IN RE BERNARD L. MADOFF INVESTMENT SECURITIES LLC

POST-TRIAL FINDINGS OF FACT AND CONCLUSIONS OF LAW DENYING APPLICATION FOR PARTIAL REFUND OF PRIOR SETTLEMENT PAYMENTS

STUART M. BERNSTEIN, Bankruptcy Judge.

At an early stage in this liquidation under the Securities Investor Protection Act ("SIPA"), Irving H. Picard, SIPA trustee (the "Trustee") of the estate of Bernard L. Madoff Investment Securities LLC ("BLMIS"), entered into a settlement agreement with Optimal Strategic U.S. Equity Ltd. ("SUS") and Optimal Arbitrage Ltd. ("Arbitrage") under which they paid the Trustee the aggregate sum of $233,750,000 (the "Optimal Settlement")1 The settlement represented 85% of the amount of the preference claims asserted against them by the Trustee. The Optimal Settlement included an equal treatment provision, or most favored nations ("MFN") clause, which required the Trustee to refund a portion of the settlement payment if he entered into similar future settlements with other defendants for a lower percentage of the claimed amount. SPV Optimal SUS Ltd. ("SPV") acquired SUS's rights and Solus Recovery Fund LP ("Solus," and with SPV, the "Applicants") acquired Arbitrage's rights under the Optimal Settlement.

The Trustee subsequently settled avoidance claims he brought against the above-captioned defendants (collectively, "JPMorgan"). JPMorgan paid the estate $325 million, or slightly more than 76% of the amounts sought (the "JPMorgan Settlement"). The Applicants contend that the JPMorgan Settlement triggered the MFN clause, and the Trustee owes each a refund. Finding the Optimal Settlement ambiguous on this point, the Court conducted a trial on July 30, 2014, and now concludes that the Applicants are not entitled to the refunds they seek.

BACKGROUND2

A. The BLMIS Accounts

The background surrounding the BLMIS Ponzi scheme has been detailed in numerous opinions, and it is unnecessary to repeat those facts beyond what is needed to explain this decision. Madoff, through BLMIS, induced investors to open discretionary trading accounts with BLMIS for the ostensible purpose of buying and selling securities. BLMIS did not engage in any trading, and instead, used newly invested funds to pay "profits" to older investors. As part of the Ponzi scheme, BLMIS generated monthly statements that purported to show valuable securities positions, but the monthly statements and the investments they showed were entirely fictitious.

SUS and Arbitrage are Bahamian trading companies that maintained accounts with BLMIS starting in the mid-1990s. (See Optimal Settlement at ¶ E.) Although their November 2008 account statements, the last statements generated by BLMIS before Madoff's arrest, showed sizeable balances—SUS and Arbitrage held $2,919,934,627.70 and $14,567,639.02, respectively—the account statements were works of fiction, the balances bloated by fictitious profits. In truth, SUS was a net loser, having deposited $1,411,084,183 more than it had withdrawn from its BLMIS account. (Id. at ¶ H.) In addition, it had withdrawn $151,831,876 from its BLMIS account within 90 days of the BLMIS December 11, 2008 filing date (the "Filing Date"). (Id. at ¶ H.)

Arbitrage, on the other hand, was a net winner. It had withdrawn $96,516,185 more than it had deposited into its BLMIS account, including $125,087,004 within the same 90-day period. (Id. at ¶ I.) Based on these facts, the Trustee asserted that both entities were liable to the estate for the withdrawals made within the 90-day preference period under Bankruptcy Code §§ 547 and 550. (Id. at ¶ M.) In addition, the Trustee contended that Arbitrage had received fraudulent transfers totaling $35,000,000 within six years of the Filing Date. (Id.) SUS and Arbitrage denied liability, and contended that as Bahamian entities, the Bankruptcy Court lacked jurisdiction over them and any judgment rendered by the Bankruptcy Court could not be enforced against them in the Bahamas. (Id. at ¶ N.)

B. The Optimal Settlement

The Trustee and SUS and Arbitrage agreed to settle their disputes. However, the case was still in its infancy, the Trustee had not established any specific settlement benchmarks and SUS and Arbitrage were concerned that the Trustee would settle similar claims in the future at a lower percentage. (Transcript of Deposition of Richard Levin, held May 16, 2014 ("Levin Deposition") at 40:15-23;3 Declaration of Marc E. Hirschfield, dated April 9, 2014 ("Hirschfield Declaration")) at ¶ 11.)4 Especially anxious about the latter possibility, SUS and Arbitrage proposed the MFN clause. It was intended to ensure that if a later-settling party received a more favorable settlement, they would receive a partial refund. (Declaration of Richard Levin Regarding Application of the Equal Treatment Provision to the Settlement Agreement Between the Trustee and JPMorgan Chase & Co., et al., dated April 9, 2014 ("Levin Declaration"), at ¶ 4 (ECF Doc. # 67); see also Hirschfield Declaration at ¶ 12.) The Trustee's counsel was "highly skeptical," concerned that an MFN clause would hamstring the Trustee's ability to enter into future settlements. (Levin Deposition at 41:15-42:6.)

