Federal Housing Finance Agency v. HSBC North America Holdings Inc. et al

Filing 863

OPINION & ORDER 104541 re: (657 in 1:11-cv-06198-DLC) MOTION for Partial Summary Judgment on the GSEs' Knowledge filed by Federal Housing Finance Agency as Conservator for the Federal National Mortgage Association, (635 in 1:11-cv-06201-DLC) MOTION for Partial Summary Judgment on the GSEs' Knowledge filed by Federal Housing Finance Agency, (853 in 1:11-cv-07010-DLC) MOTION for Partial Summary Judgment on the GSEs' Knowledge filed by Federal Housing Finance Agency, as Conservator for the Federal Home Loan Mortgage Corporation, (662 in 1:11-cv-06189-DLC) MOTION for Partial Summary Judgment on the GSEs' Knowledge filed by Federal Housing Finance Agency: FHFA's April 22, 2014 motion for summary judgment on the element and affirmative defense concerning the GSEs' knowledge is granted. (Signed by Judge Denise L. Cote on 7/25/2014) (tn) Modified on 7/29/2014 (nt).

UNITED STATES DISTRICT COURT
SOUTHERN DISTRICT OF NEW YORK
--------------------------------------- X
FEDERAL HOUSING FINANCE AGENCY,
:
:
Plaintiff,
:
:
-v:
:
:
HSBC NORTH AMERICA HOLDINGS INC., et
:
al.,
:
:
Defendants,
:
:
:
And other FHFA cases.
--------------------------------------- X

11cv6189
11cv6198
11cv6201
11cv7010

(DLC)
(DLC)
(DLC)
(DLC)

OPINION & ORDER

APPEARANCES:

For plaintiff Federal Housing Finance Agency:
Philippe Z. Selendy
Manisha M. Sheth
Andrew R. Dunlap
David B. Schwartz
QUINN EMANUEL URQUHART & SULLIVAN, LLP
51 Madison Ave., 22nd Fl.
New York, NY 10010
Marc E. Kasowitz
Christopher P. Johnson
Michael A. Hanin
Kanchana Wangkeo Leung
KASOWITZ, BENSON, TORRES & FRIEDMAN LLP
1633 Broadway
New York, NY 10019
For defendants HSBC North America Holdings Inc., HSBC USA Inc.,
HSBC Markets (USA) Inc., HSBC Bank USA, N.A., HIS Asset
Securitization Corp., HSBC Securities (USA) Inc., Neal Leonard,
Gerard Mattia, Todd White, and Jon Voigtman:
Mark G. Hanchet
John M. Conlon
Michael O. Ware
MAYER BROWN LLP
1675 Broadway

New York, NY 10019
For defendants Goldman, Sachs & Co., GS Mortgage Securities
Corp., Goldman Sachs Mortgage Co., The Goldman Sachs Group,
Inc., Goldman Sachs Real Estate Funding Corp., Howard S.
Altarescu, Kevin Gasvoda, Michelle Gill, David J. Rosenblum,
Jonathan S. Sobel, Daniel L. Sparks, and Mark Weiss:
Richard H. Klapper
Theodore Edelman
Michael T. Tomaino
Tracy Richelle High
SULLIVAN & CROMWELL LLP
125 Broad St.
New York, NY 10004
For defendants Nomura Holding America Inc., Nomura Asset
Aceptance Corp., Nomura Home Equity Loan, Inc., Nomura Credit &
Capital, Inc., Nomura Securities International, Inc., David
Findlay, John McCarthy, John P. Graham, Nathan Gorin, and N.
Dante LaRocca:
Bruce E. Clark
Katherine J. Stoller
SULLIVAN & CROMWELL LLP
125 Broad St.
New York, NY 10004
Amanda F. Davidoff
Elizabeth Cassady
SULLIVAN & CROMWELL LLP
1700 New York Ave., NW, Ste. 700
Washington, DC 20006
For defendant RBS Securities Inc.:
Thomas C. Rice
David J. Woll
Andrew T. Frankel
Alan Turner
SIMPSON THACHER & BARTLETT LLP
425 Lexington Ave.
New York, NY 10017

ii

Table of Contents
BACKGROUND..................................................... 2
I. Fannie and Freddie ......................................... 5
II. The Life of a Mortgage Loan ................................ 6
A. Underwriting.............................................. 6
B. Sampling.................................................. 8
C. Defendants’ Representations to Purchasers................ 10
D. Repurchase Obligations................................... 11
E. GSEs’ Pre-Purchase Access to Security-Specific
Information.............................................. 12
1.
2.

GSEs Did Not Trade on Material Non-Public Information,
Including Loan Files. ................................ 12
The GSEs’ Access to Loan-Level Information ............ 12

F. GSEs’ Post-Purchase Analyses............................. 15
1.
2.

Anti-Predatory Lending Reviews ........................ 15

Performance Monitoring ................................ 17

III. Originator Reviews ....................................... 20
IV. Aggregator Reviews ....................................... 24
A. Goldman Sachs............................................ 25
B. HSBC..................................................... 26
C. Nomura................................................... 27
D. RBS...................................................... 28

V. The GSEs Were Aware of the Risk of Defective Loans. ....... 28
A. GSEs Had Extensive Experience With Subprime Loans Through
Single Family Purchases.................................. 28
B. Knowledge of the Risks of Low- and No-Documentation Loans 29
C. General Knowledge of the Subprime Marketplace............ 30
D. Risk that Originator Defects Might Reach Supporting Loan
Groups................................................... 30
E. Risk of Inaccuracies in Defendants’ Representations...... 33
1.
2.

The GSEs’ Pre-Purchase Credit Risk Analyses ........... 34
Appraisal Bias ........................................ 35

VI. Single Family Discovery Rulings .......................... 37

iii

VII. FHFA’s Allegations and the Instant Motion ................ 40
DISCUSSION.................................................... 43
I. Relevant Standards ........................................ 47
A. Summary Judgment Standard................................ 47
B. Elements of Sections 11 and 12(a)(2) Claims.............. 48
C. Actual Knowledge......................................... 51

II. FHFA’s Burden ............................................. 53
III. The GSEs Did Not Have Actual Knowledge of Falsity. ....... 54
A. The GSEs’ General Knowledge.............................. 60
B. The GSEs Actually Believed Certain Representations Were
Inaccurate............................................... 66
C. GSEs Believed Some Defective Loans Might Reach SLGs...... 69
1.

Mortgage Loans May Reflect Origination Practices. ..... 69

3.

Defendants’ Use of Sampling ........................... 73

2.
4.

Expectation “Inherent” in GSEs’ Business Practices .... 71
Repurchase Obligations ................................ 77

D. Limits on Discovery...................................... 78

CONCLUSION.................................................... 85

iv

DENISE COTE, District Judge:
Plaintiff Federal Housing Finance Agency (“FHFA”), as
conservator for the Federal National Mortgage Association
(“Fannie Mae”) and the Federal Home Loan Mortgage Corporation
(“Freddie Mac”) (together, the Government-Sponsored Enterprises
or “GSEs”), brought these actions against financial institutions
involved in the packaging, marketing, and sale of residential
mortgage-backed securities (“RMBS”) purchased by the GSEs
between 2005 and 2007, alleging among other things that
Defendants 1 made materially false statements in offering
documents for the RMBS (the “Offering Documents”).

FHFA has

moved for partial summary judgment on several grounds.

At issue

in this Opinion is FHFA’s motion concerning the GSEs’ knowledge
of the falsity of these statements.
FHFA urges that no reasonable jury could find that the GSEs
knew these statements were false; accordingly, FHFA requests
partial summary judgment on Defendants’ knowledge defense under
15 U.S.C. § 77k(a) (“Section 11”) and the absence-of-knowledge
The remaining defendants are Goldman, Sachs & Co. and related
entities (“Goldman Sachs”), HSBC North America Holdings Inc. and
related entities (“HSBC”), Nomura Holding America Inc. and
related entities (“Nomura”), and RBS Securities Inc. (“RBS”)
(collectively, “Defendants”). The GSEs purchased the securities
at issue between the following dates: with respect to Goldman
Sachs, from September 7, 2005 to October 29, 2007; with respect
to HSBC, from December 20, 2005 to July 3, 2007; with respect to
Nomura, from November 30, 2005 to April 30, 2007; and with
respect to RBS, from August 31, 2006 to January 31, 2007.
1

element of FHFA’s claims under 15 U.S.C. § 77l(a)(2) (“Section
12(a)(2)”) and similar provisions of D.C.’s and Virginia’s Blue
Sky laws.

See D.C. Code § 31-5606.05(a)(1)(B) (“D.C. Blue Sky

law”); Va. Code Ann. § 13.1-522(A)(ii) (“Virginia Blue Sky law,”
and together with the D.C. Blue Sky law, the “Blue Sky Laws”).
For the reasons set forth below, FHFA’s motion is granted.
BACKGROUND
FHFA brought sixteen related actions in this district
alleging misstatements in the Offering Documents for certain
RMBS certificates purchased by the GSEs between 2005 and 2007
(the “Certificates”).

All but four actions have settled.

The

remaining actions concern 65 residential mortgage-backed
securities issued or underwritten by Defendants (the
“Securities”), which the GSEs purchased for more than $19
billion.
RMBS are securities entitling the holder to income payments
from pools of residential mortgage loans (“Supporting Loan
Groups” or “SLGs”) held by a trust.

Each of the mortgage loans

underlying the Securities at issue (the “Mortgage Loans”) began
as a loan application approved by a financial institution, known
as the loan’s originator (the “Originator”). 2

Goldman Sachs,

The Originators in these cases, none of which are parties to
the actions, include Ameriquest Mortgage Co.; Countrywide Home
Loans, Inc.; First Franklin Financial Corp.; Fremont Investment
& Loan; Greenpoint Mortgage Funding, Inc.; IndyMac Bank, F.S.B.;
2

2

HSBC, and Nomura acted as “aggregators” here, 3 purchasing Alt-A
and subprime 4 mortgage loans and then pooling them together, on
the basis of credit or other characteristics.

The loans

selected for a given securitization were transferred to a trust
created specifically for that private-label securitization
(“PLS”).
Within a given securitization, the loans were placed into
one or more Supporting Loan Groups.

For example, Goldman

Sachs’s INDX 2005-AR18 securitization, offered through a
Prospectus Supplement of September 2, 2005, was comprised of
twenty-three classes of Certificates and two Supporting Loan
Groups with an aggregated stated principal balance of over $2.4
billion.

Goldman Sachs represented that the original principal

balances of the loans in one group “conform[ed] to Fannie Mae
and Freddie Mac guidelines” that set maximum initial loan
balances, and made no such guarantee about the loans in the
second group.

New Century Mortgage Corp.; Option One Mortgage Corp.; Wells
Fargo Bank, N.A.; and WMC Mortgage Corp.

Goldman Sachs served as an underwriter but not an aggregator of
the sole remaining security in the Ally Action. RBS served as
an underwriter for four of the Securitizations at issue in the
Nomura Action.
3

Mortgage loans are often divided, by credit risk, into three
classes. In order of ascending risk, they are “prime” loans,
“Alt-A” loans, and “subprime” loans.
4

3

The trust then issued certificates to Defendants, who in
turn sold them to the GSEs, which entitled the holder to a
stream of income from borrowers’ payments on the loans in a
particular Supporting Loan Group.

Thus, a certificate’s value

depended on the ability of mortgagors to repay the loan
principal and interest and the adequacy of the collateral in the
event of default.
The certificates linked to each SLG were further subdivided
into tranches of varying seniority.

Holders of the most senior

certificates for a given SLG were paid first, after which
holders of the next-most-senior certificates received payment,
and so on.

Thus, should some borrowers in an SLG default on

their loans, certificates in the junior-most tranche would
absorb all or most of the shortfall before payments to more
senior certificates were affected.

Accordingly, the most senior

certificates were subject to less risk than were more junior
certificates.

By apportioning risk in this way, Defendants were

able to create AAA-rated securities from Alt-A and subprime
loans.

The GSEs purchased senior Certificates -- often only the

most senior -- with the highest credit ratings.
The Defendants raise arguments about the GSEs’ knowledge
concerning many facets of the origination and securitization
process.

Below are the principal facts linked to Defendants’

major arguments, as well as needed context for those facts.
4

All

reasonable inferences are drawn in favor of Defendants, as nonmovants.

Where Defendants offered a litany of similar examples,

the Court has attempted to select the strongest or most
illustrative.

For reasons given in the Discussion section, the

Court finds that many of these facts have little bearing on the
question at issue here: whether the GSEs actually knew that the
Defendants’ specific representations in the Offering Documents
at issue here were false.
I.

Fannie and Freddie
Fannie Mae and Freddie Mac are government-sponsored

enterprises created to ensure liquidity in the mortgage market.
Fannie Mae was established in 1938, Freddie Mac in 1970.

Their

primary business is to purchase mortgage loans from originators
that conform to the GSEs’ standards (“conforming loans”) and
then either hold those loans on their own books or securitize
them for offer to the public.

This side of their business is

known as the “Single Family” side.

In 2000, the GSEs began to

purchase quantities of Alt-A and subprime loans as well and
securitizing some of those purchases.

Office of Policy

Development and Research, U.S. Department of Housing and Urban
Development, Fannie Mae and Freddie Mac: Past, Present and
Future (2009), in 11 Cityscape: J. Pol’y Dev. & Res. 231, 236-37
(2009).

5

Each GSE also conducts a second business.

The GSEs

purchase and hold private-label mortgage-backed securities
(PLS), although this is a substantially smaller portion of their
activities.

It is the PLS that the GSEs purchased from the

Defendants that prompt the claims in these lawsuits.

The GSEs

held approximately $100 billion in PLS in 2002, with roughly $35
billion in subprime and $3 billion in Alt-A PLS; at their peak,
in 2005, the GSEs’ PLS holdings had grown to approximately $350
billion, with roughly $145 billion in subprime and $40 billion
in Alt-A PLS.

Cong. Budget Office, Fannie Mae, Freddie Mac, and

the Federal Role in the Secondary Mortgage Market 10 (Dec.
2010); 5 Nat’l Comm’n on the Causes of the Fin. & Econ. Crisis in
the U.S., The Financial Crisis Inquiry Report 124 fig. 7.3
(2011). 6
II.

The Life of a Mortgage Loan
The process that created a private-label security, whether

followed faithfully or not, generally worked as follows.
A.

