Phoenix Light SF Limited v. The Bank of New York Mellon Corporation

• Court: District Court, S.D. New York
• Decided: March 20, 2020
• Docket: 1:14-cv-10104
• Status: Unknown

Opinion

UNITED STATES DISTRICT COURT ELECTRONICALLY FILED DOC #: SOUTHERN DISTRICT OF NEW YORK DATE FILED: 3/20/ 2020 ------------------------------------------------------------------ X PHOENIX LIGHT SF LIMITED, et al., : : Plaintiffs, : 14-CV-10104 (VEC) : -against- : OPINION AND ORDER : THE BANK OF NEW YORK MELLON, as Trustee, : : Defendant. : ------------------------------------------------------------------ X VALERIE CAPRONI, United States District Judge: In 2014, investors in a number of Residential Mortgage Backed Securities (“RMBS”) trusts sued Defendant-Trustee (“BNYM”) for breach of contract, breach of fiduciary obligations, negligence, and violations of the Trust Indenture Act (“TIA”). On September 7, 2017, the Court granted BNYM’s motion for summary judgment in part and denied it in part. Mot. Summ. J. Op. & Order (Dkt. 201). Plaintiffs later voluntarily dismissed claims as to seven trusts. Stip. & Order (Dkt. 236). Both parties have now moved in limine to exclude the opinions of the other party’s experts. Mots. (Dkts. 291, 294). On November 13, 2019, the Court denied in large part BNYM’s motion to strike Plaintiffs’ experts’ declarations submitted in opposition to BNYM’s motion in limine. Mot. Strike Op. & Order (Dkt. 319).1 On March 9, 2020, the Court held oral argument on both sides’ motions and ruled orally on the vast majority of Plaintiffs’ motions.2 1 The Court did, however, strike Part I.A of the Adelson Declaration, paragraph twenty-nine of the Bitner Declaration, and paragraph twelve of the Mason Declaration. See Mot. Strike Op. & Order at 11. 2 By separate Order (Dkt. 336), the Court ruled on the remainder of Plaintiffs’ motion. For the following reasons, BNYM’s motion in limine is GRANTED in part and DENIED in part.3 BACKGROUND The Court presumes familiarity with its summary judgment decision, but a brief canvass

of the remaining ten claims will be helpful to the reader. Those claims—for breach of contract and TIA violations—fall into three groups of trusts. Of the trusts still at issue in this case, six (the CWALT and CWL trusts) were sponsored by Countrywide Home Loans, Inc. (the “Countrywide Trusts”). Plaintiffs claim that, as to those trusts, BNYM breached the pooling and servicing agreements (“PSAs”) by failing to disclose that complete mortgage files had not been delivered for some loans. As such, it failed to make accurate Regulation AB (“Reg AB”) disclosures. Plaintiffs assert that BNYM should have ensured that Countrywide repurchased the affected loans. As to CWALT 2006-OA3 only, Plaintiffs claim that BNYM breached the contract by failing to provide notice of breaches of representations and warranties (“R&W”) that materially and adversely affected the interests of

certificateholders and by failing to enforce loan repurchases when BNYM received written notice of R&W breaches in a letter from a certificateholder (the Baupost Letter).4 An additional three trusts still at issue were sponsored by Credit-Based Asset Servicing and Securitization LLC (the “CBASS Trusts”). Plaintiffs claim that BNYM breached their PSAs by failing to ensure that the servicer for the CBASS Trusts improved its loss mitigation practices.

3 Both parties had motions in limine that related to one aspect of Ingrid Beckles’ report – specifically whether she misinterpreted a document provided by BNYM. Those aspects of the motions in limine are stayed pending an evidentiary hearing, scheduled for April 22, 2020. 4 CWALT 2006-OA3 was the subject of a global settlement among BNYM, various investors, Bank of America, and Countrywide on June 28, 2011. That settlement is referred to herein as the “the Countrywide Settlement.” The final trust still at issue was sponsored by ECC Capital Credit Corp. and was structured using an Indenture and Servicing Agreement (the “ECR 2005-1 Trust”). Plaintiffs claim that BNYM violated Section 315(b) of the TIA by failing to give certificateholders notice of a default under the indenture within 90 days of learning of the default. DISCUSSION5

Federal Rule of Evidence 702 governs the admissibility of expert testimony. It provides that a person “qualified as an expert by knowledge, skill, experience, training, or education” may offer opinion testimony so long as: (a) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and

5 The Court will refer to the parties’ submissions using the following abbreviations. Attorneys’ submissions: Memorandum of Law in Support of BNYM’s Omnibus Daubert Motion (Dkt. 292) as “BNYM’s Mem. of Law”; Memorandum of Law in Opposition to BNYM’s Omnibus Daubert Motion (Dkt. 300) as “Pls.’ Mem. of Law”; BNYM’s Omnibus Reply in Further Support of Its Daubert Motion (Dkt. 320) as “BNYM’s Reply”; Declaration of Christopher J. Houpt (Dkt. 293) as “Houpt Decl.”; Declaration of Daniel Fisher (Dkt. 321) as “Fisher Decl.”; Declaration of Steven S. Fitzgerald (Dkt. 301) as “Fitzgerald Decl.”; Declaration of Nicole M. Clark (Dkt. 296) as “Clark Decl.” Plaintiffs’ experts’ submissions: Revised Report of Richard Bitner (Houpt Decl. Ex. 2) as “Bitner Report”; Report of Richard Bitner Concerning Damages (Houpt Decl. Ex. 3) as “Bitner Damages Report”; Declaration of Richard Bitner (Dkt. 304) as “Bitner Decl.”; Expert Report of Mark Adelson (Houpt Decl. Ex. 10) as “Adelson Report”; Rebuttal Report of Mark Adelson (Houpt Decl. Ex. 11) as “Adelson Rebuttal”; Declaration of Mark Adelson (Dkt. 302) as “Adelson Decl.”; Revised Expert Report of Ingrid Beckles (Houpt Decl. Ex. 14) as “Beckles Report”; Report of Ingrid Beckles Concerning Damages (Houpt Decl. Ex. 15) as “Beckles Damages Report”; Revised Rebuttal Expert Report of Bruce D. Spencer, Ph. D. (Houpt Decl. Ex. 18) as “Spencer Rebuttal”; Declaration of Bruce D. Spencer, Ph. D. (Dkt. 306) as “Spencer Decl.”; Revised Expert Report of Dr. Joseph R. Mason (Houpt Decl. Ex. 22) as “Mason Report”; Declaration of Dr. Joseph R. Mason (Dkt. 305) as “Mason Decl.” BNYM’s experts’ submissions: Supplemental Expert Report of Prof. Philip B. Stark (Houpt Decl. Ex. 7) as “Stark Suppl. Report”; Expert Report of Charles H. Grice (Houpt Decl. Ex. 27) as “Grice Report”; Expert Report of Faten Sabry, Ph. D. (Houpt Decl. Ex. 20) as “Sabry Report”; Rebuttal Expert Report of Faten Sabry, Ph. D. (Houpt Decl. Ex. 9) as “Sabry Rebuttal”; Declaration of Faten Sabry, Ph. D. (Dkt. 322) as “Sabry Decl.” (d) the expert has reliably applied the principles and methods to the facts . . . . The party offering expert testimony bears the burden of establishing by a preponderance of the evidence that the testimony satisfies Rule 702. The district court is, however, the “ultimate gatekeeper.” United States v. Williams, 506 F.3d 151, 160 (2d Cir. 2007). Rule 702 tasks the

