View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

United States Government Accountability Office

GAO

Report to Congressional Committees

February 2010

TROUBLED ASSET
RELIEF PROGRAM
Treasury Needs to
Strengthen Its
Decision-Making
Process on the Term
Asset-Backed
Securities Loan
Facility

GAO-10-25

February 2010

TROUBLED ASSET RELIEF PROGRAM
Accountability Integrity Reliability

Highlights
Highlights of GAO-10-25, a report to
congressional committees

Treasury Needs to Strengthen Its Decision-Making
Process on the Term Asset-Backed Securities Loan
Facility

Why GAO Did This Study

What GAO Found

The Term Asset-Backed Securities
Loan Facility (TALF) was created
by the Board of Governors of the
Federal Reserve System (Federal
Reserve) to help meet consumer
and small business credit needs by
supporting issuance of assetbacked securities (ABS) and
commercial mortgage-backed
securities (CMBS). This report
assesses (1) the risks TALF-eligible
assets pose to the Troubled Asset
Relief Program (TARP), (2)
Department of the Treasury’s
(Treasury) role in decision making
for TALF, and (3) the condition of
securitization markets before and
after TALF. GAO reviewed
program documents, analyzed data
from prospectuses and other
sources, and interviewed relevant
agency officials and TALF
participants.

TALF contains a number of risk management features that in turn likely
reduce the risk of loss to TARP funds, but risks remain. TALF was designed to
reopen the securitization markets in an effort to improve access to credit for
consumers and businesses. The Federal Reserve Bank of New York (FRBNY),
which manages TALF, is authorized to lend up to $200 billion to certain
eligible borrowers in return for collateral in the form of securities that are
forfeited if the loans are not repaid. To assist in this effort, Treasury has
pledged $20 billion of TARP funds in the form of credit protection to the
program in the event the loans are not repaid. As of December 2009, FRBNY
has made about $61.6 billion in TALF loans, of which $47.5 billion remained
outstanding. For most TALF-eligible collateral, FRBNY will stop providing
new TALF loans in March 2010, while new-issue CMBSs will be accepted as
collateral on new TALF loans through June 2010. Treasury and FRBNY
analyses project minimal, if any, use of TARP funds for TALF-related losses,
and Treasury currently anticipates a profit. While GAO found that the overall
risks TALF poses to TARP funds are likely minimal, GAO analyses showed
that CMBSs potentially pose higher risk of loss than ABSs. As shown in figure
1, ongoing uncertainty in the commercial real estate market and TALF
exposure to legacy CMBSs warrant ongoing monitoring. Finally, TALF may
present risks beyond the potential risks to TARP, such as the risk that FRBNY
might fail to identify material noncompliance with program requirements by
TALF participants. Because the Federal Reserve views TALF as a monetary
policy tool, however, statutory limitations on GAO’s authority prohibited GAO
from auditing FRBNY’s role in administering TALF.

What GAO Recommends
Treasury should (1) give increased
attention to reviewing risks posed
by CMBS, (2) strengthen its TALF
decision-making process, and
(3) determine which data are
needed to track the management
and sale of assets TALF borrowers
might surrender. To enable more
effective review of TARP, Congress
also should grant GAO authority to
audit Federal Reserve TALF
operational and administrative
activities. Treasury appreciated
GAO’s recommendations but said
GAO understated TALF’s success
and overstated the risk of CMBS.
The Federal Reserve disagreed that
there are limitations on GAO’s
authority to audit TALF activities.
GAO continues to believe that its
depiction of TALF is accurate and
its recommended actions are
necessary to strengthen the
oversight and operations of TALF.
View GAO-10-25 or key components.
For more information, contact Orice Williams
Brown at (202)512-8678 or
williamso@gao.gov.

Figure 1: Commercial Real Estate Prices and CMBS Delinquencies from January 2004 through
October 2009
Moody's/REAL Price Index
200

Percentage delinquency
5

190
180

4

Prices peaked
in October 2007.

170

3

160
150

As of November 2009, prices
show a drop of 43% since peak.
November was the first month
in which prices increased
since September 2008.

140
130
120

2
1

110
100

2004
Year

2005

2006

2007

2008

2009

0

Source: GAO presentation of Moody’s/REAL Commercial Property Price Index and Moody’s CMBS Delinquency Tracker data.

Note: Moody’s/REAL Commercial Property Price Index is a report showing the change in commercial
real estate asset sales. Moody’s CMBS Delinquency Tracker is a monthly report showing
delinquency rates for commercial mortgage-backed securities.

United States Government Accountability Office

Highlights of GAO-10-25 (continued)

Treasury has not fully documented its rationale, as part
of its decision-making processes, for reaching final
decisions related to the risks of TALF—including
decisions involving other agencies. For example, the
outcomes of Treasury’s internal analysis of the amount
of equity that TALF borrowers should hold in TALF
ABS collateral, along with other TALF program terms,
sometimes differed from FRBNY’s. However, there was
no clear documentation or explanation of how the
discrepancies were resolved or how final decisions
were made with FRBNY. Documenting the rationale
and basis for these decisions would increase
transparency and strengthen internal controls for TALF
decision-making processes. Moreover, a sound
decision-making process would help ensure that TALF
objectives are being met and that it is functioning as
intended. Unless Treasury documents the basis for
major program decisions that it made with the Federal
Reserve, it cannot demonstrate accountability for
meeting the goals of TALF and could unnecessarily
place TARP funds at risk.

