Full opinion text
OPINION SWEET, District Judge. The defendants in seven related securities fraud actions have moved for summary judgment as to various claims. The plaintiffs have opposed these motions, and certain plaintiffs have made cross-motions for summary judgment. Specifically, Kidder, Peabody & Co. (“Kidder”) and Donaldson, Lufkin & Jenrette Securities Corp. (“DLJ”), the broker defendants (collectively, the “Brokers”) in the six actions entitled ABF Capital Mgmt. v. Askin Capital Mgmt., No. 96 Civ. 2978 (the “ABF Action”), Johnston v. Askin Capital Mgmt., 97 Civ. 4335 (the “Johnston Action”), Primavera Familienstiftung v. Askin, No. 95 Civ. 8905 (the “Primavera Action”), Montpellier Resources Ltd. v. Askin Capital Mgmt., No. 97 Civ. 1856 (the “Montpellier Action”), Bambou Inc. v. Askin, No. 98 Civ. 6178 (the “Bambou Action”), and AIG Managed Market Neutral Fund v. Askin Capital Mgmt., No. 98 Civ. 7497 (the “AIG Action”) (collectively, the “Investor Actions”), have moved for summary judgment against all plaintiffs (the “Investors”) on Count II of the complaint, which is the sole count remaining against them and which alleges aiding and abetting fraud. Kidder has also moved separately against certain Investors on statute of limitations grounds, and against those Investors who invested in the Quartz Hedge Fund (“Quartz”) on grounds specific to those Investors (the “Quartz Investors”). Defendants Askin Capital Management, L.P. (“ACM”) and David J. Askin (“Askin”) (collectively, the “ACM Defendants”) have joined in the motions by DLJ and Kidder to the extent applicable. In addition, defendants Merrill Lynch, Pierce, Fenner & Smith Inc. (“Merrill”) and DLJ (collectively, the “Brokers”) have moved for summary judgment in the seventh related action, entitled Granite Partners, L.P. v. Bear Stearns & Co., Inc., No. 96 Civ. 7874 (the “Funds Action”), against Granite Partners, L.P. (“Granite Partners), Granite Corporation (“Granite Corp.”), and Quartz (collectively, the “Funds”), suing by and through the Litigation Advisory Board (the “LAB”). DLJ has moved for summary judgment on Count I of the Second Amended Complaint, which count alleges breach of contract by DLJ for wrongful margin calls. Merrill has moved for summary judgment on the three counts remaining against it, namely, Count I, alleging breach of contract for improper margin calls, Count II, alleging breach of contract for bad faith liquidations, and Count VIII, alleging commercially unreasonable liquidations in violation of Article 9 of the Uniform Commercial Code (the “U.C.C.”). Merrill has also moved to strike the expert reports submitted by the Funds in connection with the motions for summary judgment. The Funds have cross-moved against Merrill on Counts I, II and VIII. In addition, the Funds have cross-moved against DLJ on Count X of the Second Amended Complaint, which count objects to DLJ’s deficiency claim against the Funds in the related bankruptcy proceeding, and Count IX, which count alleges that DLJ failed to turn over certain principal and interest payments owed to the Funds. Finally, the ABF Plaintiffs have moved for an order removing the “confidential designation” from documents produced by DLJ and Kidder, and from the deposition testimony of present or former employees of those firms, in the Investor Actions. For the reasons set forth below, the motions will be denied in part and granted in part. The Parties and Prior Proceedings The parties in the instant actions are set forth in the prior opinions of this Court, familiarity with which is assumed. See Granite Partners, L.P. v. Bear, Stearns & Co., 17 F.Supp.2d 275 (S.D.N.Y.1998) (“Granite /”); Granite Partners, L.P. v. Bear, Stearns & Co., 58 F.Supp.2d 228 (S.D.N.Y.1999); ABF Capital Mgmt. v. Askin Capital Mgmt., L.P., 957 F.Supp. 1308 (S.D.N.Y.1997) (“ABF I”). Previous proceedings are also set forth in the prior opinions of this Court. Extensive discovery has been had in these actions, involving the exchange of tens of thousands of pages of documents and the deposition of dozens of witnesses. Proceedings relevant to the instant motions are set forth below. The summary judgment motions in the Investor Actions were filed on or about May 12, 2000, and submissions were received through September 1, 2000, at which time the matter was deemed fully submitted. The summary judgment motions in the Funds Action were filed on or about June 6, 2000, and submissions were received through July 28, 2000, at which time the matter was deemed fully submitted. Kidder’s motion in the Funds Action to strike the Funds’ expert reports was filed on July 12, 2000, and was heard and deemed fully submitted on September 20, 2000. The ABF Plaintiffs’ motion in the Investor Actions to strike the “confidential” designation from discovery materials was filed by letter of September 20, 2000, and was heard and deemed fully submitted on October 11, 2000. The Facts The following facts are drawn from the parties’ Rule 56.1 Statements and other submissions and, as required, are construed in the light most favorable to the non-movant, as applicable. They do not constitute findings of fact by the Court. Overview of the Transactions and the Funds’ Collapse The Funds were “hedge funds” which made leveraged investments in the mortgage-backed securities market, including collateralized mortgage obligations (“CMOs”). CMOs are bonds created from and collateralized by mortgage-backed securities formed from pools of residential mortgages or securities backed by such mortgages. CMOs are not listed or traded on a public exchange. The Granite Partners and Granite Corp. Funds (collectively, the “Granite Funds”) were established in 1990. Between the time of their creation and September 1991, the investment advisor for these Funds was New Amsterdam, run by Tony Estep (“Estep”). In September 1991, Askin joined the Funds as their president and investment advisor, and in September 1993, ACM was created as the Funds’ investment advisor. The Quartz Fund was established in January 1994 with ACM as its investment advisor. The Investors, who include both individuals and institutional investors, were shareholders and/or limited partners in the Funds. The earliest purchase of an interest in the Funds occurred in or about September 1990, and the latest occurred in March 1994. A number of the Investors acquired additional interests in the Funds after their initial investment. Askin and ACM purchased CMOs for the Funds from various brokers, including Merrill, DLJ, Bear Stearns, and Kidder. The brokers created the CMOS. CMOs are created from and divided into various classes, or “tranches,” each of which is entitled to a different portion of the principal and/or interest payments made by the underlying mortgage obli-gors. The tranches differ from one another with respect to their sensitivity to interest rate changes and the certainty with which their reaction to such changes can be predicted. The mortgage-backed securities market is complex and relatively illiquid. Although mortgage-backed securities, including CMOs, count among their benefits relatively high yields, these securities also carry with them certain risks. The two primary risks associated with these instruments are interest rate risk and prepayment risk. The interest rate risk is the risk that the price of the security will decrease in response to interest rate increases. The prepayment risk is the risk that homeowners may prepay their