at length the issue of foreseeability in a separate sub-heading under causation in its written opinion of March 14, 2007. The district court did not err in its application of the legal standard for proximate cause. B. Finding of Proximate Cause The district court found that it was foreseeable to a “reasonably prudent auditor” *195 that an auditor’s negligent failure to discover a Bank’s true losses and actual insolvency could result in a continuation of those losses. (J.A. at 839.) Grant Thornton’s negligence in failing to discover the fraud at Keystone allowed that fraud to continue, and the losses the FDIC seeks to recover are the foreseeable result of that ongoing fraudulent scheme. As Grant Thornton’s expert conceded, it is certainly foreseeable from the standpoint of a reasonably prudent auditor that the failure to discover fraud will result in the continuation of the fraud. (J.A. at 840.) Consequently, the district court found that the Bank’s post-audit net operating loss (operating expenses offset by operating income) for the period from April 21, 1999 (2 business days after the release of the audit) until September 1, 1999 (when Keystone was involuntarily closed by the OCC) were proximately caused by Grant Thornton’s negligence. Grant Thornton contends, however, that the Bank’s losses were not proximately caused by the audit, but were the result of the Bank’s longstanding “unprofitable securitizations and the imbalance between the Bank’s income and its interest obligations.” (Opening Br. for Grant Thornton at 24.) Because the specific facts of this case distinguish it from the typical case in which an audit is undertaken, we agree with the district court. We find it particularly significant in this case that Grant Thornton was hired to perform the audit, not in the ordinary course, but at the insistence of federal regulators who were closely watching Keystone. And Grant Thornton was well aware that factor was the reason behind its engagement. As the district court explained: “The unique position that Keystone was in at the time period in question—with federal regulators carefully watching the Bank’s actions and waiting for assurances from the outside auditor that the Bank’s financial statements were accurate—distinguish this case from any of the other cases relied upon by the parties .... ” (J.A. at 843.) A number of factual findings by the district court support its ultimate finding of proximate causation based on foreseeability. For example, there was evidence that: (1) OCC told Grant Thornton in December 1998 that Keystone had overstated its assets by about $90 million in three earlier quarterly reports; (2) by January 1999, both Buenger and Quay testified that the OCC informed them there was a “distinct possibility” that the bank would fail if the problems and weaknesses were not satisfactorily addressed and resolved, which Buenger interpreted as a “high probability” of failure (J.A. at 993-99); (3) Buenger admitted that, prior to the audit report being issued, Grant Thornton had characterized the audit as a “highest maximum risk” audit, its highest risk category (J.A. at 998); this risk category required certain additional steps and tests be conducted, some of which Buenger and Quay simply failed to perform; and (4) both Buenger and Quay “testified that their ‘fraud antenna’ were up as high as they could get.” (J.A. at 770.) These facts, among others, made it reasonably foreseeable to any prudent auditor that a failure to perform the audit with due care could result in the continued operation of a Bank that was in fact woefully insolvent and hemorrhaging losses. Additionally, as pointed out by the district court, the damages awarded were all natural and foreseeable losses as a result *196 of Keystone’s continued operations. Although Grant Thornton challenges, for example, the payment of interest on deposits received before the audit began, it is because of their recurring nature in the ordinary course of commerce that such expenses were particularly foreseeable. (See J.A. at 845 (“It was highly foreseeable that Keystone would continue to pay interest expense on deposits, dividends, legal fees, consulting fees, salaries, and other routine operating expenses.”).) Again, we see no clear error in that finding. Grant Thornton argues that affirming the finding of proximate cause in the case at bar “effectively makes the auditor an insurer for a bank’s future financial performance if it fails to recognize that the bank should close” and “would impose arbitrary and potentially breathtaking liability on auditors.” (Opening Br. for Grant Thornton at 19.) It also argues that affirmance will expose “auditors and others who serve federally-insured institutions to potentially limitless liability that is unbounded by ordinary principles of proximate causation and proportionate fault” and will “discourage prudent service providers from future dealings with federally-insured institutions — particularly those most in need of audit services.” (Id. at 18.) Again, we disagree based on the particular and unique facts of this case, primarily the specific context in which this audit occurred. Given this context, we conclude that the district court did not clearly err in finding both that the damages from the continued operation of the Bank would not have occurred but for Grant Thornton’s negligence and that they were a foreseeable result of Grant Thornton’s negligence. Cf. Hudnall, 490 S.E.2d at 61. Grant Thornton’s dire predictions of unlimited liability for auditors of insolvent banks also ignores the temporal scope of the district court’s damage determination here. (Cf. Opening Br. for Grant Thornton at 27 (referring to “crushing” and “breathtaking” liability).) Notably, the district court did not conclude that continued operating expenses for an unlimited period of time would have been foreseeable. Rather, the damages period here was for a reasonable — and foreseeable— period. 