TALF loan volume trends—suggests that the
securitization markets improved for the more
frequently traded TALF-eligible sectors after the
program’s first activity in March 2009, which is
illustrated in figure 2. Figure 2 also shows that ABS
issuances in all of the most liquid TALF-eligible sectors
dropped sharply in 2008 from their peak levels in 2006
and 2007. Consumer credit rates have not changed
significantly since TALF started. FRBNY officials said
that it is possible that without TALF, interest rates on
loans to consumers and small businesses would be
much higher than they are now. While Treasury bears
the first-loss risk from any assets that TALF borrowers
might surrender in conjunction with unpaid loans, it
has not developed measures to analyze and publicly
report on the potential purchase, management, and sale
of such assets. Without such a plan Treasury cannot
measure TALF’s success in meeting its goals under
TARP with respect to any collateral that is placed in
TALF LLC. Finally, without a plan for communicating
the findings that result from tracking and analyzing
such metrics, Treasury may not fully inform the public
of how the assets are managed and financed,
undermining Treasury’s efforts to be fully transparent
about TARP activities.

FRBNY, in consultation with Treasury,
monitors various economic indicators to measure
TALF's impact. Our analysis of such indicators—
including securitization volumes, interest rates, and

Figure 2: Fluctuations in ABS Issuances, First Quarter of 2005 through the Fourth Quarter of 2009
Dollars in billions

Enlarged inset

25

80
20
70

15

60

10

50

5
0

40

Q4
’08

30

TALF
announcement

Q1
’09

Q2

Q3

Q4

Q3

Q4

TALF
subscription

20
10
0
Q1
2005

Q2

Q3

Q4

Q1
2006

Q2

Q3

Q4

Q1
2007

Q2

Q3

Q4

Q1
2008

Q2

Q3

Q4

Q1
2009

Q2

Quarter and year
Credit card
Student loan
Auto
CMBS
Source: GAO analysis of Thomson Reuters IFR Markets data.

United States Government Accountability Office

Contents

Letter

1
Background
TALF Includes Features That Mitigate Potential Losses, but CMBSs
Could Pose Greater Risks Than Other Asset Classes
Treasury Worked with the Federal Reserve and FRBNY to Analyze
Risks Related to TALF but Did Not Fully Document Analysis
Supporting Final Decisions
Treasury’s and FRBNY’s Indicators Suggest That Credit Market
Conditions Have Shown Some Improvement, but Treasury Lacks
Performance Indicators in the Event that It Must Purchase TALF
Assets
Conclusions
Matter for Congressional Consideration
Recommendations for Executive Action
Agency Comments and Our Evaluation

36
46
48
48
49

Appendix I

Objectives, Scope, and Methodology

55

Appendix II

Methodology for Market Value Analysis of ABSs
and CMBSs

60

Appendix III

The Securitization Process Explained

63

Appendix IV

Descriptions of Asset-Backed Securities

65

Appendix V

Credit Enhancement

69

Appendix VI

Additional Information on TALF Compliance

72

Page i

4
17

32

GAO-10-25 Troubled Asset Relief Program

Appendix VII

New Securitization Volumes Have Increased since
the Inception of the TALF Program

74

Appendix VIII

Comments from the Department of the Treasury

78

Appendix IX

Comments from the Board of Governors of the
Federal Reserve System

80

Analysis of Legal Comments Submitted by the
Federal Reserve

87

GAO Contact and Staff Acknowledgments

91

Table 1: Timetable of Asset Classes Accepted As Eligible Collateral
Table 2: TALF Haircuts by Asset Class
Table 3: TALF Loans Disbursed by Asset Class, March 2009 through
December 2009
Table 4: GAO Sample Selection for ABS Extreme Market Value
Loss Analysis
Table 5: Auto Loan Securitization Volume, 2005 through Fourth
Quarter 2009
Table 6: Credit Card Receivables Securitization Volume, 2005
through Fourth Quarter 2009
Table 7: Student Loan Securitization Volume, 2005 through Fourth
Quarter 2009
Table 8: CMBS Securitization Volume, 2005 Through Fourth
Quarter 2009

7
13

Appendix X

Appendix XI

Tables

Page ii

42
60
74
75
76
77

GAO-10-25 Troubled Asset Relief Program

Figures
Figure 1: Status of TARP Apportioned Funds as of December 31,
2009
Figure 2: TALF Loan Origination and Repayment Process
Figure 3: Funding Required if a Borrower Walks Away from a Loan
Figure 4: Overall TALF Borrower Returns at Issuance for Select
Asset Classes Accepted as TALF Collateral Generally
Decreased from March through September 2009
Figure 5: Increases in Levels of Credit Enhancement on ABSs Since
the Credit Crisis Began, March 2009 through September
2009
Figure 6: Commercial Real Estate Prices and CMBS Delinquencies
from January 2004 through October 2009
Figure 7: Fluctuations in ABS Issuances, First Quarter 2005
through Fourth Quarter 2009
Figure 8: Trends in Average Finance Rates for Credit Cards, Auto
Loans at Banks, and Auto Finance Companies, June 2008
through December 2009
Figure 9: Composition of TALF Loans Disbursed by Asset Class as
of December 2009
Figure 10: Trends for AAA ABS and CMBS Spreads from January
2005 through September 2009
Figure 11: The Securitization Process

