Opinion
17-cv-5340 (LJL)
10-27-2020
AMTRUST NORTH AMERICA, INC., et al., Plaintiffs, v. KF&B, INC., Defendant.
FINDINGS OF FACT AND CONCLUSIONS OF LAW
:
Plaintiffs AmTrust North America, Inc. ("AmTrust N.A."), Wesco Insurance Company, Inc. ("Wesco"), and Technology Insurance Company, Inc. ("TIC") (collectively, "Plaintiffs" or "AmTrust") bring this action against Defendant KF&B, Inc. d/b/a KF&B Program Managers Insurance Services ("Defendant" or "KF&B") for claims of breach of contract, breach of fiduciary duty, and a declaratory judgment.
A bench trial by remote means commenced on September 21, 2020 and concluded on September 24, 2020, with additional testimony on October 1, 2020. Closing statements took place on October 6, 2020. This opinion constitutes the Court's Findings of Fact and Conclusions of Law pursuant to Fed. R. Civ. P. 52.
For the reasons that follow, the Court finds in favor of Defendant KF&B on Plaintiffs' claims.
FINDINGS OF FACT
I. The Parties
1. Plaintiff AmTrust N.A. is a wholly owned subsidiary of AmTrust Financial Services, Inc. ("AFSI"). Plaintiffs Wesco and TIC are insurance companies that were purchased and are wholly-owned by AFSI. AmTrust N.A., Wesco, and TIC are Delaware corporations with their principal places of business in New York. Prior to 2017, TIC was domiciled in New Hampshire.
2. As relevant to this action, AmTrust acted as the corporate manager for other AmTrust subsidiaries and affiliated entities, sourcing program managers, or managing general agents for specialty programs for those entities.
3. KF&B is a California corporation with its principal place of business in California. It is a licensed wholesale insurance broker and surplus agent focused mainly in the public auto sector.
II. Limousine and Taxi Program
4. Pursuant to contract, KF&B served as the program manager (also known in the industry as its managing general agent ("MGA")) for AmTrust's limousine and taxi insurance program known as the KF&B Limousine and Taxi Program (the "Program"). In that respect, and as further described below, KF&B was responsible for underwriting and selling AmTrust insurance policies to qualified limousine and taxi companies and for managing aspects of the Program. The Program had 2,400 policies and 1,500 accounts. Am. Leatzow Decl. ¶ 39.
An MGA is an individual or business entity appointed by an insurer to perform certain functions. In this case, KF&B was hired as MGA to perform underwriting functions by evaluating potential insureds to determine whether the risk fit AmTrust's risk profile. The MGA rates the proposed policy by considering certain parameters, including the loss history of the insured and the reasons for that loss history.
5. The taxi accounts in the Program were primarily handled on the AmTrust side by Michael Hynes ("Hynes") and Jerrell Fort ("Fort"), who were AmTrust in-house underwriters between 2010-2012 and 2012-2018, respectively, and Christopher Bellis ("Bellis"), who was a vice president of underwriting at AmTrust during the pendency of the Program. The Program's taxi accounts were handled on the KF&B side by Michael Howery ("Howery"), who was KF&B's lead taxi underwriter on the Program. AmTrust employed actuaries while KF&B did not. Those actuaries included Gregory Kushner ("Kushner") and John Lamendola ("Lamendola").
6. During KF&B's underwriting process, the broker served as the intermediary between KF&B and the insured, and KF&B often did not speak to the insured during the underwriting process. Tr. at 140:14-17. For the Program's taxi accounts, the broker was more often than not BB&T John Burnham Insurance Services ("BB&T"). KF&B had a history of working with BB&T from 2009 through its prior relationship with a firm named Praetorian/QBE ("Praetorian"). Am. Howery Decl. ¶¶ 57-58.
7. In this case, AmTrust seeks damages as a result of KF&B's alleged contract breaches in the underwriting of 15 of the taxi accounts in the Program.
III. The Agreements
A. Managing Producer Agreement
8. The central document in this case is the managing producer agreement ("MPA"), entered into by KF&B and AmTrust N.A., on behalf of Wesco, effective on July 1, 2011. See D-Howery-1 (the MPA, listing "Insurer" as "AmTrust North America, Inc. on behalf of: Wesco Insurance Company, a Delaware domiciled, licensed insurer (collectively, AmTrust)" and signed by Bellis on behalf of AmTrust N.A. and Wesco). The initial term of the MPA was one year from July 1, 2011 through June 30, 2012. The MPA was extended by both parties three additional times, on an annual basis for one-year terms. The parties also entered into certain amendments throughout the course of their relationship. D-Howery-1 at Endorsement 1, Endorsement 2 and Endorsement 3 to the MPA. TIC was added as a party to the MPA by written amendment effective July 1, 2013.
9. The relationship between KF&B and AmTrust ended on December 31, 2015, when AmTrust terminated the MPA.
10. The relationship between KF&B and AmTrust grew out of a prior relationship that KF&B had with Praetorian. Before bringing the limousine and taxi program to AmTrust, KF&B developed and managed a similar program for Praetorian, pursuant to an arrangement where KF&B acted as managing producer and Praetorian acted as insurer primarily for limousine accounts. In 2010, Praetorian asked KF&B to begin underwriting a small number of taxi accounts; KF&B had the ability to bind the accounts and issue policies. Through the management of that program, KF&B developed relationships with limousine and taxi accounts and brokers who made introductions to such accounts.
11. In or around 2010, after Hynes and an underwriter named Fabian Burstyn had left Praetorian and joined AmTrust, they asked KF&B whether it would be interested in bringing the taxi and limousine program to AmTrust. See Am. Howery Decl. ¶ 24; Tr. at 69:2-20. KF&B was interested and the relationship was borne.
12. Under the MPA, KF&B agreed to market AmTrust's limousine and taxi program and to solicit, underwrite, and bind accounts pursuant to guidelines set forth by AmTrust and in accordance with industry standards. AmTrust agreed to pay KF&B a fee based on the premiums AmTrust received.
13. Specifically, the MPA "appoint[ed] Producer [KF&B], on a non-exclusive basis, for the purpose of marketing, soliciting, underwriting, binding, executing, and servicing on behalf of AmTrust the types and classes of business in the states authorized in a Program." MPA, Section II.A.
14. Within and subject to AmTrust's guidelines, the MPA gave KF&B broad authority. It provided that:
Producer shall have authority to solicit, underwrite, rate, negotiate, prepare, bind, sign, distribute, and cancel Policies; to collect Premiums and pay return Premiums; to prepare and distribute all policyholder notices including endorsements, renewal and non-renewal notices, additional insured and policy certificates; to retain, oversee, and pay a loss control vendor; to cooperate with AmTrust in the selection of a claim administrator; if applicable, to respond to all Bureau Criticisms within ten (10) days of receipt of notice from AmTrust; to follow AmTrust Policy Deductible Guidelines, which are incorporated into and made part of this Agreement; to perform such additional duties as are set forth in a Program; and to serve as program manager with respect to the business described in a Program.Id.
15. The exercise of that authority was subject to standards. KF&B "agrees to perform the functions specified in a Program in accordance with the highest standards of the industry and pursuant to the standards and procedures set forth in a Program, including compliance with all Underwriting Guidelines and directives issued by AmTrust from time to time." Id., Section II.B (emphasis added).
16. It was also subject to limits. KF&B "shall have no authority to act on behalf of or bind AmTrust in any way except as expressly stated in a Program." Id., Section II.C.
17. Further, the MPA stated that KF&B "shall perform its obligations as a fiduciary of AmTrust, and shall use its best efforts to perform all acts necessary for the proper conduct of business on behalf of AmTrust." Id., Section III.
18. As noted above, KFB's compensation under the Program was tied in part to the premiums AmTrust received and to the success of the Program. KF&B's compensation increased the longer the Program continued. Under the MPA, in Program Years 1 and 2, the first two years of the Program, KF&B's fee was calculated at a rate of 20.5% of the gross written premium, reduced by returns and cancellations. Am. Howery Decl. ¶ 137; D-Howery-1 at Addendum Number 1 to MPA, Section 4(a)(i); D-Howery-1 at Endorsement 1 to the MPA, Section 4(a)(ii). That fee of 20.5% increased to 21.5% in Program Year 3 and to 22.5% in Program Year 4. Am. Howery Decl. ¶ 39; D-Howery-1 at Endorsement 2 to the MPA, Section 4. The MPA incented KF&B to underwrite and bind policies on behalf of AmTrust.
19. The MPA also incented KF&B to underwrite profitable policies. The compensation paid to KF&B assumed a provisional expected loss ratio of 60%. If the actual loss ratio was less than 60% (meaning the accounts were more profitable), the commission would be adjusted upward. If the actual loss ratio was less than 60%, the difference between that loss ratio and 60% would be multiplied by the gross earned premium and carried forward and added to losses incurred in subsequent program years for the purpose of the loss ratio calculation.
20. Not surprisingly for contracts of this type, the MPA imposes on the parties reciprocal indemnification obligations. AmTrust agreed to "indemnify, defend and hold [KF&B], its directors, officers, employees and representatives harmless against all civil and administrative liability, including reasonable attorney's fees and costs of investigation and defense, arising out of the relationship of the parties under this Agreement caused by AmTrust's act, error, or omission, except to the extent [KF&B] has caused, contributed to, or compounded such act, error, or omission." MPA, Section XIX.
21. The MPA imposes similar, and somewhat broader, indemnification obligations on KF&B. It provides:
[K&B] will, as a condition precedent to AmTrust's indemnification obligations, give AmTrust written notice upon receipt of any claim or legal action, complaints from policyholders, claimants, insurance departments, or regulators relative to any matter covered by this [MPA]. AmTrust will be entitled but not required, to participate in or assume the defense of [KF&B] in any such action. [KF&B] agrees to cooperate fully in the handling of any such action.
[KF&B] will indemnify, defend and hold AmTrust, its affiliates, and its and their directors, offices, employees and representatives harmless against all civil and
administrative liability, including reasonable attorney's fees and costs of investigation and defense, arising out of the relationship of the parties under this [MPA] caused by [KF&B]'s act, error, or omission, or the act, error, or omission of any person acting under the direction of [KF&B], including sub-agents, producers, and brokers who place business through [KF&B], except to the extent AmTrust has caused, contributed to or compounded such act, error, or omission. [KF&B] will accept and pay all fines, penalties, fees, administrative payments, and costs levied against AmTrust or [KF&B] individually or jointly, by any regulatory agency, governmental unit, or court for any violation of law or regulation attributable to the act, error, or omission of [KF&B]. [KF&B] shall be responsible to AmTrust for all acts, errors, or omissions of sub-agents, producers, and brokers. In the event [KF&B] fails to adhere to the provisions of this [MPA] or the Underwriting Guidelines when quoting, binding, underwriting or rating a Policy, whether intentional or not, [KF&B] will immediately take lawful actions necessary to remove or to reduce AmTrust's exposure on such unacceptable risks or limits, and shall indemnify AmTrust for any damages related thereto.Id. (emphasis added). The last of those provisions—italicized above—forms the basis for AmTrust's claim for damages in this case.
22. The MPA states that the "failure of AmTrust or [KF&B] to enforce any obligation under this Agreement shall not constitute a waiver of any right under the Agreement, nor shall past waiver of any provision constitute a course of conduct or a waiver in the future of the same provision." Id., Section XXII.
23. In addition, the MPA reflects the fact that the relationship between KF&B and AmTrust would not be the normal relationship between an insurer and its agent. Under "Termination Rights and Injunctive Relief," the MPA stated:
Both Producer and AmTrust understand that this Agreement places responsibilities, duties, and obligations upon Producer different from those of a normal insurance producer or agency contract. Because this Agreement has been mutually entered into for a special purpose, Producer acknowledges that this necessitates a different relationship. Producer therefore specifically agrees not to pursue any regulatory review of the termination of this Agreement or require AmTrust to comply with any statutory or regulatory provisions relating to notice, "run-off' or continuation of business written that may now or hereafter be provided by statute or regulation in recognition of a normal insurance producer or agency relationship. Furthermore, Producer shall not require compliance by AmTrust of any obligations relating to termination of this Agreement other than those provided for specifically herein.
Producer acknowledges that in the event of any threatened or actual breach of this Agreement by Producer, AmTrust will suffer immediate and irreparable injury not compensable by monetary damages and for which AmTrust will not have an adequate remedy available at law. Therefore, if AmTrust institutes an action or proceeding to enforce the provisions of this Agreement, it shall be entitled to obtain from a court of competent jurisdiction, without the posting of any bond or security, such injunctive relief, restraining orders, specific performance, or other equitable relief as may be necessary or appropriate to prevent or curtail any such threatened or actual breach. The foregoing shall be in addition to any other rights or remedies AmTrust has at law or in equity.Id., Section XXI.
24. The MPA incorporated by reference certain underwriting guidelines ("Underwriting Guidelines"). The Underwriting Guidelines are of particular note with respect to this case.
B. Underwriting Guidelines
25. The initial underwriting guidelines were dated January 1, 2011 and were updated annually through 2015.
The parties do not dispute that sections of the Underwriting Guidelines at issue in this case remained largely identical in all versions of the Underwriting Guidelines. Accordingly, citations and references to the Underwriting Guidelines are to the 2015 Underwriting Guidelines unless otherwise noted.
26. The Underwriting Guidelines distinguished between limousine accounts and taxi accounts. As a general matter, KF&B had authority—without prior approval from AmTrust—to underwrite and bind limousine accounts.
Plaintiffs have dropped their claims with respect to the limousine accounts and thus those are not discussed in this opinion.
27. By contrast, the parties agree that the Underwriting Guidelines required KF&B to refer all taxi accounts to AmTrust's in-house underwriters for their rejection or approval. If AmTrust agreed that a taxi account should be bound, it would indicate that to KF&B which would then bind the account. If AmTrust did not agree, the account would not be bound.
28. In the case of both limousine and taxi accounts, however, KF&B was obligated to follow the Underwriting Guidelines in connection with the underwriting of an account and to take on indemnification obligations for "damages related thereto" if it failed to adhere to the MPA or Underwriting Guidelines. Id., Section XIX.
29. The Underwriting Guidelines reflect that they were to serve as a "reference tool" and "framework for underwriting operations" to "guide the underwriters to select accounts having characteristics common to well run operators." Underwriting Guidelines at 1. The Introduction to the Underwriting Guidelines reflect their precatory nature:
The guideline has been prepared as a reference tool for underwriting and administrating the KFB Public Auto Program on behalf of AMTRUST UNDERWRITERS, INC.Id.
The goal as underwriting managers is to identify the most favorable accounts for the program by assessing their risks and pricing accordingly. The purpose of this manual is to guide the underwriters to select accounts having characteristics common to well run operators. The ultimate goal of the manual is to provide AMTRUST UNDERWRITERS, INC. with preferred accounts, which have been priced to provide a successful underwriting profit.
This guide is designed as a framework for underwriting operations. As such, the guidelines will be modified from time to time. Additionally discussion at all levels between KFB, INC. and AMTRUST UNDERWRITERS, INC. will be used to enhance, streamline or make any other modifications, which will improve the results for AMTRUST UNDERWRITERS, INC. and enhance the program's longevity.
30. The detailed provisions of the Underwriting Guidelines set forth in succeeding sections describe the focus of the Program, the classes of services that are prohibited under the Program, the accounts that must be submitted to AmTrust for approval before binding, and the limits of liability and coverage. The Underwriting Guidelines also have sections addressing the required contents for applications for insurance and underwriting criteria.
31. A section headed "Application" lists the information required for each submission directed to the risk associated with each prospective account. Consistent with the differing nature of the accounts and KF&B's authority with respect to those accounts, such required information differs depending on whether the account being underwritten is for taxis or limousines. The requirements for an application included, inter alia:
• "A fully completed application" that "must be signed by the producing broker and signed by the insured . . ." Id. at 5.
• "A fully completed vehicle schedule," which includes garaging location, radius of operation, and vehicle ID number. Id. at 6.
• "A fully completed Drivers Schedule" which must include the full name of driver, license number, and date of birth. Id.
• "Prior carrier loss runs, currently valued for at least 3 years (preferably 5 years) . . . other than for new ventures[.]" Id. Loss runs are used to approximate loss experience going forward.
• "Motor vehicle records." Id. at 7.
32. The Underwriting Guidelines also set forth "Underwriting Criteria." The criteria included:
• "Years in business" where "emphasis is placed on underwriting accounts that have been in business in excess of three (3) years and have shown controlled consistent growth or stability in operations." Id. at 8.