mortgages, with the result that cash flows generated by a pool of mortgages may fluctuate as homeowners pay down their mortgages at faster or slower rates. Movements in interest rates, among other factors, affect prepayment speeds. The Granite Funds were intended to take advantage of the mortgage-backed securities’ potential for high returns while being market-neutral by constructing portfolios comprised of a balanced mix of “bullish” and “bearish” securities. A bullish security is likely to increase in value when interest rates fall and decrease in value when interest rates rise, and is more volatile. A bearish security is likely to decrease in value when interest rates fall and increase in value when interest rates rise, and is more stable. The Quartz Fund was intended to be market-directional, that is, it was to maintain a bullish or bearish portfolio depending on the predicted direction of interest rates. The Granite Funds purchased many of the most complex and esoteric CMOs in existence. The more complex and esoteric CMOs are relatively illiquid. Indeed, the Brokers referred to the riskiest tranches— those most prone to large and unpredictable swings in value — as “toxic” or “nuclear waste.” The Funds primarily acquired their CMOs pursuant to repurchase agreements or “repos.” A repo is a financing mechanism that allowed the Funds to pay only a fraction of the cost of each CMO in cash, borrowing the balance from the brokers. In such a transaction, one party to the agreement agrees to sell a security to a buyer/lender for a given sum (the “repo amount”) and to buy the security back from the buyer/lender at a later date (the “buy-back date”) for the repo amount plus a market rate of interest (the “repo rate”). The buyer/lender holds the security in a “repo account.” Although repos were a means for the Funds to acquire CMOs, in these transactions the Funds acted as a “seller” and the broker acted as a “buyer” of CMOs. In effect, repos are collateralized loans. The Brokers loaned the Funds most of the purchase price for each CMO and took possession of the CMOs as security, ie., collateral, for the Funds’ performing their obligations to repurchase the CMOs, ie., repay the loan, with interest, on the buyback date. The repo buyer (the broker) obtained a security interest in the transferred securities. The use of repos bene-fitted the Brokers by allowing the Funds to increase their purchases of CMOs from the Brokers. The repo buyer (the broker) always paid (loaned) an amount less than the actual value of the security sold (transferred as security for the loan) by the repo seller (the Fund) as protection against default on the obligation to repurchase the security (pay back the loan). The difference between the amount loaned and the security’s value is the “haircut.” Thus, if a broker took a 20 percent haircut on a security valued at $100,000, then the repo amount (the amount loaned) to the Fund was $80,000 for that $100,000 security. If the value of the securities in a repo account fell below an amount agreed upon by the parties, the “margin amount,” then there was a “margin deficit” and the broker had the right to make a “margin call,” ie., to demand money or additional securities as collateral for the loan. If a proper margin call was not met, the broker had the right to liquidate the securities in the repo account. On February 4, 1994, the Federal Reserve Board raised interest rates by a quarter of a point, which was the first rate increase in approximately five years. Interest rates were raised again on March 22, 1994. As interest rates rose, prepayments on the underlying mortgage obligations for mortgage-backed securities fell. ACM magnified the effect of these market conditions by purchasing inappropriately bullish securities that were particularly sensitive to these adverse market conditions. In addition, the Funds were highly leveraged and had an excessive degree of negative convexity. The value of the Funds’ portfolios plummeted during this period. Between March 28 and March 31, 1994, the various brokers with which the Funds had entered into repo transactions issued a veritable blizzard of margin calls on the Funds, beginning with a $30 million margin call from Bear Stearns. Merrill and DLJ were two of eleven broker-dealers to make margin calls on the Funds at this time. All told, the margin demands amounted to more than $131 million. In response to the margin calls, the Funds transferred approximately $49 million in cash or unencumbered collateral to the various brokers, leaving a shortfall of almost $82 million. By March 30, 1994, the Funds’ short-term obligations exceeded their cash and unencumbered securities by approximately $60.4 million. Indeed, two days earlier, Askin had asked the Investors for an additional $120 million in capital to use, in part, to pay rapidly-mounting margin calls. As the Funds were unable to meet the margin calls, the brokers liquidated the Funds’ portfolios. The Funds collapsed and filed for bankruptcy under Chapter 11 on April 7, 1994. The Investors allegedly lost approximately $230 million in investments. The Investor Actions Between 1991 and 1993, the Funds’ portfolios were managed by Askin and John Contino (“Contino”). In 1993, Contino left ACM and was succeeded .by Richard John (“John”). Askin and, subsequently, ACM, actively marketed interest in the Funds through written materials as well as in-person through presentations by Askin and ACM’s Director of Marketing, Geoffrey Bradshaw-Maek (“Mack”). Although the alleged misrepresentations constitute different aspects of one multifaceted fraud, for ease of discussion they can be separated into two categories, namely, the “valuations fraud” and the “operations fraud.” The valuations fraud pertains to representations concerning the process by which the Funds’ securities were valued, and, specifically, whether valuations were based on broker marks, as well as representations regarding the performance of the Funds’ securities. The operations fraud pertains to representations regarding the use of computer modeling to manage the Funds’ investments. ACM’s marketing materials represented that the Granite Funds would achieve annual returns of 15% or more “by investing in a market-neutral, risk-balanced portfolio of high return, high credit quality CMO derivative securities.” These materials further represented that, while higher returns are “typically” associated with higher risk, the Granite Funds were different because, due to their “risk-balanced strategy, higher returns can be achieved with the same, or lower, levels of risk.” Thus, “stable rates of return with low risk” were promised. Market neutrality was to be achieved by acquiring balanced holdings of “bullish” and “bearish” bonds. Thus, by purchasing offsetting positions in predictable securities, the Funds would enjoy the high returns associated with rate-sensitive CMOs while hedging against the risk attendant upon interest rate fluctuations. ACM represented through a graph that the Granite Funds would earn over 10% per year even if interest rates shifted over a 600 basis point range. ACM further represented that, in order to manage the Granite Funds’ portfolios, ACM had “developed its own proprietary analytics system that is used daily.” ACM described an elaborate “structured five-step process” whereby, for “each CMO bond under consideration”: Granite subjects the bonds to extensive financial analysis. Granite imposes a variety