trial judge with “ensuring that an expert’s testimony both rests on a reliable foundation and is relevant to the task at hand.” Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 597 (1993). This gatekeeping obligation “applies not only to testimony based on ‘scientific’ knowledge, but also to testimony based on ‘technical’ and ‘other specialized’ knowledge.” Kumho Tire Co. v. Carmichael, 526 U.S. 137, 141 (1999). The threshold question for the Court is whether the “proffered expert testimony is relevant.” Amorgianos v. Nat’l R.R. Passenger Corp., 303 F.3d 256, 265 (2d Cir. 2002). If it is, the Court must then determine whether it has “a sufficiently reliable foundation to permit it to be considered.” Id. (quotations omitted). The Supreme Court has laid down several factors pertinent to this inquiry, including “whether a theory or technique . . . can be (and has been)

tested”; “whether the theory or technique has been subjected to peer review and publication”; whether uniform “standards controlling the technique’s operation” exist; and whether the theory or technique enjoys “general acceptance” within an identifiable relevant scientific or professional community. Daubert, 509 U.S. at 593–94. The Court’s ultimate objective is to “to make certain that an expert, whether basing testimony upon professional studies or personal experience, employs in the courtroom the same level of intellectual rigor that characterizes the practice of an expert in the relevant field.” Kumho Tire Co., 526 U.S. at 152. “To warrant admissibility . . . it is critical that an expert’s analysis be reliable at every step.” Amorgianos, 303 F.3d at 267. “A minor flaw in an expert’s reasoning or a slight modification of an otherwise reliable method will not render an expert’s opinion per se inadmissible,” but the “district court should undertake a rigorous examination of the facts on which the expert relies, the method by which the expert draws an opinion from those facts, and how the expert applies the facts and methods to the case at hand.” Id. “Trained experts

commonly extrapolate from existing data. But nothing in either Daubert or the Federal Rules of Evidence requires a district court to admit opinion evidence that is connected to existing data only by the ipse dixit of the expert.” Gen. Elec. Co. v. Joiner, 522 U.S. 136, 146 (1997); see also, e.g., Zerega Ave. Realty Corp. v. Hornbeck Offshore Transp., LLC, 571 F.3d 206, 213–14 (2d Cir. 2009) (“[A] trial judge should exclude expert testimony if it is speculative or conjectural . . . .”). I. Plaintiffs’ Expert Richard Bitner BNYM moves to exclude Richard Bitner’s Report and his Damages Report. Both relate to the CWALT 2006-OA3 trust. Bitner, who is Plaintiffs’ reunderwriting expert: (a) reviewed the 196 loans that had been identified in the Baupost letter to determine whether there were

material breaches of R&Ws associated with those loans; and (b) opined how long it would have taken BNYM to confirm whether the loans identified by Baupost, in fact, had breaches of their R&Ws. The Court finds that the opinions of Bitner are admissible. A. Opinions Based on Bureau of Labor Statistics Data Bitner opined that 21 loans identified in the Baupost Letter had material defects due to borrowers’ misrepresentations of their incomes. Bitner Report ¶ 99. The loans in question were “stated income” loans, i.e., loans that did not require verification of a borrower’s actual income. Id. ¶ 97. In lieu of verification, the underwriter was obligated to determine the “reasonableness” of the stated income. Id. ¶ 36. BNYM argues that Bitner’s opinion is not reliable because he compared the borrowers’ stated incomes to average-income data from the Bureau of Labor Statistics (“BLS”). In particular, BNYM advances four critiques of Bitner’s reliance on BLS data: (1) the applicable Countrywide underwriting guidelines did not require comparison of the stated income to BLS

data; (2) Bitner’s analysis wrongfully assumed no borrower’s income was above the 90th percentile of a particular job category; (3) Bitner lacks sufficient experience with BLS data to utilize it for underwriting; and (4) BLS data is not an objectively reliable source of information for income verification. BNYM’s Mem. of Law 5–8. The fact that the Countrywide guidelines did not require use of BLS data as an appropriate benchmark does not make Bitner’s use of BLS data improper. The guidelines identified no particular data set that an underwriter should use to assess the reasonableness of a borrower’s stated income. The guidelines stated only that a borrower’s stated income “must be deemed reasonable and consistent with the borrower’s occupation.” Houpt Decl. Ex. 6 at *793– 94. BNYM does not argue that this instruction precluded use of BLS data, or even that using

BLS data as the sole benchmark would have contradicted the guidelines. Indeed, BNYM admits that the guidelines did “not specify how the underwriter should make th[e] assessment” about a borrower’s stated income. BNYM’s Mem. of Law at 6. BNYM’s singular focus on Bitner’s use of BLS data is similarly misguided. Bitner concluded that income was unreasonable only if the income exceeded the 90th percentile by a substantial margin6 and only if other objective information signaled that the stated income was

6 Although the Court is less than enthusiastic about the amount of information crammed into declarations to oppose Defendant’s Daubert motion (as opposed to including the information in the expert’s report in the first instance), according to Bitner’s declaration, across the 21 loans as to which he determined “the stated income was unreasonable, the average percentage over the 90th percentile was 53.8%.” Bitner Decl. ¶ 18. Even discounting the unreasonable, such as the borrower’s assets and credit history. Houpt Decl. Ex. 4 at 347:2–18; Bitner Decl. ¶¶ 11–19. This use of BLS data is consistent with the methodology deemed admissible in U.S. Bank, National Ass’n v. UBS Real Estate Security Inc., in which the court noted that the “underwriting process draws heavily on judgment and takes into consideration the

different qualities of each loan and borrower.” 205 F. Supp. 3d 386, 405, 444, (S.D.N.Y. 2016). BNYM relies upon Ruggiero v. Warner-Lambert Co. to argue that Bitner needed to rule out alternatives before concluding that the borrower had misstated his income. 424 F.3d 249 (2d Cir. 2005). In Ruggiero, it was unreliable for an expert to use a differential medical diagnosis method to establish that a particular drug generally caused a particular ailment. Id. at 254. Although the method ruled out certain possible causes of an ailment, it did not “rule in the suspected cause,” and therefore was not appropriate for establishing general causation. Id. (quotation omitted). Ruggiero is inapplicable here. Bitner is not making a differential diagnosis or opining on general causation. In fact, he offers no opinion whatsoever on the cause of the borrowers’ stated incomes.