3 In particular, the district court concluded that, had the audit been performed properly instead of negligently, federal regulators would have closed the Bank two days after an accurate audit report had issued, or by April 21, 1999. The court limited the damages to those incurred during the period between this date and September 1, 1999, when the Bank actually closed. Cf. Thabault v. Chait, 541 F.3d 512, 519-20 (3d Cir.2008) (upholding jury verdict of almost $120 million as proximately caused by auditors’ negligent failure to discover insolvency of insurance company where the damages represented the net cost of continuing operations from the date of the audit to the date of liquidation, a period of more than nineteen months). We thus conclude that the district court’s finding of proximate cause was not in error as its determination was supported by the evidence before it and consistent with West Virginia law. C. Intervening and Superseding Cause Grant Thornton’s final challenge to the district court’s finding of proximate cause is that the actions of the Bank’s management, post audit, constituted a superseding *197 and intervening cause that extinguished Grant Thornton’s liability for damages. Grant Thornton points to evidence that Keystone’s executives were aware that the Bank was insolvent, but continued to recklessly operate Keystone and to hide its true financial condition. For example, Bank executives convinced the OCC examiners to allow the Bank to send confirmation requests to Advanta and Compu-Link, rather than the OCC sending those directly. The Bank’s management reworded the confirmation letters to its loan servicers so as to request information on loans owned not just by Keystone but also by United National Bank. Management then attempted to intercept the responses to assure the artifice was not detected. West Virginia’s Supreme Court of Appeals (‘WVSCA”) has explained the defense of an intervening cause as follows: The function of an intervening cause is that of severing the causal connection between the original improper action and the damages. Our law recognizes that an intervening cause, in order to relieve a person charged with negligence in connection with an injury, must be a negligent act, or omission, which constitutes a new effective cause and operates independently of any other act, making it and it only, the proximate cause of the injury. Sydenstricker v. Mohan, 217 W.Va. 552, 618 S.E.2d 561, 568 (2005) (internal citations, quotation marks and brackets omitted). The district court’s order contains a section titled “Intervening and Superseding Cause” in which it discusses the actions and knowledge of Keystone’s management after the audit. (J.A. at 849-873.) However, the district court analyzes the issue in terms of imputation and did not directly address the precise argument raised by Grant Thornton, which is that the actions of management need not be imputed to the FDIC to be a “superseding cause,” 4 but instead that the actions of a non-party may give rise to such a cause. (Opening Br. for Grant Thornton at 36 (citing Sydenstricker, 618 S.E.2d at 568 (the defense of intervening cause can be established based on evidence that shows “the negligence of another party or a nonparty”)).) We agree with the district court’s implicit holding that the continued effort of the Bank’s management post audit to conceal Keystone’s insolvency was not an intervening and superseding cause under West Virginia law. See Ross v. Commc’ns Satellite Corp., 759 F.2d 355, 363 (4th Cir.1985) (“An appellate court has power to determine independently whether summary judgment may be upheld on an alternative ground where the basis chosen by the district court proves erroneous.”), overruled on other grounds by Price Waterhouse v. Hopkins, 490 U.S. 228, 109 S.Ct. 1775, 104 L.Ed.2d 268 (1989). Our review of the West Virginia caselaw reveals the WVSCA would not find the post-audit acts of Keystone’s management a superseding or intervening cause. For example, in Yourtee v. Hubbard, 196 W.Va. 683, 474 S.E.2d 613 (1996), the plaintiffs decedent was killed while riding as a passenger in a stolen car, and the owner of the car was being sued because he had negligently left the car unattended with the keys in the car. Id. at 615. The WVSCA upheld the trial court’s finding that the theft of the ear by plaintiffs decedent and his friends and their subsequent acts (which included driving at high rates *198 of speeds in excess of ninety miles per hour, losing control of the vehicle, and running it into a brick wall) constituted an intervening cause that broke the chain of causation and relieved the car owner of liability. Id. at 615, 620-21. Similarly, in Harbaugh v. Coffinbarger, 209 W.Va. 57, 543 S.E.2d 338, 345-47 (2000), a decedent’s decision to play Russian roulette was an intervening cause rendering it the only proximate cause of the injury even though defendant had negligently supplied the loaded gun. Federal courts applying West Virginia law have reached similar results. See, e.g., Ashworth v. Albers Med., Inc., 410 F.Supp.2d 471, 479-81 (S.D.W.Va.2005) (concluding that a drug manufacturer was not liable for injuries caused by alleged criminal acts of third parties who introduced counterfeit versions of the manufacturer’s drug into the stream of commerce). In the foregoing cases, the two acts of negligence are unconnected and unrelated; the one could not be reasonably foreseen to be the result of the other. These cases reflect a superseding or intervening cause because the event in question was significantly independent from the initial negligence such that the separate acts of negligence had only a tangential relation to each other. By contrast, in the case at bar, the continued fraudulent conduct by the Bank’s management was not unforeseeable nor did it “operate independently” of the established fact of Grant Thornton’s negligent audit. Cf. Sydenstricker, 618 S.E.2d at 568. As noted in discussing proximate cause, we find it particularly significant that Grant Thornton was hired by Keystone— as a requirement of the Bank’s agreement with the OCC — in order to evaluate the Bank’s financial condition and that Grant Thornton knew regulators and management would rely on a clean audit report to allow the Bank to continue to operate. Thus, Bank’s management’s use of the defective audit report to continue to engage in fraudulent conduct and to stave off regulators was facilitated by Grant Thornton’s negligence. Indeed, but for the negligent audit report, the management conduct posited by Grant Thornton could not have happened. In this sense, the post-audit actions by Bank’s management are not a “new effective cause” and did not “operate! ] independently” of Grant Thornton’s negligence and thus do not constitute a superseding cause of the Bank’s damages. See Sydenstricker, 618 S.E.2d at 568; see also Wehner v. Weinstein, 191 W.Va. 149, 444 S.E.2d 27, 32-33 (1994). We thus find no error in the district court’s determination that Grant Thornton’s liability for its negligence was not absolved by a later intervening and superseding cause. III. Imputation Grant Thornton next contends that the district court erred by prohibiting it from offering certain claims or defenses, specifically (1) comparative/contributory negligence; (2) in pari delicto; and (3) “similar doctrines.” (Opening Br. for Grant Thornton at 