Federal Housing Finance Agency as Conservator for the Federal National Mortgage Association et al v. Goldman, Sachs & Co. et al

MEMORANDUM OF LAW in Support re: 729 MOTION for Partial Summary Judgment on Defendants' Due Diligence and Reasonable Care Defenses. . Document

S.D.N.Y.July 8, 2014

UNITED STATES DISTRICT COURT SOUTHERN DISTRICT OF NEW YORK FEDERAL HOUSING FINANCE AGENCY, AS CONSERVATOR FOR THE FEDERAL NATIONAL MORTGAGE ASSOCIATION AND THE FEDERAL HOME LOAN MORTGAGE CORPORATION, Plaintiff, -against- GOLDMAN, SACHS & CO., GS MORTGAGE SECURITIES CORP., GOLDMAN SACHS MORTGAGE COMPANY, THE GOLDMAN SACHS GROUP, INC., GOLDMAN SACHS REAL ESTATE FUNDING CORP., HOWARD S. ALTARESCU, KEVIN GASVODA, MICHELLE GILL, DAVID J. ROSENBLUM, JONATHAN S. SOBEL, DANIEL L. SPARKS, AND MARK WEISS, Defendants. 11 CIV. 6198 (DLC) PLAINTIFF’S MEMORANDUM OF LAW IN SUPPORT OF ITS MOTION FOR PARTIAL SUMMARY JUDGMENT ON DEFENDANTS’ DUE DILIGENCE AND REASONABLE CARE DEFENSES QUINN EMANUEL URQUHART & SULLIVAN, LLP Philippe Z. Selendy Manisha M. Sheth Sascha N. Rand Andrew R. Dunlap David B. Schwartz 51 Madison Avenue, 22nd Floor New York, New York 10010 (212) 849-7000 Attorneys for Plaintiff Federal Housing Finance Agency, as Conservator for Fannie Mae and Freddie Mac Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 1 of 76 i TABLE OF CONTENTS Page PRELIMINARY STATEMENT .....................................................................................................1 FACTUAL BACKGROUND ..........................................................................................................4 I. GSMSC CONDUCTED NO DILIGENCE ON THE WLTD SECURITIZATIONS ..........................................................................................................5 II. GOLDMAN’S DILIGENCE FOR THE WLTD SECURITIZATIONS WAS INADEQUATE....................................................................................................................6 A. Goldman Conducted No Diligence At The Time Of Securitization ........................8 1. Goldman Conducted No Diligence When Securitizing WLTD Loans, Only When Buying Them ................................................................8 2. Goldman Did Not Verify the Representations in the WLTD Securitizations’ Prospectus Supplements ..................................................10 B. Goldman’s Acquisition-Stage Diligence Was Unreasonable And Insufficient .............................................................................................................11 1. Goldman Conducted No Diligence On The Vast Majority Of Loans Included In The SLGs .....................................................................12 (a) Goldman Believed Its Internal Diligence Database Was Unreliable, Yet It Relied On That Database in Conducting Diligence ........................................................................................13 (b) Goldman Selected Samples Of Loans To Be Diligenced From Pools Of Loans It Acquired Rather Than From The Securitizations Or SLGs ................................................................14 (c) Goldman Acquiesced To The Demands Of Originators To Reduce The Number Of Loans Diligenced In The Acquisition Pools ...........................................................................16 (d) Goldman Did Not Apply Consistent Standards For Selecting Its Samples, And Did Not Include All Loans That Met Its “Adverse” Credit Criteria ..................................................19 (e) Goldman Did Not Extrapolate the Results Of Its Diligence Reviews To The Unsampled Portions Of The Acquisition Pools ...............................................................................................21 Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 2 of 76 ii 2. Goldman Failed To Follow Up On Red Flags By Conducting Additional Diligence ..................................................................................23 (a) Goldman Did Not Increase the Size of its Samples to Conduct Additional Credit Diligence ............................................23 (b) Goldman Failed To Properly Scrutinize Its Third-Party Diligence Vendors .........................................................................24 (c) Goldman Knew It Should Not Securitize Loans From Certain Originators, Yet It Did So Anyway Without Any Further Diligence ...........................................................................30 3. Goldman Failed To Use Its Own Conduit Guidelines To Diligence Loans Acquired Through Its Conduit Program Despite Representations In The Prospectus Supplements That The Loans Conformed To Those Guidelines ...............................................................31 C. Goldman Used The Results Of Its Diligence To Increase Its Own Profits While Knowingly Communicating False Information ..........................................34 1. Goldman Determined That The Actual Values Of Many Of The Mortgaged Properties Were Lower Than Reported By Originators, But Used The Actual Values Only To Negotiate Lower Prices For Itself ...........................................................................................................34 2. Goldman’s Own Originators Informed It Of The Sub-Prime Crisis In December 2006, But Goldman's Only Response Was To Begin Shorting The Market ..................................................................................36 III. GOLDMAN’S DILIGENCE FOR THIRD-PARTY SECURITIZATIONS WAS INSUFFICIENT .................................................................................................................38 ARGUMENT .................................................................................................................................40 I. GSMSC HAS NO DUE DILIGENCE DEFENSE AS A MATTER OF LAW ................42 II. GOLDMAN’S DILIGENCE WAS INADEQUATE AS A MATTER OF LAW .............42 A. As The Underwriter, Goldman Faces A Near-Absolute Standard Of Liability For Its Diligence Defense Because It Was Completely Intertwined With the Issuer, GSMSC ....................................................................43 B. Goldman Did Not Conduct A Reasonable Investigation At The Time The Prospectus Supplements Became Effective ...........................................................46 C. Goldman Did Not Reasonably Investigate Whether the Statements in the Prospectus Supplements Were True ......................................................................48 Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 3 of 76 iii D. Goldman’s Acquisition-Stage Diligence Process Was So Inadequate That Goldman Was Incapable of Performing A Reasonable Investigation ...................50 E. Goldman Did Not Look Deeper Or Question More When Confronted With Numerous Red Flags At The Acquisition Stage ....................................................52 F. Goldman Succumbed To Competitive Pressure In Securitizing Conduit Loans While Being Unable To Determine Whether Such Loans Had Been Underwritten To Goldman’s Guidelines ................................................................55 G. Goldman Knew The Information In The Prospectus Supplements Was False And Deliberately Concealed The Truth .......................................................56 H. Goldman’s Diligence on the Third-Party Securitizations Was Not Thorough Or Searching And Was Driven by the Issuers ......................................57 III. GOLDMAN AND GSMSC DID NOT EXERCISE REASONABLE CARE AS A MATTER OF LAW ...........................................................................................................59 CONCLUSION ..............................................................................................................................63 Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 4 of 76 iv TABLE OF AUTHORITIES Page Cases Alton Box Bd. Co. v. Goldman, Sachs & Co., 560 F.2d 916 (10th Cir. 1977) .................................................................................................62 Anderson v. Liberty Lobby, Inc., 477 U.S. 242 (1986) .................................................................................................................41 Celotex Corp. v. Catrett, 477 U.S. 317 (1986) .................................................................................................................40 Chris-Craft Indus., Inc. v. Piper Aircraft Corp., 480 F.2d 341 (2d Cir. 1973).................................................................................................3, 42 Demarco v. Edens, 390 F.2d 836 (2d Cir. 1968).....................................................................................................63 Dukes v. Wal-Mart Stores, Inc., 603 F.3d 571, 637 (9th Cir. 2010) (en banc) ...........................................................................52 Eng’g Contractors Ass’n of S. Fla., Inc. v. Metro. Dade Cnty., 943 F. Supp. 1546 (S.D. Fla. 1996) .........................................................................................52 In re Enron Corp. Sec., Derivative & ERISA Litig., 235 F. Supp. 2d 549 (S.D. Tex. 2002) ....................................................................................52 Ernst & Ernst v. Hochfelder, 425 U.S. 185 (1976) ...........................................................................................................42, 43 Escott v. BarChris Constr. Corp., 283 F. Supp. 643 (S.D.N.Y. 1968) ........................................................................43, 46, 49, 58 FDIC v. Giammettei, 34 F.3d 51(2d Cir. 1994)..........................................................................................................41 Feit v. Leasco Data Processing Equip. Corp., 332 F. Supp. 544 (E.D.N.Y. 1971) .................................................................................. passim FHFA v. Bank of America Corp., 2012 WL 6592251 (S.D.N.Y. Dec. 18, 2012) .........................................................................59 FHFA v. HSBC N. Am. Holdings Inc., --- F. Supp. 2d ----, 2013 WL 6480445 (S.D.N.Y. Dec. 10, 2013) ..........................................47 FHFA v. HSBC N. Am. Holdings, Inc., --- F. Supp. 2d ----, 2013 WL 6588249 (S.D.N.Y. Dec. 16, 2013) ..........................................59 FHFA v. JPMorgan Chase & Co., 2012 WL 6000885 (S.D.N.Y. Dec. 3, 2012) ...........................................................................51 Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 5 of 76 v FHFA v. UBS Americas, Inc., 858 F. Supp. 2d 306 (S.D.N.Y. 2012)................................................................................42, 57 FHFA v. UBS Americas, Inc., 2012 WL 2400263 (S.D.N.Y. June 26, 2012) .........................................................................47 Franklin Sav. Bank of N.Y. v. Levy, 551 F.2d 521 (2d Cir. 1977)...............................................................................................59, 61 Glassman v. Computervision Corp., 90 F.3d 617 (1st Cir. 1996) ................................................................................................44, 46 Golden Pac. Bancorp v. FDIC, 375 F.3d 196 (2d Cir. 2004).....................................................................................................41 Heffernan v. Pac. Dunlop GNB Corp., 965 F.2d 369 (7th Cir. 1992) ...................................................................................................60 Herman & MacLean v. Huddleston, 459 U.S. 375 (1983) ...........................................................................................................41, 42 Junker v. Crory, 650 F.2d 1349 (5th Cir. 1981) .................................................................................................60 LC Capital Partners, LP v. Frontier Ins. Grp., Inc., 318 F.3d 148 (2d Cir. 2003) ....................................................................................................53 In re Livent, Inc. Noteholder Sec. Litig., 355 F. Supp. 2d 722 (S.D.N.Y. 2005)..............................................................41, 43, 48, 54, 58 In re Metlife Demutualization Litig., 262 F.R.D. 217 (E.D.N.Y. 2009) .............................................................................................60 In re Morgan Stanley Info. Fund Sec. Litig., 592 F.3d 347 (2d Cir. 2010).....................................................................................................42 MidAmerica Fed. Sav. & Loan Ass’n v. Shearson/Am. Express, Inc., 886 F.2d 1249 (10th Cir. 1989) ...............................................................................................59 Nat’l Credit Union Admin. Bd. v. Wachovia Capital Markets, LLC, 2014 WL 1795294 (S.D.N.Y. May 6, 2014) ...........................................................................51 In re Software Toolworks Inc., 50 F.3d 615 (9th Cir. 1995) .................................................................46, 56, 57, 59, 60, 61, 62 Taniguchi v. Kan Pacific Saipan, Ltd., 132 S. Ct. 1997 (2012) .............................................................................................................50 Univ. Hill Found. v. Goldman, Sachs & Co., 422 F. Supp. 879 (S.D.N.Y. 1976) ............................................................44, 53, 54, 55, 59, 60 Wigand v. Flo-Tek, Inc., 609 F.2d 1028 (2d Cir. 1979)...................................................................................................62 Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 6 of 76 vi In re WorldCom, Inc. Sec. Litig., 346 F. Supp. 2d 628 (S.D.N.Y. 2004).............................................................................. passim In re WorldCom, Inc. Sec. Litig., 2005 WL 638268 (S.D.N.Y. Mar. 21, 2005) .............................................44, 45, 46, 50, 51, 53 Statutes and Rules 15 U.S.C. § 77b(a)(4) .....................................................................................................................42 15 U.S.C. § 77k ................................................................................................................................1 15 U.S.C. § 77k(b) .........................................................................................................................42 15 U.S.C. § 77k(b)(3)(A) ...................................................................................................48, 50, 58 15 U.S.C. § 77k(c) .........................................................................................................................58 15 U.S.C. § 77l(a)(2)..................................................................................................................1, 61 D.C. Code § 31-5606.05(a)(1)(B) ..............................................................................................1, 59 Va. Code Ann. § 13.1-522(A)(ii) ...............................................................................................1, 59 Fed. R. Civ. P. 56(e) ......................................................................................................................41 Regulations and Administrative Materials 17 C.F.R. § 230.176(g) ..................................................................................................................44 17 C.F.R. § 230.191(a)...................................................................................................................42 17 C.F.R. § 230.430B(f)(1) ............................................................................................................47 Circumstances Affecting the Determination of What Constitutes Reasonable Investigation and Reasonable Grounds for Belief Under Section 11 of the Securities Act, SEC Release No. 6335, 1981 WL 31062 (Aug. 6, 1981) ..................................................45, 55 The Regulation of Securities Offerings, 63 Fed. Reg. 67,174 (Dec. 4, 1998) ........................................................................................55 Other Authorities DBRS, Third-Party Due Diligence Criteria for U.S. RMBS Transactions, July 2011 .................22 Financial Stability Oversight Council, Macroeconomics Effects of Risk Retention Requirements (Jan. 2011) ........................................................................................51 Merritt B. Fox, Due Diligence with Residential Mortgage Backed Securities (Columbia Law & Economics Working Paper No. 462, November 25, 2013) .......................45 1 Thomas Lee Hazen, Law of Sec. Reg. § 7.4[2][A][1] (4th ed. 2002) .......................................44 Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 7 of 76 vii Permanent Subcommittee on Investigations, United States Senate, Wall Street and the Financial Crisis: Anatomy of a Financial Collapse (Apr. 13, 2011) ................................38 Webster’s New Int’l Dictionary of the English Language (2d ed. 1934) ......................................50 Webster’s Third New Int’l Dictionary of the English Language (1993) ......................................50 Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 8 of 76 Plaintiff Federal Housing Finance Agency (“FHFA”), as Conservator for the Federal National Mortgage Association (“Fannie Mae”) and the Federal Home Loan Mortgage Corporation (“Freddie Mac” and, together with Fannie Mae, the “GSEs”), respectfully submits this memorandum of law in support of its Motion for Partial Summary Judgment on Defendants’ Due Diligence and Reasonable Care Defenses.1 PRELIMINARY STATEMENT Goldman Sachs’s supposed due diligence investigation into the truth of the representations it made in Prospectus Supplements used to market and sell 42 certificates (“Certificates”) in 40 residential mortgage-backed securitizations (“Securitizations”) to the GSEs fell far short, seemingly by design, of the “reasonable investigation” and “reasonable care” required by the Securities Act of 1933 (the “Act”), 15 U.S.C. §§ 77k, 77l, the Virginia Blue Sky Act, Va. Code Ann. § 13.1-522(A)(ii), and the D.C. Blue Sky Act, D.C. Code § 31-5606.05(a)(1)(B). Goldman Sachs’ underwriter, Goldman, Sachs & Co. (“Goldman”), could not conduct the independent, adverse investigation required by law into the statements made by Goldman Sachs’s issuer, GS Mortgage Securities Corp. (“GSMSC,”) because GSMSC was fully under the control of Goldman, which selected the loans GSMSC would securitize and also controlled the diligence performed on those loans. As held in the seminal Feit v. Leasco Data Processing Equipment Corp. decision, the liability of a Section 11 defendant that is so closely related to the issuer “will lie in practically all cases of misrepresentation.” 332 F. Supp. 544, 578 (E.D.N.Y. 1971). Unable to be independent, Goldman set up a “diligence” process incapable of complying with the law. Goldman could not investigate the veracity of the Prospectus Supplements through their “effective date,” as the Act requires, because Goldman stopped examining loans once it bought them from originators, months before it decided which loans to securitize and before Goldman underwrote the offerings. And it was not possible for Goldman to investigate whether 1 “Goldman Sachs” or “Defendants” refers to The Goldman Sachs Group, Inc., together with the following subsidiaries: Goldman, Sachs & Co. (“Goldman”), GS Mortgage Securities Corp. (“GSMSC”), Goldman Sachs Mortgage Company (“GSMC”), and Goldman Sachs Real Estate Funding Corp. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 9 of 76 2 “the statements” in the Prospectus Supplements were accurate—the entire focus of its underwriting responsibilities—because neither its internal diligence team nor the third-party vendors it retained ever reviewed those Prospectus Supplements or the statements they contained. However, Goldman did use the results of its diligence process to negotiate discounts for itself while concealing those results from investors. One scheme centered on the loan-to-value (“LTV”) ratio, which compares the amount of a mortgage loan to the value of the mortgaged property. LTV ratios are critical in assessing a borrower’s ability to repay a mortgage and the adequacy of the collateral in the event of default, because the ratio indicates the amount of equity a borrower has invested in a property. It was crucial for investors to be confident that the values of the mortgaged properties were accurate. When Goldman bought mortgages it intended to securitize, it routinely concluded that the appraised property values reported to it by the loans’ originators were inflated. In such cases, Goldman generated its own “final values” for the properties, which were usually far lower than the originally-reported values. Goldman used its “final values” to negotiate a discount in the prices it paid for some of these loans because, as a result of its own valuations, Goldman had determined that the loans had higher LTV ratios and thus greater credit risk. Goldman did not disclose these discounts, or that its conclusions that the mortgaged properties it used to collateralize the Securitizations had lower values than the originators reported. Instead, Goldman presented investors such as the GSEs with LTV ratios based on the originally-reported values that Goldman had determined to be false. Goldman’s remaining processes were so ill-conceived that they prevented Goldman from knowing the credit quality of the vast majority of loans in the Securitizations. Instead of examining all loans it securitized or even a statistically valid random sample, Goldman looked only at small samples of loan pools it bought, selecting the samples so haphazardly that its own counsel acknowledged to this Court that Goldman could not draw any conclusions about the quality of the unsampled portions of the loan pools. Goldman also reduced its sample sizes below even its own internal standards, over the protests of its diligence personnel, at the demand of the Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 10 of 76 3 originators from which it was buying the loans. After it reviewed these flawed samples, Goldman made no effort to extrapolate the results to the loans outside of its samples, and never increased its samples to conduct additional credit diligence, no matter what the results showed. Goldman also did nothing when confronted with a series of red flags about the credit quality of the loans it was buying to securitize. Goldman chose to use third-party diligence vendors it did not trust—a view that a July 2007 analysis confirmed by showing that each of its vendors missed serious credit issues—yet Goldman securitized all loans those vendors deemed suitable for purchase without performing any further review. Goldman chose to rely on a diligence database that it knew to be inaccurate and unreliable. And Goldman put certain originators on its internal “watch” or “suspend” lists because Goldman did not trust their lending practices, but then chose to securitize loans it had recently purchased from those same suspect originators without performing any additional diligence. Only one red flag appears to have gotten Goldman’s attention—but not so Goldman could fulfill its duty to verify “the soundness of the security and the correctness of the registration statement and prospectus.” Chris-Craft Indus., Inc. v. Piper Aircraft Corp., 480 F.2d 341, 370 (2d Cir. 1973). Starting in 2005, Goldman partially owned two subprime originators, Lownhome and Senderra, from which it gathered intelligence about the mortgage market. In December 2006, Senderra reported that “credit quality has risen to become the major crisis in the non-prime industry. We are seeing unprecedented defaults and fraud in the market …. [including] inflated appraisals, inflated income and occupancy fraud.” Rather than conduct additional diligence or modify existing diligence processes, the Goldman employees who saw this confidential report focused on developing and implementing a strategy for Goldman Sachs to “take advantage” of the “even greater distress” it knew was occurring. Goldman Sachs promptly established massive and lucrative short positions in mortgage-related securities, while continuing to sell such securities to investors without disclosing its true assessment of the underlying loans, as later confirmed by the Senate’s Permanent Subcommittee on Investigations. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 11 of 76 4 Last, but in no way least, while Goldman reported in numerous Prospectus Supplements that the loans it purchased through its “conduit” channel had been acquired according to Goldman underwriting standards, in fact many of those loans had been originated according to more lenient standards applied by third party originators. When buying these loans, Goldman was supposed to conduct a “gap analysis” that would compare the originator guidelines with Goldman’s conduit guidelines, and identify any instances where the originators’ guidelines were more lenient than Goldman’s. However, during the relevant time period of 2005 through 2007, when it was securitizing conduit loans at a furious rate, Goldman frequently failed to perform these analyses. It simply securitized thousands of conduit loans without any basis for its representations that those loans met Goldman’s underwriting standards. Based on these undisputed facts, FHFA respectfully requests that this Court hold that Goldman cannot establish a due diligence defense under Section 11 or a “reasonable care” defense under Section 12 or the Blue Sky Acts for any of the 40 Securitizations. FHFA further requests that, because there is no evidence that any individual or corporate Defendant in this Action conducted diligence on the Prospectus Supplements, and no Defendant other than Goldman conducted any diligence that went beyond the flawed and inadequate diligence performed by Goldman, the Court enter partial summary judgment for FHFA on these defenses as to each Defendant. FACTUAL BACKGROUND From 2005 through 2007, Goldman Sachs sold 42 Certificates” from 40 Securitizations to Fannie Mae and Freddie Mac. Plaintiff’s Statement of Undisputed Material Facts (“SUF”) ¶ 1. The Securitizations fall into two groups: (1) 36 Securitizations backed by loans that Goldman’s whole-loan trading desk (the “WLTD”) bought from third-party originators (the “WLTD Securitizations”) (SUF ¶ 24);2 and (2) four Securitizations backed by loans that entities 2 Of these 36 Securitizations, 23 were backed by subprime mortgage loans and 13 were backed by prime or Alt-A mortgage loans. SUF ¶ 25. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 12 of 76 5 unaffiliated with Goldman Sachs acquired and securitized (the “Third-Party Securitizations”) (SUF ¶ 27). GSMC was the sponsor of all of the WLTD Securitizations (SUF ¶ 4), and GSMSC was depositor for 35 of the WTLD Securitizations (SUF ¶ 8). Goldman was the lead underwriter for all of the WLTD Securitizations and was lead or co-lead underwriter for all of the Third-Party Securitizations. SUF ¶ 6.3 Goldman Sachs offered and sold the GSE Certificates pursuant to registration statements that included, for each Securitization, a prospectus supplement (“Prospectus Supplement”). SUF ¶ 17. The Prospectus Supplements included representations about the Securitizations, the mortgage loans collateralizing them (the “Mortgage Loans”), and the groups of loans supporting the tranches from which the GSEs purchased the Certificates (the “SLGs”), including, among others, representations that the loans were generally underwritten in accordance with applicable underwriting guidelines. SUF ¶¶ 18-19. The Prospectus Supplements also made representations about the Mortgage Loans’ LTV ratios and combined loan-to-value (“CLTV”) ratios, the number or percentage of loans secured by owner-occupied residences, and the Certificates’ credit ratings. SUF ¶¶ 21-23. I. GSMSC CONDUCTED NO DILIGENCE ON THE WLTD SECURITIZATIONS GSMSC was entirely controlled by Goldman, and as such conducted no diligence beyond that conducted by Goldman. GSMSC “[i]ssued mortgage securities through various shelves,” but it “was not an operating company.” SUF ¶ 9. GSMSC did not have “any business operations other than securitizing mortgage assets and related activities,” and Goldman Sachs’s 30(b)(6) witness was unable to identify any “other activities” in which GSMSC engaged. Id. GSMSC had no employees apart from its officers, and all of its officers and directors were Goldman employees (SUF ¶ 10), none of whom recalled ever attending a single meeting in 3 Appendix A identifies the sponsor, depositor, underwriter and collateral type for each Securitization. Because all of the Securitizations were purchased between September 7, 2005 and October 29, 2007 (First Am. Compl. ¶ 2), the discussion herein refers only to Goldman Sachs’s practices from January 1, 2005 until October 31, 2007, unless otherwise noted. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 13 of 76 6 their GSMSC roles (see, e.g., SUF ¶ 12). All of GSMSC’s corporate functions were performed by employees of Goldman, including human resources, accounting, election of directors, appointment of officers, maintaining books and records, and providing office space. SUF ¶ 13. As GSMSC had no employees of its own, Goldman’s employees selected the loans that GSMSC securitized. Goldman employees also controlled any diligence performed on those loans when the loans were acquired by the WLTD, just as Goldman employees selected the loans that the WLTD acquired. GSMSC itself performed no diligence on the Mortgage Loans. SUF ¶ 16. II. GOLDMAN’S DILIGENCE FOR THE WLTD SECURITIZATIONS WAS INADEQUATE The only entity that purported to conduct diligence on the Mortgage Loans was Goldman. Goldman ran its residential mortgage-backed securities business from within its Mortgage Department, which was led by Jonathan Sobel from 2005 until late 2006 (SUF ¶ 29), and then by Daniel Sparks (SUF ¶ 30). In the Mortgage Department, four groups were responsible for the acquisition and securitization of residential mortgage loans: the WLTD; the Residential Mortgage Finance Group (the “Finance Group”); the Diligence Group; and the Mortgage Capital Committee (the “MCC”). SUF ¶¶ 35, 43, 66-69. From 2005 to 2007, the WLTD, located in New York (SUF ¶ 32), was run by Kevin Gasvoda, who in turn reported to Mr. Sparks (SUF ¶¶ 33-34). The WLTD bought prime, Alt-A, and subprime residential mortgage loans from originators with the aim of securitizing those loans and then selling the securities to investors. SUF ¶¶ 35, 41. The WLTD bought residential mortgage loans in two main ways (SUF ¶ 37): (1) in “bulk,” from originators selling large loan portfolios (SUF ¶ 38), and (2) via a “conduit,” through which the WLTD bought “flow loans (one at a time) and mini-bulk packages (<$80MM) from smaller originators” (SUF ¶ 39). The Finance Group, also located in New York, was run by Michelle Gill. SUF ¶ 64. The Finance Group was responsible for coordinating the final securitization process. SUF ¶ 66. In connection with the structuring, marketing, pricing, and settlement of securitizations, Finance Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 14 of 76 7 Group employees worked with external counsel, third parties such as loan originators and accounting firms, and employees in other groups within the Mortgage Department. SUF ¶ 67. The Diligence Group, sometimes referred to as “Main Street Mortgage,” was a unit within the Finance Group, and was based in St. Petersburg, Florida. SUF ¶¶ 43-44. From 2005 through 2007, Christopher Gething oversaw the Diligence Group (SUF ¶ 45), which included Linda Peterson, Brian O’Brien, and William Shuey (SUF ¶ 46). Linda Peterson directly oversaw the Diligence Group’s diligence process from 2005 until late 2007. SUF ¶¶ 47-48. Beginning in 2006, she also had responsibilities pertaining to the drafting of Goldman Sachs’s underwriting guidelines and the review of sellers’ underwriting guidelines. SUF ¶ 48. Mr. O’Brien was the Diligence Group’s Chief Underwriter, and his responsibilities included developing Goldman Sachs’s conduit underwriting guidelines and “oversee[ing] the credit policy from a credit underwriting standpoint for whole loan and bulk transactions.” SUF ¶¶ 49, 79. Mr. Shuey led the Diligence Group’s efforts on valuation diligence until late 2007, when he took over Ms. Peterson’s responsibilities of managing the entire diligence process. SUF ¶¶ 51-53. The Diligence Group also included several Transaction Managers who oversaw the diligence process for specific transactions, including Joseph Ozment, Stephanie Larson, Lauren Carter, Jon Parkinson, Troy Grehalva, and Keith Conway. SUF ¶ 54-55. Finally, the MCC was an internal, senior management committee that, as part of its “on-going monitoring and review” of Goldman’s “due diligence procedures and processes,” “approve[d] securitizations prior to going to market.” SUF ¶ 69. Senior Goldman executives such as Daniel Sparks and Jon Sobel, as well as Mark Weiss, an Executive Managing Director who headed Goldman’s conduit business, all sat on the MCC. SUF ¶ 70 & nn.5-8. Even though the MCC was tasked with the gate-keeping responsibility of approving Goldman’s Securitizations, the MCC approved only four Securitizations subject to some follow-up while approving at least 30 of the 40 Securitizations with no follow-up. See SUF ¶ 75. For the Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 15 of 76 8 remaining six Securitizations, there is no evidence the MCC reviewed the deals at all. SUF ¶ 75.4 A. Goldman Conducted No Diligence At The Time Of Securitization As Goldman’s counsel made clear to this Court, “I can’t stress enough that the way that diligence was done in this business … was to do diligence at the time of buying the loans.” SUF ¶ 176. Consequently, there is no evidence that Goldman conducted any diligence at the time it securitized the Mortgage Loans. 1. Goldman Conducted No Diligence When Securitizing WLTD Loans, Only When Buying Them As discussed in greater detail in Part I.B, infra, Goldman purchased loan pools from originators and other sellers—or, as Goldman called them, its “clients” (SUF ¶ 175)—with the intent of having GSMSC securitize the loans, and of then selling the securities to investors such as the GSEs. SUF ¶ 41. Often, Goldman combined, or “commingled,” loans from multiple acquisition pools to form a single SLG. See, e.g., SUF ¶ 182. For 26 of the 36 WLTD Securitizations, Goldman created the SLGs by commingling loans from multiple acquisition pools. Declaration of Charles Cipione in Support of Plaintiff’s Motion for Partial Summary Judgment (“Cipione Decl.”) ¶ 35; Cipione Decl. Ex. 13. After settlement, Goldman would hold the acquired pools on its books until the loans were securitized. SUF ¶ 177. Data produced by Defendants show that Goldman held many loans it placed into the SLGs on its books between one and two months before securitization. Cipione Decl. ¶ 28.5 However, the same data shows that Goldman often held loans for much longer. For example, data produced by Goldman Sachs’s expert indicates that for the 74,105 Mortgage Loans for which Goldman could locate diligence data, nearly 20,000 of them, or over a 4 These six Securitizations are: ACCR 2005-4, AHMA 2006-1, GSR 2007-AR2, INDX 2005-AR18, and INDX 2005- AR27. SUF ¶ 75 & n.12. 5 Attached as Exhibit 11 to the Cipione Declaration is a chart showing how many loans in the SLGs for the Securitizations were securitized in the intervals of 0-30 days, 31-60 days, 61-90 days, 91-120 days, 121-150 days, 151- 180 days, and over 180 days after acquisition. Conservatively, this chart presumes that diligence was performed on acquisition loan pools as of the settlement date of the acquisition, even though most diligence would have been performed prior to that date. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 16 of 76 9 quarter, were held on Goldman’s books for longer than two months. Cipione Decl. ¶ 29. Goldman held loans even longer for specific Securitizations: it took between four and five months to securitize nearly five hundred loans into the GSR 2006-OA1 Securitization and between three and four months to securitize over a thousand loans into the GSAMP 2005-HE6 Securitization. Cipione Decl. ¶ 30. And Goldman held an acquisition pool from the originator RFC, called the “RFC 5Y MAY3007 FLX” pool, on its books for over six months before securitizing 676 loans from that pool into the GSR 2007-OA2 Securitization. SUF ¶ 180. There is no evidence that Goldman conducted any further inquiry into any of these pools during the months that passed between acquisition and securitization, including when Goldman securitized those loans into the SLGs or when it prepared the Prospectus Supplements. SUF ¶ 187. The Diligence Group’s personnel recognized, however, that additional information could become available between acquisition and securitization that bore on the creditworthiness of the loans, including “extant debt that was previously undisclosed at the time that [the borrowers] took out their mortgage.” SUF ¶ 188. Further, following acquisition, servicing information could reveal additional information about whether the property was actually owner-occupied or whether the loan was in default. Id. “[P]eople are creative in hiding info at time of application,” Mr. Shuey noted in an email, directing that the Diligence Group run AppIntell’s fraud detection tools on 107 loans in the GSAMP 2006-NC2 Securitization that had gone delinquent. SUF ¶ 189.6 “[I]f you’re buying a loan of a borrower who purchased a home less than 60 or 90 days ago,” testified Ms. Carter, “often times the new address isn’t showing up accurately. So if you’re buying a loan that’s seasoned six or seven months, then that’s much easier to detect…. But in a purchase, it’s much more difficult until time has passed.” SUF ¶ 188. It was not part of Goldman’s policy to review additional information that would reveal potential disparities in a borrower’s credit or occupancy status that became available in the months after acquisition (SUF 6 AppIntell was a third-party program that Goldman Transaction Managers used to screen for borrower fraud. SUF ¶ 54. While Goldman personnel recognized that an originator using AppIntell or a similar program would “have been able to better detect fraud in the origination of loans,” Goldman never required any of its originators to do so. SUF ¶ 58. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 17 of 76 10 ¶¶ 187-88), and thus, Goldman made no effort to review any such information before securitizing the loans. 2. Goldman Did Not Verify the Representations in the WLTD Securitizations’ Prospectus Supplements Goldman allocated ultimate responsibility for “[e]nsur[ing] prospectus supplements were complete and accurate in all material respects” to the Finance Group. SUF ¶ 364. While the Finance Group reviewed certain sections of the Prospectus Supplement describing the loans, as Ms. Gill described it, SUF ¶ 368. The Finance Group did not review any loan tape data, and it did not review any loan files to verify that the Prospectus Supplements’ representations were true. SUF ¶ 364. Instead, it relied solely on the diligence the Diligence Group performed when the WLTD bought the loans. SUF ¶ 368. Separately, the Finance Group was responsible for coordinating the process by which “statistical or numeric information [for the Prospectus Supplements] was prepared by the collateral and analyst team, and verified and tied out with external accountants.” SUF ¶ 364. This process consisted of an external accounting firm’s limited review of the information presented in the Prospectus Supplements by (1) comparing a limited number of data points in the collateral tables in the Prospectus Supplements to data in the loan tapes; (2) ensuring that there were no mistakes in the calculations; and (3) for a very small number of loans, comparing certain loan tape data to certain documents in loan files provided by Goldman. SUF ¶ 366. The accounting firms did nothing to verify the accuracy of the information in the loan files, loan tapes, or the disclosures in the Prospectus Supplements. SUF ¶ 367. The firms specifically disclaimed any role in the “verification of any of the information set forth on the Preliminary Data File [defined as the ‘computer-generated mortgage loan data file and related record layout with respect to [the loan sample]’],” “ma[d]e no representations concerning the accuracy or completeness of any of the information contained therein,” and “express[ed] no opinion or any other form of assurance regarding (i) the reliability or accuracy of the data and Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 18 of 76 11 documents furnished to [them] by the Company, which were used in [their] procedures or (ii) the adequacy of the disclosures in the Offering Documents or as to whether any of the statements expressed therein omit any material facts.” Id. As Deloitte & Touche’s 30(b)(6) witness confirmed: “we would not confirm the accuracy or the validity of the information” Goldman provided. Id. According to Mr. Shuey, the Diligence Group had “no role whatsoever” in ensuring that the statements in the Prospectus Supplements were accurate. SUF ¶ 362. “Nobody on the Due Diligence Team was tasked with even corresponding with people to give them information … to make sure that the SEC disclosures were accurate.” Id.; see also SUF ¶ 361 (Mr. Gething, the head of the Diligence Group, expressing no recollection of ever being asked during the 2004- 2007 period “to review prospectus supplements for Goldman Sachs’ securitizations to determine whether representations contained in those prospectus supplements concerning due diligence were complete and accurate”). Members of the Diligence Group were not even familiar with the term “prospectus supplement.” SUF ¶¶ 360, 363. Likewise, employees and representatives of Goldman’s third-party due diligence vendors testified that they never reviewed the Prospectus Supplements to ensure their accuracy. Clayton, for example, never took “any steps to see if the loan complied with the statements in the prospectus supplements,” and “did not even receive a copy of a the prospectus supplement at the time it was conducting its credit reviews.” SUF ¶ 320. And, as discussed in Part I.B.1.b, infra, these diligence firms—which Goldman itself did not trust, see Part I.B.2.b.i, infra—only analyzed samples of loans drawn from the acquisition pools that Goldman intended to purchase from originators, not from the SLGs. B. Goldman’s Acquisition-Stage Diligence Was Unreasonable And Insufficient Rather than conduct securitization-stage investigations into the SLGs and the statements in the Prospectus Supplements, Goldman purported to run several different types of diligence only when acquiring loans. Specifically, Goldman claimed to conduct credit, compliance, and valuation diligence. SUF ¶ 131. Credit diligence was supposed to evaluate if a loan had been Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 19 of 76 12 originated pursuant to the credit requirements of the originator’s underwriting guidelines, which typically established such criteria as minimum credit scores, maximum debt-to-income (“DTI”) ratios, and maximum LTV and CLTV ratios. SUF ¶¶ 132-33. Compliance diligence was supposed to focus on whether a loan had been originated in accordance with state or federal law. SUF ¶ 134. Both credit and compliance diligence involved reviewing the loan file to ensure that Goldman Sachs was supposed to “eliminate unacceptable credit/compliance file issues.” SUF ¶ 134. Valuation diligence evaluated if the property’s appraised value was reasonable and accurate. SUF ¶ 135. Goldman had no diligence process focused on fraud review and investigations until mid-2007. SUF ¶ 56.7 1. Goldman Conducted No Diligence On The Vast Majority Of Loans Included In The SLGs Goldman knew the originators of the loans it purchased had made those loans solely to sell them to investors like Goldman, as part of an originate-to-distribute business model. See, e.g., SUF ¶ 40 (Goldman counterparty review noting originator “does not retain the loans in its portfolio for more than few months” but instead “sells the loans on a whole loan basis … to investors”). Despite this knowledge, Goldman’s Diligence Group did not review each loan the WLTD acquired. Rather, the Diligence Group used an inaccurate and unreliable database to examine only samples of loan pools, and did not use any scientific method or discernible criteria in drawing its samples. Goldman also made no attempt to extrapolate the results of its review to the loans in the acquisition pools outside of its samples, much less to the loans pulled from those pools months later to form the SLGs. 7 Goldman also engaged in a process it referred to as “collateral diligence” or “custodial diligence,” which involved the review of the files held by a custodian to “confirm existence of all original notes, mortgages, title policies[,] and assignments.” SUF ¶ 171. SUF ¶ 173. Goldman also used the same third-party diligence vendors to conduct a “data integrity” review at the acquisition stage, which simply compared data in the loan tapes—electronic spreadsheets of data describing the loans in the acquisition pools—with the loan origination files to ensure the information on the former matched the information in the latter. SUF ¶ 169. This process did not include verifying the accuracy of the information on the loan tapes or the loan origination files. SUF ¶ 170. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 20 of 76 13 (a) Goldman Believed Its Internal Diligence Database Was Unreliable, Yet It Relied On That Database in Conducting Diligence To keep track of the thousands of loans it bought, Goldman developed a “proprietary database” called “Whole Loan Asset Miner,” or “WHAM,” which it highlighted in presentations to investors such as the GSEs. SUF ¶ 190. WHAM played an integral part in Goldman’s diligence process. SUF ¶¶ 191-97. Goldman uploaded all of the data from loan tapes it received from originators into WHAM, storing hundreds of fields of data for each loan. SUF ¶ 191. Goldman used WHAM to select the samples for its acquisition-stage diligence, based on summary statistics generated from the data in WHAM. SUF ¶ 192. Goldman also stored the results of its diligence in WHAM, and was supposed to use WHAM to track loans after acquisition to determine, for example, which servicer held which loan. SUF ¶¶ 195, 197. Goldman used WHAM throughout the relevant time period. SUF ¶¶ 195, 202. As early as November 2005, however, Goldman knew that WHAM was unreliable. SUF ¶ 199. As one Transaction Manager testified, WHAM was unreliable because “data was introduced [into WHAM] from different sources. And so one could not feel completely confident that it was reliable. It was always necessary to review that data further and make a determination of whether it was real data or something that was accidentally loaded to the wrong field or however it might have been handled.” Id.; see also SUF ¶ 201 (different Goldman Transaction Manager noting that WHAM “rel[ies] heavily on the accuracy of the initial upload”). Goldman’s counsel has similarly represented to FHFA that “much of the information in the WHAM database was input manually, and the process of inputting that data was a work in progress that was not completed. As a result, we understand that the information in the WHAM database has been found at times to be both incomplete and inaccurate.” SUF ¶ 203; see also SUF ¶ 204 (June 6, 2014 letter from Goldman’s counsel stating that “[w]e also have not been able to determine whether th[e] data [in WHAM] was used in a consistent or reliable fashion”). Goldman’s diligence employees were particularly troubled by WHAM’s inability to keep track of loans that Goldman had refused to purchase during prior acquisition-stage diligence, but Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 21 of 76 14 which originators were attempting to re-sell to Goldman as part of a new acquisition pool. Rather than alerting Goldman that it was purchasing previously rejected loans, “WHAM ignores the loans that exist in WHAM and continues to upload the other loans in the pool,” such that the previously rejected loans “wouldn’t show up in the report or the summary on the current [acquisition] deal name that we’re looking to.” SUF ¶ 200. There is no evidence Goldman ever implemented policies to account for this gap, or otherwise established procedures to allow Goldman to track previously rejected loans through WHAM. Consequently, at “the first data step” in the diligence process (id.), Goldman did not know if it was reviewing accurate loan data, but there is no evidence Goldman took any steps to address these issues during the relevant time period. (b) Goldman Selected Samples Of Loans To Be Diligenced From Pools Of Loans It Acquired Rather Than From The Securitizations Or SLGs For bulk purchases, even though Goldman’s Mortgage Department Policies required “[a]ll pools purchased by GS Mortgage [to] be reviewed prior to the closing of the transaction” (SUF ¶ 77), the Diligence Group did not review all the loans in WLTD’s bulk acquisition pools. Rather, the Group performed credit and compliance diligence on only a limited sample of loans drawn from each loan pool before the WLTD purchased the pool. SUF ¶ 205. For conduit purchases, the Diligence Group purportedly conducted “[c]redit compliance and valuation diligence on 100% of loan files” up until July 2005. SUF ¶ 207. Starting in the second half of 2005, however, the WLTD used sampling for conduit purchases too, finding that “[e]xperience to date has shown that certain clients recruited by the conduit channel, are of sufficient size and sophistication, to be handled in the same way as larger bulk clients.” SUF ¶ 208. The WLTD and the conduit sales group directed the Diligence Group to sample conduit purchases in the same way it sampled bulk deals and not to review all loans in large conduit acquisition pools. Id. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 22 of 76 15 For loans not included in Goldman’s acquisition-stage diligence samples, the Diligence Group conducted no credit or compliance diligence at all. SUF ¶ 206. As William Moliski, a member of the Finance Group, made clear, “with sampling, [an originator] will be able to slip these loans [not subject to diligence] by.” Id. There is no evidence that Goldman ever increased its diligence to prevent originators from “slip[ping] these loans by” and into the Securitizations. Nor is there any evidence that Goldman did any comparison or analysis to determine how many or which of the sampled loans were included in the Securitizations, let alone in the SLGs. For 26 of the 36 WLTD Securitizations, Goldman Sachs constructed the SLGs with loans drawn from multiple acquisition pools, among which the originators, sample sizes, loan selection, and types of diligence varied. SUF ¶ 182; see also Cipione Decl. ¶ 35; Cipione Decl. Ex. 13. Moreover, in many instances Goldman placed loans from a single acquisition pool into different Securitizations. Cipione Decl. ¶ 36; Cipione Decl. Ex. 14. Yet the diligence summaries presented to the MCC, which approved the Securitizations (SUF ¶ 69), included diligence information at the acquisition pool level, not at the loan level (SUF ¶ 72), and thus provided no information about which or how many loans in the proposed SLGs had received a diligence review.8 8 The MCC was also not presented with accurate information as to the numbers of conduit loans in the securitizations it approved. For example, the MCC memorandum approving the GSAMP 2005-HE5 Securitization indicates that Goldman “purchased 100% of the Mortgage Loans via the Goldman Sachs Mortgage Conduit.” SUF ¶ 115. The Prospectus Supplement for this Securitization, however, states that “approximately 20.88% of the mortgage loans in the trust were acquired from various original loan sellers under GSMC’s mortgage conduit program.” SUF ¶ 115. Similarly, for the GSAMP 2006-HE8 Securitization, the MCC approved the transaction based on the information that the “collateral for this transaction consists of subprime loans purchased via both the Goldman Sachs Residential Mortgage Conduit (93%) and the bulk whole loan acquisition channel (7%).” SUF ¶ 125. Yet the Prospectus Supplement for this Securitization states that only “10.18% of mortgage loans in the trust (the ‘Conduit mortgage loans’) were acquired by the sponsor via the Goldman Sachs Mortgage Conduit Program” SUF ¶ 125. And the MCC memorandum for the GSAMP 2007-HE1 Securitization states that “the collateral for this transaction consists of subprime loans purchased via both the Goldman Sachs Residential Mortgage Conduit (94%) and the bulk whole loan acquisition channel (6%).” SUF ¶ 126. The Prospectus Supplement for that Securitization, however, states that only “0.16% of mortgage loans in the issuing entity (the ‘Conduit mortgage loans’) were acquired by the sponsor from various loan sellers via the Goldman Sachs Mortgage Conduit Program.” SUF ¶ 126. There is no evidence to explain why the descriptions of securitized loans made to the MCC differed so drastically from the descriptions of those same loans that Goldman then provided to investors. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 23 of 76 16 Unsurprisingly, the percentages of loans in the final SLGs that had been diligenced were often substantially smaller than the percentages of loans in the acquisition pools that Goldman’s own policies required. According to data produced by Goldman’s expert, for 16 of the 36 WLTD Securitizations, fewer than 25% of loans in the SLGs had been subjected to credit diligence (Cipione Decl. ¶ 16),9 which was Goldman’s internal “minimum” diligence sample size for Alt-A and sub-prime loan purchases (SUF ¶ 211). In many cases, the diligence was far less than Goldman’s policies required: files produced by Goldman’s expert, for example, show that Goldman conducted credit diligence on just 0.29% of the loans in the GSAMP 2005-HE5 SLG. Cipione Decl. ¶ 17. Goldman’s diligence processes resulted in swaths of loans receiving no review whatsoever. Data produced by Goldman’s expert contains the number of loans in the SLGs that Goldman selected for credit or compliance diligence. Cipione Decl. ¶¶ 13, 37. That data reveals that of the 82,903 Mortgage Loans in the SLGs, Goldman can produce loan-level data showing that it conducted credit diligence on only 26,802 Loans—less than a third of total population. Cipione Decl. ¶¶ 16, 39. Similarly, the same dataset reveals that Goldman can only product loan-level data showing that it conducted compliance diligence on 27,347 Loans—again, less than a third of the total population. Cipione Decl. ¶ 39. And the data produced by Goldman’s expert reveals that for nearly 9,400 loans within the SLGs—or over 11% of the total loans supporting the Certificates that Goldman sold to the GSEs—Goldman can produce no loan-level credit or compliance diligence data at all. Cipione Decl. ¶ 9. (c) Goldman Acquiesced To The Demands Of Originators To Reduce The Number Of Loans Diligenced In The Acquisition Pools Goldman’s internal procedures called for samples of between 25% and 50% from subprime and Alt-A loan acquisition pools. SUF ¶ 211.10 In practice, however, WLTD traders 9 Attached as Exhibit 8 to the Cipione Declaration is a chart showing the number and percentages of loans backing each SLG at issue on which the Diligence Group conducted acquisition-stage credit, compliance, and valuation diligence, according to data produced by Goldman’s expert. 10 For prime loans, Goldman’s internal policy called for a sample size of 10%. SUF ¶ 211. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 24 of 76 17 regularly stipulated to sample sizes smaller than required by Goldman’s own diligence guidelines, which it knew were “more liberal” than the rest of the market’s. SUF ¶¶ 209, 214. According to Lauren Carter, it was “typical for traders for sales at Goldman Sachs to instruct” the Diligence Group on “what percentage to sample for credit compliance” and “what valuation due diligence approach to take.” SUF ¶ 214. Often, the WLTD would give such orders at the insistence of the originators from which the WLTD was buying the loans (SUF ¶ 215), and for reasons wholly unrelated to the characteristics of the acquisition loan pools (see, e.g., SUF ¶ 213). See also SUF ¶ 214 For example, the WLTD stipulated to a 10% sample size for a purchase from Countrywide (SUF ¶ 215), prompting Mr. Ozment, a Diligence Group Transaction Manager, to ask: “Our minimum DD requirements on Alt-A is 20%. Did we really agree to 10?” Id. William Moliski, a member of the Residential Finance Group, replied: “We have been buying a significant amount of loans from Countrywide. If we are still requiring at least 20% for them, please contact me because this will prevent us from buying from Countrywide. Also, please forward the policy that states our minimum sampling requirement. Have never seen this and have only spoken about ranges. In the past we have sampled less than 20% for IndyMac in particular.” Id. (emphasis added). John Cassidy, the Diligence Group’s Compliance Manager, pushed back: “We have had significant issues across the product spectrum with C-Wide. B[efore] we go below 20% can you give me the pull-through on similar trades that justifies us going below our norm[?]” Id.(emphasis added). There is no evidence such a justification was provided. To the contrary, this pattern repeated itself throughout the relevant time period for numerous acquisition pools that included loans that wound up in the SLGs. For example, for the Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 25 of 76 18 RFC 5Y MAY3007 FLX pool, discussed in Part II.A.1, supra, which was ultimately included in the relevant SLG of the GSR 2007-OA2 Securitization, Goldman’s Chief Underwriter thought a “minimum of 50% review” of the pool was appropriate—but Goldman’s trading desk refused to agree to such a level. SUF ¶ 226. Instead, RFC allowed Goldman to sample loans outside the 25% sample size, as stipulated, so long as the extra loans were restricted only to loans Goldman had previously found to be unsuitable for securitization. SUF ¶ 227.11 Likewise, WLTD traders agreed to a 10% sample size for an American Home pool even though David Parkinson, a Transaction Manager, thought Goldman “should do 20-25% as that is standard for these pools.” SUF ¶ 213; see also id. (10% sample for acquisition pool CW COFI1 JUN302006 ARM); id. (15% sample for acquisition pool CW71 JAN302006 ARM); id. (11% sample for acquisition pool CW2_5Y MAR2907 FLX); id. (12% sample for acquisition pool CW_5Y MAR2907 FLX); id. (16.40% sample for acquisition pool CW71 DEC142005 ARM); id. (15.50% sample for acquisition pool CW101 DEC142005 ARM); id. (for all Countrywide deals in December 2005, “sample sizes were … ~16% for the Alt-A”). As one Goldman trader put it in purchasing four Countrywide pools, he “had to agree to a ~15% sample size for this trade,” instead of Goldman’s “typical” 20% sample size, “in order to get it done.” SUF ¶ 215. As a result, Goldman employees believed they had not identified all loans in the acquisition pools that might be outside its credit parameters. “How do we know we caught everything?” asked a member of the Mortgage Department, after learning that Goldman had stipulated to a sample size of 10-15% for a pool of Countrywide loans—479 of which wound up in the GSR 2006-OA1 Securitization. SUF ¶¶ 231, 233. “Beca[use] of the limited sampling on CW,” replied Transaction Manager Keith Conway, “we don’t catch everything.” SUF ¶ 233 11 According to data produced by Goldman’s expert, of the 302 loans subject to credit diligence from this RFC 5Y MAY3007 FLX pool, 235 received a final grade of EV3 or EV4 from the diligence vendor. SUF ¶ 230; see also Part I.B.2.b, infra (explaining that a third-party diligence vendor’s grade of EV3 meant that a loan did not comply with an originator’s underwriting guidelines and that a grade of EV4 meant that Goldman had overridden the vendor’s grade of EV3). While 105 of these 235 loans were dropped from the pool, another 106 were ultimately securitized in the relevant SLG of the GSR 2007-OA2 Securitization. SUF ¶ 230. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 26 of 76 19 (emphasis added) (also noting that “[t]his trade had issues with aged loans and we tried to get pay histories and were told they would not provide them.”). (d) Goldman Did Not Apply Consistent Standards For Selecting Its Samples, And Did Not Include All Loans That Met Its “Adverse” Credit Criteria Once a WLTD trader set a sample size, the Diligence Group would choose the loans within the sample; it was supposed to select a portion of the loans at random, with the rest selected based on “adverse” loan characteristics. SUF ¶ 237. Although selecting the loans was a manual process (SUF ¶ 238), the Transaction Managers did not receive any training, statistical or otherwise, in how to do so (SUF ¶ 239). See also SUF ¶ 60 (Transaction Managers did not receive any formal training when they began employment at Goldman, only on-the-job training.). As a result, the Diligence Group’s manual processes resulted in significant variances in sampling it conducted, when sampling was conducted at all. For example, there is no evidence that the Diligence Group analyzed whether its sampling procedures produced truly random samples; instead, the Diligence Group often simply used a Microsoft Excel formula to select loans. SUF ¶ 241. At times, Goldman would also substitute loans for Goldman’s “randomly- selected” sample based on whether the loans met compliance review criteria such as whether the loan was originated in a jurisdiction with particular laws on predatory lending. “We were required to do diligence so much in compliance based on regulatory and state guidelines,” explained Ms. Carter. “So sometimes you would fit that within the random [portion of the sample] as well.” SUF ¶ 242. As Goldman’s counsel summarized to this Court, Goldman’s “randomly-selected” loans were chosen “randomly but not in a statistical way.” SUF ¶ 241. As a result, even if Goldman had attempted to use the results of its samples to ascertain the quality of the remainder of the acquisition pool—and as discussed in Part I.B.1.e, infra, Goldman made no such attempt—it is undisputed that Goldman’s samples did not form a reliable basis to do so. See Mem. in Support of Defs.’ Preliminary Daubert Motion to Exclude the October 10, 2012 Expert Report of Charles D. Cowan, Ph.D, at 6 (Dkt. No. 135) (“Where an Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 27 of 76 20 expert attempts to use a sample to draw conclusions about a larger population, those conclusions are deemed reliable only if the sample is proven to be random.”).12 And because Goldman conducted diligence only at the acquisition stage, not during securitization, there is no evidence that, for the 26 Securitizations drawn from multiple acquisition pools, Goldman ever examined whether it was reviewing a truly random sample of the SLGs. Much of the time, moreover, Goldman failed to even perform a random selection. Many of the sampling spreadsheets for the acquisition pools show no evidence of any random samples being drawn. SUF ¶¶ 243-44. Goldman’s own expert confirms this trend, for example agreeing that no random sampling was ever done on the FMT DEC282006 2 pool, which contained loans that Goldman placed into the GSAMP 2007-FM2 Securitization. SUF ¶ 244. Goldman’s expert also acknowledges that Goldman conducted no random sampling on the four Third-Party Securitizations, as discussed in Part III, infra. After selecting the purportedly randomly selected component of the sample, the Diligence Group would often reduce—or “haircut”—the number of those loans that went into the final sample, to ensure that the overall sample size remained under the stipulated cap. SUF ¶ 243. These “haircut[s]” would occur even though Goldman’s policies called for the randomly selected portion of the sample to be between 10-25% of the sampled loans. SUF ¶ 242. As with the rest of the sampling procedure, SUF ¶ 243. In “adversely” selecting loans for the remainder of a sample, the Diligence Group used manual techniques to look for “what we considered to be the risk in the portfolio both from a credit and a compliance perspective.” SUF ¶ 237. But the Group had no systematic process or 12 In their June 3, 2014 Joint Memorandum of Law in Opposition to Plaintiffs’ Motion for Summary Judgment on Knowledge (“Knowledge Opp.”), Goldman, along with all other defendants in these coordinated actions, concedes that its sampling did not reveal, and was not designed to reveal, anything about the unsampled loans in the acquisition pools: “[T]he majority of loans often were not subjected to credit and compliance due diligence, and therefore would have been affected by whatever underwriting deficiencies existed in an Originator’s general pool of loans.” Knowledge Opp. 2; see also id. at 25. Had Goldman’s sampling been sufficiently random and representative, Goldman would have been able “to draw conclusions about [the] larger population,” (Dkt. No. 135 at 5), which Goldman concedes it did not do. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 28 of 76 21 method for selecting such loans until at least June 2007. SUF ¶ 245. Until then, the Diligence Group recognized that “the sample selection process was a manual function that the T[ransaction] M[anager] performed[,] which had the potential for bias and errors.” Id. The Diligence Group would identify potentially adverse loans based on various criteria, including credit scores less than 500, DTI ratios greater than 55%, LTV ratios greater than 100%, and loan balances greater than $1 million (SUF ¶ 237), but the Diligence Group would not include in the adverse portion of the sample all loans in an acquisition pool that met these criteria. A Transaction Manager would not “consider picking all of the loans with adverse characteristics for further review”; rather, “just the loans that kind of stuck out to me.” SUF ¶ 246. The Transaction Managers would not include even all “striking” loans in the final adverse sample, selecting instead only as many loans as permitted by the WLTD traders. Id. In an effort to “standardize the due diligence sample selection process,” Goldman eventually developed an automated system that would help it “hopefully become more scientific in our sample selection rather than hitting broad areas we know to be more risky but not exactly how risky[.]” SUF ¶ 245. The system had “built-in formulas” that would “filter through and flag high risk loans”—in contrast to Goldman’s incumbent “manual process.” Id. While there is evidence that Goldman was testing this new system as of mid-2007, there is no evidence that the Diligence Group applied it to any of the acquisition loan pools from which the SLGs at issue were drawn. Id. (e) Goldman Did Not Extrapolate the Results Of Its Diligence Reviews To The Unsampled Portions Of The Acquisition Pools The Diligence Group did not extrapolate the results of its review of samples to determine the characteristics of the full acquisition pools. SUF ¶ 248. “You can’t do that particularly because of adverse selection,” said Mr. Gething. “You can’t take – start out with a series of loans that were pointedly chosen as opposed to randomly chosen, pointedly chosen and then say well, I found certain things in there, therefore I’m going to extrapolate it across the universe.” Id. As a result, when asked how Goldman could “ensure that all loans with problems were detected when Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 29 of 76 22 [it was] focusing on only a subset of the entire loan population,” Stephanie Larson, a Transaction Manager explained, “I can’t ensure the – what the – what the outcome of the whole pool is based on the sample.” SUF ¶ 206.13 While Goldman employees testified that, theoretically, sampling could be generally utilized to allow for extrapolation (e.g., SUF ¶ 247), there is no evidence that Goldman ever performed that exercise. As one Diligence Group employee acknowledged, even when the diligence results indicated that the unsampled portion of the pool contained problematic loans, Goldman still purchased and securitized them: Q. In drawing a sample for the purposes of credit due diligence, if for instance you determined that 7 percent of the random sample had fatal credit issues, would you then infer that 7 percent of the entire pool had fatal credit issues? … A. Yeah, it would. … Q. In other words, if you’re drawing a sample of an entire pool and you determine that 7 percent of the random sample has credit issues, and you just told me that might mean that 7 percent of the entire pool would have credit issues, how could those loans that had not appeared within your sample be caught if they had potential credit issues? A. I don’t – well, I see what you’re saying. So, yeah, I guess you’re right. SUF ¶ 248. And because Goldman did not do any diligence at the securitization stage, it cannot present evidence it ever extrapolated its samples of individual acquisition pools to the SLGs. 13 See also DBRS, Third-Party Due Diligence Criteria for U.S. RMBS Transactions, July 2011 at 5 n.2, http://dbrs.com/research/241165/third-party-due-diligence-criteria-for-u-s-rmbs-transactions.pdf (emphasizing “the importance of the randomness of the sample” and the need to “confer[] as necessary with the due diligence firm to ensure random selection”). Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 30 of 76 23 2. Goldman Failed To Follow Up On Red Flags By Conducting Additional Diligence (a) Goldman Did Not Increase the Size of its Samples to Conduct Additional Credit Diligence Goldman did not have any standardized process in place to expand its limited sampling to conduct additional credit diligence, where warranted, on the acquisition pools that fed into the SLGs. Goldman’s policies on increasing its sample size after conducting an initial review, referred to as “upsizing,” were ad hoc at best: in order to upsize, a Transaction Manager had to request approval from Mr. Gething, the head of the Diligence Group, from the relevant trader, and from the Finance Group. SUF ¶¶ 217-18. However, until at least April 2007, Goldman did not have any formal policies or procedures regarding when upsizing was required or whose approvals were necessary in order for a sample to be upsized. SUF ¶ 219. Indeed, high “drop rates would not necessarily produce” an upsized sample, according to Lauren Carter, a Transaction Manager. SUF ¶ 218. Nor would Goldman upsize a sample if a third-party originator insisted on maintaining a smaller sample size to which the parties had stipulated. SUF ¶ 222. As a result, even though Ms. Peterson advised that “[i]f we have any ability to increase our samples particularly in subprime, we should be doing so” (SUF ¶ 210), it was “[v]ery limited, rare” for the Diligence Group to actually make such a request (SUF ¶ 220). Specific to this case, there is no evidence the Diligence Group systematically conducted additional diligence on the loans outside of its samples in light of loans identified by its vendors as being unsuitable for purchase. Of at least 276 acquisition pools from which the SLGs were drawn (Cipione Decl. ¶ 10), Goldman increased its sample sizes for only 10 pools, each time to conduct additional compliance diligence (SUF ¶ 221). There is no evidence that Goldman ever “upsized” its sample sizes to conduct additional credit diligence on the relevant acquisition loan pools for any reason. See SUF ¶ 221. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 31 of 76 24 (b) Goldman Failed To Properly Scrutinize Its Third-Party Diligence Vendors After it drew a sample from a potential loan acquisition pool, the Diligence Group retained a third-party diligence firm—generally Clayton Holdings Inc. (“Clayton”), the Bohan Group (“Bohan), MDMC, Watterson Prime, or Lydian—to conduct credit and compliance diligence on the sampled loans. SUF ¶ 288. Lydian worked exclusively on loans acquired through the WLTD’s conduit channel. SUF ¶ 289.14 Once it reviewed a loan, the diligence firm would grade the loan as “EV1,” “EV2,” or “EV3.” SUF ¶ 299. A grade of EV1 indicated the loan complied with the originator’s underwriting guidelines. SUF ¶ 300. A grade of EV2 indicated the loan did not comply with the originator’s underwriting guidelines, but purportedly had sufficient compensating factors to offset the increased credit risk such that Goldman could purchase and securitize the loan. SUF ¶ 301. A grade of EV3 indicated the loan did not comply with the originator’s underwriting guidelines, lacked sufficient compensating factors, and posed an unacceptable credit risk or presented a regulatory compliance issue. SUF ¶ 302. Loans where the Diligence Group overrode the vendors’ EV3 grades were marked in Goldman’s diligence reports as “EV4s.” SUF ¶ 314.15 Goldman set the boundaries of the third-party diligence vendors’ analyses. With Clayton, for example, Goldman instituted “particular requirements” as to which Clayton employees could analyze its loans, such that Clayton maintained an internal “Goldman do not use list.” SUF ¶ 298. Clayton generally reviewed only those loans identified by Goldman, and did not know if it was reviewing an entire acquisition pool or just a sample. SUF ¶ 294. Goldman instructed Clayton on which criteria to use in its diligence, sent the initial diligence assignment that structured the vendor’s review process, and provided the guidelines for Clayton to use for its 14 The Diligence Group also retained a number of third-party diligence firms to conduct valuation diligence, including Hansen, CoreLogic, Fiserv, and AppIntell. SUF ¶ 290. In addition, the Diligence Group relied on large banks to conduct the custodial or collateral review (SUF ¶ 291), which, as noted in n.7, supra, Goldman did not consider to be “diligence,” but rather part of the final closing-out process that simply examined whether mortgage-related documents were present. 15 Third-Party diligence vendors occasionally referred to “EV4s” as “EV2Ws.” SUF ¶¶ 314 n.23. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 32 of 76 25 review. SUF ¶ 296. Goldman established the same restrictions for its other diligence vendors. SUF ¶ 303. (i) Goldman Had Serious Concerns About The Competence Of Its Third-Party Diligence Vendors, But Continued To Use Them Without Any Increased Oversight The Diligence Group had serious concerns about the ability of its third-party diligence vendors to properly review and accurately grade loans. The Diligence Group was chronically concerned about Clayton. “I’m having real issues with them,” wrote Mr. Ozment in late 2005. “I have other trades that I’m having to really push them just to get them to perform there [sic] jobs [sic]. If I had a better alternative for DD reviews, I would certainly use them. They are like the lesser of all evils.” SUF ¶ 321. Those concerns persisted throughout the relevant time period. In early 2006, Lauren Carter informed Linda Peterson that she had “4 loans in the Ameriquest scratch and dent where the borrowers are dead[;]” however, “Clayton review states loan is delinquent, but accept[s] as is and purchase.” A year later, Clayton’s performance had not improved: “It is clear that they do not know what they are doing,” Ms. Carter wrote in early 2007, after Clayton did not reject any loans from a pool based on flags raised by AppIntell’s fraud-detection products. “We have discussed this over and over again, several of these loans had huge fico drifts as well as [‘]investigates[’] on appintell yet clayton clearly passed each one.… I keep coming to the same conclusion, these loans are not receiving the proper credit underwrite.” SUF ¶ 321.16 The Diligence Group’s view of MDMC was no better. “[I]n the sub prime arena MDMC does not adequately review the credit files” wrote Ms. Carter. “We have recently found several errors with their credit underwriting.” In fact, for a period of time, Goldman stopped using MDMC for compliance diligence because it was not meeting Goldman’s standards. SUF ¶ 324. The Diligence Group was equally concerned with Lydian, its diligence vendor for conduit purchases. In February of 2006, Mr. Gething was concerned that there might be “something 16 “‘Higher risk’ loans were those that received a … Value Verify Status of ‘High’ or “‘Investigate’ from AppIntell.” SUF ¶ 321. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 33 of 76 26 systemic with Lydian work.” SUF ¶ 325. Later, in March 2006, he wrote that “[William] Shuey and I have plenty of concern about their work quality and the fact that they can’t generate proper reporting…. I am going to redirect some of the conduit bulk flow away from Lydian to our third party vendors.” Id. Commenting on Mr. Gething’s email, Mr. Ozment wrote to Ms. Carter: “Have you seen this? Ouch! I hate Lydian/Conduit….” Id. (ellipsis in original). In November of 2006, a quality control summary regarding Lydian stated: “the quality of Lydian’s work also decreased” with specific concerns including “18% of the sample had marginal or critical errors. If this sample truly indicates the quality of work by Lydian, this would mean that 180 loans or 18% of Lydian’s entire production has marginal or critical errors,” yet “zero of the 18 loans with marginal or critical errors were escalated to a Goldman Sr. Underwriter for review,” and “quality of work deteriorated in the month of October despite Goldman submitting half of the amount of loans as September.” Id. Despite these concerns, the Diligence Group had no quality control process for its vendors during the relevant time period. Loren Morris, the Vice President of Goldman’s Residential Conduit Staff, concluded “we should devote resources to loan level quality control reviews that will conduct a full file review on a sample basis for loans purchased (usually 10%).” “This would provide us with details to assess the vendor reviews, the loans and the sellers-- shortly after purchase. The information can be a great feedback loop for DD and credit.” Goldman did not have such a process as of June 2007, however, and, as Ms. Morris wrote at that time, “we probably are about 4-6 months out.” SUF ¶ 327. Goldman’s concerns were partially validated at least as early as March 2007, when Goldman performed a post-securitization review to determine why several bonds it had sold in 2005 and 2006 were performing so poorly. Goldman looked into 77 loans that Clayton had reviewed in 2006, which included “quite a few that were [marked EV]1’s by [C]layton.” As Gregg Smigelsky, the Goldman employee who ran the review, explained, for many of the loans that Clayton had classified as EV1s in 2006—which no Goldman employee had reviewed at the time, see Part I.B.2.b.ii, infra—in 2007 “we dug deeper and found fraud.” Despite this, Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 34 of 76 27 Goldman continued to use Clayton after March 2007, including for purchases that fed into four Securitizations at issue here. SUF ¶ 322.17 In July 2007, Goldman at last “decided to put our vendors to the test.” SUF ¶ 328. In September 2007, Goldman gave the same 28 loans to Clayton, MDMC, Watterson Prime, and JCIII—Goldman opted not to test Lydian—and asked those firms to review them. SUF ¶ 329. Goldman “expected” that each firm would grade 14 loans as passing with non-material credit exceptions, 13 as failing with material credit exceptions, and 3 as failing with material compliance exceptions. SUF ¶ 330. Instead, MDMC’s overall accuracy rate was 35.71%, Clayton’s rate was 60.71%, Watterson Prime came in at 64.29%, and JCIII came in at 71.43%. SUF ¶ 331. Moreover, each vendor missed credit issues that “should have been credit ‘3’ events.” SUF ¶ 332. And the vendors missed different issues, resulting in a “random nature of the credit review.” Id. For example, Watterson Prime viewed a non-arm’s length transaction as an EV3-worthy credit issue, while Clayton, JCIII, and MDMC did not. Id. Goldman graded the vendors’ error rates in performing credit diligence as 92.31% for MDMC, 61.54% for Watterson Prime, 53.85% for JCIII, and 46.15% for Clayton. Id. The test uncovered other problems as well: • The vendors did not properly assign underwriters based on seniority, e.g., assigning a first-year underwriter to review a bank statement file for a reduced-documentation loan while assigning a 20-year veteran to review a “full doc” loan; • The vendors’ quality control process was “random;” • The vendors did not understand or uniformly test for misrepresentations; and • The vendors did not understand or uniformly test for property value. SUF ¶ 333. Goldman recognized that these “[t]est results show that we need to take action and that we are unable to continue using the same DD process.” SUF ¶ 334. Yet there is no evidence 17 Those four Securitizations are: GSAA 2007-6, GSR 2007-AR2, GSR 2007-OA1, and GSR 2007-OA2. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 35 of 76 28 that Goldman did anything to respond to these test results during the relevant time period. There is also no evidence indicating why, in light of concerns with its diligence vendors stretching back at least as far as 2005, Goldman did not review the vendors’ performance sooner or modify its use of such vendors. (ii) Goldman Performed No Quality Review Of Loans Its Vendors Graded As EV1s Or EV2s Despite the Diligence Group’s longstanding concerns about its vendors’ competence, there is no evidence the Diligence Group conducted any systematic reviews of the loans that the vendors graded as EV1s or EV2s, i.e., suitable for purchase according to the third-party vendors. SUF ¶¶ 300-01, 307-08. If a vendor graded a loan as an EV1, testified Mr. Conway, a Transaction Manager, “then I would probably assume that’s a safe loan to be included into the pool.” SUF ¶ 307. Similarly, Ms. Carter, another Transaction Manager, would “[n]ot normally” conduct any type of quality control review of loans the vendors rated as EV1s. Id. If a vendor graded a loan as an EV2, Mr. Conway “would potentially look at what the description was of that Event Level 2,” but he could not recall ever taking any further action regarding such loans. SUF ¶ 308. Diligence Group employees could not recall any policy surrounding the review of EV2 loans. Id. According to data produced by Goldman’s expert, of the 26,802 loans in the SLGs subjected to acquisition-stage credit diligence, Goldman’s vendors initially graded approximately 22,700 loans—nearly 85%—as being EV1s (75.82%) or EV2s (8.79%). Cipione Decl. ¶ 41. There is no evidence that Goldman ever reviewed or otherwise questioned the grades assigned by its vendors to these loans. Goldman simply securitized them. (iii) Goldman Actively Sought To Securitize Loans Its Vendors Graded As Not Suitable For Purchase. In marked contrast, the Diligence Group’s Transaction Managers “reviewed every loan that was tagged … with a 3” “to try to get some justification of why it was designated as an Event Level 3.” SUF ¶ 309 (emphasis added). As Goldman’s Conduit program expanded, Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 36 of 76 29 Goldman even formed a Credit Committee to review every EV3 loan on a daily basis. SUF ¶ 320. If Goldman thought a loan graded by a vendor as an EV3 had “sufficient compensating factors” or was able to “find something that perhaps [the vendor] missed,” then Goldman would override the vendor’s grade, and buy it. SUF ¶ 310. In reviewing these loans graded as EV3, the Diligence Group assured originators that Goldman’s goal was “to clear as much as possible.” SUF ¶ 312.18 The Transaction Managers did not refer to any written policies or procedures to ensure that any compensating factors identified by the Diligence Group were documented, or that the additional credit risks associated with the loans were sufficiently offset by standard compensating factors. SUF ¶ 311. “Each exception … was looked at in the context of the trade and the experience with the client,” Mr. Gething testified, “[t]here was no normal policy.” Id. The Transaction Managers could override any EV3 as long as doing so did not violate the WLTD’s bid stipulations and as long as the loan had not received the EV3 grade for compliance reasons. SUF ¶ 313. That left Goldman free, however, to purchase loans that were the “worst of the worst” from a credit perspective. SUF ¶ 317. As a result, large numbers of loans that received final grades of EV3—meaning that Goldman Sachs had reviewed the loans and agreed with its vendor that they did not comply with guidelines and lacked adequate compensating factors—found their way into the Securitizations. For example, according to data produced by Goldman’s expert: Goldman securitized (1) 35 loans that received final grades of EV3s for credit reasons into the relevant SLG of the GSAMP 2006- NC2 Securitization, Cipione Decl. ¶ 43; (2) 24 such loans into the SLG of the GSAA 2006-5 Securitization, Cipione Decl. ¶ 44; and (3) 160 such loans into the relevant SLG of the GSAMP 2006-HE4 Securitization (Cipione Decl. ¶ 46).19 18 Even if the Diligence Group decided that a loan should remain as an EV3, the WLTD traders could buy the loan anyway, as they had the ultimate authority to decide which loans to purchase. SUF ¶ 318. Goldman personnel even implemented a formal process for traders to “‘over-ride’ credit and value decisions so that the loans can be purchased,” and to identify those loans in Goldman’s WHAM system “with a unique indicator so they can be tracked.” Id. 19 To the extent Goldman argues that its diligence vendors compared some loans to Goldman’s own “client preferences” instead of or in addition to sellers’ underwriting guidelines—and there is no evidence that Goldman ever Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 37 of 76 30 (c) Goldman Knew It Should Not Securitize Loans From Certain Originators, Yet It Did So Anyway Without Any Further Diligence Goldman had a “Steering Committee” that reviewed, on an ongoing basis, originators that sold loans to it through the conduit channel. SUF ¶ 260. This Committee included senior management as well as several members of the Diligence Group. SUF ¶ 261. Members of the Diligence Group would circulate “seller status lists” to Diligence Group personnel on a monthly basis. SUF ¶¶ 262-63. These lists tracked originator performance across various criteria, such as early payment defaults, pull-through rates, and the timely receipt of financial documents. SUF ¶¶ 264, 267. The lists also assigned originators one of four statuses: “performing,” “monitor,” “watch,” or “suspend.” SUF ¶ 265. A “watch” status required a 100% review of all loans acquired from the originator to make sure the borrowers were sufficiently credit-worthy, that the loans complied with federal law, and that the appraisals matched values produced in an independent broker price opinion review. SUF ¶ 268. Any purchase from a seller on “watch” required the approval of the Steering Committee before further purchases were made. Id. A “suspend” status meant that Goldman “should not be doing a transaction” with the originator and that Goldman should not acquire any loans from it. SUF ¶ 266. Yet Goldman ignored these same “watch” and “suspend” statuses when it came time to securitize loans: there is no evidence that Goldman ever considered these “watch” or “suspend” statuses when securitizing loans it had previously acquired from originators and was holding on its books prior to securitization. For example: (1) Goldman placed NovaStar on its “watch” list on September 28, 2006 (SUF ¶ 269), but securitized over 1,100 NovaStar loans in the relevant SLGs of the GSAMP 2006-HE7 and GSAMP 2006-HE8 Securitizations in October and December 2006, respectively (SUF ¶¶ 270, 273); (2) Goldman assigned both a “watch” and “suspend” status to MLN in September 2006 and January 2007, respectively, but Goldman securitized numerous MLN loans into the GSAMP 2007-HE1 Securitization in February 2007 communicated any such preferences it might have had to a diligence firm (see SUF ¶ 297)—Clayton confirmed that, when it conducted diligence for clients using client preferences, there was no way to determine whether EV3 grades referred to breaches of those preferences or of the seller’s guidelines (SUF ¶ 303). Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 38 of 76 31 (SUF ¶¶ 270, 278-79); and (3) Goldman placed Weichert Financial on “suspend” status no later than September 28, 2006, but it securitized Weichert loans into the GSAMP 2006-HE7 Securitization just a month later, in October 2006 (SUF ¶¶ 280-81). All told, Goldman securitized over 5,800 loans into seven different SLGs from originators that, according to Goldman’s own standards, either warranted a 100% diligence review or with which it “should not be doing a transaction.” SUF ¶¶ 266, 268, 270.20 There is no evidence that Goldman ever conducted additional diligence on any of these loans—let alone the 100% diligence required by Goldman’s own policies—before securitizing them.21 Nor did Goldman ever disclose these watch or suspend lists to investors. 3. Goldman Failed To Use Its Own Conduit Guidelines To Diligence Loans Acquired Through Its Conduit Program Despite Representations In The Prospectus Supplements That The Loans Conformed To Those Guidelines Goldman’s inadequate diligence process also infected loans Goldman acquired through its conduit channel. In the Prospectus Supplements, Goldman Sachs represented that “substantially all” loans bought through its conduit program “were acquired generally in accordance” with “Goldman Sachs Mortgage Conduit Underwriting Guidelines” (the “Conduit Guidelines”). E.g., SUF ¶ 20; see also SUF ¶ 88 (Goldman Sachs Conduit Sellers Guide made a loan ineligible for purchase through the conduit unless it complied with Conduit Guidelines). In practice, however, the WTLD bought numerous loans through its conduit channel that third-party sellers had originated pursuant to the sellers’ own guidelines. SUF ¶ 90. And the Diligence Group conducted credit diligence on acquisition loan pools using those sellers’ guidelines—not the Conduit Guidelines. Id. 20 Those Securitizations are: GSAMP 2006-HE7, GSAMP 2006-HE8, GSAMP 2007-HE1, GSAMP 2007-HE2, GSAA 2007-6, GSR 2007-OA1, GSR 2007-OA2. 21 To the contrary, Goldman tried to conceal that it was securitizing loans from “suspended” originators. For example, around December 4, 2006, Goldman decided to “suspend” Sebring (SUF ¶¶ 271, 273), in part because it made “some of the worst loan credit decisions that [Brian O’Brien had] seen all year” (SUF ¶ 271). But in January 2007, Goldman “purposely” left Sebring off a list of the originators of loans in the GSAMP 2006-HE8 Securitization that Fannie Mae requested to “check whether [those originators] are approved for upcoming deals,” even though Sebring loans made up 4.67% of the loans in the GSAMP 2006-HE8 Securitization. Id. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 39 of 76 32 Before the Diligence Group could use seller guidelines, Goldman’s internal procedures required the Group to perform a “gap analysis.” SUF ¶ 91. A “gap analysis” involved “a side by side comparison … of the conduit seller’s guidelines to [the Conduit Guidelines] … compar[ing] them and highlight[ing] issues of material difference.” SUF ¶ 92. Based on the gap analysis, Goldman would decide whether it would conduct diligence using the Conduit Guidelines or using the seller’s guidelines while “potentially compensat[ing] or adjust[ing] for the gaps through maybe other mitigating factors or other compensating factors.” Id. Such adjustments were necessary because the gap analyses often revealed that the Conduit Guidelines had materially more stringent requirements than the seller’s guidelines. SUF ¶ 94. Although these gap analyses were mandatory, the Diligence Group did not always perform them. “Often, we are not able to review these [seller guidelines] in time for the trade,” explained Mr. O’Brien who, as Chief Underwriter (SUF ¶ 95), was responsible for conducting the analyses. SUF ¶ 103. In October 2006, Mr. O’Brien complained that “[t]he rate at which we are receiving new and updated guides [from conduit sellers] is increasing at an unmanageable rate.” SUF ¶ 102; see also SUF ¶ 59 (email from Christopher Gething, the head of the office that housed the Diligence Group, acknowledging that after “the mad days of ‘05 through March of 2007 when we were handling 100-120 trades per month,” Goldman’s diligence had “slipped and we have no excuse.” ).22 Consequently, the Diligence Group “cannot identify the items [where the Conduit Guidelines differ from the seller guidelines] because our DD vendors are underwriting to the seller guidelines,” and the seller guidelines “might not specify or be specific in areas that would allow the due diligence vendor to apply [Goldman’s] rules to those loans.” SUF ¶ 105. 22 Originators often pressured Goldman into using the originator’s guidelines without such an analysis. In January 2006, Grainne McNamara, Director of Operations for the conduit group, wrote to the Diligence Group that “guideline exceptions” created by the gaps between loan sellers’ guidelines and Goldman Sachs’s Underwriting Guidelines were “irritat[ing] sellers” and placing stress on Goldman’s resources. Ms. McNamara wrote that conduit sellers were “under the impression that . . . we will buy the loans that meet their guides,” and that “[w]hen we mention that we use our guides, this is a very difficult conversation at this stage in the process.” Ms. McNamara also wrote that “[h]andling such a high exception rate, finding compensations [sic] factors and discussing with sellers, puts extra strain on limited resources in underwriting.” SUF ¶ 98 (emphases in original). Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 40 of 76 33 As a result, Goldman did not know if loans acquired through the conduit actually complied with the Conduit Guidelines or not. Mr. O’Brien wrote that the conduit program was “buying loans with credit risk and features that are undisclosed to trading and finance - items that are permitted by seller guides (either outright or because not addressed).” SUF ¶ 105 (emphasis added). “Some of these [loan characteristics] may be acceptable with pricing adjustments, others may be beyond our risk tolerance and exit strategy,” Mr. O’Brien wrote. Id.; see also SUF ¶ 100 (April 2006 email from Mr. O’Brien explaining that “gap analyses” resulted in “a potential mis-match between [the Conduit Guidelines] and the loans we securitize under those guides” causing Goldman to “unwittingly buy[] loans outside [its] risk appetite”). There is no evidence that Goldman ever attempted to fix its diligence process to ensure that its gap analyses were completed or that it otherwise was able to determine which loans were acceptable under the Conduit Guidelines—which Goldman represented to the investing public were used to underwrite the loans. Instead, Goldman’s proposed solution was to approve all outstanding requests to use the sellers’ guidelines (SUF ¶ 106), and expand the Conduit Guidelines so as to “[i]nclude the most aggressive limits of our competitors and sellers guidelines” so that the “[l]oans securitized will more closely match the representation ‘underwritten to GS guides.’” SUF ¶¶ 107-08 (emphases in original). As of mid-2007, Goldman largely ceased performing “gap analyses” altogether. SUF ¶ 97. Large percentages of these improperly diligence conduit loans wound up in the Securitizations. According to internal Goldman Sachs memoranda, for example, about 44% of conduit loans in the GSAA 2006-11 Securitization were made pursuant to seller guidelines, about 48% in the GSAA 2006-8 Securitization, and about 43% in the GSAA 2006-4 Securitization. SUF ¶ 90. Yet the Prospectus Supplements for these three Securitizations each indicated that “[s]ubstantially all of the mortgage loans acquired by GSMC through its conduit program were acquired generally in accordance with the” Conduit Guidelines. SUF ¶¶ 116, 118, 121.23 There 23 As described in ¶¶ 113-30 of the SUF, conduit loans formed part of the relevant SLG for eighteen of the Securitizations: GSAA 2005-11, GSAA 2005-14, GSAA 2006-4, GSAA 2006-5, GSAA 2006-8, GSAA 2006-11, Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 41 of 76 34 is no evidence that the Diligence Group reviewed these loans—let alone all such loans in the SLGs—to see if they actually complied with the Conduit Guidelines. C. Goldman Used The Results Of Its Diligence To Increase Its Own Profits While Knowingly Communicating False Information Beyond simply conducting inadequate diligence on the WLTD Securitizations, the undisputed record also reveals that, in two distinct cases, Goldman used the results of its diligence to better its own position rather than to report accurate information to investors. 1. Goldman Determined That The Actual Values Of Many Of The Mortgaged Properties Were Lower Than Reported By Originators, But Used The Actual Values Only To Negotiate Lower Prices For Itself As noted in Part II.B, supra, Goldman conducted valuation diligence on the loans it acquired to evaluate whether the appraised values of the subject properties were reasonable and accurate. SUF ¶ 135. Goldman’s valuation diligence typically involved the use of three tools: automated valuation models (“AVMs”), broker price opinions (“BPOs”), and desk reviews. SUF ¶ 136. An AVM is a model employed to ascertain estimates, called “predicted values,” of the market value of a property. SUF ¶ 137. A BPO is an analysis of the value of the property based on three current listings and three past sales. SUF ¶ 141. A desk review is an independent review of the appraisal associated with the subject property. SUF ¶ 143. According to Goldman’s internal procedures, if, regardless of the LTV ratio, an AVM’s predicted value was within 15% of the appraised value reported by the seller, the Diligence Group would consider the appraised value of the subject property to be accurate and reasonable. SUF ¶ 138. But if an AVM’s predicted value differed from the seller’s reported value by more than 15%, or if the AVM did not return a value, the Diligence Group would then hire a third- party vendor to conduct a BPO. SUF ¶ 139. Occasionally, Goldman would use a desk review instead of a BPO, such as when “an agent couldn’t get out to the [subject] property.” SUF ¶ 143. GSAA 2007-6, GSAMP 2005-HE5, GSAMP 2006-HE3, GSAMP 2006-HE4, GSAMP 2006-HE5, GSAMP 2006-HE7, GSAMP 2006-HE8, GSAMP 2007-HE1, GSAMP 2007-HE2, GSR 2006-OA1, GSR 2007-OA1, and GSR 2007-OA2. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 42 of 76 35 If either the BPO or the desk review remained outside a 15% variance of the original appraisal value, Goldman would then engage in a review process, called a “reconciliation,” with the BPO vendor to determine a final value. SUF ¶ 145. Even if the AVM, BPO, and reconciliation all produced values that varied 15% or more from the seller’s appraised value, Goldman might still buy the loan if the seller provided reasons supporting the purchase during a “vetting session.” SUF ¶ 146. Through this reconciliation and vetting process, Goldman obtained “final values” of the subject property that it used to decide whether or not to buy the corresponding mortgage loan for eventual securitization. SUF ¶ 147; see also SUF ¶ 157 (Goldman’s expert cataloging final values for five acquisition pools).24 Even when Goldman rejected the appraised value reported by a seller in favor of a different “final value,” it still used the original appraisal value to generate the LTV ratios disclosed in the Prospectus Supplements. For example, Goldman’s expert acknowledged that the original appraisal value of Goldman Loan # , which Goldman placed in the relevant SLG of the GSAA 2006-11 Securitization, was $ , while the final reconciled value was $ . SUF ¶¶ 161-62. Goldman only reported the original value of $ for this loan in the GSAA 2006-11 loan tape, comparing that amount to the loan amount to produce a LTV ratio of 79.44% and a CLTV ratio of 100% (SUF ¶ 161); had Goldman used its final reconciled value, the LTV ratio of this loan would have been 92.74% and the CLTV ratio would have been 116.74%. SUF ¶ 163.25 Such changes in LTV ratios are especially significant because they 24 The five pools analyzed by Goldman’s expert were securitized in the following six Securitizations: FFML 2005- F11, GSAA 2006-11, GSAMP 2006-HE4, GSAMP FM2, GSAMP 2007-HE2, and GSAMP 2007-NC1. SUF ¶ 158. 25 Goldman Sachs’s expert provides numerous other examples in the five acquisition pools he examined in which Goldman failed to disclose the “final values” that were much lower—and therefore resulted in a higher LTV ratio— than what was reported in the Prospectus Supplements. For example, Goldman’s expert acknowledged that the original appraisal value of Goldman loan # —which became part of the GSAMP 2007-NC1 Securitization—was $ while the reconciled final value was $ SUF ¶ 159. Yet Goldman reported the original value of $ for this loan in its loan tape for the Securitization, comparing that amount to the loan balance of $ to obtain an original LTV of 80.18%, which was then used to produce the LTV ratio data disclosed in the Prospectus Supplement. SUF ¶ 160. Had Goldman used its “final value,” the LTV ratio for this loan would have been 84.18%. Similarly, Goldman’s expert opined that the original appraisal value of loan # which Goldman placed in the relevant SLG of the GSAMP 2007-FM2 Securitization, was $ , while the final reconciled value was $ SUF ¶¶ 164, 166. Goldman continued to only report the original value of $ in the GSAMP 2007-FM2 loan tape, which produced an LTV ratio of 90% (SUF ¶ 164); had Goldman used the final Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 43 of 76 36 increase the ratios from less than 80% to greater than 80%, which Goldman warned investors was a material change. SUF ¶ 154 (“Mortgage loans with higher original [LTV] ratios or [CLTV] ratios may present a greater risk of loss than mortgage loans with original [LTV] ratios or [CLTV] ratios of 80% or below.”)). There is no evidence that Goldman ever disclosed the “final values” it had uncovered during its diligence process to investors. There is evidence, however, that Goldman used its internal “final values” to negotiate lower prices for the loans it bought. For example, Goldman conducted additional valuation diligence on 243 of the 535 loans in RFC 5Y MAY3007 FLX pool—discussed in Parts II.A.1 and II.B.1.c, supra—obtaining “final values” for 169 of those loans. SUF ¶ 151. Goldman rejected (or “kicked”) the loans based on this valuation diligence, but then bought 11 of the loans, using the “final values” to negotiate a lower purchase price. SUF ¶¶ 152-53 (“Please note that there are 11 loans that have been re-priced with regard to accepting some of the values that were originally kicked.”).26 Nor was this an isolated practice: Goldman routinely reserved the right to renegotiate prices with originators based on its valuation diligence. SUF ¶¶ 149-50. None of this re-pricing was ever disclosed to investors.27 2. Goldman’s Own Originators Informed It Of The Sub-Prime Crisis In December 2006, But Goldman's Only Response Was To Begin Shorting The Market Beginning in 2005, Goldman took an ownership stake in two subprime originators, Lownhome and Senderra, with Goldman fully acquiring Senderra in 2007. SUF ¶ 249. Goldman used these originators to gather intelligence about the state of the market, “what loans reconciled value it possessed, the LTV ratio of this loan would have been 99.76% (SUF ¶ 167). Data produced by Goldman’s expert show that this process repeated, for just the five acquisition pools (which were fed into six Securitizations) that he examined, at least 47 times. Cipione Decl. ¶ 76; Cipione Decl. Ex. 17. 26 Nine of these eleven loans were in the relevant SLG for the GSR 2007-OA2 Securitization. SUF ¶ 156. 27 Goldman’s valuation diligence was also flawed because Goldman failed to follow its own policy, which directed it to obtain a BPO when either the AVM did not return a value or when the AVM analysis indicated a subject property had an out of tolerance original appraised value. SUF ¶ 140. For loans in the SLGs at issue, Goldman did not obtain a BPO on 20.90% of the loans where no AVM value was received. Cipione Decl. ¶ 78. Among the five pools that Goldman’s expert analyzed, Goldman did not obtain a BPO on 32.67% of loans in the SLGs after no AVM value was received. Cipione Decl. ¶ 78. And while Goldman’s valuation policy directed that it obtain a value reconciliation from a third-party diligence firm when values were indicated as out of tolerance by a BPO, Goldman failed to obtain such a reconciliation on 22.39% of such loans in the SLGs. Cipione Decl. ¶ 80. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 44 of 76 37 competitors were willing to purchase,” and to see those competitors’ underwriting standards. SUF ¶¶ 250-51. On December 10, 2006, Senderra’s CEO, Brad Bradley, sent an email to Kevin Gasvoda that highlighted “the dramatic shifts and disruption in the” subprime mortgage industry. SUF ¶ 252. Bradley informed Gasvoda that “credit quality has risen to become the major crisis in the non-prime industry. We are seeing unprecedented defaults and fraud in the market…. [including] inflated appraisals, inflated income and occupancy fraud.” Id. (emphasis added). Bradley wrote that “[a]ll in all, it is a perfect storm, the wrong way, for the origination market right now.” Id. Mr. Gasvoda forwarded this report to several Goldman executives and traders, including Daniel Sparks and Michelle Gill, informing them that this proprietary report was “INTERNAL ONLY.” Id. (capitalization in original). Four days later, another Senderra senior executive delivered a follow-up email to underscore what Mr. Bradley had written regarding the state of the subprime mortgage market. Senderra informed Mr. Gasvoda that “[w]e know that many of our competitors have not taken the necessary credit measures and will pay the price for this. Less production = less fees …. To add insult to injury we now are faced with irrational competitors chasing volume to stay in business, thus they have decreased rates on any loan which smells of quality.” SUF ¶ 253. A day later, Mr. Gasvoda forwarded this email to the same Goldman executives and traders, this time informing them “do not distribute” but that the email was “[v]ery telling.” Id. There is no evidence that Goldman took any steps after receiving these emails to increase its diligence or to otherwise account for the “perfect storm” that was occurring in the origination market. But on December 14, 2006, four days after Mr. Sparks was forwarded Senderra’s alert of “unprecedented defaults and fraud in the market” (SUF ¶ 252), Mr. Sparks met with Goldman Sachs’s senior executives—including David Viniar, Goldman Sachs’s CFO (SUF ¶ 31)— to begin discussing how Goldman could “[r]educe exposure” in the subprime market and “be in [a] position to take advantage” of the “even greater distress” he and the executives knew was occurring. SUF ¶¶ 254-55 (emphasis added). On that same day, Mr. Gasvoda instructed his staff Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 45 of 76 38 to begin selling Goldman’s RMBS inventory “even if you have to take a small loss” because “[t]here will be big opportunities the next several months.” SUF ¶ 256; see also Permanent Subcommittee on Investigations, United States Senate, Wall Street and the Financial Crisis: Anatomy of a Financial Collapse 382-86, 404-08 & nn.1556, 1651 (Apr. 13, 2011) (describing same and recounting how “[b]eginning in December 2006 and continuing through 2007, Goldman … built and profited from large net short positions in mortgage related securities”). Mr. Sparks “c[ould not] over state the importance to the business of selling [Goldman’s] positions and new issues,” (id. at 543 & n.2372), such that by March 2007 Goldman had “effectively halt[ed] new purchases of sub-prime loan pools” (SUF ¶ 257). Goldman did not stop securitizing subprime loans, however; to the contrary, it continued selling Certificates to the GSEs from seven subprime Securitizations.28 By September 2007, Goldman was publicly bragging that any “[s]ignificant losses on non-prime loans and securities were more than offset by gains on short mortgage positions.” SUF ¶ 258. III. GOLDMAN’S DILIGENCE FOR THIRD-PARTY SECURITIZATIONS WAS INSUFFICIENT In the Third-Party Securitizations, Goldman did not acquire or securitize the underlying loans. Instead, it underwrote the Securitizations and conducted diligence on the loans proposed for inclusion in the Securitization by the sponsor. See, e.g., SUF ¶ 338. Despite this difference, Goldman’s processes for conducting diligence on the Third-Party Securitizations were just as flawed as the processes for the WLTD Securitizations. For example, Goldman did not verify the representations in the Prospectus Supplements (SUF ¶¶ 360-64, 366-68), did not trust its third- party diligence vendors (SUF ¶ 321-22, 324, 326-28, 331-34), did not perform any diligence on loans outside of its samples (SUF ¶ 206), and did not upsize its samples for any of the Third- Party Securitizations (SUF ¶ 348). Moreover, the sampling spreadsheets Goldman produced for all four of the Third-Party Securitizations show no evidence that Goldman ever conducted any 28 Those seven Securitizations are: GSAMP 2006-FM3, GSAMP 2007-HE2, GSAMP 2006-HE8, GSAMP 2007- FM1, GSAMP 2007-FM2, GSAMP 2007-HE1, and GSAMP 2007-NC1. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 46 of 76 39 random sampling. SUF ¶ 350. There is likewise no evidence that Goldman ever extrapolated its results to determine what its limited analysis said about the rest of the SLG.s. Goldman’s procedures for diligence on all of the Third-Party Securitizations were in fact often worse than those for diligence conducted on the WLTD Securitizations. Goldman’s policies dictated that the Diligence Group sample only 10-15% of the loans in Third-Party Securitizations compared to 25-50% for subprime and Alt-A loans acquired for the WLTD Securitizations that would be held on Goldman Sachs’s own books. SUF ¶ 346. Goldman’s employees justified this discrepancy based on the belief that because Goldman was not buying the underlying loans, other companies determined the level of diligence that Goldman conducted. As one member of the Finance Group explained, Id. (emphases added). Goldman’ s own employees recognized the inadequacy of such non-random sampling in contemporaneous documents when reviewing its securitization of loans from RFC. “This is INSANE. I can’t believe this,” wrote Jessica Feingold, a trader, in reviewing a comparison of WLTD and third-party sample sizes. SUF ¶ 345 (capitalization in original). For two loan reviews done around the same time in 2007, Feingold noted, Goldman’s samples when buying loans for its own deals were two to three times larger than when Goldman underwrote third-party deals, and Goldman ran four to five as many BPOs on its own deals. Id. “It[’]s been pathetic” agreed Clayton DeGiacinto, a trader, in response. SUF ¶ 348. He had been pushing for Goldman to “amend procedure” and perform the same level of diligence for third-party deals as Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 47 of 76 40 for WLTD deals since 2005. See SUF ¶ 342. “Seriously,” Ms. Feingold wrote back. “I CANNOT believe it.” SUF ¶ 345 (capitalization in original). In practice, Goldman’s diligence was even worse than Goldman’s own “pathetic” policies required. For example, according to data produced by Goldman’s expert, Goldman never conducted more than 13% diligence on any of the Third-Party Securitizations, and it reviewed just 1.55% and 4.27% of the relevant SLGs in the FHLT 2006-E and ACCR 2005-4 Securitizations, respectively. Cipione Decl. ¶¶ 19-21. Similarly, for the INDX 2005-AR18 Securitization, Goldman conducted hardly any valuation diligence at all, with Goldman failing to run any AVMs and ordering BPOs on just 0.9% of the SLG. SUF ¶ 357; see also Expert Report of Charles Grice dated March 27, 2014, Ex. 13C (same). Goldman likewise waived in substantially higher numbers of loans its own vendors marked as “EV3” as compared to the WLTD Securitizations. For the ACCR 2005-4 Securitization, Goldman’s diligence vendor initially marked 47 of the loans it reviewed as EV3s, but Goldman securitized 39 loans as EV3s—even though Goldman agreed that those loans deserved EV3 grades. Cipione Decl. ¶ 48. For the FHLT 2006-E Securitization, Goldman’s diligence vendor initially marked 11 of the loans as EV3s (Cipione Decl. ¶ 47), but Goldman nonetheless securitized 10 loans with final grades of EV3, even though Goldman agreed those loans too were EV3s (Cipione Decl. ¶ 47), and thus not suitable for securitization (see SUF ¶ 302). ARGUMENT On a motion for summary judgment, “[t]he moving party bears the burden of demonstrating the absence of a material factual question,” In re WorldCom, Inc. Sec. Litig., 346 F. Supp. 2d 628, 655 (S.D.N.Y. 2004), and, as such, has “the initial responsibility of identifying those portions of the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, which it believes demonstrates the absence of a genuine issue of material fact,” id. (ellipsis and internal quotation marks omitted) (quoting Celotex Corp. v. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 48 of 76 41 Catrett, 477 U.S. 317, 323 (1986)). “‘Only disputes over facts that might affect the outcome of the suit under the governing law will properly preclude the entry of summary judgment.’” Golden Pac. Bancorp v. FDIC, 375 F.3d 196, 200 (2d Cir. 2004) (quoting Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986)). “Even if the parties dispute material facts, summary judgment will be granted unless the dispute is ‘genuine,’ i.e., unless ‘there is sufficient evidence favoring the nonmoving party for a jury to return a verdict for that party.’” Id. (quoting Anderson, 477 U.S. at 249). “When the moving party has asserted facts showing that it is entitled to summary judgment, the opposing party must ‘set forth specific facts showing that there is a genuine issue for trial,’ and cannot rest on the ‘mere allegations or denials’ of the movant’s pleading.” In re WorldCom, 346 F. Supp. 2d at 655-56 (quoting Fed. R. Civ. P. 56(e)). Where, as here, a plaintiff moves for summary judgment on a defendant’s affirmative defense, “[t]he showing required of Plaintiffs in their challenge to the legal sufficiency of [defendants’] asserted due diligence defense … is significantly lessened, for the defendants must ‘sustain the burden of proof’ as to their due diligence under the standards established by” the relevant statute. In re Livent, Inc. Noteholder Sec. Litig., 355 F. Supp. 2d 722, 729 (S.D.N.Y. 2005) (quoting Herman & MacLean v. Huddleston, 459 U.S. 375, 382 (1983)). “‘Where a plaintiff uses a summary judgment motion, in part, to challenge the legal sufficiency of an affirmative defense—on which the defendant bears the burden of proof at trial—a plaintiff may satisfy its Rule 56 burden by showing that there is an absence of evidence to support an essential element of the non-moving party’s case.’” Id. (quoting FDIC v. Giammettei, 34 F.3d 51, 54 (2d Cir. 1994)). Thus, in order to defeat summary judgment on its affirmative defense of due diligence, Goldman Sachs must “come forward with evidence supporting each essential element of [its] defense,” id., for each misrepresentation at issue, see In re WorldCom, 346 F. Supp. 2d at 680-82 (reasoning that the absence of a genuine issue of material fact as to an underwriter’s due diligence on one misrepresentation had no effect on the presence of a genuine issue on other misrepresentations). Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 49 of 76 42 I. GSMSC HAS NO DUE DILIGENCE DEFENSE AS A MATTER OF LAW GSMSC, which is liable for the misstatements and omissions in the Prospectus Supplements for 35 of the 36 WLTD Securitizations, has no due diligence defense. Goldman Sachs admits that GSMSC was the depositor for these 35 Securitizations. SUF ¶ 8. A depositor is, by both statutory and regulatory definition, an issuer for purposes of the Securities Act. See 15 U.S.C. § 77b(a)(4) (defining “issuer” to mean “the person or persons performing the acts and assuming the duties of depositor or manager pursuant to the provisions of the trust or other agreement or instrument under which such securities are issued”); 17 C.F.R. § 230.191(a) (“The depositor for the asset-backed securities acting solely in its capacity as depositor to the issuing entity is the ‘issuer’ for purposes of the asset-backed securities of that issuing entity.”); FHFA v. UBS Americas, Inc., 858 F. Supp. 2d 306, 334 (S.D.N.Y. 2012) (concluding that depositors are “issuers” for Securities Act purposes). “[T]he liability of issuers under Section 11 is ‘virtually absolute,’” id. at 329 (quoting In re Morgan Stanley Info. Fund Sec. Litig., 592 F.3d 347, 359 (2d Cir. 2010)), such that, under the plain language of 15 U.S.C. § 77k(b), an issuer has no due diligence defense, UBS, 858 F. Supp. 2d at 329. See also Herman & MacLean, 459 U.S. at 382 (issuers have no due diligence defense but instead face “virtually absolute” liability “even for innocent misstatements”); Ernst & Ernst v. Hochfelder, 425 U.S. 185, 208 (1976) (“[T]he issuer of the securities is held absolutely liable for any damages resulting from such misstatement or omission” and does not have a “due diligence” defense) (internal quotation marks omitted); Chris-Craft Indus., 480 F.2d at 370 (“A ‘due diligence’ defense to such a suit is available to all but the issuer.”). For this reason alone, there is no genuine dispute of material fact that GSMSC has absolute liability for the 35 Securitizations for which it was the issuer. II. GOLDMAN’S DILIGENCE WAS INADEQUATE AS A MATTER OF LAW There is also no genuine dispute that Goldman failed to conduct diligence adequate to support its “due diligence” affirmative defense under Section 11. While it was supposed to “‘exercise a high degree of care in investigation and independent verification of the [issuer’s] representations,’” In re WorldCom, 346 F. Supp. 2d at 662 (quoting Feit, 332 F. Supp. at 582), Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 50 of 76 43 Goldman abdicated that responsibility, violating both the plain language of Section 11 and the well-established requirement that an underwriter “must play devil’s advocate,” id. at 675 (internal quotation marks omitted) (quoting Feit, 332 F. Supp. at 582). A. As The Underwriter, Goldman Faces A Near-Absolute Standard Of Liability For Its Diligence Defense Because It Was Completely Intertwined With the Issuer, GSMSC Goldman Sachs’s underwriter, Goldman, faces an especially stringent burden of proof for its due diligence defense because it was intertwined with Goldman Sachs’s issuer, GSMSC. To defeat summary judgment, Goldman must create a genuine issue of material fact that shows both “[1] that [it] conducted a reasonable investigation and [2] that after such an investigation that [it] had reasonable ground to believe that the statements in the Registration Statements … were true.” In re WorldCom, 346 F. Supp. 2d at 682. The sine qua non of an underwriter’s diligence is “a high degree of care in investigation and independent verification of the company’s representations,” which requires a “thorough or searching inquiry.” Id. at 676, 678 (emphasis added) (internal quotation marks omitted). Even for a truly independent underwriter, “[t]he burden of proof [Goldman] must satisfy on summary judgment … is a heavy one.” In re Livent, 355 F. Supp. 2d at 733. An underwriter’s independence from an issuer is a core requirement of the Securities Act’s basic goal of “promot[ing] ethical standards of honesty and fair dealing,” Ernst & Ernst, 425 U.S. at 195. For decades, courts have made clear that “[t]he reasonable investigation- verification-requirement is simply one means of promoting the full disclosure policy of Section 11,” a policy that turns on “the importance of the role played by each participant in the scheme of distribution.” Feit, 332 F. Supp. at 577. The underwriter is supposed to be “the first line of defense with respect to material misrepresentations and omissions in registration statements.” In re WorldCom, 346 F. Supp. 2d at 662 (internal quotation marks omitted). Consequently, an underwriter may fulfill its statutorily-prescribed duties only when it “assume[s] an opposing posture with respect to [the issuer’s] management.” Feit, 332 F. Supp. at 581 (emphasis added); see also Escott v. BarChris Constr. Corp., 283 F. Supp. 643, 696 (S.D.N.Y. 1968) (“In a sense, Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 51 of 76 44 the positions of the underwriter and the company’s officers are adverse.”). According to Judge Weinstein, “[s]uch adversity is required since the underwriter is the only participant in the registration process who … is able to make the kind of investigation which will protect the purchasing public.” Feit, 332 F. Supp. at 581 (emphasis added); cf. Glassman v. Computervision Corp., 90 F.3d 617, 624 (1st Cir. 1996) (noting that financial ties between underwriter and issuer “arguably gave that underwriter a reason to inflate the offering prices”). As a result, courts recognize that the closer the ties between an issuer and underwriter, the greater the underwriter’s burden of proving a due diligence defense. E.g., In re WorldCom, 346 F. Supp. 2d at 675 (“[W]hat constitutes ‘reasonable investigation’ and a ‘reasonable ground to believe’ will vary with the degree of involvement of the [underwriter], [its] expertise, and [its] access to the pertinent information and data.”) (quoting Feit, 332 F. Supp. at 577); Univ. Hill Found. v. Goldman, Sachs & Co., 422 F. Supp. 879, 900 (S.D.N.Y. 1976) (assessing Goldman Sachs’s diligence and holding that, “[b]ased on the role it played in marketing Penn Central notes, Goldman, Sachs’[s] credit investigation must be judged by a fairly rigorous standard”); see also 17 C.F.R. § 230.176(g) (whether an underwriter’s investigation or ground for belief is reasonable turns on “the type of underwriting arrangement, the role of the particular person as an underwriter and the availability of information with respect to the registrant”). Thus, the underwriter in Feit was able to “barely” meet its due diligence defense only because of “the underwriter’s independence from the issuer.” In re WorldCom, 346 F. Supp. 2d at 675. In the new world of RMBS transactions—where issuers were alter egos of underwriters— underwriters will almost never be able to meet their burden of proof for a due diligence defense. A close analogy is the showing necessary for a director to prove a due diligence defense; in that context, the seminal decision in “BarChris created ‘a sliding scale of liability’ among directors, ‘drawing a distinction between insiders and outsiders.’” In re WorldCom, Inc. Sec. Litig., 2005 WL 638268, at *9 (S.D.N.Y. Mar. 21, 2005) (brackets omitted) (quoting 1 Thomas Lee Hazen, Law of Sec. Reg. § 7.4[2][A][1] (4th ed. 2002)). Inside directors, who have “intimate knowledge of corporate affairs and of the particular transactions[,] will be expected to make a Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 52 of 76 45 more complete investigation and have more extensive knowledge of facts supporting or contradicting inclusions in the registration statements than outside directors.” Id. (quoting Feit, 332 F. Supp. at 578); see also Circumstances Affecting the Determination of What Constitutes Reasonable Investigation and Reasonable Grounds for Belief Under Section 11 of the Securities Act, SEC Release No. 6335, 1981 WL 31062, at *12 (Aug. 6, 1981) (“SEC Rel. 6335”) (“With greater knowledge, the underwriter will be better prepared to question incomplete explanations, descriptions or reasoning and generally will be more sensitive to detecting and assessing material developments.”). Where a defendant is so heavily intertwined with the issuer, the defendant’s burden of proof is so “stringent” that “liability will lie in practically all cases of misrepresentation. [The defendant’s] liability approaches that of the issuer as guarantor of the accuracy of the prospectus.” Feit, 332 F. Supp. at 578.29 Here, Goldman’s liability is nearly absolute—it approaches that of the issuer GSMSC because, as a practical matter, Goldman was GSMSC. GSMSC had no office space, did not maintain its own books and records, had no employees, and was controlled entirely by employees of Goldman, who selected the loans that GSMSC would securitize. See Factual 29 Academic literature has reached the same conclusion. In a recent paper, Columbia Law Professor Merritt B. Fox argues that “[i]n the RMBS situation where the organization [of an RMBS] is undertaken within the same organization as the arranger and underwriter functions, this concern [over inadequate due diligence] argues in favor of absolute strict liability on the arranger and, if separate, on the much more deep pocketed underwriter entity as well.” Merritt B. Fox, Due Diligence with Residential Mortgage Backed Securities 57 (Columbia Law & Economics Working Paper No. 462, November 25, 2013), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2359679. In the RMBS context, “[t]he formal issuer – the S[pecial] P[urpose] V[ehicle] – has no operating personnel, makes no ongoing decisions, and is typically a creature of the arranger” and “is set up in a way that guarantees that in the situation that Section 11 addresses – the prospective or actual failure of a security not to make its promised payment combined with a material misstatement or omission in the registration statement – there will be no assets left to pay any resulting issuer liabilities.” Id. at 51. Rather, “[i]t is the corporate organization undertaking the arranging and underwriting functions (the arranger/underwriter organization) that makes all the decisions comparable to those of an issuer in an ordinary corporate offering.” Id. at 52. Because of the “differences between ordinary corporate offerings and RMBS ones in terms of the incentives, absent regulation, of the persons responsible for the offering disclosures to make misstatements and material omissions,” id. at 53, Professor Fox argues for strict liability for affiliated underwriter entities in the RMBS context, where “the question of what investigation an arranger actually undertook – its review of loan documentation and its testing of the accuracy what the documentation reveals – is more objectively determinable and subject to evaluation,” id. at 56. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 53 of 76 46 Background Part I, supra.30 Goldman therefore was in no position to “play devil’s advocate” against itself, or to conduct the “independent verification of [GSMSC’s] representations” required of underwriters. In re WorldCom, 346 F. Supp. 2d at 675-76 (quoting Feit, 332 F. Supp. at 582). Under the “sliding scale of liability,” Goldman’s burden of proof is therefore akin to that of an inside director with “intimate knowledge of corporate affairs and of the particular transactions[,]” In re WorldCom, 2005 WL 638268, at *9 (quoting Feit, 332 F. Supp. at 578) (internal quotation marks omitted), and Goldman’s “liability will lie in practically all cases of misrepresentation,” Feit, 332 F. Supp. at 578—including this one. B. Goldman Did Not Conduct A Reasonable Investigation At The Time The Prospectus Supplements Became Effective Whether under the near-absolute liability standard or otherwise, Goldman cannot meet its due diligence defense for any of the 36 WLTD Securitizations because there is no genuine issue of material fact that Goldman failed to investigate whether statements in the registration statement were true or whether there were omissions of material facts “at the time such part of the registration statement became effective.” 15 U.S.C. § 77k(b)(3)(A). Courts have long held that “failure by the underwriters … to consider new information up to the effective date of an offering would almost certainly constitute a lack of due diligence.” In re WorldCom, 346 F. Supp. 2d at 677 (quoting Glassman, 90 F.3d at 629); see also In re Software Toolworks Inc., 50 F.3d 615, 625 n.2 (9th Cir. 1995) (denying summary judgment on due diligence defense to underwriter that failed to investigate intra-quarterly information available after the preliminary prospectus was filed but before the effective date of the offering). In BarChris, for example, the court found the underwriter had not established a due diligence defense because an investigation through March 1961 was insufficient when the registration statement became effective in mid- May 1961. 283 F. Supp. at 697. 30 In its own briefing to this Court, Goldman Sachs is unable to distinguish between its role as issuer and underwriter, for example claiming that “the GSEs were aware that PLS issuers generally reviewed only a sample of loans underlying the Securitizations in their credit and compliance diligence.” Knowledge Opp. 25 (emphasis added). Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 54 of 76 47 Here, the Prospectus Supplements became effective on “the date each prospectus supplement was filed,” FHFA v. UBS Americas, Inc., 2012 WL 2400263, at *5 (S.D.N.Y. June 26, 2012), and the Prospectus Supplements form part of the registration statements for purposes of Section 11, FHFA v. HSBC N. Am. Holdings Inc., --- F. Supp. 2d ----, 2013 WL 6480445, at *3 (S.D.N.Y. Dec. 10, 2013) (citing 17 C.F.R. § 230.430B(f)(1)). Therefore, Goldman was required to conduct a reasonable investigation as to the accuracy and completeness of the statements in the Prospectus Supplements up until they became effective. The undisputed facts, however, are that Goldman conducted loan-specific diligence only when it acquired loans, months before the Prospectus Supplements for the WLTD Securitizations became effective. Cipione Decl. ¶ 26; Cipione Decl. Ex. 11; SUF ¶ 176 (statement of Goldman’s counsel that “I can’t stress enough that the way that diligence was done in this business … was to do diligence at the time of buying the loans.”); see also SUF ¶ 187 (Goldman employees consistently testifying they could not recall conducting any diligence at the securitization stage). While this diligence was itself so shoddy that it presents an independent ground for granting FHFA’s motion—see Part II.D, infra—there is no evidence that Goldman requested any new information about Mortgage Loans that might have been available after Goldman bought them. Goldman’s failure to conduct diligence through the effective date is particularly striking in light of Goldman personnel’s understanding that such post-acquisition information—such as servicing records showing a borrower’s occupancy status or credit reports showing recent undisclosed debts—could call into question the accuracy of the Prospectus Supplements’ representations. SUF ¶¶ 187-88. Goldman even failed to consider post- acquisition information that was already in its possession—including whether the originators of the loans it was securitizing had been placed on Goldman’s “suspend” or “watch” lists after Goldman bought those loans. See Part II.B.2.c, supra. Instead, once Goldman had acquired a pool, it simply stopped investigating the credit characteristics of that pool altogether.31 31 Moreover, as in Part II.C, infra, Goldman cannot rely on the limited analysis run by accounting firms to ensure that the data presented to Goldman was accurately transposed to the Prospectus Supplements. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 55 of 76 48 Goldman thus failed to satisfy the plain language requirement of the due diligence defense that it conduct a reasonable investigation “at the time … the registration statement became effective,” 15 U.S.C. § 77k(b)(3)(A), and—under any standard—summary judgment should be granted to FHFA regarding the 36 WLTD Securitizations on this ground alone. C. Goldman Did Not Reasonably Investigate Whether the Statements in the Prospectus Supplements Were True Goldman’s failure to investigate the truth of the statements in the Prospectus Supplements is reason enough to grant FHFA’s motion regarding all of the Securitizations, both WLTD and Third-Party. The due diligence defense applies only where an underwriter has conducted a “reasonable investigation … that the statements therein were true and that there was no omission to state a material fact required to be stated therein or necessary to make the statements therein not misleading.” 15 U.S.C. § 77k(b)(3)(A) (emphases added). The statutory text therefore requires that the underwriter review the actual statements in the registration statement and, by extension, the Prospectus Supplements. See id. Courts in this district have been unflagging in enforcing the Securities Act’s plain language. In a recent case in this District, the court granted a plaintiff’s motion for summary judgment on the due diligence defense where, inter alia, the defendant directors did not review the registration statement at all, but instead “completely relied on … outside personnel to ensure the Registration Statement’s accuracy.” In re Livent, 355 F. Supp. 2d at 737. In contrast, in Feit, the underwriter was “just barely” able to take advantage of the due diligence defense where “representatives of [the issuer] and the [underwriter] reviewed the proposed registration statement line by line.” 332 F. Supp. at 582; see also In re WorldCom, 346 F. Supp. 2d at 675- 76 (“[r]ecent Section 11 case law … shows no signs of abandoning the early courts’ demand that underwriters employ ‘a high degree of care in investigation and independent verification of the company’s representations.’” (quoting Feit, 332 F. Supp. at 582)). Thus, to meet its burden under any standard, Goldman would have to present admissible evidence that, for each Securitization, it conducted a systematic review of the statements in the Prospectus Supplements; Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 56 of 76 49 given Goldman’s “intimate knowledge “of the “particular transactions” that led to the SLGs, Goldman must show that it made “a more complete investigation … of facts supporting or contradicting inclusions in the” Prospectus Supplements than a normal Securities Act defendant. Feit, 332 F. Supp. at 578. Goldman cannot make either showing for any Securitization, as it is undisputed that Goldman structured its diligence processes such that neither the Diligence Group nor the third- party firms Goldman retained ever investigated whether the statements in the Prospectus Supplements were true. See Part II.A.2, supra. There is therefore no genuine dispute that Goldman failed to investigate or have a reasonable ground to believe that “the statements” in the Prospectus Supplements were accurate as required by Section 11’s text, and summary judgment for FHFA is appropriate. Goldman cannot escape this result by pointing to the limited efforts of its Finance Group or external accounting firms. See Part II.A.2, supra. The longstanding rule is that, “[t]o effectuate the statute’s purpose, the phrase ‘reasonable investigation’ must be construed to require more effort on the part of the underwriters than the mere accurate reporting in the prospectus of ‘data presented’ to them.” In re WorldCom, 346 F. Supp. 2d at 675 (quoting BarChris, 283 F. Supp. at 697)). Here, the Finance Group conducted a limited review of the Prospectus Supplements’ descriptions of the Mortgage Loans, in which it relied on the diligence that Goldman’s Diligence Group had conducted on the acquisition pools—months earlier and on only subsets of subsets of loans that eventually wound up in the SLGs (SUF ¶¶ 364-68). Goldman’s accounting firms likewise only reviewed whether information in the loan tapes matched the information in the loan files, expressly disclaiming any responsibility for determining if that information was accurate. SUF ¶ 367. Both processes thus relied on “data presented” and provide no basis to deny summary judgment to FHFA on Goldman’s due diligence defense. In re WorldCom, 346 F. Supp. 2d at 675. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 57 of 76 50 D. Goldman’s Acquisition-Stage Diligence Process Was So Inadequate That Goldman Was Incapable of Performing A Reasonable Investigation Summary judgment for FHFA as to all 40 Securitizations is also appropriate because Goldman’s diligence process was flawed to the point that Goldman was incapable of performing a reasonable investigation. An underwriter’s duty to conduct a “reasonable investigation” into whether the statements in offering materials were true or materially incomplete, 15 U.S.C. § 77k(b)(3)(A), requires the underwriter to engage in a “thorough or searching inquiry” that is done “with systematic attention to detail and relationship.” In re WorldCom, 346 F. Supp. 2d at 678 (internal quotation mark omitted) (quoting and citing Webster’s New Int’l Dictionary of the English Language 1306 (2d ed. 1934) and Webster’s Third New Int’l Dictionary of the English Language 1189 (1993)); see also Taniguchi v. Kan Pacific Saipan, Ltd., 132 S. Ct. 1997, 2002 (2012) (“When a term goes undefined in a statute, we give the term its ordinary meaning.”). In underwriting the Securitizations, therefore, Goldman had to conduct a thorough, systematic inquiry to that would allow it to reasonably believe that the Prospectus Supplements’ descriptions of all of the thousands of securitized Mortgage Loans were true. And given Goldman’s deep knowledge of and control over GSMSC, it was required to conduct a “complete investigation and have more extensive knowledge of facts” in order to meet its due diligence defense; any failure to do so is fatal, because, under this near-absolute standard, “liability will lie in practically all cases of misrepresentation.” In re WorldCom, 2005 WL 638268, at * 9 (quoting Feit, 332 F. Supp. at 578) (internal quotation marks omitted). Goldman cannot meet this standard as a matter of law based on its ramshackle “diligence” system, which revealed nothing about most of the loans Goldman directed GSMSC to securitize. First, Goldman conducted no diligence at all on the vast majority of the loans in the SLGs, instead relying on samples drawn from loan pools it acquired from originators. See Part II.B.1.b, supra. But because Goldman often built SLGs from multiple acquisition pools, Goldman had no idea which loans in such SLGs had received diligence reviews. Id. Nor was Goldman’s MCC, the gatekeeping entity within Goldman that gave final approval to the Securitizations, ever presented with any information regarding how many diligenced loans were in the SLGs. SUF Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 58 of 76 51 ¶ 72. The end result was that the SLGs contained far fewer diligenced loans than required by Goldman’s internal policies for acquisition pools. See Cipione Decl. ¶ 16; SUF ¶ 211. This conduct is the opposite of the “complete investigation” Goldman was required to conduct. In re WorldCom, 2005 WL 638268, at * 9. Second, Goldman’s sampling process was hopelessly flawed. Beyond using a diligence database that Goldman knew was “incomplete and inaccurate,” see Part II.B.1.a, supra, Goldman routinely drew samples that were smaller than those required by its own internal policies, and then reduced its sample sizes even more to accommodate the originators from which it was buying the loans—over the protests of its own diligence personnel, see Part II.B.1.b, supra. Reducing its sample sizes was especially inappropriate in light of Goldman’s knowledge that originators had made the loans as part of an “originate-to-distribute” model, which incentivized originators to “lower[] loan underwriting standards in ways that investors may have [had] difficulty detecting,” Nat’l Credit Union Admin. Bd. v. Wachovia Capital Markets, LLC, 2014 WL 1795294, at *3 (S.D.N.Y. May 6, 2014) (quoting Financial Stability Oversight Council, Macroeconomics Effects of Risk Retention Requirements, at 3 (Jan. 2011)). Third, while its procedures required it to select a portion of the sampled loans “randomly,” it is undisputed that Goldman failed “[t]o ensure that the sample selected for each [acquisition pool was] a random and unbiased representation of the population from which it was selected.” FHFA v. JPMorgan Chase & Co., 2012 WL 6000885, at *6 (S.D.N.Y. Dec. 3, 2012). Such haphazard sampling vitiates as a matter of law any argument that its samples were reliable or could reveal any information about the rest of the acquisition pool—to say nothing of the SLG. See Part II.B.1.d, supra. Nor can Goldman rely on its policies as evidence of the reliability of its sampling, as the undisputed record shows that Goldman routinely conducted less diligence than its own policies required at the request of loan originators, which Goldman viewed as its “clients.” See Part II.B.1.c, supra. And while Goldman’s procedures required it to select the remainder of its samples based on “adverse” criteria, it had no systematic way of doing so— choosing instead “just the loans that kind of stuck out,” in a subjective process that “had the Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 59 of 76 52 potential for bias and errors”—and even then not including all of these high-risk loans in its samples. See Part II.B.1.d, supra. Nor did Goldman ever “look deeper and question more” when confronted with evidence of potential credit concerns—loans outside of its samples that met its adverse credit criteria, or loans inside its samples graded as EV3s for credit reasons. See In re WorldCom, 346 F. Supp. 2d at 677. Fourth, Goldman then systematically ignored the results of its diligence reviews, never extrapolating them to the unsampled loans in its acquisition pools or in the SLGs, see Part II.B.1.e, supra, and never increasing its sample sizes to perform additional credit diligence—no matter what the results of its diligence review of the sampled loans showed, see Part II.B.2.a, supra, or what red flags Goldman faced, see Parts II.B.2.b & c, supra.32 By failing to ever determine what its unreliable and unrepresentative samples said about the pools of loans Goldman acquired (to say nothing of the SLGs the Prospectus Supplements actually discussed), Goldman’s diligence was unreasonable as a matter of law. E. Goldman Did Not Look Deeper Or Question More When Confronted With Numerous Red Flags At The Acquisition Stage Goldman’s due diligence defense also fails as a matter of law because Goldman did not “‘look deeper and question more’ when confronted with red flags,” as required by Section 11. In re WorldCom, 346 F. Supp. 2d at 677 (quoting In re Enron Corp. Sec., Derivative & ERISA Litig., 235 F. Supp. 2d 549, 707 (S.D. Tex. 2002)). Goldman should know this rule better than most: nearly 40 years ago, another court in this District concluded that the existence of “red flags” or “storm warnings” functioned “to render Goldman, Sachs’ normal procedures 32 Even if there were evidence that Goldman had extrapolated the results of its diligence review of the sample of an acquisition pool to the entire acquisition pool—and, as discussed in Part II.B.1.e, supra, there is none—Goldman cannot establish that its results could be reliably extrapolated to the SLG, because the SLG is only a subset of those pools. See, e.g., Eng’g Contractors Ass’n of S. Fla., Inc. v. Metro. Dade Cnty., 943 F. Supp. 1546, 1560 n.16 (S.D. Fla. 1996) (noting that where groups combined for statistical analysis are “more heterogeneous,” it is “less likely the … analysis will be reliable”), aff’d, 122 F.3d 895, 919 n.4 (11th Cir. 1997); Dukes v. Wal-Mart Stores, Inc., 603 F.3d 571, 637 n.12 (9th Cir. 2010) (en banc) (Ikuta, J., dissenting) (discussing same), majority decision rev’d, 131 S. Ct. 2541 (2011). As Goldman previously argued to this Court, “[e]ach [C]ertificate is a separate security, generally collateralized by a unique set of loans and governed by a unique set of disclosures about those loans.” Dkt. No. 135, at 11. As a matter of statistics, Goldman therefore could not simply assume that its diligence on several pools could provide any robust estimate about the “unique set of loans” in the SLG. Id. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 60 of 76 53 inadequate and to require more concrete verification.” Univ. Hill Found., 422 F. Supp. at 902 (when confronted with red flags, Goldman Sachs’s “exclusive reliance on publicly available data and the unverified representations of management [was] inadequate to meet its obligation to the investing public”). Under the stringent burden of proof Goldman must meet to prove its due diligence defense, any failure by Goldman to follow up on any red flag must result in summary judgment in FHFA’s favor. See In re WorldCom, 2005 WL 638268, at *9. There is no genuine issue of material fact preventing this outcome for either the WLTD or Third-Party Securitizations. First, Goldman’s use of its “incomplete and inaccurate” WHAM database to conduct due diligence, see Part II.B.1.a, supra, was itself a red flag that “Goldman, Sachs’ normal procedures [were] inadequate [and] … required more concrete verification.” Univ. Hill Found., 422 F. Supp. at 902. This is especially true for the WHAM database’s inability to display loans that Goldman had previously rejected. That Goldman had previously decided a loan was unsuitable for securitization is undoubtedly an “‘alert[] … that something is seriously wrong’” with a loan, In re WorldCom, 346 F. Supp. 2d at 673 (quoting LC Capital Partners, LP v. Frontier Ins. Grp., Inc., 318 F.3d 148, 155 (2d Cir. 2003)), yet there is no evidence that Goldman took any steps to further identify or investigate such red flags. Second, Goldman ignored persistent red flags created by its own diligence personnel throughout the relevant time period about the competence of Goldman’s third-party diligence vendors. See Part II.B.2.b.i, supra. Goldman’s blind faith in data presented by others is “not … sufficient to ward off liability,” In re WorldCom, 346 F. Supp. 2d at 672; rather, “more concrete verification” is required, Univ. Hill Found., 422 F. Supp. at 902. Despite chronic concerns about its vendors’ competence, prior to late 2007 Goldman took no action to confirm that loans graded by those vendors as suitable for purchase were, in fact, suitable for purchase. See Parts II.B.2.b.i & ii, supra. Nor is there any evidence to show that Goldman took steps to modify its oversight of its diligence firms or made any effort to change the diligence processes employed by these firms, despite Goldman’s complete control of the parameters and personnel these third-party Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 61 of 76 54 firms used. See Part II.B.2.b, supra. Far from conducting the “complete investigation” required by Goldman’s deep ties to GSMSC, see Feit, 332 F. Supp. at 578, Goldman did not so much as question the grades provided by its vendors for nearly 85% of the loans in the SLGs—over 22,000 of them—that its vendors reviewed See Part II.B.2.b.ii, supra. And when Goldman finally did put its vendors to the test in late 2007, all of the tested vendors produced error rates of over 45% (SUF ¶¶ 334-35), revealing a systemic problem of which Goldman was on notice throughout the three years that make up the relevant time. Goldman’s “blind reliance” on vendors that it did not trust to identify loans that it would securitize is insufficient as a matter of law to satisfy its due diligence defense. In re Livent, 355 F. Supp. 2d at 738. Third, Goldman ignored warning signs that loans outside of its samples could be as defective as loans graded as EV3s within its samples. In marked contrast to its hands-off approach to loans graded by its vendors as EV1s and EV2s, Goldman reviewed every loan that the vendors graded as an EV3, seeking to “overturn” those grades so that it could buy and securitize as many of those loans as possible. Goldman’s own reviews often showed that loans in its samples were unsuitable for purchase for credit reasons. Moreover, Goldman had “adverse” criteria that identified the loans it believed had the greatest credit risk, yet it did not include all loans that met these criteria within its samples. Despite such warnings that loans outside its samples could have the same potential credit issues as “adversely-selected” loans, or the same confirmed credit issues of EV3s within its samples, Goldman did not systematically conduct additional credit diligence outside of its samples. See Part II.B.2.a, supra. Goldman’s decision to ignore these red flags “rendered [its] ongoing … investigation unreasonable,” Univ. Hill Found., 422 F. Supp. at 902. Finally, Goldman securitized more than 5,800 loans by at least 20 separate originators that were on Goldman’s “watch” or “suspend” lists at the time of securitization—without conducting any diligence beyond the standard diligence it performed when buying the loans months before. See Part II.B.2.c, supra. The “watch” and “suspend” status of these originators were quintessential “warning signals to [an] underwriter that something more might be in order,” In re Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 62 of 76 55 WorldCom, 346 F. Supp. 2d at 677 (quoting Univ. Hill Found., 422 F. Supp. at 900) (internal quotation marks and brackets omitted), but Goldman did nothing. Consequently, for the seven Securitizations where Goldman securitized loans from originators on its “watch” or “suspend” lists,33 Goldman’s due diligence defense fails as a matter of law. F. Goldman Succumbed To Competitive Pressure In Securitizing Conduit Loans While Being Unable To Determine Whether Such Loans Had Been Underwritten To Goldman’s Guidelines Goldman’s due diligence defense further fails as a matter of law for the 18 Securitizations containing loans purchased through Goldman’s conduit program because Goldman failed to underwrite these loans against its conduit guidelines or perform the required “gap analysis” that compared its conduit guidelines to seller guidelines. An underwriter must take the time necessary conduct a reasonable investigation before associating itself with an offering, such that “an underwriter is ‘never compelled to proceed with an offering until he has accomplished his diligence.’” In re WorldCom, 346 F. Supp. 2d at 669 (emphasis in original) (quoting SEC Rel. 6335, 1981 WL 31062, at *10). The SEC has “‘expressly rejected the consideration of competitive timing and pressures when evaluating the reasonableness of an underwriter’s investigation.’” Id. at 670 (quoting The Regulation of Securities Offerings, 63 Fed. Reg. 67,174, 67,231 (Dec. 4, 1998)). Goldman nonetheless routinely securitized loans purchased through its conduit program without accomplishing its diligence in the form of a gap analysis, as Goldman’s Chief Underwriter admitted. See Part II.B.3, supra. Goldman was “not able to review [seller guidelines] in time for trade” because they were coming in “at an unmanageable rate.” Id. Rather than finish its diligence as required by the Securities Act, Goldman instead allowed competitive timing and pressures to win out, securitizing loans even though the Diligence Group was informing Goldman that it “cannot identify” whether a loan was consistent with Goldman’s “risk tolerance and exit strategy,” let alone the Conduit Guidelines Goldman represented would 33 These seven Securitizations are identified in n.20, supra. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 63 of 76 56 govern. Id. Nor did Goldman ever disclose in the Prospectus Supplements that the Mortgage Loans purchased through Goldman’s conduit program were underwritten to originators’ guidelines, rather than the Conduit Guidelines. E.g., SUF ¶¶ 116, 118, 121. By failing to conduct the gap analyses that were required by its own policies, there is no genuine dispute of material fact that Goldman cannot meet its burden of showing that it acted reasonably, or with the requisite “high degree of care,” in conducting diligence on the 18 WLTD Securitizations containing conduit loans.34 In re WorldCom, 346 F. Supp. 2d at 662 (quoting Feit, 332 F. Supp. at 582). G. Goldman Knew The Information In The Prospectus Supplements Was False And Deliberately Concealed The Truth Another independent reason to grant FHFA’s motion is Goldman’s undisputed failure to disclose to investors its belief that much of the information in the Prospectus Supplements was inaccurate. This was especially egregious with respect to Goldman’s knowledge that the LTV ratios stated in the Prospectus Supplements were based on inaccurate appraisal values. An underwriter cannot meet its due diligence defense when it “had access to all information that was available and deliberately chose to conceal the truth.” In re Software Toolworks, 50 F.3d at 625. That rule is at its apex here, where Goldman had the greatest access to the information about the SLGs given its complete control over GSMSC. See Feit, 332 F. Supp. at 578. It is undisputed that Goldman reviewed the appraised property values provided by the Mortgage Loans’ originators and, in many cases, determined that the “final” property values were lower than the originators’ values. See Part II.C.1, supra. Goldman used these “final values” to negotiate lower prices for itself, as principal, when buying certain loans from originators, but it reported LTV ratios to investors based on the originators’ stated values— which Goldman believed to be false. Id. Moreover, had Goldman used the “final values” it would have materially altered Prospectus Supplements descriptions of the Certificates’ risk, since many of the “final values” would have pushed the LTV ratios for numerous loans over 34 These 18 Securitizations are identified in n.23, supra. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 64 of 76 57 80%, a key threshold that Goldman itself recognized materially affected the quality of the collateral. Id. Goldman thus “had access to all information that was available and deliberately chose to conceal the truth,” and cannot prevail on its due diligence defense. In re Software Toolworks, 50 F.3d at 625. Similarly, Goldman had unique and confidential access to the dire warnings directly provided by its own affiliate and proprietary originators that the sub-prime mortgage market was experiencing “unprecedented defaults and fraud,” yet Goldman failed to respond, except to the extent in sought to profit through shorting the market. See Part II.C.2, supra. Vibrantly red flags such as these “trigger[] a duty on the part of defendants to scrutinize the loans included in their securitizations more closely,” UBS, 858 F. Supp. 2d at 321, but there is no evidence in the record that Goldman ever did so. Just as it was unreasonable for a defendant to “fail[] to inquire into issues of particular prominence in [their] own internal valuations … or in the financial press,” In re WorldCom, 346 F. Supp. 2d at 683, it was unreasonable for Goldman to have done nothing in response to the “[v]ery telling” warnings of “unprecedented defaults and fraud in the market” it received (SUF ¶ 253). Rather than look deeper or question more, Mr. Sparks—the head of Goldman’s Mortgage Department—instead began discussions with senior Goldman Sachs executives about how to “take advantage” of the market’s “even greater distress.” SUF ¶¶ 254- 55. H. Goldman’s Diligence on the Third-Party Securitizations Was Not Thorough Or Searching And Was Driven by the Issuers There is also no genuine dispute of material fact that Goldman did not conduct a reasonable investigation into the loans underlying the Third-Party Securitizations, let alone have a reasonable ground to believe that the representations regarding the loans in the Prospectus Supplement for each Securitization was accurate. See In re WorldCom, 346 F. Supp. 2d at 679. As with the WLTD Securitizations, for the Third-Party Securitizations there is no evidence that Goldman performed diligence designed to verify the accuracy of statements in the Prospectus Supplements, see Part II.A.2, supra, but there is evidence that Goldman used the same ad hoc Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 65 of 76 58 diligence process, see Part II.B.1, supra, untrusted vendors, see Part II.B.2.b.i, supra, and flawed sampling procedures, see Part II.B.1.d, supra. Nor did Goldman conduct any random sampling at all on the Third-Party Securitizations. SUF ¶ 350. For the reasons stated above, these failures render Goldman’s diligence on all of the Third-Party Securitizations unreasonable as a matter of law. Goldman’s diligence of the Third-Party Securitizations is additionally unreasonable because Goldman systematically conducted diligence on many fewer loans for third-party deals than it did for the WLTD Securitizations. See Part III, supra. For every securitization it underwrote, Goldman had to conduct a reasonable investigation. 15 U.S.C. § 77k(b)(3)(A). To be reasonable, Goldman’s investigation must be “that required of a prudent man in the management of his own property.” 15 U.S.C. § 77k(c). Under that standard, the fact that Goldman sampled even fewer loans for its Third-Party Securitizations as compared to the WLTD Securitizations means that its investigations were unreasonable as a matter of law, especially when considered against Goldman’s nuanced knowledge of the mortgage market, see Part II.C.2, supra. The admission by Goldman’s diligence personnel that it was “INSANE” for Goldman to have lower standards for third-party deals because other underwriters, and not Goldman, were “driving” the diligence process (SUF ¶¶ 345), also flunks Section 11’s requirement that Goldman independently verify the issuers’ representations. See In re Livent, 355 F. Supp. 2d at 738 (“[B]lind reliance on others does not substitute for the due diligence that defendants must demonstrate to avoid Section 11 liability.”); BarChris, 283 F. Supp. at 697 (“The other underwriters, who did nothing and relied solely on [the lead underwriter] and on the [lead underwriter’s] lawyers, are also bound by it.”). Goldman’s actual diligence practices on the Third-Party Securitizations bears out that Goldman’s investigation was unreasonable as a matter of law, with Goldman conducting credit diligence on less than five percent of the SLGs for two of the Third-Party Securitizations and valuation diligence on less than one percent of the SLG for another. See Part III, supra. Consequently, summary judgment for FHFA on Goldman’s due diligence defense is appropriate for the Third-Party Securitizations. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 66 of 76 59 III. GOLDMAN AND GSMSC DID NOT EXERCISE REASONABLE CARE AS A MATTER OF LAW Summary judgment for FHFA is also appropriate on GSMSC and Goldman’s “reasonable care” defenses under Section 12 of the Securities Act and the Virginia and D.C. Blue Sky laws, because neither Defendant analyzed the loans in each Securitization’s actual SLGs, reviewed the statements in the Prospectus Supplements, or did anything more than rely on the untrustworthy assurances of originators and third-party due diligence firms.35 While Section 12’s “reasonable care” defense is less “exacting” than Section 11’s “reasonable investigation” requirement, In re WorldCom, 346 F. Supp. 2d at 663, there is substantial overlap between the two: the Second Circuit has recognized that a defendant fails to exercise reasonable care where it fails to “conduct[] an ongoing investigation of [the issuer’s] financial condition.” Franklin Sav. Bank of N.Y. v. Levy, 551 F.2d 521, 527 (2d Cir. 1977); see In re Software Toolworks Inc., 50 F.3d at 621 (because Section 11 standard is “similar, if not identical” to Section 12 standard, “analysis of each on summary judgment is the same”). Consequently, a failure to conduct a “reasonable investigation” under Section 11 constitutes, in almost all cases, a lack of “reasonable care” under Section 12. See, e.g., Franklin Sav. Bank, 551 F.2d at 527 (underwriter “failed to exercise reasonable professional care in assembling and evaluating the financial data … no matter how honestly but mistakenly [underwriter’s belief in the accuracy of the financial data was] held”); Univ. Hill Found., 422 F. Supp. at 902 (defendant, treated as underwriter under Section 11, could not meet Section 12 “reasonable care” defense where its “exclusive reliance on publicly available data and the unverified representations of management [was] inadequate to meet [defendant’s] obligations to 35 Both the D.C. and the Virginia Blue Sky Acts contain language that is identical to Section 12’s “reasonable care” defense, see D.C. Code § 31-5606.05(a)(1)(B); Va. Code Ann. § 13.1-522(A)(ii), and this Court has held that, where the statutory language in the two Blue Sky Acts is identical to Section 12, “the D.C. and Virginia securities laws are generally interpreted in accordance,” FHFA v. Bank of America Corp., 2012 WL 6592251, at *7 n.8 (S.D.N.Y. Dec. 18, 2012). See also FHFA v. HSBC N. Am. Holdings Inc., --- F. Supp. 2d ----, 2013 WL 6588249, at *2, *4 (S.D.N.Y. Dec. 16, 2013) (recognizing that both Blue Sky Acts contain the identical “reasonable care” defense as in Section 12). The same analysis under Section 12 therefore applies to the Blue Sky Act claims. See MidAmerica Fed. Sav. & Loan Ass’n v. Shearson/Am. Express, Inc., 886 F.2d 1249, 1255 n.3 (10th Cir. 1989). This Court has also already ruled that neither GSMSC nor Goldman has a loss causation defense under either Blue Sky Act. HSBC N. Am. Holdings, 2013 WL 6588249, at *4. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 67 of 76 60 the investing public”); cf. In re Metlife Demutualization Litig., 262 F.R.D. 217, 235 (E.D.N.Y. 2009) (“A defendant may show ‘reasonable care’ [under Section 12] by introducing evidence that the allegedly fraudulent prospectus was the result of reasonable investigation.”). Under any conception of the “reasonable care” defense, there is no genuine issue of material fact that Goldman has none. First, Goldman’s “uniquely close” relationship with GSMSC (see Factual Background Part I , supra) means that its burden is “correspondingly far greater.” Univ. Hill Found, 422 F. Supp. at 901; see, e.g., Heffernan v. Pac. Dunlop GNB Corp., 965 F.2d 369, 373 (7th Cir. 1992) (“The defendant’s position gives content to the term ‘reasonable care.’ For instance, reasonable care for a director requires more than does reasonable care for an individual owning a few shares of stock with no other connection to the corporation.”); Junker v. Crory, 650 F.2d 1349, 1361 (5th Cir. 1981) (underwriter’s “familiar[ity] with the financial operations of” the issuers meant that “the exercise of reasonable care would have required him to seek an appraisal or at least some knowledgeable estimate of the [issuer’s] property in assessing its true value”). Goldman’s wholesale intertwining with GSMSC thus again places Goldman’s burden of proof for its reasonable care defense at its apex. Second, Goldman did not exercise “reasonable care” because its Diligence Group failed to conduct any investigation after Goldman acquired the loans that GSMSC would securitize. By so limiting its inquiry, Goldman did not avail itself of relevant information arising after it purchased loans but before it securitized them. See Part II.A.1, supra. “Its failure to inquire more closely into the basis for the statements [made by the originators] … rendered Goldman Sachs’[s] ongoing … investigation unreasonable, and its representation to the [GSEs] untrue within the meaning of § 12[a](2).” Univ. Hill Found., 422 F. Supp. at 902; see also Franklin Sav. Bank, 551 F.2d at 527 (underwriter must “conduct[] an ongoing investigation” to take advantage of reasonable care defense (emphasis added)); In re Software Toolworks, 50 F.3d at 625-26 & n.2 (failure to analyze intra-quarterly data prior to effective date vitiated underwriter’s Section 11 and 12 defenses). Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 68 of 76 61 Third, Goldman did not examine the Prospectus Supplements to determine whether they contained any “untrue statement of a material fact or omit[ted] to state a material fact necessary in order to make the statements … not misleading.” 15 U.S.C. § 77l(a)(2). Because Goldman never determined whether the statements in the Prospectus Supplements were accurate—as opposed to inadequately investigating individual acquisition pools prior to securitization (see Part II.A.2, supra)—Goldman cannot meet its burden of proof. Fourth, the limited inquiry Goldman conducted when acquiring loans is insufficient as a matter of law to qualify as reasonable care. There is no genuine issue of disputed fact that “Goldman, Sachs failed to exercise reasonable professional care in assembling and evaluating the … data” regarding the Mortgage Loans that it would describe—months later—in the Prospectus Supplements. Franklin Sav. Bank, 551 F.2d at 527. As discussed in Part II.B.1.b, supra, Goldman examined only subsets of the acquisition pools, which themselves were subsets of the SLGs, leaving entire swaths of loans in the SLGs unexamined: Goldman’s expert admits there is no data in the record showing that Goldman conducted diligence on over 9,400 Mortgage Loans (Cipione Decl. ¶ 9), Goldman failed to have its third-party diligence firms review over 66% of the Mortgage Loans for credit or compliance issues (Cipione Decl. ¶¶ 16, 39), and there is no evidence that Goldman ever questioned the grades its vendors gave to over 85% of the loans that Goldman instructed its vendors to examine (Cipione Decl. ¶ 41, SUF ¶¶ 307-08). As even Goldman’s own counsel admits that its sampling procedures were not statistically reliable (SUF ¶ 241), there is no evidence in the record showing that Goldman did, or even could, evaluate the acquisition pools, to say nothing of the SLGs. Fifth, Goldman ignored numerous red flags that should have stripped it of its confidence in the data it was getting from originators and its third-party diligence firms, see Part II.B.2, supra, and its failure to investigate these storm warnings vitiates Goldman reasonable care defense just as surely as it does Goldman’s due diligence defense. See Franklin Sav. Bank, 551 F.2d at 527 (unreasonable for underwriter to vouch for quality of securities “particularly in view of the worsening condition of [the issuer]”); In re Software Toolworks, 50 F.3d at 626 (no Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 69 of 76 62 summary judgment on defendants’ “due diligence” and “reasonable care” defenses when underwriters “did little more than rely on [the issuer’s] assurances that the transactions were legitimate” despite red flags that they were not). In sum, whatever differences there may be between the “reasonable care” and “due diligence” defenses, “the analysis of each on summary judgment is the same,” id. at 621, and Goldman is not entitled to either. Sixth, there is no genuine issue of material fact that Goldman failed to use “reasonable care” by using seller guidelines to conduct diligence on the loans it acquired through its conduit program while representing in the Prospectus Supplements that the loans were underwritten to the Conduit Guidelines—and then not conducting the “gap analyses” required by its own internal procedures to see if the two sets of guidelines matched up. See Part II.B.3, supra. Consequently, Goldman “[can]not shoulder[ ] its burden of proving that it was reasonable for it to have determined, on [the effective date], that the [issuer’s] notes were credit worthy,” because “[i]n fact the record shows that it assumed no burden at all.” Alton Box Bd. Co. v. Goldman, Sachs & Co., 560 F.2d 916, 924 (10th Cir. 1977). Seventh, Goldman did not exercise “reasonable care” when it listed in the Prospectus Supplements LTV ratios that Goldman knew, based on its own investigation, were materially overstated. See Part II.C.1, supra. As with Section 11, an underwriter loses its reasonable care defense under Section 12 where it has “access to all information that was available and deliberately cho[o]se[s] to conceal the truth.” In re Software Toolworks, 50 F.3d at 625; see, e.g., Wigand v. Flo-Tek, Inc., 609 F.2d 1028, 1034 (2d Cir. 1979) (seller failed to exercise reasonable care when seller “knew that the company had lost money in 1970 and was losing money in 1971, and failed to tell [the purchaser] this”). Goldman’s use of its “final values” to negotiate lower prices with originators demonstrated its knowledge that the original stated values—and thus the LTVs in Prospectus Supplements based on those values—were false. Finally, GSMSC is also not entitled to a reasonable care defense. There is no evidence that GSMSC was even capable of exercising reasonable care, since it “was not an operating company.” SUF ¶ 9. To the extent that GSMSC was capable of any actions, because Goldman’s Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 70 of 76 63 employees controlled GSMSC (SUF ¶¶ 10-13) Goldman’s failure to conduct reasonable care is fully attributable to GSMSC. Even if GSMSC were somehow able to take actions independent of Goldman, there is no evidence that GSMSC ever asked any questions or inquired into any aspect of the quality of the loans it was securitizing and Goldman was selling. Cf. Demarco v. Edens, 390 F.2d 836, 842-43 (2d Cir. 1968) (finding reasonable care where issuer conducted in- depth inquiry into quality of underwriter). GSMSC therefore is also fully liable under Section 12 and the Virginia and D.C. Blue Sky Acts. CONCLUSION For the reasons set forth above, FHFA respectfully requests that the Court grant its motion for partial summary judgment on Defendants’ Due Diligence and Reasonable Care defenses. Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 71 of 76 64 Respectfully Submitted DATED: New York, New York June 10, 2014 By: /s/ Philippe Z. Selendy Philippe Z. Selendy Manisha M. Sheth Sascha N. Rand Andrew R. Dunlap David B. Schwartz QUINN EMANUEL URQUHART & SULLIVAN, LLP 51 Madison Avenue, 22nd Floor New York, New York 10010 (212) 849-7000 Attorneys for Plaintiff Federal Housing Finance Agency, as Conservator for Fannie Mae and Freddie Mac Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 72 of 76 A1 Appendix 1: The Securitizations The WLTD Securitizations No. Securitization Tranche SLG Sponsor/Seller Depositor Lead Underwriter Collateral Type 1. AHMA 2006-1 1A1 Group 1 Goldman Sachs Mortgage Company American Home Mortgage Assets LLC Goldman, Sachs & Co. Alt-A 1A2 Group 1 2. FFML 2005-FF11 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 3. FFML 2005-FF8 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 4. FFML 2006-FF13 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 5. GSAA 2005-11 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A 6. GSAA 2005-14 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A 1A2 Group 1 7. GSAA 2005-15 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A 8. GSAA 2006-11 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A 9. GSAA 2006-2 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 10. GSAA 2006-4 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 73 of 76 A2 No. Securitization Tranche SLG Sponsor/Seller Depositor Lead Underwriter Collateral Type 11. GSAA 2006-5 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A 12. GSAA 2006-8 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A 13. GSAA 2007-6 3A1A Group 3 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A 14. GSAMP 2005-AHL2 A1A Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 15. GSAMP 2005-HE5 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 16. GSAMP 2005-HE6 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 17. GSAMP 2005-WMC2 A1A Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 18. GSAMP 2005-WMC3 A1A Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 19. GSAMP 2006-FM1 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 20. GSAMP 2006-FM2 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 21. GSAMP 2006-FM3 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 22. GSAMP 2006-HE3 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 23. GSAMP 2006-HE4 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 24. GSAMP 2006-HE5 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 25. GSAMP 2006-HE7 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 26. GSAMP 2006-HE8 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 74 of 76 A3 No. Securitization Tranche SLG Sponsor/Seller Depositor Lead Underwriter Collateral Type 27. GSAMP 2006-NC2 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 28. GSAMP 2007-FM1 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 29. GSAMP 2007-FM2 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 30. GSAMP 2007-HE1 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 31. GSAMP 2007-HE2 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 32. GSAMP 2007-NC1 A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 33. GSR 2006-OA1 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A 34. GSR 2007-AR2 6A1 Group 6 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Prime 35. GSR 2007-OA1 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A 36. GSR 2007-OA2 1A1 Group 1 Goldman Sachs Mortgage Company GS Mortgage Securities Corp. Goldman, Sachs & Co. Alt-A Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 75 of 76 A4 The Third-Party Securitizations No. Securitization Tranche SLG Sponsor/Seller Depositor Lead Underwriter Collateral Type 1. ACCR 2005-4 A1 Group 1 Accredited Home Lenders, Inc. Accredited Mortgage Loan REIT Trust Goldman, Sachs & Co. Subprime 2. FHLT 2006-E 1A1 Group 1 Fremont Investment & Loan Fremont Mortgage Securities Corp. Goldman, Sachs & Co. Subprime 3. INDX 2005-AR18 1A1 Group 1 IndyMac Bank, F.S.B. IndyMac MBS, Inc. Goldman, Sachs & Co. Alt-A 4. INDX 2005-AR27 2A1 Group 2 IndyMac Bank, F.S.B. IndyMac MBS, Inc. Goldman, Sachs & Co. Alt-A Case 1:11-cv-06198-DLC Document 769 Filed 07/08/14 Page 76 of 76