Page iii

6
14
16

23

26
29
38

41
43
45
64

GAO-10-25 Troubled Asset Relief Program

Abbreviations
ABS
AIG
AUP
CBLI
CMBS
CUSIP
EESA
Federal Reserve
FFELP
FICO
FRBNY
LIBOR
NRSRO
OIS
PIMCO
RMBS
SBA
SIGTARP
SUBI
TALF
TARP
Treasury

asset-backed security
American International Group, Inc.
agreed upon procedure
Consumer and Business Lending Initiative
commercial mortgage-backed security
Committee on Uniform Security Identification
Procedures
Emergency Economic Stabilization Act of 2008
Board of Governors of the Federal Reserve System
Federal Family Education Loan Program
Fair Isaac Corporation
Federal Reserve Bank of New York
London Interbank Offered Rate
nationally recognized statistical rating organization
Overnight Indexed Swap
Pacific Investment Management Company, LLC
residential mortgage-backed security
Small Business Administration
Special Inspector General for TARP
special units of beneficial interest
Term Asset-Backed Securities Loan Facility
Troubled Asset Relief Program
Department of the Treasury

This is a work of the U.S. government and is not subject to copyright protection in the
United States. The published product may be reproduced and distributed in its entirety
without further permission from GAO. However, because this work may contain
copyrighted images or other material, permission from the copyright holder may be
necessary if you wish to reproduce this material separately.

Page iv

GAO-10-25 Troubled Asset Relief Program

United States Government Accountability Office
Washington, DC 20548

February 5, 2010
Congressional Committees
The recent financial crisis disrupted the securitization markets—which
have funded an increasing share of consumer credit and small business
loans in recent years—making credit less available to households and
small businesses.1 This “credit crunch” has been a major factor behind the
weak economy and slow recovery. The Term Asset-Backed Securities
Loan Facility (TALF), which the Board of Governors of the Federal
Reserve System (Federal Reserve) first announced in November 2008, was
created to help restore the securitization markets. The Department of the
Treasury (Treasury) provides credit protection for this program under the
Troubled Asset Relief Program (TARP), as authorized under the
Emergency Economic Stabilization Act of 2008 (EESA).2 The Federal
Reserve Bank of New York (FRBNY), which manages TALF, is authorized
to lend up to $200 billion in 3- or 5-year nonrecourse loans to investors to
purchase AAA-rated asset-backed securities (ABS) and commercial
mortgage-backed securities (CMBS), which are, in turn, pledged as
collateral for the loans.3 Nonrecourse loans are provided against collateral,
and if the loans are not repaid, FRBNY’s only recourse is to assume
control of the pledged assets.4 If borrowers do not repay their TALF loans,
Treasury will fund the purchase of up to $20 billion of the collateral
backing the loans from FRBNY, creating exposure that places TARP funds,
and therefore taxpayers, at risk.

1

Securitization is a process where financial assets are brought together into interestbearing securities that are sold to investors. Banks receive financing for future loans from
securitizations and may rely on the ability to securitize in making certain types of loans.
2

Pub. L. No. 110-343, Div. A., 122 Stat. 3765 (2008) (codified at 12 U.S.C. §§ 5201 et seq.).

3

The Federal Reserve originally announced that TALF would lend up to $200 billion,
although it subsequently stated that it would be willing to increase the limit to as much as
$1 trillion if needed. As of June 2009, Federal Reserve officials suggested that an increase
to $1 trillion is unlikely.
4

A nonrecourse loan essentially means that in the event a borrower defaults on the loan,
the borrower’s exposure is limited to the assets used to secure nonrecourse loans.
Accordingly, the lender, in this case FRBNY, cannot seek repayment of the outstanding
debt from any of a borrower’s other assets. However, if a participant in TALF breaches the
loan agreement by, for example, misrepresenting its eligibility for a TALF loan, then the
loan loses its nonrecourse status and the borrower is personally liable for the outstanding
debt.

Page 1

GAO-10-25 Troubled Asset Relief Program

EESA, which authorized TARP, also provided GAO with broad oversight
authorities for actions taken under TARP and requires GAO to report at
least every 60 days on TARP activities and performance.5 To fulfill our
statutorily mandated responsibilities, we have been monitoring and
providing updates on TARP programs, including TALF, in prior 60-day
reports.6 This report expands on previously reported TALF activities.
The Federal Reserve created TALF under Section 13(3) of the Federal
Reserve Act, which permits the Federal Reserve to authorize a Federal
Reserve Bank to lend to nondepository institutions in “unusual and
exigent circumstances.” The Federal Reserve viewed creation of TALF as
part of its monetary policy activities because, having already reduced the
federal funds target rate to close to zero, it needed additional mechanisms,
like TALF, to provide liquidity directly to borrowers and investors in key
credit markets. GAO is prohibited by statute from auditing deliberations,
decisions, or actions by the Federal Reserve or any Federal Reserve bank

5

Section 116 of EESA, 122 Stat. at 3783 (codified at 12 U.S.C. § 5226).

6

See GAO, Troubled Asset Relief Program: Additional Actions Needed to Better Ensure
Integrity, Accountability, and Transparency, GAO-09-161 (Washington, DC: Dec. 2, 2008);
GAO, Troubled Asset Relief Program: Status of Efforts to Address Transparency and
Accountability Issues, GAO-09-296 (Washington, DC: Jan. 30, 2009); GAO, Troubled Asset
Relief Program: March 2009 Status of Efforts to Address Transparency and
Accountability Issues, GAO-09-504 (Washington, DC: Mar. 31, 2009); GAO, Troubled Asset
Relief Program: June 2009 Status of Efforts to Address Transparency and Accountability
Issues, GAO-09-658 (Washington, DC: June 17, 2009); GAO, Troubled Asset Relief Program:
One Year Later, Actions Are Needed to Address Remaining Transparency and
Accountability Challenges, GAO-10-16 (Washington, DC: Oct. 8, 2009); and GAO, Financial
Audit: Office of Financial Stability (Troubled Asset Relief Program) Fiscal Year 2009
Financial Statements, GAO-10-301 (Washington, DC: Dec. 9, 2009).