20.) Relying in large part on the decision of the WVSCA in Cordial v. Ernst & Young, 199 W.Va. 119, 483 S.E.2d 248 (1996), the district court held that Grant Thornton was barred from asserting these or similar claims or defenses that involved the imputation of the knowledge or actions of Keystone’s management to the FDIC. Consequently, the court granted the FDIC’s motion to dismiss the affirmative defenses asserted by Grant Thornton, and dismissed related counts of Grant Thornton’s third party complaint. We review de novo a district court’s decision to strike a defendant’s affirmative defenses or dismiss a defendant’s counterclaims. Cf. Murray *199 v. United States, 215 F.3d 460, 463 (4th Cir.2000) (conclusions of law are reviewed de novo). As an initial matter, we note that state law controls what defenses are available against the FDIC when the agency is acting as the receiver of a failed financial institution. See supra at Section II.A (citing O’Melveny & Myers, 512 U.S. at 89, 114 S.Ct. 2048; and Resolution Trust Corp., 37 F.3d at 154-55). Accordingly, the FDIC simply “ ‘steps into the shoes’ of the failed” financial institution and is then subject to whatever defenses state law provides. O’Melveny, 512 U.S. at 86-87, 114 S.Ct. 2048 (citation omitted). Accordingly, West Virginia law governs the issue of whether the knowledge or conduct of the Bank’s management can be imputed to the FDIC here. Although the parties have not cited to any West Virginia cases directly addressing whether the actions of a bank’s management can be imputed to the FDIC as receiver, two WVSCA decisions offer guidance in predicting how that court would rule on this issue. Because those cases point to seemingly conflicting conclusions, we examine them in some detail. The first case, Wheeling Dollar Sav. & Trust Co. v. Hoffman, 127 W.Va. 777, 35 S.E.2d 84 (1945), involved an insurance company and a loan association, both of which had been placed in receivership. See id. at 86. The insurance-company receiver sued the loan-association receiver. As a defense, the loan-association receiver asserted fraud and the doctrine of unclean hands, predicated on facts showing that (1) the insurance company and loan association were run by the same secretary-treasurer, id. at 87, and (2) “the whole system of accounts between the[] corporations and the official reports made on their behalf seem[ed] to be permeated with deliberate fraud.” Id. at 88. Because the two company’s accounts were “created and preserved by the common manager,” the WVSCA held that the whole system of fraudulent accounts was “chargeable to the officers and directors of each [company], either through actual knowledge or the gross ignorance or neglect of their official duties, and, hence, to the corporations themselves.” Id. The WVSCA further held that the knowledge and/or negligence of the corporations was chargeable to their receivers, as “[t]he rights of the respective receivers rise no higher than those of the corporations which they represent.” Id. Accordingly, the WVSCA applied the doctrine of unclean hands and based on the “high probability of fraud in the whole subject of th[e] litigation ... refuse[d] [to] consider[ ] ... the plaintiffs bill.” Id. at 89. Wheeling Dollar thus favors Grant Thornton’s position in the present case. It is not, however, the WVSCA’s most recent holding regarding the defenses available against a government entity serving as the receiver of a failed financial institution. In Cordial v. Ernst & Young, 199 W.Va. 119, 483 S.E.2d 248 (1996), the accounting firm Ernst & Young was hired by both Blue Cross and Blue Shield (“BCBS”) and the West Virginia Commissioner of Insurance (“the Commissioner”) to perform external audits of BCBS’s accounts. See id. at 251-52. The Commissioner later placed BCBS into receivership, see id. at 255, and filed suit against Ernst & Young alleging various causes of action including negligence, breach of fiduciary duty, breach of contract, and fraud. Id. Citing Wheeling Dollar, Ernst & Young argued that the Commissioner acting as receiver could “assert only those claims that [BCBS] could itself have brought.” Id. at 256 & 257 n. 9. For several reasons, the WVSCA disagreed. First, the court *200 cited a prior case in which it recognized the Commissioner acting as “ ‘[Receiver is a government official charged with authority to protect not only the shareholders of the corporation, but also policyholders, creditors and the public.’ ” Id. at 257 (quoting Clark v. Milam, 192 W.Va. 398, 452 S.E.2d 714, 720 (1994)). In other words, “[r]ather than being deemed to solely represent the interests of the corporation, the Insurance Commissioner as [Receiver represents a broad array of interests, including those of the public.” Id. (quotation and emphasis omitted). Second, the WVSCA relied on the reasoning of the United States District Court for the Northern District of Illinois in a federal case involving a suit brought by a financial institution receiver “against accountants for an improper audit.” Id. (citing Resolution Trust Corp. v. KPMG Peat Marwick, 845 F.Supp. 621 (N.D.Ill.1994)). The WVSCA agreed with the Illinois district court that “ ‘[p]ublic policy concerns mandate a finding that the duty of FDIC to collect on assets of a failed institution runs to the public and not to the former officers and directors of the failed institution,’ ” and that “ ‘it is the public which is the intended beneficiary of FSLIC, just as it is the public which is the beneficiary of the common law duty imposed upon officers and directors to manage properly the institutions entrusted to their case.’” Id. (quoting KPMG Peat Marwick, 845 F.Supp. at 623) (additional citations omitted). Third, the WVSCA cited West Virginia’s “comprehensive scheme of insurance regulation” as evidence of a “broad public interest in the sound administration of insurance firms.” Id. The Commissioner, as receiver, thus “carries] out a duty that runs to the public in pursuing the claims of policyholders, creditors, shareholders or the public.” Id. (internal quotation marks omitted). Accordingly, the Commissioner “acts as the representative of interested parties, such as the defunct insurer, its policyholders, creditors, shareholders, and other affected members of the public” and “has standing ... to bring an action ... to vindicate the rights of such interested parties.” Id. Critically, at the end of this discussion, the WVSCA inserted a footnote addressing Ernst & Young’s claim under Wheeling Dollar that “[t]he rights of the respective receivers rise no higher than those of the corporations which they represent.” Id. at 257 n. 9 (quotation marks omitted). This