On April 6, 2009, Richard Levin, Esq., the attorney for SUS and Arbitrage, circulated the first draft of the MFN clause. (Hirschfield Declaration at ¶ 19; see TX 5.)5 According to the draft, the MNF clause applied to settlements with (1) former BLMIS customers (a) whose statements showed an account balance "in excess of $100 million at any time" (b) who had received preference payments or fictitious profits6 (defined as the "Total Recoverable Amount") in excess of $25,000,000 and (c) had sufficient assets to satisfy the Total Recoverable Amount as of December 10, 2008, (2) the Trustee's other claims against the customer were less than 50% of the Total Recoverable Amount, (3) the settlement was paid before trial or a motion for summary judgment, (4) the settlement was for less than 80% of the Total Recoverable Amount7 and (5) in the Trustee's reasonable judgment, the customer's jurisdictional contacts with the United States were less than SUS's. If a settlement met these criteria, SUS and Arbitrage would be entitled to a refund measured by the difference between the percentage of the Total Recoverable Amount that the customer paid in settlement and the percentage that SUS had paid.

Levin subsequently sent the first full draft of the settlement agreement to the Trustee's counsel, Marc E. Hirschfield, Esq., on April 13 or 14, 2009 (the "April 13 draft"). (Levin Declaration at ¶ 7; Hirschfield Declaration at ¶ 22; see TX 6.) The MFN clause appeared in ¶ 12. (TX 6.) This draft increased the settlement percentage from 80% to 85%, but made no other material changes.

Hirschfield sent the next draft on May 4, 2009 to Levin with a revised MFN clause that now appeared in ¶ 13 (the "May 4 draft"). (TX 7.) This version of the MFN clause applied to the Trustee's settlement of a "Similar Claim" against a "similarly Situated Person or Entity." A "Similar Claim" involved a claim against a former customer for at least $100 million on account of preferences or fictitious profits. The term "Similarly Situated Person or Entity" identified several additional criteria beyond the financial criteria included in the prior drafts as well as the collectability and jurisdictional contacts required or referred to in the prior drafts. Finally, the May 4 draft granted the Trustee the ability to remove a finite number of settlements, at his option, from the coverage of the MFN clause.

Levin returned the next draft (the "May 7 draft") (TX 9) with his May 8, 2009 email. (TX 8). The draft did not alter the substance of paragraph 13, but added language that the Trustee had conducted a confirmatory investigation and "concluded that the Optimal Companies and their affiliates were not complicit in the fraud that BLMIS and Madoff perpetrated on BLMIS's customers and did not have actual knowledge of the fraud." (TX 9 at ¶ P.)

Although the May 7 draft did not alter the MFN clause in the Trustee's May 4 draft, the parties still disagreed. Levin sent a new draft of paragraph 13 on May 9, 2009 (the "May 9 draft"). (TX 11.) It reduced the amount of a qualifying settlement from $100 million to $40 million, and instead of defining "Similar Claim" and "Similarly Situated Person or Entity," the May 9 draft stated that the 85% benchmark applied "if the circumstances of the Avoiding Power Claims and the Qualifying Settlement are similar to the circumstances underlying this Agreement." The May 9 draft adopted five "non-exclusive factors" to be used in determining whether the circumstances of the Avoiding Power Claims and the Qualifying Settlement were similar: (i) the defendant's ability to pay, (ii) "the nature of the Avoiding Power Claims," (iii) the jurisdictional connections with the United States, (iv) the defendant's knowledge of BLMIS's fraud, and (v) the stage of litigation by the Trustee against the defendant. According to Levin, he was trying to break the "logjam," (TX 10), and "bridge the gap" between the broader clause proposed by SUS and Arbitrage and the narrower provision that the Trustee wanted. (Levin Deposition at 57:15-58:17.)

The parties finally reached their settlement on May 22, 2009.8 SUS agreed to pay the Trustee $129,057,094.60 and the Arbitrage agreed to pay $106,323,953.40. (Optimal Settlement at ¶¶ 2, 4.) These amounts represented 85% of the Trustee's 90-day preference claims. Paragraph 13 of the Optimal Settlement included the following MNF clause:

The parties further agreed that they did not intend to generate collateral litigation over whether a particular settlement was made under similar circumstances, and any dispute would be heard in summary fashion by the Court. (Id. at ¶ 13(d).) They also stipulated that their agreement would be governed by New York law, (id. at ¶ 22), and the captions were inserted only for convenience and were not intended to define, limit or describe the scope of the agreement or any of its provisions. (Id. at ¶ 24.)