• "New ventures" that were not the target risk, but AmTrust was "willing to accommodate them when carefully underwritten." Id.
• "Radius of operations will not be a limiting factor" but the "[t]erritory in which an account is located or operated will be considered as part of the pricing of the risk." Id.
• "Inspections" were required to be conducted by a KF&B-approved inspection company on all new accounts once bound, except for accounts in California, Florida, and Ohio, which had different requirements. Id. at 8-9.
• "Loss ratio analysis," which is a measure of "earned to incurred [to] be completed regularly on all accounts including the experience from the prior years, if any." Id. at 9.
• "Claim frequency analysis," which stated that "[p]referred taxi accounts should have a claim frequency no higher than 1 claim per unit year of experience." Increased frequency is "normally a sign of deteriorating operations linked possibly to management changes, increased operations, changes in operations, new drivers, etc." Id.
• "Liability deductible (taxi)" required taxi fleet accounts to be written "subject to a preferred minimum liability deductible or [self-insured retention] of $25,000" but "[l]ower retention levels may be considered on a submit[ed] basis." Id. at 9-10.
• "Rating approach and variations" that covers "deductible accounts" and "collateralization of deductible obligations." Id. at 10.
33. Under the section for "rating approach and variations," the subsection "deductible accounts" is of particular importance. That subsection sets forth the rating approach for deductible accounts: "The underwriter is to use the current ISO rules and rates. In the event the rates charged are more than ISO, appropriate consent to rate forms are obtained." Id. The taxi accounts at issue were all deductible accounts.
IV. Termination of the Program
34. The limousine and taxi program was profitable for its first few years of operation. However, in subsequent years, coinciding with a downturn in the limousine and taxi business generally, the Program turned unprofitable. In particular, after initially enjoying profits from the Program's largest taxi account, Administrative Services Cooperative, Inc. ("ASC"), AmTrust began to experience significant losses as claims and indemnity losses outstripped deductibles and premiums on the account. AmTrust grew concerned about the ability to turn around the Program. See Tr. at 189:21-190:11; see also id. at 137:11-18; PX 30.
35. AmTrust terminated the Program with KF&B effective the end of December 2015. On August 7, 2015, Bellis of AmTrust sent a notice letter to Howery at KF&B, also on behalf of Wesco and TIC, terminating the Program effective December 31, 2015.
36. This action followed two years later.
V. Procedural History
37. Plaintiffs first brought this action on June 9, 2017 in Supreme Court of New York and served Defendants on June 15, 2017. Dkt. No. 1 ("Complaint" or "Compl."). Defendant removed the action to this Court on July 14, 2017. Id. The action proceeded under the Honorable J. Paul Oetken.
38. The Complaint alleged three causes of action: breach of contract; breach of fiduciary duty; and declaratory judgment.
39. On July 21, 2017, KF&B filed an Answer with counterclaims. Dkt. No. 9. KF&B amended its counterclaim on September 8, 2017, Dkt. No. 22, and AmTrust moved to dismiss. On June 11, 2018, the Court denied Plaintiffs' motion to dismiss the amended counterclaim. Dkt. No. 44. Plaintiffs filed an Answer to the amended counterclaim on June 25, 2018. Dkt. No. 45.
Defendant did not pursue its counterclaim during trial or offer any evidence in support of it. To the extent that any evidence offered related to the counterclaim, it plainly is insufficient to establish the counterclaim. Accordingly, the Court deems the claim abandoned and will enter judgment for AmTrust on it.
40. The action was transferred to the undersigned on February 4, 2020.
41. On April 2, 2020, KF&B moved for partial summary judgment in five separate motions on the following topics: breach of fiduciary duty and declaratory judgment; referred program accounts; maintenance of underwriting files; reputational damage or harm; and causation. Dkt. Nos. 109, 112, 115, 118, 122.
42. On April 3, 2020, the Court held a status conference and scheduled the case for trial to begin on September 14, 2020. Dkt. No. 125.
43. On May 19, 2020, Plaintiffs filed one omnibus brief in opposition to KF&B's motions for partial summary judgment. Dkt. No. 136. KF&B submitted five separate replies on June 1, 2020. Dkt. Nos. 138, 140, 142, 143, 144.
44. On July 22, 2020, the Court held a conference with the parties to address the forthcoming bench trial. Defense counsel noted that certain members of the trial team were high-risk for COVID-19 and that having an in-person trial in New York would expose them to an increased risk of COVID-19. Counsel also represented that its witnesses were all out of state and were in locations that would require them to undergo a 14-day quarantine before trial. Defense counsel therefore expressed the preference that the Court hold the trial, with the witnesses and lawyers all appearing remotely. For their part, Plaintiffs' counsel requested a speedy trial. Although they expressed a preference for an in-person trial, they acknowledged the challenges with lawyers and witnesses who were from out-of-state appearing in person during the COVID-19 pandemic. Ultimately, Plaintiffs' counsel, like defense counsel, consented to a remote bench trial in September.
45. On July 29, 2020, the Court issued proposed virtual hearing procedures and invited the parties to submit objections or alternate procedures one week later. Dkt. No. 150. The parties provided input but did not make objections. In response to the input, the Court modified its procedures, including by permitting a party not bearing the burden of proof to move to supplement the declaration of a witness it intended to offer in defense in its case-in-chief. Dkt. No. 155.
46. On September 11, 2020, the Court granted KF&B's motion for partial summary judgment on Plaintiffs' claim for reputational damage or harm. Dkt. No. 167.
47. On September 14, 2020, the Court granted KF&B's motion for partial summary judgment on Plaintiffs' claims related to the maintenance of the underwriting files, finding no causation or damages. Dkt. No. 168.
48. Also on September 14, 2020, the Court granted KF&B's motion for partial summary judgment on Plaintiffs' breach of fiduciary duty claim, finding it was duplicative of the contract claim, and on Plaintiffs' declaratory judgment claim. Dkt. Nos. 169. It thus dismissed the second and third causes of action.
49. Also on September 14, 2020, the Court denied KF&B's motion for partial summary judgment on the referred program accounts, but deemed certain facts as established. Dkt. No. 170.
50. The Court held the final pretrial conference on September 17, 2020. At that conference, the Court laid out the trial schedule and addressed questions regarding the order of witnesses and introduction of exhibits. The Court decided a series of motions to strike and motions in limine on September 16, 2020.
51. The trial was held between September 21, 2020 and September 24, 2020. Before trial, the Court denied the summary judgment motion on causation. See Tr. at 3:1-8.
52. Plaintiffs offered fact witnesses Mark Murphy ("Murphy") and Kathleen Smith ("Smith") and expert witnesses Donald Bendure ("Bendure") and Michael Scruggs ("Scruggs"). Plaintiffs also offered into evidence the deposition testimony of Fort, Andrew Charrette (an AmTrust auditor), Masoud Shahri at BB&T, Howery, Lawrence Kalior (a KF&B employee), Dubin, and Leatzow. Defendant offered fact witnesses Howery and Hynes and expert witnesses Michael Dubin ("Dubin") and Jim Leatzow ("Leatzow"). Defendant also offered into evidence the deposition testimony of Fort.
53. The Court received testimony on direct by declaration but permitted the parties to amend their declarations shortly before the witness testified to eliminate testimony that was no longer relevant to the issues being tried. Near the end of trial, the Court denied Defendant's motion to strike the trial testimony of Scruggs, finding that it fell within the Court's broad latitude to permit rebuttal testimony. But, because Defendant had eliminated portions of the Dubin declaration offered in the case-in-chief in an attempt to forestall the Scruggs testimony, the Court permitted Defendant to recall Dubin at the end of trial through a revised declaration on which Dubin was cross-examined by Plaintiffs on October 1, 2020. Dkt. Nos. 187, 188.
The only witnesses who were called were within the control of the respective party who called them to testify.
54. Closing arguments were on October 6, 2020.
55. At the beginning of trial, the Court directed the parties to inform the Court of any issues with respect to the remote proceedings. With the exception of a few occasions when the transmission "froze," which were quickly addressed to the satisfaction of the parties, no party indicated any objection to the remote nature of the proceeding. Throughout the trial, both parties indicated satisfaction with the remote technology being used to transmit testimony and argument. Counsel were able to examine uninterrupted and the Court was able to hear and observe the witnesses and make credibility findings. At the conclusion of trial, in response to the Court's inquiry, neither counsel raised any issues with respect to the remote nature of the trial.
VI. Plaintiffs' Breach of Contract Claim
56. Plaintiffs allege that KF&B breached the MPA and the Underwriting Guidelines by performing the calculations used to derive a premium for the select deductible taxi accounts with the result that KF&B recommended for underwriting and binding taxi accounts that would not pay sufficient premium to cover their projected losses. Specifically, Plaintiffs allege that KF&B calculated the premiums in a manner that reduced the premium specified for an account and concealed or minimized the risk associated with the account, in violation of: the specific provision of the Underwriting Guidelines requiring KF&B "to use the current . . . rules and rates" of the Insurance Services Offices ("ISO"); KF&B's obligations to underwrite consistent with the "highest standards of the industry"; and KF&B's contractual fiduciary obligations. The details of these alleged violations are addressed below.
Plaintiffs initially alleged that Defendant falsified territory codes, class information, and other inputs in order to further reduce premiums. The Court previously determined that it was not genuinely in dispute that Defendant did not manipulate territory codes in connection with taxi accounts that were referred by KF&B to AmTrust—which are the only accounts now at issue in the case—and treated that fact as established in the case pursuant to Fed. R. Civ. P. 56(g). See AmTrust North Am., Inc. v. KF&B, Inc., 2020 WL 5513458, at *2 n.1 (S.D.N.Y. Sept. 14, 2020). In addition, when asked whether there was testimony that Plaintiffs were no longer offering, Plaintiffs agreed to strike and/or withdraw information as to the detrend factor and compound calculation errors from the declaration of their expert, Michael Scruggs. Tr. at 681:9-683:2. They agreed that those alleged errors no longer needed to be addressed in the case. See Tr. 819:9-821:6. Plaintiffs also brought claims that KF&B failed to comply with its collateral and deductible obligations but have since withdrawn those claims. See Dkt. No. 185.
VII. KF&B's Alleged Manipulation of Pricing Formulas to Reduce Premiums
57. AmTrust's claims arise from a two-part process that KF&B used to determine whether to recommend to AmTrust for binding or for AmTrust to determine whether to underwrite and bind a taxi account. In the first part of that two-part process, KF&B would collect information from the broker for a prospective account, evaluate that information, and insert it into a prescribed formula to calculate the premium that AmTrust would be permitted to charge for a taxi policy under the various formulas and rates AmTrust had agreed with the states. In the second part, KF&B would conduct its own analysis of the account, based on the insured's loss history, to derive a figure for the risk of the account and to compare it to the calculated premium. AmTrust alleges that KF&B breached duties to AmTrust in the first part.
58. Specifically, AmTrust alleges that, as part of that process: (1) KF&B's use of an algorithm for calculating premiums from 2011 to 2014 did not reflect an accurate and faithful interpretation of the current rules contained in the ISO Commercial Lines Manual ("ISO-CLM"), which AmTrust alleges was in violation of the Underwriting Guidelines' requirement that KF&B use the current ISO rules; (2) even after the error was corrected, KF&B continued to miscalculate premiums in order to underwrite accounts that it knew would be unprofitable; and (3) KF&B made errors in the calculation of a specific portion of the formula for premium, relating to the past loss experience of the insured against which a deductible should be applied in a manner, which gave a misleading impression that the account was more profitable in the past than it actually was. AmTrust claims that, as a result of these errors, KF&B recommended accounts with premiums that did not sufficiently cover the risk AmTrust was agreeing to assume. Had AmTrust been aware of the true risk profiles of certain accounts, AmTrust argues it never would have issued the policies, let alone at the premiums that the policies were written and AmTrust would not have suffered any losses from these accounts.
59. Initially, prior to trial, AmTrust sought relief with respect to 25 accounts, of which 15 were taxi accounts and 10 were limousine accounts. During trial, AmTrust abandoned its claim for relief with respect to the 10 limousine accounts. As such, it seeks relief only with respect to the 15 taxi accounts in connection with underwriting losses. Each of these 15 accounts was referred to AmTrust with the exception of Bay Area Metro in the 2014-2015 policy year, for which AmTrust seeks damages only with respect to the alleged ISO 98 error.
60. The Court first discusses the relevant formulas and procedures for calculating premiums and then turns to the underwriting process as a whole. It then addresses each of the claimed errors.
A. ISO Rules and Rates
61. By law, each insurance company seeking to issue insurance in a particular state must be authorized in that state and file with the state the underwriting and actuarial program they are going to use, including the rates and algorithms. See Tr. at 592:25-593:18, 631:15-632:17. Those rates may be based on the insurer's own actuarial model that it reviews with the state or based on models and formulas provided by the industry organization, ISO.
62. The benefit of using ISO is that an insurer does not have to recreate an actuarial model and may rely on ISO's information-gathering. ISO collects experience from various insureds and provides information, including rates and forms, based on that experience. Id. at 528:21-530:6, 631:15-632:17. Specifically, it provides information on risk classifications, "rates" for each classification, and the general rules by which to calculate premiums for each classification. It also provides industry averages that insurance companies can reference when they negotiate with individual insurance departments within each state. ISO files with each individual state, but these filings have no effect until an insurance company decides to adopt that filing and it is accepted by the state. ISO employs teams of actuaries who perform these aggregate analyses. Id. at 631:15-632:7. ISO's rating methodology for calculating insurance rates and premium can be reduced to a series of formulas called rating models. See Am. Bendure Decl. ¶ 33. Different rating models are used in different circumstances.
63. An insurer who uses ISO rules and rates may accept the rules and rates as is or may negotiate with the state regarding how those rules and rates will be calculated and what calculations to use. Tr. at 631:15-632:17, 714:6-18. As the witnesses emphasized, different states follow different rules with respect to ISO. Id. at 619:1-19. Moreover, there is no such thing as "an ISO rate" in isolation. Rather, there are individual ISO rates that each insurer negotiates with each state. Id. at 631:15-632:7.
64. AmTrust was an ISO subscriber and agreed to use the ISO rules and rates as filed with each state. See, e.g., id. at 592:24-25, 631:15-632:7. That means that AmTrust would follow the ISO formulas in each state, as well as the rates for factors, such as loss costs, Increased Limits Factor (defined below), and deductible factors, all of which are set by ISO actuaries for each state and then used as inputs for rating algorithms. Am. Bendure Decl. ¶ 24.
65. The errors in this case are alleged to have arisen in the calculation of the premium required for a taxi account with a deductible, and specifically in the rules that were applied to calculate that premium. There is no dispute that KF&B accurately used the rates as set by ISO for each state that served as the input for the rating algorithm. The dispute is as to the rule (i.e., the algorithm) itself.
66. The calculation of premium for a deductible taxi account under ISO is a function of several different factors that seek to set a baseline for a "basic" account in that particular state or territory and then adjust that baseline for the size of the account, any deductible associated with it, and its individual risk.
67. The building block in calculating the premium is the "basic rate." The basic rate calculates a per unit (i.e., per vehicle) cost of insurance assuming a policy limit of $100,000 without a deductible. It is the product of two components: loss costs and the loss cost multiplier. Tr. at 21:4-14; Am. Bendure Decl. ¶ 29. The "loss costs" or "basic limits loss costs" is the ISO figure for the unit cost in a particular state and territory for a taxi on a policy, assuming a policy limit of $100,000. Tr. at 20:21-21:3; Am. Bendure Decl. ¶ 29; Am. Howery Decl. ¶ 73. The "loss cost multiplier" is a unique figure that each insurance company files with the state which adjusts the "loss costs" for factors, such as insurance company expenses unique to the particular insurer. Am. Howery Decl. ¶¶ 73-75.
68. After a basic rate is calculated, the formula must take into account the effect of a deductible and the size of the program. A "deductible factor," promulgated by ISO, is applied to account for the fact that the unit cost (or risk) for a $100,000 policy that contains a deductible will, of course, be less than the unit cost for a policy that does not contain a deductible. In simple terms, the loss must be greater on a deductible account before it is absorbed by the insurer.