contention formed part of Ernst & Young’s greater argument that because BCBS’s “managers indisputably knew what its financial condition was, ... they could not bring a valid claim against E & Y for failure to disclose such.” Id. The Cordial Court rejected this proposition on two grounds. First, the court explained “that Wheeling Dollar was decided prior to the adoption” of West Virginia’s comprehensive insurance regulatory scheme. Id. “Consequently, [the case was not] relevant to the issues at bar.” Id. “Moreover,” the court explained, “since [the] Commissioner, acting as receiver, is vindicating the lights of the public, including the Blue Cross creditors, policyholders, providers, members, and subscribers, [there was] no merit in this contention.” Id. (emphasis added). Footnote 9 of the Cordial opinion thus indicates that the WVSCA would not limit the Commissioner, as receiver, only to the rights of the represented corporation because he also serves to “vindicate] the rights of the public.” Id. It also suggests that Wheeling Dollar is no longer good law, at least in the context of government receiverships. Furthermore, in formulating the controlling public policy involved, the WVSCA heavily relied on — and *201 favorably referred to — KPMG Peat Mar-wick, which applied similar logic to the FDIC in a suit against auditors. We see no principled difference between the Commissioner’s role as receiver in Cordial and that of the FDIC in the case at bar. We therefore find no error in the district court’s decision not to allow Grant Thornton’s affirmative defenses of in pari delicto, comparative negligence, or other defenses or claims that relied on imputation of the Bank’s management to the FDIC. IV. Settlement Credit Based on the FDIC’s settlement of its claims against Kutak, the district court awarded a percentage of the settlement amount as a settlement credit to Grant Thornton. On appeal, Grant Thornton posits two challenges to the calculation of the credit. First, Grant Thornton contends it was entitled to a $22 million setoff based on a computation in the Kutak proceeding. In the alternative, Grant Thornton argues that the district court improperly calculated the settlement credit by basing it only on amounts actually received by the FDIC, and instead should have based the credit on the face value of the Kutak settlement. We address each argument in turn after some additional factual background. Before trial, Grant Thornton sought leave to file a contribution claim against the Bank’s outside counsel, Kutak. The district court held that the FDIC’s settlement with Kutak over Kutak’s liability in Keystone’s failure extinguished Grant Thornton’s contribution claim. However, the court noted that Grant Thornton might be entitled to a settlement credit to be resolved in a later proceeding. After trial, the district court delayed entry of judgment against Grant Thornton and held a hearing to determine the amount of any settlement credit. As described by the district court, the Kutak settlement agreement provided that Kutak’s primary insurer, Executive Risk Indemnity, Inc., would immediately pay the FDIC the remaining policy limits of $8 million. Kutak also signed a $4 million promissory note bearing 3% interest, to be paid in installments. The settlement agreement further provided that Kutak and the FDIC would cooperate in pursuing a $10 million claim on Kutak’s excess insurance policy with Reliance Insurance Company, which was in receivership. If the FDIC received less than $8 million from Reliance, then Kutak would make up a portion of the shortfall according to a formula set forth in a second promissory note. The maximum amount that Kutak would be required to pay under this second promissory note was $2,750,000. Before the district court, Grant Thornton argued it was entitled to a credit of $22 million, the full face amount of the Kutak settlement agreement ($8 million plus $4 million plus $10 million) because: (1) Kutak was jointly responsible for the operating losses for which Grant Thornton had been held liable and (2) the settlement agreement did not allocate the proceeds among joint and alleged non-joint claims. The FDIC took the initial position that Grant Thornton was due no settlement credit because the FDIC had planned to sue Kutak only for damages associated with Keystone’s securitizations, a matter for whieh Grant Thornton was not liable. The FDIC later admitted that there was some overlap between the damages against Kutak and against Grant Thornton, but argued that, if a settlement credit was due, the amount should be based on the amount actually recovered from Kutak, rather than the full face value of the agreed settlement amount. The district court ruled that: (1) Kutak was responsible for $292,899,625 in dam *202 ages to the FDIC, including the $25,080,777 of post-audit net operating losses for which it was jointly liable with Grant Thornton; and (2) in determining the credit due Grant Thornton, the FDIC/Kutak settlement should be allocated proportionally by dividing the $292,899,685 amount by the $25,080,777 figure, yielding a 8.563% settlement credit ratio. Put differently, the district court ruled that the overlapping damages — the indivisible loss — accounted for only 8.563% of the total damages caused by Kutak. Using this formula, the district court then calculated the settlement credit based upon the funds actually received by the FDIC from Kutak. These funds included what the FDIC was “guaranteed” to receive from the primary insurer and Kutak’s promissory notes (a total calculated at $15,692,521), rather than the higher stipulated settlement amount of $22 million. The settlement credit was thus determined to be $1,343,750.57 (8.563% of $15,692,521). The district court also held that Grant Thornton should receive an additional credit equal to 8.563% of any additional future payments made by Kutak to the FDIC under the settlement agreement. A. Grant Thornton first contends that West Virginia law requires a setoff equal to the full face amount in the settlement agreement between FDIC and Kutak, $22 million. The accounting firm argues that West Virginia law and Section 4 of the Uniform Contribution Among Tortfeasors Act (“UCATA”), which Grant Thornton contends has been adopted by court decision in West Virginia, require it. 5 Further, Grant Thornton makes several policy arguments as to why West Virginia’s rules governing partial settlements in multi-party cases “can be meaningfully applied only if the allocation is included in the settlement agreement.” (Opening Br. for Grant Thornton at 47.) Lastly, Grant Thornton argues that hearings on “allocation” of damages, such