Page 2

GAO-10-25 Troubled Asset Relief Program

that are related to monetary policy matters.7 Accordingly, although the
Federal Reserve and FRBNY voluntarily gave us access to TALF-related
documents and staff, our review of TALF was limited and we did not
evaluate the actions of the Federal Reserve with respect to TALF. For
example, we did not assess FRBNY’s system of internal controls or the
role of TALF participants—such as agents, borrowers, or auditors—in
certifying and validating compliance with the program’s terms.8 We also
conducted only limited verification of factual information provided by
Federal Reserve officials, which we used for background description
purposes and to assess Treasury’s involvement in TALF.
The objectives of this report are to (1) analyze the risks that TALF
presents to TARP funds and therefore to taxpayers, (2) evaluate how
Treasury analyzed the risk of TALF assets and used this information in
making decisions on TALF with the Federal Reserve and FRBNY, and
(3) assess changes in securitization and credit market conditions before
and after TALF’s implementation, based on indicators tracked by Treasury
and FRBNY.
To meet the report objectives, we analyzed data from TALF prospectuses
and data providers to understand securitization volumes, prices, and
spreads. We also reviewed program operation and design documents,
including documentation regarding the selection of eligible asset classes,
the amount of equity a TALF borrower is required to hold in a security,

7

Section 714 of Title 31 of the U.S. Code limits GAO’s authority to audit certain Federal
Reserve activities. Specifically, GAO audits of the Federal Reserve and Federal Reserve
banks may not include, among other things, “deliberations, decisions, or actions on
monetary policy matters, including discount window operations, reserves of member
banks, securities credit, interest on deposits and open market operations . . . , or
transactions made under the direction of the Federal Open Market Committee” 31 U.S.C. §
714 (b)(2)-(3). This prohibition limits GAO’s ability to audit the Federal Reserve’s actions
taken with respect to TALF. The Helping Families Save Their Homes Act of 2009, enacted
on May 20, 2009, amended Section 714 to provide GAO authority to audit the Federal
Reserve’s actions taken under Section 13(3) of the Federal Reserve Act, 12 U.S.C. § 343,
with respect to a single and specific partnership or corporation. See Pub. L. No. 111-22, 123
Stat. 1632, 1662-63, codified at 31 U.S.C. § 714(e). Among other things, this amendment
provides GAO with authority to audit Federal Reserve actions taken with respect to three
entities also assisted under TARP—Citigroup, Inc., American International Group, Inc., and
Bank of America Corporation—but does not provide GAO with authority to audit Federal
Reserve monetary policy actions taken with respect to TALF. We have raised this concern
about audit authority in previous testimonies and reports. See GAO, Troubled Asset Relief
Program: Status of Efforts to Address Transparency and Accountability Issues,
GAO-09-1048T (Washington, DC: Sept. 2009); GAO-10-16; and GAO-09-658.

8

See appendix VI for more information on the compliance program for TALF.

Page 3

GAO-10-25 Troubled Asset Relief Program

and requirements for borrowers and collateral.9 In addition, we analyzed
agreements between the relevant agencies on TALF. We interviewed
officials from the federal agencies responsible for the program, along with
policy analysts, economists, and attorneys. We also interviewed current
and potential TALF program participants, including investors,
underwriters, issuers, and sponsors. For additional information on the
scope and methodology for this engagement, see appendix I.
We conducted this performance audit from May 2009 to January 2010 in
accordance with generally accepted government auditing standards. Those
standards require that we plan and perform the audit to obtain sufficient,
appropriate evidence to provide a reasonable basis for our findings and
conclusions based on our audit objectives. We believe that the evidence
obtained provides a reasonable basis for our findings and conclusions
based on our audit objectives.

Background

Securitization is a process by which similar debt instruments—such as
loans, leases, or receivables—are aggregated into pools, and interestbearing securities backed by such pools are then sold to investors. These
ABSs provide a source of liquidity for consumers and small businesses
because financial institutions can take assets that they would otherwise
hold on their balance sheets, sell them as securities, and use the proceeds
to originate new loans, among other purposes.10 During the recent financial
crisis, the value of many ABSs dropped precipitously, bringing the
securitization markets to a virtual halt. As a result, households and small
businesses found themselves unable to access the credit that they needed
to, among other things, buy homes and expand inventories.
TALF was designed to reopen the securitization markets in an effort to
improve access to credit for consumers and businesses. The program
provides loans to certain institutions and business entities in return for
collateral in the form of securities that are forfeited if the loans are not
repaid. To assist in this effort, Treasury provides credit protection for the
program as part of TARP’s Financial Stability Plan under the Consumer

9

Eligible asset classes are those types of ABSs accepted for TALF. The amount of equity
borrowers are required to hold in the securities pledged as collateral to secure TALF loans
is calculated as a percentage of the pledged collateral’s value (either its par value, market
value, or a value determined from FRBNY internal risk assessments).

10

For additional information on the securitization process, see appendix III.