The Trustee filed a motion to approve the Optimal Settlement on May 26, 2009. (See SPV 2.) The motion informed the Court that the 90-day withdrawals represented "textbook" preference, (id. at ¶ 22), ample evidence supported the preference claims, and there were no defenses. (Id. at ¶ 11, 13.) The Trustee's agreement to accept 85% of the preference claims was intended to avoid the cost and delay of uncertain, protracted litigation. (Id. at ¶¶ 21-22.) He also noted that Arbitrage might be liable under the New York Debtor & Creditor Law for the $35 million it had withdrawn between 2003 and 2007, (id. at ¶¶ 12-13), but released the fraudulent transfer claims for no additional consideration. In addition, SUS and Arbitrage were located in the Bahamas, and the motion cited potential jurisdictional and other issues relating to the Trustee's ability to collect any judgment. (Id. at ¶ 22.) Finally, the Trustee had conducted an investigation of SUS and Arbitrage, and had concluded that they were neither complicit in nor had actual knowledge of the fraud perpetrated by BLMIS. (Id.) Consequently, he had no reason to believe that he had grounds to assert claims other than the avoiding power claims against them or to disallow any claims they might have against the estate, but reserved the right to void the agreement and return the payments if he subsequently discovered information that materially affected his decision to settle. (Id. at ¶ 23; see Optimal Settlement at ¶ 14.) The Court approved the Optimal Settlement on June 16, 2009. (TX 2.)

C. The Settlements with JPMorgan

By early 2014, JPMorgan was fighting on three fronts. The Government was threatening criminal prosecution relating to JPMorgan's activities as BLMIS' banker. In addition, the Trustee had commenced this adversary proceeding against JPMorgan on December 2, 2010. (Complaint: Second Redacted Version, filed April 14, 2011 ("Complaint") (ECF Doc. #18).)9 He asserted two types of claims: (1) to avoid and recover $425 million of fraudulent transfers (the "Avoidance Claims"), and (2) to recover damages on account of JPMorgan's aiding and abetting fraud, aiding and abetting breach of fiduciary duty, conversion, unjust enrichment and contribution (the "Common Law Claims"). The Avoidance Claims, brought under §§ 544, 547-48 and 550 of the Bankruptcy Code, SIPA § 78fff-2 and §§ 270-81 of the New York Uniform Fraudulent Conveyance Act, sought to avoid and recover at least $149,079,881.50 in initial transfers made on account of loan repayments, interest on the loans and account fees, (Complaint at ¶ 277 & Ex. A), and $276,346,673.84 received as subsequent transfers through JPMorgan's redemption of its interests in three BLMIS feeder funds: Fairfield Sentry Limited, Fairfield Sigma Limited and Herald Fund SPC. (Id. at ¶¶ 279-86 & Ex. E.) Following the withdrawal of the reference, JPMorgan moved to dismiss the complaint, the District Court dismissed the Common Law Claims on the ground that the Trustee lacked standing to assert them, and returned the Avoidance Claims to this Court. Picard v. JPMorgan Chase & Co. (In re BLMIS), 460 B.R. 84, 106 (S.D.N.Y. 2011).10 The Second Circuit affirmed the dismissal of the Common Law Claims, Picard v. JPMorgan Chase & Co. (In re BLMIS), 721 F.3d 54 (2d Cir. 2013), and Picard filed a petition for a writ of certiorari. The petition was denied, but only after the global settlement described shortly. See Picard v. HSBC Bank PLC, 134 S.Ct. 2895 (2014).

After the District Court dismissed the Common Law Claims, Stephen and Leyla Hill and Paul Shapiro filed two class action complaints (collectively, the "Class Action") against JPMorgan in the District Court. The Class Action essentially asserted the dismissed Common Law Claims on behalf of the net losers, i.e., investors who had deposited more with BLMIS than they had withdrawn. The District Court eventually placed the Class Action on its suspense docket pending the Supreme Court's decision Roland v. Green, 675 F.3d 503 (5th Cir. 2012), cert. granted sub nom. Chadbourne & Parke LLP v. Troice, 133 S.Ct. 977 (U.S. 2013).

Facing possible criminal and civil liability, JPMorgan and its adversaries entered into a global resolution on January 6, 2014 involving three simultaneous but separately negotiated settlements. See Shapiro v. JPMorgan Chase Co., No. 11 Civ. 8331 (CM) (MHD), 2014 WL 1224666, at *1 (S.D.N.Y. Mar. 24, 2014). First, under a deferred prosecution agreement (the "DPA") with the United States Attorney for the Southern District of New York, (SPV 7), JPMorgan consented to the filing of a two-count Information charging it with the failure to maintain an effective money laundering program in 2008 and the failure to file a suspicious activity report in October 2008. (DPA at ¶ 1 & Ex. B.) It agreed to forfeit $1,700,000,000 to the United States. (Id. at ¶ 3.) In addition, JPMorgan agreed to "accept and acknowledge responsibility for its conduct" as described in the Information and the 85 stipulated facts included in a Statement of Facts attached to the DPA as Exhibit C, (id. at ¶ 12), and further, JPMorgan agreed that "it shall not, through its attorneys, agents, or employees, make any statement, in litigation or otherwise, contradicting the Statement of Facts or its representations in this Agreement." (Id. at ¶ 17.)

