Wesley, Circuit Judge:

In the wake of the Great Depression, Congress took measures to protect the U.S. economy from suffering another catastrophic collapse. Congress's first step in that endeavor was the Securities Act of 1933 (the "Securities Act" or "Act"), ch. 38, 48 Stat. 74 (codified as amended at 15 U.S.C. § 77a et seq.). The Act's chief innovation was to replace the traditional buyer-beware or caveat emptor rule of contract with an affirmative duty on sellers to disclose all material information fully and fairly prior to public offerings of securities. That change marked a paradigm shift in the securities markets. See Ernst & Ernst v. Hochfelder, 425 U.S. 185, 194-95, 96 S.Ct. 1375, 47 L.Ed.2d 668 (1976).

This case demonstrates the persistent power of the Securities Act's full-disclosure requirement in the context of the Great Recession. The height of the housing bubble in the mid-2000s saw an explosion in the market for residential mortgage-backed securities ("RMBS"). See Adam J. Levitin & Susan M. Wachter, Explaining the Housing Bubble, 100 GEO. L.J. 1177, 1192-202 (2012). In the midst of that market frenzy, two government-sponsored enterprises, the Federal Home Loan Mortgage Corporation ("Freddie Mac" or "Freddie)" and Federal National Mortgage Association ("Fannie Mae" or "Fannie") (collectively, the "GSEs"), purchased a subset of RMBS known as private-label securitizations ("PLS") from a host of private banks. Defendants-appellants Nomura1 and RBS2 (collectively, "Defendants")3 sold the GSEs seven of these certificates (the "Certificates") in senior tranches of PLS (the "Securitizations") using prospectus supplements (the "ProSupps"). Each ProSupp described the creditworthiness of the loans supporting the Securitization, including an affirmation that the loans "were originated generally in accordance with the underwriting criteria."

The housing market began to collapse in 2007 and the value of PLS declined rapidly. Shortly thereafter, plaintiff-appellee the Federal Housing Finance Agency (the "FHFA"), the statutory conservator of Freddie and Fannie,4 brought sixteen actions in the U.S. District Court for the Southern District of New York against financial institutions that sold PLS certificates to the GSEs, alleging that the offering documents used in those transactions overstated the reliability of the loans backing the securitizations, in violation of the Securities Act and analogous provisions of certain "Blue Sky laws,"5 the Virginia Securities Act, as amended, VA. CODE ANN. § 13.1-522, and the District of Columbia Securities Act, D.C. CODE § 31-5606.05.6 Sixteen of the FHFA's actions were coordinated before District Judge Denise Cote. Fifteen of those cases settled, resulting in more than $20 billion in recovery for the FHFA. The case on appeal was the only one to go to trial.

After issuing multiple pre-trial decisions and conducting a bench trial, the District Court filed a 361-page trial opinion rendering judgment in favor of the FHFA. The court found that Defendants violated Sections 12(a)(2) and 15 of the Securities Act, see 15 U.S.C. §§ 77l(a)(2), 77o, and analogous provisions of the Virginia and D.C. Blue Sky laws, see VA. CODE ANN. § 13.1-522(A)(ii); D.C. CODE § 31-5606.05(a)(1)(B), (c), by falsely stating in the ProSupps that, inter alia, the loans supporting the Securitizations were originated generally in accordance with the pertinent underwriting guidelines. As a result, the court awarded the FHFA more than $806 million in recession-like relief. Special App. 362-68.

Defendants appeal multiple aspects of the District Court's trial opinion, as well as many of the court's pretrial decisions. We find no merit in any of Defendants' arguments and AFFIRM the judgment. The ProSupps Defendants used to sell the Certificates to the GSEs contained untrue statements of material fact — that the mortgage loans supporting the PLS were originated generally in accordance with the underwriting criteria — that the GSEs did not know and that Defendants knew or should have known were false. Moreover, the FHFA's claims were timely, the District Court properly conducted a bench trial, Defendants are not entitled to a reduction in the FHFA's award for loss attributable to factors other than the untrue statements at issue, Defendants NAAC and NHELI were statutory sellers, and the FHFA exercised jurisdiction over Blue Sky claims.

BACKGROUND

I. Legal Framework

A. The Securities Act

"Federal regulation of transactions in securities emerged as part of the aftermath of the market crash in 1929." Ernst & Ernst, 425 U.S. at 194-95, 96 S.Ct. 1375. The first set of regulations came in the Securities Act, which was "designed to provide investors with full disclosure of material information concerning public offerings of securities in commerce, to protect investors against fraud and, through the imposition of specified civil liabilities, to promote ethical standards of honesty and fair dealing." Id. at 195, 96 S.Ct. 1375 (citing H.R. REP. No. 85, at 1-5 (1933)). Shortly thereafter, Congress passed a series of companion statutes, including the Securities Exchange Act of 1934 (the "Exchange Act"), ch. 404, 48 Stat. 881 (codified as amended at 15 U.S.C. § 78a et seq.), which was intended "to protect investors against manipulation of stock prices through regulation of transactions upon securities exchanges and in over-the-counter markets, and to impose regular reporting requirements on companies whose stock is listed on national securities exchanges." Ernst & Ernst, 425 U.S. at 195, 96 S.Ct. 1375 (citing S. REP. No. 792, at 1-5 (1934)). Congress's purpose for this regulatory scheme "`was to substitute a philosophy of full disclosure for the philosophy of caveat emptor ... in the securities industry.'" Basic Inc. v. Levinson, 485 U.S. 224, 234, 108 S.Ct. 978, 99 L.Ed.2d 194 (1988) (quoting SEC v. Capital Gains Research Bureau, Inc., 375 U.S. 180, 186, 84 S.Ct. 275, 11 L.Ed.2d 237 (1963)).

The Securities Act regulates the use of prospectuses in securities offerings. A prospectus is "any prospectus, notice, circular, advertisement, letter, or communication, written or by radio or television, which offers any security for sale or confirms the sale of any security," with certain exceptions not applicable here. 15 U.S.C. § 77b(a)(10). Section 5(b)(1) of the Securities Act provides that it is unlawful "to make use of any means or instruments of transportation or communication in interstate commerce or of the mails to carry or transmit any prospectus relating to any security" unless the prospectus meets certain disclosure requirements. 15 U.S.C. § 77e(b)(1); see 17 C.F.R. § 230.164. Section 5(b)(2) provides that it is unlawful "to carry or cause to be carried through the mails or in interstate commerce any such security for the purpose of sale or for delivery after sale, unless accompanied or preceded by a prospectus" that meets additional disclosure requirements. 15 U.S.C. § 77e(b)(2).

Section 12(a)(2) of the Act, as amended, 15 U.S.C. § 77l, accords relief to any person (1) who was offered or purchased a security "by means of a prospectus or oral communication"; (2) from a statutory seller; (3) when the prospectus or oral communication "includes an untrue statement of a material fact or omits to state a material fact necessary in order to make the statements, in the light of the circumstances under which they were made, not misleading"; and (4) the plaintiff did not "know[] of such untruth or omission" at the time of sale (the "absence-of-knowledge element"). 15 U.S.C. § 77l(a)(2); see In re Morgan Stanley Info. Fund Sec. Litig. (Morgan Stanley), 592 F.3d 347, 359 (2d Cir. 2010). Scienter, reliance, and loss causation are not prima facie elements of a Section 12(a)(2) claim. Morgan Stanley, 592 F.3d at 359.

Section 12 authorizes two types of mutually-exclusive recovery. See 15 U.S.C. § 77l(a); Wigand v. Flo-Tek, Inc., 609 F.2d 1028, 1035 (2d Cir. 1979). If the plaintiff owned the security when the complaint was filed, Section 12 authorizes rescission — the plaintiff returns the security to the defendant and the defendant refunds the plaintiff the purchase price with adjustments for interest and income. See 15 U.S.C. § 77l(a); Wigand, 609 F.2d at 1035. If the plaintiff no longer owned the security when the complaint was filed, Section 12(a)(2) permits the plaintiff to recover "damages." 15 U.S.C. § 77l(a); see Wigand, 609 F.2d at 1035.

Section 12 contains two affirmative defenses. First, a plaintiff will not be entitled to relief if the defendant "did not know, and in the exercise of reasonable care could not have known, of [the] untruth or omission" at issue. 15 U.S.C. § 77l(a)(2). This is known as the "reasonable care" defense. Morgan Stanley, 592 F.3d at 359 n.7.

Second, a defendant may seek a reduction in the amount recoverable under Section 12 equal to

15 U.S.C. § 77l(b). This is known as the "loss causation" defense, Iowa Pub. Emps.' Ret. Sys. v. MF Glob., Ltd., 620 F.3d 137, 145 (2d Cir. 2010), or "negative loss causation," In re Smart Techs., Inc. S'holder Litig., 295 F.R.D. 50, 59 (S.D.N.Y. 2013). Unlike the Exchange Act, which generally requires plaintiffs to prove loss causation as a prima facie element, see 15 U.S.C. § 78u-4(b)(4), the Securities Act places the burden on defendants to prove negative loss causation as an affirmative defense, see McMahan & Co. v. Wherehouse Entm't, Inc., 65 F.3d 1044, 1048 (2d Cir. 1995).

Section 12 is closely related to Section 11 of the Securities Act, as amended, 15 U.S.C. § 77k, which "imposes strict liability on issuers and signatories, and negligence liability on underwriters," for material misstatements or omissions in a registration statement. NECA-IBEW Health & Welfare Fund v. Goldman Sachs & Co. (NECA), 693 F.3d 145, 156 (2d Cir. 2012). Both provisions are limited in scope and create in terrorem7 liability. See id.; William O. Douglas & George E. Bates, The Federal Securities Act of 1933, 43 YALE L.J. 171, 173 (1933). The loss causation defense in Section 12 was adapted from the loss causation defense in Section 11(e) of the Securities Act. See S. REP. No. 104-98, at 23 (1995).

Finally, Section 15 of the Act, as amended 15 U.S.C. § 77o, provides that "[e]very person who ... controls any person liable under ... [Section 12(a)(2)] shall also be liable jointly and severally with and to the same extent as such controlled person." 15 U.S.C. § 77o(a). "To establish [Section] 15 liability, a plaintiff must show a `primary violation' of [Section 12] and control of the primary violator by defendants." See In re Lehman Bros. Mortg.-Backed Sec. Litig., 650 F.3d 167, 185 (2d Cir. 2011) (quoting ECA, Local 134 IBEW Joint Pension Tr. of Chi. v. JP Morgan Chase Co., 553 F.3d 187, 206-07 (2d Cir. 2009)).

In this case, the District Court awarded the FHFA rescission-like relief against all Defendants under Section 12(a)(2) and found NHA, NCCI, and the Individual Defendants control persons under Section 15 for the seven PLS transactions at issue. FHFA v. Nomura Holding Am., Inc. (Nomura VII), 104 F.Supp.3d 441, 598 (S.D.N.Y. 2015). Defendants appeal the District Court's decisions as to each prima facie element of the Section 12(a)(2) claims (except that the sales were made by means of a prospectus) and as to both affirmative defenses.8

B. The Blue Sky Laws

The Commonwealth of Virginia and District of Columbia have enacted Blue Sky laws modeled on the Securities Act as originally enacted in 1933. Andrews v. Browne, 276 Va. 141, 662 S.E.2d 58, 62 (2008); see Forrestal Vill., Inc. v. Graham, 551 F.2d 411, 414 & n.4 (D.C. Cir. 1977) (observing that the D.C. Blue Sky law was based on the Uniform Securities Act); see also Gustafson v. Alloyd Co., Inc., 513 U.S. 561, 602-03, 115 S.Ct. 1061, 131 L.Ed.2d 1 (1995) (Ginsburg, J., dissenting) (observing that the Uniform Securities Act was based on the Securities Act of 1933). These Blue Sky laws contain provisions that are "substantially identical" to Sections 12(a)(2) and 15. Dunn v. Borta, 369 F.3d 421, 428 (4th Cir. 2004); see Hite v. Leeds Weld Equity Partners, IV, LP, 429 F.Supp.2d 110, 114 (D.D.C. 2006).9 As relevant to this appeal, the Blue Sky laws are distinct only in that each requires as a jurisdictional element that some portion of the securities transaction at issue occurred in the State. D.C. CODE § 31-5608.01(a); see Lintz v. Carey Manor Ltd., 613 F.Supp. 543, 550 (W.D. Va. 1985) (citing Travelers Health Ass'n v. Commonwealth, 188 Va. 877, 51 S.E.2d 263 (1949)).

The District Court awarded the FHFA relief under the D.C. Blue Sky law for the sale of one Certificate and relief under the Virginia Blue Sky law for the sales of three other Certificates. Nomura VII, 104 F.Supp.3d at 598.

II. Factual Background10

This case centers on the RMBS industry of the late 2000s. RMBS are asset-backed financial instruments supported by residential mortgage loans. A buyer of an RMBS certificate pays a lump sum in exchange for a certificate representing the right to a future stream of income from the mortgage loans' principal and income payments. PLS are RMBS sold by private financial institutions. See Pension Benefit Guar. Corp. ex rel. St. Vincent Catholic Med. Ctrs. Ret. Plan v. Morgan Stanley Inv. Mgmt. Inc. (Pension Benefit Guar.), 712 F.3d 705, 713-14 (2d Cir. 2013).

This case touches on nearly every aspect of the PLS securitization process — from the issuance of mortgage loans through the purchase of a securitization. Because of the size and complexity of this case, in addition to the fact that the final order rule requires us to review a number of the District Court's pre-trial rulings, see 28 U.S.C. § 1291, there is much to consider. We think it best to begin with a summary of the securitization process from 2005 to 2007, the time period relevant to this case, and then to introduce the parties and the transactions at issue. Issue-specific facts are addressed in more detail in the discussion sections below.

A. The PLS Securitization Process

1. Originating a Mortgage Loan Using Underwriting Guidelines

The first step in the PLS process was the issuance of residential mortgage loans. Mortgage loans were issued to borrowers by entities known as originators. Originators issued loans according to their loan underwriting guidelines, which listed the criteria used to approve a loan. See United States ex rel. O'Donnell v. Countrywide Home Loans, Inc. (O'Donnell), 822 F.3d 650, 653 n.2 (2d Cir. 2016). These guidelines helped each originator assess the borrower's ability to repay the loan and the value of the collateral. Originators balanced those two criteria to determine a potential loan's credit risk.

Following the underwriting guidelines, originators required each prospective borrower to complete a loan application, usually on the Uniform Residential Loan Application (the "URLA"). The URLA required borrowers to disclose, under penalty of civil liability or criminal prosecution, their income, employment, housing history, assets, liabilities, intended occupancy status for the property, and the sources of the funds they intended to use in paying the costs of closing the loan. Originators used this information to determine objective factors relevant to the borrower's credit risk, such as a credit score according to the Fair Isaac Corporation's model (a "FICO score"), credit history, and debt-to-income ratio. Once each borrower submitted the URLA, the originator kept it and other related documentation in the borrower's loan file.

The underwriting guidelines required originators to assess the reasonableness of the borrower's assertions on the URLA. This was easiest when borrowers supported their URLA applications with corroborating documentation. Some applications required verification of both the borrower's assets and income, while some required verification only of the borrower's assets. Other borrowers submitted stated-income-stated-assets ("SISA") applications, which did not require verification of income or assets, or no-income-no-assets ("NINA") applications, which were complete without the borrower even stating his or her income or assets. SISA and NINA applications were more difficult to assess, but not categorically ineligible to receive loans.

The underwriting guidelines generally permitted originators to accept SISA and NINA applications and to make other exceptions to the underwriting criteria if there were compensating factors that indicated the borrower's ability and willingness to repay the loan. The guidelines set forth the specific conditions under which exceptions would be permitted. Originators were required to mark the borrower's loan file whenever an exception to the underwriting criteria had been granted and to explain the basis for that decision.

After forming an opinion about a borrower's creditworthiness based on the URLA and related documentation, originators assigned the transaction a credit risk designation, which affected the interest rate for the loan. When an applicant had good credit, the transaction was labeled "prime." When an applicant had materially impaired credit, the transaction was labeled "subprime." And when an applicant's credit fell between good and materially impaired, the transaction was labeled "Alt-A." See Pension Benefit Guar., 712 F.3d at 715.

Once they had assessed the borrower's credit, originators balanced that assessment against the value of the collateral (i.e., the present market value of the residence the borrower wanted to purchase or refinance), as determined by an appraiser, to measure the overall credit risk of the loan. Originators compared the amount of the loan against the value of the collateral to develop a loan-to-value ratio, a key indicator of credit risk. It was common in the RMBS industry to use a loan-to-value ratio of 80% as a benchmark. Relative to loans at that ratio, a loan worth between 80% and 90% of the collateral value was 1.5 times more likely to default and a loan worth between 95% and 100% of the collateral value was 4.5 times more likely to default. A loan-to-value ratio of more than 100% meant that the loan exceeded the value of the residence and the borrower was "underwater."

If the originator was comfortable with the overall credit risk after reviewing the buyer's creditworthiness, the value of the collateral, and the loan-to-value ratio, the loan would be approved.

The underwriting guidelines and loan files were crucial throughout and beyond the origination process. Supervisors employed by the originators could check loan files against the underwriting guidelines to ensure that loan issuance decisions met important criteria. For example, the District Court found that "[c]ompliance with underwriting guidelines ensure[d] ... an accurate calculation of the borrower's [debt-to-income] ratio, which is a critical data point in the evaluation of a loan's risk profile." Nomura VII, 104 F.Supp.3d at 536. After the loan issued, originators used the information in the loan file to describe the loan characteristics for financial institutions interested in purchasing it.

2. Creating a PLS

The next step in the PLS process was the aggregation and securitization of the residential mortgage loans into an RMBS. Originators compiled their issued loans into "trade pools" and then solicited bids from PLS "sponsors" or "aggregators" to purchase them. The originators provided prospective bidders with a "loan tape" for each pool — "a spreadsheet that provided data about the characteristics of each loan in the trade pool" including "loan type (fixed or adjustable rate), ... original and unpaid principal balance, amortization term, borrower's FICO score, the mortgaged property's purchase price and/or appraised value, occupancy status, documentation type and any prepayment penalty-related information." J.A. 4385.

The sponsor that prevailed in the bidding process was given access to a limited number of loan files to conduct a due diligence review of the originators' underwriting and valuation processes before final settlement.11 The sponsor was entitled prior to closing to remove from the trade pool any loans that did not meet its purchasing requirements, such as those below a minimum FICO score or exceeding a maximum debt-to-income ratio. Upon closing, the prevailing sponsor acquired title to the loans in the trade pools and gained access to the complete set of loan files. The prevailing sponsor was also given a copy of the underwriting guidelines the originators used to issue the loans.

The sponsor then sold the loans to a "depositor," a special purpose vehicle created solely to facilitate PLS transactions. The true sale from sponsor to depositor was intended to protect the future PLS certificate-holders' interests in the loans in the event that the sponsor declared bankruptcy. It was common in the RMBS industry for the depositor and sponsor entities to act at the direction of the same corporate parent.

The depositor then grouped the loans into supporting loan groups ("SLGs") and transferred each group of loans to a trust. In exchange, the trust issued the depositor certificates that represented the right to receive principal and interest payments from the SLGs. The trustee managed the loans for the benefit of the certificate holders, often hiring a mortgage loan servicing vendor to manage the loans on a day-to-day basis. The depositor then sold most of the certificates to a lead underwriter, who would shepherd them to the public securities markets; a few certificates remained under the ownership of the depositor. It was also common in the industry for the lead underwriter to be controlled by the same corporate parent that controlled the sponsor and depositor.

3. Preparing a PLS for Public Sale

The final steps in the PLS process were the preparation and sale to the public of the certificates. The lead underwriter, sponsor, and depositor (collectively, "PLS sellers") worked together to structure the securitization, to solicit credit ratings for the certificates principally from three major credit-rating agencies, Moody's Investors Service, Inc. ("Moody's"), Standard & Poor's ("S & P"), and Fitch Ratings ("Fitch") (collectively, the "Credit-Rating Agencies" or "Rating Agencies"), and to draft and confirm the accuracy of the offering documents. Once those tasks were completed, the lead underwriter would market the certificates to potential buyers.

The PLS sellers structured securitizations with two credit enhancements that distributed the risk of the loans unequally among the certificate holders. The first was subordination. The PLS certificates were organized into tranches, ranked by seniority. Each SLG supported one or more tranches of certificates and distributed payments in a "waterfall" arrangement. This arrangement guaranteed senior certificate-holders first claim to all principal and interest payments. Once all the senior certificate-holders were satisfied, the SLGs' payments spilled over to junior certificate-holders, who would receive the remaining balance of the payments.