Page 4

GAO-10-25 Troubled Asset Relief Program

and Business Lending Initiative (CBLI). Treasury has pledged $20 billion
for TALF LLC—a special purpose vehicle created by FRBNY—to purchase
the underlying collateral associated with TALF loans in the event the loans
are not repaid. Because of Treasury’s role in protecting these taxpayer
funds committed through TARP, Treasury has consulted with the Federal
Reserve on TALF’s design. Under TALF, FRBNY is currently authorized to
extend up to $200 billion in nonrecourse loans to eligible borrowers
pledging eligible collateral. TALF is authorized to extend new loans
against nonmortgage-backed ABSs and legacy CMBS collateral until March
31, 2010, and against newly issued CMBS collateral until June 30, 2010. As
of December 2009, FRBNY has made about $61.6 billion in loans under
TALF. Of that amount, $47.5 billion in TALF loans remained outstanding as
of the end of December 2009. The amount of loans outstanding may be less
than the amount of loans extended due to loan prepayments by the TALF
borrower or paydown of principal.
TALF accounts for a small proportion of TARP funds (see fig. 1). As of
December 31, 2009—of the $20 billion committed—Treasury had loaned
TALF LLC $100 million: $16 million for administrative expenses and $84
million for potential asset purchases. This amount is less than 1 percent of
the $25.3 billion apportioned to the CBLI program, which itself represents
5 percent of apportioned TARP funds. TALF will receive TARP funds
beyond the $100 million already loaned if additional funding is required by
TALF LLC to purchase surrendered or seized collateral resulting from
unpaid TALF loans.

Page 5

GAO-10-25 Troubled Asset Relief Program

Figure 1: Status of TARP Apportioned Funds as of December 31, 2009
(Dollars in billions)

Consumer and Business Lending Initiative
Home Affordable Modification Program

$45.5

Public Private Investment Program

$68.0

Automotive Industry Financing Program

5%

$25.3

$91.8

8%

38%

12%

17%
21%

Othera

$115.0

Capital Purchase Program

Of the $25.3 in
Consumer and
Business Lending
Initiative funds
apportioned, less than
1%, or $100 million, has
gone to TALF.

$208.0

Source: GAO presentation of Treasury Office of Financial Stability data.
a

The “other” category consists of the Targeted Investment Program, American International Group
Investment, and the Asset Guarantee Program.
Note: Percentages may not total due to rounding.

FRBNY created TALF LLC, a special purpose vehicle, to purchase and
manage any assets that TALF borrowers surrender or the FRBNY seizes. 11
TALF LLC also holds any excess accumulated interest from TALF loans
and the $100 million funded portion of the Treasury loan for administrative
expenses and collateral purchases, plus interest earned from permitted
investments.
A portion of the interest earned by FRBNY on all TALF loans—called the
“excess interest”—is paid to TALF LLC as a fee for TALF LLC’s
commitment to purchase the assets received by FRBNY in conjunction
with a TALF loan. As of December 31, 2009, TALF LLC had accumulated
approximately $198 million in excess interest, with roughly $30 million
added each month (according to FRBNY officials), based on the current
loan portfolio.

11

Special purpose vehicles are legal entities, such as limited liability companies, created to
carry out a specific financial purpose or activity.

Page 6

GAO-10-25 Troubled Asset Relief Program

If accumulated fees and interest earned on TALF LLC’s investments are
insufficient to cover any asset purchases, Treasury will provide additional
TARP funds to TALF LLC to finance up to $20 billion of asset purchases.
Subsequently, TALF LLC will finance any additional purchases by
borrowing funds from FRBNY. The TARP loan is subordinate to the
FRBNY loan, thus the TARP funds provide credit protection to FRBNY.

Asset Classes Eligible for
Use as TALF Collateral

When TALF was first announced, the Federal Reserve made a number of
asset classes eligible for use as collateral in consultation with Treasury,
adding more as the program evolved (see table 1). Initially, securities
backed by automobile, credit card, and student loans, as well as loans
guaranteed by the Small Business Administration (SBA), were deemed
eligible because of the need to make credit in these sectors more widely
available.12 For most TALF-eligible collateral, FRBNY will stop providing
new TALF loans in March 31, 2010, while new-issue CMBSs will be
accepted as collateral on new TALF loans through June 30, 2010.

Table 1: Timetable of Asset Classes Accepted As Eligible Collateral
Asset class

Description

Auto, credit card
receivables, student
loans, and small
a
business

•

•
•
•

Mortgage servicing
advances, business
equipment, vehicle
fleet, and floor planb

•

•
•

•

Date announced

Auto includes retail loans and leases relating to cars, light trucks,
motorcycles, and other recreational vehicles originated on or after October 1,
2007.
Credit card receivables include both consumer and corporate credit card
receivables maturing in 2009 or the first quarter of 2010.
Student loans include federally guaranteed (including consolidation loans)
and private student loans with a disbursement date on or after May 1, 2007.
Small business includes fully guaranteed SBA 7(a) and 504 loans,
debentures or pools originated on or after January 1, 2008.

November 2008

Mortgage servicing advances are receivables created by principal and
interest, tax and insurance, and corporate advances made by Fannie Mae- or
Freddie Mac-approved residential mortgage servicers for mortgages
originated on or after January 1, 2007.
Business equipment includes retail loans and leases originated on or after
October 1, 2007.
Vehicle fleet include commercial and government fleets and commercial
loans secured by vehicles and the related fleet leases and subleases of such
vehicles to rental car companies, all originated on or after October 1, 2007.
TALF-eligible floor plan ABS include revolving lines of credit to finance dealer
inventories originated on or after January 1, 2009.

March 2009

12

For additional information on these asset classes, see appendix IV.