69. The Incremental Factor ("IF") is a separate factor published by ISO that accounts for the size of the program, in particular, the incremental effect of a deductible on the increased limits of loss of a policy over $100,000. The incremental impact of a deductible varies and decreases with every additional dollar of coverage over $100,000. That is, with each additional dollar of coverage, there is less deductible to cover an anticipated loss. The IF is intended to measure and calculate that effect.
70. The IF is different from a separate ISO value called the Increased Limits Factor ("ILF") which is used when calculating premiums for non-deductible accounts above $100,00 to account for—as the title suggests—the increased limits of loss.
71. The IF is always one digit less than the ILF. For example, if the ILF is 1.28, then the IF is 0.28. Tr. at 39:14-23.
72. Two other factors are used to calculate the premium on a taxi account with a deductible. The "public fleet factor" is an ISO figure used for taxis (i.e., public fleets). In the Program, the factor was routinely 1.1. Am. Howery Decl. ¶ 76.
73. The "total modification factor" accounts for relevant facts idiosyncratic to the individual insured account. As noted above, the "loss costs multiplier" takes into account factors and experience idiosyncratic to the insurer. The total modification factor, by contrast, focuses on the insured. The total modification factor, in turn, is a product of three other modification factors: the schedule factor, the experience factor, and the expense reduction factor.
74. The schedule factor involves modification of manual rates either upward (i.e., schedule debits) or downward (i.e., schedule credits) by the underwriter to reflect the individual risk characteristics of the insured. For example, an insured with good management (i.e., cooperation with the insurance company and with insurer recommendations), demonstrated training and supervision of employees, good equipment, and demonstrated commitment to safety through meetings and accident reports and records might receive a credit. An insured without those characteristics might receive a debit. The range of schedule credits or debits that can be issued (and in which category of risk) is limited by ISO on a state-by-state basis. Tr. at 580:17-582:17. The range is a percentage that is then converted to a numerical value. A schedule factor below 1 will reduce the premium. A schedule factor above 1 will increase it.
75. The experience factor, also known as the experience modification factor ("EMF"), measures the difference between the insured's actual past experience and the expected or actual experience of the class of risk. An EMF of 1 would not result in any change to the policy premium. Am. Howery Decl. ¶ 88. An EMF above 1 would operate as a debit, increasing the premium for an account and an EMF below 1 would operate as a credit, decreasing the premium for an account. Id.
76. The expense reduction factor generally provides a credit to the premium based upon reduced claim adjustment expenses through the use of a third party administrator for claims.
77. ISO requires all of these factors to be considered in rating a taxi account and determining the premium. Expressed mathematically, the rule set forth in the ISO-CLM would ordinarily require the premium for vehicle accounts with a deductible to be a function of the application of the following ISO 98 formula:
Premium (per unit) = ((Basic Rate x Deductible Factor) + (Basic Rate x Incremental Factor)) x Public Fleet Factor x Total Modification FactorAm. Bendure Decl. ¶ 34.
ISO itself does not set forth the formula. It contains a textual description of the analysis to be applied that the insurer must reduce to a formula filed with the state.
78. In plain English, the basic rate is multiplied by the deductible factor to account for the impact of the deductible as to the first $100,000 limit of loss. That product is then added to the product of the basic rate and the IF to account for the incremental effect of the deductible on the additional limits of loss. The sum of those two figures is then multiplied by both the public fleet factor and the total modification factor to arrive at the rate or premium.
79. To calculate the premium for vehicle accounts without a deductible, ISO requires application of a different formula:
Premium (per unit) = Basic Rate x Increased Limits Factor x Public Fleet Factor x Total Modification FactorId. ¶ 33. Neither the deductible factor nor the IF is used. Rather the basic rate is simply adjusted for the fact that there are increased limits of loss in the policy.
80. The first error alleged by AmTrust inheres in the fact that between 2011 and 2014, the Program used something in between the two formulas to calculate premium for taxi accounts with a deductible (the "2011-2014 Rating Model"):
Premium (per unit) = Basic Rate x Deductible Factor x Increased Limits Factor x Public Fleet Factor x Total Modification FactorBendure Decl. ¶ 101. As will be discussed further below, the 2011-2014 Rating Model thus made two errors with respect to this formula: it used an ILF for deductible taxi accounts (rather than the IF), and it failed to adjust the basic rate by the IF.
81. The result of this formula was to calculate a lower premium than that which would be derived from application of the ISO formula.
B. The Underwriting Process
1. The Calculation of the Premium
82. As noted above, the calculation of the projected premium was the first of two parts of the underwriting process designed to lead—in the case of taxi accounts—to a determination whether to recommend an account to AmTrust. As accounts were renewed on an annual basis, the underwriting process was conducted every year for each account.
83. The first component—calculating the projected premium—required KF&B to use information provided by AmTrust as well as information developed by KF&B itself. The basic rate, deductible factor, IF, ILF, and algorithm all came from AmTrust. Although, as discussed below, KF&B initially had developed its own algorithm, in the execution of the Program, it employed the algorithm used by AmTrust. For its part, KF&B identified and calculated the factors that were unique to the particular account requesting to be written and to make a recommendation to AmTrust based on those factors. In essence, KF&B measured and made determinations with respect to the components of the total modification factor: the schedule factor, the experience factor, and the expense reduction factor, and then input that information to the formula it based off of the AmTrust formula to arrive at the premium. The schedule factor required the greatest underwriter judgment.
84. In this process, KF&B, after having been approached by the broker with an account to consider insuring, would collect information about the prospective insured required to be collected by the Underwriting Guidelines, including basic biographical information, such as the insured's federal identification number, name, and address, the insured's operations, equipment and maintenance information, safety information, limits of liability, coverages, deductibles, and vehicle information. This information would go, in part, towards the schedule factor.
85. KF&B also would collect the loss history provided by either the insured or the broker. The loss history would generally include the current year and the prior years, including the number of claims, the indemnification paid on the claims, and information regarding the cost of handling the claims. This information would go both towards the EMF and the loss rating analysis, discussed later in this opinion.
86. KF&B then sent the submission documents to an office manager at KF&B to upload the documents onto KF&B's system.
87. To begin the underwriting process, KF&B would enter information about the insured into an Excel spreadsheet that Howery, the lead underwriter at KF&B for large fleet taxi business, prepared and provided to AmTrust for its review prior to AmTrust approving a taxi account for binding (the "Referral Spreadsheet"). The Referral Spreadsheet was based on a similar document which KF&B used with Praetorian. Am. Howery Decl. ¶ 45; Tr. at 69:15-70:4, 71:7-11.
88. The Referral Spreadsheet contained Excel formulas that would automatically generate a premium based on information provided by the insured. Certain variables in those formulas—such as the loss cost, deductible factor, ILF, IF, and public fleet factor—however, were based on the ISO rates that AmTrust had adopted with each state. To obtain the ISO rates for these factors, KF&B used AmTrust's rating system (the "CPP"), which was housed on the AmTrust system.
89. KF&B would enter the insured's account information into the CPP system, which generated a printable document called a "ratings worksheet" that showed the ISO rates to be used for that particular account. For example, based on the zip code of an account, CPP would generate a territory code for that account. That territory code would be associated with a specific loss cost for that territory based on ISO, and the loss cost would be used to calculate the "basic rate." Am. Howery Decl. ¶¶ 71, 73. That ratings worksheet also showed the calculated premium per unit (before application of the total modification factor).
90. Next, KF&B entered the applicable ISO rates from the CPP's ratings worksheet into its Referral Spreadsheet. KF&B also entered the loss cost multiplier developed by AmTrust used to calculate the basic rate.
91. Under the 2011-2014 Rating Model, the Referral Spreadsheet automatically multiplied the basic rate by the deductible factor by the public fleet factor to equal the "basic limits premium." Id. ¶ 78. The basic limits premium was then multiplied by the increased limits factor to equal the increased limits premium.
92. The Referral Spreadsheet also calculated the total modification factor, which as previously stated, is based on three factors: EMF, schedule factor, and expense reduction factor.
Plaintiffs do not submit that the expense reduction factor contained mistakes.
93. The Referral Spreadsheet automatically calculates the EMF based on the insured's account information and on the ISO rates adopted by AmTrust. Specifically, it is driven by the basic limits premium (i.e., basic rate multiplied by the deductible factor multiplied by the public fleet factor) and calculated based upon the insured's prior three years of experience, including information such as historical annual units (i.e., the number of vehicles) and losses and expenses. Calculation of the EMF also involved an actuarial determination called a "loss development factor." Id.; see, e.g., D-Howery-23. All loss development factors used in the Referral Spreadsheet were taken from ISO and/or given to Howery by AmTrust's actuaries. Am. Howery Decl. ¶ 81; Tr. at 103:11-18. The EMF was shown both on the "Rating" and the "Experience" tabs of the Referral Spreadsheet.
94. As of this step—and before calculating the schedule factor or expense reduction factor—the Referral Spreadsheet was performing calculations based on the insured's loss history and on ISO values that came from AmTrust's CPP system. Am. Howery Decl. ¶¶ 74-78; Tr. at 77:1-20.
95. The end result of this analysis would be the premium that AmTrust would be permitted to charge under the formulas it had negotiated with each respective state.
2. The Determination of Whether to Recommend An Account
96. The calculation of the premium was only the first, and more mechanical part, of KF&B's work in connection with whether to recommend a taxi account. As noted, the formula and many of the values it used in the first part came from AmTrust and were required to be consistent with filings AmTrust made with the states. Tr. at 104:3-23. KF&B had the right to exercise judgment with respect to certain aspects of the total modification factor, including the schedule factor, but even there its judgment was constrained by AmTrust's agreement with the states.
97. The second, and more subjective, part of KF&B's job was its analysis of whether at the prescribed premiums the account would generate a profit, and its work, if the calculated premium was not acceptable to the insured, to generate an alternative calculated premium based on modifications to the values in the formula it was permitted to make. Id. at 104:3-23.
98. The work on the second part was driven by KF&B's "loss rating" analysis. Am. Howery Decl. ¶¶ 91-92. In the loss rating process, KF&B would generate a figure for the average expected loss amount per unit (e.g., per vehicle in the account being insured) in order to determine whether AmTrust should underwrite the account. Id. ¶ 92. KF&B would trend and develop the insured's past losses to calculate an expected average loss amount per unit and compare that to the projected premium amount per unit to determine whether the premium would cover the expected losses. Specifically, KF&B would determine whether the ratio between the two fell within the insurer's acceptable risk limits based on the "projected loss ratio" that would be calculated off those two figures.
99. The loss rating analysis took place in the "Insurance History" and "Experience" tabs of the Referral Spreadsheet. As part of this analysis, Howery would perform various calculations in order to determine the projected loss ratio for the proposed policy year at the burn rate (i.e., the expected losses per unit). Howery generally considered a projected loss ratio equal to or below 55% as acceptable, which was lower than the MPA's acceptable loss ratio of 60%. Anything over 55% warranted further consideration according to Howery. Id.
100. The loss rating analysis allowed KF&B to determine whether the proposed premium was commensurate with the risk presented by the insured. Tr. at 150:9-17. The increased limits premium that had been calculated by the ISO factors, without the EMF and schedule rating, would be balanced against the premium suggested by the loss rating analysis. Tr. at 152:10-25. That premium might be higher than the market or than the loss rating analysis dictated the premium should be. Id. at 108:23-25.
101. Based on a review of the insured's prior loss runs and on this loss rating analysis, Howery might have concerns regarding particular losses on some of the claims. See, e.g., Am. Howery Decl. ¶ 162. His declaration provides examples of when he would request additional information from the broker on the losses to determine whether a loss was problematic. See, e.g., id. ¶ 165. For example, he might review the police report on an accident to determine whether the insured was at fault. Id.
102. The process was an iterative one, both between KF&B and AmTrust and between KF&B and the broker representing the insured. If KF&B determined that the projected loss ratio was acceptable, it would send the proposed per unit premium calculation to AmTrust. If the projected loss ratio was not acceptable, KF&B would use one or both of the two tools it had to adjust the calculated premium. It could adjust the schedule rating, based on factors unique to the insured, to reduce the calculated premium and make it more attractive to the insured. It could also consider whether the premium was generated in part by an anomalous large figure affecting the EMF, and if so, with AmTrust's approval, KF&B had discretion to override the EMF generated by the formula by manually entering a new EMF and to use one that the parties believed would sufficiently and appropriately cover the risk.
103. For example, KF&B might, with AmTrust's permission, decide to use a prior year's lower EMF rather than the calculated EMF if it felt it was a more accurate representation of how the account performed as a whole and was not inflated due to a particularly bad year in losses. See, e.g., id. ¶ 342. Howery gave as another possibility that AmTrust could permit KF&B to adjust the EMF if, for example, the ISO loss cost for a particular territory code had increased, thereby increasing the premium, and the parties wanted the per unit premium to be commensurate with the account's actual performance. See id. ¶ 282.
104. If KF&B determined the account was worth signing and the risk reasonable to pursue, Howery would send his proposed quote to the AmTrust underwriter for approval, either Hynes or Fort. He might also copy AmTrust's actuaries or Murphy. In the early years of the Program, Howery sent the Referral Spreadsheet to AmTrust by email. After October 2012, he sent the Referral Spreadsheet via AmTrust's server through file transfer protocol ("FTP") and would email AmTrust to let it know he had done so. Id. ¶¶ 65, 112-113.
105. In these emails, Howery would draw attention to certain details, such as pricing or unique characteristics of the insureds. If he had adjusted the EMF, he usually did not note the adjusted EMF in the body of the email, though the adjustment was both noted and seen in the Referral Spreadsheet on both the "Rating" tab (which calculated the per unit premium and final premium) and the "Experience" tab (which calculated the EMF). Depending on the account, AmTrust might approve the quote or there might be additional back-and-forth between KF&B and AmTrust. Because KF&B could not issue a final quote or bind a taxi account, Howery was required to get approval from AmTrust for every policy.
106. Once AmTrust approved the premium, KF&B would negotiate with the broker to assess whether the projected premium amount would be accepted by the insured. KF&B might make further adjustments if the premium was not acceptable, again after consultation with AmTrust. The end objective was to agree to a premium within a range that would be acceptable to both the insured and the insurer, AmTrust, based on the risk. Id. ¶¶ 93-96. In some instances, Howery would have to go back to the broker for additional information. The broker might also ask for a reduction in the premium or the deductible, which would then cause Howery to go back to AmTrust for its permission. KF&B might change the premium by excluding certain years of losses from the insured's loss history, changing the deductible option or the schedule factor, or removing a certain dollar value off the final per premium amount. See, e.g., id. ¶¶ 541-44, 953-59, 1441-443. In addition, Howery might also ask the broker, before sending the proposed quote to AmTrust, the upper limits of the per unit premium that the insured would be willing to accept for that policy year.
107. Howery testified that the underwriting process was a complicated process with numerous changes, and he spoke regularly with Fort or Hynes through this process. Tr. at 152:10-25. The deposition testimony from Fort is consistent with Howery's description of their interactions.
108. Fort's deposition testimony stated that he would review and analyze the documentation on the taxi accounts submitted by Howery. Fort Dep. at 24:10-22. When asked what went into his analysis, Fort testified:
A variety of different things. Looking at his pricing recommendations, loss, the loss run documentation that he provided, any details he had about the account, I would review that. The taxi accounts, the pricing I would probably work with the actuary to look at some of those from time to time.Id. at 26:12:18.
109. If Fort believed he was missing information, he would obtain that information from Howery before he authorized and released a quote. Id. at 177:3-14. When asked if he would have approved an account or a quote if he was missing information needed to complete his review of a referral, Fort responded, "If there was some missing information, I would have went back to [KF&B] before I would authorize and release any quote." Id. at 177:3-5.
110. For example, if "the loss run information did not match up with what Mike [Howery] had summarized in the pricing model," Fort "would go back to Mike" and say that it "doesn't add up, it seems there might be a discrepancy, you may want to look at it again." Id. at 39:22-40:3. In another instance, Fort alerted Howery to a formula error in one of the Referral Spreadsheets with respect to the 2015-2016 policy period for Bay Area Metro. D-Howery-605. He noticed that one of the cells in the "Insurance History" tab referenced the wrong cell in its formula. Id.
111. Fort also testified that he would sometimes work with AmTrust's actuaries to review the pricing. They would look for:
Variety of different things, whether the factors that made sense that was in his model for this particular account. The projected premium that Mike was recommending, did it make sense, the collateral amount make sense. Those are the type of things that we would be looking for. . . Actuarial trend factors on losses, claim accounts, those type of things.Fort Dep. 28:9-15.