The Court also finds meritless BNYM’s argument that Bitner lacks sufficient experience to map borrowers’ job descriptions onto the categories in BLS data. He has used other income- reporting data sources, such as Salary.com, but chose to use BLS data here, after reviewing BLS’s data-gathering methodology, because he concluded that BLS had better historical data. See Houpt Decl. Ex. 4 at 317:10–23; Bitner Decl. ¶ 22. BNYM has not shown that the BLS database has salient differences from other online salary databases that would require unique experience to interpret when one has extensive experience using other similar data sets.

probative value of that statistic a little because of the amount of mischief that can be hidden in an average, it suggests that using the 90th percentile cut-off did not significantly affect his conclusion. The Court finally finds that the BLS data is not itself so unreliable that it taints Bitner’s opinion. BLS data has flaws, but those flaws are similar to the flaws in other online salary databases that underwriters frequently use, see Bitner Decl. ¶ 22, and both U.S. Bank and Nomura allowed reunderwriting expert testimony that relied upon BLS data, see U.S. Bank, 205

F. Supp. 3d at 444; FHFA v. Nomura Holding Am., Inc., 104 F. Supp. 3d 441, 527–28 (S.D.N.Y. 2015), aff’d, 873 F.3d 85 (2d Cir. 2017). Moreover, as noted above, Bitner’s opinion is not based solely on BLS data, limiting the impact of imperfections. See Bitner Decl. ¶ 19. His opinion, in short, is based on adequate data. BNYM’s discussion of United States v. Percoco mischaracterizes this Court’s critiques of the Percoco expert’s specific comparisons between the defendant’s job and BLS’s job categories. No. 16-CR-776, 2018 WL 879499, at *3–4 (S.D.N.Y. Feb. 13, 2018); see BNYM Reply at 2. Those criticisms were not wholesale critiques about the coarseness of BLS data, but focused on the lack of rigor used by the purported expert in, inter alia, selecting appropriate comparable job titles. Although BNYM’s statistics expert, Philip Stark, criticized Bitner’s

opinions for using high-level BLS categories that contain large variability (i.e., they were “quite coarse”), Bitner built in a buffer above 90th percentile to reduce potential error caused by such coarseness. Moreover, even according to Stark, see Stark Suppl. Report ¶¶ 10–14, Bitner used a high-level BLS category that had large variability in just two instances out of 196. See Nomura, 104 F. Supp. 3d at 528 (holding that reliance on BLS data was not unreasonable because, in part, the reunderwriting expert “was conservative in using BLS data, comparing borrowers’ stated income only to the 90th percentile of income recorded by the BLS, the highest available figure.”). Unlike what was evident in Percoco, neither Stark nor BNYM has identified particular problems with the matching of job categories between the borrower’s declared job and the BLS occupation category Bitner selected. B. Opinions About Title Policy Defects Bitner also opined that 100 loans breached the PSAs because they had insufficient title insurance coverage. Bitner Report ¶ 142. BNYM argues that his opinion should be excluded

because he has no experience dealing with title insurance coverage. BNYM’s Mem. of Law at 8. The Court finds that Bitner’s underwriting experience qualifies him. He need not have actually processed a title policy claim to be able to determine that the loans in the CWALT 2006- OA3 trust were required to have title insurance, that all of the loans in the trust had negative amortization features, that the negative amortization feature required an appropriate endorsement on the title insurance policy to deal with that characteristic of the loan, and that certain loans in the trust did not have the necessary endorsements. C. Opinions About Incomplete Loan Files In reaching his opinion, Bitner assumed that if a required document was not in the loan file provided to him then the document likely was never in the file. Bitner Report ¶ 123; Bitner

Decl. ¶ 32. BNYM argues that is an unwarranted assumption because Bitner had no information about the loan files’ chain of custody. BNYM’s Mem. of Law at 9. The Court finds that Bitner’s inference is reasonable, and BNYM’s concerns go to the weight of his testimony, not its admissibility. Judge Castel drew a similar inference in U.S. Bank, concluding that documents’ “absence from the loan files makes it more likely than not that they were not used in the underwriting process” because “the contents of the loan files were the basis for the loans’ underwriting and approval.” 205 F. Supp. 3d at 457. If BNYM wishes to challenge Bitner’s conclusions regarding particular loan files, see id. at 501, it may do so, but the Court sees no reason to exclude Bitner’s testimony as unreliable. Chain of custody issues go to the weight, not the admissibility, of his testimony. D. Opinions on Materiality of Alleged Breaches Bitner asserts that 156 loans contained in CWALT 2006-OA3 had at least one material breach of an R&W, totaling 454 material breaches. Bitner Report ¶¶ 2, 9, 82. BNYM argues

that he has no basis for opining on the materiality of the breaches because he has no RMBS investing or loan valuation experience. BNYM’s Mem. of Law at 11. The Court disagrees that investing and valuation experience are the only relevant experiences that would qualify an expert to opine on materiality. Bitner has sufficient experience to determine materiality from the perspective of a loan originator, who was selling loans to investors. See Bitner Decl. ¶¶ 36–37. From that experience, Bitner has a basis to determine whether certain defects would dissuade a reasonable investor from purchasing a loan. The standard Bitner applied was both consistent with this Court’s interpretation of the PSA, see Mot. Summ. J. Op. & Order at 19, and consistent with expert testimony in other cases by loan origination and underwriting experts regarding the materiality of breaches, see, e.g., Assured

Guar. Mun. Corp. v. Flagstar Bank, FSB, 920 F. Supp. 2d 475, 510–11 (S.D.N.Y. 2013). Indeed, BNYM’s own reunderwriting expert appears to accept that a materiality determination is at least within his and Bitner’s ken. See Grice Report ¶¶ 52–53. Although BNYM’s expert challenges Bitner’s materiality determinations regarding particular breaches, see, e.g., id. ¶¶ 120–22, that, again, is a question of weight, not admissibility. E. Opinions About Duration of the Reunderwriting Process Finally, Bitner opined on the amount of time it would have taken to reunderwrite the loans listed in the Baupost letter. He provided two estimates: one for a comprehensive review and one for a review only of the specific deficiencies mentioned by Baupost. See Bitner Damages Report ¶¶ 6–7. BNYM challenges Bitner’s opinion that a comprehensive reunderwriting effort of defective loans would have taken six to nine months, and that a less comprehensive effort would have taken three to six months.7 BNYM Mem. of Law at 12. BNYM argues that the only basis for this opinion—Bitner’s own experience—contradicts

his conclusions. Bitner and his team, according to BNYM, took two years to reunderwrite the loans he claims BNYM could have reunderwritten in six to nine months, citing deposition testimony from Bitner. But that is not what Bitner testified. The original project to reunderwrite approximately 2000 loans began in early 2016, but that project was put on hold until after this Court’s summary judgment opinion. Following this Court’s decision, Bitner was reengaged to reunderwrite only 196 loans, which took his team four to five weeks. Fitzgerald Decl. Ex. 1 at 30:8–19; Bitner Decl. ¶ 41. In addition, Bitner’s extensive underwriting experience is adequate to form a basis for his opinions about how long a comprehensive and less comprehensive review would have taken. BNYM’s concern that Bitner underestimated the length of time it would have taken BNYM to obtain loan files for reunderwriting go to the weight of his testimony only, not

its admissibility. II. Plaintiffs’ Expert Mark Adelson Mark Adelson, Plaintiffs’ securitization expert, opined regarding the actions BNYM could and should have taken as an RMBS trustee, even without direction from a certificateholder. Adelson’s testimony is intended to elucidate the “prudent person” standard contained in the applicable PSAs and to demonstrate that BNYM did not meet that standard. The Court finds that the Adelson Report is admissible in part and inadmissible in part.