of economic scenarios on each security to identify how it would perform should there be changes in interest rates or prepayments. The results of this analysis are used to generate cash flow models for each CMO. Granite assigns probabilities to the possible outcomes for the CMOs under a variety of possible interest rate, prepayment, volatility and spread scenarios. With this information, Granite ascribes total return profiles to each CMO bond. ACM further described how it made use of its “analytic models” to determine how each bond would combine with others to “form a hedged, lower-risk portfolio,” and to “continually” monitor the portfolio and maintain its balance. ACM criticized other money managers for their lack of internal “sophisticated analytics capabilities” and their need to rely on third parties for analysis. Each Investor executed a subscription agreement, or Private Placement Memorandum (“PPM”), with respect to his investment. The PPMs made similar, but less specific, representations as to the Granite Funds’ investment strategy and analytic tools. The PPMs described the “investment objective” of the Granite Funds as “earnfing] a consistently high rate of return ... that is relatively stable over time ... by using market-neutral mortgage investing.” The strategy was designed to produce these stable results “whether interest rates are rising, falling or remaining essentially unchanged.” The investment advisor would use “carefully constructed and researched” “computer models to project the investment performance of different Mortgage-Backed Securities under different interest-rate scenarios and [to] search for securities with appropriately offsetting return profiles,” and to “actively” manage the portfolios. Askin sent investors monthly letters discussing the Funds’ performance (the “Performance Letters”). In January 1992, As-kin reported that he had corrected “a ‘bearish’ tilt” in the portfolios established by his predecessor, New Amsterdam. According to Askin, this “effective implementation of [the Funds’] portfolio strategy”— or “=balancing of the hedges” — resulted in “outstanding performance.” The following month, Askin wrote that analysis from his “proprietary pricing models” led to another adjustment in hedging positions and the maintenance of the “market-neutral portfolio.” In March of that year Askin announced that the “risk-balanced strategy produced an attractive total rate of return.” Similar representations were made during the ensuing two years. Askin pointed out on more than one occasion that there was much volatility in the interest rate markets, such that recognized (and bullish) fixed income indices suffered occasional monthly losses. According to Askin, his “analytics technology” enabled the Funds to maintain their market-neutral posture and, thus, achieve stable, positive returns virtually every month. Mack and Askin testified that on numerous occasions they represented to the Investors through in-person presentations that ACM valued and would continue to value the Funds’ portfolios and calculated returns based on “marks” provided by the brokers. Many Investors have testified to receiving such representations. As mentioned previously, CMOs are not traded on a public exchange. Therefore, when an institution that owns CMOs wishes to determine the value of its CMO position, it marks that position to market. A mark is an estimate of the price at which a security will sell. The Funds’ auditors, Price Waterhouse, issued annual audited financial statements in which it was stated that “100% of the [Funds’] investments were valued on the basis of a price quotation provided by principal market makers.” The market makers for CMOs are broker-dealers. The Price Waterhouse statements also listed the use of such price quotations for valuing the Funds’ securities as the first of the Funds’ “significant accounting policies.” Finally, the Price Waterhouse statements provided that “other assets ... for which market quotations are not readily available are valued at their fair value as determined in good faith.” Askin ratified the 1992 audited financials as “accurate and complete.” Potential investors routinely received from ACM the audited financial statements for the year prior to the contemplated investment, and then received subsequent audited statements. The Performance Letters also referenced the use of broker marks. Askin’s very first letter noted that poor performance in recent months resulted from “unfavorable dealer marks.” Another letter referenced Askin’s belief that there was “a disparity between the economic worth of many of our security holdings and the marks placed on them by some of the dealers.” Similar representations were made in other letters. The PPMs stated that securities traded “over-the-counter” would be valued based on the “closing bid” price, while other unidentified securities would be valued based on the “estimated fair value ... as determined in good faith by the General Partner,” in the case of Granite Partners, or “the fair market value ... as determined by the Directors in consultation with the Board of Advisors,” in the case of Granite Corp. The General Partner for Granite Partners was Dashtar Corporation, a company wholly owned by Askin. Askin was one of two Directors of Granite Corp. Granite Partners also had a Limited Partnership Agreement (“LPA”) which provided for “good faith” valuations. The PPMs made certain risk disclosures. The first page warned that “[t]he shares offered herein involve a high degree of risk. No one should invest who cannot afford to lose his entire investment.” Like all prospectuses, the PPMs included a “Risk Factors” section, which section repeated that there was a risk of complete loss of the investment, stated that “any investment in securities” entails “substantial risks,” and noted as risk factors prevailing interest rates, the investment advisor’s.ability to predict changes in those rates and in the real estate market, and the use of leverage by the advisor. Part of the investment strategy was described as using a “high rate” of leverage. The PPMs also cautioned, “there can be no assurance that risks will actually be reduced to the extent predicted by the [computer] models.” Finally, the PPMs disclosed that “the effect of fluctuations in interest rates on the value of Mortgage-Backed Securities, particularly including Residuals, is often complicated due to the prepayment characteristics of these instruments.” The PPMs disclosed that the Granite Funds were designed to exploit mispricing opportunities growing out of the difficulties undergone by thrifts that had purchased CMOs in the 1980’s. When Askin and Contino arrived in 1991, the -Funds were using a computer program called Wall Street Analytics. This program has some “option adjusted spread” (OAS) capability. OAS was developed as a technique for capturing the complexities of mortgage-backed securities better than had previous analytical tools, in particular, the dynamic factors of interest-rate changes and prepayment speeds. The extent to which OAS analysis was used at that time by institutions and individuals analyzing such securities is disputed. Askin and Contino wanted some of the features of the program to be proprietary to them, and worked with Andrew Chasen (“Chasen”), a third-party consultant to develop a product. Ultimately, Chasen licensed his “Amalgamated Bivariate” program (the “Amalgamator”) to ACM. Whether or not this