Underwriting

A loan file that contained the documents assembled by an
originator for a given Mortgage Loan was reviewed in a process

Available at http://www.cbo.gov/sites/default/files/cbofiles/
ftpdocs/120xx/doc12032/12-23-fanniefreddie.pdf (last visited
July 25, 2014).
5

Available at http://www.gpo.gov/fdsys/pkg/GPO-FCIC/pdf/GPOFCIC.pdf (last visited July 25, 2014).
6

6

called “underwriting” at one or more times before the Loan was
placed in a securitization.

In the first instance, the

originator underwrote each Loan it approved, confirming that it
met applicable underwriting guidelines, was valued reasonably
and accurately, and was not fraudulent.

Originators sometimes

made case-by-case exceptions to their underwriting guidelines
when a loan application failed to meet a certain guideline but
appeared to nonetheless qualify for a mortgage program based on
compensating factors.
Defendants, who functioned as “aggregators” of the Mortgage
Loans, acquired loans from Originators in order to pool them
into Supporting Loan Groups that would be tied to securities.
Before purchasing these loans and issuing securities backed by
them, the GSEs understood that Defendants -- or third-party
vendors on Defendants’ behalf -- conducted due diligence review
of the loan pools to confirm that the loans met the relevant
underwriting guidelines as well as to confirm other
characteristics of the loans that were described in the
prospectus supplements that accompanied each securitization,
such as the loans’ loan-to-value (“LTV”) ratios 7 and combined

The loan-to-value ratio is the ratio of the loan amount to the
appraised value of the property securing the loan. 80% was a
common benchmark used to divide lower- and higher-risk loans.
7

7

loan-to-value (“CLTV”) ratios 8 and the owner-occupancy rates for
the properties underlying the loans.

Based on these reviews,

Defendants might refuse to purchase some loans.
B.

Sampling

As the Defendants described their due diligence practices
to one or both GSEs, when Defendants conducted due diligence
review of a pool of loans, they often reunderwrote a sample of
the loans.

HSBC, for example, disclosed to PLS investors that

it reunderwrote the loan file against not only the originator’s
underwriting guidelines but also HSBC’s own eligibility
standards.

It employed a “10% minimum adverse and random

sample” for prime and Alt-A loans, and “25% minimum adverse and
random sample” (20% adverse and 5% random) for subprime loans.
“An adverse sample . . . [wa]s selected based on the layered
risks inherent in the loans (i.e., those posing the greatest
default and loss [risk] in the pool).”

This adverse sample was

created by a “proprietary model, which w[ould] risk-rank the
mortgage loans in the pool” based on a dozen indicia of credit
risk.

The selection of the sample depended as well on HSBC’s

evaluation of the “[o]verall risk level of the mortgage pool,”

The combined loan-to-value ratio applies to properties securing
more than one loan. It is the ratio of the sum of all loans
secured by the property to the appraised value of the property.
8

8

“[p]rior transaction due diligence results,” and the
“[f]inancial standing of the seller.”
A 2006 Freddie Mac review of Goldman Sachs reported that
Goldman Sachs’s “due diligence sampling model . . . relies
primarily on adverse selection to ensure they see the loans with
the highest probabilities of default.”

The review found that

Goldman Sachs’s “Alt A and Subprime sample levels depend on
seller”; the sample size was “[t]ypically 50%” but could be “2025% for more seasoned sellers.”

These samples were “typically

85-90% adversely selected.”
Similarly, Fannie Mae reported in a 2006 review that RBS’s
“typical sample size” for non-prime loans was 25%,
“predominantly adversely selected.”

For prime and Alt-A loans,

“sampling size [wa]s determined by a statistical calculation
intended to obtain a 95 percent confidence interval, a less than
10 percent error rate, and precision of five percent or
greater.”

RBS “require[d] additional adverse selection for

compliance [red flags], high loan balance, low FICO [credit]
score, seasoning, or other abnormal loan characteristics.” 9
A Freddie Mac report on Nomura’s diligence practices in
March 2006 found that Nomura conducted property and compliance
“Compliance” generally refers to a loan’s compliance with
statutory and regulatory requirements, including the Securities
and Exchange Commission’s Regulation AB, which governs assetbacked securities.
See 17 C.F.R. § 229.1100-1123.
9

9

due diligence on 100% of loans.

Nomura conducted credit due

diligence on 100% of loans in pools with amounts less than $25
million, and on 20% of loans in pools with greater amounts.
C.

Defendants’ Representations to Purchasers

In the offering documents for each security, Defendants
made representations to purchasers, like the GSEs, concerning
the mortgage loans’ adherence to applicable guidelines and the
loans’ characteristics.

The offering documents included a Shelf

Registration Statement filed with the Securities and Exchange
Commission (“SEC”), as well as the relevant Prospectus and
Prospectus Supplements. 10
For instance, with respect to Supporting Loan Group I in
Goldman Sachs’s Securitization FFML 2006-FF13, 11 an SLG that
backed a senior Certificate purchased by Fannie Mae, Goldman
Sachs represented that:
(1)

“[t]he mortgage loans were originated or acquired
generally in accordance with the underwriting
guidelines described in th[e] prospectus
supplement”;

The representations at issue in these actions appear in the
Securities’ Prospectus Supplements. Goldman Sachs issued and
underwrote forty Securitizations pursuant to Shelf Registration
Statements and Prospectus Supplements and underwrote one
Securitization issued by a subsidiary of Ally Financial Inc.
HSBC issued seventeen Securitizations pursuant to Shelf
Registration Statements and Prospectus Supplements. And Nomura
issued seven Securitizations pursuant to Shelf Registration
Statements and Prospectus Supplements.
10

In its supplemental opposition brief, Goldman Sachs leads with
a discussion of FFML 2006-FF13.
11

10

(2)

67% of the loans (or 68.94% of the pool by
principal balance) had an LTV ratio of 80% or
lower, and the same percentage had a CLTV ratio
of 80% or lower;

(3)

100% of the underlying properties were owner
occupied; and

(4)

the Certificate would be “rated . . . in the one
of the four highest [credit] rating categories.” 12

D.

Repurchase Obligations

Many of the Offering Documents for the Securities at issue
included a repurchase provision requiring Originators or
Defendant sponsors to buy back a “defective” loan in certain
circumstances.

In many cases, the Originator or sponsor was

obligated to repurchase a loan if it was originated as a result
of fraud, negligence, or a misrepresentation or omission.

For

example, with respect to FFML 2006-FF13, First Franklin
Financial Corp. (“First Franklin”), as Originator, or Goldman
Sachs, as sponsor, was “obligated to repurchase, or substitute
for, [any] mortgage loan” should it happen that, “with respect
to [that] mortgage loan[,] any of [First Franklin’s or Goldman
Sachs’s] representations and warranties . . . are breached in
any material respect as of the date made.”

More specifically, the Term Sheet for FFML 2006-FF13 stated
that Fannie Mae’s Certificate -- in the most senior class for
its Supporting Loan Group -- was expected to receive a “AAA”
credit rating from Standard & Poor’s and a “Aaa” rating from
Moody’s.
12

11

E.

GSEs’ Pre-Purchase Access to Security-Specific
Information
1.

GSEs Did Not Trade on Material Non-Public
Information, Including Loan Files.

Given federal securities laws prohibiting insider trading,
the GSEs understood that their PLS traders and any others
involved in purchase decisions could not be privy to material
nonpublic information about the Securities.

In addition to

purchasing PLS, the GSEs purchased mortgage loans directly
through the Single Family side of their businesses.

But,

because of the securities laws, “there was a firewall between
[each GSE’s PLS business] and [its] single family” business.
For example, Fannie Mae’s Single Family Counterparty Risk
Management Group (“SFCPRM”) recognized that it “must exercise
extreme caution when disseminating information relating to PLS
Counterparties,” since it was barred from “sharing non-public
information with the Mortgage Portfolio Traders” at PLS.
Likewise, Freddie Mac required that non-public information be
scrubbed from data about a given PLS deal before PLS traders
could review it.

In particular, the GSEs understood that

federal securities laws barred them from “pre-settlement access
to the loan files.”
2.

The GSEs’ Access to Loan-Level Information

The GSEs’ PLS traders were often given access only to
aggregated information concerning a securitization or supporting
12

loan group as a whole.

This information typically included the

distributions of LTV and CLTV ratios, owner-occupancy rates,
FICO credit scores, no- and low-documentation loans, mortgage
type (e.g., fixed-rate or amortizing adjustable rate), and Alt-A
and subprime loans.
There is no evidence that the GSEs were ever given prepurchase access to the loan files for the Mortgage Loans.

But,

in at least one PLS deal in March 2006, HSBC sent Paul Norris
(“Norris”), who was responsible for Fannie Mae’s PLS portfolio,
loan-level information including each loan’s originator, LTV
ratio, and property street address. 13

In July and September

2007, Fannie Mae used a proprietary automated valuation model
(“AVM”), called the Retrospective Property Service (“RPS”), to
test the appraised property values in three prospective PLS
deals, none of which is at issue in these actions.

To run an

RPS test, Fannie Mae needed to have, among other data, the
street addresses for the mortgage properties.

An AVM can offer

insight into the risk of appraisal bias in a given loan. 14

David

The loan-level information provided to Fannie Mae on that
occasion also included note origination date, loan product,
original term, maturity date, loan purpose, original and current
loan amount, interest rate, property type, borrower’s income and
debt-to-income (“DTI”) ratio, monthly P&I, and occupancy type.
13

Freddie Mac Risk Officer Ray Romano (“Romano”) testified that
he believed a deviation of “5 to 10 percent” between AVMs and
appraisers would be “reasonable . . . to expect”; a greater
difference could signal a greater risk of bias.
14

13

Gussman (“Gussman”), a Fannie Mae Vice President in the Capital
Markets Strategy division, testified that Fannie Mae “ran [RPS]
sort of on an ad hoc basis on a couple of deals” in or around
September 2007 and was “investigating” whether to use it “on all
PLS deals going forward” but never did so “because we never got
that data we’re talking about” for enough deals.

In a September

2007 email discussing the use of RPS in connection with another
PLS deal, a Fannie Mae employee noted that “[f]or seasoned
[i.e., older] deals I don’t think we have address data.”
The only evidence with respect to Freddie Mac indicates
that loan-level information, including property street
addresses, was not provided to PLS traders, although others at
both GSEs had access to loan-level information to review the
extent to which loans would meet affordable housing goals set
for the GSEs by the U.S. Department of Housing and Urban
Development and to request that loans that did not meet housing
goals be removed from a prospective Supporting Loan Group. 15
There is no evidence that the GSEs ever used property street
addresses to conduct a pre-purchase test of the Mortgage Loans
at issue in these actions for appraisal bias.

These housing goals included, for example, the purchase of
“loans to lower income borrowers that are owner occupied and in
metro areas.”
15

14

F.

GSEs’ Post-Purchase Analyses
1.

Anti-Predatory Lending Reviews

The GSEs made efforts to discourage predatory lending
practices such as prepayment penalties and excessive fees and
points, or lending without regard to a borrower’s ability to
repay the loan.

But, “because of SEC related restrictions

related to public securities, the [pre-purchase] loan level due
diligence [for anti-predatory lending compliance] [wa]s managed
by the seller of securities, not Fannie Mae Legal.”

These

sellers were required to make representations and extend
warranties about compliance with these anti-predatory policies.
The GSEs had the right to “put back” any loans that violated the
“reps and warranties.”

Consequently, Fannie Mae’s Legal

Department took steps after the purchase of a PLS certificate to
confirm that the PLS sellers had complied with Fannie Mae’s
anti-predatory lending policy.

The Legal Department undertook

operational reviews of Originators after the settlement date of
PLS purchases and required transaction-specific due diligence 90
days after the settlement date.

These reviews were not shared

with Fannie Mae’s PLS traders.
As of May 2005, Fannie Mae’s anti-predatory lending review
was conducted as follows.

Once Fannie Mae’s Supporting Loan

Group was finalized, and prior to settlement, Fannie Mae
required that the issuer or underwriter have its due diligence
15

provider “generate a Fannie Mae compliance exception report”
(“Fannie Mae Exception Report”).

Loans identified as exceptions

had to be cleared with the diligence provider or removed from
the SLG prior to settlement.

Ninety days after settlement, the

Fannie Mae Exception Report was sent to Fannie Mae’s Legal
Department.

The Legal Department would review the Report and

require immediate repurchase of any loan exceptions that
remained in the SLG. 16
Defendants have submitted two such Fannie Mae Exception
Reports.

The first, issued in April 2006 concerning Goldman

Sachs Securitization GSAMP 2006-FM1, noted that the “borrower
does not have sufficient ability to pay” in 10 loans (1% of the
sample reviewed), as well as several other compliance
violations, including excessive finance charges or certain
documents missing from the loan file.

The second, issued in

July 2005 concerning a securitization not at issue in these
actions underwritten by RBS, did not find that any borrowers had
insufficient ability to repay the loan.

Defendants misleadingly characterize Fannie Mae’s Legal
Department’s anti-predatory lending reviews, arguing that
“Fannie Mae also reviewed a sample of SLG loans after purchase
to confirm that they complied with its anti-predatory lending
policies.” The evidence cited by Defendants does not suggest
that the Legal Department ever conducted its own review of a
sample of loan files; it merely reviewed the Exception Report
generated by the issuer or underwriter’s due diligence vendor.
16

16

2.

Performance Monitoring

After purchasing RMBS, the GSEs monitored the performance
of their Securities.

They received monthly “remittance reports”

that reflected delinquencies and early payment defaults (“EPDs”)
in the underlying Mortgage Loans.

Freddie Mac created a

“Weakest Deals Report” listing struggling securities and a
“Mortgage ABS Portfolio Report” that tracked delinquency rates
by loan type.

Similarly, Fannie Mae created a monthly “Watch

List” identifying poorly performing PLS deals.

Several of the

GSEs’ former employees testified that EPDs may suggest
misrepresentations in the underlying loan applications.
In the late summer and fall of 2006, the GSEs saw an
increase in EPDs, including in the PLS they held.

On August 25,

2006, a Fannie Mae weekly email sent by PLS trader Shayan
Salahuddin (“Salahuddin”) concerning news in the subprime market
noted that H&R Block, the parent of Option One Mortgage (“Option
One”), took losses “due to an increase in early payment defaults
on Option One originations.”