The second of these credit enhancements was overcollateralization. The total outstanding balance of all of the mortgage loans supporting an entire PLS often exceeded the outstanding balance of the loans supporting the publicly available PLS certificates. As a result, some loans in the PLS were tethered to certificates owned by the depositor or sponsor and were not available for public purchase. These non-public loans served as a loss-saving measure by making payments to the public certificate-holders (in order of seniority) in the event that the loans supporting their public certificates defaulted.

After structuring the PLS, the PLS sellers would solicit a credit rating for each tranche. Because, as the District Court explained, PLS "were only as good as their underlying mortgage loans," Nomura VII, 104 F.Supp.3d at 465, the Credit-Rating Agencies based their determinations primarily on the quality of the certificates' supporting loans. They did this by modeling the credit risk of the SLGs using information from the loan tape, provided by the PLS sellers. The Rating Agencies also evaluated the certificates' credit enhancements.

The Rating Agencies' review included examining draft offering documents for representations that the supporting loans were originated in accordance with originators' underwriting criteria. This was standard in the industry, as the Rating Agencies agreed that compliance with the underwriting guidelines was an important indicator of a loan's credit risk. More credit enhancements were required to secure an investment-grade rating for any certificate backed by loans that either did not comply with the underwriting guidelines or were missing documentation from their loan files.

The PLS sellers explained these credit enhancements, credit ratings, and other important features of the PLS to the public primarily in three offering documents — a shelf registration, a free writing prospectus, and a prospectus supplement. The shelf registration was a pre-approved registration statement filed with the Securities and Exchange Commission (the "SEC") that contained generally applicable information about PLS. See 17 C.F.R. §§ 230.409, 230.415. The shelf registration enabled the lead PLS underwriter to make written offers to potential buyers using a free writing prospectus. See id. § 230.433(b)(1). The free writing prospectus broadly described the characteristics of the certificate and the supporting SLGs. If an offeree was interested after reading the description, it could commit to purchasing the certificate. Title in the certificate and payment were exchanged within approximately a month of that commitment. The PLS sellers sent the buyer a prospectus supplement and filed the same with the SEC near the date of that exchange.12

The prospectus supplement contained the most detailed disclosures of any of the offering documents. This document provided specific information regarding the certificate, the SLGs, and the credit quality of the underlying loans. It warranted the accuracy of its representations regarding loan characteristics. And, crucially, it affirmed that the loans in the SLGs were originated in accordance with the applicable underwriting guidelines. As the District Court noted, "whether loans were actually underwritten in compliance with guidelines was extremely significant to investors." Nomura VII, 104 F.Supp.3d at 536. The prospectus supplement ordinarily disclosed that some number of loans in the SLG may deviate substantially from, or violate, the applicable underwriting guidelines.13

B. The PLS Transactions at Issue

1. The Parties

a. The Sellers

Defendants sold the Certificates to the GSEs. Subsidiaries of Defendant NHA were the Certificates' primary sellers. Defendant NCCI served as the sponsor for all seven of the transactions at issue. Defendant NAAC served as the depositor for one Securitization, and Defendant NHELI served as the depositor for the remaining six. And Defendant Nomura Securities, served as the lead or co-lead underwriter for three of the Securitizations.

Defendant RBS served as the lead or co-lead underwriter for four of the Securitizations.14

b. The Buyers15

Fannie and Freddie purchased the Certificates. Both GSEs are privately-owned corporations chartered by Congress to provide stability and liquidity in the mortgage loan market. Fannie was established in 1938. See National Housing Act Amendments of 1938, ch. 13, 52 Stat. 8. Freddie was established in 1970. See Emergency Home Finance Act of 1970, Pub. L. No. 91-351, 84 Stat. 450. They were at the time of the transactions at issue, and remain today, "the dominant force[s]" in the mortgage loan market. See Town of Babylon v. FHFA, 699 F.3d 221, 225 (2d Cir. 2012).

The primary way the GSEs injected liquidity into the mortgage market was by purchasing mortgage loans from private loan originators. See O'Donnell, 822 F.3d at 653. This side of the GSEs' operations was known as the "Single Family Businesses." By purchasing loans from originators, the Single Family Businesses replenished originators' capital, allowing originators to issue new loans. The Single Family Businesses held the loans purchased from originators on their books and sometimes securitized them into agency RMBS, similar to a PLS, to be offered for public sale. See Pension Benefit Guar., 712 F.3d at 714-15; Levitin & Wachter, supra, at 1187-89.

The Single Family Businesses contained due diligence departments. These departments conducted due diligence of specific loans prior to purchase. They also periodically reviewed their originator counterparties' general underwriting practices, and PLS sellers' due diligence practices, including Defendants'.16

As a secondary element of their businesses, the GSEs operated securities trading desks that purchased PLS. PLS purchases created liquidity in the mortgage market by funneling cash back through PLS sponsors and underwriters to loan originators for use in future loans. The GSEs' PLS traders generally operated out of Fannie's headquarters in Washington, D.C. and Freddie's headquarters in McLean, Virginia.

The GSEs played a significant role in the PLS market despite the relatively minor role it occupied in their businesses. The GSEs' PLS portfolios reached their heights in 2005, when they owned approximately $350 billion worth of PLS, with $145 billion backed by subprime loans and $40 billion backed by Alt-A loans (loans that were rated lower than prime loans but higher than subprime loans). The GSEs bought approximated 8% of the $3 trillion dollars' worth of PLS sold from 2005 to 2007. PLS traders working for the GSEs purchased the Certificates at issue.

2. The Transactions

Between 2005 and 2007, the GSEs purchased Certificates from Defendants in seven PLS Securitizations — NAA 2005-AR6, NHELI 2006-FM1, NHELI 2006-HE3, NHELI 2006-FM2, NHELI 2007-1, NHELI 2007-2, and NHELI 2007-3. These transactions were executed generally in accordance with the standard practices at the time, as described in the previous sections. The supporting loans are predominantly Alt-A or subprime. Each Certificate is in a senior tranche of its respective Securitization. Combined, the Certificates cost approximately $2.05 billion and, at times of sale, had expected value of $2.45 billion.17

Defendants sold the Certificates by means of shelf registrations, free writing prospectuses, and the ProSupps.18 The ProSupps provided detailed information regarding the loans in the SLGs. They described the risks inherent in subprime and Alt-A loan transactions and provided the credit ratings for each tranche. They included charts displaying the objective characteristics for loans in each SLG, such as aggregate remaining principal balances, FICO scores, and loan-to-value ratios. Five ProSupps promised that "[i]f ... any material pool characteristic differs by 5% or more from the description in this [ProSupp], revised disclosure will be provided either in a supplement or in a Current Report on Form 8-K." E.g., J.A. 9120.

Most importantly for purposes of this appeal, every ProSupp stated that "the Mortgage Loans ... were originated generally in accordance with the underwriting criteria described in this section," (the "underwriting guidelines statement"). J.A. 9117; see J.A. 6884, 7174, 7527, 7895, 8296, 8718.19 The ProSupps then described the underwriting criteria used by originators that contributed loans to the SLGs and stated that the originators may have made "certain exceptions to the underwriting standards ... in the event that compensating factors are demonstrated by a prospective borrower." E.g., J.A. 9117. Six of the ProSupps described the specific underwriting guidelines for each originator that alone contributed more than 20% of the loans in the SLGs. For these originators, the ProSupps typically also stated that the loans were issued "generally" in accordance with the underwriting guidelines. E.g., J.A. 7520.

Six of the ProSupps stated that some loans were issued under "Modified [Underwriting] Standards." E.g., J.A. 9118. The ProSupps stated that these modified standards permitted originators, for example, to issue loans to foreign nationals, who might lack reliable sources to verify their credit score or lack a score altogether, or use "less restrictive parameters" in issuing loans, such as "higher loan amounts, higher maximum loan-to-value ratios, ... the ability to originate mortgage loans with loan-to-value ratios in excess of 80% without the requirement to obtain mortgage insurance if such loans are secured by investment properties." E.g., J.A. 9119. The ProSupps disclosed the number of loans issued under the modified standards.20

C. The Housing and Financial Crisis21

The GSEs purchased the Certificates from Defendants during a period when the markets for mortgage loans and associated securities were exploding. A combination of factors including low interest and unemployment rates, an increased use of adjustable-rate mortgages and other innovative loan products, and government policies encouraging home ownership heated the housing market. Home prices increased, and aggregate mortgage debt in the U.S. more than doubled between 2000 and 2008.

During this period, originators also relaxed underwriting standards. Subprime lending jumped from 9.5% of all new mortgage loans in 2000 to 20% of all new mortgage loans in 2005; Alt-A lending also grew substantially. Originators also began to approve loans that failed to meet the underwriting guidelines with an eye towards securitizing these loans quickly, thus transferring the credit risk of the loans from originators to PLS certificate-holders. See Levitin & Wachter, supra, at 1190.

Securitization fueled the credit bubble. As described above, securitization enabled originators to shift credit risk to the financial markets and turn the prospect of future loan repayment into instant cash for new loans. In 2000, the PLS market was worth less than $150 billion. By 2005-2006, the PLS market was worth more than $1.1 trillion. Once it began, the securitization frenzy built on itself — securitizations of subprime mortgages increased the quantity of new subprime mortgage originations. Those new mortgages were in turn securitized, and the cycle started over.

The housing market began its decline in 2006. Increased mortgage interest rates led to a spike in prices that made many homes too expensive for potential buyers, decreasing demand. An oversupply of housing also put downward pressure on home prices. U.S. housing prices started to fall in April 2006. From April 2007 through May 2009, they fell almost 33%.

Default and delinquency rates increased with the decline in housing prices. By 2009, 24% of homeowners, many of whom had purchased homes during the mid-2000s boom, were left with negative equity: mortgages with outstanding principal balances greater than the homes' current valuations. Shoddy underwriting practices, which approved loans for borrowers who could not afford to repay, and spikes in adjustable mortgage rates also contributed to an increase in defaults. With rising interest rates, refinancing was difficult. Defaulting on mortgage loans became an attractive option for homeowners. Each default and resulting foreclosure sale depressed the prices of surrounding homes further, sending the housing market into a vicious downward cycle.

Increased default rates had an adverse impact on investment products tied to mortgage loans, and on the entire financial system as a result. As principal and interest payments slowed over the course of 2007, the value of these securities declined. One bank in August 2007 reported that the decrease in mortgage securitization markets' liquidity made it "impossible" to value certain RMBS instruments. J.A. 5419. Banks that had invested heavily in RMBS sold off their positions (driving down the value of those assets further) and closed related hedge fund divisions. Credit tightened, interbank lending ceased, and concerns about financial institutions' liquidity and solvency led to runs on financial institutions. Several major financial institutions, including Lehman Brothers, Bear Sterns, and Merrill Lynch, experienced significant financial stress.

In December 2007, the U.S. entered a one-and-a-half-year recession, the longest since the Great Depression. U.S. real gross domestic product contracted by about 4.3% during that time. Unemployment rose to 10% in 2009, more than double the 2007 rate.

III. Procedural History

In the aftermath of the financial crisis, Congress passed the Housing and Economic Recovery Act of 2008 (the "HERA"), Pub. L. No. 110-289, 122 Stat. 2654, out of concern for the GSEs' financial condition. See UBS II, 712 F.3d at 138. The HERA created the FHFA, an "independent agency of the Federal Government," 12 U.S.C. § 4511(a), to serve as a conservator for Fannie, Freddie, and other GSEs in financial straits, see id. § 4617(a). The HERA empowered the FHFA to "collect all obligations and money due the [GSEs]," id. § 4617(b)(2)(B)(ii), and take other actions necessary to return them to solvency. Id. § 4617(b)(2)(B)(i).

On September 2, 2011, the FHFA initiated sixteen actions that were eventually litigated together in the Southern District of New York, including the instant "Nomura action," against financial institutions that sold PLS certificates to Fannie Mae and Freddie Mac. These cases were consolidated before Judge Cote. They all settled before trial, with the exception of this case.

The FHFA began the Nomura action by bringing claims under Sections 11, 12(a)(2), and 15 of the Securities Act and Virginia and D.C. Blue Sky analogs based on alleged misstatements in the PLS offering documents. The FHFA alleged that Defendants' offering documents falsely stated (1) the underwriting guidelines statement, (2) the supporting loans' loan-to-value ratios, (3) whether mortgaged properties were occupied by the mortgagors, and (4) that the Credit-Rating Agencies were provided with accurate information regarding loan characteristics before issuing ratings decisions. The FHFA initially demanded a jury trial for "all issues triable by jury." J.A. 409.

The District Court issued numerous pre-trial decisions. Defendants appeal from the following:

Trial was originally slated to be held before a jury to decide the Section 11 claims, while the District Court would decide the Section 12 claims, with the jury's determination controlling overlapping factual issues. Roughly a month before pretrial memoranda were due, the FHFA voluntarily withdrew its Section 11 claim. As a result, the District Court, over Defendants' objection, conducted a four-week bench trial on the Section 12, Section 15, and Blue Sky claims.22

One month after trial concluded, the District Court issued a detailed 361-page opinion systematically finding for the FHFA on each claim. See generally Nomura VII, 104 F.Supp.3d 441. The court held that Defendants violated Section 12(a)(2) because each ProSupp contained three categories of false statements of material information: (1) the underwriting guidelines statements, (2) the loan-to-value ratio statements, and (3) the credit ratings statements. See id. at 559-73. Our focus on appeal, on this point, is devoted solely to the statements regarding underwriting guidelines, which are sufficient to affirm the court's judgment. See 15 U.S.C. § 77l(a)(2) (authorizing relief if the offering documents contain just one untrue statement of material fact); N.J. Carpenters Health Fund v. Royal Bank of Scot. Grp., PLC (N.J. Carpenters Health Fund II), 709 F.3d 109, 116, 123 (2d Cir. 2013) (allowing a Section 11 lawsuit to proceed on the allegation that RMBS offering documents falsely stated that the loans adhered to the underwriting guidelines).

The court also rejected Defendants' loss causation defense, see Nomura VII, 104 F.Supp.3d at 585-93, found that Defendants violated the analogous provisions of the Virginia and D.C. Blue Sky laws, see id. at 593-98, and held that NHA, NCCI, and the Individual Defendants were control persons under Section 15, see id. at 573-83. The court awarded the FHFA $806,023,457, comprised of roughly $555 million for violations of the Blue Sky laws and roughly $250 million for violations of the Securities Act. See id. at 598.23

This appeal followed.

DISCUSSION

Our discussion proceeds in two parts. The first addresses issues the District Court resolved before trial: (A) whether the FHFA's claims were timely under the statutes of repose; (B) whether in light of the GSEs' generalized knowledge and experience in the mortgage loan market (1) the FHFA's claims were timely under the statutes of limitations and (2) the FHFA was entitled to summary judgment holding that the GSEs did not know the ProSupps' underwriting guidelines statements were false; (C) whether the FHFA was entitled to summary judgment holding that Defendants failed to exercise reasonable care; and (D) whether the Seventh Amendment entitled Defendants to a jury trial. The second addresses issues resolved after trial: (A) whether the FHFA is entitled to relief under Section 12(a)(2) because (1) each Defendant is a statutory seller, (2) the underwriting guidelines statements were false, (3) those statements were material, and (4) Defendants failed to make out an affirmative defense of loss causation; as well as (B) whether the FHFA is entitled to relief under the analogous Virginia and D.C. Blue Sky provisions.

I. Pretrial Decisions24

A. Statutes of Repose

Defendants appeal the District Court's denial of their motion for summary judgment on the ground that the FHFA's claims, which were filed on September 2, 2011 (more than three years after the Securitizations were sold), were time-barred by the Securities Act, Virginia Blue Sky, and D.C. Blue Sky statutes of repose. See 15 U.S.C. § 77m (three-year period of repose); VA. CODE ANN. § 13.1-522(D) (two-year period of repose); D.C. CODE § 31-5606.05(f)(1) (three-year period of repose).25 The District Court held that the statutes of repose were displaced by an extender provision in the HERA, codified at 12 U.S.C. § 4617(b)(12), which permits the FHFA to bring any "tort claim" within three years and any "contract claim" within six years of its appointment as the GSEs' conservator on September 6, 2008.26 See FHFA v. UBS Ams., Inc. (UBS I), 858 F.Supp.2d 306, 313-17 (S.D.N.Y. 2012) (holding all coordinate cases brought by the FHFA before September 6, 2011 timely under the HERA), aff'd, UBS II, 712 F.3d 136 (2d Cir. 2013); see also HSBC III, 2014 WL 4276420, at *1. On appeal, Defendants argue that while the HERA displaces otherwise applicable statutes of limitations, it does not affect statutes of repose.

In UBS II, a 2013 decision in an interlocutory appeal in one of the FHFA's parallel coordinated actions, a panel of this Court held that § 4617(b)(12) "supplants any other [federal or state] time limitations that otherwise might have applied" to the FHFA's actions, including the Securities Act and Blue Sky statutes of repose. 712 F.3d at 143-44. This conclusion was compelled by the definitive language in § 4617(b)(12), which makes clear that "the applicable statute of limitations with regard to any action brought by the [FHFA]... shall be" time periods provided in the HERA, see UBS II, 712 F.3d at 141-42 (internal quotation marks omitted) (quoting 12 U.S.C. § 4617(b)(12)), and was corroborated by the purpose of the HERA to permit the FHFA to "`collect all obligations and money due' to the GSEs[] to restore them to a `sound and solvent condition,'" id. at 142 (quoting 12 U.S.C. §§ 4617(b)(2)(B)(ii), (D)). We considered that reading § 4617(b)(12) to preclude and pre-empt all types of time-limitation statutes, including statutes of repose, was consistent with Congress's intent because it allowed the FHFA more "time to investigate and develop potential claims on behalf of the GSEs." Id.

Ordinarily, UBS II would end our inquiry. See Lotes Co., Ltd. v. Hon Hai Precision Indus. Co., 753 F.3d 395, 405 (2d Cir. 2014) ("[A] panel of this Court is `bound by the decisions of prior panels until such time as they are overruled either by an en banc panel of our Court or by the Supreme Court.'" (quoting In re Zarnel, 619 F.3d 156, 168 (2d Cir. 2010))). But one year after UBS II was decided, the Supreme Court handed down CTS Corp. v. Waldburger, ___ U.S. ___, 134 S.Ct. 2175, 189 L.Ed.2d 62 (2014), which held that 42 U.S.C. § 9658,27 a provision in the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (the "CERCLA") that imposes a federal commencement date for state statutes of limitations, does not pre-empt state statutes of repose. See 134 S.Ct. at 2188. Defendants' sole argument in the present appeal is that CTS abrogated UBS II.

This is not the first case in this Circuit to consider the impact of CTS on UBS II. In FDIC v. First Horizon Asset Sec., Inc. (First Horizon), 821 F.3d 372 (2d Cir. 2016), cert. denied, ___ U.S. ___, 137 S.Ct. 628, 196 L.Ed.2d 518 (2017), we held that CTS did not disturb the portion of UBS II's holding that held § 4617(b)(12) precludes the federal Securities Act's statute of repose. Id. at 380-81. That forecloses Defendants' argument insofar as it applies to the FHFA's claims under the Securities Act.28 See Lotes, 753 F.3d at 405.

It remains an open question in this Circuit whether CTS undermined the portion of UBS II's holding that held § 4617(b)(12) pre-empts the Virginia and D.C. Blue Sky laws' statutes of repose. Cf. Church & Dwight Co., Inc. v. SPD Swiss Precision Diagnostics, GmBH, 843 F.3d 48, 64-65 (2d Cir. 2016) (observing that pre-emption analysis does not control preclusion analysis).29 "[C]oncerns about the primacy of federal law and the state-federal balance" that are unique to the pre-emption context presented here distinguish it from preclusion context in First Horizon. Church & Dwight Co., 843 F.3d at 64 (internal quotation mark omitted) (quoting POM Wonderful LLC v. Coca-Cola Co., ___ U.S. ___, 134 S.Ct. 2228, 2236, 189 L.Ed.2d 141 (2014)). Still, some aspects of our earlier preclusion analysis aid in deciding the pre-emption issue on this appeal. Cf. id. ("[P]re[-]emption principles can be `instructive' in the ... preclusion context...." (quoting POM Wonderful, 134 S.Ct. at 2236)).30

Nothing about CTS seriously undermines UBS II. The Supreme Court's analysis in CTS focused primarily on four considerations. First, § 9658 provides that state law will be the default rule for time limitations and that a federal commencement date will operate as a limited "exception" to that rule. This suggested to the Court that Congress intended § 9658 to leave many of the state time-limitation rules in place. See CTS, 134 S.Ct. at 2185 (majority opinion). Second, § 9658 refers explicitly to a "statute of limitations" but does not mention a "statute of repose." Although this was not dispositive of the ultimate issue, the Court took this as an indication that Congress did not intend § 9658 to reach statutes of repose. Id. at 2185-86. Third, Congress, in debating the CERCLA, considered a report that recommended language providing for explicit pre-emption of state statutes of repose, but chose not to include the proposed language in the final statute. Id. at 2186. Fourth, § 9658 defines the state provisions it preempts as the "applicable limitations period[s]" during "which a civil action may be brought" and provides for equitable tolling in certain circumstances, two concepts inapplicable to repose analyses. Id. at 2187-88 (internal quotation marks omitted). For these reasons, the Supreme Court held § 9658 did not reflect clear congressional intent to pre-empt overlapping state statutes of repose. Id. at 2188.