as occurred here, encourage protracted legal proceedings and are inherently unfair, since the settling party is not involved in the proceedings and the non-settling party is not in as good a position to contest the plaintiffs claims against the settling party. Grant Thornton thus argues that where a settlement (like that between the FDIC and Kutak) covers both joint and non-joint liabilities and does not allocate among joint and non-joint claims, “the nonsettling party is entitled to a credit equaling the entire settlement amount.” (Opening Br. for Grant Thornton at 49 (quoting Cohen v. Arthur Andersen LLP, 106 S.W.3d 304, 310 (Tex.App.2003)).) We agree with the district court that the determination of setoff is a complex question and that Grant Thornton’s simple solution of deducting the $22 million face settlement amount from the verdict against it, while “appealing for its *203 simplicity of application,” is “simple, neat, and wrong.” (J.A. at 958.) Under West Virginia law, the threshold question of whether or not Grant Thornton is entitled to any settlement credit is based on whether the loss is a single, indivisible loss. See Biro v. Fairmont Gen. Hosp., Inc., 184 W.Va. 458, 400 S.E.2d 893, 896 (1990) (“In order to permit a verdict reduction reflecting a prior settlement, Zando held that there must be a ‘single indivisible loss arising from the actions of multiple parties who have contributed to the loss.’ ”) (citation omitted). Where there is an indivisible injury, then a setoff is appropriate. See id. at 896. But if there are divisible injuries causing loss, then no set-off will be allowed. Cf. id. at 897 (concluding that an injury from negligently performed surgery was divisible from injuries from a fall in the hospital while recovering from that surgery and thus no offset was warranted). But in the case at bar we have a partial overlap of the damages contemplated in the FDIC’s settlement agreement with Kutak and the damages found to have been caused by Grant Thornton. That is, the total damages sought against Kutak by the FDIC included the full $25 million of the Bank’s post-audit net operating loss for which Grant Thornton was also found responsible. Thus, that overlapping portion of the damages related to the operating loss is indivisible, and a setoff is appropriate. Zando, 390 S.E.2d at 809 (“a single indivisible loss arising from the actions of multiple parties” entitles the non-settling defendant to a reduction). We agree with the district court, however, that “the FDIC’s claims against Kutak for the $25 million in damages for which Grant Thornton has been found liable are divisible” from the FDIC’s claim for damages against Kutak for the remaining losses to the Bank because “one person [Kutak] caused all of the damages and another person [Grant Thornton] caused only part of the damages.” (J.A. at 955.) See Restatement (Third) of Torts § 26 cmt. f (2000) (“[divisible damages can occur ... when one person caused all of the damages and another person caused only part of the damages.”). Simply giving Grant Thornton a credit equal to the $22 million FDIC settlement with Kutak requires an assumption that the entire amount of the Kutak settlement was meant to pay for net operating expenses after April 21, 1999, an assumption not supported by the record or logic. Indeed, as found by the district court, Kutak’s involvement was for a much longer period of time than Grant Thornton and resulted in damages not attributable to Grant Thornton at all (such as damages incurred as a result of the failed securitization programs). While the FDIC and Kutak did not allocate specific damages in the settlement agreement, to give Grant Thornton the full $22 million credit would not be in accord with the principles set out by the WVSCA in Zando, governing verdict credits for nonsettling defendants. 6 *204 In short, we find no error in the district court’s approach in allocating the damages between Kutak and Grant Thornton or in its conclusion that only a portion of those damages overlapped and were indivisible. Accordingly, we affirm the district court’s decision that Grant Thornton is not entitled to a full $22 million reduction, but only to a portion of that amount. As to the amount of damages attributed by the district court to Kutak, (ie., its arrival at the $292 million figure and the resulting 8.563% calculation), those findings are reviewed only for clear error. We cannot say based on the record before this Court that these findings were “clearly erroneous.” B. Grant Thornton alternatively contends that the district court separately erred in basing the settlement credit on the amounts actually recovered by the FDIC, rather than on the stipulated amount in the settlement agreement. Put differently, Grant Thornton argues that even if the 8.563% credit ratio is correct, the amount of the credit should have been 8.563% of $22 million (the full amount of the stipulated agreement), rather than 8.563% of $15,692,521 (the amount the district court described as money the FDIC had received to date “plus additional guaranteed recovery” (J.A. at 971)). Neither party has pointed to a case from West Virginia where the agreed settlement amount and the amount actually paid or recovered from that settlement differed, and it does not appear that the issue has been directly addressed by any West Virginia published decision. In Hardin v. New York Central Railroad Co, 145 W.Va. 676, 116 S.E.2d 697 (1960), the WVSCA stated the rule as: “[I]f one joint tortfeasor makes a settlement with the plaintiff the amount of the settlement, if presented properly during the trial or after the trial, should be deducted either by the jury or by a court.” Id. at 701 (emphasis added). This language supports Grant Thornton’s position that it is the amount of the agreed settlement that governs. However, there is somewhat contradictory language in Tennant v. Craig, 156 W.Va. 632, 195 S.E.2d 727 (1973), which instructs that “[wjhere a payment is made, and release obtained, by one joint tort-feasor, other joint tort-feasors shall be given credit for the amount of such payment in the satisfaction of the wrong.” Id. at 730 (emphasis added); see also Savage v. Booth, 196 W.Va. 65, 468 S.E.2d 318, 323 (1996) (citing Zando for the proposition that the non-settling joint tortfeasor is “entitled to receive credit for the settlement amount paid”). Because there was no discrepancy between the face value of the settlement and the amount of the settlement payment in Zando, Tennant, or Hardin, we are left as a federal court seeking to apply state law to forecast how the WVSCA would determine the issue. See Ellis, 530 F.3d at 287 (in