7 Bitner offered his opinion in aid of Plaintiffs’ damages expert’s analysis of when the trusts would have received money from loan repurchases. A. Opinions That a Trustee Should Review Loans and Demand Repurchase Following an Event of Default BNYM challenges two related opinions from the Adelson Report. Adelson first opined that: “[t]here are a number of things that a non-agency MBS trustee [like BNYM] can do once an Event of Default occurs,” including “initiat[ing] a review of defaulted mortgage loans and demand[ing] repurchase of those that have breaches of R&Ws.” Adelson Report ¶ 172. Adelson’s second opinion is prescriptive: “it would not be prudent for a non-agency MBS trustee to merely provide notice and then do nothing except wait for a direction from certificateholders”; and, because that is what BNYM did, in his opinion, BNYM acted imprudently. Id. ¶ 174; Adelson Rebuttal ¶¶ 31–32. The Court finds that Adelson’s first opinion is admissible, but the

second opinion is not because it lacks a sufficient factual basis. BNYM argues that Adelson has not explained how his industry experience supports his first opinion about all of the actions that BNYM, as a trustee, could have taken. BNYM’s Mem. of Law at 14. “Rule 702 specifically contemplates that expert testimony may be based upon experience.” Pension Comm. of Univ. of Montreal Pension Plan v. Banc of Am. Sec., LLC, 691 F. Supp. 2d 448, 473 (S.D.N.Y. 2010) (citing Fed. R. Evid. 702 Advisory Committee Note). Nonetheless, “the witness must show why that experience provides a reliable foundation for his or her testimony.” Id. Adelson (somewhat loosely) connects his 30 years of experience as an RMBS industry analyst, consultant, and attorney to his knowledge of all the options RMBS trustees have in certain hypothetical scenarios. See Adelson Report ¶¶ 10–21. He stated in his

deposition that reviewing defaulted loans and demanding repurchase “is, in fact, what investors have done . . . many, many times.” Houpt Decl. Ex. 12 at 130:6–12. That observation is supported by his experience as head of structured finance research at Nomura Securities International and head of the RMBS rating team at Moody’s, at which Adelson “monitored developments in the RMBS industry closely and observed the reactions of investors and securitization professionals.” Adelson Decl. ¶¶ 15–16. Alone, this may not be a sufficient basis for his opinion—which focuses on what a trustee, not an investor, can do—but Adelson also provides several examples of trustees taking affirmative action to protect certificateholders’

interests. He discusses trustees who independently filed claims in bankruptcy proceedings, Fitzgerald Decl. Ex. 2 at 142:23–143:22; trustees who sued to enforce repurchase obligations, Adelson Report ¶¶ 95, 156; a white paper published by Moody’s discussing a trustee that “uncovered serious management problems in the allocation and reporting of funds” in specific transactions, id. ¶¶ 85–86; and GSEs demanding servicing improvements when they act as trustees, id. ¶ 172. Adelson’s experience coupled with supporting examples provide a sufficient basis for his opinion about what an RMBS trustee could do in hypothetical loan default scenarios. The fact that Adelson provides no examples of RMBS trustees actually taking the actions he discusses does not render his opinion that they could have done so inadmissible. The PSA’s

standard that Adelson’s opinion is meant to inform references a “prudent person,” not a “prudent trustee.” Thus, his examples need not be limited to actions he has seen a trustee take. The fact that his examples do not include trustees is fertile grounds for cross-examination but not exclusion. Industry practice at the time relative to defaulting loans does not necessarily control the standard of care, particularly when “the industry was comprised of only a few participants . . . .” FHFA v. Nomura Holding Am., Inc., 873 F.3d 85, 131–32 (2d Cir. 2017). Put differently, following one’s industry colleagues in a race to the bottom does not mean that the bottom reflects prudent conduct. As to Adelson’s prescriptive opinion, the Court finds that it lacks a reliable basis. A substantial portion of Adelson’s opinion, including the contractual PSA language on which he relies, only supports the proposition that RMBS trustees had a duty to act prudently. See Adelson Report ¶¶ 65–114. That proposition is uncontested. The relevant issue that Adelson’s

challenged opinion addresses is whether a trustee acts prudently when it waits for direction from certificateholders to take action. As to that portion of his opinion, Adelson does not connect his experiences to his opinion that a prudent trustee should act without direction, and, on the flip side, that it is imprudent for a trustee to provide notice to certificateholders and then wait for directions on how to proceed. See id. ¶¶ 174–75. His opinion is not supported by any of the examples he provides as support, and to the extent it appeals to “common sense,” it is unhelpful to the trier of fact. Houpt Decl. Ex. 12 at 130:11. The example Adelson provides that is closest to this one factually—a trustee suing a servicer to enforce its repurchase obligation—involved a trustee acting at the direction of certificateholders. See, e.g., Complaint at 1, Bear Stearns Mortg. Funding Tr. 2006-SL1 v. EMC

Mortg. LLC, No. 7701 (Del. Ch.) (July 16, 2012); Complaint at 1, HSBC Bank USA, N.A. v. Merrill Lynch Mortg. Lending, No. 652793/2016 (NY Sup. Ct., filed Aug. 24, 2016). Two discrete pieces of documentary evidence offer some general support for Adelson’s opinion, but they do not provide a sufficient basis for his much more specific opinion that a trustee should independently review and demand repurchase of loans in the event of default without direction from the certificateholder. See, e.g., SEC v. Lek Sec. Corp., 370 F. Supp. 3d 384, 419 (S.D.N.Y. 2019) (finding that “three examples d[id] not provide any adequate basis for an expert’s [broad conclusions based on this non-representative sample]”). Adelson references sanctions that the U.S. Securities and Exchange Commission (“SEC”) imposed on trustees for being negligent with respect to the National Century Financial Enterprise (“NCFE”) default; but he does not explain how the trustees were negligent. Adelson Report ¶ 87; Houpt Decl. Ex. 12 at 213:24–214:7. Adelson also cites to a Federal Deposit Insurance Corporation manual stating that “[t]he trustee must ensure that the collateral, a pool of loans for example, is properly serviced, payments are

properly allocated to principal and interest, and delinquencies are identified and resolved.” Adelson Decl. ¶ 88 (emphasis added). That instruction is as consistent with BNYM’s position as it is with Plaintiffs’ position. The Court thus excludes Adelson’s opinion that a prudent person would have acted without certificateholder direction. B. Opinion About Notices of Defaults Creating Market Pressure Adelson opined that had BNYM publicly disclosed servicers’ non-compliance, the adverse market reaction “would have created substantial pressure for other transaction parties to have addressed the defaults,” resulting in repurchases of defective loans. Adelson Report ¶ 7, 189, 197; Adelson Decl. ¶ 17. This opinion extends to the Countrywide Trusts, as to which BNYM allegedly failed to disclose, pursuant to Reg AB, the magnitude of non-compliance with servicing criteria relating to missing mortgage file documents.8 And it extends to the ECR 2005-