One similarity between § 4617(b)(12) and § 9658 is that both refer to statutes of limitations but neither references statutes of repose. See First Horizon, 821 F.3d at 376, 379. While this might suggest on first glance that neither statute reaches repose statutes, we reasoned in UBS II that an explicit statutory reference to repose statutes is not a sine qua non of congressional intent to pre-empt such statutes. See 712 F.3d at 142-43. CTS confirmed — rather than undermined — that reasoning. See 134 S.Ct. at 2185. CTS observed that usage of the terms "limitations" and "repose" "has not always been precise." Id. at 2186; accord UBS II, 712 F.3d at 142-43 ("Although statutes of limitations and statutes of repose are distinct in theory, the courts... have long used the term `statute of limitations' to refer to statutes of repose...."). Indeed, although Congress has indisputably created statutes of repose in the past, it "has never used the expression `statute of repose' in a statute codified in the United States Code." First Horizon, 821 F.3d at 379 (observing that 15 U.S.C. § 77m, titled "Limitation of actions," creates a three-year repose period); see also Cal. Pub. Emps.' Ret. Sys. v. ANZ Sec., Inc. (CalPERS), ___ U.S. ___, 137 S.Ct. 2042, 2049, 198 L.Ed.2d 584 (2017) (analyzing federal statute to determine whether it included a statute of limitation or statute of repose). As a result, CTS cautioned, while the presence of the term "statute of limitations" in a federal statute may be "instructive" of Congress's intended pre-emptive scope, it is not "dispositive." See 134 S.Ct. at 2185. That reinforces UBS II's refusal to resolve its pre-emption inquiry based solely on the bare text of § 4617(b)(12). See First Horizon, 821 F.3d at 376.31

Defendants also argue that, under CTS, § 4617(b)(12)'s repeated use of the words "claim accrues" indicates that it was meant only to pre-empt statutes of limitations. In CTS, the Supreme Court noted that § 9658 pre-empts the "commencement date" for any "applicable limitations period" under state law, 42 U.S.C. § 9658(a)(1), and defines the "applicable limitations period" as the period when "a civil action [alleging injury or damage caused by exposure to a hazardous substance] may be brought," id. § 9658(b)(2). See 134 S.Ct. at 2187. That indicated to the Court that Congress intended to displace only the commencement date for statutes of limitations because a "statute of repose... `is not related to the accrual of any cause of action.'" Id. (quoting 54 C.J.S., LIMITATIONS OF ACTIONS § 7, p. 24 (2010)).

Section 4617 uses some similar language. It provides that the new filing period for claims brought by the FHFA is at least six years for any "contract" claim and three years for any "tort" claim, "beginning on the date on which the claim accrues." 12 U.S.C. §§ 4617(b)(12)(A)(i)(I), (ii)(I). It also describes how to determine "the date on which a claim accrues" for purposes of the HERA. Id. § 4617(b)(12)(B). Defendants argue that this language — specifically the words "claim accrues" — carries the same indication of congressional intent as § 9658's definition of the "applicable limitations period."

We disagree. CTS does not stand for the proposition that whenever "accrue" appears in a federal statute it is a talismanic indication of congressional intent to pre-empt only statutes of limitations. Context is crucial. Congress used the phrase "a civil action ... may be brought" in § 9658 in defining the class of state statutes it intended to pre-empt. In contrast, Congress used the words "claim accrues" in § 4617(b)(12) in defining the time limitation the HERA newly created for claims brought by the FHFA. Put another way, the HERA's use of the word "accrues" "tells us ... that [§ 4617(b)(12)] is itself a statute of limitations" but does not "provide[]... guidance on the question whether [§ 4617(b)(12)] displaces otherwise applicable statutes of repose...." First Horizon, 821 F.3d at 379.

The only remaining argument against pre-emption of the state statutes of repose is that both § 9658 and § 4617(b)(12) pre-empt certain time limitations for state claims while leaving untouched "other important rules governing civil actions." CTS, 134 S.Ct. at 2188. "`The case for federal pre-emption is particularly weak where Congress has indicated its awareness of the operation of state law in a field of federal interest, and has nonetheless decided to stand by both concepts and to tolerate whatever tension there is between them.'" Id. (brackets omitted) (quoting Wyeth v. Levine, 555 U.S. 555, 575, 129 S.Ct. 1187, 173 L.Ed.2d 51 (2009)). But § 9658 leaves in place far more of state law than § 4617(b)(12). Section 9658 provides only a federally mandated accrual date for state limitations periods and leaves unchanged "States' judgments about causes of action, the scope of liability, the duration of the period provided by statutes of limitations, burdens of proof, [and] rules of evidence." CTS, 134 S.Ct. at 2188. Section 4617(b)(12), by contrast, provides a comprehensive, singular time limitation for all actions brought by the FHFA. See UBS II, 712 F.3d at 141-42. It governs entirely the rules regarding when the FHFA may bring its claims — from the moment the filing period commences, see 12 U.S.C. § 4617(b)(12)(B), through the length of the period for each type of the claim, see id. § 4617(b)(12)(A). Congress has not stood by any state time-limitation rules when it comes to claims brought by the FHFA as the GSEs' conservator.

In all other respects, CTS and UBS II arose in substantially different contexts. Section 9658's legislative history reveals that Congress specifically considered and decided against using language that would explicitly pre-empt statutes of repose. See CTS, 134 S.Ct. at 2186. There is no similar legislative history for Section 4617(b)(12). See UBS II, 712 F.3d at 143. Section 9658 "describ[es] the [pre-empted] period in the singular," which "would be an awkward way to mandate the pre-emption of two different time periods." CTS, 134 S.Ct. at 2186-87. Section 4617(b)(12) applies "to any action brought by the [FHFA]," 12 U.S.C. § 4617(b)(12)(A) (emphasis added), "`including claims to which a statute of repose generally attaches.'" UBS II, 712 F.3d at 143 (quoting UBS I, 858 F.Supp.2d at 316-17). Section 9658 contains a provision for equitable tolling, an important characteristic of statutes of limitations that distinguishes them from statutes of repose. See CTS, 134 S.Ct. at 2187-88. There is no similar provision in § 4617(b)(12).

In sum, "CTS's holding is firmly rooted in a close analysis of § 9658's text, structure, and legislative history." First Horizon, 821 F.3d at 377. None of those statute-specific considerations undermines UBS II's close analysis of § 4617(b)(12), which differs significantly from § 9658. We reaffirm our prior holding that Congress designed § 4617(b)(12) to pre-empt state statutes of repose.32

B. Knowledge Issues — Statutes of Limitations and Knowledge of the ProSupps' Underwriting Guidelines Misrepresentations

Defendants next raise two pre-trial issues that turn on the extent to which the GSEs were or should have been aware that the ProSupps' underwriting guidelines statements were false. The first is the statute of limitations. In addition to the statute of repose discussed above, Section 13 of the Securities Act contains a statute of limitations that bars any action not brought within one year after the plaintiff learned or should have learned of the material misstatement or omission giving rise to the claim. 15 U.S.C. § 77m; see CalPERS, 137 S.Ct. at 2049 (2017) (discussing three-year time bar).33 The HERA extended the filing period only for contract claims that were valid on (or became valid after) September 6, 2008, the date when the FHFA assumed conservatorship. See 12 U.S.C. §§ 4617(b)(12)(A), (12)(B); J.A. 341-42. Thus, any FHFA claim that was time-barred by Section 13 on that date remained time-barred under the HERA. On the FHFA's motion for summary judgment, the District Court held that the FHFA's claims were timely as a matter of law. The court concluded that no reasonable jury could find the GSEs knew or should have known as of September 6, 2007, one year before the HERA extender became effective, that ProSupps' underwriting guidelines statements were false, despite widespread PLS credit downgrades in the summer of 2007 and the Single Family Businesses' generalized experience with mortgage loan originators and PLS aggregators. Nomura I, 60 F.Supp.3d at 502-09; see also UBS I, 858 F.Supp.2d at 321-22. Defendants contest that decision on appeal.

The second, related issue is whether the FHFA was entitled to summary judgment on the purchaser's absence-of-knowledge element of a Section 12(a)(2) claim. See 15 U.S.C. § 77l(a)(2).34 The District Court granted the FHFA summary judgment on this element, holding again that the Single Family Businesses' expertise in the general mortgage loan market did not provide adequate knowledge of the specific untruths in the ProSupps. See HSBC II, 33 F.Supp.3d at 480-93. Defendants also contest this decision on appeal.

We address these issues in tandem, as the relevant facts and legal questions overlap in large part.

1. Factual Summary35

a. The Single Family Businesses' Due Diligence

The GSEs' Single Family Businesses, in their capacities as aggregators and sponsors of RMBS instruments, gathered a significant amount of information about the mortgage loan market and mortgage loan originators. Fannie's Single Family due diligence division was the Single Family Counterparty Risk Management Group (the "SFCPRM"); Freddie's Single Family due diligence division was the Alternative Market Operations Group (the "AMO"). Through the work of the SFCPRM and AMO, the GSEs amassed "more knowledge about the mortgage market than probably anybody else." J.A. 1317.

The SFCPRM and AMO conducted counterparty reviews of originators with whom the GSEs regularly did business. These reviews involved desk audits and on-site visits to originators' offices. Often the GSEs hired Clayton Advisory Services, Ltd. ("Clayton"), a third-party mortgage diligence vendor, to re-underwrite a sample of the originators' issued loans and assess the originators' compliance with their underwriting guidelines. The GSEs also analyzed originators' adherence to appraisal protocols, capability to detect fraud, and ability to meet repurchase obligations. If an originator received a positive result from this review, the GSE placed, or maintained, it on a list of approved originators.

The SFCPRM and AMO conducted counterparty reviews for at least five originators that issued loans backing the Certificates in this case; we note some pertinent results of those reviews below:

The GSEs' knowledge about the mortgage loan industry required a delicate information sharing arrangement between their Single Family Businesses and their PLS traders.

On the one hand, the GSEs did not want to purchase loans or securitizations supported by loans that they knew were originated or aggregated by companies they did not trust. The Single Family Businesses' research proved helpful to the PLS traders in that regard; and indeed each GSE required that any originator that individually contributed more than a certain percentage (10% for Fannie, 1% for Freddie) of the total unpaid principal balance of a PLS be on its list of approved originators.

On the other hand, the GSEs were concerned that its PLS traders would violate federal insider-trading laws if, before purchasing PLS, they reviewed the certain loan-specific information the Single Family Businesses considered in making purchases for their own aggregation practices. The GSEs accordingly limited their PLS traders' access to only the Single Family Businesses' reviews of originators' general practices. Fannie's PLS traders were given the final lists of approved originators; Freddie's were given the full counterparty review paperwork. PLS traders were not given access to any specific loan-level information for the transactions at issue.

The SFCPRM and AMO also evaluated PLS sellers and maintained a list of approved PLS counterparties. Both Nomura and RBS were placed on the GSEs' lists of approved PLS sellers. In August 2004, the AMO rated Nomura's due diligence program "Satisfactory" based on Nomura's "good due diligence methodologies, reasonable valuation processes and sound controls." Id. at 3170. In a November 2006 review, the SFCPRM noted it had access to somewhat limited information to review RBS's diligence, but apparently accepted RBS's characterization of its practices as robust. Nomura I, 60 F.Supp.3d at 491.

Despite ensuring that they purchased loans from approved originators and PLS sellers, the GSEs knew that there was still a risk that some defective loans could creep into SLGs for PLS certificates they purchased. The heads of the GSEs' PLS portfolios acknowledged in deposition testimony that they believed that loans in an SLG "would reflect the general underwriting practices of the originators responsible for those loans." J.A. 10323. That meant that "if an originator was not following its own guidelines and was contributing loans to the collateral for the pool," the GSEs "would have expected that loans not underwritten to the originator's guidelines would then end up in the" SLGs. Id. at 10325 (emphasis added; internal quotation marks omitted). To limit that possibility, the GSEs required "rep[resentation]s and warrant[ie]s" from the approved PLS sellers for each certificate they purchased, believing that they could rely on those institutions to limit the number of the defective loans to an immaterial level. Id. at 1063; see also HSBC II, 33 F. Supp. 3d at 471 ("[A Fannie employee] testified that Fannie Mae's `process basically relied on the dealers and originators providing it with reps and warranties as to the validity of how these loans were underwritten.'" (brackets omitted)).

b. The GSEs' Awareness of PLS Market Trends

GSEs were also familiar with public information about the overall RMBS market in 2006 and 2007. This information included a growing number of reports of borrower fraud and lower underwriting standards among mortgage loan originators. Beginning in July and August of 2007, it also included reports that the three primary credit-rating agencies, Moody's, S & P, and Fitch, began to accelerate their negative views of RMBS.

On July 10, 2007, Moody's downgraded the junior tranches of many RMBS — including Securitizations NHELI 2006-FM1 and NHELI 2006-FM2. The credit ratings for the senior tranches in these Securitizations did not change. Moody's attributed its downgrades to "a persistent negative trend in severe delinquencies for first lien subprime mortgage loans securitized in 2006." Nomura I, 60 F.Supp.3d at 498 (internal quotation marks omitted). Moody's noted that the supporting loans "were originated in an environment of aggressive underwriting" and that increased default rates were caused in part by "certain misrepresentations ... like occupancy or stated income and appraisal inflation." Id. (internal quotation marks omitted; brackets omitted).

That same day, S & P placed on negative rating watch a host of RMBS — but none of the Securitizations — citing "lower underwriting standards and misrepresentations in the mortgage market." Id. (internal quotation mark omitted). S & P questioned the quality of the data "concerning some of the borrower and loan characteristics provided during the rating process." Id. S & P made clear that, going forward, its ratings for RMBS certificates would hew more closely to their seniority within the securitization.

After expressing doubt on July 12, beginning in August of 2007 Fitch downgraded hundreds of RMBS. On August 3, 2007, Fitch downgraded junior tranches in Securitizations NHELI 2006-FM2 and NHELI 2006-HE3, but Fitch did not downgrade the senior tranches in those Securitizations at that time.

On August 17, 2007, S & P downgraded junior tranches in Securitization NAA 2005-AR6. As with Moody's and Fitch's downgrades, S & P did not change its rating for the senior tranches in the Securitization at that time.

The Rating Agencies took no further action on the Securitizations through September 6, 2007. As of that date, none of the GSEs' senior-tranche Certificates had been downgraded, but junior tranches in NHELI 2006-FM2 had been downgraded by two Rating Agencies, and junior tranches in NAA 2005-AR6, NHELI 2006-FM1, and NHELI 2006-HE3 had each been downgraded by one Rating Agency.

The GSEs monitored these junior tranche downgrades. The GSEs understood that the credit risks of the all of the tranches in a Securitization were connected. At least one Fannie employee during the summer of 2007 attempted to ascertain whether the GSE owned any Certificates in Securitizations that had been downgraded. On August 17, 2007, a Fannie employee circulated internally "a short eulogy for the subprime RMBS market." Id. at 499.

2. Analysis

a. Statutes of Limitations

Section 13's statute of limitations extinguishes any action not "brought within one year after the discovery of the untrue statement or the omission, or after such discovery should have been made by the exercise of reasonable diligence." 15 U.S.C. § 77m. The filing period commences "when the plaintiff discovers (or should have discovered) the securities-law violation." CalPERS, 137 S.Ct. at 2049. A securities-law violation is discovered when the plaintiff learns "sufficient information about [the violation] to ... plead it in a complaint" with enough "detail and particularity to survive a [Federal Rule of Civil Procedure] 12(b)(6) motion to dismiss." MBIA, 637 F.3d at 175. A plaintiff is charged with knowledge of any fact that "a reasonably diligent plaintiff would have discovered." Id. at 174 (internal quotation mark omitted) (quoting Merck & Co., Inc. v. Reynolds, 559 U.S. 633, 653, 130 S.Ct. 1784, 176 L.Ed.2d 582 (2010)).

"[W]hen the circumstances would suggest ... the probability that" a violation of the securities laws has occurred — a situation sometimes called "storm warnings" — we deem the plaintiff on inquiry notice and assume that a reasonable person in his or her shoes would conduct further investigation into the potential violation. Lentell v. Merrill Lynch & Co., Inc., 396 F.3d 161, 168 (2d Cir. 2005) (internal quotation marks omitted) (quoting Levitt v. Bear Stearns & Co., Inc., 340 F.3d 94, 101 (2d Cir. 2003)). Under prior Circuit law, the Section 13 limitations period could begin to run as early as the moment a plaintiff knew or should have known of storm warnings that placed it on inquiry notice. See Staehr v. Hartford Fin. Servs. Grp., 547 F.3d 406, 426 (2d Cir. 2008).36 The Supreme Court's decision in Merck changed that rule. See 559 U.S. at 650-53, 130 S.Ct. 1784. After Merck, we still assume a reasonable plaintiff on inquiry notice would conduct further investigation, but the limitations period begins to run only when, in the course of that investigation, the reasonable plaintiff would have discovered sufficient information to plead a securities-law violation adequately. See id. at 651, 130 S.Ct. 1784; MBIA, 637 F.3d at 174.37

A storm warning "need not detail every aspect of the alleged" securities-law violation. Staehr, 547 F.3d at 427. Information triggers the duty to inquire if it "`relates directly to the misrepresentations and omissions the [p]laintiff[] ...allege[s] in [its] action against the defendants,'" id. (alteration omitted) (quoting Newman v. Warnaco Grp., Inc., 335 F.3d 187, 193 (2d Cir. 2003)), and is, in the totality of the circumstances, "specific enough to provide an ordinary investor with indications of the probability (not just the possibility) of" a violation. Id. at 430 (emphases added; citations omitted). For example, we have found that an insurance company taking three substantial "reserve charges" followed by a national periodical publishing an article about the company's issues with reserves triggered a duty to inquire about the company's concealment of a negligent practice to under-reserve for insurance claims. See LC Capital Partners, LP v. Frontier Ins. Grp., Inc., 318 F.3d 148, 155 (2d Cir. 2003). We have also found a plaintiff on inquiry notice regarding a bank's concealment of conflicts of interest when a magazine article described one of an affiliated financial research analyst's conflicts. See Shah v. Meeker, 435 F.3d 244, 249-51 (2d Cir. 2006). But we have found "generic articles" regarding structural conflicts of interest in the financial services industry insufficient to trigger a duty to inquire about specific instances of knowing and intentional fraud in that industry. See Lentell, 396 F.3d at 170.

In this case, Defendants argue the GSEs became aware of two categories of storm warnings before September 6, 2007, one year prior to the effective date of the HERA's extender provision. First, Defendants argue that the GSEs, through their Single Family Businesses, knew first-hand that originators that issued loans supporting the Securitizations had subpar underwriting practices. That knowledge, the argument goes, would have caused a reasonable investor in the GSEs' shoes to conduct an investigation into whether the loans in the SLGs supporting the Securitizations were poorly underwritten. Second, Defendants contend that the credit downgrades of junior tranches in the Securitizations in the summer of 2007 put the GSEs on notice that the supporting loans were not as trustworthy as the ProSupps portrayed.

We are not persuaded. The Single Family Businesses' generalized experience with originators in the mortgage loan market did not trigger inquiry notice to investigate the specific representations in the ProSupps. The Single Family Businesses clearly knew or should have known that some originators who issued loans backing the Certificates were, as a general matter, less-than-rigorous in adhering to underwriting guidelines. But they reasonably believed that not every loan issued by those originators was defective, that the SLGs backing the Certificates did not contain all of the originators' loans, and that the SLGs were not representative samples of the originators' entire loan pools. The SLGs contained specific loans that Defendants specifically selected from a larger population of loans issued by the originators.