Page 7

GAO-10-25 Troubled Asset Relief Program

Asset class
Insurance premium
finance, new-issue
CMBSs, and legacy
CMBSsc

Description
Insurance premium finance includes loans originated for the purposes of
paying premiums on property and casualty insurance originated on or after
January 1, 2009.
•
New-issue CMBSs issued on or after January 1, 2009.
•
Legacy CMBSs issued before January 1, 2009.
•

Date announced
May 2009

Source: GAO presentation of information gathered from the FRBNY Web site.
a

Under the 7(a) program, SBA provides lenders with guarantees on up to 85 percent of the value of
loans to qualifying small businesses in exchange for fees to help offset the costs of the program.
Under the 504 program, which generally applies to small business real estate and other fixed assets,
SBA also provides certified development companies with a guarantee on up to 40 percent of the
financing of the projects’ costs in exchange for fees, while the small business borrowers and other
lenders provide the remaining 60 percent of the financing with no guarantee. For additional
information on these programs, see GAO, Small Business Administration’s Implementation of
Administrative Provisions in the American Recovery and Reinvestment Act, GAO-10-507R
(Washington, DC: Apr. 16, 2009).

b

Floor plan loans are made to auto and nonauto dealers to finance their inventories.

c

CMBSs are securities backed by mortgages for commercial real estate, such as office buildings or
shopping centers.

All TALF-eligible ABSs must be denominated in U.S. dollars, must be rated
AAA by at least two TALF-eligible nationally recognized statistical rating
organizations (NRSRO), must not have a credit rating below the highest
investment-grade rating category from a TALF-eligible NRSRO, and must
not be on review for a potential rating downgrade.13 In general, borrowers
must be U.S.-based businesses, investment funds, or U.S.-insured
depository institutions, although foreign banks with U.S. branches that
maintain reserves with a Federal Reserve bank are also eligible. However,
all or substantially all of the eligible collateral’s underlying credit
exposures must be
•

for newly issued ABSs—originated by U.S.-organized entities or
institutions, or U.S. branches or agencies of foreign banks—and

13
Nationally recognized statistical rating organizations, which are registered with the
Securities and Exchange Commission (SEC), have designed credit ratings to provide
investors with information about a security’s credit quality. SEC requires the agencies to
undergo an initial application review and subsequent examinations. The rating scales are
comparable, ranging from AAA, or its equivalent (excellent), to D, or its equivalent (poor).
Anything rated BBB or higher is considered investment grade, and anything below BBB is
considered noninvestment grade or speculative. Credit ratings are assigned without regard
to whether a security is TALF-eligible or not. The TALF-eligible rating agencies are Fitch
Ratings; Moody’s Investors Service; Standard & Poor’s; DBRS, Inc.; and Realpoint LLC.

Page 8

GAO-10-25 Troubled Asset Relief Program

•

for all ABSs—made to U.S.-domiciled obligors or located in the U.S. or one
of its territories, in the case of real property.
Interest rates for TALF loans are either fixed or floating and vary
according to the collateral securing the loan, as has been determined by
FRBNY.
In order to constitute eligible collateral for a TALF loan, both the issuers
and sponsors of the proposed collateral must provide certification
documents stating that, among other things,

•

they have reviewed TALF’s terms and conditions;

•

the collateral is TALF-eligible;

•

an independent accounting firm was provided consent, in certain cases, to
contact the TALF compliance fraud hotline if it suspects fraud or illegal
acts; and

•

purchasers of the securities that are affiliated with the originators, issuers,
or sponsors cannot use these securities as TALF collateral.14
In addition, external auditors review certain representations made by
issuers and sponsors about the TALF eligibility of the collateral.15 If any of
these representations change, the issuer and sponsor must provide public
notice. If any of the certifications are found to be incorrect, TALF LLC and
FRBNY can recover damages and the issuer and sponsor will be subject to
review by Treasury, the Special Inspector General for TARP (SIGTARP),
and GAO.16

How TALF Works

A number of entities help administer the TALF program.

14

An issuer is an entity that sells ABSs backed by its assets and can be a special purpose
vehicle or similar legal entity. A sponsor is an entity that facilitates such a transaction. For
additional information on the securitization process, see appendix III.

15

For more information on the certification process and the role of auditors, see appendix
VI.

16

Congress authorized SIGTARP as an oversight entity for TARP. SIGTARP has made
numerous recommendations to improve the implementation of TARP and TALF. See EESA
§ 121, 122 Stat. at 3788 (codified at 12 U.S.C. § 5231).

Page 9

GAO-10-25 Troubled Asset Relief Program

•

TALF agents, which are primary dealers or designated broker-dealers that
operate under FRBNY’s Master Loan and Security Agreement. The agents’
responsibilities include conducting due diligence on TALF borrowers and
making representations to FRBNY regarding eligibility of TALF borrowers
and their collateral, submitting TALF loan requests and supporting
documentation to FRBNY and the TALF custodian on behalf of borrowers,
delivering administrative fees and collateral from TALF borrowers to
FRBNY, and distributing the TALF borrower’s share of principal and
interest payments paid on the collateral backing the TALF loan.17

•

The Bank of New York Mellon, which serves as custodian of the program
and is responsible for administering TALF loans, holding and reviewing
collateral, collecting payments and administrative fees, disbursing cash
flows, maintaining the program’s books and records, and assisting other
TALF entities with the pricing of collateral.

•

Collateral monitors—Trepp LLC and Pacific Investment Management
Company LLC (PIMCO)—which check the pricing and ratings of
securities; provide valuation, modeling, reporting, and analytical support;
and advise on related matters.