1 Trust, as to which BNYM allegedly failed to disclose, pursuant to TIA, that loans were in default as a result of a master servicer’s failure to enforce repurchases and non-compliance with servicing standards.9 In support of his opinion, Adelson relies on his experience in the RMBS

8 Reg AB disclosures were made with respect to CWALT 2006-OA3, CWALT 2006-OA6, CWALT 2006- OA11, CWL 2006-4, CWL 2006-9 and CWL 2006-14. See Adelson Report ¶ 190 n.81. 9 As to ECR 2005-1, the sole trust subject to TIA, the purported defaults were fourfold: (1) the master servicer’s failure to enforce a seller’s repurchase obligations; (2) the master servicer’s failure to act when the servicer liquidated loans that should have been repurchased; (3) the master servicer’s failure to compel a servicer to abide by accepted servicing standards; and (4) the master servicer’s failure to report (1) – (3) in its annual assessment of compliance with servicing criteria. See id. ¶ 188. industry and contact with structured finance investors. Adelson Decl. ¶¶ 15–17. Adelson also references two “bad news events” as to which he witnessed strong negative market reactions— the NCFE fraud disclosure and the Conseco Financial bankruptcy—and his role in drafting the “autopsy reports” for those events. Houpt Decl. Ex. 12 at 248:2–11; Fitzgerald Decl. Ex. 2 at

265:3–16; Adelson Decl. ¶¶ 18–24. Adelson’s experience does not form a sufficient basis for his opinion. “In order to opine on a hypothetical scenario, an expert must have ‘acted reasonably in making assumptions of fact upon which he would base his testimony.’” MF Glob. Holdings Ltd. v. PricewaterhouseCoopers LLP, 232 F. Supp. 3d 558, 576 (S.D.N.Y. 2017) (quoting Boucher v. U.S. Suzuki Motor Corp., 73 F.3d 18, 21 (2d Cir. 1996)). Adelson stated at his deposition that he has “seen the market react to . . . bad news events plenty of times.” Houpt Decl. Ex. 12 at 248:2–11. He also asserts that investors and other market participants would have found the fact that there were incomplete mortgage files “to be a very important disclosure.” Adelson Decl. ¶ 17. It is, of course, generally true that a “bad news event” can generate market pressure. But what matters here is

whether the specific Reg AB and TIA disclosures would have. Without more, this Court cannot countenance Adelson’s ipse dixit assertions that greater disclosure would have generated significant pressure to comply with the PSAs and that the pressure would have succeeded. His opinion hinges entirely on the NCFE and Conseco events. See Adelson Decl. ¶¶ 18– 24. But Adelson fails to explain how the NCFE fraud and Conseco bankruptcy are analogous to his theory of what would have happened if BNYM had disclosed the extent to which loan files were missing documents. NCFE and Conseco involved different market failures and present no evidence of market pressure launching servicers into action, let alone spurring the repurchase of millions of dollars’ worth of loans. NCFE had issued “roughly $3.35 billion of (supposedly) asset-backed securities . . . that were to have been backed by healthcare receivables.” Id. ¶ 20. In truth, nothing backed those securities. Id. When NCFE’s fraudulent scheme was exposed, the market purportedly reacted with “hysteria.” Id. ¶¶ 20–21. There is no evidence, however, that the market’s reaction caused NCFE to do anything.

The Conseco bankruptcy is even farther afield: a parade of horribles precipitated an investor exodus from whole manufactured housing (“MH”) loan securities. As competition increased in the manufactured housing sector, lenders relaxed their standards in a race to the bottom. Id. ¶ 23. Conseco and other loan servicers filed for bankruptcy, and MH loans deteriorated until the MH securities transactions themselves defaulted and the entire sector was eviscerated. Id. ¶ 23. Although Adelson points to Conseco as evidence that market pressure would have led to the repurchase of loans, the Conseco example illustrates only that news of the bankruptcy of a large player in a sector can lead to investor flight. That is both obvious and irrelevant to the opinion he wishes to offer. At best, the NCFE and Conseco examples are evidence that the disclosures Plaintiffs

contend were required would have led to an earlier demise of the industry. Without adequate support, Adelson’s opinion about what would have followed more detailed disclosures is just “surmise and conjecture.” Amorgianos, 137 F. Supp. 2d at 191. III. Plaintiffs’ Expert Ingrid Beckles Ingrid Beckles is Plaintiffs’ loan servicing expert. BNYM challenges most opinions expressed in the Beckles Report and the Beckles Damages Report. The challenged opinions fall into three categories: (1) servicers of the disputed trusts’ loans did not engage in prudent servicing, (2) servicers would have improved their performance had BNYM provided more oversight, and (3) 18,939 loans had uncured document exceptions that had not been corrected as of December 12, 2018. The Court finds that Beckles’ first two opinions are admissible; BNYM’s motion to exclude the third is stayed pending an evidentiary hearing. A. Opinions About Prudent Servicing BNYM first argues that Beckles provides no basis for her opinion that the standards announced by Government Sponsored Entities (“GSE”) “largely define the concept of prudent

servicing” and that “GSE servicing requirements represent[] prudent and industry standard practices.” BNYM’s Mem. of Law at 19–20. The thrust of BNYM’s challenge is that the PSAs do not incorporate the GSE servicing guidelines; thus, relying upon those guidelines contradicts the plain language of the PSAs. Id. at 20. That argument is a non sequitur. The fact that the PSAs do not expressly incorporate the GSE servicing guidelines does not mean that those guidelines cannot reflect the industry standard of prudence. The PSAs state, to take one example, that the servicer “shall service and administer the Mortgage Loans in accordance with this Agreement and the normal and usual standards of practice of prudent mortgage servicers.” Clark Decl. Ex. 15 § 3.01. That language invites external benchmarks to determine what “prudent” mortgage servicers did during the time

frame at issue in this case. It does not contradict Beckles’s opinion that parts of the GSE guidelines are a relevant benchmark.10 See Fitzgerald Decl. Ex. 13 at 153:27–154:10. Although BNYM takes issue with Beckles’s extrapolation from the standards required for servicing of GSE-owned loans to the industry standards of prudence for servicing privately-owned loans, which are the loans at issue in this case, her opinion stands on firm ground. She has extensive experience with servicing both categories of loans, see Beckles Decl. ¶¶ 3–12, and BNYM, in