Generalized knowledge that originators issued some defective loans alone would not cause a reasonable investor to believe necessarily that his or her particular PLS certificates were backed by such loans. A reasonable investor's suspicions would be raised only if Defendants' loan-selection processes were also defective such that the shoddily underwritten loans would slip past their screens and into the SLGs. In this case, there was little indication of that, as both Nomura and RBS were approved by the GSEs as PLS counterparties.

Neither do the acknowledgments by leaders in the GSEs' PLS trading departments that they expected the SLGs to contain some defective loans indicate that a reasonable investor in their shoes would have investigated whether the ProSupps contained false statements. See HSBC II, 33 F.Supp.3d at 471. Those statements reflect an understanding that due diligence processes are never perfect and a reasonable expectation that those processes may fail to excise an immaterial number of defects from the SLGs. Knowledge of a risk of immaterial deviations is quite different from knowledge of a risk of material deviations. For a material portion of the SLGs, a reasonable investor would do exactly as the GSEs' did — "rel[y] on the dealers and originators providing ... reps and warranties as to the validity of how these loans were underwritten." Id. (internal quotation mark omitted; brackets omitted).

Defendants argue that the GSEs were not entitled to rely on Defendants' diligence and should have assumed that the loans in the SLGs were representative of the originators' entire loan pools because the ProSupps did not represent that the loans in the SLGs would be "the cream of the crop." RBS's Br. 35. While it is true that the ProSupps made no representations about the loans in the SLGs relative to other loans the originators issued, the ProSupps did represent that the loans in the SLGs "were originated generally in accordance with the underwriting criteria." E.g., J.A. 6884. A reasonable investor in the GSEs' shoes would take that statement for all that it was worth: an affirmation that, regardless of the quality of the median loan in the residential mortgage market, these specific loans in these specific SLGs met the underwriting criteria.

Neither would the credit downgrades of junior tranches cause a reasonable investor in the GSEs' shoes to investigate whether the ProSupps contained material misstatements or omissions. To be sure, the Credit-Rating Agencies' bearish turn on RMBS expectations revealed that they had begun to doubt the strength of the loans in the downgraded securitizations' SLGs, and those doubts would cause some concern for every reasonable certificate-holder regardless of seniority. As a product of the subordination for senior PLS certificates, a single SLG supported junior and senior-tranche certificates simultaneously. Thus, concerns about the SLGs' creditworthiness could reach the senior tranches of any Securitization that had downgraded junior tranches. See Nomura I, 60 F.Supp.3d at 499 ("The GSEs recognized that, generally, downgrades to junior tranches increased the risk of a future downgrade to the GSEs' senior tranches.").

It does not follow, however, that the summer 2007 credit downgrades would cause a reasonable senior-certificate holder to believe the PLS offering documents contained false statements that were material. See Staehr, 547 F.3d at 430 (observing that a storm warning triggers inquiry notice only when it indicates a probability of a full securities violation). The senior and junior certificate-holders did not have the same risk exposure. Certificate-holders were entitled to distributions of principal, interest, and collateral in the supporting loans in descending order of seniority. A reasonable senior certificate-holder might understand the Rating Agencies' decisions to downgrade junior tranches while maintaining the senior-tranche ratings to mean that any misrepresentation in the offering documents was mild enough that the subordination and over-collateralization still insulated them from loss. On that understanding, tranche-specific downgrades might seem material to a reasonable investor in a junior certificate but not to a reasonable investor in a senior certificate.

Finally, under Merck, it was Defendants' burden to prove that a reasonable investor in the GSEs' shoes would have conducted a fulsome investigation and uncovered information sufficient to make out a plausible claim for relief by September 6, 2007 — just weeks after the credit downgrades. See MBIA, 637 F.3d at 174. Defendants adduced "no evidence of ... how long it would take a reasonably diligent investor in the GSEs' position to investigate the [instant Section 12(a)(2)] claims such that it could adequately plead them." Nomura I, 60 F.Supp.3d at 509; see also Pension Tr. Fund, 730 F.3d at 279 (concluding that it would have taken a reasonable institutional investor in RMBS using a "proprietary process" that involved analyzing "court filings" two months to uncover loan-quality misrepresentations in offering documents). Their failure to establish this indispensable piece of the statute of limitations defense dooms their argument on appeal.

b. Absence-of-Knowledge Element

Section 12(a)(2) requires the plaintiff to prove that it did not "know[]" of the material misstatement in the prospectus. 15 U.S.C. § 77l(a)(2); see Healey v. Chelsea Res., Ltd., 947 F.2d 611, 617 (2d Cir. 1991). This is an actual knowledge standard. See Casella v. Webb, 883 F.2d 805, 809 (9th Cir. 1989). In contrast to the reasonable care affirmative defense (discussed below), Section 12 does not require the plaintiff to undertake any investigation or prove that it could not have known the falsity of the misstatement at issue. See 15 U.S.C. § 77l(a)(2) (precluding recovery if the defendant "did not know, and in the exercise of reasonable care could not have known, of such untruth or omission"). Section 12 requires plaintiffs to prove only that they in fact lacked knowledge of the falsity. See N.J. Carpenters Health Fund v. Rali Series 2006-QO1 Tr., 477 Fed. Appx. 809, 813 n.1 (2d Cir. 2012) (summary order); cf. N.J. Carpenters Health Fund II, 709 F.3d at 127 n.12 (observing that Section 11 creates an analogous "affirmative defense where a defendant can prove that `at the time of ... acquisition,' the purchaser `knew' of the alleged `untruth or omission'" (quoting 15 U.S.C. § 77k(a))).

Actual knowledge may be proven or disproven by direct evidence, circumstantial evidence, or a combination of the two. See Desert Palace, Inc. v. Costa, 539 U.S. 90, 100, 123 S.Ct. 2148, 156 L.Ed.2d 84 (2003). Publicly available information may provide relevant circumstantial evidence of actual knowledge. See id. However, Section 12's amenability to circumstantial evidence of actual knowledge should not be viewed as creating a constructive knowledge standard. The mere "[a]vailability elsewhere of truthful information cannot excuse untruths or misleading omissions in the prospectus." Dale v. Rosenfeld, 229 F.2d 855, 858 (2d Cir. 1956) (emphasis added). A plaintiff is entitled to recover under Section 12 if it was genuinely unaware of the falsity no matter how easily accessible the truth may have been.

Furthermore, Section 12 requires the plaintiff to prove only that it did not know that the specific statement at issue in the prospectus or oral communication was false. See 15 U.S.C. § 77l(a)(2) ("[T]he purchaser not knowing of such untruth or omission...." (emphasis added)). This is to be distinguished from knowing that there was a risk that the statement was false and from knowing that other similar statements in the same prospectus or other prospectuses were false. Section 12(a)(2)'s absence-of-knowledge element focuses on the buyer's actual knowledge of the truth-in-fact of the particular statement at issue.

For substantially the same reasons that undergird our statute of limitations ruling above, we conclude that the GSEs lacked actual knowledge of the falsity of the specific underwriting guidelines statements in the ProSupps. Defendants failed to link the GSEs' generalized knowledge about the mortgage loan origination industry to the ProSupps' specific statements regarding the quality of the loans in the SLGs. Section 12 permitted the GSEs to rely on the ProSupps' representations that the specific loans backing the Securitizations were originated generally in accordance with the underwriting criteria, regardless of the existence of other poorly issued loans in the market at the time. The Securities Act placed the sole burden on Defendants to ensure that representation was correct. See Basic, 485 U.S. at 234, 108 S.Ct. 978 (observing that the Securities Act replaces caveat emptor with "a philosophy of full disclosure" (quoting SEC v. Capital Gains Research Bureau, Inc., 375 U.S. 180, 186, 84 S.Ct. 275, 11 L.Ed.2d 237 (1963))).

Two cases that bear directly on the absence-of-knowledge issue warrant further discussion.

Defendants' absence-of-knowledge argument relies on an analogy to In re Initial Public Offerings Securities Litigation (IPO), 471 F.3d 24 (2d Cir. 2006). IPO was an appeal under Federal Rule of Civil Procedure 23(f) to review the certification of a class of investors in an action against underwriters of initial public offerings ("IPOs"). Id. at 27, 31. The class alleged that the underwriters violated Section 10(b) of the Exchange Act and Section 11 of the Securities Act by "condition[ing] allocations of shares at the [IPO] price on agreements to purchase shares in the aftermarket." Id. at 27. This scheme allegedly inflated secondary share prices and, consequently, the underwriters' compensation. See id. Part of the class's burden was to establish that it could provide common proof that each plaintiff lacked actual knowledge of the underwriters' aftermarket-purchase scheme. Id. at 43.

We held that the class failed to meet its burden. Id. at 43-44. The class initially based its allegations of the aftermarket-purchase scheme on an "industry-wide understanding" that IPO underwriters secured agreements to make such purchases, gleaned from customer interactions with those underwriters and from publicly available information in an SEC bulletin and news reports. Id. at 43. Those allegations of widespread knowledge led us to require individual inquiries into which members of the class had been exposed to that information before participating in an IPO. Id. at 43-44. We also concluded in a footnote that knowing about the aftermarket-purchase agreements was the functional equivalent of knowing about the scheme to inflate secondary securities prices because one could reasonably infer knowledge of the latter from knowledge of the former. Id. at 44 n.14.

Drawing on IPO, Defendants argue that the GSEs could have reasonably inferred that the ProSupps contained false statements from their Single Family Businesses' experience with the mortgage loan originators. That argument reads IPO too broadly. IPO — in the course of de-certifying the class — held that the widespread public information in that case made it too difficult to determine as a common question whether the plaintiffs had actual knowledge of the material misstatements at issue. We did not rule that public information alone can prove actual knowledge for all plaintiffs. See id. at 43-44. The district court on remand still had to examine whether each plaintiff had actual specific knowledge notwithstanding the public information.

We did not suggest in IPO that generalized public information plus a "reasonable inference" establishes specific knowledge. We stated that if a plaintiff actually knew about the aftermarket-purchase agreements, it was reasonable to infer that the plaintiff knew those agreements would result in inflated secondary market prices. Id. at 44 n.14. In other words, once a plaintiff had actual knowledge of a specific fact, a fact-finder could reasonably infer that the plaintiff knew of the natural specific consequences of that fact. For example, if the GSEs actually knew that the loans in the SLGs were not originated generally in accordance with the underwriting criteria, then under IPO, it would be reasonable to infer that the GSEs knew those loans were more likely to default and the value of the Certificates would likely fall. Defendants, however, attempt to establish actual knowledge of a specific fact (that the loans in the SLGs were defective) by drawing a "reasonable inference" from generalized knowledge about the mortgage loan industry. IPO cannot bear that weight.

Viacom International, Inc. v. YouTube, Inc. (Viacom), 676 F.3d 19 (2d Cir. 2012), is more on point. There, Viacom and other content-providers alleged that YouTube committed direct and secondary copyright infringement by hosting vast amounts of unlicensed copyrighted material on its website. Id. at 28-29. One issue was whether YouTube had actual and specific knowledge of the copyrighted material Viacom accused it of hosting. Id. at 32-34. Record evidence revealed that YouTube knew, based on internal surveys of its website, that between 75% and 80% of its content contained copyrighted material. Id. at 32-33. We concluded that those surveys were "insufficient, standing alone, to create a triable issue of fact as to whether YouTube actually knew, or was aware of facts or circumstances that would indicate, the existence of particular instances of infringement." Id. at 33. More evidence was required to establish that YouTube had actual knowledge of the copyrighted material specified in Viacom's complaint. See id. at 33-34.

The best case scenario for Defendants is no better than the survey evidence in Viacom. At most, the GSEs were aware that many PLS were supported by loans that were not originated in accordance with the underwriting guidelines. There is no evidence that the GSEs knew whether the specific PLS at issue were within or without the class of infected PLS. Without that crucial piece of information, Viacom precludes a reasonable jury from holding that the GSEs actually knew of the specific misstatements in the ProSupps.38

C. Reasonable Care Defense

Defendants appeal the District Court's grant of the FHFA's motion for summary judgment seeking to preclude Defendants from asserting a reasonable care defense at trial. Nomura II, 68 F.Supp.3d at 444-46.

Section 12(a)(2) provides a complete defense to any defendant who "did not know, and in the exercise of reasonable care could not have known," that the misstatement at issue was false. 15 U.S.C. § 77l(a)(2); see Morgan Stanley, 592 F.3d at 359 n.7.39 This raises a classic, mixed-law-and-fact question of reasonableness, usually committed to a jury. For that reason, only in the rare case can a court, viewing the facts in light most favorable to defendants, resolve the reasonable care defense as a matter of law. We are aware of only two other federal decisions, one of which was recently decided in a similar RMBS case, holding on summary judgment that a Section 12 defendant cannot pursue this defense. See Nat'l Credit Union Admin. Bd. v. UBS Sec., LLC, Nos. 12-2591, 12-2648, 2017 WL 411338, at *4-6 (D. Kan. Jan. 31, 2017) (granting partial summary judgment where "defendants essentially offered no evidence of due diligence," id. at *5); see also Plunkett v. Francisco, 430 F.Supp. 235, 241 (N.D. Ga. 1977).40

Nevertheless, the District Court held this was an "exceptional" case "where no reasonable, properly instructed jury could find" that Defendants should not have known that the ProSupps' statements affirming that the loans in the SLGs adhered to the underwriting guidelines were false. Nomura II, 68 F.Supp.3d at 445.41 Defendants argue on appeal that a reasonable jury could find that Defendants met the reasonable care standard because their due diligence complied with PLS industry practices at the time.

1. Factual Summary42

a. Nomura

Nomura's Transaction Management Group oversaw the process of purchasing and conducting due diligence of loans intended for securitization. Individual Defendants John P. Graham and N. Dante LaRocca both served, at different times, as the head of this group; Individual Defendant David Findlay, Nomura's Chief Legal Officer, also played a role in supervising this group. Nomura's Trading Desk purchased loans from originators, and Nomura's Due Diligence Group reviewed those loans. The Diligence Group consisted of between three and five employees, including its group leader, initially Joseph Kohout and later Neil Spagna.

The Trading Desk purchased a few loans individually, but a vast majority of the loans it securitized were purchased in trade pools. A trade pool with an aggregate principal balance greater than $25 million was known as a "bulk pool." All other trade pools were "mini-bulk pools." The SLGs at issue here were comprised of 15,806 loans, 14,123 (=89%) of which came from 54 bulk pools and 1,561 (=10%) of which came from 140 mini-bulk pools. The remaining 122 (=1%) loans in the SLGs were purchased individually. When an originator solicited bids for a trade pool, it made only the loan tape available to Nomura and other PLS aggregators. Traders did not review individual loan files before bidding on a pool.

After Nomura won a bid to purchase loans but before final settlement, the originators made available some number of loan files for Nomura's Diligence Group to review. Consistent with industry practices at the time, this pre-acquisition review was the only round of diligence Nomura conducted prior to offering the Securitizations to the public. The Diligence Group directed, inter alia, a credit review and a compliance review of the loans. The credit review examined whether the loans were originated in accordance with the originators' underwriting guidelines. The compliance review examined whether the loans complied with the relevant federal, state, and municipal regulations.

The Diligence Group conducted credit and compliance reviews for approximately 40% of the loans in the SLGs at issue. Nomura reviewed each loan purchased individually, virtually every loan purchased in a mini-bulk pool, and virtually all of the loans in 24 of the bulk pools. For the remaining 30 bulk pools (which contributed 82.1% of the total loans in the SLGs), the Diligence Group reviewed only a sample. Nomura's Trading Desk — not the Diligence Group — sometimes entered into agreements with counterparty originators limiting the size of the samples, which ranged from about 20% to 50% of the pool. Some of those agreements placed a hard cap on the size of the sample, while others affixed the size of the sample but entitled Nomura to request additional loans, a process known as "upsizing" the sample. Nomura did not upsize any of the samples at issue in this case.

Nomura used a non-random process to compile their samples. The Diligence Group selected 90% of the sample using a proprietary computer program created by S & P known as LEVELS. LEVELS employed adverse sampling, a process which involves combing through the loan tape to select for review the loans with the highest credit risk in a trade pool based on debt-to-income ratio, FICO score, loan-to-value ratio, and outstanding principal balance. The remaining 10% of the sample was selected "in an ad hoc fashion" based on similar risk factors. Id. at 451.

Kohout warned Nomura employees in an internal email that Nomura's use of LEVELS "is a non industry standard approach," J.A. 2631, and "does not conform to what is generally deemed to be effective by industry standards," id. at 2632. He stated that "when presenting our process to both internal and external parties, it will have to be made clear that [the Diligence Group's] role in both the sample selection and management of risk on bulk transactions has been diminished to the point of that of a non effective entity pursuant to our limited role in the process." Id. at 2631-32. The Single Family Businesses' counterparty reviews of PLS sponsors revealed that several other sponsors also used LEVELS to compose portions of their due diligence samples.

After selecting the sample, the Diligence Group deputized a third-party vendor, often Clayton or American Mortgage Consultants Inc. ("AMC"), to perform the credit and compliance reviews, with occasional oversight and assistance from Nomura employees. This was consistent with industry practices. The vendor used the sample loan files to re-underwrite the loans according to the originators' underwriting guidelines, additional criteria provided by Nomura, and applicable laws. The vendor gave each loan an "Event Level" ("EV") grade on a scale from 1 to 3, 1 indicating that the loan met all of the review criteria and 3 indicating that the loan materially deviated from the criteria or lacked critical documentation. The vendor then transmitted those grades to Nomura's Diligence Group on a document titled "Individual Asset Summaries."

The Diligence Group reviewed all of the vendor's EV2 and EV3 grades and as many as half of its EV1 grades. This review was limited to examining the "Individual Asset Summaries"; Nomura did not examine any loan files. The Diligence Group possessed the authority to issue client overrides that vacated the vendor's grade and to direct the vendor to re-grade the loan. With respect to the loans drawn from the 54 bulk pools that contributed to the SLGs here, the Diligence Group directed the vendor to change roughly 40% of the EV3 grades to EV2 grades.

The record contains one audit of Nomura's pre-acquisition review vendors, which LaRocca, then-head of Nomura's Transaction Management Group, reviewed. The audit report is dated August 24, 2006 (before four of the Securitizations settled). It finds that in a sample of 109 loans previously graded EV1 or EV2, seven of these should have received an EV3 grade and another 29 should have received no grade at all given the lack of supporting documentation. There is no evidence that Nomura changed its credit and compliance review processes after this audit.

After it received the final results of the third-party review, Nomura purchased all of the EV1 and EV2 loans — and acquired their loan files. Nomura intended to "kick out" (i.e., remove from the trade pool) all of the EV3 loans, although approximately 2.6% of the loans backing the Securitizations had been sampled and received an EV3 grade. In an internal email, Spagna stated that "typical" kick-out rate ranged from 7% to 8% of the sample and a rate of 12.12% was "much higher" than average. Id. at 2639. The average kick-out rates for the trade pools at issue was 15.2%.

Nomura held most of the purchased loans for between two and five months. During that time, the Trading Desk grouped the purchased loans into SLGs. Nomura's traders made loan-by-loan selections using a non-random process designed to create SLGs that would meet market demands. The traders based their evaluations of the loans on factors such as credit scores, geographic concentrations, and loan-to-value ratios. Nomura conducted no review of the SLGs' creditworthiness as a whole.

Nomura's Transaction Management Group wrote the ProSupps after the SLGs were formed. The ProSupps made representations about the characteristics of the SLGs. For three of the Securitizations, there is no specific evidence that Nomura verified the accuracy of these representations. For four of the Securitizations, Nomura's verification process consisted of the Transaction Management Group reviewing a "Due Diligence Summary" — a single page created by the Diligence Group listing the percentage of loans to be securitized that had been reviewed and the kick-out rates for the trade pools. Each summary included a disclaimer: "The material contained herein is preliminary and based on sources which we believe to be reliable, but it is not complete, and we do not represent that it is accurate." J.A. 2876.

b. RBS

RBS, the lead or co-lead underwriter for four of the Securitizations, also reviewed the loans in the SLGs. RBS's due diligence was led by Brian Farrell, the Vice President of RBS's credit risk department.

For two of the Securitizations it underwrote, RBS conducted no independent review. This practice was common among underwriters in the PLS industry. RBS's review of NHELI 2006-HE3 diligence consisted of reviewing three documents created by Nomura — an aforementioned Due Diligence Summary, an additional summary of collateral characteristics, and a list of the names of the originators that contributed more than 5% of the loans in the SLGs. RBS also relied on Nomura-provided data integrity studies that affirmed the ProSupps contained no input errors or mathematical miscalculations, as well as a "negative assurance letter" from Nomura's counsel that stated counsel was unaware of any facts that would render the ProSupps misleading.