Second, JPMorgan and the Trustee entered into the JPMorgan Settlement. (SPV 5.) JPMorgan agreed to pay the Trustee $325 million in settlement of the Avoidance Claims. (Id. at ¶ J.)

Third, JPMorgan settled the Class Action, agreeing to pay $218 million. (SPV 6 (the "Class Action Settlement"), ¶ O.) The Trustee did not receive any additional sums under the Class Action Settlement. He did, however, release his Common Law Claims (which were the subject of the still-pending petition for a writ of certiorari), (id. at ¶ 12), and the Class Action Settlement required the approval of Bankruptcy Court as well as the District Court. (Id. at ¶¶ 1-2.)

The Trustee filed separate motions to approve the JPMorgan and Class Action Settlements on January 7, 2014. In response, the Applicants argued that the JPMorgan Settlement triggered the MFN clause in the Optimal Settlement, and accordingly, they were entitled to refunds of more than $7 million. (Limited Response of SPV Optimal SUS Ltd. to the Proposed Settlement Agreement Between the Trustee and JPMorgan Chase & Co., et al., filed Jan. 28, 2014 ("SPV Limited Response"), at 4-5 (ECF Doc. # 42).) The $325 million exceeded $40 million, represented 76.32% (less than 85%) of the $425 million that the Trustee sought to avoid and recover from JPMorgan, and satisfied the "similarity" factors listed in the MFN clause. (Id. at 5-6.) The Trustee disagreed.

During the February 4, 2014 hearing, the Court stated that it did not understand the distinction that the parties were trying to make between the various types of avoidance actions or the relevance of the knowledge and complicity factor, and found the Optimal Settlement ambiguous "at least" in this respect. (TX 17, at 28:1-23.) As the Applicants did not object to the JPMorgan Settlement or the Class Action Settlement, the Court approved both, and left the question of the Applicants' right to a refund for future determination. (Order Pursuant to Section 105(a) of the Bankruptcy Code and Rules 2002 and 9019 of the Federal Rules of Bankruptcy Procedure Approving Settlement of Avoidance Claims By and Between the Trustee and JPMorgan, filed Feb. 5, 2014, at 3 (ECF Doc. # 51).)

The Court subsequently held an evidentiary hearing on July 30, 2014. The evidence consisted primarily of the various drafts of the Optimal Settlement and the testimony of the two attorneys involved in drafting the agreement. For the most part, they testified about what they contended the MFN clause meant, but did not provide evidence, beyond the exchange of drafts, that touched on their negotiations, or specifically, who said what to who about the MFN clause and its various iterations.

D. The Parties' Contentions Regarding the MFN Clause

1. The Applicants

The Applicants argue that four of the five "similarity" factors in the MFN clause support their right to a refund, and "[n]o single factor is per se determinative, nor is equal weighting of the factors necessarily appropriate." (SPV Brief at 13.) Their holistic approach is expressed in terms of "negotiating leverage." They concede that the nature of the avoiding power claims favors the Trustee, (id. at 14), but contend that the other factors enhanced the Trustee's negotiating leverage to a point that was equal to or greater than the leverage he had during his negotiations with SUS and Arbitrage. For example, JPMorgan stipulated to its "knowledge of the BLMIS fraud," (SPV Brief at 17), and its admission "acts as a counterweight," to the nature of the claim and "cuts strongly in favor of" the Applicants. (Id. at 17-18.) In addition, SUS and Arbitrage settled with the Trustee before a suit was filed, but JPMorgan settled "after years of pre-complaint discovery and . . . motion practice at every level of the federal judiciary." (Id. at 18-19.) This factor favors the Applicants because an early settlement is "worth more" to the Trustee, (id. at 18), and if SUS and "had settled only in 2014, rather than in 2009, one would have expected the Trustee . . . to settle . . . at a higher percentage." (Id. at 20.) In addition, the jurisdictional connections of the defendant and the enforceability of a judgment favor the Applicants because SUS and Arbitrage were Bahamian companies with no assets or presence in the United States "other than its BLMIS accounts," and the Trustee would have faced jurisdictional obstacles that were not present in its suit against JPMorgan. (Id.)

Finally, the Applicants assert that SUS and Arbitrage on the one hand and JPMorgan on the other had the ability to satisfy any judgment. This would appear to render this factor neutral. The Applicants nevertheless maintain that it cuts against the Trustee because JPMorgan's "ability to pay gave the Trustee no reason to reach a deal with JPM at a lower percentage than with [SUS and Arbitrage]." (Id. at 21.)