112. The "purpose of actuarial reviewing projected premium with respect to a taxi referral" was:
[T]o be certain, I want to have some kind of level of confidence that the premium that Mike [Howery] was recommending was going to be an adequate premium for that particular risk.Id. at 36:15-23.
113. KF&B also called Hynes to testify at trial. His declaration submitted as his direct testimony recounted a similar experience with KF&B.
114. His declaration stated that he had "positive dealings with KF&B throughout my career," that KF&B has "a good reputation in the industry," and that KF&B had "good underwriters and Mike Howery in particular is very knowledgeable." Hynes Decl. ¶ 29.
115. Hynes said that when he received a referral from KF&B, he would "review the account, review the losses and loss ratios, and review what the exposures were." Id. ¶ 20. Depending on the kind of risk, he would "also look for other indicators of a well run account," id. ¶ 20, because although the "program underwriting guidelines were a good start for determining whether a risk was right for the program," there were "some accounts which did not fit perfectly within the general guidelines [that] could still be quite profitable," id. ¶ 23. As a result, he would "try to review these accounts and assess the risk." Id. ¶ 24. Hynes stated that when reviewing:
Sometimes my assessment was that the risk was not right under any circumstances and decline approval. Other times, I would make modifications to the insurance quote or conditions to account for the increased risk presented. It all depends on the account.Id. ¶¶ 24-26, 30.
For me, the primary questions were does this risk make sense and are we getting the right price for it.
Assessing risks and making modifications of this nature is extremely subjective and I did so based on my many years of experience, both with the insurers I worked at before AmTrust, and also during my time as VP of Underwriting.
. . . .
The very nature of underwriting is that it is a judgment call—insurance is about trying to hedge against uncertainty, and there is no guarantee of success for any insurer, no matter how well they underwrite their risks.
116. When deciding whether to approve KF&B's referral, Hynes stated that:
If I approved a referral, it was because I had all the information I needed to make that decision and also because I believed that the risk had the potential to be profitable for the program.Id. ¶¶ 27-28 (emphasis added).
If I did not have enough information, or if I was not comfortable with the risk, I would decline.
117. Notably, AmTrust chose not to examine Hynes or to elicit from him any testimony regarding Underwriting Guidelines or MPA violations allegedly committed by KF&B. Nor did AmTrust present any of AmTrust's other underwriters or actuaries that worked closely with KF&B in the Program to testify regarding violations by KF&B or information that they believed KF&B misrepresented or failed to provide them.
118. If AmTrust gave approval for a quote, Howery would issue a quote to the broker. If the broker accepted the quote, the broker would give Howery permission to bind the policy, and Howery would work with the broker to get an updated vehicle list. Am. Howery Decl. ¶¶ 116-117. Howery would get final information on the vehicle count and types of vehicle and input them into CPP. Id. ¶ 118; Tr. at 167:25-168:7. Howery would then bind the account, CPP would issue a PDF of the policy, and Howery would save the final Referral Spreadsheet. Am. Howery Decl. ¶¶ 119-121. Howery would generally send a copy of the materials to the broker and AmTrust's third-party claims administration North American Risk Services ("NARS") so that NARS could set up a claim file and claim fund agreement. Id. ¶ 122. The underwriting process for that policy would end by Howery sending a binder of the documentation to the broker.
C. ISO 98
119. Plaintiffs allege that the premium generated by the Referral Spreadsheet and used to underwrite and bind the accounts failed to comply with the "current" rules of ISO and that, as a result, KF&B breached its contractual duties and is liable in indemnification without regard to fault. The claim turns upon the language of the MPA and Underwriting Guidelines and on the application of ISO 98 in calculating the premium for policies. Plaintiffs further allege that when both parties discovered the error in 2014, and KF&B was instructed to correct it, KF&B made certain manual overrides to the EMF in its formula that failed to comply with ISO.
1. Genesis and Discovery of the 2011-2014 Rating Model Error
120. The basic limits premium and the increased limits premium in the Referral Spreadsheet were based on the formula provided by AmTrust and in the CPP. AmTrust later concluded and KF&B agreed that both AmTrust's system and KFB's Referral Spreadsheet contained an incorrect algorithm for deductible taxi accounts. The algorithm used the ILF applicable to non-deductible accounts rather than the IF applicable to deductible accounts and it failed to adjust the basic rate by the IF.
121. The error in KF&B's system had its genesis with AmTrust and the CPP system. KF&B adapted and modified the Referral Spreadsheet to reflect the values KF&B was receiving from AmTrust's CPP system after the Program was moved from Praetorian to AmTrust.
122. Specifically, when KF&B first introduced its program to AmTrust, KF&B already had already developed an algorithm for ISO-compliant taxi programs in connection with the Praetorian program. KF&B's program at Praetorian initially covered only limousines, used independent rates, and did not cover ISO-compliant taxi accounts, Tr. at 72:4-18, but in 2010, Praetorian asked KF&B to begin writing a small taxi business according to ISO, id. at 69:5-11, 72:4-18. Praetorian gave KF&B access to the ISO-CLM, and Howery—who was not an actuary or an expert in ISO—reviewed ISO and attempted to develop the algorithms necessary to apply mathematically the guidelines set forth in the ISO-CLM. He developed an algorithm which he believed complied with ISO 98 in an Excel spreadsheet that also contained the loss rating analysis (the "Praetorian Model"). Id. at 71:21-73:17.
123. The Praetorian Model was included in the model and rating templates KF&B presented to AmTrust before the AmTrust program commenced. AmTrust's actuaries provided KF&B with "edits and enhancements [and] additional information for the rating template to better model the loss rating portion of the risk." Id. at 174:3-8. But AmTrust did not suggest edits or enhancements to the ISO 98 portion and that portion was not discussed or modified as a result. Id.at 174:15-20. AmTrust also did not take information from the Praetorian Model and put it into their systems. Id. at 76:17-77:16.
124. The Praetorian Model ultimately developed into the Referral Spreadsheet that Howery used while underwriting at AmTrust. Howery used the Referral Spreadsheet as a "tool" that "would follow the same logic as AmTrust's system and it would kick out the new rate." Id. at 65:18-25. He testified that because AmTrust was an insurer for large taxi fleets, some of them in excess of a thousand vehicles, AmTrust's "rating would take hours to run a premium." Id. at 65:13-14. Howery testified that that time could be problematic if, for example, AmTrust ran a premium for an insured for 500 vehicles but needed to requote the same insured for 553 vehicles a few days later. Id. at 65:10-66:1, 184:21-185:14. But with the Referral Spreadsheet, he was able to run a premium on an account with hundreds or thousands of vehicles in significantly less time. Moreover, the Referral Spreadsheet permitted Howery to perform his "loss rating" analysis.
125. In 2010 or 2011, "shortly after starting to work with AmTrust" but before AmTrust had bound an account through the Program, Howery "realized that [his] numbers in [his] spreadsheet did not equal the output in their rating system. And so [he] updated [his] spreadsheet to equal the same as their system, so [his] premium would equal their premium." Id. at 79:3-7, 84:3-8. Specifically, the "Rating" tab in his spreadsheet was not calculating the same premium that came out of AmTrust's CPP system. As described above, the premium calculated by the Referral Spreadsheet should have been equal to the premium calculated by the CPP system, assuming no modifications to the EMF, schedule factor, or any other possible discretionary factors.
126. Howery essentially reverse-engineered. He determined that the reason for the discrepant results was a difference between how AmTrust calculated the premium and how KF&B calculated the premium, and he adapted the Referral Spreadsheet to track how the CPP system calculated the premium. Id. at 79:21-80:7, 81:12-25. Hence the error in the KF&B system was borne—it came from AmTrust.
127. Howery testified credibly that he followed the AmTrust formula because AmTrust had "the regulatory people, the ISO people, the statisticians, the programmers" and he "assume[d] they are correct." Id. at 82:15-23; see also id. at 82:12-23, 86:6-87:1. Howery also assumed that how KF&B had calculated the premium was incorrect, and that AmTrust was correct, because AmTrust has "a compliance department, they have a regulatory department, they report to ISO. I assume if they are doing it wrong, ISO is going to come back and tell them they are doing it wrong." Id. at 101:19-22. Regardless, he testified, he was required to follow the AmTrust formula and not his own independent views of what he thought ISO might require. Id. at 263:4-10; see also id. at 107:15-19, 178:20-179:9.
128. Importantly, there is no evidence that the Praetorian Model or Referral Spreadsheet were relied on by AmTrust in setting the ISO 98 formula that AmTrust used. Instead, as Howery testified, the Referral Spreadsheet was his own "underwriting tool to do the loss rating" and "that piece is not on their system." Id. at 77:13-16.
129. Howery used the Referral Spreadsheet with the premium generated by it for three to four years without complaint from AmTrust or any issues arising with respect to the ISO 98 formula.
130. On June 4, 2014, however, AmTrust changed the ISO algorithm in its CPP system. See D-Howery-16 (September 4, 2014 email from Smith at AmTrust to Howery stating that "the algorithm correction was put in place effective 6/4/2014"; August 21, 2014 email from Becky Sheffler at AmTrust to Aimee Moore at AmTrust stating that "there was a change to the rating algorithm for policies with deductible liability (loaded to live 6-4-2014)"); see also Tr. at 116:19-21, 121:16-125:9.
131. AmTrust did not complain to KF&B about the algorithm or even alert KF&B to the change. It did not ascribe responsibility to KF&B for any error with respect to the algorithm.
132. Howery only found out about the change weeks later and as a result of confusion among certain insureds as to the premium they were required to pay for additional vehicles added to an account. Specifically, as a result of the change, the system generated higher premiums for additional vehicles for insureds who had an account with AmTrust than the premiums of which they had been informed and that were generated with the old algorithm. Tr. at 122:3-20; see D-Howery-16 (Smith noting the new algorithm resulted in "premium discrepancies on some recently quoted policies as well as in-force policies").
133. Three weeks after the change, on June 30, 2014, Howery—who was unaware of the change—emailed systems support at AmTrust that he was trying to bind a quote but that the premium was being calculated at $5,208 per unit "when it should be $3,894 per unit." PX-043. He did not know the reason for the discrepancy. Unbeknownst to Howery, the difference was as a result of the algorithm change. Howery asked AmTrust systems support to check the formula. On July 2, 2014, systems support responded and informed Howery that there was a change to the rating algorithm. It stated that the algorithm "was previously wrong and it is now correct." Id. Howery did not previously know of the change to the algorithm. Tr. at 116:19-21, 121:16-125:9.
134. Howery forwarded the email chain to Fort with the note, "FYI, I'm going to have to do some manual adjustments to make premiums balance." PX-0718. There is no response from Fort in the email thread.
135. Separately, Howery responded to AmTrust systems support to ask, "Any chance I can get the algorithm (perhaps in Excel) so I can match my underwriting worksheet?" PX-043. Systems support responded that it would ask IT and get back to him.
136. On August 22, 2014, Fort received an internal AmTrust email about an "algorithm change that was made in the system on 6/4" that was changing the calculated premium. D-Howery-14. The email stated that AmTrust's testing unit managers would be asked to provide more information on the algorithm change.
137. On August 27, 2014, Fort responded and stated that he had "checked the [ISO] CLM regarding the deductible" and he did not see where the algorithm had changed in the way that the testing unit manager had described. He included "an example noted in the ISO-CLM to illustrate how the deductible factors [are] calculated and used." Id.
138. On August 29, 2014, Fort received a reply that the new algorithm now followed ISO 98 and informing him of the factors that were missing from the old formula. Id.
139. On September 2, 2014, Fort forwarded the email chain to Howery stating that he wanted to give him a "heads up on this" and that he was "scheduled to meet with others" to discuss. Id. He noted that "it appears there has been a change in our system algorithm [a]s to how the deductible credit is calculated" and that "how it was calculated before was incorrect in our system." Id. He further stated, "This calculation method should also be updated in your referral model." Id.
140. Fort's language is telling. He described the error as being in AmTrust's system, not KF&B's. He directed Howery and KF&B to use the formula AmTrust was using. He did not ask KF&B for its views regarding what ISO required. Nor did he give KF&B discretion to use a different formula. The email did not blame KF&B for the incorrect algorithm or suggest that the mistake came from KF&B. His email reflected an understanding that the formula came from AmTrust and that the current ISO rules that KF&B was required to follow were what AmTrust deemed to be the current ISO rules, without questioning or second-guessing.
141. Howery responded the same day, thanking Fort for the update, saying that he "absolutely agree[s] with the correction to the algorithm and ha[s] changed [his] referral rating sheet to match." D-Howery-15. Howery added that the "problem is the change was made on 6/4 and as noted should have been done on a future effective basis." Id. Howery's testimony too reflected the shared understanding with Fort that the correct algorithm was the responsibility of AmTrust, which filed with the state, and that KF&B's responsibility was to follow AmTrust's directions and the ISO rules AmTrust followed. He testified that when AmTrust told him about the error, he "assumed they were right," Tr. at 106:7, because he "agree[s] with the carrier. If they say they're wrong, if they've reviewed it and they've reviewed ISO and they're changing it, then [he] agree[s] with them," id. at 106:17-19.
2. EMF Override
142. The change in algorithm created immediate issues for both AmTrust and KF&B. KF&B had difficulty matching "the premiums generated at the time of renewal to any negotiated range within the insured's broker" after the change because "the revised formula generated much higher premiums from the prior year's premium on an otherwise unchanged account." Am. Howery Decl. ¶ 100. As a result, for some accounts, KF&B manually overrode the calculated EMF—provided that the premium with the revised EMF still projected a profitable loss ratio. Id. Howery testified he did this "to keep the business and with AmTrust's express understanding and approval." Id. Plaintiffs argue that KF&B did not alert AmTrust to these manual overrides and that they were not consistent with the MPA and Underwriting Guidelines.
143. Howery admitted that AmTrust did not give him a global endorsement to override the calculated EMF. Rather, he testified and contends that he was given such consent on an account-by-account basis based on emails and telephone conversations with AmTrust. Tr. at 213:12-215:6. He also testified he would document the EMF he was directed to use in the "Experience" tab of his Referral Spreadsheet, which was sent to AmTrust prior to approval of a referral and which reflected both the calculated EMF and the EMF that was utilized as a result of the manual override. Id. at 215:4-6.
144. AmTrust initially alleged, through its expert Scruggs, that there were "at least 22 different accounts" both before and after the ISO 98 change where there were manual overrides. Am. Scruggs Decl. ¶ 61. That figure was always an error. Scruggs's declaration refers to 22 Referral Spreadsheets and not 22 accounts. The 22 Referral Spreadsheets correspond to 12 different accounts, not 22 accounts. There are multiple Referral Spreadsheets for each account, corresponding to different policy years. Moreover, the number of accounts was later narrowed at trial to cover only taxi accounts and only those policy years after the ISO 98 change. Thus, Plaintiffs now submit that a total of eight taxi accounts—covering nine policy periods after the ISO 98 change—involved EMF manual overrides.
Plaintiffs seek damages for the following accounts and policy periods that were also in Scruggs' declaration: AAA (2015-2016), D-Howery-105; ASC (2015-2016), D-Howery-889; City Service (2015-2016), D-Howery-273; SDTLA (2014-2015), D-Howery-1182; UITD (2014-2015; 2015-2016), D-Howery-169, D-Howery-195. Plaintiffs also seek damages for the following accounts that were not listed by Scruggs as having had an EMF manual override: MKBS (2015-2016), D-Howery-707; San Gabriel Transit, Inc. (2014-2015), D-Howery-1344; G&S Transit Management, Inc. (2014-2015), D-Howery-1344.
145. The documentary record admitted in evidence reflects that in almost every instance, AmTrust—through Fort or others— in approving the quote also approved the EMF override. AmTrust did so by stating that it reviewed the "loss rating" analysis, or by acknowledging that it reviewed the "materials" or specifically the Referral Spreadsheet, which was either attached to the email from KF&B to AmTrust or uploaded to AmTrust's system via FTP. Although for certain accounts and policy years, AmTrust approved a quote in email without reference to the materials, the evidence does not contradict Howery's testimony that he always received approval. There is no reason to believe that AmTrust limited itself to the highlights of the email or did not review the materials sent it by KF&B when seeking approval. The Court also credits Howery's testimony that he had "numerous conversations" with Fort "by telephone," and that such changes were done on an account-by-account basis. Tr. at 214:4-215:6. Significantly, AmTrust failed to offer a single fact witness to testify that AmTrust was not informed of a manual override. Both AmTrust underwriters on the Program testified that they always received the information that they wanted and needed. The gap in AmTrust's proof is revealing. The Court finds that AmTrust has failed to prove that there was ever an instance in which it was not told and did not approve the manual overrides.