product was proprietary to ACM is disputed. Askin has testified that the output from the Amalgamator was used to evaluate the sensitivity of individual securities and whole portfolios to changes in market conditions, to measure investment worth, and to assess the “tilt” of a portfolio and compliance with market neutrality. John, however, found the Chasen product too cumbersome to be usable and ineffective to monitor market neutrality. Chasen believed his product provided only “basic an-alytics,” not a CMO “model.” John Richardson (“Richardson”), an expert for the Investors, opined that the Amalgamator was incapable of running analyses which Richardson maintains are needed to run a market-neutral portfolio, including OAS, effective duration, and effective convexity, for each bond and the entire portfolio. Lawrence Wiener (“Wiener”), however, also an expert for the Investors, concluded that the Amalgamator was capable of computing present values of securities in different interest scenarios, and of constructing different prepayment scenarios — a description consonant with the PPMs’ claims. Late in 1993, ACM updated Chasen’s system with the Derivative Solutions system. Weiner opined that, in combination, the Chasen and Derivatives Solution systems are a powerful analytic tool that can perform sophisticated analysis on both an individual security and portfolio-level basis. However, many of the Funds’ bonds were never input into the Derivatives Solution system. Askin communicated with the dealers on a monthly basis regarding valuation of the CMOs owned by Funds. Typically, this process involved ACM sending a “marks sheet” to each dealer with a list of the CMOs which ACM wanted that dealer to mark. Generally, the securities listed were CMOs that the Granite Funds had either purchased from that dealer, or which that dealer was then financing under a repurchase agreement. The valuation or marking of CMOs is a complex process which was conducted by the brokers’ traders. Marking requires the exercise of informed judgment. The brokers marked the Funds’ CMOs to market on a monthly basis, devoting considerable time and effort to this process. Askin believed that dealer marks should be scrutinized and challenged. This belief was reflected in minutes of some of the meetings of the Granite Funds’ Investment Committee. The Performance Letters also reflected that Askin was seeking to get the brokers to establish “more consistent” marks or marks that better reflected what Askin believed to be the true value of the Funds’ portfolios. In meetings with investors in 1992 and 1993, Askin contrasted his approach with that taken by his predecessor, Estep, which he characterized as unquestioning with respect to the brokers’ valuations. Kidder began trading CMOs with Granite Partners and Granite Corp. in 1990. Kidder was one of at least 16 dealers trading CMOs with the Funds. Over the course of its dealings with Askin, Kidder provided some 600 marks for the Granite Funds. Beginning in May 1992, Askin placed monthly calls to challenge Kidder’s marks and request revised ones from William O’Connor (“O’Connor”), a Kidder salesperson. The first time Kidder agreed to revise a mark for a bond held by Granite Corp. was in May 1992, and the first time it agreed to revise a mark for Granite Partners was in March 1993. Kidder did not provide revised marks to Quartz. In total, Kidder provided ACM with 86 revised marks. Recorded conversations between O’Con-nor and Askin reflect O’Connor’s readiness to provide Askin with revised marks. In one, O’Connor stated to Askin, “[a]lright sir, I will adjust these.” In another, O’Connor stated, “I will remark these and send them over in five minutes.” In a third, O’Connor assured Askin, “[mjonth-end marks ... are completely negotiable ... I will work with you, however you want, on month end marks.” On March 3, 1994, O’Connor expressed concern regarding the revisions sought by Askin for reporting February 1994 performance, stating, “If you want it there I’ll mark it there but I am going to go on record and say I can sell you that bond cheaper.... [Although] as long as we are not, you know, as I said hemispheres apart, it’s not going to be an issue.” Ultimately, Kidder did not provide revised marks in March 1994. This was the month the brokers made the margin calls and conducted the liquidations that wiped out the Funds’ accounts. DLJ began trading CMOs with the Funds in late 1991, Beginning in January 1992, and continuing through March 1994, Asian called DLJ salesperson Betsy Com-erford (“Comerford”) on a monthly basis and requested revised marks. After each call, DLJ supplied ACM with a revised month-end mark sheet reflecting the prices specified by Askin. DLJ never indicated in any way that the revised month-end mark sheets contained something other than DLJ’s real month-end marks. DLJ revised approximately 400 marks, or over half of all marks provided to ACM. Kidder did not provide ACM with revised marks in March 1994. DLJ did. Ultimately, Askin used many of his own, “manager marks” to report performance for February 1994. O’Connor, arguing with colleagues on March 25, 1994, whether Kidder should “pull the plug” on ACM by making margin calls, warned that they should not do that because “we are in bed with ACM.” The Brokers did not see ACM’s marketing materials and were not present at ACM’s presentations to Investors. However, both Brokers reviewed the Price Wa-terhouse statements reporting a policy of using “100%” broker marks, and both Brokers annually confirmed their marks to the Funds’ auditors. O’Connor acknowledged to Askin his understanding that marks for “performance purposes and repo purposes ... are two different things,” and discussed with Askin their mutual understanding that when As-kin sought revised marks he would not take those marks as “indications of bids or offers [for] real trading.” O’Connor complimented Askin for not “believ[ing] everything on the [revised mark] sheet.” O’Connor also told Eric Kieter (“Kieter”), a CMO trader at a buy-side firm, that Askin didn’t object when Kidder “mark[ed] things to the bone for repo” because “then we’d have performance marks.” In a conversation on March 14, 1994, O’Connor explained to Michael Vranos (“Vranos”), Managing Director and head CMO trader for Kidder, “[t]his is where he [Askin] wants the marks to be. This is where you marked them for month end. This is not for repo purposes. This is performance marks.” Vranos replied, “I don’t want to be defrauding his investors.” On March 21, 1994, O’Connor, Vranos, and David Barrett (“Barrett”), also of Kidder, discussed the issue again. O’Connor commented, [T]he beautiful thing about Askin [is that] he doesn’t sit there and make us use the performance marks as his repo marks. From a credit perspective we’re covered. To which Vranos replied: Right. Just from a liability standpoint we’re not because we are defrauding investors. But, other than that, it’s no big deal. Remember, Dave, you’re an officer so your ass is going to be on the line. Kidder’s largest revision occurred when it agreed to “schmear” a downward correction in valuation for a particular CMO, the “Pru-Home” bond, over November and December 1993, where the value of that bond had been dramatically, and erroneously, overstated in October 1993. If Kidder had reported the Pru-Home bond’s actual value in November 1993, the Funds’ would have reported a loss in net asset value (“NAV”) for that month