•

CW Capital, which provides underwriting advisory services related to
certain commercial mortgage loans backing newly issued CMBSs.
FRBNY announces monthly subscription periods, during which potential
borrowers apply for loans and funds are disbursed. FRBNY has a
precertification process to streamline the process for certain eligible
borrowers. TALF precertification documents indicate that borrowers must
be deemed to be top-tier financial entities—that is, they must be seen as
industry leaders and be ranked among the largest entities in the industry
or have some of the largest operations.18 During our audit, FRBNY officials
told us that they review loan requests from all borrowers that meet general
eligibility criteria, and that not all borrowers need to be precertified;
indeed, most are not precertified. Applicants must work through a TALF
agent throughout the application process. Because the agents must

17

The primary dealers are 18 banks and securities broker-dealers that trade in U.S.
government securities with FRBNY. FRBNY had added 4 more broker-dealers to serve as
agents for borrowers accessing TALF, and as of December 31, 2009, there were a total of 22
TALF agents, but currently not all have executed the Master Loan and Security Agreement.

18

These documents also indicated that FRBNY created a proprietary list of such “top-tier”
entities for its own internal proposes and has not shared it with the TALF agents.

Page 10

GAO-10-25 Troubled Asset Relief Program

demonstrate that they know a potential borrower and must vouch for its
reputation, they put applicants through a “Know Your Customer” program
based on provisions in the Patriot Act.19 Once this process is completed,
TALF agents submit a loan request package to FRBNY that includes:
•

borrower identification information, such as name, address, and tax
identification number;

•

loan information, such as the requested loan amount ($10 million
minimum), the term of the loan, the loan rate, and the type of interest rate
(fixed or floating) that corresponds to the type of collateral offered;

•

collateral information, such as the CUSIPs of the securities,20 asset class
and subclass, price and face value of the collateral, the weighted average
life of the collateral,21 and the haircut amount—a percentage assigned to
the loan based on the asset class, or subsector where appropriate, of the
collateral and its weighted average life;

•

any appropriate filing documents, including a prospectus and offering
documents of the securities expected to be pledged; and

•

proof of purchase for the ABSs and CMBSs that are being offered as
collateral.
Next, the Bank of New York Mellon and FRBNY verify the data that the
TALF agents provide and, among other things, ensure that the ratings
submitted for the securities are the most recent. For legacy CMBS
collateral, FRBNY evaluates whether the price the TALF borrower paid

19

The 2001 USA Patriot Act established new and enhanced measures to prevent, detect, and
prosecute money laundering and terrorism. One of the more important measures for
financial institutions was addressed in section 326, Verification of Identification, more
commonly referred to as “Know Your Customer.” Pub. L. No. 107-56, 115 Stat. 272 (Oct. 26,
2001).
20

CUSIP stands for the Committee on Uniform Security Identification Procedures. CUSIP
numbers are alphanumeric identifiers assigned to individual securities.

21
In general, the weighted average life is the average number of years that unpaid principal
is outstanding. The TALF program’s definition varies according to asset class and
assumptions that FRBNY developed about the likelihood of prepayment for the different
ABS classes. The weighted average life for CMBSs is based on the assumption that each
loan amortizes according to schedule and pays in full on its maturity date, without
prepayment. FRBNY’s Web site has further details and the formulas used to compute the
weighted average life of collateral.

Page 11

GAO-10-25 Troubled Asset Relief Program

was reasonable based on pricing information FRBNY receives from the
collateral monitors and Bank of New York Mellon.22 The collateral
monitors, Trepp LLC and PIMCO, also conduct stress tests on pledged
legacy CMBS collateral to help ensure that the loan amounts will not
exceed the stressed value of the pledged securities. Three weeks prior to
each ABS subscription, a newly issued ABS undergoes a risk assessment
by FRBNY (with PIMCO’s support) to determine if it is likely to be
accepted as TALF collateral. The issuer must provide FRBNY any
information it provided the NRSROs so FRBNY can conduct its own credit
risk assessment. The issuer must also consent to permitting NRSROs to
discuss all aspects of the rating with FRBNY, including credit quality of the
ABS, modeling, and methodology, among other things.
On each TALF loan’s settlement date, the borrower must deliver the loan
collateral and administrative fees to FRBNY’s custodian, the Bank of New
York Mellon, which holds the collateral for the life of the loan. The
administrative fees vary by asset class and cover FRBNY’s administrative
costs for the facility.23

Determining the Loan Amount

If FRBNY deems the collateral eligible, it will lend an amount calculated
by subtracting a designated haircut percentage from the lesser of either
par or market value of the pledged collateral (or, in the case of legacy
CMBSs, a value based on an internal risk assessment). This percentage, or
“haircut,” in effect sets the amount of equity the TALF borrower holds in
the collateral. Haircuts vary by FRBNY’s assessment of market risks for
each sector and subsector.24 The haircut represents the difference between
the value of the proposed collateral and the value of the loan (table 2).25

22

Trepp LLC and PIMCO assist FRBNY by providing valuation, modeling, analytics and
reporting, and advising on these matters. Trepp LLC focuses only on CMBSs. PIMCO
analyzes both mortgage-backed and nonmortgage-backed ABSs.