Kathleen Smith, a vice president of underwriting at AmTrust, submitted a declaration in which she stated that "neither I nor AmTrust authorized KF&B to perform any manual EMF overrides." Smith Decl. ¶ 11. But the Court does not credit her testimony with regard to AmTrust generally. Smith admitted that AmTrust's underwriters and actuaries on the Program, Hynes, Lamendola, and Kushner, did not report to or work for her, and provided no basis upon which she would have personal knowledge of what the underwriters and actuaries permitted or did not permit. Tr. at 512:4-13, 513:14-24. Only Fort "may have reported to me during some of th[e] time" he managed the Program but she did not describe her interactions or her role with respect to either Fort or KF&B. Tr. at 512:17-20. As the Court explains below, the MPA and Underwriting Guidelines gave the AmTrust and KF&B underwriters discretion to manually override the EMF.
146. Plaintiffs' evidence and argument at trial focused on the manual EMF override in connection with the referral of the account with the biggest losses in the Program, ASC, for the policy year of 2015-2016. The process that resulted in KF&B's recommendation of the account to AmTrust for the policy year 2015-2016 and to AmTrust's decision to accept that recommendation and instruct KF&B to bind the account began in mid-October 2014 and concluded in December 2014.
147. On October 17, 2014, Howery received the KFBA submission number and began the underwriting process for the 2015-2016 policy year. That date was after the June 2014 date when AmTrust changed the algorithm for deductible taxi accounts and the September 2014 date when KF&B was told to implement that change in the Referral Spreadsheets. Am. Howery Decl. ¶¶ 1096-99.
148. On November 5, 2014, as was his regular practice, Howery emailed Fort to let him know the Referral Spreadsheet had been uploaded to the FTP site. D-Howery-885. Based on the loss history provided by ASC, the Referral Spreadsheet calculated an EMF of 0.799 and KF&B elected to issue a schedule credit of 25%. The Referral Spreadsheet projected a per unit premium of $2,942. Am. Howery Decl. ¶ 1110; D-Howery-884. On November 6, 2014, Fort approved the quote, saying: "Since this is a large account I reviewed your pricing with actuary. You are authorized to release quote as per your suggested pricing . . . The pricing you are suggesting is where we need to be and we don't feel there is any room to drop pricing lower. That's our observation." D-Howery-886.
149. Howery issued the quote to ASC, but ASC was disappointed and indicated that it did not believe the quote accurately reflected its own performance. Am. Howery Decl. ¶ 1116. As a result, Howery had meetings with ASC to discuss steps it could take to improve its operations (and thereby reduce the premium it would be charged). Id.
150. One month later, on December 9, 2014, Howery emailed Fort to advise that he had twice met with ASC to discuss its current loss ratio situation and see if there were any areas that ASC could improve. D-Howery-888. Howery included steps that ASC was taking to improve its risk management. He concluded that "[o]ur currently proposed rate increase is 23.6%," but "[b]ased on the insured's initiative to improve operations and reduce claims, his cooperation with me and NARS, I am suggesting reducing the rate increase to 14.3% with the understanding that more will come next year if the numbers do not improve." D-Howery-890.
151. Howery also attached a revised Referral Spreadsheet that proposed reducing the per unit premium to $2,715, from the $2,942 per unit premium in the earlier Referral Spreadsheet. D-Howery-889.
152. The lower rate increase was the function, in part, of Howery's election not to use the calculated EMF of 0.799 but to use the prior year's calculated EMF of 0.654. However, at the same time, Howery issued a lower schedule credit of 15%, rather than the previous 25% schedule credit, which partially offset a decrease in the per unit premium. Howery's analysis and the choices he made were clearly reflected in the Referral Spreadsheet. The "Experience" tab of the Referral Spreadsheet showed a calculated EMF of 0.799 and, underneath that cell, a manually entered value of 0.654 with text next to it that said "2012 2013 2014 factor." Am. Howery Decl. ¶ 1119; D-Howery-889. The new schedule credit was shown on the "Schedule Rating" tab of the Referral Spreadsheet. D-Howery-889.
153. The same day, Fort responded, "I reviewed material below and the revised loss rating. . . [C]ontingent on the changes noted below [i.e., the steps ASC was taking to improve its risk management] and cooperation with you and NARS, you are authorized to release revised quote as noted. This is paper thin and I don't think we should go any lower than the revised figures." D-Howery-890. Howery issued the update quote to BB&T that day and received the final vehicle list for the binder on December 31, 2014. Am. Howery Decl. ¶¶ 1123, 1125.
154. Plaintiffs argue that AmTrust did not approve Howery's use of the manually-entered lower EMF because Howery's email does not expressly reference the lower EMF and Fort's email does not say that AmTrust approved the lower EMF. Tr. at 222:4-17. But the email chain between Fort and Howery also does not discuss other information plainly relevant to the quote and that Fort necessarily reviewed and approved, including the dollar value of the per unit premium, and there is no reason to suspect that if Fort reviewed the Referral Spreadsheet for the premium, he would not also review the Referral Spreadsheet for the components that resulted in the premium, including the EMF, and approve the use of those components. That information was plainly disclosed to Fort and to AmTrust. Indeed, Fort's December email states that he reviewed the "revised loss rating," which, Fort testified at his deposition, also required an analysis of the rest of the Referral Spreadsheet which contained the information about the EMF. Specifically, when reviewing the Referral Spreadsheet, Fort would "[s]ee if the loss run information matched up with what [Howery] had input or summarized in the pricing model." Fort Dep. at 39:14-21. In other words, Fort looked at the full pricing model.
155. Even without Fort's reference to reviewing the loss rating in the email, Fort testified that when "assessing whether to authorize Mike to release a quote for the taxi program," he would review always the "pricing model," in addition to "the loss runs that [Howery] submitted with the referral, the application, any other documentation he had about the account." Id. at 39:7-13.
156. The Court concludes and finds that AmTrust knew of and approved the use of the revised EMF. Howery testified that he always had either an oral or a written conversation in which the revised EMF was approved and AmTrust has offered no evidence to contradict that testimony. Tr. at 246:14-247:1. The correspondence regarding ASC supports Howery's testimony and his recollection. Indeed, to take Plaintiffs' argument to its logical conclusion would mean that Fort did not approve anything other than the quoted percentage increase to the per unit premium.
157. The Court further finds that the only evidence offered by Plaintiffs to attempt to refute Howery, the expert testimony of Scruggs, is unreliable. In some instances, Scruggs incorrectly cites to spreadsheets that he claims show manual overrides to the EMF when they show no such thing. In other instances, while the spreadsheets show a modification of the EMF, that modification is not due to a manual override but rather is due to updated information that led the algorithm to calculate a new EMF.
See Beverly Hills for 2015-2016 (PX-193 [KFB_00467237], "Rating 1," "Rating 2," and "Experience" tabs).
See LA Checker for 2014-2015 (compare D-Howery-723, and D-Howery-740, and Am. Howery Decl. ¶¶ 953-59, and D-Howery-741, with PX-206 [ANA00411849]).
158. There are a few examples where there is no documentary evidence that KF&B sent the revised Referral Spreadsheet with the new EMF to AmTrust for approval or where KF&B in its papers to this Court cites the wrong Referral Spreadsheet. However, these examples are few and far between, and are easily explained by Howery's testimony that he had regular conversations with AmTrust by telephone. Tr. at 214:4-8. In the majority of instances, KF&B sent the Referral Spreadsheet to AmTrust's underwriters, and sometimes its actuaries, for approval through email or uploaded by FTP, and that spreadsheet showed both the generated EMF and the manually entered EMF and clearly indicated which one had been selected.
See AAA for 2014-2015 (Am. Howery Decl. ¶¶ 263-68; compare D-Howery-81, and D-Howery-89, with PX-217 [KFB_00455319]); AAA for 2015-2016 (Am. Howery Decl. ¶¶ 289-91; D-Howery-105, D-Howery-106, D-Howery-109; compare D-Howery-117, with PX-221 [KFB_00462965]); SDTLA for 2015-2016 (Am. Howery Decl. ¶¶ 1472-79; compare D-Howery-1233, with PX-195 [KFB_00470844]).
See MKBS for 2015-2016 (Am. Howery Decl. ¶¶ 908-10; D-Howery-707; D-Howery-711).
See City Service for 2015-2016 (compare D-Howery-273, and D-Howery-274, with PX-194 [KFB_00468478]); Irving Holdings for 2015-2016 (compare D-Howery-409, and D-Howery-414, with PX-207 [ANA00416586]); UITD for 2014-2015 (compare D-Howery-169, and, D-Howery-171, and D-Howery-177, with PX-220 [KFB_00462919]); UITD for 2015-2016 (compare D-Howery-195, and D-Howery-198, and D-Howery-199, with PX-196 [KFB_00471402]); SDTLA for 2014-2015 (D-Howery-1181; D-Howery-1182); San Gabriel and G&S Transit for 2014-2015 (D-Howery-1344; D-Howery-1345).
159. In the end, the Court finds credible the testimony and evidence that KF&B obtained approval for the manual overrides. Plaintiffs have not proven their case that the manual overrides were not approved.
D. ALAE and EMF
160. Plaintiffs allege that KF&B committed a further error in the calculation of the premium for the taxi accounts and in particular, in the calculation of the EMF.
161. In order to accurately measure the EMF for accounts with a deductible, ISO rules require the insurer to apply the deductible against incurred losses. A deductible can reduce the indemnity that the insurer pays to the claimant on behalf of the insured. In short, the loss to the insurer on a claim is less on an account with a deductible than on an account with no deductible. That same logic, however, does not apply to so-called allocated loss adjustment expenses ("ALAE"), which are expenses incurred to third parties in connection with handling a particular claim with a particular account, such as litigation expenses. In the case of ALAE, ISO requires the EMF to be based on the full ALAE without any adjustment for the deductible. A deductible does not reduce the insurer's own costs of adjusting a claim. Thus, ISO gives the insurer no credit against ALAE for any deductible.
Expenses that are not paid to third parties, such as company overhead, are considered sunk costs and are not included in ALAE. If the expenses are not in connection with a particular claim, they are considered to be "unallocated." Tr. at 125:10-19.
162. Rule 5J.2 of the ISO Rules ("ISO 5J.2") provides:
Risks Written on a Deductible Basis. For risks whose operations are to be written on a deductible basis, the losses to be included in the rating shall be the total of the amounts calculated in Paragraphs F.1 and F.2, except that incurred losses experienced under full coverage shall be reduced to an equivalent deductible amount by subtracting the deductible amount from the indemnity payment prior to applying the MSL limitation. Allocated claim adjustment expenses shall be included in full.D-Howery-13a (emphasis added). The two referenced paragraphs direct the insurer how to calculate indemnity losses. Paragraph F.1 requires the losses in the rating to include paid and outstanding losses (including allocated claim expense) with the indemnity amount limited to basic limits and with the indemnity amount and allocated claim expense resulting from any single occurrence limited by a figure called the maximum single loss ("MSL") value (specified elsewhere by ISO) based on the company subject loss cost. Id. Paragraph F.2 requires adjustments to reflect the ultimate level of losses for each year in the experience period in certain situations. Id. In plain English, the incurred losses are to be reduced by the deductible but the ALAE is not. The rule is silent with respect to the ultimate projected losses that have not yet been incurred.
163. KF&B calculated the EMF based on the loss history submitted by the insured in its initial application and its renewal application each year thereafter, which KF&B then entered into the Referral Spreadsheet. KF&B listed this information on the tab "Losses" of the Referral Spreadsheet and across columns, including "DOL" (i.e., date of each loss), "indemnity paid," "indemnity reserve," "ALAE paid," and "ALAE reserved." Many times, the loss runs submitted by the insured contained values for both losses and ALAE. However, on some occasions, the loss runs would not contain a value for ALAE.
164. It is undisputed that when the loss runs included a figure for ALAE, KF&B included the ALAE "in full" as specified by ISO 5J.2, without reducing it for the effect of the deductible. That calculation complied with ISO. Tr. at 125:24-126:4.
165. However, when the loss runs did not include a figure for ALAE, KF&B did not include any amount for ALAE and applied the deductible to the full amount listed for losses.
166. Howery did not ask AmTrust actuaries what to do when ALAE was not broken out. He opined that he "would think, as the AmTrust actuaries are . . . looking at [his] data," that if "they'd seen years with no ALAE, they might have suggested [he] do something," but he did not have "anything directing [him] to do when there's no ALAE." Id. at 137:1-4. Howery consulted the ISO-CLM, but it did not tell him to treat ALAE and losses when an insured did not provide data on each. Id. at 128:7-23. As a result, Howery followed what he believed to be the ISO rule—he included ALAE in full when there was ALAE but when there was no ALAE reported, he applied the deductible to the indemnity loss.
167. AmTrust argues that, in those instances, KF&B violated ISO, the Underwriting Guidelines, and the MPA by not separately breaking out a figure for ALAE when there was no ALAE reported. Its argument proceeds in two parts.
168. First, AmTrust argues that where the loss history did not include a separate figure for ALAE, ISO required KF&B to make an adjustment for the lack of data (for example, by using average ALAE rates or loss elimination ratios) or to alert AmTrust to what AmTrust claims was the missing information. Plaintiffs point to an ISO circular on taxes and expense analysis for companies. That ISO circular provided an industry average of the ratio of ALAE to losses incurred, around 20%. Id. at 240:1-21.
169. Plaintiffs state that KF&B also could have excluded the data altogether. On cross-examination of KF&B's expert, Michael Dubin, Plaintiffs pointed to ISO Rule 4B ("ISO 4B") entitled "Other Company's Experience and Self-Insured Experience." That rule states, "If the risk has been self-insured or insured from which the experience is not obtainable, the experience may be used if submitted to the company in the form of a statement signed by the insured. Experience in such form may be excluded from the rating if its reliability cannot be established." PX-323. Plaintiffs' counsel tried to make the point that if the figure given by the insured did not contain a separate number for ALAE, KF&B could have decided to ignore the data and use an EMF of 1. Defense expert Dubin testified that ISO 4B was available in certain situations, but that rule was limited, required a signed statement by the insured, and was applied only when the data was determined to be unavailable and unreliable. Tr. at 742:23-745:12.
170. Plaintiffs' expert, Scruggs, also alluded to an ISO rule that he claimed would have permitted KF&B simply to ignore years where no ALAE was reported. See id. at 708:7-12. But Plaintiffs' counsel clarified in closing that this statement was a reference to ISO Rule 4A ("ISO 4A"), which provides no such thing. ISO 4A states, "The experience modification shall be determined from the latest available three years' experience incurred by the company establishing the rating in this state, or in all states, for the forms of insurance to be rated. In the event the experience for the full experience period is not available, at least one completed policy year shall be used." Id. at 803:25-804:10; see PX-323. It thus applies only when there is no experience data for a policy period and does not permit the underwriter to selectively ignore experience data because there is no ALAE recorded.
171. Second, AmTrust argues that even where the ALAE figure was listed as zero, there were circumstances where the insured necessarily must have incurred ALAE and KF&B would have known of that fact and should have tried to break out ALAE from the insured's reported loss history. For example, during cross-examination, Plaintiffs' counsel showed Howery the Referral Spreadsheet for ASC for its 2015-2016 policy year. Tab 258. The "Experience" tab listed the loss history for ASC for the three prior years. ASC had, in the prior year, ALAE of $342,518 and losses of $1,151,483; in the second prior year, ALAE of $617,503 and losses of $1,678,490; and in the third prior year, ALAE of $0 and losses of $1,560,185. Id.; see Tr. at 196:14-198:20. Plaintiff's counsel questioned how Howery could assume, when calculating the EMF for the upcoming policy year, that ALAE was $0 in the third prior year even though it had been almost 30% of the total loss amount in the prior and second prior years. Tr. at 196:14-198:20.
172. AmTrust also relies on Howery's testimony, under cross-examination, that although "[t]he vast majority of claims in this taxi world are small and they don't have ALAE," one is "typically only going to have ALAE when you have larger, more complicated claims," id. at 130:20-25, so that if "there's a very large account, yes, there's probably ALAE in there," id. at 135:21-23.