rather than, as was reported, a gain. O’Connor cannot recall any customers other than Askin who regularly asked for marks to be changed. In Comerford’s view, DLJ’s initial marks were “correct,” “market” prices, whereas the revised prices sought by As-kin were not. Other DLJ personnel shared that view. Comerford did not believe Askin could price CMOs better than DLJ, and never saw any evidence supporting Askin’s prices. Nonetheless, each month, DLJ provided ACM with another version of the most recent month-end mark sheet, reflecting the precise prices that Askin had requested. Most of the time Askin sought upward revisions, but at times he also sought — and obtained — downward ones. O’Connor reported that Askin told him “I still got a lot of unrealized gains — why do you think I call you back every month and make you write down some prices.... I got a shit load of rainy day money if I need it.” Comerford testified that she provided revised marks to Askin so that calculations could be made of the yield of the Funds’ bonds “as if’ they had the value proposed by Askin. The revised mark sheets, however, provided Askin’s prices without yields. Comerford’s assistant believed the month-end sheets for Askin contained “pricing,” not yield calculations. When DLJ delivered the revised mark sheet to ACM it sometimes included a cover note stating that “month end prices” were enclosed or placed the words “with price changes” or “original” on the document. Comerford has also testified that “I know now I shouldn’t have done it this way,” and “I screwed up.” Comerford testified that DLJ provided month-end marks “for every single customer,” knowing that “all marks are used to make some kind of reporting on the progress of the fund.” In addition, ACM entered into a contract with DLJ in which DLJ agreed to provide “accurate” marks each month. The contract stated that the marks were “critical” to ACM’s “ability to report [the Funds’] performance [to investors] as soon after the end of each month as possible,” that “poor portfolio performance” would “cost [ACM’s] customers,” and warned that DLJ’s failure to honor the contract would cause ACM “to reduce significantly the amount of business” it gave to DLJ. ACM personnel knew that Askin challenged broker marks, but were aware of neither the scope of the marks revisions obtained by Askin nor the readiness with which the brokers supplied those revisions. Ronald Augustin (“Augustin”), of ACM, knew that at times Askin obtained revised marks but perceived these as corrections of substantive “mistakes.” Mack was “flabbergasted” by the “massive changes” involved, after hearing a taped conversation between Askin and O’Connor. John considered the process revealed by tapes of Askin’s conversations with O’Connor to be inconsistent with his understanding of the process. John felt “morally compromised” in March 1994 when he learned of Askin’s use of “manager marks” to report performance for February 1994, because John had believed the Funds used “broker-dealers’ marks ... for our performance”. Indeed, John and other senior ACM staff threatened to resign in March 1994 when they learned of Askin’s intention to use his own prices. Beginning with Askin’s arrival, the reported returns for both Granite Partners and Granite Corp. improved significantly. An expert for the Investors, Jed Kaplan (“Kaplan”), concluded that the Brokers’ revised marks did not represent fair market value. Kaplan further concluded that, if Askin had used DLJ’s and Kidder’s initial marks rather than the revised ones, between January 1992 and February 1994, there would have been fifteen months with negative changes in NAV for Granite Corp. and sixteen months with negative changes in NAV for Granite Partners. ACM reported no losing months for this period. An expert for the Brokers, Lee Errickson (“Errickson”), quarrels with these numbers, although Errickson’s report also concludes that there would have been losing months reported. Kidder’s revisions, standing alone, would have turned a losing month into a winning month on only one occasion, November 1993, when it agreed to “schmear” the Pru-Home bond correction between two months. In addition, of themselves, Kidder’s revisions would have had little or no impact on the reported volatility, duration, “Sharpe Ratio” (a performance/risk measure), leverage, or targeted annual returns. The reported monthly returns were between positive NAV 1.1% and 2.8% for Granite Corp., and between positive 1% and 3% for Granite Partners. Price Waterhouse, in conducting its audits, took certain steps to determine whether the marks provided by the Brokers were corroborated by actual sales, and concluded that the valuations of the CMOs at year-end were “reasonable.” Andrew Carrón (“Carrón”), an expert for the Brokers, concluded that Kidder’s revised marks were “valid indicators of market value.” The Bankruptcy Trustee concluded that Kidder’s initial marks and revised marks were “[Generally ... consistent,” with a difference of 3% or less in most cases. Comerford had read the statement of Granite’s investment objective, contained in its marketing materials, as “[a] stable high absolute level of return of 15 percent per year across a broad range of interest rate scenarios.” Donald Peskin (“Pes-kin”), also of DLJ, understood that Askin “was [running a market-neutral fund]; he had marketed his fund as a market-neutral fund.” Peskin could not be sure, but assumed this understanding came from Com-erford. Deborah Reynolds (“Reynolds”), a DLJ trader, knew the Funds were “supposed to be ... zero duration.” DLJ performed six analyses of the Funds’ portfolios between January 1992 and September 1993, each of which showed the Funds to be bullish. There is a dispute as to whether DLJ was provided with complete information regarding the portfolios’ composition before doing these anal-yses. DLJ executives were aware the Funds’ persistent lack of neutrality. John Friel (“Friel”), head of DLJ’s finance desk, concluded that the portfolios were “sensitive” to interest rate increases. Leon Pollack (“Pollack”), head of DLJ’s fixed income department, testified that “anyone who had [the Funds’] position^] would be in trouble [when] rates were going up.” Reynolds testified that DLJ’s view was that the Funds “had a fair amount of duration.” When Askin told O’Connor about the new Quartz Fund, he told O’Connor, “[t]he main difference in Quartz relative to the Granite is it’s not constrained to being market neutral.” Exchanges among Kidder personnel and between Kidder and Askin personnel indicate Kidder’s awareness that the Granite Funds were persistently bullish, and that this was contrary to the way they were supposed to be structured. O’Connor stated to other Kidder personnel) “[t]his guy’s made 85% or more for his investors in the last two straight and he didn’t do that by being ... duration neutral.... [M]y only danger is ... my biggest guy blowing himself up.” Vra-nos laughed when he said the Granite Funds are “structured as [] zero duration.” O’Connor stated to John, of ACM, “You and I both know that the portfolios are not market neutral.” ACM was one of Kidder’s biggest customers, and was DLJ’s single largest mortgage-backed securities customer. The Brokers recommended and sold vast quantities of “inverse 10” to ACM. In the third and fourth quarters of 1993, for example, Kidder sold over $120 million in inverse 10 and over $300 million in other bullish securities to ACM. As