RBS's review of NHELI 2006-FM2 consisted primarily of reviewing reports from AMC that described the loans. Before transmitting it to RBS, Nomura reviewed these reports and discovered that the SLGs contained 19 EV3 loans, despite Nomura's policy against purchasing such loans. Spagna, who took over Nomura's Diligence Group after Kohout, emailed AMC and requested that it "mark these loans as client overrides Credit Event 2s for all 19 loans in question" and then "forward to me the updated set of reports for these two deals." J.A. 2878. The vendor complied and Nomura sent RBS the reports as revised. After noting one issue based on experience with a particular originator, RBS approved the vendor's reports.

RBS also participated in a teleconference with RBS's counsel, Nomura (represented in part by Spagna), Nomura's counsel, and other underwriters to discuss diligence on NHELI 2006-FM2. Spagna recalled to a fellow Nomura employee that RBS asked two questions about Nomura's diligence processes, that he "took the liberty to bullshit them," and that he thought "it worked." Id at 2881.

After NHELI 2006-FM2 had closed, an RBS employee emailed Farrell to discuss RBS's diligence for this deal. Farrell wrote: "We did not perform actual diligence on this. Diligence was performed by another company for Nomura. We signed off on their results." Nomura II, 68 F.Supp.3d at 460. The RBS employee responded: "How frequently is this done?" Id. Farrell replied: "Since being employed, this is the only review type I was involved in where due diligence results were reviewed and a new diligence was not ordered." Id. (brackets omitted).

RBS did conduct independent reviews of sample loans from NHELI 2007-1 and NHELI 2007-2. RBS selected samples using adverse sampling in part and "semi-random" sampling in remaining part. J.A. 2606. The semi-random technique grouped the remaining loans by unpaid principal balance and selected randomly from within those groups. For NHELI 2007-1, RBS's sample contained 5.8% of the adjustable-rate loans in a group, part of which eventually composed the relevant SLG. For NHELI 2007-2, Farrell requested RBS employees to form a larger sample, preferably 25% of the loan pool, because he thought the loans were "crap." Id. at 2886. In the end, RBS sampled 6% of the loans from the NHELI 2007-2 SLG.

RBS's diligence as an underwriter was similar to Nomura's as a PLS sponsor.43 RBS outsourced its credit and compliance reviews to Clayton, which used loan files to re-underwrite the loans in each sample subject to client overrides. The re-underwriting analyses for NHELI 2007-1 yielded 33 loans (=32% of the sample) graded "3," the equivalent of EV3. Within an hour and six minutes after Clayton transmitted that information to RBS, RBS issued overrides for 30 of those grades and ordered that the loans be reclassified as acceptable for purchase. The re-underwriting analysis for NHELI 2007-2 yielded 50 grade-3 loans (=16.2% of the sample), all of which RBS overrode.

RBS provided no objective record evidence to support these overrides. An RBS employee testified that the decision-making process for issuing a client override consisted of "review[ing] a loan file to see if there were compensating factors for exceptions" by "flip[ping] through the pages" for between "20 minutes" and "three hours" depending on whether he "thought it was important." Nomura II, 68 F.Supp.3d at 462. Farrell testified that he reviewed six of the overridden loans in NHELI 2007-1 and found them to have "sufficient compensating factors." Id. He justified the rest of the overrides in NHELI 2007-1 with similar reasoning.

2. Analysis

Section 12's reasonable care defense is available to any defendant who did not know and in the exercise of reasonable care could not have known of the material misstatement in the prospectus. See 15 U.S.C. § 77l(a)(2). Congress did not explicitly define the duty of reasonable care under Section 12. But one can discern the term's meaning by reference to related administrative guidance, non-statutory indicators of congressional intent, such as the section's legislative history and statutory context, and common-law principles. See Mohamad v. Palestinian Auth., 566 U.S. 449, 132 S.Ct. 1702, 1709, 182 L.Ed.2d 720 (2012) ("Congress is understood to legislate against a background of common-law adjudicatory principles." (quoting Astoria Fed. Sav. & Loan Ass'n v. Solimino, 501 U.S. 104, 108, 111 S.Ct. 2166, 115 L.Ed.2d 96 (1991))); Demarco v. Edens, 390 F.2d 836, 842 (2d Cir. 1968) (looking to common-law principles to define reasonable care under Section 12).

Section 12 imposes negligence liability. See NECA, 693 F.3d at 156. "Negligence, broadly speaking, is conduct that falls below the standard of what a reasonably prudent person would do under similar circumstances...." Fane v. Zimmer, Inc., 927 F.2d 124, 130 n.3 (2d Cir. 1991). "[I]t is usually very difficult, and often simply not possible, to reduce negligence to any definite rules; it is `relative to the need and the occasion,' and conduct which would be proper under some circumstances becomes negligence under others." W. Page Keeton et al., Prosser and Keeton on Torts § 31 at 173 (5th ed. 1984) (quoting Babington v. Yellow Taxi Corp., 250 N.Y. 14, 18, 164 N.E. 726 (1928) (Cardozo, C.J.)).

Courts have explored negligence liability for securities offerors in the analogous context of Section 11. See In re Software Toolworks Inc. (Software Toolworks), 50 F.3d 615, 621 (9th Cir. 1994); In re WorldCom, Inc. Sec. Litig. (WorldCom), 346 F.Supp.2d 628, 663 (S.D.N.Y. 2004) (Cote, J.). But see Glassman v. Computervision Corp., 90 F.3d 617, 628 (1st Cir. 1996) ("The law on due diligence is sparse...."). SEC guidance advises that "the standard of care under Section 12(a)(2) is less demanding than that prescribed by Section 11." Securities Offering Reform, SEC Release No. 75, 85 SEC Docket 2871, available at 2005 WL 1692642, at *79 (Aug. 3, 2005).44 Still, Section 11 law is persuasive in defining reasonable care under Section 12.45 See SEC Release No. 75, 2005 WL 1692642, at *79 ("[W]e believe that any practices or factors that would be considered favorably under Section 11, including pursuant to Rule 176, also would be considered as favorably under the reasonable care standard of Section 12(a)(2)."); H.R. REP. NO. 73-85, at 9 (1933) (discussing jointly the duties of care under Sections 11 and 12).

Section 11, like Section 12, imposes a negligence standard. See Herman & MacLean v. Huddleston, 459 U.S. 375, 383-84, 103 S.Ct. 683, 74 L.Ed.2d 548 (1983); NECA, 693 F.3d at 156. Section 11 achieves this by providing a defense to any underwriter defendant who "had, after reasonable investigation, reasonable ground to believe and did believe ... that the statements [at issue] 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). For a defendant's investigation to be reasonable, its actions must conform to those of "a prudent man in the management of his own property." 15 U.S.C. § 77k(c); see WorldCom, 346 F.Supp.2d at 663.

The measures a reasonably prudent person would take in the management of his property are context dependent. Under Section 12, they are a function of, inter alia, (1) the nature of the securities transaction, (2) the defendant's role in that transaction, (3) the defendant's awareness of information that might suggest a securities violation and its response(s) upon learning of such information, and (4) industry practices. See WorldCom, 346 F.Supp.2d at 674-77; 17 C.F.R. § 230.176 (listing relevant considerations in deciding whether an investigation was reasonable under Section 11).

The reasonable care standard adapts to the context of each transaction. The SEC has issued a rule regarding the due diligence review that issuers of asset-backed securities should conduct before making public offerings. See 17 C.F.R. § 230.193; Issuer Review of Assets in Offerings of Asset-Backed Securities, SEC Release No. 9176, 100 SEC Docket 706, available at 2011 WL 194494 (Jan. 20, 2011).46 The SEC requires issuers to adopt due diligence policies that provide reasonable assurance that the offering documents' descriptions of the assets are accurate in all material respects. See SEC Release No. 9176, 2011 WL 194494, at *6. Specific review standards depend on the type of product offered. See id. For RMBS, the SEC requires issuers to provide reasonable assurance of the truth of all information related to the supporting loans that is required to be in a prospectus or prospectus supplement, including representations of the loans' "credit quality and underwriting." Id. at *7. Sometimes that may require reviewing all of the supporting loans. But an RMBS issuer also may review a sample of the loans if the loan pool is so large that reviewing all of the loans is prohibitive and the sample is "representative of the pool." Id. at *6.

The nature of the defendant's position within a given transaction also affects the standard of care. See 2 THOMAS LEE HAZEN, THE LAW OF SECURITIES REGULATION § 7:45 (7th ed., 2016) ("Reasonable care imparts a sliding scale of standards of conduct...."). As Congress explained when it initially passed the Securities Act, "[t]he duty of care to discover varies in its demands upon participants in security distribution with the importance of their place in the scheme of distribution and with the degree of protection that the public has a right to expect." H.R. REP. No. 73-85, at 9. Those closest to the offered securities — issuers, for example — are more likely to come into contact with material information, and thus may be required to exercise more care to assure that disclosures are accurate. See Feit v. Leasco Data Processing Equip. Corp., 332 F.Supp. 544, 577-78 (E.D.N.Y. 1971). In an RMBS distribution, the depositor as the formal issuer, and the affiliated entities that control it, such as the sponsor and affiliated underwriters, occupy this position of closeness to the offered products. See H.R. REP. No. 73-85, at 12.

Unaffiliated underwriters are often the sole adversarial entities in a securities distribution. As a result, they assume a unique role. See Feit, 332 F.Supp. at 581-82. The Securities Act places upon underwriters "the primary responsibility for verifying the accuracy and completeness of information provided to potential investors." Chris-Craft Indus., Inc. v. Piper Aircraft Corp., 480 F.2d 341, 369-70 (2d Cir. 1973). That special responsibility guides the standard of care for underwriters under Section 12 mandates. See Sanders v. John Nuveen & Co., Inc., 619 F.2d 1222, 1228 n.12 (7th Cir.1980) ("The fact that [Section 12] does not expressly single out underwriters ... for a higher standard of liability does not mean that this status is irrelevant to determining what specific actions [an underwriter must] show to prove its exercise of reasonable care.").

Whether a defendant learns or should learn of alarming information that suggests a violation of the securities laws — so-called "red flags" — and how the defendant responds are perhaps the most important considerations in assessing reasonable care. See WorldCom, 346 F.Supp.2d at 679. Reasonable care requires a context-appropriate effort to assure oneself that no such red flags exist. If a defendant encounters red flags, reasonable care mandates that it examine them to determine whether the offering documents contain a material falsehood and, if so, to correct it. Cf. Lentell, 396 F.3d at 168 ("Inquiry notice ... gives rise to a duty of inquiry when the circumstances would suggest to an investor of ordinary intelligence the probability that [there has been a violation of the securities laws]." (internal quotation marks omitted) (quoting Levitt, 340 F.3d at 101)). An RMBS seller must conduct "further review" when "warranted in order to provide reasonable assurance that [the offering documents are] accurate in all material respects." SEC Release No. 9176, 2011 WL 194494, at *6.

Finally, industry standards and customs are highly persuasive in setting the standard of care, but they are not controlling. See In re City of New York, 522 F.3d at 285. As Judge Hand famously explained in The T.J. Hooper, in exceptional cases "a whole calling [or industry] may have unduly lagged in the adoption of new and available devices. It never may set its own tests, however persuasive be its usages. Courts must in the end say what is required." 60 F.2d 737, 740 (2d Cir. 1932). The reasonable care standard will not countenance an industry-wide "`race to the bottom' to set the least demanding standard to assess [its] conduct." SEC v. Dain Rauscher, Inc., 254 F.3d 852, 857 (9th Cir. 2001). Thus, particularly where "the industry was comprised of only a few participants who controlled the practice," id., and where industry practices have not previously survived judicial scrutiny, see Chasins v. Smith, Barney & Co., 438 F.2d 1167, 1171 (2d Cir. 1970), custom is less persuasive evidence of reasonable prudence. But see In re City of New York, 522 F.3d at 285 ("Courts will not lightly presume an entire industry negligent.").

In this case, no reasonable jury could find that Defendants exercised reasonable care. Nomura, as the sponsor, depositor, and occasional underwriter, was given access to the loans — and the loan files — prior to purchase and later owned the loans themselves. That uniquely positioned Nomura to know more than anyone else about the creditworthiness and underwriting quality of the loans. As a result, investors relied on Nomura's review of the loans and representations about the loans' likelihood to default. In making those representations, Nomura fell below the standard of conduct Section 12 requires.

Nomura could not be reasonably sure of the truth of any statements in the ProSupps regarding the loans' adherence to the underwriting guidelines. The single round of diligence Nomura conducted involved credit reviews for only a sample of the loans. At the direction of its Trading Desk, Nomura limited that sample to about 40% of the trade pool. Nomura then used a combination of ad hoc selections and LEVELS, the adverse sampling program, to compile its samples. These selection procedures chose a sample of the "riskiest" loans rather than a sample that was representative of the entire loan pool.

The criteria LEVELS used to identify "risky" loans was not tied to the loans' adherence to the underwriting guidelines. LEVELS relied solely on loan-tape information, such as loan-to-value and debt-to-income ratios, to form its adverse samples. These characteristics may be indicators of general credit risk, but Nomura provided no evidence whatever to suggest that they are indicators of the likelihood that a loan met the underwriting criteria. "As Kohout [later] explained at trial," LEVELS's singular reliance on the loan tape "made it impossible to select a sample based on a prediction of which loans were more likely to have `adverse' characteristics, such as a misstated LTV ratio or DTI ratio, an unreasonable `stated' income, or to find loans that deviated from the originator's underwriting guidelines." Nomura VII, 104 F.Supp.3d at 473.

The problems with Nomura's sample selection were compounded by its failure to conduct reliable credit and compliance reviews. The audit Nomura commissioned of its credit and compliance reviews, however, raised serious red flags about the efficacy of its due diligence procedures. Nomura learned that approximately 30% of a sample of 109 loans receiving a final grade of EV1 or EV2 after the loan-level reviews should have received an unacceptable grade of EV3 or no grade at all. There is no evidence that Nomura took any action to correct that deficiency in its procedures.

Similarly, the high kick-out rates for the trade pool samples observed by Nomura should have raised suspicions about whether its due diligence was reliable. Spagna considered a 7% to 8% kick-out rate to be standard and a 12% kick-out rate to be higher than normal, yet Nomura observed a 15.2% kick out rate for the trade pools at issue. In other words, Nomura's samples contained nearly double the normal amount of loans that failed credit or compliance review. A reasonable investor in that scenario would have upsized the sample to determine if this problem pervaded the entire trade pool. Nomura did not.

Nomura's SLG compilation procedures were also problematic. The Trading Desk grouped the loans into SLGs without any assistance from the Diligence Group. Nomura performed no review of the SLGs after they were compiled. The only due diligence the Trading Desk reviewed was a single-page summary describing diligence for the loan pool, attached to which was an express disclaimer that the information contained therein should not be taken as complete and accurate. Moreover, the Trading Desk's methodology for selecting loans broke the inferential chain between the results of its sample testing and the representations in the ProSupps. The ProSupps described the loans as SLGs, yet Nomura compiled SLGs using non-random and ad hoc selection procedures that turned on the trader's instincts about market demand. Despite its representations in PLS offering documents, in reality Nomura had no way to know the credit risk of any given SLG.47

RBS's conduct was no better. For NHELI 2006-HE3 and NHELI 2006-FM2, RBS relied entirely on Nomura's diligence. That did not adequately discharge RBS's responsibility as an underwriter to verify independently the representations in the offering documents. Spagna's conduct with regard to NHELI 2006-FM2 is a revealing example. Without RBS's knowledge, Spagna retroactively changed the pre-acquisition grades for 19 purchased loans from EV3 to EV2 before sending the due diligence reports to RBS. And when RBS asked Spagna about Nomura's due diligence, he "bullshit[ted]" them. Nomura II, 68 F.Supp.3d at 460. RBS was blind to these acts of malfeasance. See Nat'l Credit Union Admin. Bd., 2017 WL 411338, at *4-6; Mass. Mut. Life Ins., 110 F.Supp.3d at 301.

For NHELI 2007-1 and NHELI 2007-2, RBS conducted some diligence but not enough to meet the standard of reasonable care. RBS sampled just 5.8% of a group of loans from which Defendants composed the SLG backing the NHELI 2007-1 Certificate and just 6% of the loans in NHELI 2007-2 even though it believed the loans in the latter Securitization were "crap." Nomura II, 68 F.Supp.3d at 461 (internal quotation marks omitted). RBS compiled those samples in part using non-representative adverse selection. Its re-underwriting analyses revealed that =32% of the loans in NHELI 2007-1 and =16.2% of the loans in NHELI 2007-2 deserved a failing grade for credit or compliance review even after Nomura's pre-acquisition screening. But instead of requesting a larger sample to determine if this problem was consistent for the entire trade pool or further questioning Nomura about this issue, RBS overrode all, or nearly all, of those failing grades in short time periods — in the case of NHELI 2007-1 just over an hour. RBS provided no objective justification for any of those override decisions and only specific subjective justification for six. That conduct fell well below the standard of reasonable care.48

Defendants' primary contention on appeal is that their conduct could not be unreasonable as a matter of law because it conformed to industry practices at the time. They argue that LEVELS was an industry standard adverse selection software,49 most PLS sellers conducted only one round of pre-acquisition diligence, it was standard for PLS sellers to outsource loan-level diligence to third parties such as Clayton, and many PLS underwriters relied on the aggregator's diligence representations.

We are not persuaded a properly instructed jury could find Defendants' conduct reasonable based on these standards. This argument is tellingly limited. Defendants do not contend that every choice they made was in keeping with best practices in the PLS industry, nor do they suggest that their actions, on the whole, were consistent with industry customs. They pick and choose instances of conduct that they claim met the standards of the industry. A seller's scattershot compliance with industry custom does not deprive a plaintiff of a Section 12 remedy. That Defendants' use of sampling or LEVELS or a third-party vendor complied with industry customs does not mean their conduct taken as a whole was reasonable under the circumstances.

Moreover, our analysis is only informed by industry standards, not governed by them. See In re City of New York, 522 F.3d at 285. The RMBS industry in the lead up to the financial crisis was a textbook example of a small set of market participants racing to the bottom to set the lowest possible standards for themselves. See Dain Rauscher, 254 F.3d at 857. Accordingly, even if Defendants' actions on the whole complied with that industry's customs, they yielded an unreasonable result in this case.

Defendants also argue that use of adverse sampling cannot be unreasonable because the SEC has advised that asset due diligence may vary depending on the circumstances, in lieu of adopting a proposed rule that would require RMBS sellers to use representative samples in all cases. See SEC Release No. 9176, 2011 WL 194494, at *4, *6. This argument is not persuasive either. SEC's refusal to ban adverse sampling in all cases is not inconsistent with our holding that, in this particular case, Defendants' use of non-representative sampling contributed in part to a course of unreasonable conduct.

Finally, we have no doubt that, had they exercised reasonable care, Defendants could have learned that a material number of the loans were not originated in accordance with the underwriting guidelines. This is not a case where Defendants incorrectly forecasted a future occurrence or inaccurately assessed the future impact of a past event. The relevant information in this case was static and knowable when Defendants securitized the loans and wrote the ProSupps. At that time, the manner in which the loans were originated had already occurred — they had been issued either in accordance with the underwriting criteria or not. And it was possible for Defendants, who owned the loans and regularly conducted business with third-party vendors that perform re-underwriting analyses, to learn whether they were.

D. Jury Trial

After the FHFA withdrew its Section 11 claim, the District Court conducted a bench trial on the remaining Section 12(a)(2), Section 15, and analogous Blue Sky claims. See Nomura IV, 68 F.Supp.3d at 496-98.50 Defendants contend that the bench trial violated their right to a jury trial under the Seventh Amendment.

The Seventh Amendment to the United States Constitution preserves the right of any party to a civil action to compel a jury trial in "Suits at common law." "The phrase `Suits at common law' refers to `suits in which legal rights were to be ascertained and determined, in contradistinction to those where equitable rights alone were recognized, and equitable remedies were administered.'" Eberhard v. Marcu, 530 F.3d 122, 135 (2d Cir. 2008) (emphasis in original) (quoting Granfinanciera, S.A. v. Nordberg, 492 U.S. 33, 41, 109 S.Ct. 2782, 106 L.Ed.2d 26 (1989)). Determining whether an action is a "Suit[] at common law" requires two steps. Id. The first assesses "whether the action would have been deemed legal or equitable in 18th century England." Id. (internal quotation marks omitted) (quoting Germain v. Conn. Nat'l Bank, 988 F.2d 1323, 1328 (2d Cir. 1993)). The second and "more important" step asks "whether `the remedy sought ... is legal or equitable in nature.'" Id. (alteration in original) (quoting Granfinanciera, 492 U.S. at 42, 109 S.Ct. 2782).