Salahuddin continued:

Investors should carefully consider how this affects
their view of origination standards and quality
control in the subprime sector, as Option One has long
been considered on[e] of the better originators.
However, from a purely economic perspective, this
issue is more poignant to originators than to
investors as originators are required to repurchase
loans that default within the first couple of payment
cycles.

17

Salahuddin testified that “in late 2006 . . . the increase in
defaults in deals in general was . . . concerning.”

A December

15, 2006 email from Kin Chung (“Chung”), Fannie Mae’s Director
of PLS Surveillance, noted, based on preliminary data, that “our
2006 vintage [of PLS] is showing much faster ramping in
delinquencies than previous vintages” and that “the 2006Q2
[second-quarter 2006] vintage is performing significantly worse
than the 2006Q1 vintage.”
In November 2006, an internal Freddie Mac document titled
“Summary of Freddie Non-Agency Portfolio and the Subprime
Market” reported, with respect to the “2006 Vintage” of loans,
that “2006 60+ day delinquencies (‘DQ’) [i.e., loans where
monthly payments are more than 60 days past due] are greater
than all other vintages at comparable seasoning, including 2000
which was the worst vintage of this decade.”

In particular, the

summary notes that the 60+ day DQ rate was equivalent to 90
basis points (“bps”) for the 2003 vintage, 160 bps for the 2005
vintage, and 240 bps for the 2006 vintage.

It also reports that

“more leveraged 2nd liens and high LTV loans have performed
worse in 2006 relative to any other vintage,” that there is a
“significant incidence of foreclosure in the 2006 vintage at
such an early stage,” and that “there has been a noticeable
decline in underwriting standards (higher CLTVs, lower FICOs).”
The summary states that “[u]nderwriting is a major concern as
18

originators are stretching for volume.”

With respect to EPDs in

particular, the summary reported that “[e]arly payment defaults
are up significantly, especially for loans with piggyback
seconds [second mortgages],” with “Option One, Accredited and
WAMU hav[ing] reported significantly higher levels of EPD.”
Freddie Mac’s December 2006 “MABS [Mortgage Asset Backed
Securities] Industry Review” reported “a pervasive loosening of
credit standards in originators’ underwriting guidelines.”

It

determined that “the combination of lower HPA [home price
appreciation] and aggressive underwriting practices ha[d] begun
to show up in delinquency and EPD figures.”

The review noted

that Fitch projected that “2006 vintage subprime mortgages will
perform substantially worse in 2007, with cumulative lifetime
losses ultimately reaching 7%, the worst collateral losses of
any vintage to date,” and that “[s]erious subprime delinquencies
ha[d] increased almost 50% year-over-year.”
These trends continued in 2007.

For example, a June 8,

2007 Fannie Mae “PLS Credit Trend Report” for the month of May
noted that “Fannie Mae 2006 investments continue to demonstrate
the higher delinquency rate as well as cumulative losses than
other vintages at the same age,” that “Fannie Mae 2007 subprime
investments are tracking the 2006 vintage in 60+ days
delinquency rate,” and that “Fannie Mae 2007 Alt-A acquisitions
perform worse than the 2006 vintage at the same age.”
19

The

report also found that “Fannie Mae 2006 and 2005 investments
have significantly higher Foreclosure rate and REO 17 percentage
compared to other vintages at the same age.

This is consistent

with the looser underwriting standards during that period.”
III. Originator Reviews
The GSEs’ Single Family businesses investigated and
approved originators before purchasing mortgage loans from them.
The PLS operations at the GSEs relied on those reviews, or the
results of those reviews, from the Single Family operations in
making their trading decisions.
As of August 2006, Fannie Mae required PLS to confirm,
among other things, that any originator of more than 10% of the
underlying mortgage loans was on a list of Fannie Mae approved
originators.

Its SFCPRM was tasked with approving originators.

SFCPRM reviews, some of which included on-site visits, primarily
assessed “[c]ounterparty risk,” which was “the risk of financial
loss to Fannie Mae resulting from [the counterparty]’s failure
to meet its contractual obligation[s],” including inability to
meet repurchase obligations.

SFCPRM also examined originators’

underwriting protocols, appraisal practices, and fraud
detection.

In some instances SFCPRM, itself or through a third-

“REO,” or “Real Estate Owned,” refers to properties that have
been foreclosed upon and are owned by the mortgage holder.
17

20

party vendor, conducted loan-level reviews of a sample of the
originators’ loans.
Pre-approval from Freddie Mac’s Alternative Market
Operations Group (“AMO”) was required for any originator that
originated loans constituting more than 1% of the unpaid
principal balance of Freddie’s PLS portfolio.
Freddie’s Single Family business.

AMO was part of

AMO’s counterparty reviews of

originators examined, among other things, the originator’s
adherence to its underwriting guidelines and its appraisal
protocols.

AMO originator reviews typically included on-site

visits and reunderwriting of fifty loan files.

While PLS

traders at Freddie Mac had access to AMO’s originator reviews,
it appears that Fannie Mae’s traders were only given a list of
approved originators.
SFCPRM conducted an on-site operational review of
Originator First Franklin on May 24 and 25, 2006. 18

Defendants

list First Franklin as the sole “major originator” in eight of
the 65 Securities at issue and as one of several “major
originators” in two of the Securities.

Among the counterparty

risks SFCPRM evaluated were the “[r]isk of poor underwriting
performance and deterioration of credit quality in originations”
Defendants also highlight a second originator review of First
Franklin from August 2007, but this postdates the GSEs’ last
purchase of a Certificate in which Defendants identify First
Franklin as a “major originator.”
18

21

and the “[r]isk of poor due diligence around property valuation
and collateral support.”

SFCPRM’s summary findings concerning

First Franklin’s origination practices were as follows:
First Franklin has acceptable operational
practices and procedures to mitigate the risk
associated with the origination of Subprime
collateral. FF utilizes quality control and quality
assurance findings to enhance their processes and
procedures. Sales and branch office compensation is
based heavily on loan performance and QC [quality
control] findings, rather than simply on loan volume.
This creates companywide awareness around quality, not
quantity.
The appraisal review process utilizes CoreLogic’s
LoanSafe to mitigate the risk associated with
fraudulent and inflated appraised values. The use of
CoreLogic and other vendors is considered a best
practice within the industry. FF’s automated
underwriting system, Easy Writer, is rules based and
allows for loan scenarios to be placed in the best
available loan program. FF’s [sic] is very conscious
of predatory lending and has hard stops built into
their LOS [loan origination system] to prevent the
origination of loans which violate state and federal
predatory guidelines. FF also tracks all loan
exceptions and reviews the exceptions in a targeted
quality control sample each quarter to determine any
unfavorable risk trends associated with specific
exceptions.
Reviewing the performance of older First Franklin loans, SFCPRM
found that First Franklin
ARM [adjustable-rate mortgage] vintages are performing
consistently at the top tier against Subprime peers.
Although the 2001, 2002, and 2003 vintages are
trending downward, the ranking indicates a well
performing pool of loans. . . . Trending analysis
should be reviewed periodically to predict any
downward trends.

22

On SFCPRM’s behalf, diligence vendor Clayton reviewed 450 19
recently originated loan files and found 95 credit exceptions.
Forty-three of these exceptions were “immaterial,” resulting
from issues like “outdated” documents, reducing the number of
“true level 3 findings” (i.e., loans that fail to meet
guidelines and “would result in an outright denial for
purchase”) to 52.

These 52 credit exceptions included

insufficient credit history, missing housing verification,
insufficient assets, and exceeded DTI and LTV ratios.

SFCPRM

characterized these findings as “better than average as compared
to other Subprime originators” and as “acceptable.”

SFCPRM also

noted that First Franklin “sets a 15% annual tolerance level for
branch underwriting errors,” including “[b]oth clerical and
substantive errors.”
Other Originator reviews were less positive.

For example,

Freddie Mac’s AMO reviewed Residential Mortgage Assurance
Enterprise, LLC (“ResMae”) in April 2004 and gave it an overall
rating of “Marginal” and recommended that Freddie Mac “Proceed
With Caution.”

Defendants list ResMae as a “major originator”

in one of the 65 Securities at issue, which Freddie Mac
purchased on April 30, 2007.

AMO found that ResMae’s “credit

Notably, although SFCPRM reports that Clayton reviewed 450
loan files, the percentages for exceptions suggest Clayton
reviewed 4,500 loan files.
19

23

philosophy is marginal, and closely resembles the traditional
‘story loan’ approach formerly used in subprime,” which approves
loans based “more on the borrower’s individual story than . . .
on objective lending criteria.”

AMO also found that “[q]uality

control is inadequate, as it relies entirely on investor due
diligence.”
One of the most extreme examples is AMO’s review of First
NLC Mortgage Corporation (“First NLC”), which Defendants name as
a “major originator” in three of the 65 Securities at issue,
purchased by Freddie Mac December 29, 2005, August 31, 2006, and
January 31, 2007.

AMO’s review, conducted on January 31 and

February 1, 2005, awarded First NLC an overall rating of “Poor.”
AMO found that First NLC “allow[ed] almost any borrower with a
FICO [credit score] of 500 or greater to obtain a mortgage” and
determined that “weaknesses in First NLC’s processes and
controls lead to high levels of risk layering,” which was
“particularly evident in their treatment of credit scores, which
combines a bad business practice with poor execution.”

AMO

warned that “[t]his results in loans with credit attributes that
would be greatly misrepresented on a pricing tape.”
IV.

Aggregator Reviews
The GSEs reviewed and approved aggregators as well as

originators.

Their reviews of the four aggregators whose

Securities are at issue reflect the following assessments.
24

A.

Goldman Sachs

AMO reviewed Goldman Sachs’s conduit program 20 on May 18,
2005 and rated it “Satisfactory” across the board.

AMO noted

that Goldman Sachs conducted “[d]ue diligence on 100% of loan
files using Goldman Sachs guidelines,” which included “[r]eunderwrit[ing] 100% of all loan files, re-calculat[ing] all loan
data, review[ing] for benefit to borrower and capacity to
repay.”

With respect to appraisals, AMO found that Goldman

“[p]erform[ed] extensive property-by-property valuations using
Automated Valuation Models (AVM), Broker Price Opinions (BPO)
and field reviews,” and “[e]valute[d] 60% of the appraisals.”
AMO issued a review of Goldman Sachs as an aggregator
generally on June 15, 2006, rating its due diligence
“Satisfactory” and rating it “Above Average” overall.

As noted

above, Goldman Sachs’s “due diligence sampling model . . .
relie[d] primarily on adverse selection to ensure they see the
loans with the highest probabilities of default,” with sample
sizes “[t]ypically 50%.”

With respect to appraisal review in

particular, AMO noted that “Goldman Sachs takes a proactive and
unique approach to the appraisal review process that starts in

Under its conduit program, Goldman Sachs acquired, for
securitization, mortgage loans from originators and others
pursuant to a given set of Goldman Sachs underwriting guidelines
that would be set out in the securitization’s registration
statement.
20

25

their adverse sample selection.”

AMO explained that “GS closely

monitors national trends and looks for pockets of weakness in
the national real estate markets, these areas are identified by
zip codes and become a part of their adverse selection sample.”
Goldman “runs an AVM on the entire sample selection,” which it
selected “based on . . . the ‘robustness’ of data provided by
the AVM vendor.”

Moreover, AMO reported that “GS views vacant

and listed properties as major risks” and thus uses “zip code
tables and other internal adverse indicators” to identify
properties on which to “order occupancy checks and run MLS
listings to verify the status.”

A credit approval processed in

May 2007 shows that AMO rated Goldman Sachs as an aggregator as
“Above Average.”
B.

HSBC

Fannie Mae issued a counterparty credit review of HSBC
(HASCO) on May 3, 2007.

The review found that HSBC’s “subprime

shelf is acceptable based upon (1) the quality and quantity of
loan file due diligence and (2) strong pool performance of the
bonds over a period of two years and 17 deals.”

Concerning

HSBC’s diligence processes, Fannie Mae found that HSBC:
•

Has a well defined and comprehensive loan
acquisition due diligence process that includes
o in depth originator and servicer review and
validation process which incorporates the
experience of the HSBC organization[; and]

26

o an above average sampling size and the active
participation of HSBCs’ [sic] personnel in the
entire loan review process including loan level
review of exceptions in valuations, credit, and
compliance.
•

Established its own sub-prime loan underwriting
guidelines, which has allowed the aggregator to
minimize the inclusion of exceptionally risky loan
types.

Fannie Mae did note, however, that “[a]lthough [HSBC] claim[s]
to conduct intensive and rigorous counterparty due diligence,
they have a track record of buying collateral from some poorly
capitalized companies including New Century.”

The report

recommended approving HSBC as a PLS aggregator.
C.

Nomura

Freddie Mac’s AMO issued an aggregator operational review
of Nomura on March 14, 2006.

“Based upon the combination of

good due diligence methodologies, reasonable valuation processes
and sound controls, AMO rate[d] Nomura subprime as Satisfactory
overall.”

AMO found that “Nomura’s due diligence program is

well managed,” and “found no issues with Nomura[’s] appraisal
process, which is solid.”

AMO noted that “Nomura takes the

property evaluation process seriously and places a high priority
on collateral valuation.”

27

D.

RBS

Fannie Mae issued an aggregator review of RBS Greenwich
Capital in November 2006. 21

The review notes that RBS employed

Clayton, The Capital Group, and Watterson-Prime to “conduct loan
level due diligence on its acquisitions.”

RBS reviewed loans

“pursuant to seller’s guidelines,” and “stated that its program
to monitor seller lending matrices [in connection with their
guidelines] [wa]s robust,” although Fannie Mae was not provided
“in-depth detail regarding this program.”

RBS was found to

“perform[ ] credit reviews through a process designed to
determine that the loans generally comply with the lender’s
underwriting guidelines through a check of borrower income and
asset documentation, review of credit reports and credit scores,
and recalculation of debt to income ratios.”

Fannie Mae

understood that RBS employed the sampling methodology described
above.
V.

The GSEs Were Aware of the Risk of Defective Loans.
A.

GSEs Had Extensive Experience With Subprime Loans
Through Single Family Purchases.

The GSEs’ Single Family businesses bought a large number of
subprime and Alt-A mortgage loans directly from Originators.