case governed by state law, if the state’s highest court “has spoken neither directly nor indirectly on the particular issue before us, we are called upon to predict how that court would rule if presented with the issue”) (citation omitted). We conclude that the best indication of how the WVSCA would resolve this issue is set forth in Zando by virtue of the approval of both UCATA Section 4 and citation to Tommy’s Elbow Room, Inc. v. Kavorkian, 754 P.2d 243 (Alaska 1988). See Zando, 390 S.E.2d at 805 (West Virginia’s “practice with regard to verdict re *205 duction basically comports with Section 4 of the UCATA”). Based on the WVSCA’s statement that its practice “basically comports” with UCATA Section 4, id., we predict that court would adopt the language of the uniform act, including that the settlement credit to the nonsettling defendant is the amount “stipulated by the release or the covenant, or ... the amount of the consideration paid for it, whichever is the greater.” Cf. 12 U.LA. at 98 (1975), quoted in Zando, 390 S.E.2d at 803. This conclusion is bolstered by the WVSCA’s statement in Zando after discussing UCATA Section 4 “to have the verdict reduced by the amount of any good faith settlements previously made with the plaintiff by other jointly liable parties.” Id. at 806. The plain language of UCATA Section 4 directs that the settlement credit should be based on the amount stipulated by the release with the settling defendant, if that is greater than the amount paid. Applying that provision here, the setoff should have been calculated based on the face amount of the settlement, $22 million. Similarly, Tommy’s Elbow Room interpreted the same language, which had been adopted by statute in Alaska. 754 P.2d at 244-45 (citing Alaska Stat. § 09.16.040(1)). Indeed, the issue in Tommy’s Elbow Room was whether a nonsettling defendant’s liability should be reduced by the face amount stipulated in the settlement agreement or by the amount paid. Id. The Alaska Supreme Court concluded that the proper amount for the setoff was the (likely) higher settlement amount, rather than the amount actually recovered in settlement. Id. at 246-47. The district court here noted that other jurisdictions had criticized that rule, as adopted in Tommy’s Elbow Room, and concluded that the WVSCA would not follow it, either. However, we find Zando’s favorable reference both to Tommy’s Elbow Room and to UCATA § 4 to be the better indicators of how the WVSCA would rule. Accordingly, we predict that the WVSCA would apply the plain language of UCATA Section 4, just as the Tommy’s Elbow Room court did, and would base the setoff amount here on the face value of the settlement. 7 Thus, the proper settlement credit here should be 8.563% of $22 million, or $1,883,860. Thus, the final judgment against Grant Thornton, adjusted for the proper settlement credit, should have been $23,196,917 ($25,-087,777 minus $1,883,860), not $23,737,026.43. V. Prejudgment Interest The district court denied the FDIC’s request for an award of prejudgment interest on the judgment amount owed by Grant Thornton. It first examined the federal statute, 12 U.S.C. § 1821(i), which authorizes the award of interest in a bank receivership proceeding: In any proceeding related to any claim against an insured depository institution’s director, officer, employee, agent, attorney, accountant, appraiser, or any other party employed by or providing services to an insured depository institu *206 tion, recoverable damages determined to result from the improvident or otherwise improper use or investment of any insured depository institution’s assets shall include principal losses and appropriate interest. 12 U.S.C. § 1821(0 (emphasis added). 8 The district court concluded that the award of prejudgment interest was not appropriate because, under applicable West Virginia law, W. Va.Code § 56-6-81, “the court does not believe that the damages the FDIC seeks to recover are ‘special or liquidated damages’ ” warranting prejudgment interest. (J.A. at 886.) The FDIC contends in its cross-appeal that this determination was error. 9 The FDIC essentially argues that the statutory language “recoverable damages ... shall include principal losses and appropriate interest” means that prejudgment interest must be awarded in all cases; it is never discretionary. Conversely, Grant Thornton argues that the district court correctly interpreted § 1821(0 to mean that “if prejudgment interest is available under West Virginia law in a case of this nature, the court may award prejudgment interest.” (Cf. J.A. at 884.) There is a dearth of case law applying this statute, and none of the cases that reference the provision expressly address the arguments raised by the parties here. See, e.g., Fed. Deposit Ins. Corp. v. Mijalis, 15 F.3d 1314, 1326-27 (5th Cir.1994) (in case where the parties disputed only the rate of interest, court seemingly interpreted “appropriate interest” as meaning the “appropriate rate of interest,” but declined to address the propriety of the interest awarded because defendants did not properly preserve the issue); Fed. Deposit Ins. Corp. v. UMIC, Inc., 136 F.3d 1375, 1385 (10th Cir.1998) (prejudgment interest could not be awarded under this section because the conduct upon which the judgment was based occurred before the enactment of Section 1821(i)). Likewise, we have not found any cases specifically discussing whether similarly-worded provisions require the award of prejudgment interest in all cases or simply allow it in appropriate cases. See, e.g., 12 U.S.C. § 1787(i) (“recoverable damages ... shall include principal losses and appropriate interest”); 12 U.S.C. § 4617(h) (same); 12 U.S.C. § 5390(g) (same). In construing the statute, we must give the words therein their ordinary meaning. United States v. Abdelshafi, 592 F.3d 602, 607 (4th Cir.2010) (statutory interpretation requires that the court “strive to implement congressional intent by examining the plain language of the statute” and to give a statute its “plain meaning,” which in turn is “determined by reference to its words’ ... ‘ordinary, contemporary, [and] common meaning’ ”) (citations omitted). As an initial matter, we note that nothing in § 1821(i) refers to prejudgment interest, but simply to “appropriate inter *207 est.” The FDIC argues that to interpret “interest” as referring only to post-judgment interest would render the language superfluous, since there is already a statute providing for the award of post-judgment interest to all successful plaintiffs in civil cases, 28 U.S.C. § 1961(a). On this basis, we conclude that the reference to “appropriate interest” in § 1821(i) may include