O’Connor put it, “I try to shove'inverse IOs down [Askin’s] throat.” Wiener concluded that all of these securities were bullish. During the samé period, DLJ sold approximately $115 million in inverse IOs, approximately the same amount in other securities which Wiener concluded were bullish, and one $3.5 million straight 10 that was bearish. The vast majority of all the securities sold by the Brokers to ACM was bullish. The Brokers represented the inverse IOs as bearish at a time when they were bullish. Jeffrey Lewis (“Lewis”), DLJ’s head derivatives trader, characterized inverse IOs are “bullish when you want [them] to be bearish and bearish when you want [them] to be bullish.” Comerford testified that inverse IOs are not a substitute for straight IOs, and that it required “a lot of expertise” to understand what ACM purchased. Vranos testified that As-kin was buying “high risk” securities from Kidder. Vranos did not think inverse IOs were bearish in February 1994, during which time Kidder was selling inverse IOs to ACM and representing them as bearish. O’Connor feared the Funds would “blow up” due to their bullish tilt. ACM made it possible for the Brokers to sell their entire CMO offerings by purchasing “deal-driver” tranches of exotic securities created by the brokers themselves. In each of these deals, the Brokers sold hundreds of millions of dollars worth of CMOs. The Brokers perceived Askin as one of the few buyers in the market for these deal-driving tranches. Vranos testified that Askin was a buyer of deal-drivers, and O’Connor stated, “You can count the number of inverse 10 buyers on three fingers.” Richard Whiting (“Whiting”), of DLJ, testified that Askin was a “very significant” DLJ client due to his “unique” requests and willingness to buy deal-driving “tranches ... new issue CMO securities.” The commissions received by O’Connor and Comerford were related in part to the riskiness of the security, with riskier securities generating higher commissions. O’Connor described how he received increased commissions on the “nuclear waste” sold to Askin. Comerford also testified to the relationship between her commission rate and the nature of the securities. The Brokers provided very favorable financing for ACM’s CMO purchases. DLJ routinely loaned ACM 95% of the purchase price for CMOs in their reverse repo transactions, creating a 19-to-l debt/equity ratio. Kidder provided ACM with a special type of credit facility. O’Connor described this arrangement to Askin, stating, “[W]e have actually set up a different credit facility for handling you where Vra-nos signs a sheet that says ... I absorb any losses here brought about by ... an Askin account blowing up and us not maintaining a haircut.” Vranos testified to the “special account facility pool” set up to do “asset-based lending” and that ACM was a “large fund that fell under the special account facility ... policy.” Although the Funds’ CMO holdings were complex and volatile, DLJ was able to sell many of the Funds’ securities within a two-day period following the DLJ liquidation. Moreover, experts for the Funds, in the Funds Action, have opined that the prices obtained by DLJ, as well as by Merrill in its liquidation auction, were lower than what could have been obtained. The brokers were aware of the claims made regarding computer modeling in the PPMs, but did not see the marketing materials and were not present at ACM’s in-person presentations. O’Connor stated to Vranos, “Askin has no model,” and joked that the “model” consisted of “wett[ing] his finger and put[ting] it in the air.” On another occasion, ACM’s John joked to O’Connor, “I am matching up wits with Vranos on my HP.” DLJ’s Comerford believed that DLJ had an obligation to “know [its] customer.” DLJ knew that several customers had advanced analytical capabilities. DLJ knew that ACM was buying complex securities that could be understood only with sophisticated modeling. DLJ personnel did not see evidence that ACM had such modeling. However, DLJ personnel, unlike Kidder, did not actually express a view that ACM had inadequate modeling capabilities, or “no model” at all. The Investors include wealthy individuals, money management firms, hedge funds or “funds of funds” (established to institute in other hedge funds), pension plans, and insurance companies. Unless otherwise permitted by the Granite Funds, Investors in those funds were required to invest a minimum of $1,000,000. The PPMs which each Investor signed, stated, “[t]he undersigned has the necessary knowledge and experience in financial and business matters to enable him to evaluate the merits and risks of this investment.” The PPMs also included a disclaimer that “no representations or warranties have been made to [the Investor],” and “in entering into this transaction [the Investor is] not relying upon any information other than that contained in the Offering Memorandum [i.e., the PPM] and the results of [the Investor’s] own independent investigation.” The PPM also confirmed that there was an opportunity for the Investor to ask questions of ACM, and receive responses, as part of her investigation. Potential and actual investors were given the opportunity to obtain information about the Funds and how they were run. The Investors were free to examine ACM’s offices and its computer models, or to interview its personnel, including Askin. They were invited to observe the computer modeling on ACM’s screens and were provided with printouts generated by the Am-algamator. They were provided with the Performance Letters and had access to lists of the Funds’ holdings. The Price Waterhouse statements also set forth each security owned by the Funds, broken down by type. These statements were given to prospective as well as current investors. Mack testified that an investor request “to see something or speak to someone” was never denied. ACM also provided responses to due diligence questions by those Investors who inquired. One such Investor, Commonwealth/Providian, concluded that ACM had “sophisticated and comprehensive computer capabilities.” The Price Waterhouse statements broke the Funds’ securities into broad categories, such as “interest only tranches,” “principal only tranches,” “other tranches,” and “residual interests.” These categories corresponded with categories used in ACM materials, in which ACM represented that interest-only CMOs and residuals were bearish, thus balancing (or hedging) the bullish principal-only tranches. In the annual audited statements, the reported market value of the bullish categories roughly approximated the reported market value of the bearish ones. Some of the tranches reported as “interest only” in the audited financials were in fact inverse IOs, which are not bearish. The Price Waterhouse statements did list the individual securities themselves, so that an Investor could have obtained a prospectus for that security. In addition, some Investors requested, and received, “Current Holdings Reports” which identified securities in more detail, e.g., as “Interest Only Inverse Floater” or “Super P/O.” DLJ obtained statements containing this level of detail in order to perform its portfolio analyses for ACM. Investors testified that they read the Performance Letters and .Price Water-house statements as reflecting that ACM’s market-neutral, low-volatility approach was working. Each of the Investors has stated in a sworn declaration that he would not have invested, or retained