10 Nor does it mean, however, that every variance from the GSE servicing guidelines constitutes either imprudent conduct or a variance from industry standards. any event, has not articulated any reason why industry standards for servicing varied based on who owned the loans (let alone provided any evidence that they did vary). B. Opinion About Improved Servicing BNYM seeks to exclude Beckles’s opinion that, if acting prudently, BNYM would have exercised greater oversight of the servicers of the CBASS and ECR 2005-1 Trusts after there had

been events of default, and that greater oversight would have led to servicing improvements within 3 to 6 months.11 See Beckles Report ¶ 124; Beckles Damages Report ¶ 4; Beckles Decl. ¶ 28. BNYM argues that her experience at Freddie Mac—the basis for her opinion—is inapt because Freddie Mac is a large government-sponsored mortgage company with greater leverage to pressure servicers to improve their servicing practices than it had. BNYM’s Mem. of Law at 22. While Freddie Mac’s leverage in the industry was undoubtedly significantly greater than BNYM’s, BNYM has not shown that difference makes Beckles’s opinion unreliable. The contexts are analogous; indeed, it is somewhat self-evident that oversight generally leads to better outcomes, even if reasonable minds can disagree about how much better a result oversight in this context would have generated. Beckles has direct experience monitoring servicing

practices as a master servicer, witnessed the outcomes from oversight, and provides examples of the sort of oversight she contends was required and would have made a difference. See Beckles Report ¶¶ 8–9, 96–102; Beckles Decl. ¶¶ 15–26. In any event, to the extent BNYM takes issue with the transferability of the effects of oversight from Freddie Mac to the effects of oversight from a private trustee, that is fair grist for cross-examination.

11 Those improved servicing practices would, according to Beckles, in turn have reduced the severity of the losses experienced in those trusts. Beckles Report ¶ 91. IV. Plaintiffs’ Rebuttal Expert Bruce Spencer BNYM moves to exclude three opinions of Plaintiffs’ rebuttal damages expert, Bruce Spencer. The Court finds that Spencer’s opinions are admissible. A. Criticism of BNYM’s Damages Expert’s Opinion BNYM challenges Spencer’s qualifications, but the Court finds that his opinion is well

within his expertise. BNYM’s expert, Faten Sabry, opined that Plaintiffs had no reason to infer from the default or loss rates of loans in the disputed trusts that there were underwriting or servicing problems associated with the loans. Sabry Report ¶ 11(d). Sabry bases her opinion on a statistical analysis of the loans in this case relative to loans in comparable loan pools as of 2009 or 2010. Id. ¶ 11(d)(ii)–(iii). Spencer criticized Sabry’s opinion for failing to account for the effect of trustee oversight of servicers that Plaintiffs contend was required. Spencer Rebuttal ¶¶ 14–20. According to Spencer, Sabry’s control group lacks the variable upon which Plaintiffs’ theory of liability hinges—proper servicer oversight. Id. The Court disagrees with BNYM that Spencer must have experience with or understanding of servicer oversight to criticize Sabry on this front. See BNYM’s Mem. of Law

at 24. Spencer is a Professor and Former Chairman of the Department of Statistics at Northwestern and has worked as a statistical consultant for the U.S. Census Bureau, the Department of Justice, and the General Accounting Office, among other agencies. Spencer Rebuttal ¶¶ 3, 7–8. He opined on the statistical import of Sabry’s analysis, relying upon his expertise of how matching estimators are developed and utilized. Id. ¶¶ 13, 15–16, 20. That a statistician must select the appropriate test and control groups to isolate a variable (in this case, servicer oversight) is Statistics 101, and it does not require subject-matter expertise in that particular variable. BNYM’s argument that Spencer may not rely on Beckles and her servicing experience to identify loans that received appropriate servicing oversight is also unpersuasive. See BNYM’s Reply at 15. Spencer applied what he knows—statistical concepts and methods—and Beckles supplied the raw materials for his analysis. BNYM’s position “would, in essence, disqualify all

[statisticians] from conducting a damage assessment unless they first acquire ‘expertise’ in the specific industry in which they purport to opine on damages.” Washington v. Kellwood Co., 105 F. Supp. 3d 293, 309 (S.D.N.Y. 2015) (quotation omitted); see also Arista Records LLC v. Lime Grp. LLC, No. 06-CV-5936, 2011 WL 1674796, at *10 (S.D.N.Y. May 2, 2011) (“[C]ourts have accepted expert testimony from financial and economic analysts even when those witnesses lack industry specific knowledge.” (quotation omitted)). B. Opinion Regarding Loss Severity Similarly, BNYM argues that Spencer’s opinion regarding loss severity is based on an unreliable control group.12 BNYM’s Mem. of Law at 25. BNYM first takes issue with how that control group was selected. Spencer apparently relied upon Beckles to identify the population of loans containing the relevant variables and upon

a consulting group to pull the relevant data files. See Spencer Decl. ¶¶ 14–18. The Court finds no merit in BNYM’s argument. Others may assist so long as the expert’s opinion is within his or her expertise, based upon sufficient facts, and the product of a reliable methodology. BNYM offers no reason to conclude that the data files or Beckles’s assistance compromised Spencer’s analysis. Spencer built the control group and instructed his support staff how to compute loss severity. Id. ¶¶ 18–20. That is enough.

12 “Loss severity” is defined as the “ratio of the total liquidated loss to the outstanding balance of a loan.” Spencer Rebuttal ¶ 26. BNYM’s second argument, that Spencer’s control group contains different loan characteristics than his test group, is also unpersuasive. BNYM asserts that Spencer’s control group contains all prime loans, but the test group contains mostly subprime loans. BNYM’s Mem. of Law at 26. BNYM also asserts that the matched loans are interest-only and yet the

loans in the control group are fully-amortizing, fixed-rate, full-documentation loans with no interest-only or balloon features. Id. Although BNYM has highlighted potential differences between the groups, it has not explained why those differences undermine Spencer’s methods. “Prime” and “subprime,” for example, are general categories encompassing various loan attributes, and, even though used colloquially in the industry, they are imprecise terms concealing the pertinent loan attributes. See Beckles Report ¶ 104; Beckles Decl. ¶ 29. In lieu of relying on descriptions like “prime” and “subprime,” Spencer went deeper: he matched nine specific loan attributes that Beckles identified as impacting loss severity, such as FICO scores and loan-to-value ratios (some of the attributes generally map to the distinctions between “prime” and “subprime” loans). See Spencer Rebuttal ¶ 36; Spencer Decl. ¶ 13. BNYM and