2. he Trustee's Contentions

The Trustee's principal argument is that the MFN clause only applies to settlements of straightforward avoidance claims for the recovery of preferences or fictitious profits transferred to BLMIS customers that are difficult to defend. (Trustee's Memo at 19, 21.) The earlier drafts were limited to this narrow class of settlements, the parties never discussed a broader reach, and the switch to the five-factor test was intended to "bridge the gap" between the Trustee's and SUS' and Arbitrage's versions of the MFN clause. (Id. at 20.) Furthermore, while ¶ 24 of the Optimal Settlement states that paragraph headings do not define, limit or describe the scope of the agreement, the ¶ 13 heading language was never changed ("Equal Treatment for SUS and Arbitrage with Other Similar BLMIS Customers"). This serves as additional evidence that the change in the MFN clause from "customers" to "defendants" was not intended to expand the scope of the provision. (Id. at 25-26.) The circumstances of the settlements were also starkly different. The JPMorgan Settlement was part of a complex "global resolution" among the Trustee, the class representatives involving avoidance, criminal, and common law claims, whereas Optimal Settlement covered "simple preference claims" involving only SUS and Arbitrage. (Id. at 29, 31.) In addition, the settlements with JPMorgan were part of a global resolution in which the Trustee released the Common Law Claims as well as the Avoidance Claims. (Id. at 37.) The stage of litigation was also dissimilar because Optimal settled before discovery and the filing of a complaint, (id. at 37), but JPMorgan settled after the Trustee suffered adverse rulings, warranting a lower settlement amount. (Id. at 24; see id. 37-38.) Lastly, the Trustee asserts that there was little extrinsic evidence regarding the collectability of a judgment against SUS and Arbitrage or their jurisdictional contacts with the United States.11 (Id. at 25.)

DISCUSSION

This is a contract dispute governed by New York law. "The fundamental, neutral precept of contract interpretation is that agreements are construed in accord with the parties' intent." Greenfield v. Philles Records, Inc., 780 N.E.2d 166, 170 (N.Y. 2002). "A written agreement that is clear, complete and subject to only one reasonable interpretation must be enforced according to the plain meaning of the language chosen by the contracting parties." Brad H. v. City of New York, 951 N.E.2d 743, 746 (N.Y. 2011). If the contract is ambiguous, the court may resort to extrinsic evidence. In determining the meaning of an ambiguous provision, reasonable certainty rather than absolute certainty is the goal. Travellers Int'l AG v. Trans World Airlines, Inc., 722 F.Supp. 1087, 1102 (S.D.N.Y. 1989). The court should not add or excise terms, or "imply a term where the circumstances surrounding the formation of the contract indicate that the parties, when the contract was made, must have foreseen the contingency at issue and the agreement can be enforced according to its terms." Reiss v. Fin. Performance Corp., 764 N.E.2d 958, 961 (N.Y. 2001).

The Court must look to the objective manifestations of intent as gathered from the parties' words and deeds. SR Int'l Bus. Ins. Co. v. World Trade Ctr. Props., LLC, 467 F.3d 107, 125 (2d Cir. 2006). "Although a party's uncommunicated subjective intent cannot supply the ultimate meaning of an ambiguous contract," id. at 125, "the parties' reasonable expectations of a contract's meaning are determined based on an objective understanding of the surrounding circumstances in which the agreement was made, and `a party's subjective understanding, while not controlling, may shed light on the state of th[e] negotiations and could bear on that party's objective actions.'" Chesapeake Energy Corp. v. Bank of New York Mellon Trust Co., 957 F.Supp.2d 316, 342 (S.D.N.Y. 2013) (quoting World Trade Ctr. Props., 467 F.3d at 126). Finally, to the extent that a trial and other aids to interpretation do not resolve the ambiguities, they must be construed against the drafters under the doctrine contra proferentem. 151 West Assocs. v. Printsiples Fabric Corp., 460 N.E.2d 1344, 1345 (N.Y. 1984) ("It has long been the rule that ambiguities in a contractual instrument will be resolved contra proferentem, against the party who prepared or presented it.").

When the Court approved the settlements on February 4, 2014, the Applicants interposed a limited objection relating to their right to a partial refund. The Court did not rule on the issue then, but opined that the Optimal Agreement was ambiguous. The Court focused on the five-factor test, but further exposition revealed other ambiguities, the principal one centering on the meaning of "similar." The parties were attempting to target more favorable settlements of "similar" claims; the Trustee insisted on a narrow MFN clause and SUS and Arbitrage a broader one. The five-factor test was intended to "break the logjam" and "bridge the gap." It did not succeed in bringing clarity as this dispute shows, and to the extent the MFN clause still remains unclear, it must be construed against the Applicants' predecessors, the drafters.

The MFN clause uses the word "similar" in three contexts. First, the Avoiding Power Claims subject to the Qualifying Settlement must be "similar" to the Avoiding Power Claims asserted against SUS and Arbitrage. (Optimal Settlement at ¶ 13(a).) Second the "circumstances" of the Avoiding Power Claims must be "similar to the circumstances underlying this Agreement." (Id. at ¶ 13(b)(iii).) Third, the "circumstances" of the Qualifying Settlement must also be "similar to the circumstances underlying this Agreement." (Id.) The five-factor test applies to determine "whether the circumstances of the Avoiding Power Claims and the Qualifying Settlement are `similar,'" (id. at 13(c)), implying that it does not govern the initial decision under ¶ 13(a) whether the Avoiding Power Claims are similar. The Avoiding Power Claims asserted against SUS and Arbitrage sounded in preference and fictitious profits, and the Trustee has argued that these are the only types of claims encompassed within the MFN clause. In fact, these were the only types of avoidance claims mentioned in the earlier drafts prior to the introduction of the five-factor test.