Although RBS has settled FHFA’s claims against it in the Ally
Action, it has not settled the claims against it in the Nomura
Action, where RBS was an underwriter for four of the
Securitizations at issue.
21

28

The GSEs purchased these loans either in bulk (“bulk purchases”)
or individually, through a “flow” of single loans (“flow
purchases”).

In some instances, the GSEs performed pre-purchase

due diligence on the loans.

Defendants have identified

documents showing that Fannie Mae conducted pre-purchase due
diligence on bulk purchases of subprime loans since 2006.
For example, on September 11, 2006, Fannie Mae purchased
779 subprime loans with an unpaid principal balance of
approximately $120 million directly from Originator First
Franklin.

Later that month, Fannie Mae purchased a Certificate

in Goldman Sachs’s FFML 2006-FF13 Securitization; First Franklin
originated “[s]ubstantially all” of the Mortgage Loans in the
corresponding Supporting Loan Group.
B.

Knowledge of the Risks of Low- and No-Documentation
Loans

Because of disclosures made in documents provided to the
GSEs prior to their purchase of PLS, they were informed that
some of the SLGs contained low- and no-documentation mortgage
loans, and thus, stated-income stated-asset (“SISA”) loans and
no-income no-asset (“NINA”) loans.

The GSEs understood that

low- and no-documentation loans posed heightened risks of
borrower fraud, and considered the percentage of these loans in
the Supporting Loan Groups when making PLS purchase decisions.

29

C.

General Knowledge of the Subprime Marketplace

In addition to the knowledge gained through the GSEs’
first-hand experience in the Alt-A and subprime loan markets,
the GSEs received a number of articles in 2006 and 2007 about
weakness in the subprime market.

Some of these articles

reported a rise in borrower fraud and less rigorous underwriting
by originators.

Fannie Mae PLS traders drafted “subprime weekly

updates” that summarized news concerning the subprime market.
D.

Risk that Originator Defects Might Reach Supporting
Loan Groups

Some of the GSEs’ former employees testified about the
risks of defective loans being sold to the aggregators and
reaching Supporting Loans Groups.

Michael Aneiro (“Aneiro”),

Freddie Mac’s Vice President of Non-Agency Portfolio Management,
who headed Freddie Mac’s PLS portfolio, agreed that, “if an
originator was not following its own guidelines and was
contributing loans to the collateral for a pool, [he] would
expect that loans not underwritten to the originator’s
guidelines would then end up in the collateral for a pool.”
Aneiro also agreed that it “would be [his] general expectation”
that “when an originator contributed collateral to a PLS deal
. . . the collateral reflected the originator’s general
origination practices.”

30

Aneiro’s counterpart at Fannie Mae, Norris, agreed when
asked if he would “expect the loans backing PLS [he was] buying
to be reflective of an originator’s general underwriting
practices.”

Norris also testified that Fannie Mae’s “process

basically relied on the dealers and originators providing [it]
with reps and warrant[ie]s as to the validity of how these loans
were underwritten.”

Norris explained that, after receiving a

report prepared within the Single Family business that
Originator New Century may not have been following its own
guidelines, Fannie Mae’s PLS division did not investigate
whether New Century was following their guidelines “as to
originations in PLS,” because Fannie insisted on “reps and
warrant[ie]s on every deal that we purchase.”
The attorney in Fannie Mae’s Legal Department who headed
its anti-predatory lending reviews, John Ingram, testified that
if the Legal Department saw “a trend . . . with a particular
originator,” it may have “let the trading desk know that we
don’t want to see any more deals with that originator.”

Ingram

also testified that “the issuer would have been engaged with the
due diligence provider in looking at the anti-predatory lending
results and removing loans presettlement directly with the due
diligence provider,” because “[t]hat was their responsibility.”
Peter Niculescu (“Niculescu”), Fannie Mae’s Vice President
for Portfolio Strategy, when asked if he “expect[ed] that the
31

originator’s loans would perform similarly whether they were
sold as whole loans or placed in PLS,” testified:

“Yes, I can

see no reason why the loan performance would be different,
depending on what vehicle they were subsequently placed into.”
Niculescu also testified that Fannie Mae “purchased [the
Securities] through established underwriters who had assuredly
good due diligence processes on the underlying loans,” and
therefore “would not have had at the time much reason to be
concerned . . . about the incidence of fraud in the issues of
securities that we had purchased.”
Perri Henderson (“Henderson”), a Freddie Mac PLS trader,
when asked if she would “expect the loans backing [a] PLS deal
to generally reflect th[e] originator’s practices,” answered,
“Unless they stated otherwise I would assume that.”

Henderson

was not the PLS trader responsible for any of the Certificates
purchased from the Defendants.
Speaking about Freddie Mac’s own use of sampling to review
prospective bulk loan purchases for the Single Family business,
Ronald Feigles (“Feigles”), the Freddie Mac employee responsible
for due diligence on Freddie’s bulk purchases of Alt-A mortgage
loans, agreed that one could “reasonably assume” that at least
some of the unsampled loans contained defects.

And Feigles

agreed that it was “probably accurate” that Freddie Mac would
“typically . . . purchase between 70 and 85 percent of the loans
32

that it looked at in the Alt-A bulk purchases.”

The percentage

of loans purchased among those offered for sale is referred to
as the “pull-through rate.”

A February 2006 memorandum

concerning a Fannie Mae Single Family purchase from Originator
New Century noted that Fannie Mae was told New Century’s
“standard pull through rate is 90%” and further noted that
“[o]ther lenders in subprime expect pull through rate (e.g.
First Franklin) in the mid and upper 90’s.”
E.

Risk of Inaccuracies in Defendants’ Representations

There is no evidence that the GSEs ever questioned
Defendants’ representations about any of the Securities at
issue.

At a few points, however, each GSE questioned the

reliability of the descriptions of certain sets of PLS loans.
First, in July 2006, a Freddie Mac credit approval analysis
authored by Kevin Palmer (“Palmer”) stated, concerning a
security purchased from dismissed defendant UBS Americas, Inc.
(“UBS”), that the “deal has about 47% of the collateral
concentrated right at 80% CLTV.

We believe that the dealer is

not reporting true CLTV accurately and therefore we have shifted
all of this collateral into the 95+ CLTV bucket for costing
purposes.”

Palmer reported, “Even under this extreme scenario,”

the deal warranted approval.

On August 21, 2006, Palmer wrote

about another security purchased from dismissed defendant UBS
that the “deal contained a high percentage of collateral right
33

at 80% LTV, this was an indicator to us that the reported CLTV
number could have been low.”

And in another July 2006 email,

Palmer noted that Freddie Mac “requested loan level information
(. . . ensur[ing] that this information was made available to
all AAA investors who requested it),” where a Merrill Lynch SLG
had “the highest CLTV’s (95.99%) that we have seen in the past
six months” and “a fairly high percent of NINA [no-income noasset loans].”

Palmer noted Freddie Mac was “concerned about

the accuracy” of its internal risk estimates based on aggregate
pool-wide information, believing that they overestimated credit
risk.
On March 26, 2007, Fannie Mae’s Norris wrote regarding a
pool of loans originated by New Century, “Given these are likely
the last loans made by New Century, we ask that you make some
liberal assumptions with regard to CLTV, income, dti, etc.”

The

GSEs never purchased a security backed by these loans.
1.

The GSEs’ Pre-Purchase Credit Risk Analyses

The GSEs’ pre-purchase credit analysis often included
modeling multiple scenarios to determine a Certificate’s risk
profile.

For instance, as of April 2007, Fannie Mae’s pre-

purchase review included a “50 scenario analysis.”

As Caijiao

Zhao (“Zhao”), the Director of PLS Analytics, testified,
“[t]here’s always the best case analysis, which is using the
tape [with Defendants’ representations], using the corporate
34

approved model, but there’s another set of alternative . . .
‘what if’ analys[e]s,” where “[w]e say, okay, what if LTV is 10
points higher than what’s reported to us and things like that.”
Similarly, Palmer testified that Freddie Mac conducted “scenario
analysis . . . illustrating what the performance would be if
[certain loan characteristics, like] CLTVs were different” than
as described by sellers.

As Palmer explained, “our assumption

of the LTVs [wa]s that the information provided to us was
accurate,” but Freddie Mac considered “how the [Certificate]
would perform under different LTVs” as an exercise in “prudent
risk management.”
2.

Appraisal Bias

The GSEs were generally aware of the risk of appraisal
bias, which, if uncorrected, can distort reported LTV and CLTV
ratios.

For example, Gary Kain (“Kain”), the Vice President of

Freddie Mac’s Mortgage Investments and Strategy Group
responsible for overseeing both Single Family and PLS
activities, was asked if he became concerned in 2006 and 2007
about the risk of appraisal bias in PLS.

He testified that “at

some point in this process” he “became . . . aware” that “there
was potentially an issue with appraisal bias.”

In an August

2005 memorandum, the Vice President of Mortgage Credit Risk
Management at Fannie Mae, Shelley Poland, listed as a “risk
factor[] and practice[] prevalent in today’s mortgage credit
35

environment” the “general degrading of appraisal quality as well
as the proliferation of automated appraisal valuations.”
By 2005, Fannie Mae’s Single Family Research & Analytics
group created quarterly “Lender Valuation Bias Reports” that
reported appraisal bias by region and by originator for the top
twenty-five originators, based on a comparison of originatorreported values for loans purchased by Fannie Mae’s Single
Family business against values generated by the GSE’s
Retrospective Property Valuation System (RPS).

The report for

the second quarter of 2007 found that the total median valuation
bias for all originators was 6.2% and the total median valuation
bias for Originator Countrywide, for example, was 7.5%.
The risk of appraisal bias, generally, appears not to have
been separately accounted for in the GSEs’ models, at least by
Fannie Mae. 22

“[S]ubprime collateral had not been analyzed for

appraisal bias historically [at] Fannie Mae.”

Zhao testified

that Fannie Mae did “no[t] account for” the “risk of appraisal
bias in the tool [it] created to assess or quantify credit risk

Some at Fannie Mae believed, however, that the risk of
appraisal bias was partially captured by its models. For
example, Mark An, who worked with Zhao on analytics, explained
that Fannie Mae’s model “probably” failed to capture appraisal
bias risks “fully,” but opined that it was “probably not far off
either” because the model was built using historical data from
1988 to 2005 which included “loans originated from high bias
environment[s] (the early 1990s, for example).”
22

36

of PLS,” because they did not know “how much it [was]” for a
given loan pool.
VI.

Single Family Discovery Rulings
Defendants complain in their opposition to FHFA’s motion

that the limits the Court imposed on discovery in these actions,
particularly in connection with the GSEs’ Single Family
businesses, have prejudiced their ability to establish actual
knowledge.

The parties litigated the proper scope of Single

Family discovery over several years, beginning with a conference
in July 2012.

After treating these issues in six subsequent

conferences over the course of the next year, the Court received
briefing and issued an opinion on June 28, 2013 (the “June 28
Opinion”), in which it set out the appropriate limits.

FHFA v.

UBS Americas Inc., 11cv5201 (DLC), 2013 WL 3284118, at *14-15
(S.D.N.Y. June 28, 2013).

The June 28 Opinion includes a

lengthy description of the history of these disputes as of that
date, which is incorporated by reference here.

The Opinion set

out a procedure for the parties to follow in the event they
needed to request further targeted discovery.

Id. at *25.

Defendants moved for reconsideration on July 12, which the Court
denied on September 25, 2013.

FHFA v. JPMorgan Chase & Co.,

11cv6188 (DLC), 2013 WL 5354212 (S.D.N.Y. Sept. 25, 2013).
These issues were raised again on several occasions.

For

instance, they were raised in the context of the Defendants’
37

requests for certain discovery from Clayton, a due diligence
vendor.

This past December, they were raised in the context of

expansive demands for additional discovery from the GSEs
themselves.

After a lengthy conference with the parties on

December 9, 2013, certain of the Defendants issued further
subpoenas to third parties in an effort to obtain the type of
material that the Court had already ruled was not properly
discoverable in these actions.

On December 18, the Court

ordered that “defendants cease efforts to obtain from third
parties the types of documents and information that this Court
has previously ruled they may not obtain from the GSEs or
others.”

Most recently, these issues returned in January of

this year when Defendants argued that they should be permitted
to use Single Family discovery barred in these actions but
produced in connection with a related case in California.

An

Order of January 8, 2014, prohibited the use of such documents.
As the record before the Court on the instant motion
attests, Defendants were granted extensive discovery of the
GSEs’ Single Family businesses.

FHFA produced three categories

of documents from the Single Family side: (1) “documents
considered in connection with the purchases at issue,” (2)
“documents held by custodians who were required to give the PLS
traders such information,” and (3) “documents that went to the
GSEs’ risk management committees with supervisory responsibility
38

over the PLS trading.”

2013 WL 3284118, at *3.

As detailed in

the June 28 Opinion, the general categories of documents
produced to the Defendants also encompass a broad array of
documents concerning Originators:
FHFA has provided several categories of documents
regarding Originators. These include documents that
discuss risk associated with an Originator, an
Originator’s underwriting guidelines, or an
Originator’s adherence to its underwriting guidelines
if those documents were provided to or accessible to
PLS traders, their supervisors, or senior risk
custodians, regardless of whether the documents
actually pertained to PLS. FHFA has also produced
operational reviews of Originators conducted by Single
Family, regardless of whether they were provided to
PLS traders, their supervisors, or senior risk
custodians. FHFA has also produced reports or studies
generated by Single Family that show appraisal bias on
the part of Originators, again regardless of whether
PLS traders, supervisors, or senior risk custodians
received them. Finally, FHFA has produced any
documents relating to Originators that the PLS
businesses actually used, including Alternative Market
Operations (“AMO”) reviews and Counterparty and Credit
Risk Management group scorecards from Freddie Mac and
Counterparty Credit Reviews, Private Label Securities
Counterparty Approval Reports, and daily surveillance
reports from Fannie Mae.
Id. at *10.
The June 28 Opinion also outlined the production of
documents from GSE committees and custodians who had
responsibility for Single Family operations:
FHFA has also provided documents from sources
that span both the Single Family and PLS sides,
including documents from committees and particular
custodians. From Freddie Mac, FHFA has produced
documents from the Enterprise Risk Management
Committee, the Credit Risk Subcommittee, the Market
39

Risk Subcommittee, and Asset/Liability Management
Committee. From Fannie Mae, FHFA has produced
documents from the Private Label Advisory Team, the
Credit Risk Committee, the Market Risk Committee, and
the Risk Policy and Capital Committee. FHFA has also
produced documents from twenty-five custodians who had
some Single Family responsibilities, including nine
who were directly involved in counterparty reviews.
Id.
Fact discovery closed on December 6, 2013.