173. Specifically, AmTrust claims that by failing to apply the industry average where no ALAE was listed, KF&B understated the EMF and generated inappropriately low premiums that were deficient by $17.5 million. Am. Scruggs Decl. ¶ 41. AmTrust argues that because these premiums would have been significantly higher than what was offered, many of these accounts would not have been bound, and AmTrust would not have suffered any losses from these accounts. Id. ¶ 44.
174. For its part, KF&B has several responses. First, it argues that ISO does not permit the insurer to substitute an average figure or any other estimate for ALAE when an insured reports loss history that indicates that there has been no ALAE. According to KF&B, the calculation of the EMF under ISO 5J.2 is a mechanical process. The "incurred losses" are reduced by the deductible but not ALAE. The ALAE is included in full when the insured reports ALAE. When the insured does not report ALAE or reports ALAE of zero, no ALAE is included in the calculation. If the insurer or underwriter believes there has been a mistake or that the algorithm as calculated does not account for the full economic cost and risk to the insurer, it has three options: (1) it can use the data as-is provided by the insured; (2) it can decline to use the data pursuant to ISO 4B but only if the requirements of ISO 4B are satisfied; or (3) it can appeal to a rating committee for a change to the algorithm. What it cannot do is calculate a figure for ALAE that has not been included in the loss runs. Tr. at 763:24-765:4.
175. Second, KF&B argues that Scruggs' methodology leads to demonstrably inaccurate results. In instances where an insured has reported ALAE of zero, using the industry average would not only be wrong but would misstate EMF. In fact, KF&B argues that there is no evidence that any taxi account incurred ALAE without reporting it or that KF&B included ALAE within losses when it should have been excluded.
176. Third, KF&B's witnesses credibly testified that it was not unreasonable to assume that a reported ALAE of zero was accurate. KF&B was not required to follow options 2 or 3 above. Howery testified that a claim might be settled early without incurring any adjustment expense (because taxi claims tended to be small) or, alternatively, the insurer might handle any adjustment in-house, which would result in zero ALAE. Tr. at 247:7-16. For example, as to the ASC account, on redirect, Howery testified that even if an insured had a $1.5 million loss in one year, if they settled that claim without involving attorneys, or if the claims were handled in-house, there would be no ALAE. Id. The loss runs also might not include ALAE because the insurer did not want to disclose that figure for competitive or other reasons. In addition, Dubin testified that whether a particular expense was allocated and thus generated ALAE, or unallocated and did not generate ALAE, would be a matter of the insurer's internal accounting policies. Id. at 766:4-6.
177. The Court credits KF&B's testimony. Plaintiffs did not demonstrate that KF&B knew that there was hidden ALAE that should have been broken out. Plaintiffs did not point to an ISO rule or other provision that required KF&B to estimate the ALAE amount if not broken out. Rule 5J.2 requires that where there is ALAE it must be included "in full." It does not state that where the insured does not report ALAE, the insurance company (or its underwriter) is required (or even permitted) to estimate what the ALAE might be. Plaintiffs did not point to any provision that required such estimation. As Dubin testified, there can be "no error in using zero if zero was the actual data recorded to ISO by another company for their allocated loss adjustment expense." Id. at 635:4-7.
178. Plaintiffs did not establish that KF&B was required to apply ISO 4A or 4B to determine ALAE and failed to do so. The Court credits Dubin's testimony and the language of ISO 4B that indicates ISO 4B is applied only when the insured provides a sworn statement that the experience data is unobtainable and the insurer or underwriter determines that the experience data is unreliable. Plaintiffs have not demonstrated that Defendant or the various insureds knew that there was ALAE in the indemnity amount that was excluded in the accounts at issue. Scruggs did not point to ALAE that existed and was ignored, instead saying that it was "extremely unlikely that the ALAE for certain policies would have actually been zero." Am. Scruggs Decl. ¶ 30. Plaintiffs also did not point to any discussion between AmTrust and KF&B that directed KF&B to deviate from ISO with respect to ALAE.
E. Other Allegations
179. Plaintiffs' expert, Donald Bendure, lists a number of other account-specific errors in the underwriting process, including that the underwriting files did not contain the information required to be maintained. Prior to trial, the Court granted summary judgment against AmTrust's claim that it suffered damages as to KF&B's independent obligation to maintain the underwriting files. Dkt. No. 168. There was no evidence that any such failure to maintain a file caused damages to AmTrust in light of the undisputed evidence that AmTrust received all of the information it requested with respect to the referred taxi accounts. Id. at 7-8; see also AmTrust N. Am., Inc. v. KF&B, Inc., 2020 WL 5513458, at *2 (S.D.N.Y. Sept. 14, 2020). The Court, however, permitted (and did not exclude) the evidence with respect to the underwriting process generally and the collection of information from the insureds to the extent such evidence was relevant to AmTrust's other breach of contract claims that KF&B recommended accounts that it should not have recommended.
180. AmTrust withdrew its claim for damages based on any account-specific errors and omissions, but, in any event, Bendure's credibility as to these and certain other allegations was dismantled at trial. For example, Bendure's deposition testimony was played during trial in which he realized he missed information that KF&B had in fact submitted or reviewed when underwriting certain policies. Tr. at 353:1-17. As another example, Bendure accused KF&B of manipulating the EMF because he asserts that it used an EMF of 0.799 for ASC that was not justified by the facts. But on cross-examination, KF&B brought to his attention documentary evidence that showed that the EMF of 0.654 was ultimately used, not 0.799, thus rendering his analysis on that point incorrect. And at trial, Bendure admitted that KF&B had in fact used 0.654, not 0.799. Id. at 363:20-364:6. As a third example, Bendure listed the fact that a safety meeting was not held for a particular account as an underwriting deviation. When KF&B's counsel noted that the account was a one-driver, one-vehicle account and asked with whom the driver was supposed to hold meetings, Bendure responded, "Himself." Id. at 317:10-21. When counsel pressed, "And so because he didn't document a meeting with himself, you found that was an underwriting deviation," Bendure responded, "Not showing up was a problem." Id. at 317:22-24. Bendure was not a credible witness.
CONCLUSIONS OF LAW
I. Breach of Contract
181. A plaintiff seeking to recover for breach of contract under New York law must prove, "by a preponderance of the evidence, (1) the existence of a contract between itself and that defendant; (2) performance of the plaintiff's obligations under the contract; (3) breach of the contract by that defendant; and (4) damages to the plaintiff caused by that defendant's breach." Diesel Props S.r.l. v. Greystone Bus. Credit II LLC, 631 F.3d 42, 52 (2d Cir. 2011).
The parties do not dispute the applicability of New York law. See also MPA, Section XXIV ("This Agreement shall be interpreted and governed by the laws of the State of New York.").
182. Plaintiffs allege that KF&B breached three provisions. First, they allege KF&B breached the Underwriting Guidelines, and thus the MPA, when it failed to "use the current ISO rules and rates." Underwriting Guidelines at 2. Second, they allege KF&B breached its obligation under the MPA to perform "in accordance with the highest standards of the industry." MPA, Section II.B. Third, they allege that KF&B failed to "perform its obligation as a fiduciary of AmTrust" under the MPA. Id., Section III.
A. "Current ISO Rules and Rates"
1. ISO 98
183. In this case, analysis of Plaintiffs' claims begins and ends with a topic as to which the parties did not spend much time: contract interpretation.
184. Plaintiffs' first set of arguments depends on the Underwriting Guidelines. The seemingly simple language of the Underwriting Guidelines betrays an underlying complexity. The Underwriting Guidelines state: "The underwriter is to use the current ISO rules and rates." Plaintiffs argue that the reference to "current ISO rules and rates" requires the underwriter to calculate a premium according to the correct interpretation of the ISO rules and rates such that if it is later determined that the rules or rates it used did not comply with ISO, the underwriter would be strictly liable under the indemnification clause. Defendant claims, by contrast, that there is no correct ISO rule or rate in some absolute sense. It argues that the language of the Underwriting Guidelines requires it to use the ISO rules and rates filed by the insurance company with the state. The "current" rules and rates are whatever the rules and rates are that AmTrust files with the state. Therefore, in its view, KF&B is required to use what AmTrust represents has been filed with the state as compliant with ISO and KF&B is not required to make an independent determination of the correct ISO rules and rates. As long as it uses the figures provided to it by AmTrust, KF&B argues that it is not liable under the indemnification clause. The Court concludes that KF&B has the better interpretation.
185. Under New York law, "words and phrases in a contract should be given their plain meaning," and the contract "should be construed so as to give full meaning and effect to all of its provisions." Shaw Grp. Inc. v. Triplefine Int'l Corp., 322 F.3d 115, 121 (2d Cir. 2003) (internal citation and quotation marks omitted). A court interpreting a contract is "to give effect to the intent of the parties as revealed by the language they chose to use." Bolt Elec., Inc. v. City of New York, 223 F.3d 146, 150 (2d Cir. 2000). "Contracts 'should be examined in light of the business purposes sought to be achieved by the parties and the plain meaning of the words chosen by them to effect those purposes.'" Shipping & Fin., Ltd. v. Aneri Jewels LLC, 2019 WL 5306979, at *2 (S.D.N.Y. Oct. 21, 2019) (quoting Newmont Mines Ltd. v. Hanover Ins. Co., 784 F.2d 127, 135 (2d Cir. 1986)). They should "not be interpreted to produce a result that is absurd, commercially unreasonable or contrary to the reasonable expectations of the parties." Greenwich Capital Fin. Prod., Inc. v. Negrin, 903 N.Y.S.2d 346, 348 (1st Dep't 2010).
186. Under those rules, the better interpretation of the clause is that it expresses a shared understanding of AmTrust and KF&B that the accounts to be underwritten—and based upon which KF&B could earn a commission—were to be consistent with ISO as AmTrust and the states construed the requirements of ISO, and that KF&B was to use the ISO rules and rates as directed by AmTrust. That is, the "current ISO rules and rates" are those that AmTrust files with the state, regardless whether they reflect—in hindsight—the best interpretation and whether an alternative interpretation might be possible. The clause did not require KF&B to independently determine the algorithms or rates that ISO required, and it did not hold KF&B accountable, in indemnification or otherwise, for mistakes that AmTrust made in the algorithm or in information it provided to KF&B.
187. The Court reaches that conclusion for several reasons. First, the undisputed testimony at trial is that there is no such thing as an "ISO" rule or rate in isolation. Both Scruggs and Dubin testified that the ISO rule or rate was determined by the state's acceptance of what AmTrust submitted to be compliant with ISO. Scruggs testified that the obligation in the Underwriting Guidelines was to use the ISO rules and rates as currently filed with the state. Tr. at 713:24-714:18. When asked what it means to require KF&B to use the ISO rules and rates filed by AmTrust, Scruggs responded, "ISO files loss cost and other rating components, rates, rules, forms" and "companies who are members of ISO, once a state has approved ISO's plan, like the commercial auto plan, they can then file to accept it." Id. at 714:6-11. As a result, once an insurance company, like AmTrust, files to accept those ISO rules and rates—with or without deviation—AmTrust must use the ISO rules and rates that are filed with the state. Dubin testified that the only way that the ISO 98 formula would be applied incorrectly would be if it did not match the formula an insurance company had negotiated with an insurance department and "only AmTrust and the insurance department would know what that negotiation led to," not KF&B. Id. at 632:8-633:1.
188. Second, this interpretation is faithful to a textual analysis of the clause and its interpretation as a whole. On its face, the clause requires two things of the underwriter in parallel: (1) that the underwriter use the "current ISO . . . rates" and (2) that the underwriter use the "current ISO rules." At argument, AmTrust agreed, as it must, that the requirement to use "current ISO . . . rates" imposed on KF&B an obligation to use the ISO values for loss costs, IF, ILF, public fleet factor, and the like, generated by ISO with respect to each state and reported by AmTrust to KF&B. KF&B had neither the responsibility nor the authority to second-guess the values and factors that AmTrust required it to use. It had to use the rates given to it by AmTrust as the current ISO rates. The "current" rates meant the rates "occurring in or existing at the present time," as used by AmTrust. Merriam-Webster.com, https://www.merriam-webster.com/dictionary/current (last visited Oct. 27, 2020); see Oxford English Dictionary, https://www.oed.com/view/Entry/46097?rskey=aTKvYu&result=2 (last visited Oct. 27, 2020) ("Belonging to the current week, month, or other period of time."). "Current" thus refers to what AmTrust reports to KF&B as the current rule or rate.
189. It follows, therefore, that if the "current ISO . . . rates" mean the rates provided by AmTrust to KF&B as the ISO rates AmTrust is currently using, the "current ISO rules" must mean the rules provided by AmTrust to KF&B as the current rules that AmTrust is required to use under ISO. These terms must be construed in pari materia such that both terms are given the same meaning. See United States v. Freeman, 44 U.S. 556, 556 n. 1, 3 (1845) (provisions "are in pari materia which relate to the same person or thing, or the same class of persons or things"); In re Hawker Beechcraft, Inc., 515 B.R. 416, 430 (S.D.N.Y. 2014) ("[A]djacent statutory subsections that refer to the same subject matter should be read harmoniously."). Howery testified to essentially such an understanding without putting it in legal terms, stating that it was his "understanding to follow the ISO rules that AmTrust uses . . . the word 'current' in there, for example isn't necessarily accurate. It's up to AmTrust whether or not they adopt current ISO rules and rates. I am required to follow the directives of AmTrust. If they don't adopt the current rule or rate, I still have to use what AmTrust dictates." Tr. at 263:4-10.
190. Third, ISO is an actuarial formula. KF&B had only underwriters, not actuaries. In contrast, AmTrust had actuaries that determined the ISO rates with each state. KF&B's obligation to use the "current ISO rules and rates" was not for KF&B to interpret what those rules and rates were. It was to access that information from AmTrust's actuaries, through the CPP system.
191. Fourth, KF&B's interpretation is consistent with the MPA as a whole. As KF&B stressed at argument, the MPA not only required KF&B to comply with the Underwriting Guidelines but also required it "to produce Policies in accordance with applicable AmTrust filings and to abide by the requirements of the insurance laws and regulations of the states in which it will operate on behalf of AmTrust." MPA, Section II.B. Those filings, however, include AmTrust's agreements with the states with respect to how an ISO-compliant rate will be calculated. They represent the rules by which AmTrust has agreed to be bound in each particular state and govern such matters as to whether any particular rate or premium is discretionary. See Tr. at 631:15-633:3.
192. An interpretation that KF&B was required to use, under the Underwriting Guidelines, the rules and rates KF&B independently determined were required by ISO without regard to what AmTrust had filed with the state would put the provisions of the MPA (and the Underwriting Guidelines incorporated therein) at war with one another. KF&B would be required to use a rule or rate it believed was ISO-compliant under the Underwriting Guidelines even if the use of such rule or rate put AmTrust out of compliance with its filings with the state and KF&B therefore out of compliance with the body of the MPA. The Court is required, however, to read terms of a contract in harmony so that they are consistent with one another and not to interpret a contract to require in one provision what it is implicitly forbidden in another. See, e.g., Net2Globe Int'l, Inc. v. Time Warner Telecom of New York, 273 F. Supp. 2d 436, 445 (S.D.N.Y. 2003) (courts are directed to "interpret a contract so as to give effect to all of its provisions and 'cannot and should not accept an interpretation that ignores the interplay of the terms, renders certain terms inoperable, and creates a conflict where one need not exist'") (quoting Pearce, Urstadt, Mayer & Greer Realty Corp. v. Atrium Dev. Assocs., 571 N.E.2d 60, 64 (1991)).
193. AmTrust's responses in this regard miss the mark. It claimed in summation that the Court need not worry about Section II.B of the MPA because there is no claim in this case that AmTrust violated state regulations or that any error in ISO 98 caused it to violate its agreements with the states. It also claimed, and in the same breath, that KF&B had not proven that the original (incorrect) ISO 98 application was consistent with AmTrust's filings. It could be, AmTrust suggested, that the ratings by both AmTrust and KF&B based on the ISO 98 error were inconsistent with the state filings. But the point is not whether KF&B violated Section II.B of the MPA. It is that the Court is required to interpret the MPA so its provisions are in harmony and to avoid, if possible, an interpretation of one part of the MPA (the Underwriting Guidelines) that would permit and sometimes require KF&B to take action that could be inconsistent with another provision of the MPA (Section II.B). The way to do that, consistent with the plain language, is to interpret the Underwriting Guidelines to use the rules and rates that AmTrust asks it to use. Moreover, to the extent that there was an absence of evidence of the relation between the original (incorrect) ISO 98 application and what was in any particular state filing, that absence of proof cannot create a case on behalf of AmTrust. AmTrust, as a plaintiff, bore the burden of proof on the claim of contract breach. It is undisputed that AmTrust alone was responsible for the filings with the state, including the determination with the state as to how to apply ISO. It follows that if KF&B used what AmTrust represented as the rules and rates that KF&B should use, that KF&B cannot be liable under the Underwriting Guidelines and the indemnification obligation that follows those underwriting Guidelines.