his investment, if he had known of Askin’s practice of obtaining revised marks from the brokers. DLJ’s Comerford testified that the Funds’ reported returns in 1992 would have led her to conclude that the portfolios were neutral. However, an expert for Kidder, David Ross (“Ross”), opined that one cannot draw conclusions about duration or market sensitivity from the reported returns. Ross compared the Funds’ reported returns to the performance of benchmarks and concluded that the Funds’ returns were more volatile than Treasuries and were correlated with interest rate movements. Matthew Richardson (“Richardson”), an expert for the Investors, performed a similar comparison and concluded that the Funds did not achieve market neutrality. Richardson’s use of regression analysis has been criticized by a DLJ expert, Raj Mehra (“Mehra”), as inadequate for making the measurements required to reach his conclusion. Wiener analyzed the securities listed in the statements provided to DLJ for purposes of the DLJ portfolio analyses. Wiener analyzed the effective duration and effective convexity of each bond within the portfolios and concluded that the Funds were never market-neutral. Some Investors made their investments before Askin’s arrival and, thus, did not have contact with Askin, ACM, or the ACM marketing materials before making their investments: Employee Retirement Income Plan of Minnesota Mining and Manufacturing Company (“3M”), The David I. Chemerow 1992 Trust (the “Chemerow Trust”), Robert Johnston (“Robert Johnston”), and Richard Johnston (“Richard Johnston”) as Trustee for the Demeter Trust (the “Demeter Trust”). These Investors contend that assurances made by Askin upon his arrival, or other ACM representations, induced them to maintain their investments. These Investors did receive the PPMs. Some Investors made their initial investment before Askin’s arrival, but made other investments during Askin’s tenure: Lionel Sterling (“Sterling”) and Antaeus Enterprises, Inc. (“Antaeus”). Investor Roma Malkani (“Malkani”) invested after Askin’s arrival but did not have contact with Askin or anyone else at ACM before investing. A number of Investors did not know and did not ask how the Funds’ portfolios were valued at the time they made their investments: L.H. Rich Companies (“L.H.Rich”), Primavera Familienstiftung (“Primavera”), International Asset Management Limited (“IAM”), ABF Capital Management (“ABF”), CoriFrance, Hedged Investment Partners (“HIP”), Global Hedge Fund (“Global”), Malkani, 3M, Diversified Income Strategies, Excelsior Investment Fund and Excelsior Qualified L.P. (collectively, “Excelsior”), Obling-ter, W Finance Arbitrage (“W France”), the Regency Fund, Montpellier Resources Limited, Robert Johnston, Sofa Partners, the Demeter Trust, Gilla (B.V.I.) Limited (“Gilla”), and Nemrod Leverage Holdings Limited (“Nemrod”). Although all the Investors claim to have reviewed and relied upon the Price Water-house statements, fewer than half of the sixty Investors were able to produce copies of those statements, and only thirteen testified to having seen the “100%” broker marks language, with only seven having seen the language before making their investments. Some Investors perceived, based on the PPMs or other information provided by Askin, that Askin had some discretion in reporting the value of the Funds’ securities: Glenwood Balancing Fund, L.P. Glenwood Trust Company as Custodian for Walker Art Center, Glenwood Trust Co. Group Trust II, Glenwood Partners, L.P., Samta, Inc. (“Samta”), Bambou, Inc. (“Bambou”), Loukoum, Inc. (“Loukoum”), FIDR Investors 1996 Trust, HNM First Investors 1996 Trust, HNM Second Investors 1996 Trust, SDI, Inc., Commonwealth Life Insurance Company, Providian Life & Health Insurance Company (“Commonwealth/Providian”), Neutral Strategies, L.P., Pine Equities, L.P., and the Chemer-ow Trust. Some Investors knew that Askin talked to the brokers about revising marks: Sterling, Commonwealth/Providian, Oakwood Associates, Rosewood Associates, and the Chemerow Trust. A number of Investors had little understanding of ACM’s computer modeling capabilities and never asked for a demonstration or documentation: Levitt Family Trust, Spirit Debt Limited, Spirit Neutral Limited, 3M, Neutral Strategies, L.P., Zimmerman Family Trust I, Zimmerman Income Partners, IAM, Bambou, Louk-oum, Samta, Hubert Looser, L.H. Rich, Primavera, ABF, CoriFrance, HIP, Conservation Securities Limited Partnership, Trans-Resources, Inc., Excelsior, Gilla, Nemrod, Robert Johnston, William Mona-ghan, Oblingter, and W Finance. One Investor, William Cook (“Cook”), took note of Askin’s comments that he asked the brokers to reconsider their marks. Cook tested the accuracy of the valuations by comparing the reported marks and subsequent sale prices of the bonds. Cook, an investment strategist for a large insurance company, was specifically familiar with CMOs. Cook believed that, while Askin sought revisions, he ultimately accepted the broker’s price if he was unable to convince them of their error. Facts Relevant to the Statute of Limitations On March 10, 1994, Askin reported a two percent loss to the Investors for all three Funds. On March 25, 1994, ACM restated the February 1994 loss, acknowledging that the correct figure was approximately twenty percent. On March 24, 1995, Primavera filed a putative class action complaint in the United States District Court for the Northern District of California. This suit asserted federal claims and was brought “on behalf of a class consisting of all persons and entities who purchased, directly or beneficially, securities issued by any of the Granite Funds ... from January 26, 1993 through the date the Granite Funds went bankrupt.” In April 1995, the Funds’ Chapter 11 trustee filed a preliminary report with the bankruptcy court regarding his investigation of the Funds’ demise. This report indicated the possible involvement of the Brokers in Askin’s fraudulent scheme. On September 20, 1995, the Primavera complaint was amended to add the Brokers as defendants, and on October 18, 1995, the Primavera Action was transferred to the Southern District of New York. On March 1998, class certification in the Primavera Action and the later-filed Montpellier Action was denied. Providian Life & Health Insurance Company (“Providian”) is an insurance company with its principal place of business in Pennsylvania. Providian is a plaintiff in ABF Action, filed on March 27, 1996. Sterling is a Connecticut resident. Sterling is a plaintiff in the Johnston Action, filed on June 9,1997. The Demeter Trust maintains its principal place of business in Connecticut. The Demeter Trust is a plaintiff in the Johnston Action, filed on June 9,1997. Cook is a Connecticut resident. Cook is a plaintiff in the AIG Action, filed on October 21,1998. Arbor Place, L.P. (“Arbor”), maintains its principal place of business in Massachusetts. Arbor is a plaintiff in the Montpel-lier Action, pursuant to the amended complaint filed on June 2, 1997. Global Hedge Fund (“Global”) is domiciled in Jersey, Channel Islands. Global is a plaintiff in the Montpellier Action, pursuant to the amended complaint filed on June 2,1997. Malkani is a Maryland resident. Malka-ni is a plaintiff in the Montpellier Action, pursuant to the amended complaint filed on June 2,1997. The Funds Action The Repurchase Transactions DLJ and each of the Funds executed a standard industry agreement referred to as the PSA (“Public