FHFA and the

GSEs produced more than 19 million pages of documents in
discovery; defendants in the sixteen related actions produced
more than 212 million pages, and third parties produced more
than 14 million pages.

Defendants took depositions over 46

days, questioning 29 GSE and FHFA witnesses.
VII. FHFA’s Allegations and the Instant Motion
FHFA alleges misstatements and omissions in the Offering
Documents concerning four attributes of the Securities: (1) the
Originators’ adherence to applicable underwriting guidelines in
originating the Mortgage Loans; (2) the LTV ratios and CLTV
ratios of the Mortgage Loans; (3) the owner-occupancy rates of
the properties securing the Mortgage Loans; and (4) the credit
ratings of the purchased Certificates, which correspond to
particular tranches of the Securities.

For instance, in

connection with Nomura’s securitization NHELI 2007-3, 23 Nomura

NHELI 2007-3 is one of Defendants’ strongest examples, as 77%
of the Supporting Loan Group for Freddie Mac’s Certification
were originated by ResMae. As described above, Freddie Mac’s
23

40

made the following representations about Supporting Loan Group
I:
(1)

“All of the Mortgage Loans . . . were originated
generally in accordance with the underwriting
criteria described” in the Prospectus Supplement;

(2)

58.49% of the loans (and 59.31% of the Group I
pool by aggregate remaining principal balance)
had an original LTV ratio of 80% or lower, and
that 18.88% of the loans (and 27.05% of the pool
by remaining balance) had an original CLTV ratio
of 80% or lower;

(3)

90.03% of the properties securing the Group I
loans (corresponding to 89.21% of the loan pool
by aggregate remaining principal balance) were
owner-occupied; and

(4)

the Certificate would not issue unless it was
assigned an “AAA” credit rating by Standard &
Poor’s, a “Aaa” rating by Moody’s, and an “AAA”
rating by DBRS.

FHFA claims that these misrepresentations and omissions caused
hundreds of millions of dollars in damages to Fannie Mae and
Freddie Mac when the Mortgage Loans experienced defaults and
delinquencies at a higher rate than they would have had the
Mortgage Loans been as the Offering Documents described.
FHFA brings its claims for these alleged misstatements and
omissions under Sections 11 and 12(a)(2) of the Securities Act
of 1933 (the “Securities Act”), as well under substantially
identical provisions of the D.C. and Virginia Blue Sky Laws.

To

AMO reviewed ResMae in 2004 and recommended that Freddie Mac
“Proceed With Caution,” in part because its “[q]uality control
is inadequate, as it relies entirely on investor due diligence.”
41

prevail on these claims, FHFA must prove that the Securities’
registration statements (for Section 11) or prospectuses (for
Section 12(a)(2) and the Blue Sky Laws) contained an untrue
statement of a material fact or omitted to state a material
fact.

See 15 U.S.C. § 77k(a); id. at § 77l(a)(2); D.C. Code

§ 31-5606.05(a)(1)(B); Va. Code Ann. § 13.1-522(A)(ii).
Defendants are not liable, however, if the GSEs knew of the
untruth or omission.

Under Section 11, the burden is on

Defendants to prove the GSEs’ knowledge as a defense to
liability; under Section 12(a)(2) and the Blue Sky Laws, the
burden is on FHFA to prove the absence of the GSEs’ knowledge as
an element of its claims.
On April 22, 2014, FHFA brought the instant motion for
partial summary judgment on the ground that no reasonable jury
could find that the GSEs knew of the scores of alleged
misstatements in the Defendants’ Offering Documents.
was fully submitted on June 24.

The motion

Defendants contend that

sufficient evidence exists to permit a jury to find the GSEs
knew that conditions existed in the marketplace -- particularly
in regard to the Originators’ underwriting practices -- that
created a risk of false representations appearing in the
Offering Documents.

42

DISCUSSION
Defendants contend that the GSEs’ knowledge of origination
practices and of the subprime market -- including the rise of
EPDs among 2006 loans in the GSEs’ own PLS, first identified in
late summer 2006 -- suffices to support a finding that the GSEs
actually knew of the “very matters” that FHFA asserts made the
Mortgage Loans false.

In particular Defendants identify seven

categories of information that, they argue, constitute
circumstantial evidence that the GSEs were “well aware of
information FHFA” now claims rendered the Defendants’ Offering
Documents “misleading”:
(1)

the GSEs’ counterparty reviews of Originators;

(2)

the GSEs’ direct purchases of subprime and Alt-A
loans;

(3)

the GSEs’ awareness of risks associated with lowand no-documentation loans;

(4)

the fact that Originators were obligated to
repurchase “defective loans”;

(5)

the GSEs’ general knowledge of “the subprime
marketplace,” which saw declines in RMBS
performance, lesser adherence to underwriting
standards, and increasing borrower fraud;

(6)

the results of Fannie Mae’s anti-predatory
lending reviews; and

(7)

the performance of similar loans from the same
Originators or dealers.

As described below, this evidence could not support a finding of
the GSEs’ actual knowledge of falsity of the alleged false

43

statements in the Offering Documents, and is insufficient to
raise a question of fact that would prevent summary judgment.
Fatal to most of Defendants’ knowledge theories is the
simple fact that, whatever the GSEs’ knowledge of an
Originator’s general practices, the Offering Documents contained
specific representations about particular sets of Mortgage
Loans.

These were Mortgage Loans that the Defendants had

purchased and chosen to populate the Supporting Loan Groups for
a Certificate.

The Defendants’ study and selection of that set

allowed them to describe the quality of these loans in detail.
The level of detail included the percentage of loans in a
Supporting Loan Group with an original LTV ratio of 80% or lower
and the percentage with an original CLTV ratio of 80% or lower.
For instance, Nomura represented in the NHELI 2007-3 Offering
Documents that 58.49% of the loans (and 59.31% of the pool by
aggregate remaining principal balance) had an original LTV ratio
of 80% or lower, and that 18.88% of the loans (and 27.05% of the
pool by remaining balance) had an original CLTV ratio of 80% or
lower.

Thus, whatever the GSEs’ knowledge of general practices

or trends, neither of them had access to the files or data that
would give them knowledge that those detailed representations
about the Mortgage Loans in a Supporting Loan Group were false.
As described below, the Defendants have identified no evidence
-- either direct or circumstantial -- of the extremely
44

improbable scenario they posit: the GSEs, when investing
billions of dollars in Securities, knew of the falsity of
Defendants’ specific representations about the Mortgage Loans in
the Supporting Loan Groups underlying the Certificates at issue.
Instead, Defendants argue in generalities about the GSEs’
knowledge concerning, among other things, weaknesses in some
Originators’ underwriting practices and a rise in early payment
defaults and foreclosures among loans in the GSEs’ PLS
originated in 2006 and 2007.

Such general knowledge could not

support a finding that, for example, Freddie Mac actually knew
that, contrary to Nomura’s representations concerning Supporting
Loan Group I in the NHELI 2007-3 Offering Documents, it was not
the case that 58.49% of the loans (and 59.31% of the pool by
aggregate remaining principal balance) had an original LTV ratio
of 80% or lower.

Defendants offer no evidence suggesting that

Freddie Mac actually knew that the true percentage of loans with
such an LTV ratio was other than 58.49%, or that Freddie Mac
knew the true percentage could not be 58.49%, as opposed to say,
for instance, 50% or 70%. 24
This argument is most plain in the case of LTV ratios and
owner-occupancy percentages, but it applies as well to
Defendants’ representations concerning adherence to applicable
underwriting guidelines and credit ratings. The GSEs’ general
knowledge did not give them actual knowledge that Defendants
would place a sufficient number of defective loans in a given
Supporting Loan Group to render false their representations
concerning adherence to guidelines. Nor did it give them actual
24

45

Nor do the Defendants offer any evidence concerning what
the true numbers were for these Securities.

In fact, at this

late stage of the litigation -- well after the close of
discovery in Goldman Sachs, HSBC, and Ally, and only weeks
before the parties’ Joint Pretrial Orders are due in these
actions on September 2 -- Defendants continue to maintain that
their representations were not, in fact, false. 25

Indeed, they

“vigorously deny FHFA’s allegations that the Offering Materials
contained any misrepresentations or omissions.”

Of course, it

cannot both be the case that the representations were true and
that the GSEs had actual knowledge of their falsity.
Defendants’ failure to raise arguments concerning the
specific representations at issue for each Certificate presents
a challenge to the Court in properly construing Defendants’
arguments.

Defendants’ joint counterstatement of material facts

is comprised of 461 paragraphs, many of them setting out
specific bits of information that are never discussed in
Defendants’ briefs or linked to specific Certificates at issue.
The Court has endeavored to adequately credit Defendants’

knowledge that the warranted Standard & Poor’s credit rating for
a given Certificate was, for example, “A” or “AA,” as opposed to
“AAA.”
Fact discovery is completed in the Nomura Action, but expert
discovery is ongoing. The Nomura Joint Pretrial Order will be
filed on November 21, 2014.
25

46

general arguments while holding Defendants to the specific
evidence proffered here and applying these arguments to the
specific representations at issue.
I.

Relevant Standards
A.

Summary Judgment Standard

Summary judgment may not be granted unless all of the
submissions taken together “show[ ] that there is no genuine
dispute as to any material fact and the movant is entitled to
judgment as a matter of law.”

Fed. R. Civ. P. 56(a).

The

moving party bears the burden of demonstrating the absence of a
material factual question, and in making this determination, the
court must view all facts in the light most favorable to the
non-moving party.

Eastman Kodak Co. v. Image Technical Servs.,

Inc., 504 U.S. 451, 456 (1992); Holcomb v. Iona Coll., 521 F.3d
130, 132 (2d Cir. 2008).

Once the moving party has asserted

facts showing that the non-movant’s claims cannot be sustained,
the opposing party must “set out specific facts showing a
genuine issue for trial,” and cannot “rely merely on allegations
or denials” contained in the pleadings.

Fed. R. Civ. P. 56(e);

see also Wright v. Goord, 554 F.3d 255, 266 (2d Cir. 2009).

Nor

may a party “rely on mere speculation or conjecture as to the
true nature of the facts to overcome a motion for summary
judgment,” as “[m]ere conclusory allegations or denials cannot
by themselves create a genuine issue of material fact where none
47

would otherwise exist.”

Hicks v. Baines, 593 F.3d 159, 166 (2d

Cir. 2010) (citation omitted).

“A submission in opposition to

(or in support of) summary judgment need be considered only to
the extent that it would . . . be[] admissible at trial.”

Doe

ex rel. Doe v. Whelan, 732 F.3d 151, 157 (2d Cir. 2013)
(citation omitted).

Only disputes over material facts -- “facts

that might affect the outcome of the suit under the governing
law” -- will properly preclude the entry of summary judgment.
Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986).
B.

Elements of Sections 11 and 12(a)(2) Claims

This Court has previously discussed at length the structure
and aims of the Securities Act, see In re WorldCom, Inc.
Securities Litigation, 346 F. Supp. 2d 628, 656-64 (S.D.N.Y.
2004), and that discussion is incorporated by reference.

In

brief, the private rights of action in the Securities Act were
“designed to assure compliance with [its] disclosure provisions
. . . by imposing a stringent standard of liability on the
parties who play a direct role in a registered offering.”
Herman & MacLean v. Huddleston, 459 U.S. 375, 381-82 (1983)
(discussing Section 11).

Through Sections 11 and 12(a)(2), the

Securities Act “provides the purchasers of registered securities
with strict liability protection for material misstatements or
omissions in registration statements filed with the SEC,” In re
Lehman Bros. Mortgage-Backed Sec. Litig., 650 F.3d 167, 175 (2d
48

Cir. 2011), as well as “misstatements or omissions in a
prospectus.”

NECA-IBEW Health & Welfare Fund v. Goldman Sachs &

Co., 693 F.3d 145, 156 (2d Cir. 2012).
To establish a claim under Section 11, a plaintiff must
prove that it:
(1)

. . . purchased a registered security, either
directly from the issuer or in the aftermarket
following the offering;

(2)

the defendant participated in the offering in a
manner sufficient to give rise to liability under
section 11; and

(3)

the registration statement “contained an untrue
statement of a material fact or omitted to state
a material fact required to be stated therein or
necessary to make the statements therein not
misleading.”

In re Morgan Stanley Information Fund Sec. Litig., 592 F.3d 347,
358-59 (2d Cir. 2010) (quoting 15 U.S.C. § 77k(a)).

“Issuers

are subject to virtually absolute liability under section 11.”
Id. at 359 (citation omitted).
In defense, a defendant may “prove[] that at the time of
such acquisition [the purchaser] knew of such untruth or
omission.”

15 U.S.C. § 77k(a).

Non-issuers may also prove

that, “after reasonable investigation, [they had] reasonable
ground to believe and did believe, at the time such part of the
registration statement became effective, that the statements
therein were true and that there was no omission to state a
material fact required to be stated therein or necessary to make
49

the statements therein not misleading.”

Id. at § 77k(b)(3).

The “standard of reasonableness” is “that required of a prudent
man in the management of his own property.”

Id. at § 77k(c).

Similarly, to make out a prima facie case under Section
12(a)(2), a plaintiff must prove:
(1)

the defendant is a “statutory seller”;

(2)

the sale was effectuated “by means of a
prospectus or oral communication”; and

(3)

the prospectus or oral communication “include[d]
an untrue statement of a material fact or
omit[ted] 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” [“(the purchaser not knowing of such
untruth or omission)”].

In re Morgan Stanley, 592 F.3d at 359 (quoting 15 U.S.C. §
77l(a)(2)).

Where the burden is on a defendant to prove the

purchaser’s knowledge as a defense to liability under Section
11, under Section 12(a)(2) and the Blue Sky Laws, the burden is
on the purchaser to prove the absence of its knowledge as an
element of its claims.

15 U.S.C. § 77l(a)(2); D.C. Code § 31-

5606.05(a)(1)(B); Va. Code Ann. § 13.1-522(A)(ii).

Plaintiffs

bringing Section 11 and 12(a)(2) claims need not establish
“scienter, reliance, or causation.”