194. Finally, while neither party argued that the agreements are ambiguous, to the extent they can be perceived as such, the interaction between the parties and communications after the 2014 change supports this reading of the parties' intent. "Where a contract term is ambiguous, we look to extrinsic evidence to determine the intention of the parties . . . includ[ing] the parties' apparent intention, what would be commercially reasonable, and 'the parties' interpretation of the contract in practice, prior to litigation.'" Matter of MPM Silicones, L.L.C., 874 F.3d 787, 796-97 (2d Cir. 2017) (internal citations omitted).
195. AmTrust has not submitted evidence that its CPP system was connected to or otherwise relied on KF&B's Referral Spreadsheet (or Praetorian Model) at any point. And when Fort was alerted in September 2014 to the June 2014 change to the algorithm in the CPP system, he told Howery that AmTrust changed the methodology and instructed him to use the new formula. Fort did not ask Howery to confirm that the old formula in the CPP system was wrong and that the new formula was now corrected. Fort did not ask Howery to check the formula in his Referral Spreadsheets. Fort did not ask Howery for his opinion or discretion on the change. Instead, Fort told Howery that "there has been a change in our system's algorithm" and "[i]t looks like how it was calculated before was incorrect in our system," and then directed Howery to modify his Referral Spreadsheet to match AmTrust's calculation method: "This calculation method should also be updated in your referral model." D-Howery-16. That interaction demonstrates that, prior to the litigation, the parties understood the clause to mean what the Court now has determined the plain language also suggests: KF&B was to obtain its information and instructions as to the current ISO rules and rates through AmTrust. If those rules and rates represented an incorrect application of ISO or there might be a different interpretation of ISO that could be used, that was a matter for AmTrust to raise in its state filings. It was not something that KF&B could unilaterally alter through its Referral Spreadsheets.
196. Notably also, the Court's conclusion does not read the language "ISO" out of the Underwriting Guidelines for several reasons. By referring to ISO, the Underwriting Guidelines reflect the shared understanding that the rules and rates to be used were ISO. That would be important to both AmTrust and KF&B: the two had a shared interest in the premiums that would be generated by any accounts to be signed and thus an interest in understanding the basis on which the premiums were to be calculated even if responsibility for negotiating and filing the formulas and rules resided with AmTrust. KF&B could not complain that the rules and rates being used were ISO rates, even if it might sign more accounts and generate more premium by—against AmTrust's desires—departing from ISO. The Underwriting Guidelines told both AmTrust and KF&B what rates to use: ISO rates. Moreover, to as those inputs such as the schedule factor which it would be KF&B's responsibility to calculate, it told KF&B where to look in order to determine the acceptable range of schedule credits or debits that could be issued: ISO.
197. Thus, the evidence does not support the claim that KF&B violated the Underwriting Guidelines by using the 2011-2014 Rating Model that contained the algorithm that AmTrust had in its CPP system. Plaintiffs also do not allege that the values of the ISO factors themselves that KF&B used were incorrect, or that KF&B did anything other than what AmTrust directed KF&B to do with respect to ISO. This claim fails.
2. EMF Override
198. KF&B's use of a figure for the EMF other than what was calculated also did not violate the MPA, but for other reasons. As discussed in detail in Section VII.B.1. of this opinion, the EMF is not a figure generated by ISO or provided by AmTrust, but its method of calculation is dictated by ISO and it incorporates ISO rates. Tr. at 211:14-20, 212:25-11, 591:11-16. The calculation of the EMF is based on the basic limits premium, the insured's prior three years of experience, and AmTrust's actuarial loss development factor.
199. As the Court found, the manual override to the EMF was clearly indicated in the Referral Spreadsheet, KF&B sent the Referral Spreadsheet to AmTrust's in-house underwriter (and in some cases, also its actuaries) either by email or FTP, and AmTrust approved the quote and analysis in that Referral Spreadsheet either by email or telephone. The evidence is clear, and the Court so finds, that the EMF was modified in order to ensure that AmTrust assess the risks and prices appropriately and to retain a preferred account, and with the shared understanding and objective that AmTrust generate a successful underwriting profit if the calculated EMF was determined not to appropriately assess risks and prices. This deviation from the ISO-compliant ECF calculation was authorized by AmTrust, as permitted under the Underwriting Guidelines.
200. The MPA and Underwriting Guidelines did not require AmTrust and KF&B to ignore what they collectively believed to be the best outcome for AmTrust, consistent with its risk tolerances. The Underwriting Guidelines were stated to be "guidelines," i.e., an "indication or outline of policy or conduct." Merriam-Webster.com, https://www.merriam-webster.com/dictionary/guidelines (last visited Oct. 27, 2020); accord The American Heritage Dictionary, https://ahdictionary.com/word/search.html?q=guidelines (last visited Oct. 27, 2020) ("A statement or other indication of policy or procedure by which to determine a course of action."). The introductory language to the Underwriting Guidelines makes clear that they are not to be applied in a rigid manner that precludes discussion and modification and at the expense of attracting accounts that would generate an underwriting profit. That language states that the Underwriting Guidelines are to be a "reference tool," that the goal is to "identify the most favorable accounts for the program by assessing their risks and pricing accordingly," and that the goal of the Underwriting Guidelines is to "guide the underwriters to select accounts having characteristics common to well run operators." Underwriting Guidelines at 1. "The ultimate goal of the [Underwriting Guidelines] manual is to provide AMTRUST FINANCIAL GROUP with preferred accounts, which have been priced to provide a successful underwriting profit." Id. To that end, the Underwriting Guidelines were stated to be "designed as a framework" and could and would be "modified from time to time." Id. This introduction makes clear that "discussion at all levels between KFB, INC. and AMTRUST FINANCIAL GROUP will be used to enhance, streamline, or make any other modifications, which will improve the results for AMTRUST FINANCIAL GROUP and enhance the program's longevity." Id.
201. The Court must "construe the [contract] in a way that 'affords a fair meaning to all of the language employed by the parties in the contract and leaves no provision without force and effect.'" Consol. Edison Co. of New York v. Allstate Ins. Co., 774 N.E.2d 687, 693 (2002) (quoting Hooper Assoc. v. AGS Computers, 548 N.E.2d 903, 905 (1989)). "An interpretation of a contract that has the effect of rendering at least one clause superfluous or meaningless . . . is not preferred and will be avoided if possible." LaSalle Bank Nat'l Ass'n v. Nomura Asset Capital Corp., 424 F.3d 195, 206 (2d Cir. 2005) (internal quotation marks and citation omitted). Here, the language of the introduction that the parties are to engage in discussion at all levels to make "modifications, which will improve the results for AMTRUST FINANCIAL GROUP" is entitled to equal dignity as the other provisions of those guidelines and the MPA. It cannot be disregarded.
202. AmTrust failed to prove that the changes to the EMF—to the extent that they can be construed to be inconsistent with any language of the Underwriting Guidelines—were other than modifications mutually agreed by the parties on a case-by-case basis to "improve the results" for AmTrust, assess the risks and price appropriately, and to ensure that AmTrust would retain a preferred account that would "provide a successful underwriting profit." Underwriting Guidelines at 1. In certain cases, the parties mutually judged wrong and the accounts did not generate a profit. But that did not make KF&B the insurer of AmTrust.
203. Finally, any argument that this interpretation is inconsistent with the so-called no-waiver clause in the MPA would be meritless. That clause that states: "The failure of AmTrust or [KF&B] to enforce any obligation under this Agreement shall not constitute a waiver of any right under the Agreement, nor shall past waiver of any provision constitute a course of conduct or a waiver in the future of the same provision." MPA, Section XXII. Under it, AmTrust does not waive by failure to enforce any of KF&B's obligations such as the obligation to produce policies in accordance with applicable AmTrust filings, to maintain its licenses, to maintain a sufficient staff of competent and trained personnel, or to obtain AmTrust approval for a number of other actions. Nor does KF&B waive by its failure to enforce any of AmTrust's obligations, such as AmTrust's obligations to compensate KF&B. But the question here is not whether AmTrust waived what otherwise would have been a requirement under the MPA. The question is whether, reading the MPA as a whole, KF&B violated any provision of the MPA by modifying the EMF—with AmTrust's consent—in order to retain an account believed to be profitable. The Court concludes that based on the plain language of the Underwriting Guidelines and the MPA that KF&B did not violate either.
3. ALAE
204. Plaintiffs have also failed to establish that KF&B's accounting for ALAE and the calculation of the EMF based on that accounting violated the Underwriting Guidelines and therefore the MPA.
205. AmTrust's claim turns, and fails, on the language of ISO. If KF&B's accounting and calculation of EMF did not violate ISO (and there is no showing that it was inconsistent with any state filing or instruction from AmTrust), it follows that KF&B cannot have violated the provision of the Underwriting Guidelines requiring that ratings be based on ISO.
206. ISO 5J.2 does not direct that an insurer or underwriter must include ALAE when the insured and the broker do not report that they have incurred any ALAE. As noted above, ISO 5J.2 on its face requires that ALAE be included in full and without any adjustment for a deductible when ALAE is reported on the loss runs. Thus, it provides:
Risks Written on a Deductible Basis. For risks whose operations are to be written on a deductible basis, the losses to be included in the rating shall be the total of the amounts calculated in Paragraphs F.1 and F.2, except that incurred losses experienced under full coverage shall be reduced to an equivalent deductible amount by subtracting the deductible amount from the indemnity payment prior to applying the MSL limitation. Allocated claim adjustment expenses shall be included in full.D-Howery-13a (emphasis added).
207. Defense expert Dubin explained, in testimony that the Court finds credible, that underwriters are to assume that the data reported by the insured is correct "whether it's zero ALAE or a number for ALAE" and thus including the data as reported by the insured would be synonymous with including the data in full. Tr. at 739:19-21, 746:22-748:10.
208. The only exceptions under ISO offered by Plaintiffs were under ISO 4A and ISO 4B. Under ISO 4A, when the experience data for the full experience period is unavailable (i.e., all three years), the experience data available for at least one policy year (if available) can be used. See id. at 804:25-10; PX-323. Under ISO 4B, when the insured submits a signed statement saying that the risk is self-insured or the experience data is not obtainable, and the insurer or underwriter finds the data unreliable, the data may be excluded and an EMF of 1 may be used. See Tr. at 742:23-745:12. If the underwriter does not believe that such rules are sufficient to address its concerns and believes that EMF is being calculated in a manner that does not capture the true risk, it is not free to disregard the rules. Its recourse is to use an appeals process to change the ISO rules. See Tr. at 763:24-765:4.
209. The Court finds that Dubin's interpretation of ISO 5J.2 is a sensible one and that the interpretation given by Plaintiffs is not only incredible but would introduce subjectivity and uncertainty into a calculation that demands certainty. ISO 5J.2 on its face calls for a mathematical calculation based on information provided to the insurer. If that information reflects that ALAE was incurred, ALAE is to be included without adjustment for a deductible. If it does not reflect that ALAE was incurred but rather includes all of the losses as incurred losses, then the insurer is not required to second-guess that information or conduct its own examination and include losses as ALAE that have not been reported to it as ALAE. AmTrust's argument that this result is illogical is based on the false premise that the Rule is intended to capture the full economic cost to the insurer of a claim. But that premise is undermined by the very language of ISO 5J.2 itself. It permits the insurer to ignore costs of a claim that are not separately allocated. If the insurer determines in its discretion to manage a claim internally rather than hire outside lawyers and vendors, that economic cost will not be captured by ISO 5J.2. In addition, ISO 5J.2 is silent with respect to how to treat ultimate projected losses and whether the deductible should be applied against those projected losses. The deductible would reduce a projected loss the same as it would reduce a claim that was already paid but yet ISO 5J.2 does not address how ultimate losses should be handled. In those circumstances, it is not illogical to interpret ISO 5J.2 as its plain language and Dubin's testimony establishes to requires only a mathematical calculation based on provided information with all of the strengths and weakness that such an imprecise approach implies.
210. By contrast, AmTrust's interpretation of ISO 5J.2 would sow confusion and uncertainty into an area that values certainty. Under AmTrust's interpretation, a company that has loss history for all three years (and thus cannot use ISO 4A by its terms) which the insurer or underwriter does not believe is unreliable (thus making it ineligible for ISO 4B treatment) would nonetheless be permitted (and indeed required) if those loss runs did not include ALAE to adopt one of three approaches: it could disregard entirely the year where there was no ALAE (the ISO 4A approach), it could use an EMF of 1 (the ISO 4B approach), or it could simply lift the industry average from ISO (an approach found nowhere in the rules). But, as Dubin explained, in the instance where there is data and it has not been determined to be unreliable under ISO 4B, there is no necessary reason (or permission) for assuming it is unreliable and inserting any ALAE at all. Nor, under AmTrust and Scruggs' approach, is there any rule that would determine which approach to use over another—each of which could lead to dramatically different rules. Under AmTrust's interpretation, the EMF used would differ not only based on the information provided by the insured, but on the approach taken by the underwriter itself with respect to that information. Indeed, when pressed as to how KF&B was required to act in that situation, Scruggs himself admitted that the methodology he proposed, which was an "actuarial approach," was not one that KF&B was required to employ and that his method in fact violated ISO 5F.1 with respect to certain accounts. See id. at 693:15-694:11. The Court finds more credible the testimony of Dubin that ISO 5J.2 requires ALAE to be included in full when ALAE is reported and does not permit the invention of a figure for ALAE when no ALAE is reported. See Grdinich v. Bradlees, 187 F.R.D. 77, 81-82 (S.D.N.Y. 1999) (excluding expert testimony on industry standards as unsupported speculation where expert "base[d] his conclusions on his own authority" and on "common-sense" guidelines).
211. Plaintiffs also suggest that KF&B could have appealed to a rating committee or could have raised with AmTrust the fact that ISO 5J.2 did not account for situations where accounts in certain policy periods that reported no or zero ALAE. But there is no documentary or other evidence that KF&B thought ISO Rule 5J.2 as written needed to be altered or avoided or that KF&B or AmTrust had suggested a rule change. ISO 5J.2 admits of the possibility that there might be no ALAE recorded, and where there is no ALAE recorded, it does not require ALAE to be included in full.
212. That interpretation of ISO is fatal to its claim here. Plaintiffs did not submit evidence that the experience data for the accounts at issue were inaccurate such that it contained hidden ALAE. Nor did they argue or put forth evidence that the data upon which KF&B based its calculations was unreliable. There is no evidence that ISO 4A or 4B could have been used. Accordingly, KF&B committed no violation.
213. Although AmTrust accuses KF&B of knowing or at least suspecting that certain accounts showing zero ALAE had actually improperly included that ALAE in its indemnity because Howery testified that he thought there could "[p]otentially [be] ALAE there" and "[i]f there's a very large account, yes, there's probably ALAE in there," see, e.g., Tr. at 135:10-23, 804:15-805:10, the evidence falls short of establishing, for any particular account, that the data was unreliable, that KF&B failed to include ALAE that should have been included in full, or that the resulting EMF was incorrectly calculated. The only contrary evidence was from Scruggs and the Court finds him to be unreliable on this issue.
214. In sum, Plaintiffs have not established a breach of this provision in the Underwriting Guidelines.
B. "Highest Standards of the Industry"
215. Plaintiffs' case is not assisted by framing it in terms of the obligation to act consistent with the "highest standards of the industry." As noted above, under the MPA, KF&B "agree[d] to perform the functions specified in a Program in accordance with the highest standards of the industry and pursuant to the standards and procedures set forth in a Program, including compliance with all Underwriting Guidelines and directive issued by AmTrust from time to time." MPA, Section II.B. That language holds KF&B to a high standard of care in the performance of the functions of the Program. It cannot behave negligently in the selection of accounts to refer, the calculation of a schedule factor, or the negotiation of a premium. But the language does not assign to KF&B functions that are not otherwise assigned to it under the MPA.