For years, there was little doubt that an action under Section 12(a)(2) was not a "Suit[] at common law," id., within the meaning of the Seventh Amendment. A Section 12 action operates much like an 18th century action at equity for rescission, which extinguished a legally valid contract that had to "be set aside due to fraud, mistake, or for some other reason." 12A C.J.S. CANCELLATION OF INSTRUMENTS § 1 (2017); see Randall v. Loftsgaarden, 478 U.S. 647, 655, 106 S.Ct. 3143, 92 L.Ed.2d 525 (1986) (describing Section 12's remedy of "rescission" upon "prospectus fraud").51 The Supreme Court and this Court have recognized that a Section 12(a)(2) action is the Securities Act-equivalent of equitable rescission. See Gustafson, 513 U.S. at 576, 115 S.Ct. 1061 ("[Section] 12(2) ... grant[s] buyers a right to rescind...."); Pinter v. Dahl, 486 U.S. 622, 641 n.18, 108 S.Ct. 2063, 100 L.Ed.2d 658 (1988) ("Section 12 was adapted from common-law (or equitable) rescission...."); Deckert v. Indep. Shares Corp., 311 U.S. 282, 288, 61 S.Ct. 229, 85 S.Ct. 189 (1940) (concluding that a Section 12(a)(2) claim "states a cause for equitable relief"); Royal Am. Managers, Inc. v. IRC Holding Corp., 885 F.2d 1011, 1019 n.4 (2d Cir. 1989) ("An equitable claim such as rescission [under Section 12(a)(2)] is for the court, not the jury, to decide."). Commentators have also consistently analogized an action under Section 12(a)(2) to equitable rescission. See, e.g., 69A AM. JUR. 2D Securities Regulation — Federal § 982 (2016); 2 HAZEN, THE LAW OF SECURITIES REGULATION § 7:56; Harry Shulman, Civil Liability and the Securities Act, 43 YALE L.J. 227, 243-44 (1933).

In 1995, Congress added the loss causation affirmative defense to Section 12(a)(2). Private Securities Litigation Reform Act of 1995, Pub. L. No. 104-67, § 105(3), 109 Stat. 737, 757 (codified at 15 U.S.C. § 77l(b)). Defendants' primary argument is that the amendment altered the nature of the Section 12(a)(2) remedy to that of damages for the injury arising from the false statement — a decidedly legal remedy — and therefore a Section 12 action now triggers the Seventh Amendment jury trial right.

There is some credence to Defendants' position. At 18th century common law, equitable rescission required the seller to refund the buyer the full original purchase price in exchange for the purchased item, regardless of its present value. See Pinter, 486 U.S. at 641 n.18, 108 S.Ct. 2063; Lyon v. Bertram, 61 U.S. 149, 154-55, 20 How. 149, 15 S.Ct. 847 (1857) ("`Where a contract is to be rescinded at all, it must be rescinded in toto, and the parties put in statu quo.'" (quoting Hunt v. Silk (1804) 5 East 449, 452 (Lord Ellenborough, C.J.))). In other words, the seller bore the risk of depreciation unrelated to the misrepresentation. Section 12(a)(2) with a loss causation defense shifts the risk burden to the buyer by authorizing the seller to refund the original purchase price less any reduction in the item's present value not attributable to a material misstatement. See Iowa Pub. Emps'. Ret. Sys., 620 F.3d at 145.

Furthermore, in the Section 10(b) context, this Court has "described loss causation in terms of the tort-law concept of proximate cause." Lentell, 396 F.3d at 172; see also Nomura VII, 104 F.Supp.3d at 585 ("`Loss causation is the causal link between the alleged misconduct and the economic harm ultimately suffered by the plaintiff ... [and] is related to the tort law concept of proximate cause.'" (alterations in original) (quoting Lattanzio v. Deloitte & Touche LLP, 476 F.3d 147, 157 (2d Cir. 2007))). Proximate cause generally defines the scope of a defendant's legal liability. CSX Transp., Inc. v. McBride, 564 U.S. 685, 692-93, 131 S.Ct. 2630, 180 L.Ed.2d 637 (2011); Lattanzio, 476 F.3d at 157. When a judgment imposes personal legal liability on a defendant, even occasionally in the context of a restitution claim, it can create a legal remedy. See Great-West Life & Annuity Ins. Co. v. Knudson (Knudson), 534 U.S. 204, 213-14, 122 S.Ct. 708, 151 L.Ed.2d 635 (2002).

Nevertheless, the addition of the loss causation defense did not transform Section 12(a)(2)'s equitable remedy into a legal one. The limited degree to which the modern Section 12(a)(2) remedy differs from common-law rescission does not change the fact that, fundamentally, it is equitable relief. Section 12(a)(2) has never provided exactly the same relief as 18th century equitable rescission. Section 12(a)(2) has traditionally been more buyer-friendly than its common-law counterpart because it authorizes recovery even after the buyer no longer owns the security at issue. See Pinter, 486 U.S. at 641 n.18, 108 S.Ct. 2063; Shulman, supra, at 244. The availability of an alternative damages remedy never stood as a barrier to considering Section 12(a)(2)'s rescission-like remedy equitable for purposes of the Seventh Amendment. Nor does the loss causation defense, which merely tilts the balance of equities in the modern Section 12(a)(2) remedy slightly back toward sellers.

Likewise, our suggestion in the Section 10(b) context that loss causation is akin to proximate cause does not mean that Section 12(a)(2) with a loss causation defense necessarily provides a legal claim. Equitable rescission permits a court to order "the nullification of a transfer of property between the claimant and the defendant ... and ... a mutual accounting in which each party pays for benefits received from the other in consequence of the underlying exchange and its subsequent reversal." RESTATEMENT (THIRD) OF RESTITUTION AND UNJUST ENRICHMENT § 54 cmt. a. Loss causation in Section 12(a)(2) serves the latter function — it is a mutual accounting that prevents the buyer from reaping an unjust benefit at the expense of the seller. This restores the parties to the status quo ante the securities transaction at issue while ensuring that the terms of the rescission are just (in Congress's view), a hallmark of equitable recessionary relief. See Marr v. Tumulty, 256 N.Y. 15, 22, 175 N.E. 356 (1931) (Cardozo, C.J.).

Defendants' further arguments come up short. As an initial matter, none of the Defendants' remaining arguments rely on changes in the law that would upset the long-established consensus that Section 12(a)(2) is an equitable claim that authorizes equitable relief. See Pinter, 486 U.S. at 641 n.18, 108 S.Ct. 2063. Moreover, Defendants' arguments are unpersuasive on the merits.

Defendants contend that because Section 11 and Section 12 claims are similar and Section 11 claims are considered legal for purposes of the Seventh Amendment, Section 12 claims ought to be considered legal too. While Sections 11 and 12(a)(2) are "Securities Act siblings with roughly parallel elements," Morgan Stanley, 592 F.3d at 359, they are not identical twins when it comes to the nature of relief each authorizes; indeed, sometimes they are quite different. See id. ("Section 12(a)(2) [and Section 11] provide[] similar redress...." (emphasis added)). Section 12 authorizes two forms of relief: A buyer who retains ownership over the security may sue under Section 12 for equitable rescission, which limits recovery to "the consideration paid for such security." 15 U.S.C. § 77l(a). A buyer who no longer owns the security may sue under Section 12 for "damages," id., "the classic form of legal relief," Knudson, 534 U.S. at 210, 122 S.Ct. 708 (internal quotation mark omitted) (quoting Mertens v. Hewitt Assocs., 508 U.S. 248, 255, 113 S.Ct. 2063, 124 L.Ed.2d 161 (1993)). See also Wigand, 609 F.2d at 1035 ("If [a Section 12(a)(2)] plaintiff owns the stock, he is entitled to rescission...."). Section 11 authorizes only legal "damages." 15 U.S.C. § 77k(e).

Defendants argue further that since a plaintiff who no longer owns the security at issue is entitled to a legal remedy under Section 12(a)(2), the remedy for a plaintiff who still owns the security must be of the same nature. In Defendants' view, a plaintiff should not have the power to manipulate a seller's constitutional right to a jury trial by choosing, through the act of selling or retaining the security, whether the suit will sound in law or in equity. Assuming Defendants are correct that an action for money damages under Section 12(a)(2) is a "Suit[] at common law,"52 Eberhard, 530 F.3d at 135, this case does not involve that situation. Here, the FHFA still owns and can physically return the Certificates as it would be required to do on an equitable rescission claim. Indeed, in issuing its final judgment, the District Court ordered the FHFA to "deliver" the Certificates to Defendants in exchange for the amounts recoverable. Special App. 365-67. Moreover, Defendants' contention that a buyer should not have the power to decide the form of relief sought overlooks the express language of Section 12(a)(2), which authorizes the buyer to sue "either at law or in equity." 15 U.S.C. § 77l(a).

Finally, Defendants urge that, at common law, a court of equity could issue an order only against persons who actually "possessed the funds in question and thus were ... unjustly enriched." Pereira v. Farace, 413 F.3d 330, 339 (2d Cir. 2005). Defendants argue that the non-underwriter Defendants cannot be subject to equitable rescission because they did not in fact sell the Certificates nor did they receive funds from the GSEs in exchange for the Certificates. The Supreme Court has made clear that "there is no reason to think that Congress wanted to bind itself to the common-law notion of the circumstances in which rescission [under Section 12(a)(2)] is an appropriate remedy." Pinter, 486 U.S. at 647 n.23, 108 S.Ct. 2063. "Congress, in order to effectuate its goals, chose to impose [rescission-like] relief on any defendant it classified as a statutory seller, regardless of the fact that such imposition was somewhat inconsistent with the use of rescission at common law." Id. As discussed further below, all of the Defendants were statutory sellers.

Accordingly, we reaffirm that, even after the addition of the loss causation defense, a Section 12(a)(2) action allows for equitable relief where the plaintiff still owns the securities and the remedy sought is literal rescission. Such an action is not a "Suit[] at common law," Eberhard, 530 F.3d at 135, for purposes of the Seventh Amendment.53

II. Trial Decision54

A. Section 12(a)(2) Claims

1. Statutory Sellers

Defendants contest the District Court's finding that NAAC and NHELI, the PLS depositors for the transactions at issue, were statutory sellers for purposes of Section 12(a)(2). See Nomura VII, 104 F.Supp.3d at 554-55; UBS I, 858 F.Supp.2d at 333-34. Defendants argue that PLS depositors cannot be statutory sellers because they have no direct involvement in passing title in PLS to buyers.55

Section 12(a)(2) requires proof that the defendant is a "statutory seller" within the meaning of the Securities Act. Pinter, 486 U.S. at 641-42, 108 S.Ct. 2063; see 15 U.S.C. § 77l(a)(2).56 The Securities Act does not define "statutory seller," however. See Pinter, 486 U.S. at 642, 108 S.Ct. 2063. Judicial precedent has settled that an entity is a statutory seller if it "(1) `passed title, or other interest in the security, to the buyer for value,' or (2) `successfully solicited the purchase of a security, motivated at least in part by a desire to serve [its] own financial interests or those of the securities' owner.'" Morgan Stanley, 592 F.3d at 359 (brackets omitted) (quoting Pinter, 486 U.S. at 642, 647, 108 S.Ct. 2063). SEC Rule 159A provides that, for purposes of Section 12(a)(2), an "issuer" in "a primary offering of securities" shall be considered a statutory seller. 17 C.F.R. § 230.159A(a). The Securities Act in turn defines "issuer" to include "the person or persons performing the acts and assuming the duties of depositor." 15 U.S.C. § 77b(a)(4). SEC Rule 191 further clarifies that "[t]he depositor for ... 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." 17 C.F.R. § 230.191(a).

The combination of this statutory provision and administrative direction makes clear that PLS depositors, such as NAAC and NHELI, are statutory sellers for purposes of Section 12(a)(2). Each is a "depositor for ... asset-backed securities," specifically RMBS. See 17 C.F.R. § 230.191. PLS depositors are thus "issuers." See 15 U.S.C. § 77b(a)(4). And, as "issuers," PLS depositors fall within the definition of statutory seller. See 17 C.F.R. § 230.159A.

Defendants' only avenue of attack on appeal is to contest the validity of Rules 159A and 191. "[A]mbiguities in statutes within an agency's jurisdiction to administer are delegations of authority to the agency to fill the statutory gap in reasonable fashion." Nat'l Cable & Telecomms. Ass'n v. Brand X Internet Servs. (Brand X), 545 U.S. 967, 980, 125 S.Ct. 2688, 162 L.Ed.2d 820 (2005). "Chevron requires a federal court to accept [a federal] agency's construction of [a] statute" so long as the statute is ambiguous and the agency's interpretation is reasonable. Id. (citing Chevron, U.S.A., Inc. v. Nat. Res. Def. Council, 467 U.S. 837, 843-844 & n.11, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984)). "Only a judicial precedent holding that [a] statute unambiguously forecloses [an] agency's interpretation ... displaces a conflicting agency construction." Id. at 982-83, 125 S.Ct. 2688.

Defendants do not and cannot argue that SEC Rules 159A and 191 are unreasonable. Instead, they cite Pinter v. Dahl as "a judicial precedent holding that" the Securities Act "unambiguously forecloses" SEC Rules 159A and 191. See Brand X, 545 U.S. at 982-83, 125 S.Ct. 2688. We disagree. Pinter actually stands for the proposition that the Securities Act is ambiguous as to the definition of statutory seller. See 486 U.S. at 642-47, 108 S.Ct. 2063. Pinter acknowledged that, given the lack of clear guidance from Congress, statutory seller must include "[a]t the very least ... the owner who passed title, or other interest in the security, to the buyer for value." Id. at 642, 108 S.Ct. 2063. But it also observed that Section 12 "is not limited to persons who pass title" for value and that related statutory terms "are expansive enough" for Section 12 "to encompass the entire selling process." Id. at 643, 108 S.Ct. 2063 (quoting United States v. Naftalin, 441 U.S. 768, 773, 99 S.Ct. 2077, 60 L.Ed.2d 624 (1979)). The only element of the statutory seller provision Pinter found unambiguous is that "Congress did not intend to impose [Section 12] rescission... on a person who urges the purchase but whose motivation is solely to benefit the buyer." Id. at 647, 108 S.Ct. 2063.

SEC Rules 159A and 191 locate depositors within the selling process for PLS. As the District Court explained, depositors play an essential role in PLS distribution schemes — at the direction of the PLS sponsor, they "purchase the loans ... and deposit them in a trust," which "creates a true sale of the assets, thereby protecting certificate-holders against the risk of a subsequent bankruptcy by the sponsor." Nomura VII, 104 F.Supp.3d at 463. Rules 159A and 191 therefore accord with Pinter's understanding of the expansive definition of statutory seller. See 486 U.S. at 643, 108 S.Ct. 2063.

2. Falsity

Defendants contest the District Court's finding that the underwriting guidelines statements were false.

Section 12(a)(2) requires proof that the prospectus at issue contains at least one "untrue statement of a ... fact or omit[ted] to state a ... fact necessary in order to make the statements, in the light of the circumstances under which they were made, not misleading." 15 U.S.C. § 77l(a)(2); see Morgan Stanley, 592 F.3d at 359.57 "[W]hether a statement is `misleading' depends on the perspective of a reasonable investor: The inquiry ... is objective." Omnicare, Inc. v. Laborers Dist. Council Constr. Indus. Pension Fund, ___ U.S. ___, 135 S.Ct. 1318, 1327, 191 L.Ed.2d 253 (2015) (discussing misleading omissions in the context of Section 11). The falsity inquiry "requires an examination of `defendants' representations, taken together and in context.'" Morgan Stanley, 592 F.3d at 366 (quoting DeMaria v. Andersen, 318 F.3d 170, 180 (2d Cir. 2003)). "The literal truth of an isolated statement is insufficient." Id. "[W]hen an offering participant makes a disclosure about a particular topic, whether voluntary or required, the representation must be `complete and accurate.'" Id. (quoting Glazer v. Formica Corp., 964 F.2d 149, 157 (2d Cir. 1992)). Both false statements of fact and false statements of opinion are actionable under Section 12(a)(2). See Omnicare, 135 S.Ct. at 1325-27.

a. Factual Summary

This case turns on the following statement, which appeared in each of the ProSupps: "The Mortgage Loans [in the SLGs] have been purchased by the seller from various banks, savings and loan associations, mortgage bankers and other mortgage loan originators and purchasers of mortgage loans in the secondary market, and were originated generally in accordance with the underwriting criteria described in this section." J.A. 6884 (emphasis added).58

Each ProSupp described that underwriting process:

Id. at 6884-85.

Each ProSupp also included a warning regarding possible deviations from the underwriting guidelines:

Id. at 6886.

NHELI 2007-3 contained an additional warning regarding originator ResMAE:

Id. at 9069 (emphasis added).

b. Procedural Summary

The District Court examined the above language in detail.59 The court interpreted the underwriting guidelines statements as asserting that the supporting loans, with a few immaterial exceptions, were originated in accordance with the underwriting guidelines the originators used to issue the loans.

The court then set out to determine whether in fact the loans in the SLGs were originated generally in accordance with the underwriting guidelines. (As the underwriting guidelines statement is unquestionably one of provable fact, the District Court did not need to consider Defendants' subjective belief in, inquiry into, or knowledge of the truthfulness of the statement. See Omnicare, 135 S.Ct. at 1325-26.) The court relied on the testimony of one of the FHFA's experts, Robert Hunter, a consultant with "expertise in residential loan credit issues." Nomura VII, 104 F.Supp.3d at 456. Hunter conducted a forensic re-underwriting of 723 sample loans including "100 or close to 100 ... loans for six of the seven SLGs, and 131 ... loans for the relevant SLG in NAA 2005-AR6." Id. at 522.60

Hunter's review entailed comparing "the loan file for each loan to the originator's guidelines." Id. at 522. The parties stipulated for the most part to an applicable set of guidelines that were representative of the originators' guidelines at the time the loans were issued. When they did not, Hunter re-underwrote the sample loans using "originators' guidelines that were dated between 30 to 90 days prior to the closing of the loan." Id. When those were not available, Hunter analyzed the loans using what he styled "minimum industry standards." Id. Hunter's "industry standards" were "the most lenient standards employed for subprime and Alt-A loans between 2002 and 2007" drawn "from the many guidelines he examined and from his professional experience." Id. Hunter also used these industry standards to supplement gaps in the originators' guidelines.

Hunter concluded that approximately 66% of the sample loans contained material deviations from the originators' underwriting criteria that negatively affected the creditworthiness of the loans. Id. at 523. Hunter also found that "the level of underwriting defects in the [s]ample was so severe that it was unlikely that any of the loans in the seven SLGs ... was actually free of defects," id. at 541, although some of the defects in the sample were immaterial to credit risk.

Defendants called Michael Forester, founder of "a regulatory compliance, loan review, and internal audit services firm," id. at 457, as an expert to contest Hunter's findings. After reviewing Forester's analysis in detail, the District Court concluded that many of his complaints about Hunter's work were "essentially irrelevant." Id. at 525. The court also rejected Defendants' objections to Hunter's analysis.

The District Court ultimately credited the bulk of Hunter's analysis. See id. at 531. The court, acting as a fact-finder and guided by the expert testimony, conducted its own loan-by-loan underwriting analysis. The court confirmed that, as a "conservative" measurement, at least 45% of the loans in each SLG "had underwriting defects that materially affected credit risk." Id. at 533. As a result, it found that the ProSupps' descriptions of the supporting loans "as having been `originated generally in accordance' with originators' guidelines" were false. Id.

c. Analysis

On appeal, Defendants contend that the District Court misinterpreted the underwriting guidelines statements. They also argue that the District Court improperly credited Hunter's analysis. Neither argument is persuasive.

1. The District Court's Interpretation of the Underwriting Guidelines Statements61

Defendants attack the District Court's interpretation of the ProSupps on four grounds. First, they contend the District Court misinterpreted the phrase "the underwriting criteria described in this section" as referring to the underwriting criteria the originators used in issuing the loans. Defendants argue that the ProSupps meant to refer to the underwriting criteria described in the ProSupps themselves.62 Because the District Court and Hunter re-underwrote the sample loans according to the originators' guidelines, Defendants conclude, their findings are fundamentally flawed.

This argument makes no sense. Defendants urge us to read the ProSupps as stating that the loans in the SLGs "were originated" in accordance with underwriting guidelines that the PLS sellers wrote after purchasing and securitizing the loans — that is, after the loans were originated. Of course, loans cannot be originated in accordance with guidelines that do not exist until after their creation. In a similar vein, the principal reason why the six later-issued ProSupps included descriptions of the underwriting guidelines was that SEC Regulation AB requires RMBS sponsors in their offering documents to describe "the ... underwriting criteria used to originate ... pool assets." 17 C.F.R. § 229.1111(a)(3) (emphasis added).63 It would make little sense to read the ProSupps as stating that guidelines written after loan origination were "used to originate" the loans.