City of Pontiac Policemen’s

& Firemen’s Ret. Sys. v. UBS AG, 752 F.3d 173, 182-83 (2d Cir.
2014).

50

C.

Actual Knowledge

The standard for actual knowledge under Sections 11 and
12(a)(2), as well as the Blue Sky Laws, 26 has already been
decided in these actions.

The June 28 Opinion held that the

purchaser must have “actual,” “specific knowledge of the falsity
of the particular statements at issue.”

2013 WL 3284118, at

*14-15.
The “[a]vailability elsewhere of truthful information
cannot excuse untruths or misleading omissions in the
prospectus.”
1956).

Dale v. Rosenfeld, 229 F.2d 855, 858 (2d Cir.

The Second Circuit has rejected the idea that “merely

available, as opposed to widely known, public information” could
establish a plaintiff’s knowledge, as Section 11’s affirmative
defense requires that “the defendant . . . prove actual
knowledge.”

New Jersey Carpenters Health Fund v. Royal Bank of

Scotland Grp., 709 F.3d 109, 127 n.12 (2d Cir. 2013).
The June 28 Opinion explained that “the GSEs were entitled
to rely on the representations made in the offering documents
and to believe that the work done by defendants and their
diligence firms made the resulting Supporting Loan Group
stronger than the general set of loans being sold by a
particular Originator.”

2013 WL 3284118, at *18.

Accordingly,

The parties agree that Sections 11 and 12(a)(2)’s standard
applies to the Blue Sky Laws.
26

51

there was “no necessary connection between an Originator's
general way of doing business and the characteristics of a
particular group of loans that have been examined and assembled
into a securitization by a defendant entity.”

Id. at *19.

Thus, the GSEs’ “kn[owledge] of problems with an Originator will
not establish that the GSEs had actual knowledge that the
specific representations in the prospectus supplements issued by
the defendants were false.”

Id. at *18.

More fundamentally, knowledge about a general population -here, the set of all loans generated by a particular Originator
-- cannot be conflated with knowledge concerning a specific
subset of that population, like loans from that Originator
selected and securitized into a particular Supporting Loan Group
by a Defendant.

For example, the Second Circuit recently held

that YouTube’s internal estimates showing that 75-80% of video
streams it hosted contained copyrighted material were
“insufficient, standing alone, to create a triable issue of fact
as to whether YouTube actually knew . . . [of] the existence of
particular instances of infringement” for purposes of copyright
infringement claims against it.

Viacom Int’l, Inc. v. YouTube,

Inc., 676 F.3d 19, 32-33 (2d Cir. 2012).
A plaintiff under Sections 11 and 12(a)(2) is under no duty
to investigate a seller’s representations.

Accordingly,

“[c]onstructive knowledge cannot bar a purchaser’s recovery”
52

under Sections 11 or 12(a)(2), Casella v. Webb, 883 F.2d 805,
809 (9th Cir. 1989) (Section 12(a)(2)); see N.J. Carpenters
Health Fund, 709 F.3d at 127 n.12 (similar for Section 11).
While circumstantial evidence of actual knowledge can be just as
probative as direct evidence, see Desert Palace, Inc. v. Costa,
539 U.S. 90, 100 (2003), that evidence must be directed to
actual knowledge; a case for constructive knowledge, on a record
that establishes no actual knowledge, will not suffice.
Actual knowledge of falsity means just that.

Mere

awareness of the ever-present risk that an issuer is mistaken,
and that certain representations might be inaccurate, will not
support a finding of actual knowledge of falsity.

Even

suspicion of falsity, before it ripens into actual knowledge,
will not suffice.

Cf. GSC Partners CDO Fund v. Washington, 368

F.3d 228, 242 (3d Cir. 2004) (noting, in the context of an
issuer’s forward-looking statement, that defendant’s “lack of
certainty or confidence” in its prediction that its statement
was false is inconsistent with defendant’s “actual knowledge” of
falsity).
II.

FHFA’s Burden
FHFA bears the burden on its Section 12(a)(2) and Blue Sky

Law claims to show that the GSEs possessed no actual knowledge
that the specific representations of the Defendants in the
Offering Documents were false.

To carry this burden it has
53

relied on testimony of the GSEs’ former employees, who uniformly
stated that they did not know of any false statements in the
Offering Documents and many of whom said they relied on
Defendants’ representations.

FHFA has relied as well on a

comprehensive written record comprised of 19 million pages
produced by FHFA and the GSEs that shows the GSEs did not
receive loan files for the Mortgage Loans, did not express
suspicion of Defendants’ representations concerning the
Securities at issue, and in fact relied on these representations
in making their purchasing decisions.

FHFA has presented

sufficient evidence from which a jury could find that the GSEs
had no actual knowledge of the falsity of the representations in
the Offering Documents at issue here.

For the reasons below,

Defendants have not created a genuine issue of material fact on
the GSEs’ knowledge.
III. The GSEs Did Not Have Actual Knowledge of Falsity.
Defendants “vigorously deny” that the “Offering [Documents]
contained any misrepresentations or omissions” and reject FHFA’s
claims that the Mortgage Loans were “riddled with underwriting
defects, inflated appraisals and borrower fraud.”

But, “[t]o

the extent these conditions existed” in the Mortgage Loans,
Defendants argue that “the GSEs knew of them.”

That is,

Defendants claim to have identified voluminous circumstantial
evidence that, if it is true that the Mortgage Loans were
54

“riddled with underwriting defects” and the like, the GSEs knew
of the conditions that created those defects.

Knowledge of

conditions creating a risk of falsity, however, is not actual
knowledge of falsity.

And, despite the benefit of millions of

pages of discovery, Defendants have uncovered no evidence,
circumstantial or otherwise, of the GSEs’ actual knowledge of
the falsity of Defendants’ representations.

Consequently,

Defendants have failed to meet their burden under Section 11 or
raise a triable issue of fact with respect to the GSEs’
knowledge as to Section 12(a)(2) and the Blue Sky Laws.
Instead, Defendants have amassed evidence of the GSEs’
general knowledge 27 of Originator practices and the subprime
market at large, including a pattern of early payment defaults
in PLS purchased by the GSEs emerging in late summer 2006.
Through counterparty reviews and a history of close interaction,
the GSEs knew much about the major players in the RMBS market,
including the Originators of the Mortgage Loans and Defendants
themselves as aggregators.

The GSEs also knew and considered

the fact that low- and no-documentation loans carried more risk
The parties dispute the extent to which knowledge held by the
GSEs’ Single Family businesses was shared with, or is imputable
to, the traders who made the GSEs’ PLS purchasing decisions.
FHFA also argues that several categories of evidence relied upon
by Defendants are inadmissible. Because the information
identified by Defendants is insufficient to establish actual
knowledge no matter who was aware of it, the Court need not
reach these issues.
27

55

than full-documentation loans, and were aware of the existence
of appraisal bias in loans originated by certain Originators.
In addition to this general information about the
marketplace and Originators, the GSEs had certain pre-purchase
information about the Mortgage Loans underlying their PLS,
including the descriptions contained in the Offering Documents.
In the case of one PLS not at issue here, the GSEs were given
detailed loan-level information, although not loan files. 28

But,

as will be discussed below, there is no evidence that either of
the GSEs ever knew any of Defendants’ many representations about
the Mortgage Loans that are at issue here to be false.
Because of federal securities laws, the GSEs were not
permitted to conduct pre-purchase reviews of loan files for the
Supporting Loan Groups underlying the Securities.

But, even if

the GSEs had been permitted to investigate more fully the bona
fides of the Defendants’ representations in the Offering
Documents, the securities laws imposed no burden on them to do
so.

Instead, it fell to Defendants, who purchased the loans

from the Originators and determined which loans to pool into
Supporting Loan Groups, to ensure that their representations in
the Offering Documents were accurate.

The GSEs understood that

Fannie Mae received property street addresses in connection
with three prospective PLS deals in July and September 2007, but
none of these are at issue in these actions.
28

56

the Defendants conducted pre-securitization due diligence before
making their representations about the Mortgage Loans to the
market, including the GSEs.

That due diligence provided

Defendants with a complete defense to strict liability claims
under Section 12(a)(2) if they exercised “reasonable care,” and
provided non-issuer Defendants with a complete defense to
Section 11 claims, if the diligence met the standards of “a
prudent man in the management of his own property.”
77k(b)(3), (c); 15 U.S.C. § 77l(a)(2).

15 U.S.C. §

The GSEs were entitled

to rely on those representations, and all the evidence suggests
that they did so.

But, as explained above, the Securities Act

and the Blue Sky Laws do not require any showing of reliance.
Instead, they forbid false statements.
Defendants’ arguments founder on this fact.

The GSEs’

knowledge of infirmities in a particular Originator’s several
underwriting practices does not support an inference that the
GSEs knew the Defendants were placing misrepresentations in the
Offering Documents about the quality of the loans they had
placed in a Supporting Loan Group.
Consider, for instance, one of the strongest examples for
Defendants, Nomura’s NHELI 2007-3 securitization.

Freddie Mac

purchased a Certificate tied to NHELI 2007-3 on April 30, 2007.
More than three-quarters of the underlying loans (77.61%) were
originated by ResMae.

After conducting due diligence (to the
57

best of the GSEs’ knowledge), Nomura represented in the NHELI
2007-3 Offering Documents that 58.49% of the loans (and 59.31%
of the pool by aggregate remaining principal balance) had an
original LTV ratio of 80% or lower, and that 18.88% of the loans
(and 27.05% of the pool by remaining balance) had an original
CLTV ratio of 80% or lower.

Nomura also represented that 90.03%

of the properties securing the loans (corresponding to 89.21% of
the loan pool by aggregate remaining principal balance) were
owner-occupied.

Nomura has offered no evidence that Freddie Mac

actually knew any of these very specific representations about
this SLG were false.
It is true that three years earlier Freddie Mac’s AMO
reviewed ResMae, rating it as “Marginal” and recommending that
Freddie Mac “Proceed With Caution.”

Defendants urge that this

knowledge, together with general knowledge of weakness in the
subprime market and rising early payment defaults in similar
securities, can support a finding that Freddie Mac knew to be
false Nomura’s representation that 58.49% of NHELI 2007-3’s
loans had an original LTV ratio of 80% or lower.

It cannot.

Although Nomura emphasizes the limitations of its due diligence,
Freddie Mac’s AMO issued an aggregator review of Nomura a little
more than a year before this purchase, finding “good due
diligence methodologies, reasonable valuation processes and
sound controls.”

AMO reported that “Nomura’s due diligence
58

program is well managed” and “found no issues with Nomura[’s]
appraisal process, which is solid.”

In particular, AMO noted

that Nomura conducted property and compliance due diligence on
100% of loans and credit due diligence on 20% of pools as large
as NHELI 2007-3’s SLG.
the board.

AMO rated Nomura “Satisfactory” across

There is simply no evidence that Freddie Mac

actually knew -- or even believed -- that weaknesses in ResMae’s
general origination practices would render false Nomura’s
specific representations about the specific loans in the SLG.
Defendants offer four chief arguments in opposition to
FHFA’s motion, each of which is discussed below.

First,

Defendants argue that, despite Defendants’ due diligence, the
GSEs should have known -- given their reviews of Originators and
reports of Originator-specific appraisal bias, knowledge of the
marketplace, and knowledge of the early payment default rates in
2006 for 2006 PLS purchased by the GSEs and of loans purchased
by the GSEs’ Single Family businesses from some of the
Originators -- that poorly originated loans would reach the
Supporting Loan Groups.

But constructive knowledge is

irrelevant, and this evidence does not raise a triable issue of
fact as to actual knowledge.
Defendants next argue that the GSEs actually believed that
defective loans would reach the SLGs.

Their evidence supports,

at most, a belief of inaccuracy concerning securities not at
59

issue in these actions.

Thus, as noted below, the Court need

not reach the question of whether or when a showing of belief
can create a triable issue as to knowledge.
Third, Defendants argue that the GSEs were aware that the
SLGs “may contain” defective loans.

The GSEs were, of course,

aware of the risk that any of Defendants’ representations might
prove inaccurate, and took steps to evaluate and protect against
such risk, but there is no evidence they actually knew any
particular representation was false concerning any of the
Certificates at issue.

In particular, Defendants lean on the

fact that the GSEs knew Defendants’ diligence teams reviewed
only a sample of the loans in a pool.

Yet Defendants offer no

evidence that the GSEs were concerned -- let alone knew -- that
Defendants’ use of sampling had resulted in SLGs that did not
match the Offering Documents’ description of the Mortgage Loans.
Finally, Defendants complain that they have not been
afforded a full opportunity to take discovery concerning the
Single Family side of the GSEs’ businesses.

Defendants have had

that opportunity, and the record makes clear that the additional
information requested about the Single Family businesses of the
GSEs fails to meet Federal Rule of Procedure 56(d)’s standard.
A.

The GSEs’ General Knowledge

Defendants argue that the GSEs’ general knowledge
constitutes circumstantial evidence of the GSEs’ actual
60

knowledge of the falsity of the Defendants’ representations.
They point to evidence of knowledge concerning some Originators
and related phenomena, such as a growing recognition from August
2006 onward that their PLS portfolio was experiencing an unusual
or even distressing number of early payment defaults (EPDs).
The Defendants argue that courts applying the actual knowledge
standard “have relied heavily on circumstantial evidence,
including general knowledge,” and “have found that actual
knowledge of a misstatement can be inferred from a party’s
understanding of limited facts about the alleged misstatements.”
Defendants are mistaken.
Defendants principally rely on the Second Circuit’s opinion
in In re Initial Public Offerings Securities Litigation, 471
F.3d 24 (2d Cir. 2006) (“In re IPO”).
28 Opinion, In re IPO is inapposite.
*15.

As explained in the June
See 2013 WL 3284118, at

The Second Circuit simply held that class certification

was inappropriate where “widespread knowledge” of aftermarket
purchase requirements among a putative class of purchasers would
necessitate “individual inquiries as to the knowledge of each
member of the class.”

In re IPO, 471 F.3d at 44.

As a logical

matter, aftermarket purchase requirements inflate a security’s
price.

Thus, the court went on to note that “it would surely be

at least a reasonable inference . . . that a requirement to
purchase in the aftermarket would artificially inflate
61

securities prices,” and consequently knowledge of the
requirement established knowledge of the artificial inflation,
making an individual inquiry of each class member mandatory.
Id. at 44 n.14.