Sabry do not take issue with his method, only his chosen comparators. See Sabry Rebuttal ¶¶ 74–77. After choosing different comparators, Sabry re-ran Spencer’s model, showing that BNYM’s own expert believes the model itself is reliable. See id. ¶ 77 & Ex. 13. For example, unlike Spencer, Sabry also matched loans in the test and control groups on “collateral type,” i.e., prime or subprime. See id. ¶¶ 75, 77. Potential flaws with Spencer’s control group thus go to the weight of his testimony, not its admissibility.13

13 Sabry also points to a report finding “that Prime and Alt A loans have higher recovery rates [and lower severity rates] than subprime or option ARMs.” Sabry Rebuttal ¶ 76. Although tending to cast Spencer’s chosen comparators into doubt, this general finding is additional fodder for cross-examination. Lastly, BNYM points to a purported admission in Spencer’s deposition that certain unmatched variables could impact his analysis; but Spencer so stated only after BNYM asked him to “assume” that “those variables affect[] severity.” Houpt Decl. Ex. 19 at 172:10–174:24. Spencer’s answer does not amount to an admission that those variables affect loss severity—

BNYM’s question does not prove its premise. C. Extrapolation of Loss Severity to Unmatched Loans In his analysis, Spencer attempted to match each loan in the trusts at issue in this case to loans in a control group along factors that affect loss severity. The only pertinent difference, according to Spencer, between the test and control groups was that the latter included the kind of servicing oversight Plaintiffs contend was required while the former did not. Spencer Rebuttal ¶¶ 33, 36–39. Spencer thus claims to have isolated the effect of servicer oversight on loss severity, which he calls the “severity rate loss differential.” Id. ¶¶ 13(c), 40. Although Spencer was not able to match 468 out of 3,548 loans, he imputed the average severity rate differential for matched loans to the unmatched loans. Id. ¶¶ 34, 43. BNYM primarily argues that his imputation was improper because Plaintiffs must prove

trustee breaches “loan-by-loan and trust-by-trust.” Ret. Bd. of the Policemen’s Annuity and Benefit Fund of the City of Chi. v. Bank of N.Y. Mellon, 775 F.3d 154, 162 (2d Cir. 2014); see also Mot. Summ. J. Op. & Order at 13–14. In essence, BNYM seeks to extend Retirement Board’s reasoning about the evidentiary burden to prove liability to evidentiary rulings on proof of damages. But the Second Circuit stated that “misconduct must be proved loan-by-loan and trust-by-trust.” 775 F.3d at 162 (emphasis added). Retirement Board, in this case, does not “preclude[] the use of sampling . . . particularly to establish damages”; for extending a “loan-by- loan” requirement to damages would be unwarranted. Royal Park Invs. SA/NV v. Bank of N.Y. Mellon, No. 14-CV-6502, 2019 WL 6117533, at *3 (S.D.N.Y. Nov. 18, 2019). Indeed, courts in this circuit have approved statistical sampling for proving damages in RMBS cases. Deutsche Bank Nat’l Tr. Co. for Morgan Stanley Structured Tr. I 2007-1 v. Morgan Stanley Mortg. Capital Holdings LLC, 289 F. Supp. 3d 484, 496–97 (S.D.N.Y. 2018) (collecting cases). The Court is more concerned with BNYM’s second argument that the unmatched loans in

Spencer’s analysis appear to have different credit attributes that affect loss severity compared to the matched loans. See Sabry Rebuttal ¶¶ 80, 83. BNYM’s rebuttal expert, Sabry, opined that if the sampled portion is not a representative subset of the unmatched loans along attributes that affect loss severity, then the factual basis for Spencer’s extrapolation evaporates. Id. ¶ 83; see also BNYM’s Mem. of Law at 27–28. Sabry conducted a representativeness analysis of Spencer’s two data sets and found that “unmatched loans have statistically significant different credit attributes than the matched loans . . . across loan and borrower characteristics that may impact severity.” Sabry Report ¶¶ 80, 83. Nonetheless, the Court is not persuaded that this renders Spencer’s sampling analysis unreliable. Whether Spencer can impute the average severity rate differential does not depend on

whether the matched loans are representative across characteristics that affect loss severity, as Sabry opined. Rather, it depends on whether they are representative across characteristics that affect the severity rate differential. Spencer’s analogy to medical studies provides a helpful illustration. See Spencer Rebuttal ¶¶ 29–31; Spencer Decl. ¶ 30. If a clinical trial determines that a new drug has a certain average effect on a sample population, then it is valid to predict that the drug will have that same average effect on a broader population. But just as a drug may have a different effect depending on specific attributes of the person taking the drug, servicer oversight may have a different effect depending on the attributes of the loan. Thus, Spencer, using a “bias adjustment” procedure, ran a regression analysis of factors that correlate to the severity rate differential in order to account for how those factors could distort his extrapolation to the unmatched loans. See Spencer Rebuttal ¶ 41; Spencer Decl. ¶¶ 20, 32; see also Oral Arg. Tr. (Mar. 9, 2020) at 161:9–162:19, 174:12–175:5. BNYM does not identify any issues with the bias adjustment procedure or unaccounted for factors. Reasonable minds can differ on whether

his calculation appropriately deals with the unmatched loans or, as BNYM’s rebuttal expert opined, the sample used in his calculation is fatally unrepresentative.14 Because “[t]he law is well-settled that only serious flaws in an expert’s reasoning or methodology will warrant exclusion,” MBIA Ins. Corp. v. Patriarch Partners VIII, LLC, 09-CV- 3255, 2012 WL 2568972, at *16 (S.D.N.Y. July 3, 2012), the appropriate battleground on which BNYM can attack Spencer’s analysis is “vigorous cross examination” and “presentation of contrary evidence” at trial, Rexall Sundown, Inc. v. Perriqo Co., 651 F. Supp. 2d 9, 28 (E.D.N.Y. 2009) (quoting Daubert, 509 U.S. at 596). V. Plaintiffs’ Expert Joseph Mason Joseph Mason is Plaintiffs’ damages expert. He was retained to “calculate damages

arising from BNYM’s alleged failure to perform duties Plaintiffs contend BNYM was obligated to perform.” Mason Report ¶ 1. Plaintiffs’ case presents a “but-for” world in which BNYM successfully carried out those alleged duties as to the ten at-issue trusts. Those duties, as they

14 Moreover, Sabry’s critique is further deflated when one considers that statistical significance is merely evidence that something besides random error is at work. It does not necessarily mean that those differences are practically important. See, e.g., Waisome v. Port Auth., 948 F.2d 1370, 1376 (2d Cir. 1991) (“[T]hough the disparity was found to be statistically significant, it was of limited magnitude.”). Assuming BNYM is correct that the mismatch along attributes affecting loss severity is cause for concern, the critical question would be the degree to which the different loan attributes Sabry identified affect loss severity. Although Sabry’s assertion that 13% of the loans (i.e., the unmatched set) represents almost 45% of the losses calculated by Mason (under his second damages scenario, but not his first, it should be noted) goes a long way to proving that the findings from the matched set cannot be extrapolated to the unmatched set, it does not explain how differences in the characteristics of the two sets warped Spencer’s extrapolation. See Sabry Rebuttal ¶¶ 79–80, 84. pertain to damages, primarily center on enforcing loan repurchases, which, in Plaintiffs’ view, would have avoided Plaintiffs’ losses from loans that instead liquidated. To calculate damages, Mason models the loans’ principal and interest that Plaintiffs would have received in the “but- for” world and compares those returns to what Plaintiffs did, in fact, receive. Id. ¶ 4.