Other provisions in the Optimal Settlement render this reading suspect. First, the definition of "Avoiding Power Claims" is broad, and includes turnover proceedings under Bankruptcy Code § 542 as well as Bankruptcy Code §§ 544, 547 and 550.12 Claims to turnover property of the estate under Bankruptcy Code § 542 are distinct from preference and fraudulent transfer claims. See Savage & Assocs., P.C. v. BLR Servs. SAS (In re Teligent, Inc.), 307 B.R. 744, 751 (Bankr. S.D.N.Y. 2004) ("The trustee cannot compel the turnover of non-estate property under 11 U.S.C. § 542, and circumvent the more restrictive fraudulent transfer claim in the process."). Second, the statutory references in the definition are broad enough to encompass more than just preference and fictitious profits claims. Third, in determining whether the "circumstances" are similar, the Court must consider "the nature of the Avoiding Power Claims (such as whether they are for recovery of a preference or for recovery of principal or fictitious profits)." (Optimal Settlement at ¶ 13(c)(ii).) This factor implies that the Qualifying Settlement can involve Avoiding Power Claims that are not preference or fictitious profits claims. At a minimum, the Qualifying Settlement can involve a claim for the return of principal. Fourth, even the Trustee concedes that nature of the avoiding power claim is not dispositive. (Trustee's Memo at 21-22 ("[C]laims for preferences and claims for fictitious profit may both be `similar' to the Optimal settlement but neither is necessarily or dispositively `similar.'").) The Court must give effect to all of the provisions of the Optimal Settlement, and even though the parties' earlier drafts focused on preference and fictitious profits claims, the final version cannot be read as limited to those types of claims. For similar reasons, I also reject the Trustee's position that the Qualifying Settlement must involve a claim against a customer. The earlier drafts referred to claims against customers and included objective criteria that only a customer could meet. For example, the settling customer had to maintain a BLMIS account that achieved a balance of at least $100 million at some point. The later drafts, including the final version of the MFN clause, dropped the requirement that the settling party had to have an account or a minimum balance. Furthermore, the final version substituted "defendant" for "customer" to identify the settling party. And while Levin did not modify the heading of the MFN clause (which refers to claims against customers), his testimony at trial implied that it was an oversight. (7/30/14 Trial Transcript ("Tr.") at 13:18-23.) Finally, ¶ 24 of the Optimal Agreement stated that the captions were inserted for convenience and did not affect the content of the agreement.

While I reject the Trustee's narrower reading of the Optimal Settlement, I nevertheless conclude that the JPMorgan Settlement did not trigger the refund obligation. Notwithstanding the ambiguities in the MFN clause, the parties were obviously attempting to craft criteria that would identify settlements that would be considered similar to the Optimal Settlement. The MFN clause's factors correspond to the considerations that normally drive a settlement: the risks and rewards of litigation (i.e., establishing liability) and collectability of a judgment.13 Consideration of the nature of the Avoiding Power Claims subsumes the principal issues that affect the risks and rewards, and this case highlights why the two sets of claims are so dissimilar.

The Trustee asserted a preference claim against SUS and preference and fictitious profits claims against Arbitrage. These were essentially strict liability claims, and even SPV's counsel conceded that they are "difficult to defend." (Levin Deposition at 17:2-11.) Moreover, their willingness to settle for 85% without a fight is telling. The 15% was essentially a tip that I assume the Trustee gave to encourage SUS and Arbitrage to write a check and spare him the time and effort of recovering and enforcing a judgment. And while SUS and Arbitrage had raised issues of personal jurisdiction, which the Trustee mentioned in his application to approve the settlement, these issues did not appear to raise any particular concerns. Indeed, subsequent litigation involving other defendants has confirmed that defendants who invested directly or indirectly with BLMIS and received payments from BLMIS as initial transferees or as subsequent transferees of those initial transferees were subject to the Court's personal jurisdiction. E.g., Picard v. Bureau of Labor Ins. (In re BLMIS), 480 B.R. 501, 517 (Bankr. S.D.N.Y. 2012) ("In a nutshell, BLI invested tens of millions of dollars in Fairfield Sentry with the specific purpose of having funds invested in BLMIS in New York, and intended to profit from this U.S.-based investment. As such, BLI cannot claim a violation of its due process rights from having to appear in a New York court to defend itself in a suit arising from activities with a clear New York nexus."); Picard v. Maxam Absolute Return Fund, L.P. (In re BLMIS), 460 B.R. 106, 118 (Bankr. S.D.N.Y. 2011) ("[S]everal of the Trustee's claims `arise out of or relate to' these contacts with the United States such that Maxam Limited `should reasonably anticipate adjudication of these transactions to take place' here. [Citation omitted.] Without Maxam Limited's transferring assets to and from the United States, there could not be claims for subsequent transfers against it."); Picard v. Chais (In re BLMIS), 440 B.R. 274, 279-82 (Bankr. S.D.N.Y. 2010) (personal jurisdiction existed over Israeli defendant who maintained two accounts with BLMIS in New York, the Trustee's claims arose out of or related to these accounts and but for the transfers to and from the BLMIS accounts, there could be no fraudulent transfer claims against her).