Here, by contrast, there is no necessary

relationship between deficiencies in an Originator’s
underwriting practices and the falsity of a Defendant’s
particular representations in the Offering Documents supporting
a Securitization about a set of particular Mortgage Loans that
had been reviewed, purchased, and securitized by the Defendant.
Defendants have found no evidence that the GSEs even
mistrusted Defendants’ representations about the Mortgage
Loans, 29 let alone any evidence that the GSEs actually knew that
any particular representation was false and were content
nonetheless to make their purchases blind.

It is with good

reason that Defendants title this prong of their argument “The
GSEs’ Knowledge About Originators’ Practices and Loans Permitted
Inferences About the Loans in the Supporting Loan Groups.”
(Emphasis added.)

It is not enough that certain information

“permitted inferences” (however shaky) to be drawn; the question
is whether the GSEs did, in fact, draw these inferences.
As evidence of this mistrust, the Defendants have principally
pointed to a statement in a July 2006 credit analysis by Freddie
Mac’s Palmer that “[w]e believe that the dealer is not reporting
true CLTV accurately” and another statement by Palmer in August
2006 that “the reported CLTV number could have been low.”
Neither statement concerned the Defendants’ Securities.
29

62

Defendants repeatedly emphasize that circumstantial
evidence can be as probative as direct evidence.

This is true.

But, Defendants have pointed to no circumstantial evidence of
actual knowledge.
To the contrary -- although FHFA need not establish
reliance to prove its claims under Sections 11 and 12(a)(2) or
the Blue Sky Laws -- it is undisputed that the GSEs used the
loan characteristics reported by Defendants to evaluate the
Certificates and model their risk profiles.

Based on those risk

profiles, the GSEs sometimes rejected pools or insisted that a
tranche be further subdivided to mitigate those risks.

Indeed,

because the securities laws prevented the GSEs from conducting
pre-purchase loan-level diligence, the GSEs were all the more
dependent upon aggregators’ representations concerning the
Mortgage Loans.
Moreover, Defendants’ proposed “inferences” are themselves
illogical.

Defendants have offered no reason the GSEs would

have believed that Defendants were incapable of identifying
defective loans and either eliminating them from the Supporting
Loan Groups or taking their existence into account when
composing their descriptions of the SLGs for the Prospectus
Supplements.

To the contrary, the GSEs conducted reviews of

each of the Defendant aggregators and reported that they were
quite capable in their diligence.
63

Second, Defendants’ arguments bear no relation to the
specific representations at issue and thus offer no reason that
the GSEs’ general Originator and RMBS knowledge would have
caused the GSEs to doubt -- let alone recognize as false -particular figures.

Taking as an example the lead

securitization featured in Goldman Sachs’s supplemental
opposition memorandum, Goldman Sachs represented that 67% of the
Mortgage Loans underlying FFML 2006-FF13 had an LTV ratio below
80%.

Goldman Sachs has offered no reason why 67% should have

appeared improbably high to Fannie Mae in September 2006, as
opposed to, say, 50% or 80%.

Even if PLS traders knew of

SFCPRM’s operational review of the Originator of the underlying
loans, First Franklin, in May 2006, it is difficult to see why
SFCPRM’s finding that First Franklin had “acceptable operational
practices and procedures to mitigate the risk associated with
the origination of Subprime collateral” would cause Fannie Mae
to infer that the percentage of Mortgage Loans in FFML 2006FF13’s Supporting Loan Group with an LTV ratio below 80% was not
67%.
Third, a contradiction lies at the heart of Defendants’
current litigation position.

Because Defendants have found no

evidence suggesting that the GSEs actually knew their
representations were false, they effectively argue that poor
origination practices and the declining performance of subprime
64

and Alt-A RMBS should have made it obvious that each of the
pertinent representations in the Offering Documents for the
Defendants’ 65 Securitizations was false.

At the same time,

however, Defendants continue to contend that each of these
representations was true, and that their due diligence defense
will shield them, in part, from liability from any inaccuracy
that may have crept into the Offering Documents.

Although

Defendants are perfectly entitled to plead and argue in the
alternative, their positions here are inconsistent.

It bears

emphasis that at this late stage -- long after the close of fact
discovery and as the parties prepare their Pretrial Orders for
three of these four cases -- Defendants continue to argue both
that their representations were true and that underwriting
defects, inflated appraisals and borrower fraud were so endemic
as to render their representations obviously false to the GSEs.
Using the example just given, Goldman Sachs argues both that
Fannie Mae knew that the percentage of loans with an LTV ratio
below 80% was not 67%, but also that the true figure was, in
fact, 67%.
Here, the Defendants’ evidence concerning the GSEs’ general
knowledge does not create a triable issue of fact on the GSEs’
actual knowledge of the falsity of the specific
misrepresentations alleged in the Offering Documents.

There can

be no question that the GSEs’ general knowledge of weaknesses in
65

some Originator practices, of the risk profiles of low- and nodocumentation loans, of the rise in early payment defaults in
the GSEs’ PLS as of August 2006, and of similar phenomena, 30 does
not support a finding that the GSEs actually knew that the
Defendants’ specific representations about the specific Mortgage
Loans underlying the Securities were false. 31
B.

The GSEs Actually Believed Certain Representations
Were Inaccurate.

Defendants proffer evidence that, they argue, supports a
finding that beginning in mid-2006, “GSE personnel believed that
the data provided by issuers of PLS might not be wholly
accurate.”

A belief that issuers’ data, as a whole, “might” not

This encompasses as well knowledge of compliance exceptions
found by Fannie Mae’s Legal Department in post-purchase antipredatory reviews of PLS.
30

Defendants also argue that the breadth of the discovery that
this Court permitted the Defendants to take of the GSEs is
implicit recognition that “‘generalized knowledge’ can
demonstrate the requisite actual knowledge.” But, permitting
the Defendants to take extensive discovery of the GSEs does not
alter the standards of proof at trial. Indeed, the June 28
Opinion makes clear that “material has been discoverable despite
its limited relevance, given the particularity of the knowledge
on the part of the GSEs the defendants will have to establish at
trial.” 2013 WL 3284118, at *18 (emphasis added).
31

Defendants make a similar, general argument about the Court’s
rulings “sustain[ing] FHFA’s claims at the pleading stage.” It
is difficult to analyze this argument since the Defendants do
not point to any specific passage in a prior Opinion in these
actions. FHFA’s allegations were supported by Certificatespecific information. In any case, Defendants do not explain
why pleading standards should determine a party’s burden of
proof at trial.
66

be “wholly” accurate, is a far cry from actual knowledge that a
specific representation made by the Defendants in a particular
Offering Document is false.

Indeed, Defendants have found no

evidence of suspicion concerning the Securities at issue in the
above-captioned actions, much less any evidence sufficient to
support a finding of actual knowledge that Defendants’ specific
representations were false.
Defendants’ chief evidence on this front, two documents
authored by Palmer, concern securities no longer at issue in
these actions, as they were purchased from dismissed defendant
UBS.

Palmer’s July 2006 credit approval analysis explains,

“[w]e believe that the dealer is not reporting true CLTV
accurately” because the “deal has about 47% of the collateral
concentrated right at 80% CLTV.”

As a result, Palmer modeled

the deal with an assumption of 95% or greater CLTV to show that
“[e]ven under this extreme scenario,” credit approval was
appropriate.

On August 21, 2006, Palmer wrote about another UBS

security that the “deal contained a high percentage of
collateral right at 80% LTV, [and] this was an indicator to us
that the reported CLTV number could have been low.”

(Emphasis

added.)
These documents suggest that Freddie Mac was wary about
deals with a high concentration of loans with a reported CLTV of
exactly 80% and that Freddie recognized the risk that some of
67

these loans contained silent second mortgages which reduced
their credit quality.
account for such risks.

Freddie modeled “extreme scenario[s]” to
Far from raising a question of fact

concerning some general suspicion of the representations made by
every bank in the RMBS industry, these documents show that,
where Freddie Mac had concerns about the reported
characteristics of the Mortgage Loans by a particular bank -here, UBS -- Freddie Mac’s employees discussed these concerns
and expressly asked that they be incorporated into their models.
This is not to suggest that a belief of falsity is tantamount to
actual knowledge of falsity; because Defendants have discovered
no evidence of similar beliefs about the Securities at issue,
the Court need not address when belief ripens into knowledge.
The gap between the GSEs’ precautions and any actual
knowledge of falsity is nicely illustrated in some of the other
documents cited by Defendants.

For instance, Norris instructed

that analysts in March 2007 “make some liberal assumptions with
regard to CLTV, income, dti, 32 etc.” when it came to “the last
loans [originated] by New Century,” which backed a security the
GSEs never purchased.

Norris’s concerns about New Century

prompted caution, but do not show that he knew any particular
representations by Defendants were false.

32

“DTI” refers to a borrower’s debt-to-income ratio.
68

Likewise, in April 2007, Fannie Mae’s Director of PLS
Surveillance, Chung, expressed concern about “‘liar loans’ and
biased appraisals” and asked for models that could calculate
“what severity increment [i.e., what magnitude of negative
effects from these factors] is needed to break a bond.” 33

As

Zhao, the Director of PLS Analytics, explained, “there’s always
wanting to know what if and the impact” and “potential risk
factor[s] you want to run some analysis on.”

No reasonable jury

could find from the fact that the GSEs were aware that there was
a risk that a bank’s representations might turn out to be
incorrect, and modeled and attempted to account for this risk,
that the GSEs knew any particular representation by the
Defendants for these Securities to be false. 34
C.

GSEs Believed Some Defective Loans Might Reach SLGs.
1.

Mortgage Loans May Reflect Origination Practices.

Defendants point to a handful of statements from former GSE
employees indicating that, in general, they expected loans
To “break a bond” is to diminish or wipeout the flow of income
due to the bondholder. Here, because the GSEs purchased senior
Certificates in Defendants’ Securitizations, a number of loans
in the Supporting Loan Group could default before the GSEs’
Certificate payments were affected.
33

For this reason, Defendants’ contention that “the GSEs knew
that the owner-occupancy percentages in the Offering [Documents]
may have been incorrect” because “borrowers could lie about
their intentions” does not raise a triable issue of fact
concerning the GSEs’ actual knowledge that specific
representations concerning owner-occupancy percentages in
specific SLGs were false. (Emphasis added.)
34

69

backing PLS to “reflect” an Originator’s general underwriting
practices.

Many of these statements are ambiguous.

And no one

testified that they believed a material number of defective
loans would be present in particular Supporting Loan Groups, or
so much as suggested that such an expectation caused them to
doubt any of Defendants’ representations about specific SLGs.
These statements do not support a finding that the GSEs actually
knew the Defendants’ representations to be false.
Defendants’ strongest employee testimony came from Aneiro,
Freddie Mac’s Vice President of Non-Agency Portfolio Management.
Aneiro was asked, “[I]f an originator was not following its own
guidelines and was contributing loans to the collateral for a
pool, you would expect that loans not underwritten to the
originator’s guidelines would then end up in the collateral for
a pool?”

Aneiro agreed.

Aneiro did not testify that he knew

any of the Originators here were not following their own
guidelines, or that he believed the Defendants here would
purchase loans from such Originators and permit loans from such
Originators to become “collateral for a pool.”

And, even in

this hypothetical, Aneiro did not testify that he believed a
sufficient number of loans would fail to meet guidelines such
that Defendants’ specific representations regarding the
Securities at issue here would be rendered false.

70

Aneiro was then asked, “[W]hen an originator contributed
collateral to a PLS deal, was it your understanding that the
collateral reflected the originator’s general origination
practices?”

He agreed that that “would be [his] general

expectation, yes.”

This testimony concerning Aneiro’s “general

expectation” about a hypothetical originator’s “general
origination practices” does not suggest that Aneiro believed
this would be true of Originators with inadequate origination
practices, or that the Defendants would agree to purchase and
securitize loans from such Originators, let alone that
Defendants would fail to identify and remove defective loans
from Supporting Loan Groups or falsely describe the SLGs. 35
2.

Expectation “Inherent” in GSEs’ Business
Practices

Defendants next argue that the mere fact that the GSEs
required originator reviews before purchasing PLS, and rejected
certain deals because the principal originators had loose
underwriting standards, shows the GSEs’ belief that an
As noted above, Defendants elicited similar testimony from
Henderson, as well as Norris. Norris went on to testify that
Fannie Mae’s “process basically relied on the dealers and
originators providing [it] with reps and warrant[ie]s as to the
validity of how these loans were underwritten.” Norris
explained that, after receiving a report prepared within the
Single Family business that Originator New Century may not have
been following its guidelines, Fannie Mae’s PLS division did not
investigate whether New Century was following their guidelines
“as to originations in PLS,” because Fannie insisted on “reps
and warrant[ie]s on every deal that we purchase.”
35

71

originator’s general underwriting practices would be reflected
in the SLGS.

This argument again conflates the GSEs’ awareness

of the risk that Defendants might let slip defective loans into
the SLGs with the GSEs’ actual knowledge that Defendants had
slipped a material number of defective loans into any given SLG
and then falsely described the SLGs in the Offering Documents.
Defendants’ arguments that “the GSEs knew” that certain
representations in the Offering Documents “may have been
incorrect” go to recognition of risk, not knowledge of falsity.
(Emphasis added.)
There was, of course, a risk that Defendants’ due diligence
would fail to catch all defective loans in any given
securitization.

As purchasers of Certificates backed by those

Mortgage Loans, the GSEs were exposed to that risk.

Like any

business, the GSEs acted to mitigate that risk, refusing to
purchase securities where that risk was greatest -- i.e., where
the GSEs believed the originators had become too loose with
their underwriting guidelines -- and obtaining “reps and
warranties” from PLS sellers.

Far from supporting the GSEs’

actual knowledge of a material number of defective loans in a
given SLG, these practices show that the GSEs took affirmative
steps to prevent the purchase of securities backed by defective
loans.

In any event, this evidence does not raise a question of

72

fact regarding the GSEs’ actual knowledge of the false
representations.
3.

Defendants’ Use of Sampling

Defendants argue that the fact that they reviewed only a
sample of loans, and that the GSEs knew this, proves that the
GSEs knew some defective loans “might” reach the SLGs.

An

examination of the GSEs’ knowledge of the Defendants’ due
diligence practices does not suggest that the GSEs knew the
Defendants’ representations in the Offering Documents were
false.
First, the Defendants did not report to the GSEs that they
were simply taking a random sample of a certain size.