216. Thus, to the extent that AmTrust's claim is understood to rest upon the calculation of the premium for deductible accounts under ISO 98, or the choice of years to use in the calculation of the EMF, or in the approach to ALAE, that claim fails for the same reason that the claims brought directly under the Underwriting Guidelines fail. Plaintiffs failed to prove that KF&B was deficient in its performance of any of those functions.
217. Perhaps recognizing that challenge, Plaintiffs presented an alternative framing of the alleged violation at trial. Counsel argued that, KF&B should have informed AmTrust that: the CPP system was not calculating the same per unit premium as the Referral Spreadsheet when AmTrust discovered that discrepancy; that the manually modified EMF calculation was inappropriate; and that KF&B was not separating out a figure for ALAE when no figure was broken out in the loss runs. None of these arguments succeeds on the evidence presented to the Court.
218. As the Supreme Court has held in an analogous context, when a party asks a court to impose meaning or liability based on "known customs or usages in a particular industry . . . the parties must prove those customs or usages using affirmative evidentiary support in a given case." M & G Polymers USA, LLC v. Tackett, 574 U.S. 427, 439 (2015); see also Pension Comm. of Univ. of Montreal Pension Plan v. Banc of Am. Secs., LLC, 691 F. Supp. 2d 448, 463-64 (S.D.N.Y. 2010) (admitting expert testimony on industry standards). That principle is as relevant in giving meaning to a contractual reference to the "standards of the industry" as it is to interpreting ambiguous language that implicitly references the standards of and meaning in the industry. A party cannot simply assert industry practice or ask the factfinder to speculate. Cf. Lippe v. Bairnco Corp., 2002 WL 15630, at *2 (S.D.N.Y. Jan. 7, 2002) (experts who testify as to "the customs and standards of an industry, and [] opine as to how a party's conduct measured up against such standards . . . must do more than aver conclusorily") (Chin, J.) (citation omitted). It must introduce evidence of industry standards. Plaintiffs failed to do so.
219. Plaintiffs' counsel, more than their own experts, made the point that Howery should have gone back to AmTrust to inform it that the CPP system was not calculating the same per unit premium as his Referral Spreadsheet. But there is no affirmative evidence that industry standards would have required an underwriter—who did not have actuaries and who was obliged to follow the formula that the insurer, who did have actuaries, filed with the state—to alert the insurer that the underwriter's own prior formula was inconsistent with that of the insurer. The underwriter was required to follow the insurer formula whether it wanted to or not. Indeed, the most Plaintiffs offered was the testimony of Bendure that KF&B also "could have" hired outside actuaries to review the Referral Spreadsheet and that KF&B should have gotten it fixed, but Bendure did not say whether that was required or industry standard. Tr. at 359:18-360:2. Scruggs did not opine on the error.
KF&B's expert Jim Leatzow said that if he personally knew there was a "problem or an issue going on" with respect to the information coming out of a system such as CPP, or if he knew there was something being done that was not compliant with ISO, he would bring it to the insurer's attention. See Tr. at 577:1-14, 587:21-24, 579:8-18, 598:17-25. At the same time, he also testified that "KF&B is not the ISO subscriber, AmTrust is. AmTrust owes whatever obligation to ISO, but KF&B is basically using whatever is given to them or whatever is available to them from the insurer they're representing." Id. at 559:23-560:5; see also id. at 595:7-596:5. None of Leatzow's testimony went to industry standards either way.
220. Plaintiffs also did not provide expert testimony on industry standards with respect to the EMF manual override. Notably, Plaintiffs did not call or submit testimony from AmTrust's in-house underwriters, Fort and Hynes, or its actuaries, Lamendola and Kushner, that stated KF&B departed from industry standards when manually overriding the EMF. Nor did Plaintiffs submit testimony that AmTrust did not have all of the documentation it needed to bind a taxi account. In fact, as discussed earlier, Fort testified as to the opposite—that he would ask KF&B for missing or additional documentation.
221. Finally, Plaintiffs offered no industry evidence to support that if KF&B received loss history that showed zero ALAE it was industry practice to return to the broker, in turn, to ask the insured whether its accounting team had accurately separated out ALAE and indemnity. The Court credits Dubin's testimony that says that KF&B was stuck with the data given by the insured.
C. "Obligations as a Fiduciary"
222. Finally, Plaintiffs argue that KF&B violated its contractual duty to satisfy its fiduciary obligations when it failed to alert AmTrust to the aforementioned errors.
The Court previously dismissed Plaintiffs' breach of fiduciary duty claim concerning allegations based solely on the contractual language of Section III of the MPA. See AmTrust N. Am., Inc. v. KF&B, Inc., 2020 WL 5519197, at *2 (S.D.N.Y. Sept. 14, 2020). --------
223. Section III of the MPA required that:
[KF&B] shall perform its obligations as a fiduciary of AmTrust, and shall use its best efforts to perform all acts necessary for the proper conduct of business on behalf of AmTrust. [KF&B] shall advise AmTrust if it, or any officer or employee is convicted of a felony crime involving dishonesty or breach of trust, and shall in addition notify AmTrust of any suspected fraudulent acts in connection with this Agreement.MPA, Section III; see also id., Section XXI (the MPA "places responsibilities, duties, and obligations upon [KF&B] different from those of a normal insurance producer or agency contract").
224. "A fiduciary relationship may arise [] where the parties to a contract specifically agree to such a relationship." Poon v. Roomorama, LLC, 2009 WL 3762115, at *4 (S.D.N.Y. Nov. 10, 2009) (citation omitted). The duties owed by KF&B under the contract were identical to those under tort law generally. The contract did not add or subtract from those duties. As a fiduciary, KF&B had a duty of loyalty and good faith and a duty to disclose material matters within the scope of its relationship.
225. There is no allegation here, and was no evidence, that KF&B breached a duty of loyalty. Although the MPA's compensation structure gave KF&B an interest in generating premiums and overall business, that was an interest KF&B shared with AmTrust and cannot give rise to a violation of the duty of loyalty. The interests of KF&B did not conflict with those of AmTrust but were aligned. See, e.g., Boston Consulting Grp., Inc. v. NCR Corp., 2020 WL 5731963, at *3 (S.D.N.Y. Sept. 24, 2020) (breach of fiduciary duty of loyalty involves "blatant self-dealing" and "situations in which a fiduciary's personal interest possibly conflicts with the interest of those owed a fiduciary duty") (quoting Birnbaum v. Birnbaum, 539 N.E.2d 574, 576 (1989)).
226. Plaintiffs also did not prove that, in the discharge of its obligations with respect to the underwriting, rating, and binding of the taxi accounts at issue, KF&B violated a duty of good faith or a duty of care. The Court assumes that as AmTrust's "appointed agent," D-Howery-1 at Endorsement 1 to the MPA, charged with acting as a fiduciary, KF&B owed a duty to AmTrust "to act with the care, competence, and diligence normally exercised by agents in similar circumstances," Restatement (Third) of Agency § 8.08 (2020). The Court has already concluded that KF&B did not violate the specific obligations imposed by the Underwriting Guidelines or the more general obligation to act according to the highest standards of the industry in its underwriting, rating and binding of those accounts. Although the MPA requires KF&B to perform its duties with care and as a fiduciary, AmTrust did not argue—and the evidence would not support—that Section III of the MPA imposed different obligations with respect to the formula for rating deductible taxi accounts (and ISO 98) or with respect to the calculation of EMF (and the treatment of ALAE) than imposed by the more specific sections of the MPA. And the same evidence that supports that KF&B acted consistent with the highest standards of the industry also supports that KF&B acted in good faith and did not violate any duty of care with respect to its discharge of its functions with respect to those accounts.
227. AmTrust's most substantial argument is that KF&B violated a duty to disclose to AmTrust the circumstances surrounding the ISO 98 error and the calculation of ALAE. With respect to the former, AmTrust adduced evidence that Howery, having previously developed what he believed to be an accurate algorithm for deductible taxi accounts and observing that AmTrust used a different formula, nonetheless adopted AmTrust's formula without informing AmTrust that AmTrust's formula deviated from KF&B's own formula. With respect to the latter, AmTrust adduced evidence that Howery had consulted the ISO-CLM to determine how to treat loss runs where no ALAE was reported and decided that it did not provide a direct answer, without informing AmTrust or asking AmTrust's actuaries for the answer.
228. After careful consideration, however, the Court concludes that such evidence is insufficient to support the claim that KF&B violated its duties "to perform its obligations as a fiduciary of AmTrust, and [to] use its best efforts to perform all acts necessary for the proper conduct of business on behalf of AmTrust." MPA, Section III. As a general matter, an agent has a "duty to use reasonable effort to provide the principal with facts that the agent knows, has reason to know, or should know when . . . subject to any manifestation by the principal, the agent knows or has reason to know that the principal would wish to have the facts or the facts are material to the agent's duties to the principal." Restatement (Third) of Agency § 8.11; accord Qosina Corp. v. C & N Packaging, Inc., 948 N.Y.S.2d 308, 310 (2d Dep't. 2012). As agent and fiduciary, KF&B thus owed a "duty to disclose material facts." Grumman Allied Indus., Inc. v. Rohr Indus., Inc., 748 F.2d 729, 738-39 (2d Cir. 1984) ("[U]nder New York law, a duty to disclose material facts is triggered . . . where the parties enjoy a fiduciary relationship.").
229. A breach of this duty to disclose, however, is "limited to matters relevant to affairs entrusted." Rush v. Oppenheimer & Co., Inc., 681 F.Supp. 1045, 1055 (S.D.N.Y. 1988) (Sweet, J.); see Conway v. Icahn & Co., Inc., 16 F.3d 504, 510 (2d Cir.1994) (an agent "has a duty to use reasonable efforts to give its principal information relevant to the affairs that have been entrusted to it"). Thus, whether KF&B reached a duty to AmTrust "depends on the scope of the trust extended to [KF&B] at the time of the alleged misconduct." BNY Capital Markets, Inc. v. Moltech Corp., 2001 WL 262675, at *8 (S.D.N.Y. Mar. 14, 2001) (Lynch, J.)
230. Plaintiffs' claim with respect to ISO 98 is that KF&B knew that the ISO 98 formula used by AmTrust was different from that calculated by KF&B itself and that KF&B had believed was compliant with ISO and did not bring that to AmTrust's attention. But, as the Court has found, the determination of the appropriate ISO rule and rate to use, and its translation into an algorithm, was the responsibility of AmTrust—the party with sole involvement in and responsibility for the filings with the states. KF&B did not have a responsibility to determine the correct rule; its responsibility aside from underwriting was to make sure that its ratings and underwritings were consistent with what was represented to it by AmTrust to be the right rule and algorithm to use. Indeed, as Howery testified and as this Court has found, KF&B did not have actuaries and was not hired or qualified to determine the ISO actuarial formulas for AmTrust, which AmTrust negotiated with each state on its own. Instead, KF&B was to use the ISO actuarial formulas, as determined by AmTrust, in performing its own underwriting analysis. It thus follows that while KF&B could gratuitously have offered observations with respect to the algorithm, there was no duty in law or contract that required it to do so. See, e.g., Ross v. FSG PrivatAir, Inc., 2004 WL 1837366, at *7 (S.D.N.Y. Aug. 17, 2004) (no breach where defendant, who held itself out as an expert in "aircraft sales and acquisitions" and in "negotiating contracts for the purchase and sale of aircraft," drafted contract that prevented plaintiff from recovering his deposit); BNY Capital Markets, 2001 WL 262675, at *10 (no breach where defendant was hired for services in connection with one transaction, not to review other transactions or "to conduct broad due diligence to determine every possible scenario under which the [] transaction at issue might trigger a double-fee obligation"; "[i]ndeed, the parties' course of conduct demonstrates that [plaintiff] assumed exclusive obligations for making such determinations").
231. Nor does the evidence support that KF&B should have inferred that the CPP system's calculation of ISO 98 was incorrect and that AmTrust had made the mistake. AmTrust was in a far better position than KF&B to evaluate and determine the accuracy of the actuarial formulas. See BNY Capital Markets, 2001 WL 262675, at *10 (rejecting argument that defendant "should have inferred a problem" from the discrepancies because plaintiff was in a "far better position than [defendant] to evaluate their significance"). Similarly, Plaintiffs have not adequately demonstrated that KF&B's failure to alert AmTrust to the discrepancy between the CPP system and the Referral Spreadsheet violated the "best efforts" clause such that it was done for the purpose of self-dealing or was an act of bad faith. See Impax Media, Inc. v. Ne. Advert. Corp., 2018 WL 3962841, at *7 (S.D.N.Y. Aug. 17, 2018) (Sweet, J.) (analyzing "best efforts" clause under duties of good faith and loyalty). The Court finds credible KF&B's evidence that it did not know, or have reason to believe, that AmTrust's own CPP system and actuaries had the wrong ISO formula.
232. Similarly, AmTrust has not proved that KF&B failed to provide information about its treatment of ALAE that KF&B knew or had reason to know that AmTrust would have wanted to know or that was material to KF&B's discharge of its duties to AmTrust. KF&B disclosed, and AmTrust had, the loss history for the respective accounts when deciding whether to accept KF&B's referral and bind the account. Those were reflected in the Referral Spreadsheets and were in the possession of AmTrust. Significantly, AmTrust knew KF&B's methodology in providing for ALAE because AmTrust had the back-up loss history data for each insured and also KF&B's Referral Spreadsheets that showed this data and automatically calculated ALAE and indemnity from the data. These Referral Spreadsheets were sent to AmTrust's in-house underwriters, and in some cases its actuaries, every time an account was bound. AmTrust could thus see that the insured reported no or zero ALAE and that the Referral Spreadsheet used this same ALAE. See Aaron Ferer & Sons Ltd. v. Chase Manhattan Bank, Nat'l. Ass'n, 731 F.2d 112, 123 (2d Cir. 1984) ("While there is evidence that [defendant] did not present a detailed explanation of" the third-party agreement at issue, "plaintiffs failed to establish that [defendant's] knowledge was 'superior', because all of the information that plaintiffs now claim was concealed from them was either a matter of public record, was not pursued by plaintiffs, or was disclosed, at least in part, by [defendant]").
233. Moreover, AmTrust's own underwriters, Hynes and Fort, both testified that they had the information that they needed to decide whether to bind the account. Even after it was known, at least through this case, precisely how KF&B treated losses that did not contain ALAE, neither they nor any of AmTrust's actuaries testified that the loss information or any other information was missing when they made the decision to accept a referral.
234. Indeed, AmTrust offered only two fact witnesses, Mark Murphy, a vice president of finance at AFSI, and Kathleen Smith, a vice president of underwriting at AmTrust. Neither witness was involved in the underwriting of the accounts or the decisions to accept the referrals. Murphy was responsible for financial administration of the Program, Am. Murphy Decl. ¶¶ 5-6, and Smith has not described her interaction with KF&B or involvement in the Program, see supra. Thus, neither witness was able to testify, from their own knowledge, that there was information KF&B had that was not disclosed to AmTrust, much less that such information would have been important or even relevant to it in making the decision to insure the accounts.
235. Plaintiffs' claim thus comes down to the assertion that it would have been important to AmTrust to know that KF&B was not estimating and calculating a separate figure for ALAE when the loss runs did not include such figure because KF&B's approach was inconsistent with ISO or it could have adopted a different approach under ISO. But the Court has found that AmTrust has not proven that KF&B's approach was inconsistent with ISO or that any other method was available. As Dubin testified, an underwriter would not consider it to be an error to record zero ALAE if that was the actual data reported by the insured as part of their internal accounting. Significantly, there is no evidence that any of the loss runs at issue showing zero ALAE were incorrect, unreliable, or otherwise faulty, or that KF&B knew or believed that they were incorrect, unreliable, or otherwise faulty. Thus, AmTrust has not shown that KF&B was in possession of any information, not otherwise shared with AmTrust, that AmTrust would have wanted to know or that would have been important to AmTrust and therefore also has not shown that KF&B breached any duty to disclose.
236. Having concluded that there was no breach and KF&B satisfied its duties to AmTrust, the Court need not and does not analyze causation or damages.
ORDER
237. For the reasons stated, judgment should be entered for Defendant on Plaintiffs' claims and for Plaintiffs on Defendant's amended counterclaim. The parties are directed to jointly prepare a form of judgment consistent with this opinion.
238. The Clerk of Court is respectfully directed to terminate all pending motions.
SO ORDERED. Dated: October 27, 2020
New York, New York
/s/_________
LEWIS J. LIMAN
United States District Judge