The claims asserted against JPMorgan were of a different and more demanding ilk. The initial transfer claims arose from the repayment of loans made in 2005 and 2006, and related interest and account fees, only a small portion of which were recoverable as a preference. JPMorgan obviously gave value to the extent of the loans, and the Trustee's ability to recover the repayment of principal faced several obstacles. In its motion to dismiss the Complaint, JPMorgan argued that the repayment of the loan was effected through a setoff, and a setoff is not a "transfer" that is subject to avoidance under Bankruptcy Code §§ 544, 547 or 548.14 (TX 18, at 56-59.) In addition, JPMorgan was fully secured, its repayment did not deplete the estate or harm creditors, and consequently, could not be avoided and recovered. (Id. at 59-60, 61-62.) Finally, any claim to recover loan repayments as fraudulent transfers was barred by the Second Circuit's decision in Sharp Int'l Corp. v. State Street Bank & Trust Co. (In re Sharp Int'l Corp.), 403 F.3d 43 (2d Cir. 2005) absent proof that the transferee knowingly participated in the borrower's fraud. (TX 18, at 62-66.)

Although not part of the motion to dismiss, the subsequent transferee claims faced other hurdles. The Trustee would first have to demonstrate that the initial transfers to the three feeder funds were avoidable.15 He would then have to trace the avoidable BLMIS transfers from the feeder funds to JPMorgan to show that JPMorgan was a subsequent transferee. See IBT Int'l, Inc. v. Northern (In re Int'l Admin. Servs., Inc.), 408 F.3d 689, 708 (11th Cir.2005). Finally, the Trustee cannot recover from a subsequent transferee "that takes for value, including satisfaction or securing of a present or antecedent debt, in good faith, and without knowledge of the voidability of the transfer avoided." 11 U.S.C. § 550(b)(1). Thus, the issue of JPMorgan's knowledge of and complicity in the BLMIS fraud, a consideration under the five-factor test, increased the risk that the Trustee would fail to prove JPMorgan's liability for the initial transfers, the subsequent transfers, or both.

The Applicants say "no problem"; JPMorgan admitted its knowledge of the BLMIS fraud in the DPA. They directed the Court's attention to certain of the stipulated facts incorporated into the DPA that JPMorgan could not challenge in the litigation brought by the Trustee, including that

Without minimizing the significance of JPMorgan's admissions, they suggest that JPMorgan willfully blinded itself to the "red flags" generated by Madoff's investment advisory business, but do not necessarily mean that JPMorgan had actual knowledge of the BLMIS Ponzi scheme. See Picard v. Merkin (In re BLMIS), 515 B.R. 117, 139-40 (Bankr. S.D.N.Y. 2014) (comparing actual knowledge and willful blindness). In addition, the stipulated facts regarding JPMorgan's suspicions of fraud primarily centered on events in 2007 and 2008, (DPA, Ex. C at ¶¶ 21, 38-61), and in particular, an internal October 16, 2008 JPMorgan internal memorandum. (DPA, Ex. C at ¶¶ 54-61.) The Information attached to the DPA focused exclusively on 2008. (DPA, Ex. B.) JPMorgan, however, made the loans to BLMIS in 2005 and 2006, and BLMIS repaid the $145 million loan to JPMorgan on June 1, 2006. (Complaint, Ex. A at 2.) Under Sharp International, JPMorgan's knowledge that BLMIS had used the proceeds of a fraudulent scheme to repay the loan would not render the loan repayment avoidable except as a preference, and the repayment fell outside of the 90-day preference window. Thus, JPMorgan's admissions did not transform the Trustee's case into a "slam dunk" that was "difficult to defend," and even the Applicants conceded that "the trustee had a harder case to prove against JPMorgan [and] [w]ho knows if he would have done it or not." (Tr. at 42:15-17.)