Rather,

they represented to the GSEs that they employed sophisticated
tools to ensure that their sampling was representative and
permitted them to make representations about the loan pool with
confidence.

The information that the Defendants presented to

the GSEs about their due diligence programs has been described
above.

The GSEs understood, for instance, that HSBC employed a

“25% minimum adverse and random sample” -- 20% adverse and 5%
random -- for subprime loans.

This adverse sample was comprised

of the loans “posing the greatest default and loss [risk].”
HSBC created its adverse sample with a “proprietary model, which
w[ould] risk-rank the mortgage loans in the pool” based on a
dozen indicia of credit risk.

The selection of the sample
73

depended as well on HSBC’s evaluation of the “[o]verall risk
level of the mortgage pool,” “[p]rior transaction due diligence
results,” and the “[f]inancial standing of the seller.”
Similarly, RBS determined the sample size for Alt-A loans with
“a statistical calculation intended to obtain a 95 percent
confidence interval.”
Defendants have offered no reason, and proffered no
evidence, to suggest that they led the GSEs to understand that
their sampling protocols undermined Defendants’ ability to
accurately represent the characteristics of the SLGs.

And

Defendants have offered no evidence that the GSEs believed
Defendants did not employ sampling correctly.

Instead,

Defendants collect bits of available information suggesting
that, in some cases, PLS aggregators reviewed a small percentage
of the loans.

For instance, Norris of Fannie Mae testified in

2013 that although he did not have “specific[]” knowledge about
the way aggregators employed sampling, he recalled that
aggregators’ sample sizes varied and “remember[ed] a range,
possibly 10 percent to 50 percent,” of loans in a Supporting
Loan Group were reviewed.

This testimony was untethered to the

practices of the Defendants and does not contradict the
contemporaneous descriptions which the Defendants gave to the
GSEs of the due diligence practices during the period 2005 to
2007.

For the reasons explained above, this does not suggest
74

actual knowledge that Defendants’ representations in the
Offering Documents were false.
Defendants also point to testimony from Feigles, the
Freddie Mac employee responsible for due diligence on Freddie’s
Alt-A bulk purchases, about Freddie Mac’s own use of sampling to
review prospective bulk loan purchases for its Single Family
business.

Feigles agreed one could “reasonably assume” that at

least some of the loans contained defects.

Feigles did not

suggest, however, that a material number of loans would contain
defects.

Nor does he suggest that, if the rate of defects was

unusually large in a given sample, that Freddie Mac would not
enlarge its sample size so as to keep defects for the whole
loans it was purchasing within tolerance levels.

Much less does

it suggest that Freddie Mac expected that the Defendants would
fail to scrutinize their Mortgage Loans with sufficient care to
ensure that representations they were making in the Offering
Documents they filed with the SEC were accurate.
Defendants also assert that “[f]rom the GSEs’
conversations with Originators, they understood that the ‘pull
through rate’ during subprime due diligence typically was 70-85%
-- meaning the majority of most pools were deemed suitable for
purchase by issuers.”

As an initial matter, the evidence cited

by Defendants does not support this assertion.

Defendants cite

only to testimony from Feigles that it was “probably accurate”
75

that Freddie Mac would “typically . . . purchase between 70 and
85 percent of the loans that it looked at in the Alt-A bulk
purchases” for its Single Family business.

Feigles did not

testify that he understood this rate to be typical for purchases
by aggregators for their securitization business, or that it
arose from any “conversations with Originators.”
Elsewhere, Defendants cite evidence that Fannie Mae’s
Single Family side learned in February 2006 from New Century
that “their standard pull through rate is 90%” and knew that
“[o]ther lenders in subprime expect pull through rate (e.g.
First Franklin) in the mid and upper 90’s.”

Defendants do not

identify evidence of a particular pull-through rate with respect
to any of the Securitizations at issue here, or the GSEs’
knowledge of such a rate.
In any case, the pull-through rate should depend upon many
factors, such as the parameters a purchaser chose for its buying
program, the price a purchaser was willing to pay for the loans,
the quality of the loans reviewed, and the purpose for which the
loans were purchased -- for instance, whether they were to be
held or securitized.

The Defendants have not pointed to

evidence that a particular pull-through rate implies, by itself,
that a material number of defective loans were “pulled through”
into Supporting Loans Groups of the Securitizations.

76

Defendants have identified no evidence that the GSEs
mistrusted the Defendants’ representations because of their use
of sampling, let alone actually knew any one of those many
representations to be false.

Accordingly, the fact that the

GSEs knew Defendants employed sampling in their due diligence
does not support a finding of actual knowledge of the falsity of
Defendants’ particular representations.
4.

Repurchase Obligations

Similarly, the repurchase obligations included in many of
the Certificates’ Offering Documents, which required the
Originator or sponsor to repurchase a loan if any of Defendants’
representations were materially false as to that loan, could not
support a finding of actual knowledge of the falsity of
particular representations.

The GSEs’ recognition of the risk

that some defective loans might reach a Supporting Loan Group is
not the same as actual knowledge that Defendants’
representations concerning the loan pool as a whole were false.
Defendants’ arguments concerning risk disclosures in the
Offering Documents fail for the same reasons.
Individuals purchase accident and travel insurance every
day not because they know they will be in a car crash or that
illness will prevent them from flying away for vacation, but
because they recognize that risk exists and seek to protect
themselves from the financial consequences of the realization of
77

that risk.

Ensuring that sellers of PLS provided “reps and

warranties” for the Mortgage Loans reflects the same prudence.
Considering all of the evidence put forward by the parties,
both singly and together, Defendants have not raised a genuine
issue of material fact as to the GSEs’ actual knowledge of the
specific misrepresentations and omissions alleged by FHFA in the
Offering Documents for the Securities at issue.
D.

Limits on Discovery

On the final page of their joint thirty-five page brief,
Defendants argue that “the limitations on discovery in these
Actions have foreclosed defendants from obtaining discovery that
would further show the GSEs’ knowledge of the alleged defects.” 36
In support of this conclusory assertion they refer to a
declaration that summarizes additional discovery “precluded by
orders of the Court” 37 that they assert “would further
demonstrate the existence of disputed issues of fact.”

Fact discovery has closed in all actions, and expert discovery
has closed in all actions but Nomura. Nomura offers no reason
to believe that the final weeks of expert discovery will produce
evidence of the GSEs’ actual knowledge.
36

The declaration incorrectly characterizes the record of
discovery rulings in this complex litigation, including
assertions that the Court “denied discovery” altogether on
particular issues. The record of the Court’s discovery rulings
is available in its Opinions, endorsements on correspondence,
and the transcripts from conferences. Given the existence of
that record, which includes the Court’s explanations for those
rulings, there is no need to try to correct or place in context
the assertions in the declaration. It is also not necessary to
37

78

Pursuant to Rule 56(d), Fed. R. Civ. P., where a party
opposing summary judgment “shows by affidavit or declaration
that, for specified reasons, it cannot present facts essential
to justify its opposition, the court may: (1) defer considering
the motion or deny it; [or] (2) allow time to . . . take
discovery.”

That declaration must detail, among other things,

“what facts are sought” and “how these facts are reasonably
expected to create a genuine issue of material fact.”

Miller v.

Wolpoff & Abramson, LLP, 321 F.3d 292, 303 (2d Cir. 2003).

A

court is free to reject a non-movant’s Rule 56(d) requests “if
the evidence sought would be cumulative or if the request is
based only on speculation as to what potentially could be
discovered; and a bare assertion that the evidence supporting
[non-movant]’s allegations is in the hands of the moving party
is insufficient to justify the denial of summary judgment.”

In

re Dana Corp., 574 F.3d 129, 148-49 (2d Cir. 2009) (citation
omitted).
The Defendants’ effort to avoid summary judgment by
reference to additional discovery fails for two related reasons.
First, they rest their assertion that additional discovery would
have been of aid to them upon an incomplete and even misleading
description of the fulsome discovery they did take and never
revisit any of those rulings to address the issues presented by
this motion for partial summary judgment.
79

confront their failure to discover evidence suggesting actual
knowledge from that body of material.

Second, they fail to meet

the Rule 56(d) standard.
Document discovery in these actions has been massive: FHFA
and the GSEs produced more than 19 million pages of documents in
discovery.

Defendants received all relevant Single Family

documents.

Defendants are incorrect when they charge that they

were “almost entirely precluded from obtaining information
related to the GSEs’ Single Family businesses.”

Defendants were

granted extensive discovery concerning the GSEs’ Single Family
businesses, including all “documents considered in connection
with the purchases at issue,” “documents held by custodians who
were required to give the PLS traders such information,” and
“documents that went to the GSEs’ risk management committees
with supervisory responsibility over the PLS trading.”
3284118, at *3.

2013 WL

They also received counterparty reviews of

Originators created within the Single Family businesses.
Additional documents of little relevance -- documents neither
considered in connection with the purchases at issue nor
possessed by the management committees that jointly supervised
PLS trading and Single Family activities -- would not aid
Defendants in their opposition to the instant motion.

The

Defendants were never able to explain, for instance, why their
requests for data of Single Family transactions concerning whole
80

loans that were purchased under a different set of guidelines
than used by PLS aggregators and that were not, by definition,
in the Defendants’ SLGs were anything other than oppressive and
irrelevant.
The Defendants complain in their joint opposition brief
about a second, related and unsuccessful request that they made
for Single Family individual loan records.

They complain that

FHFA successfully resisted producing, “on grounds of burden,”
evaluations by the Single Family businesses of individual loans
that they rejected and which ended up instead as loans
populating the SLGs.

The Defendants are correct; the Court

upheld FHFA’s objection to this production.

Like their other

request for transaction records of the loans that the GSEs
actually purchased, these documents would have been of
absolutely no assistance to the Defendants in resisting this
motion for summary judgment.

The Defendants’ request for this

production was rejected not only on the ground of the enormous
burden it would impose on FHFA to attempt to locate such
records, but also because any records that might be located
would be utterly irrelevant.
The Defendants speculated that some number of loans in the
SLGs might have been previously offered to the GSEs’ Single
Family businesses for purchase and rejected for failure to meet
underwriting guidelines.

There were roughly 1.1 million loans
81

in the SLGs in the coordinated cases pending in this district,
and it would have been a herculean (and perhaps impossible) task
to determine whether any of them had previously been offered to
and rejected by the GSEs’ Single Family businesses.

Since the

GSEs had their own guidelines for their buying programs,
however, there was also no reason to believe that the reasons
for rejection would have told us anything about the issues at
stake here.

But assuming there might be some overlap in the

standards causing rejection of the loan by the GSEs and the
alleged misrepresentations by the Defendants in the Offering
Documents, there was no reason to believe that the GSEs had in
their possession any record of their reason for rejecting the
loan, much less had any requirement to record each and every
reason for rejecting a loan.

But, most significantly, from the

perspective of the GSEs’ traders who decided to purchase the
PLS, they did not know which loans populated the SLGs and
therefore could not have drawn any conclusions about the
accuracy of the representations in the Offering Document, which
were in any event representations about the characteristics of
the set of loans that populated the SLGs and not representations
about a single loan.
Defendants also complain that limitations on depositions
left them with “no testimony from some individuals whose
understanding of the risks of the Securitizations could
82

establish knowledge.”

Defendants took depositions over 46 days,

questioning 29 GSE and FHFA witnesses.

None of that testimony

so much as suggested actual knowledge that Defendants’
representations were false.

Because an understanding “of the

risks” would not support a finding of actual knowledge of the
falsity of specific representations for the reasons explained
above, such testimony would be irrelevant as well as cumulative.
The hundreds of hours of testimony Defendants did take, and the
millions of pages of discovery Defendants received, leave no
doubt that the GSEs lacked actual knowledge of the falsity of
Defendants’ representations.
Nor have the Defendants met their burden under Rule 56(d).
This attempted showing comes after the conclusion of discovery,
not in the midst of it.

It comes after many lengthy conferences

and countless submissions in which the Defendants were given an
opportunity to be heard and to explain their requests for
further discovery.

The June 28 Opinion provided a roadmap for

the Defendants to follow in the event they wished to renew any
discovery request and once again explained the legal standards
that underlay the Court’s decisions to grant some of the
parties’ requests and to deny others.

Against that background,

Defendants’ declaration fails to establish how the requested
discovery would be “reasonably expected to create a genuine

83

issue of material fact” as required by Rule 56(d).
F.3d, at 303.

Miller, 321

Two examples will suffice.

The Defendants complain that UBS, the defendant scheduled
to go to trial first, was given authority over the timing of the
FHFA depositions, and that those depositions were therefore
scheduled before FHFA had completed the production of its
documents.

The declaration, however, does not point to any

later-produced document that entitled them to reopen a
deposition, and a denial of a request to do so.
The declaration complains about missing documents.

It

explains that Fannie Mae failed to produce all of its
operational reviews of Nomura and Goldman Sachs and that Freddie
Mac had not retained all of its email for the period before
November 2007.

To the extent that the Defendants believed that

there was any sanctionable discovery abuse by FHFA, they had the
opportunity to raise that issue through the appropriate
application during the discovery period.

They did not do so

then, and they have not done so now.
The declaration argues that the operational reviews would
have shown that Fannie Mae was aware that Nomura performed 20%
credit diligence on large bulk pools and that Goldman Sachs
performed due diligence on a sample of the loans that it
securitized.

Both of those points were captured in the

operational reviews that FHFA did locate and produce in
84

discovery and have been described in this Opinion.

As for the

missing Freddie Mac emails, the declaration speculates that the
email destruction has deprived Defendants of “potentially
important” evidence relating to the GSE’s purchase of
Certificates and its actual knowledge of false statements in the
Offering Documents.

This naked speculation falls far short of

the showing required by Rule 56(d).

As described above, the

Defendants have failed to raise a question of fact regarding
Freddie Mac’s actual knowledge of false statements based on all
of the voluminous discovery provided in this lawsuit.

The

absence of a complete repository of Freddie Mac emails for the
relevant period of time does not cast doubt on that finding, nor
does it suggest that reopening the discovery period would cure
the problem.
CONCLUSION
FHFA’s April 22, 2014 motion for summary judgment on the
element and affirmative defense concerning the GSEs’ knowledge
is granted.

SO ORDERED:
Dated:

New York, New York
July 25, 2014
__________________________________
DENISE COTE
United States District Judge

85