Defendants' own actions belie their argument. When Nomura hired Clayton and AMC to conduct pre-acquisition credit and compliance reviews, Nomura instructed it to compare the loan files against the originators' underwriting guidelines. See Nomura II, 68 F.Supp.3d at 451. Furthermore, trial testimony from Nomura employees and others confirms that Defendants, other RMBS issuers and underwriters, as well as Moody's, S & P, and Fitch all understood the underwriter guidelines assertion in the ProSupps to refer to originators' guidelines. E.g., J.A. 4392, 4491, 5355, 6295-96, 6299-300.

Second, Defendants argue that the ProSupps merely describe the procedures the originators' used to issue the underlying loans, rather than promise that the loans met the originators' guidelines criteria. It would have been "meaningless" to promise compliance with that criteria, Defendants contend, because "investors did not know what those guidelines said." Nomura's Br. 39.

The central flaw in this argument is that it is a-textual. The ProSupps affirm that the loans "were originated ... in accordance with the underwriting criteria." Defendants' argument reads the word "criteria" out of that sentence.

Moreover, it would not be meaningless to read the ProSupps as promising that the loans complied with the underwriting guidelines, regardless of whether the reader is familiar with the details of those guidelines. See ACE Sec. Corp., Home Equity Loan Tr., Series 2006-SL2 v. DB Structured Prods., Inc., 25 N.Y.3d 581, 596, 36 N.E.3d 623 (2015) (observing that PLS sponsors generally "warrant[] certain characteristics of the loans"). PLS consumers and the Credit-Rating Agencies — the primary audience for the ProSupps — considered it important that a sponsor warrant in offering documents that loans in the SLGs met the originators' underwriting criteria. This affirmed that the loans in the SLGs survived the gauntlet of the originators' underwriting reviews for creditworthiness, which bore directly on the loans' risk of default. The statement also assured investors that Defendants, through their diligence departments, independently checked that loans satisfied the originators' guidelines criteria. A mere description of the origination process would not accomplish that effect.

Third, Defendants argue that the word "generally" — as in, the loans "were originated generally in accordance with the underwriting criteria" — put readers of the ProSupps on notice that loans in the SLGs may deviate materially from the underwriting guidelines. The District Court, by contrast, interpreted "generally" to warn only that the SLGs may contain loans with "certain immaterial exceptions" to the underwriting guidelines. Nomura VII, 104 F.Supp.3d at 563 (emphasis added; internal quotation marks omitted) (quoting Nomura II, 68 F.Supp.3d at 485).

We agree with the District Court. Defendants' interpretation of "generally" would render the underwriting guidelines statement essentially meaningless. As noted above, readers of the ProSupps looked to this representation for an affirmation that the loans met the underwriting criteria. They would find cold comfort in a promise that contained the significant hedge Defendants urge. Furthermore, Defendants' interpretation of "generally" is undermined by the view of their own expert, Forester, who testified:

J.A. 6125.64

Fourth, Defendants argue that the District Court failed to accord proper weight to the explicit warning in the ProSupp for NHELI 2007-3 that ResMAE's weak "financial condition ... at the time of origination may have ... adversely affected its ability to originate mortgage loans in accordance with its customary standards." J.A. 9069. They argue that this specific hedge superseded the more general statements about the quality of the supporting loans writ large. See Omnicare, 135 S.Ct. at 1330 ("[A]n investor reads each statement... in light of all its surrounding text, including hedges....").

The problem with this argument is that the warning was too equivocal to hedge adequately against the ProSupps' later statements regarding compliance with underwriting guidelines. The vague warning that ResMAE's bankruptcy "may have ... adversely affected its ability to originate mortgage loans in accordance with its customary standards" was insufficient to put the reader on notice that a critical mass — nearly 50% — of the loans in the pertinent SLG were not originated properly. J.A. 9069. Furthermore, despite the warning the ProSupp affirmed that ResMAE "fully reviews each loan to determine whether [its underwriting] guidelines ... are met." Id. at 9113. That watered down any of the marginal ameliorative effect the ProSupp's earlier warning might have had.

2. The District Court's Falsity Findings65

Defendants also challenge the District Court's crediting of Hunter's expert testimony and finding based thereon that at least 45% of the loans in the SLGs were originated with underwriting defects.

Their arguments, at best, marginally undercut the substance of Hunter's analysis.66 We find in them no basis to second guess the District Court's adoption of Hunter's findings.

Defendants further argue that it was improper for the District Court, which lacks the expertise of Hunter and Forester, to conduct its own confirmatory re-underwriting analysis. We disagree. The court conducted this analysis in its capacity as fact-finder. A fact-finder is not required to make a binary choice between adopting an expert's conclusion in full or rejecting it entirely. See United States v. Duncan, 42 F.3d 97, 101 (2d Cir. 1994) (explaining that expert testimony should not "tell the jury what result to reach" but "aid the jury in making a decision") (emphasis in original). Furthermore, any error the District Court committed in crediting only a portion of Hunter's testimony would be harmless. See 28 U.S.C. § 2111. The court made clear that "[i]f limited to the stark choice between Hunter's expert testimony and Forester's, [it] would unhesitatingly accept Hunter's." Nomura VII, 104 F.Supp.3d at 531.67

For the foregoing reasons, Nomura offers no basis to reverse the District Court's finding that the ProSupps' underwriting guidelines assertion was false.

3. Materiality

Defendants contest the District Court's finding that the underwriting guidelines statements were material.

Section 12(a)(2) requires proof that each false statement or omission was material. See 15 U.S.C. § 77l(a)(2); Morgan Stanley, 592 F.3d at 359. Whether a statement or omission is material is an objective, totality-of-the-circumstances inquiry. TSC Indus., Inc. v. Northway, Inc., 426 U.S. 438, 445, 449, 96 S.Ct. 2126, 48 L.Ed.2d 757 (1976). A material fact is one that "assume[s] actual significance" for a reasonable investor deciding whether to purchase the security at issue, but it need not be outcome-determinative. Id. at 449, 96 S.Ct. 2126; see Folger Adam Co. v. PMI Indus., Inc., 938 F.2d 1529, 1533 (2d Cir. 1991). "In the Second Circuit," a statement or omission is material "if a reasonable investor would view [it] as `significantly altering the "total mix" of information made available.'" Stadnick v. Vivint Solar, Inc., 861 F.3d 31, 36 (2d Cir. 2017) (brackets omitted) (quoting TSC Indus., Inc., 426 U.S. at 449, 96 S.Ct. 2126); see Basic, 485 U.S. at 231-32, 108 S.Ct. 978.

Here, the District Court easily found that the ProSupps' underwriting guidelines statements were material. Nomura VII, 104 F.Supp.3d at 557-59, 570-73.68 The court began by presuming materiality for any description in the ProSupps that deviated by 5% or more from the loans' true characteristics. See id. at 558. It drew the 5% figure from two sources: First, five of the ProSupps promised that Defendants would issue supplementary disclosures in the event that "any material pool characteristic differs by 5% or more from the description in this [ProSupp]." Id.; see also Asset-Backed Securities, SEC Release No. 8518, 84 SEC Docket 1624, available at 2004 WL 2964659, at *235 (Dec. 22, 2004) (requiring supplemental disclosure "if any material pool characteristic of the actual asset pool at the time of issuance of the asset-backed securities differs by 5% or more ... from the description of the asset pool in the prospectus"). Second, SEC administrative guidance, which we have repeatedly cited with approval, counsels that 5% falsity for statements in offering documents may provide "a preliminary assumption" of materiality. SEC Staff Accounting Bulletin No. 99, 64 Fed. Reg. 45,150, 45,151 (Aug. 19, 1999); see Litwin v. Blackstone Grp., L.P., 634 F.3d 706, 717 (2d Cir. 2011); ECA, Local 134 IBEW Joint Pension Tr. of Chi. v. JP Morgan Chase Co., 553 F.3d 187, 197-98 (2d Cir. 2009); Ganino v. Citizens Utils. Co., 228 F.3d 154, 163-64 (2d Cir. 2000). The court found that the underwriting guidelines statements far exceeded that threshold, as at least 45% of the loans in the SLGs did not adhere to the originators' underwriting criteria. Nomura VII, 104 F.Supp.3d at 571. The court then confirmed its presumption of materiality by demonstrating how loans that do not adhere to underwriting criteria have higher default rates, and as a result, affect a reasonable investor's view of the value of PLS supported by such loans. Id.

On appeal, Defendants raise five challenges to the District Court's materiality analysis — one procedural, two substantive, and two evidentiary.69 We address each in turn.

a. Procedural Challenge: Use of a Numerical Threshold

Defendants argue that the District Court employed a legally erroneous process for deciding materiality because it relied in part on a numerical threshold. See Nomura VII, 104 F.Supp.3d at 558.

Although "we have consistently rejected a [purely] formulaic approach to assessing the materiality of an alleged misrepresentation," Hutchison v. Deutsche Bank Sec. Inc., 647 F.3d 479, 485 (2d Cir. 2011) (alteration omitted) (quoting Ganino, 228 F.3d at 162), we have permitted courts to conduct materiality analyses that are partially quantitative, see Litwin, 634 F.3d at 717. A numerical threshold is no substitute for a fulsome materiality analysis that also considers qualitative factors, but it can provide "`a good starting place for assessing the materiality of [an] alleged misstatement.'" Hutchison, 647 F.3d at 487 (quoting ECA, Local 134 IBEW Joint Pension Trust of Chi., 553 F.3d at 204); see also id. at 485. Indeed, an "integrative" materiality analysis will consider both quantitative factors and qualitative factors to determine whether a reasonable investor would have considered the misstatement or omission significant in making an investment decision. Litwin, 634 F.3d at 717.

The District Court in this case did exactly what we require. The court began with a reasonable quantitative analysis, using 5% falsity as a threshold for materiality. See Nomura VII, 104 F.Supp.3d at 558. The court then turned to qualitative factors. It found "overwhelming, and essentially undisputed, evidence that" the ProSupps' false underwriting guidelines statements "would be viewed by the reasonable PLS investor as significantly altering the total mix of information available." Id. at 570. Indeed, Defendants' own witnesses agreed that, as a general matter, adherence to underwriting criteria is a reliable indicator of mortgage loan default rates, and the return for a PLS certificate is a function of the degree to which such loans are repaid. The court therefore concluded that a reasonable investor deciding whether to invest in PLS would consider the underwriting guidelines statements crucial to his or her investment decision. See id. at 570-71. The court buttressed its qualitative materiality conclusion by noting that defense counsel admitted in summation that the supporting loans' rate of adherence to the underwriting guidelines "could be material to an investor." Id. at 571 n.185.

The District Court's opinion is a textbook example of an integrative materiality analysis that considers "both quantitative and qualitative factors." See Litwin, 634 F.3d at 717 (internal quotation marks omitted) (quoting SEC Staff Accounting Bulletin No. 99, 64 Fed. Reg. at 45,151). We find no legal error in the court's use of a numerical threshold to inform its decision.

b. Substantive Challenge: The Trade Date

Defendants challenge the substance of the District Court's materiality decision first on the ground that none of the ProSupps' statements could have been material because the GSEs did not receive the ProSupps until after the so-called "trade dates."

1. Factual Summary

The Securities Act requires virtually every written offer of securities to qualify as a prospectus under Section 10. 15 U.S.C. § 77e(b)(1). Section 10 provides for two types of permissible prospectuses. The default type is a written offer that meets intensive disclosure requirements listed in Section 10(a), sometimes called a "Section 10(a) prospectus." Id. § 77j(a); see 17 C.F.R. § 229.1100 et seq. Alternatively, Section 10(b) permits the SEC to promulgate rules expanding the definition of a Section 10 prospectus to include offerings that "omit[] in part or summarize[] information" required by Section 10(a), sometimes called a "Section 10(b) prospectus." 15 U.S.C. § 77j(b).70

For years after the passage of the Securities Act, the SEC did not promulgate any rules pursuant to Section 10(b). During that time, every written offer of securities needed to comply with the detailed requirements of Section 10(a). See FHFA v. Bank of Am. Corp., No. 11cv6195, 2012 WL 6592251, at *3-4 (S.D.N.Y. Dec. 18, 2012).

In 2005, the SEC invoked its Section 10(b) power for the first time when it promulgated Rule 164 and associated rules. These rules liberalize the offering process by permitting certain issuers to make initial written offers of securities using "free writing prospectuses." 17 C.F.R. §§ 230.164, 230.405; see Securities Offering Reform, SEC Release No. 75, 2005 WL 1692642, at *37-38. Free writing prospectuses may be used only if, inter alia, (1) the offered security is subject to a filed registration statement and to a base prospectus, 17 C.F.R. § 230.433, and (2) the issuer transmits a Section 10(a) prospectus to the SEC "no later than the second business day following ... the date of the determination of the offering price" of the security, id. § 230.424(b)(5). The information in a free writing prospectus and the information in the final Section 10(a) prospectus "shall not conflict." Id. § 230.433(c)(1).

Defendants sold the Certificates at issue here in a fluid process that relied on the use of free writing prospectuses. They contacted GSE traders to offer a PLS certificate sale, and if a trader was interested, transmitted a free writing prospectus containing some (but not all) of the information regarding the loans in the SLG. After reviewing the free writing prospectus, the GSE trader and Defendants made mutual commitments to purchase and to sell the Certificate described in it. The date of this commitment is known as the "trade date."

Within roughly a month following the trade date, the GSE transferred payment to Defendants, who in turn transferred title in the Certificate to the GSE, on what is known as the "settlement date." Defendants filed a ProSupp with the SEC within one day of the settlement date and delivered the ProSupp to the GSE shortly thereafter. The ProSupp contained the balance of the detailed information regarding the supporting loans and served as Defendants' final Section 10(a) prospectus for purposes of 17 C.F.R. § 230.424(b)(5).

Each transaction was conditioned on Defendants' promise that the ProSupp would not reveal a material difference between the true character of the supporting loans and those described in the free writing prospectus. Cf. id. § 230.433(c)(1) (providing that a free writing prospectus and prospectus supplement "shall not conflict"). Conditional agreements of this sort were common in the market for asset-backed securities at the time. As comments to the SEC explained, "asset-backed securities offerings involved conditional contracts where investors agreed to purchase securities before they had all the prospectus information." Securities Offering Reform, SEC Release No. 75, 2005 WL 1692642, at *75 n.407. If a ProSupp revealed "new or changed information" that differed materially from the loan descriptions in the free writing prospectus, the GSE would be "given the opportunity to reassess [its] purchase decision[]." See id.

2. Analysis

With that context in mind, it is clear that the ProSupps, although transmitted after the GSEs initially committed to purchase the Certificates, could be material to the GSEs' purchase decisions. See, e.g., N.J. Carpenters Health Fund II, 709 F.3d at 125-28 (holding statements in RMBS prospectus supplements could be material); Plumbers' Union Local No. 12 Pension Fund v. Nomura Asset Acceptance Corp., 632 F.3d 762, 773 (1st Cir. 2011) (same). The ProSupps served dual functions of filling informational gaps left by the free writing prospectus offerings while also confirming that the loan quality representations in those initial offering documents were truthful in all material respects. In so doing, the ProSupps assumed the material role of convincing the GSEs to finalize the transactions. Cf. Field v. Trump, 850 F.2d 938, 948 (2d Cir. 1988) (concluding that misstatements or omissions that "lull" plaintiffs "into forgoing" a unilateral right are material).

A contrary result would undermine the Securities Act's "philosophy of full disclosure." See Basic, 485 U.S. at 234, 108 S.Ct. 978 (quoting SEC v. Capital Gains Research Bureau, Inc., 375 U.S. 180, 186, 84 S.Ct. 275, 11 L.Ed.2d 237 (1963)). It is fundamental to the Act that every sale of registered securities must be preceded or accompanied by a Section 10(a) prospectus without any material misstatements or omissions on pain of civil liability. See 15 U.S.C. §§ 77e(b)(2), 77l. The ProSupps were the sole Section 10(a) prospectuses delivered in these transactions. If they were categorically immaterial because of their dates of transmission, Defendants could be held to account only for statements made in free writing prospectuses, which may "omit[] in part or summarize[] information," 15 U.S.C. § 77j(b), and would no longer face the possibility of civil litigation for failing to satisfy the full disclosure requirements of Section 10(a). The Act does not permit such an outcome.

c. Substantive Challenge: The Reasonable Investor Standard

Defendants further attack the substance of the court's materiality holding by arguing that the ProSupps' underwriting guidelines statements would not have "assumed actual significance" to a reasonable investor in the GSEs' shoes. See TSC Indus., 426 U.S. at 449, 96 S.Ct. 2126. Defendants contend that, given the GSEs' unique power in the RMBS market, the analysis in this case should have focused on whether a reasonable investor with the GSEs' knowledge and investment purposes, rather than a reasonable generic buyer of PLS certificates, would have considered the underwriting guidelines statements material. This more-specific reasonable investor, Defendants claim, would have valued less the credit quality of the loans backing the Certificates because the GSEs' driving purpose for purchasing PLS certificates was to meet a statutorily-mandated goal of devoting a percentage of their loan portfolio to low- and moderate-income housing, not to secure a return on investment. Defendants further argue that, to the extent the GSEs valued such a return, the credit enhancements of the GSEs' senior tranche Certificates meant that the quality of the loans would have no more than a de minimis impact on their returns on these investments.

1. Factual Summary

In 1992, Congress imposed on the GSEs "an affirmative obligation to facilitate the financing of affordable housing for low- and moderate-income families in a manner consistent with their overall public purposes, while maintaining a strong financial condition and a reasonable economic return." Federal Housing Enterprises Financial Safety and Soundness Act, Pub. L. No. 102-550, § 1302(7), 106 Stat. 3491 (codified at 12 U.S.C. § 4501(7)). Congress delegated authority to administer this mandate to the U.S. Department of Housing and Urban Development ("HUD").71

HUD set annual requirements for the percentage of the GSEs' loan portfolios that were required to be devoted to low- and moderate-income housing. See Federal Housing Enterprises Financial Safety and Soundness Act, § 1331, 106 Stat. at 3956 (codified as amended at 12 U.S.C. § 4561). In 1993, HUD required the GSEs to devote 30% of their portfolios to low- and moderate-income housing. See 58 Fed. Reg. 53048, 53049 (Oct. 13, 1993). By 2006, HUD's requirement grew to 53%. The penalties for failing to meet HUD's low-income housing goals were severe. The GSEs' executives' compensation was tied to meeting HUD's goals. HUD could also send the GSEs cease-and-desist letters and assess civil monetary penalties against them.

The GSEs were entitled to count loans backing PLS toward HUD's low- and moderate-income housing goals. See 24 C.F.R. § 81.16(c)(2). The GSEs negotiated with Defendants and other PLS sellers for the right to select certain loans for the SLGs backing the Certificates to ensure that those loans met HUD's criteria. The GSEs knew that mortgage loans issued to borrowers with lower income came with an increased risk of default. Hence, they secured credit enhancements to protect their investments in the Certificates.

2. Analysis

"The question of materiality, it is universally agreed, is an objective one, involving the significance of an omitted or misrepresented fact to a reasonable investor." TSC Indus., 426 U.S. at 445, 96 S.Ct. 2126. For that reason, the GSEs' HUD-mandated investment goals have no role to play in the reasonable investor test in this case. A court is not required to import the subjective motives of a particular plaintiff into its materiality analysis.

The reasonable investor was designed to stand in for all securities offerees, whose purposes for investing and experiences with financial products may vary. Limiting the reasonable investor's intentions and knowledge to the plaintiff's subjective features would undermine that design. See Basic, 485 U.S. at 234, 108 S.Ct. 978.

Defendants' definition of the reasonable investor is not compelled by the rule that a court assessing the materiality of a statement must consider the offering documents "taken together and in context." See Rombach v. Chang, 355 F.3d 164, 172 n.7 (2d Cir. 2004) (quoting I. Meyer Pincus & Assocs., P.C., v. Oppenheimer & Co., Inc., 936 F.2d 759, 761 (2d Cir. 1991)). A court must, of course, consider the statement at issue in the context of the objective features surrounding the sale and the seller. That context includes, for example, all facts related to the statement or omission, its surrounding text, the offering documents, the securities, the structure of the transaction, and the market in which the transaction occurs. See Omnicare, 135 S.Ct. at 1330 ("[A reasonable] investor takes into account the customs and practices of the relevant industry."); Freidus v. Barclays Bank PLC, 734 F.3d 132, 140 (2d Cir. 2013) (considering the materiality of misstatements and omissions in light of the "deteriorating credit market"). The context Defendants contend the District Court improperly ignored is different. They argue that the District Court should have considered subjective facts about the buyers and their motives for engaging in the transaction. We find no support for that position.