The issue of knowledge of the BLMIS fraud is not only relevant to the nature of the Avoiding Power Claims; the defendant's knowledge and complicity was included as a separate factor under the five-factor test. The parties offered different reasons for its inclusion at trial. According to Levin, SUS and Arbitrage feared that the Trustee could settle multiple claims against the same party, allocate the largest portion of the settlement to the preference and fictitious profit claims and evade the MFN clause. (Tr. at 16:23-17:7.) Levin included the knowledge and complicity factor to prevent the Trustee from allocating the largest portion of a multi-claim settlement to the preference claim even though he had a solid case to recover an intentional fraudulent transfer under Bankruptcy Code §§ 548(a)(1)(A) and 548(c). (Levin Declaration at ¶ 21(b).) Furthermore, the defendant's inability to sustain a good faith defense under § 548(c) gave the Trustee "leverage in negotiating a settlement that was at least as strong as the leverage he had when negotiating with SUS and Arbitrage, making the circumstances of any resulting settlement more similar." (Id.) Hirschfield testified that the knowledge and complicity factor, when read in conjunction with the nature of the Avoiding Power Claim, was intended to limit "similar" claims to the type asserted against SUS and Arbitrage— "those for which no intent of the transferee need be shown—i.e., preferences or fictitious profits." (Hirschfield Declaration at ¶ 48.) Cases requiring proof of knowledge or complicity required "higher standards of pleading and proof and thus the Trustee wanted the flexibility to accept less than 85% without triggering the most favored nation provision." (Id.)

The Court originally found this provision to be ambiguous based on the Applicants' argument that the factor imposed an obligation on the Trustee to demand higher settlements from morally guilty defendants. (See SPV Limited Response at 8-9 (If "the defendant knew of the BLMIS fraud or was complicit in it in any way . . . the Trustee should not be willing to settle for less. . . . [T]he Trustee should not be willing to settle for less than he settled with SUS, because no other defendant could have been less complicit than the Trustee found SUS.").) The Applicants have abandoned this argument, and their concerns regarding the misallocation of the settlement do not apply to the JPMorgan Settlement. Furthermore, neither side offered any evidence regarding negotiations over this factor. In the end, I resolve any ambiguity in the Trustee's favor and conclude that this factor, though separate, was intertwined with the nature of the Avoiding Power Claim, and related to the difficulty in proof. If the settled claim involved questions of knowledge and complicity and the evidence favored the Trustee's case, the settled claim was more "similar" to the claims against SUS and Arbitrage; if he lacked the proof, the settled claims would be less "similar." For the reasons discussed, the Trustee faced a greater risk of litigating the Avoidance Claims against JPMorgan than he faced suing SUS and Arbitrage to avoid and recover indefensible preferences. Consequently, this factor weighs against the finding that the two sets of claims are "similar" within the meaning of the MFN clause.

The jurisdictional, enforceability and collectability factors received mention, but apparently, limited consideration. SUS and Arbitrage had contacts with the United States, receiving transfers from BLMIS accounts maintained in New York. They are or were multi-million dollar hedge funds, and the Applicants have conceded that their ability to pay and JPMorgan's ability to pay were similar. (SPV Brief at 21.) There is certainly no suggestion, much less any evidence, that after BLMIS collapsed they invested all of their substantial assets in localities beyond the ultimate reach of an American judgment. Nor does the record show that the Trustee thought long or hard about his ability to sue SUS and Arbitrage in New York or enforce the inevitable judgment, and judging by their willingness to settle for 85% before the Trustee fired a shot, neither did SUS nor Arbitrage.

The last factor, the stage of the litigation, is difficult to assess. The Trustee settled with SUS and Arbitrage early on and before any litigation. The Trustee settled with JPMorgan after extensive litigation, and as part of a global resolution involving the claims asserted by the class representatives and the Government as well as the Trustee. The Applicants imply that the Trustee should have been expected to settle with JPMorgan for a higher percentage than 85% because of the advanced stage of the litigation. (SPV Brief at 20 ("Had [SUS and Arbitrage] settled with the Trustee after having put the Trustee through as much litigation as JPM did, and had settled only in 2014, rather than in 2009, one would have expected the Trustee to have agreed to settle only at a higher percentage.") This argument is based on the "common practice in bankruptcy cases that early settlements of avoiding power litigation are usually less than settlements at later stages." (Levin Declaration at ¶ 21(e).)

I cannot take judicial notice of any such common practice. It may be true that a defendant who takes a hard line on settlement at an early stage softens and is willing to pay more as the case progresses and he spends more money defending it. By the same reasoning, a plaintiff might take a tough settlement stance at the outset but relent and accept less as the case drags on. Here, there is no reason to suppose that the settlement value of the claims against JPMorgan increased as the litigation progressed. By the time of the JPMorgan Settlement, the Common Law Claims had been dismissed, and the motion to dismiss the Avoidance Claims was still pending. In addition, JPMorgan possessed significant assets and the incentive to defend the Trustee's $425 million lawsuit. It is more plausible that the pressure of a criminal conviction coupled with the desire to reach a global settlement of its "Madoff problem" was the catalyst that triggered JPMorgan's willingness to settle when it did for the amounts it agreed to pay.

In summary, the JPMorgan Settlement did not trigger the Applicants' right to a refund. The nature and circumstances of the Avoiding Power Claims asserted against JPMorgan were significantly different, more difficult to prove and the outcome less certain than the preference and fictitious profit claims that SUS and Arbitrage settled. Finally, there is no evidence that their status as Bahamian entities affected the amount of the settlement or any reason to believe that it did.