both postjudgment and prejudgment interest. See Comeau v. Rupp, 810 F.Supp. 1172, 1180-81 (D.Kan.1992). However, the mere fact that “appropriate interest” could include both prejudgment and post-judgment interest, does not lead to the conclusion that the statute mandates prejudgment interest must always be awarded. See id. at 1180 (“Notwithstanding the authority to award prejudgment interest under [the Financial Institutions Reform, Recovery, and Enforcement Act of 1989, Pub.L. No. 101-73, 103 Stat. 183, (“FIRREA”) ],” the court will still have to decide “the appropriateness of such an award.”). Unsurprisingly, the parties offer competing interpretations for this dilemma, but both essentially add language to the statute in doing so. The FDIC’s interpretation of “appropriate interest” is “an appropriate rate of interest” or “an appropriate amount of interest.” Grant Thornton’s interpretation is that “appropriate interest” means “interest, if appropriate” or “interest, in an appropriate case” or “interest, if otherwise appropriate.” We conclude that Grant Thornton has the better argument and its interpretation is the most harmonious with a natural reading of the statute. See 62 Cases, More or Less, Each Containing Six Jars of Jam v. United States, 340 U.S. 593, 596, 71 S.Ct. 515, 95 L.Ed. 566 (1951) (“[O]ur problem is to construe what Congress has written. After all, Congress expresses its purpose by words. It is for us to ascertain — neither to add nor to subtract, neither to delete nor to distort.”). First, the common dictionary definition of “appropriate” more easily comports with Grant Thornton’s interpretation. Most often, “appropriate” means “specially suitable: fit, [or] proper.” Webster’s Third New Int’l Dictionary 106 (1961). We could easily substitute those definitional words for “appropriate” and the statute would continue to mean what Grant Thornton and the district court interpreted it to mean. That is, “damages shall include ‘specially suitable’ interest” or “ ‘fitting’ interest” or “ ‘proper’ interest.” By contrast, to interpret “appropriate” as specifically referring only to the rate or amount of interest would require additional words with a different substantive meaning being written into the statute. Second, the FDIC hinges its argument primarily on the fact that the statute contains the word “shall,” a mandatory and not permissive term. (See Principal Br. for FDIC at 62 (citing United States v. Monsanto, 491 U.S. 600, 607, 109 S.Ct. 2657, 105 L.Ed.2d 512 (1989)).) But while Congress used the language “shall,” it also included the word “appropriate” for a purpose. Nothing in the statute suggests that “appropriate” refers to a rate or amount of interest. Indeed, other statutes providing that interest shall be an element of damages do not include the limitation “appropriate.” See, e.g., 28 U.S.C. § 1961(a) (stating that postjudgment interest “shall be allowed on money judgment in a civil case recovered in a district court” and specifying in detail how interest is to be calculated); 7 U.S.C. § 2564 (in infringement of plant variety protection, the court “shall award damages adequate to compensate for the infringement but in no event less than a reasonable royalty for the use made of the variety by the infringer, to *208 gether with interest and costs as fixed by the court”). Notably, moreover, a statutory review makes clear that Congress knows how to specify rates of interest or to refer to certain factors in setting interest when it chooses to do so. See, e.g., 42 U.S.C. § 9607(a)(4) (setting forth damages in CERCLA cases and explaining what damages the interest applies to, the dates of accrual and referring to specific rates of interest); 15 U.S.C. § 15 (in antitrust actions, instructing “[t]he court may award ... simple interest on actual damages,” describing prejudgment interest period, and allowing such an award “if the court finds that the award of such interest for such period is just in the circumstances”). In those statutes, Congress did not simply say that an “appropriate” rate should be used, but instead gave specific particulars about the rate, the time period for interest, and/or or how to calculate it. In contrast, no such directions appear in § 1821(0For these reasons, we find the word “appropriate” is best read as a limitation as to when prejudgment interest should be provided, not as a reference to any particular “rate” or amount of interest nor that its award is mandated in all cases. Thus, we conclude that the award of prejudgment interest under § 1821 (0 is discretionary, ie., that it need only be awarded if appropriate. The FDIC next contends that, even if we adopt Grant Thornton’s interpretation of the statute, the district court nonetheless abused its discretion in denying interest because it looked solely to West Virginia law to determine whether interest was appropriate in this case. For support, the FDIC cites to United States v. Dollar Rent A Car Systems, Inc., 712 F.2d 938 (4th Cir.1983) (“Dollar”). The FDIC’s reliance on Dollar is misplaced. In Dollar, this Court concluded that the district court abused its discretion when it awarded prejudgment interest and considered itself bound by the lower rate established by state law. Id. at 941. This was an abuse of discretion because federal law, not state law, governed that case and federal law in fact granted discretion to award a higher rate. Id. Dollar does not stand for the general proposition, as cited by the FDIC, that a district court abuses its discretion when it analyzes whether interest is warranted under state law. The district court’s analysis here suggests that it interpreted the language as meaning that it would award interest if appropriate under West Virginia law. Although the court’s opinion does not explicitly state that it was constrained by West Virginia law, it analyzed the issue under West Virginia law and concluded as follows: Even assuming that the FDIC is entitled to prejudgment interest under FIR-REA, the court does not believe that the damages the FDIC seeks to recover are “special or liquidated damages” within the meaning of W. Va.Code § 56-6-31. Accordingly, the FDIC’s request for an award of prejudgment interest is denied. (J.A. at 886; see also id. at 885-86 (concluding that the damages here were neither special nor liquidated damages).) The district court’s reference to state law here was not an abuse of discretion. Particularly in view of our earlier discussion of O’Melveny & Myers, where the Supreme Court held the FDIC in FIR-REA cases was to “work out its claims under state law,” 512 U.S. at 87, 114 S.Ct. 2048, the district court correctly looked to the applicable