In any event, we would affirm even assuming arguendo that a reasonable investor would have shared the GSEs' subjective purpose of purchasing PLS certificates to meet HUD-mandated housing targets. Materiality casts a net sufficiently wide to encompass every fact that would significantly alter the total mix of information that a reasonable investor would consider in making an investment decision. See Basic, 485 U.S. at 231-32, 108 S.Ct. 978. An interest in whether the loans backing a particular PLS met HUD's definition of low- and moderate-income housing does not exist to the exclusion of a profit motive. Indeed, the fact that the GSEs sought credit protection for their investments indicates that they cared whether the PLS certificate would yield a reliable return. And, as explained above, a reasonable investor in PLS would consider the creditworthiness of the supporting assets material to his or her projection of the securities' total return.

Defendants similarly misplace their reliance on the GSEs' interest in credit protections. This argument erroneously implies a zero-sum game where, on the one hand, an investor either has no credit protection and therefore cares deeply about the credit quality of the loans or, on the other, has strong credit protection and therefore considers the credit quality of the loans irrelevant. Credit enhancement is one important factor that a reasonable investor would consider when deciding whether to invest in PLS. But credit enhancement is not so important that, alone, it would cause an investor to ignore entirely the quality of the loans in the SLG. As one of Defendants' witnesses explained, "[i]nvestors balanced the degree of credit enhancement against the expected losses on the underlying collateral, which generally depended on ... collateral characteristics." J.A. 5226. In other words, the riskier the sponsor represents the loans to be, the more credit protection an investor will seek. It is crucial that a reasonable investor know the true nature of the collateral to ensure that her credit protection is appropriately tethered to the risk of default.

d. Evidentiary Challenges

Finally, Defendants argue that the District Court erred in excluding two categories of evidence related to materiality. First, Defendants argue the court improperly excluded evidence that showed the GSEs, through their Single Family Businesses, knew of the shoddy mortgage origination processes. Second, Defendants argue the court improperly excluded evidence of the GSEs' HUD-mandated housing targets, which they contend are relevant for the reasons described above.

The District Court granted the FHFA's motion in limine to exclude the above evidence under Federal Rule of Evidence 403 because the court found its probative value substantially outweighed by the prejudicial effect of injecting the issue of reliance into the trial. Nomura III, 2014 WL 7229361, at *3-4; see also Morgan Stanley, 592 F.3d at 359 ("[P]laintiffs bringing claims under sections 11 and 12(a)(2) need not allege ... reliance...."). At the time the court rendered its initial Rule 403 decision, the FHFA's Section 11 claims were still in the case, and thus the case was still set for a jury trial. After the trial was converted into a bench trial, the court maintained that the evidence violated Rule 403 and held in the alternative that such evidence was irrelevant. Nomura VII, 104 F.Supp.3d at 593.

We conclude that the District Court did not abuse its discretion on the basis that the challenged evidence was irrelevant to whether the ProSupps' false statements regarding underwriting guidelines were material.

The GSEs' general knowledge of the mortgage market was irrelevant to materiality. As explained above, the GSEs were entitled to treat Defendants' loan quality representations as promises that the loans in these specific SLGs were not a representative cross-section of available mortgage loans but rather a select group of loans with the qualities described in the ProSupps. That the loans differed from those qualities would have affected a reasonable investor's view of the Certificate regardless of that investor's knowledge about mortgage market generally.

The GSEs' housing mandates were similarly irrelevant. However important HUD's housing mandates were to the GSEs' PLS investment decisions, they would not render immaterial to a reasonable investor in the GSEs' position whether or not the investment would produce a financial return.

4. Negative Loss Causation

Defendants appeal the District Court's denial of their negative loss causation defense.

Section 12(b) permits a defendant to seek a reduction in the plaintiff's Section 12 award equal to the depreciation in value of the security not resulting from the material misstatement or omission at issue. See 15 U.S.C. § 77l(b); Morgan Stanley, 592 F.3d at 359 n.7. The text of Section 12(b) plainly provides that loss causation is an affirmative defense to be proven by defendants, not a prima facie element to be proven by plaintiffs. See 15 U.S.C. § 77l(b) (placing the burden of proof on "the person who offered or sold [the] security"); McMahan, 65 F.3d at 1048. The burden to prove negative loss causation is "heavy," given "Congress' desire to allocate the risk of uncertainty to the defendants in [Securities Act] cases." Akerman v. Oryx Commc'ns, Inc., 810 F.2d 336, 341 (2d Cir. 1987); see also NECA, 693 F.3d at 156 (observing that the Securities Act creates in terrorem liability designed to encourage full disclosure by offerors).

Defendants relied on the testimony of two experts, Kerry Vandell and Timothy Riddiough, to meet their burden. Both experts opined that the entirety of the Certificates' losses were attributable to macroeconomic factors related to the 2008 financial crisis and not attributable to the ProSupps' misrepresentations. Faced with the "all-or-nothing proposition" that the Certificates' losses either were or were not "caused entirely by factors other than any material misrepresentations," the court sided with the FHFA. Nomura VII, 104 F.Supp.3d at 541. The court agreed that the financial crisis played a role in the Certificates' reductions in value, but concluded that Defendants failed to disaggregate the crisis from the ProSupps' misstatements. As a result, the macroeconomic financial downturn provided no basis to reduce the FHFA's award. See id. at 585-93. On appeal, Defendants reiterate their arguments that the Certificates lost value as a product of macroeconomic factors related to the 2008 financial crisis, and that the ProSupps' misstatements or omissions are not causally linked to that crisis.72

Although Defendants have maintained that, "through trial, six of the seven Certificates at issue paid ... every penny, and on the seventh, realized losses were $25 million," Nomura's Br. 72, it is clear that the Certificates have suffered loss. "[T]he value of a security may not be equivalent to its market price." McMahan, 65 F.3d at 1048. In the context of RMBS,

NECA, 693 F.3d at 166.

The District Court's task was to determine the cause of that loss. Given that Defendants bore the burden of proof on this issue, the court correctly began with the presumption that "any decline in value" was "caused by the [ProSupps'] misrepresentation[s]." See McMahan, 65 F.3d at 1048. Defendants could break that causal link only by proving that "the risk that caused the loss[es] was [not] within the zone of risk concealed by the misrepresentations and omissions." See Lentell, 396 F.3d at 172 (emphasis omitted). In other words, they were required to prove that "the subject" of the ProSupps' misstatements73 and omissions was not "the cause of the actual loss suffered." See Suez Equity Inv'rs, L.P. v. Toronto-Dominion Bank, 250 F.3d 87, 95 (2d Cir. 2001).

We agree with the District Court that Defendants failed to break the link between the Certificates' reduction in value and the ProSupps' misstatements. We previously suggested that "there may be circumstances under which a marketwide economic collapse is itself caused by the conduct alleged to have caused a plaintiff's loss, although the link between any particular defendant's alleged misconduct and the downturn may be difficult to establish." Fin. Guar. Ins. Co. v. Putnam Advisory Co., LLC (Putnam Advisory), 783 F.3d 395, 404 n.2 (2d Cir. 2015).74 The largely uncontested record evidence suggests that this was such a case. The District Court found that "shoddy [mortgage-loan] origination practices" of the sort concealed by the ProSupps' misstatements "contributed to the housing bubble" that created the 2008 financial crisis. Nomura VII, 104 F.Supp.3d at 587; id. at 536-40; see also Bubb & Krishnamurthy, supra, at 1550-55 (arguing that overinflated expectations of expansions in the housing market created a bubble, which in turn led to the financial crash); Levitin & Wachter, supra, at 1202-10 (arguing that the housing bubble was the product of the PLS market providing an oversupply of housing finance).75 Defendants agreed "that there is a link between the securitization frenzy associated with those shoddy practices and the very macroeconomic factors that they say caused the losses to the Certificates." Nomura VII, 104 F.Supp.3d at 587. They therefore failed to rupture the causal connection between "the subject" of the ProSupps' misstatements and the loss the GSEs suffered. See Suez Equity Inv'rs, 250 F.3d at 95.

The District Court concluded that the 2008 financial crisis was, if anything, an impediment to Defendants' attempt to carry their burden to prove negative loss causation. See Nomura VII, 104 F.Supp.3d at 586-87. That was consistent with our prior statements regarding loss causation and macroeconomic crises. A financial crisis may stand as an impediment to proving loss causation because it can be difficult to identify whether a particular misstatement or macroeconomic forces caused a security to lose value in the fog of a coincidental market-wide downturn. See Lentell, 396 F.3d at 174. When a plaintiff alleges a violation of the Exchange Act, defendants benefit from the opacity of a financial crisis because the burden is on the plaintiff to prove loss causation as a prima facie element. See id. at 172. When a plaintiff alleges a violation of the Securities Act, loss causation is not a prima facie element but an affirmative defense. McMahan, 65 F.3d at 1048. The burden is then on defendants to prove loss causation, and any difficulty separating loss attributable to a specific misstatement from loss attributable to macroeconomic forces benefits the plaintiff. See id. (presuming absent proof to the contrary that any decline in value is caused by the misstatement or omission in the Securities Act context).

Defendants argue that the record clearly refutes the District Court's findings. They contend that testimony from Riddiough, Vandell, and FHFA loss-causation expert James Barth, as well as the GSEs' statements in legal briefs in other cases, SEC filings, and internal documents, all reveal that market-wide forces caused the Certificates to lose value. Even accepting Defendants' view of the trial evidence, we find no basis for reversal. It is uncontested that the housing market and related macroeconomic forces were partial causes of the Certificates' losses. The crucial point that doomed Defendants' loss causation defense is that those macroeconomic forces and the ProSupps' misstatements were intimately intertwined. The financial crisis may have been an important step in between the ProSupps' misstatements and the Certificates' losses, but all three events were linked together in the same causal chain. See Nomura VII, 104 F.Supp.3d at 592 ("[The financial crisis] cannot be `intervening' if [D]efendants' misrepresentations, and the underlying facts they concealed, were part and parcel of it.").76

Finally, we reject Defendants' argument that the ProSupps' misstatements and the financial crisis were not connected because any contribution the ProSupps made to that crisis was "[t]iny." Nomura's Br. 85. Rarely, if ever, is it the case that one can point to a single bad actor or a single bad act that brought an entire financial system to its knees. Financial crises result when whole industries take unsustainable systemic risks. See John C. Coffee, Jr., Systemic Risk after Dodd-Frank: Contingent Capital and the Need for Regulatory Strategies Beyond Oversight, 111 COLUM. L. REV. 795, 797 (2011) ("In 2008, ... a localized economic shock in [the U.S.] subprime mortgage market ... nearly caused the meltdown of worldwide capital markets as that shock was transmitted through counterparties and global markets with the speed of a tsunami."); Kathryn Judge, Fragmentation Nodes: A Study in Financial Innovation, Complexity, and Systemic Risk, 64 STAN. L. REV. 657, 670-77 (2012) (explaining the systemic risk in the market for home-loan securitizations); see also Steven L. Schwarcz, Systemic Risk, 97 GEO. L.J. 193, 204 (2008) (defining systemic risk). The ProSupps' misstatements contributed to the systemic risk in the PLS market in the mid-2000s. Defendants may not hide behind a market downturn that is in part their own making simply because their conduct was a relatively small part of the problem.

B. Blue Sky Claims

Even when a plaintiff prevails under Section 12(a)(2), the analogous Virginia and D.C. Blue Sky provisions require proof of an additional element to trigger relief — that the securities transaction(s) at issue occurred within the regulating jurisdiction. The District Court found that the FHFA met its burden of proof on this element. Nomura VII, 104 F.Supp.3d at 595-97. Defendants contest that finding.77

1. Blue Sky Jurisdiction

"[B]lue-sky laws ... only regulate[] transactions occurring within the regulating States." Edgar v. MITE Corp., 457 U.S. 624, 641, 102 S.Ct. 2629, 73 L.Ed.2d 269 (1982); see UNIF. SEC. ACT § 414(a) (1956); D.C. CODE § 31-5608.01(a) (providing that the D.C. Blue Sky law applies "when an offer to sell is made in [D.C.] or an offer to purchase is made and accepted in [D.C.]"); Lintz, 613 F.Supp. at 550 (observing that the Virginia Blue Sky law applies only to securities transactions that occurred in Virginia). A securities transaction occurs where each party "incur[s] irrevocable liability." Absolute Activist Value Master Fund Ltd. v. Ficeto (Absolute Activist), 677 F.3d 60, 68 (2d Cir. 2012). That may be more than one location. For example, if the buyer "incur[s] irrevocable liability ... to take and pay for a security" in New York and the "seller incur[s] irrevocable liability ... to deliver a security" in New Jersey, the transaction occurs in both New York and New Jersey. See id.

It is undisputed that Defendants did not incur liability to deliver the Certificates in either D.C. or Virginia. The FHFA triggered Blue Sky liability by proving that Fannie incurred irrevocable liability to purchase NAA 2005-AR6 in D.C. and that Freddie incurred irrevocable liability to purchase NHELI 2006-FM2, NHELI 2007-1, and NHELI 2007-2 in Virginia.

2. The D.C. PLS Transaction

The District Court found that the NAA 2005-AR6 transaction occurred in D.C. based on the following facts: Fannie's principal place of business was D.C.; Fannie's PLS traders worked in D.C.; Nomura emailed offering materials to Fannie's PLS traders' work email addresses; and Nomura sent a physical confirmation of purchase to Fannie's D.C. headquarters. Nomura VII, 104 F.Supp.3d at 597. On appeal, Defendants do not contest those findings, but argue they fail to provide a sufficient basis for D.C. Blue Sky liability.

First, Defendants argue that the mere fact that Fannie's principal place of business is in D.C. "does not affect where the transaction occur[red]." Nomura's Br. 93 (internal quotation mark omitted) (quoting Absolute Activist, 677 F.3d at 69).78 That is accurate, but is insufficient to require reversal. The District Court's finding that Fannie purchased a Certificate in D.C. did not rely solely on Fannie's principal place of business. Rather, the court relied on the location of Fannie's principal place of business in addition to testimonial evidence that Fannie's PLS traders worked in the D.C. office. Those two facts taken together adequately support the court's inference for purposes of our review. See Absolute Activist, 677 F.3d at 68 ("[T]he location of the broker [is] relevant to the extent that the broker carries out tasks that irrevocably bind the parties to buy or sell securities....").

Second, Defendants assert that the email addresses on which the District Court relied are "non sequitur[s]" because they "do not reveal anything about the geographic location of the addressee." Nomura's Br. 93 (internal quotation mark omitted) (quoting Shrader v. Biddinger, 633 F.3d 1235, 1247-48 (10th Cir. 2011)). There is a kernel of truth to this argument as well, but it misses the mark. An email address may not reveal much about geographic location of the addressee on its own, but the fact that an addressee received an email at his work email address can support the inference that the addressee opened the email at work. And that fact in turn, taken together with the District Court's finding that Fannie's PLS traders worked in D.C., supports the inference that Nomura's emails were opened in D.C. These findings are further buttressed by the fact that Nomura sent a physical copy of an after-sale confirmation to Fannie's D.C. headquarters. Where Nomura sent an after-sale confirmation is not irrefutable evidence of where the antecedent sale occurred. But the destination for that confirmation supports the inference that the entire Certificate transaction — including the initial offering, the sale, and the after-sale confirmation — occurred between Nomura's New York office and Fannie's D.C. office.

Finally, Defendants argue that the District Court improperly shifted the burden of proof when it observed that that "Defendants have offered no affirmative evidence that the offers to sell were not made in and/or accepted in ... D.C." Nomura VII, 104 F.Supp.3d at 597. Defendants misunderstand the District Court's statement. In deciding whether the evidence showed that the sale occurred in D.C., the District Court merely noted that Defendants offered no evidence to counterbalance the evidence in the FHFA's favor. Balancing evidence, a task well within the fact-finder's competence, is not the same as shifting the burden of proof.

3. The Virginia PLS Transactions

The District Court found that the NHELI 2006-FM2, NHELI 2007-1, and NHELI 2007-2 transactions occurred in Virginia based on similar facts: Freddie's principal place of business was in Virginia; Freddie's PLS traders worked in Freddie's Virginia office; Defendants sent PLS offering materials to Freddie's PLS traders at their work email addresses; and Defendants sent a physical confirmation of sale to Freddie's Virginia headquarters. Id.

Defendants' arguments regarding Virginia Blue Sky jurisdiction largely track their D.C. Blue Sky arguments above and are rejected for the same reasons. Defendants offer two new arguments with regard to the Virginia PLS sales. First, Defendants fault the District Court for not requiring the FHFA to "present[] testimony from someone who ... had direct knowledge about how and where [Freddie's PLS traders] executed the trades" at issue. RBS's Br. 59. While perhaps good advice for the FHFA going forward, that is no argument for clear error. There was more than one correct way for the FHFA to prove its case. Second, Defendants make much of the fact that two Freddie employees stated that Freddie's PLS traders purchased PLS certificates "generally" — instead of "always" — from an office in McLean, Virginia. That testimony may not be the best evidence that Freddie purchased the Certificates at issue in Virginia, but clear error requires more than pointing out that a plaintiff could have, in theory, offered stronger evidence. See Krist, 688 F.3d at 95.

CONCLUSION

"It requires but little appreciation of the extent of the [securities industry]'s economic power and of what happened in this country during the [Great Depression] to realize how essential it is that the highest ethical standards prevail" in financial markets. Silver v. N.Y. Stock Exch., 373 U.S. 341, 366, 83 S.Ct. 1246, 10 L.Ed.2d 389 (1963). In passing the Securities Act, Congress affixed those standards of honesty and fair dealing as a matter of federal law and authorized federal courts to impose civil remedies against any person who failed to honor them. See Ernst & Ernst, 425 U.S. at 195, 96 S.Ct. 1375. And now, in the wake of the Great Recession, the mandate of Congress weighs heavy on the docket of the Southern District of New York. The district court's decisions here bespeak of exceptional effort in analyzing a huge and complex record and close attention to detailed legal theories ably assisted by counsel for all parties.

The judgment is AFFIRMED.

APPENDIX A

APPENDIX B

Securitizatio      Buyer     Sponso     Deposito        Lead
     n                          r          r         Underwriter(
                                                         s)

    NAA           Fannie      NCCI       NAAC          Nomura
  2005-AR6                                           Securities

   NHELI          Freddi      NCCI       NHELI         Nomura
  2006-FM1          e                                Securities

   NHELI          Freddi      NCCI       NHELI         RBS &
  2006-HE3          e                                 Nomura
                                                     Securities

   NHELI          Freddi      NCCI       NHELI          RBS
  2006-FM2          e

   NHELI          Freddi      NCCI       NHELI          RBS
   2007-1           e

   NHELI          Freddi      NCCI       NHELI          RBS
   2007-2           e

   NHELI          Freddi      NCCI       NHELI       [nonparty]
   2007-3           e

APPENDIX C

Securitization      Purchase         Principal       Interest
                      Price          Payments79       Payments

     NAA            $65,979,707     $42,801,327     $17,517,513
  2005-AR6

    NHELI          $301,591,187     $282,411,183    $23,756,542
  2006-FM1

    NHELI          $441,739,000     $331,937,382    $34,559,137
  2006-HE3

    NHELI          $525,197,000     $346,402,921    $42,099,996
  2006-FM2

    NHELI          $100,548,000     $53,271,881      $8,701,219
   2007-1

    NHELI          $358,847,000     $235,700,674    $29,010,757
   2007-2

    NHELI          $245,105,000     $127,924,783    $19,350,587
   2007-3

[Editor's Note: The preceding image contains the reference for footnote79].

APPENDIX D

Securitization       ProSupp      Settlement       Filing
                     Date80         Date81         Date82

     NAA           11/29/2005     11/30/2005     11/30/2005
   2005-AR6

    NHELI           1/27/2006      1/31/2006      1/31/2006
   2006-FM1

    NHELI           8/29/2006      8/31/2006      8/30/2006
   2006-HE3

    NHELI          10/30/2006     10/31/2006     10/31/2006
   2006-FM2

    NHELI           1/29/2007      1/31/2007      1/31/2007
    2007-1

    NHELI           1/30/2007      1/31/2007       2/1/2007
    2007-2

    NHELI           4/27/2007      4/30/2007       5/1/2007
   2007-3

[Editor's Note: The preceding image contains the references for footnotes80,81,82].