Securities Association”) Agreement (the “PSA Agreement”) with respect to their repo transactions. Merrill and Quartz executed the PSA Agreement, but Merrill and Granite Partners and Granite Corp., respectively, did not. The PSA Agreement provides with respect to the Funds’ obligation as to margin maintenance, and the brokers’ right to demand additional collateral: If at any time the aggregate Market Value of all Purchased Securities subject to all Transactions in which a particular party hereto is acting as Buyer [the broker] is less than the aggregate Buyer’s Margin Amount for all such Transactions (a “Margin Deficit”), then Buyer may by notice to Seller [the Fund] require Seller in such Transactions, at Seller’s option, to transfer to Buyer cash or additional Securities reasonably acceptable to Buyer (“Additional Purchased Securities”) so that the cash and aggregate Market Value of the Purchased Securities, including any such Additional Purchased Securities, will thereupon equal or exceed such aggregate Buyer’s Margin Amount (decreased by the amount of any Margin Deficit as of such date arising from any Transactions in which such Buyer is acting as Seller. PSA Agreement ¶ 4(a). The “Buyer’s Margin Amount,” ie., the amount of collateral which the Fund was required to maintain in a repo account so as not to have a margin deficit, is “the amount obtained by application of a percentage ... agreed to by Buyer and Seller prior to entering into the transaction, to the Repurchase Price for such Transaction.” PSA Agreement ¶ 2(c). The PSA Agreement provides that the Fund may obtain the return of collateral that is in excess of the required margin amount: If at any time the aggregate Market Value of all Purchased Securities subject to all Transactions in which a particular party hereto is acting as Seller [the Fund] exceeds the aggregate Seller’s Margin Amount for all such Transactions at such time (a “Margin Excess”), then Seller may by notice to Buyer [the broker] require Buyer in such Transactions, at Buyer’s option, to transfer cash or Purchased Securities to Seller, so that the aggregate Market Value of the Purchased Securities, after deduction of any such cash or any Purchased Securities so transferred, will thereupon not exceed such aggregate Seller’s Margin Amount [ ]. PSA Agreement § 4(b). The PSA Agreement defines “market value, with respect to any Securities as of any date,” as, the price for such Securities on such date obtained from a generally recognized source agreed to by the parties or the most recent closing bid quotation from such a source, plus accrued income to the extent not included therein ... as of such date (unless contrary to market practice for such Securities). PSA Agreement ¶ (2)(h). If a margin call remains unsatisfied one business day after notice was given, the buyer, ie., the broker, has the right to liquidate the repo positions of the seller, ie. the Fund, in default. See PSA Agreement ¶ 11. In order to accomplish this liquidation the broker has two options, referred to hereinafter as “Option A” and “Option B”: (A) immediately sell, in a recognized market at such prices as the nondefault-ing party may reasonably deem satisfactory, any or all Purchased Securities subject to such Transactions and apply the proceeds thereof to the aggregate unpaid Repurchase Prices and any other amounts owing by the defaulting party hereunder [“Option A”] or (B) in its sole discretion elect, in lieu of selling all or a portion of such Purchased Securities, to give the defaulting party credit for such Purchased Securities in an amount equal to the price therefor on such date, obtained from a generally recognized source or the most recent closing bid quotation from such a source, against the aggregate unpaid Repurchase Prices and any other amounts owing by the defaulting party hereunder [“Option B”]. PSA Agreement ¶ ll(d)(i). In the case of the DLJ repo transactions, the margin maintenance requirement was set in relation to the haircut percentage. According to the report of Funds’ expert John Y. Campbell (“Campbell”), this requirement was equal to the (repo amount) * (100% + haircut percentage). For example, one of the securities held in Granite Corp.’s repo account with DLJ was FHLMC 1415 S. The repo or loan amount for this security was $1,951,000. The haircut was 10%. The margin amount required to avoid a margin deficit was $2,146,100, which is 110% of the repo amount. The haircut amounts for the securities held by DLJ ranged from 5% to 25%. This margin maintenance obligation was not the same as the repurchase obligation due on the buy-back date. As explained earlier, the repurchase obligation was equal to the repo amount plus the market rate of interest. The PSA Agreement provides expressly that the parties “intend that all Transactions hereunder be sales and purchases and not loans.” PSA Agreement ¶ 6. As mentioned earlier, Merrill entered into a PSA Agreement with Quartz but not with Granite Corp. or Granite Partners. Merrill and each of the Funds — Granite Corp., Granite Partners, and Quartz — entered into a “reverse repurchase confirmation” agreement (a “Repo Trade Confirmation”) for each repo transaction. The Repo Trade Confirmations all have identical terms. With respect to the level of collateral required to be maintained in the repurchase accounts, the confirmations provide: If on a business day the market value of the securities for a transaction is less than the agreed upon percentage of the outstanding purchase price, the purchaser may demand a mark to market.... Margin percentage shall at all times be equal to 102% of the repurchase principal plus accrued repurchase' interest to date unless otherwise agreed. Repo Trade Confirmation at 2. The confirmations also contain certain provisions concerning adequate assurance of performance: If any time prior to the repurchase date, reasonable ground for insecurity shall arise with respect to performance by a party hereto, the other party may demand from such party that adequate assurance of due performance by such party be provided. A party is in default if ... it fails to provide adequate assurance of due performance upon demand by the other party. If either party is in default, the other party may without notice ... sell the securities.... Repo Trade Confirmation at 2. Under the Repo Trade Confirmations, the Funds were to repurchase the repoed securities from Merrill by April 25, 1994. Unlike the PSA Agreement, the Repo Trade Confirmations do not provide expressly that the transactions are intended to be sales and purchases rather than loans. Michael Aneiro (“Aneiro”), Augustin, John, Contino (“Confino”), Stephen J. Dendinger (“Dendinger”), and Askin, all of whom were employees of either Merrill or the Funds, testified that they were not aware of and did not view there to be any differences between the Merrill/Granite transactions, in which no PSA Agreement was executed, and the Merrill/Quartz transactions, in which both a PSA Agreement and Repo Trade Confirmations were executed — or between the Merrill/Granite transactions and the transactions between the Funds and the other broker-dealers, in which only PSA Agreements were executed. Some of these same witnesses also testified that they understood the repo transactions as loan arrangements, with the securities serving as collateral for those loans. Merrill’s internal policy manual describes repos as “collateralized loan[s],” and the internal d