state law in order, to determine whether prejudgment interest was “appropriate” in this case. See United States v. Am. Mfrs. Mut. Cas. Co., 901 F.2d 370, 372-73 (4th Cir.1990) (in the *209 absence of “explicit standards for the allowance of pre-judgment interest,” federal statutes are “treated as incorporating the applicable state law on [the] issue”); Quesinberry v. Life Ins. Co. of N. Am., 987 F.2d 1017, 1030 (4th Cir.1993) (“absent a statutory mandate the award of prejudgment interest is discretionary with the trial court”). Moreover, we find no abuse of discretion in the district court’s determination that prejudgment interest here was not warranted. Although the court ultimately arrived at a damages award, until it did so, it was impossible for Grant Thornton to know how much it owed. 10 See Lockard v. City of Salem, 130 W.Va. 287, 43 S.E.2d 239, 243 (1947) (“interest is denied when the demand is unliquidated for the reason that the person liable does not know what sum he owed and therefore cannot be in default for not paying.”) (citation and quotation marks omitted); Bond v. City of Huntington, 166 W.Va. 581, 276 S.E.2d 539, 550 (1981) (prejudgment interest applies where case involves “only pecuniary losses that are subject to reasonable calculation that exist at the time of the trial”). Similarly, under federal law, courts may deny prejudgment interest where “a legitimate controversy existed” regarding the amounts ultimately deemed to be owed. Moore Bros., 207 F.3d at 727. Instead, the court must “weigh the equities in a particular case to determine whether an award of prejudgment interest is appropriate.” Id. In short, we find no error in the district court’s denial of prejudgment interest. VI. Conclusion For the foregoing reasons, we affirm the district court’s judgment in all respects except the amount of the settlement credit. As to that issue, we reverse and remand for further proceedings consistent with this opinion. AFFIRMED IN PART AND REVERSED AND REMANDED IN PART. Footnotes [1] The FDIC’s claims were first asserted below as counterclaims in Grant Thornton v. FDIC, No. 1:00-0655, and in a complaint in intervention in Gariety v. Grant Thornton, LLP, No. 2:99-0992. The district court subsequently severed the FDIC’s claims against Grant Thornton from the other claims in Gariety, assigned a new case number (No. 1:03-2129), and consolidated them for trial with the FDIC’s claims in No. 1:00-0655. [2] The general factual background set forth in Ellis and repeated herein tracks to a large degree the district court’s findings of fact in this case. [3] At oral argument, the FDIC acknowledged that Grant Thornton’s liability would not have extended indefinitely, but instead could have been naturally limited by any subsequent audit required to be conducted by federal regulators at regular intervals. [4] The question of whether the district court correctly held that the actions of Keystone’s management could not be imputed to the FDIC is addressed in detail infra at Section III. [5] Section 4 of UCATA provides: When a release or a covenant not to sue or not to enforce judgment is given in good faith to one of two or more persons liable in tort for the same injury ... [i]t does not discharge any of the other tortfeasors from liability for the injury....; but it reduces the claim against the others to the extent of any amount stipulated by the release or the covenant, or in the amount of the consideration paid for it, whichever is the greater. Bd. of Educ. of McDowell Cnty. v. Zando, Martin & Milstead, Inc., 182 W.Va. 597, 390 S.E.2d 796, 803 n. 6 (1990) ("Zando ") (quoting 12 U.L.A. at 98 (1975)) (emphasis added). As described by the WVSCA in Zando, this results in a “pro tanto, or dollar-for-dollar, credit for partial settlements against any verdict ultimately rendered for the plaintiff.” Id. at 805. [6] In making the allocation determination, the district court emphasized that it was attempting to adhere to the principle expressed in Zando that "a plaintiff is entitled to one, but only one, complete satisfaction for his injury.” 390 S.E.2d at 803. The court also discussed favorably the Supreme Court of Virginia’s decision in Tazewell Oil Co., Inc. v. United Va. Bank, 243 Va. 94, 413 S.E.2d 611 (1992), in which a plaintiff sued three banks for various acts of creditor misconduct resulting in various harms to the plaintiff. Id. at 617. After settling with two banks, plaintiff obtained a jury verdict against the third bank, and the trial court allowed a credit against the verdict for the full amount of the settlements. Id. The Supreme Court of Virginia reversed, remanding for an allocation among the multiple injuries and instructing that "the court must look at the injury or damage covered by the release and, if more than a single injury, *204 allocate, if possible, die appropriate amount of compensation for each injury.” Id. at 622. [7] We understand the criticism of this rule, which may make it more difficult for a plaintiff to obtain a complete satisfaction. However, the same competing policy issues would have been evident to the WVSCA when it discussed UCATA § 4 in Zando. We further note, as did the Tommy's Elbow Room court, that any time a plaintiff settles a claim, the plaintiff assumes the risk that some amount of the settlement will not be recoverable. 754 P.2d at 245. The fact that it ultimately is not recoverable and that the non-recovery works to the detriment of the plaintiff, is the result of the plaintiff's own bargain in agreeing to the settlement amount and its terms of payment. [8] The district court did not address the question under § 1821(Z) of whether the damages at issue here resulted from an "improvident or otherwise improper use of ... [Keystone’s] assets.” Instead, the court assumed, without deciding, that this condition was satisfied. In light of our conclusion that the district court properly concluded that prejudgment interest was inappropriate in the case at bar, we too assume, without deciding, the damages satisfy this condition. [9] The parties appear to agree that the district court’s decision as to whether to award prejudgment interest is reviewed for abuse of discretion. (See Principal Br. for FDIC at 64; Response/Reply Br. for Grant Thornton at 48-49 (citing Moore Bros. Co. v. Brown & Root, Inc., 207 F.3d 717, 727 (4th Cir.2000)).) [10] In addition to numerous disagreements at trial regarding a proper measure of damages, here a special separate hearing was held after the trial to determine the Kutak setoff amount, if any. Until that hearing was held and the district court issued its ruling, Grant Thornton could not know what amount it owed.