The Court finds that Mason’s opinions are admissible. To the extent those opinions rely upon opinions that this Court excludes, the inputs into Mason’s damages model must change, but his testimony remains admissible. BNYM challenges four aspects of the Mason Report. BNYM first challenges Mason’s assumption that loans in CWALT 2006-OA3 would have been repurchased in his “but-for” world, notwithstanding the Countrywide Settlement. BNYM’s Mem. of Law at 31. Mason’s model subtracts the Countrywide Settlement proceeds from the losses that Plaintiffs would have avoided had the loans subject to the settlement been repurchased instead of liquidated, offsetting the settlement’s recovery from Plaintiffs’ total damages. In other words, Plaintiffs recover the delta between the losses avoided in Mason’s “but-for” world and the recovery Plaintiffs received from the settlement. BNYM would instead

have Mason remove the loans subject to the settlement altogether from his model to avoid what it calls “double-counting” the proceeds from those loans.15 BNYM’s argument is in fact a challenge to Plaintiffs’ legal theory. BNYM has repackaged its affirmative res judicata defense based on the loans subject to the Countrywide Settlement as a motion to exclude Mason’s opinion. Doubts about the legal sufficiency of Plaintiffs’ theory of liability is not grounds for excluding expert testimony that assumes the

15 BNYM also asserts that Mason’s model “double counts” the Countrywide proceeds. BNYM Mem. of Law at 29. That is incorrect. Mason’s model subtracts what Plaintiffs in fact received from what they would have received in Plaintiffs’ but-for world. What Plaintiffs received includes the settlement payments. viability of that theory. See In re Pfizer Inc. Sec. Litig., 819 F.3d 642, 661 (2d Cir. 2016); see also Olin Corp. v. Lamorak Ins. Co., No. 84-CV-1968, 2018 WL 1901634, at *21 (S.D.N.Y. Apr. 18, 2018). On this motion, whether the settlement payments should be included in the model goes to weight—Mason’s model is still able to produce a reliable result in either BNYM’s

or Plaintiffs’ scenario. See BNYM’s Reply at 19. BNYM next challenges as arbitrary two assumptions in Mason’s damages model about the timing of repurchases. Those assumptions, according to BNYM, are that: (1) repurchases would not have begun until late 2008 and repurchases of already liquidated loans would have occurred within six months of that date, and (2) no defective loan would have been repurchased until its date of liquidation (in other words, Mason backs out the loss on that date to simulate repurchase).16 BNYM’s Mem. of Law at 31–32; BNYM’s Reply at 19–20; see also Mason Report ¶¶ 69–71. A consequence of the second assumption is that Plaintiffs would recover on all the returns from a performing loan up until the date it was liquidated in addition to the recovery on the repurchase, leading to greater damages than if the performing loan were repurchased

before the liquidation date. Although BNYM takes issue with certain inputs, arguing, for example, that Mason should have assumed that repurchases would have occurred sooner, BNYM does not take issue with the model itself. Indeed, BNYM’s expert, Sabry, employed Mason’s model to reproduce his analysis using BNYM’s preferred inputs. See Sabry Rebuttal ¶ 3; Sabry Decl. ¶¶ 2–16;

16 The Court notes that Plaintiffs dispute BNYM’s treatment of Mason’s second assumption. See Pls.’ Mem. of Law at 35. BNYM apparently conflates the date a loan is repurchased with the date that repurchase is demanded, creating the appearance that some repurchases fall beyond the statute of limitations. See BNYM’s Mem. of Law at 33. But Mason, according to Plaintiffs, assumes that repurchases would have occurred in the “but-for” world on the date the loan was liquidated. That date is not the same as the date on which Plaintiffs allege BNYM should have taken steps to enforce repurchases—between 2009 and 2010, or, in other words, within the statute of limitations. Whichever version of Mason’s timing assumption is correct is not determinative of admissibility. Fitzgerald Decl. Ex. 17 at 13:2–14:22. BNYM is free to argue at trial that the timing assumptions underlying Sabry’s version of Mason’s model is better. Although BNYM’s argument relative to timing may be persuasive to a finder of fact, Mason’s choice to ground his model on the date of liquidation is not so unreasonable as to warrant exclusion: it is an objective,

known date when loss—which Plaintiffs avoid in the but-for world—was recorded. BNYM’s reliance upon Washington is misplaced. In Washington, an expert prepared a model calculating damages based on “financial projections and forecasts [from] expected performance and/or prior historical performance.” 105 F. Supp. 3d at 319. The model purported to be a “manifestation of the projections themselves.” Id. at 322–23. But the court found that those projections were pure fabrication. Id. at 322. As such, the damages analysis had no basis other than speculation. Id. at 323. In other words, the expert’s false assumptions were essential to the model itself, and without those assumptions, the model had nothing to offer. Not so here. Although Mason’s assumptions may be flawed, those flaws do not undermine the reliability of his model. Whereas the excluded model in Washington could not be salvaged, BNYM does not

contend that Mason’s assumptions are so essential that the model itself is premised upon “subjective belief or unsupported speculation.” Daubert, 509 U.S. at 590. The alternative inputs that Sabry demonstrates Mason could have used evince the opposite. In sum, BNYM’s motion amounts to an argument that its own assumptions should be substituted into Mason’s model in order to prevent that model from “inflat[ing] the alleged damages.” BNYM’s Mem. of Law at 31. But probing weaknesses in the model’s assumptions in favor of alternative inputs is appropriate grist for cross-examination and for BNYM’s experts’ rebuttal testimony, not for exclusion. See Olin Corp. v. Certain Underwriters at Lloyd’s London, 468 F.3d 120, 134 (2d Cir. 2006) (citing Amorgianos, 303 F.3d at 267). Accordingly, BNYM’s motion to exclude Mason’s opinion is denied. CONCLUSION For the foregoing reasons, BNYM’s motion in limine is GRANTED in part and DENIED in part. No later than April 3, 2020, the parties must submit a joint letter proposing next steps, including a schedule for summary judgment on the four issues proposed at oral argument and, assuming the Court declines to allow a second motion for summary judgment, a schedule for trial. If the parties have identified additional issues for summary judgment, they must include those issues in their joint letter as well. The Clerk of Court is respectfully directed to close the open motion at docket entry 291. SO ORDERED.

Date: March 20, 2020 VALERIE CAPRONI New York, New York United States District Judge