23

Administrative fees due on the settlement date are equal to 10 basis points (bps) or 0.10
percent of the loan amount for nonmortgage-backed ABS collateral, and 20 bps for CMBS
collateral.
24

A sector is a group of securities that are similar with respect to maturity, type, rating,
industry, or coupon.
25

Generally, the percentage is applied to the lesser of the market value of the collateral or
the par value. In addition, to account for potential risks in legacy CMBSs, the
FRBNY conducts a risk assessment of pledged legacy CMBS collateral and applies the
haircut to the least of the market-, par-, or assessment-values. For more details, see
http://www.newyorkfed.org/markets/talf_faq.html.

Page 12

GAO-10-25 Troubled Asset Relief Program

Table 2: TALF Haircuts by Asset Class
Weighted average life in years for ABSs
Sector
Auto

0 to less
than 1

1 to less
than 2

10%

Subsector

11%

Prime retail lease

2 to less 3 to less
than 3
than 4
12%

13%

4 to less
than 5
14%

5 to less 6 to less
than 6
than 7
–

–

Prime retail loan

6

7

8

9

10

–

–

Subprime retail loan

9

10

11

12

13

–

–

Motorcycle or other
recreational vehicles

7

8

9

10

11

–

–

Commercial and
government fleets

9

10

11

12

13

–

–

12

13

14

15

16

–

–

Rental fleets
Credit Card
Equipment
Floor plan

Prime

5

5

6

7

8

–

–

Subprime

6

7

8

9

10

–

–

Loans and leases

5

6

7

8

9

–

–

Auto

12

13

14

15

16

–

–

Nonauto

11

12

13

14

15

–

–

Premium finance

Property and casualty

5

6

7

8

9

–

–

Servicing
advances

Residential
mortgages

12

13

14

15

16

–

–

Small business

SBA loans

5

5

5

5

5

6

6

Student loan

Private

8

9

10

11

12

13

14

Government
guaranteed

5

5

5

5

5

6

6

New-issue
CMBSs

15

15

15

15

15

–a

–a

Legacy CMBSs

15

15

15

15

15

–b

–b

Source: GAO presentation of information gathered from the FRBNY Web site.

Note: For ABSs benefiting from a government guarantee with average lives of 5 years and beyond,
haircuts will increase by 1 percentage point for every 2 additional years (or portion thereof) of average
life at or beyond 5 years. For all other ABSs with average lives of 5 years and beyond, haircuts will
increase by 1 percentage point for each additional year (or portion thereof) of average life at or
beyond 5 years.
a

For newly issued CMBSs with average lives beyond 5 years, collateral haircuts increase by 1
percentage point of par for each additional year of average life. No newly issued CMBS may have an
average life of more than 10 years.

b

For legacy CMBSs with average lives beyond 5 years, haircuts increase by 1 percentage point of par
for each additional year of average life. No legacy CMBS may have an average life of more than 10
years.

Loan Repayment Process

As discussed earlier, the Bank of New York Mellon holds the collateral
throughout the life of the loan. As the collateral securities generate cash
flows, the Bank of New York Mellon makes all principal and interest

Page 13

GAO-10-25 Troubled Asset Relief Program

payments to FRBNY on behalf of the TALF borrower. The borrower may
earn returns from the collateral after all loan obligations have been
satisfied. As shown in figure 2, any returns that the collateral assets earn
beyond the required principal and interest payments is delivered to the
borrower via the TALF agent after all monthly loan payment obligations
are met.26 FRBNY retains a portion of the interest that is calculated using
Overnight Indexed Swap (OIS) rates plus 25 basis points (approximately
the FRBNY’s cost of funds).27 The remaining part of the TALF loan interest
payment is transferred to TALF LLC, which would use the funds to
purchase surrendered collateral before accessing funds from Treasury.
This accumulated interest from TALF loans is held in an account called
the cash collateral account, and the Bank of New York Mellon is
authorized to invest these funds on behalf of TALF LLC to earn interest
income.
Figure 2: TALF Loan Origination and Repayment Process

Custodian and administrator
Bank of New York Mellon

Assets pledged as collateral

1

Principal and interest payment
TALF
borrower

TALF loan

FRBNY TALF account

(flow of TALF loan from
FRBNY TALF account)

Principal and
interest received

3

2
Portion of interest payment retained by FRBNY

FRBNY

3a
3b

TALF
agent

Excess interest
paid to TALF
LLC
TALF LLC

3c

Returns from collateral after
meeting loan payment obligationsa

Source: GAO.

26

Under certain circumstances, the excess interest earned on ABS or CMBS above the
TALF loan interest payable will be remitted to the TALF borrower only up to a proportion
of the original haircut amount, with the remainder of such amount applied to the TALF
loan principal. FRBNY refers to this as a net carry diversion, and it accelerates the
repayment of the TALF loan and ensures that the TALF borrower maintains a significant
equity stake in the ABS or CMBS collateral over the life of the loan.

27

A basis point represents .01 percent, or one-hundredth of a percent. In other words, 100
basis points equals 1 percent. For example, the difference between 1.00 percent and 1.50
percent would be expressed as 50 basis points. The OIS rate is a type of interest rate swap
that is based on daily federal funds rates.

Page 14

GAO-10-25 Troubled Asset Relief Program

a

If the collateral is not generating enough interest to satisfy the loan obligation, the borrower must
eventually make up the difference or surrender collateral.

TALF offers nonrecourse loans, which allow borrowers to walk away—or
stop paying a loan—with no personal exposure for the unpaid portion of
the debt.28 In this case, borrowers would surrender their collateral through
a TALF agent, which would submit a collateral surrender form to FRBNY.
Within 10 days of receiving the surrender notice, the Bank of New York
Mellon cancels the outstanding balance of the loan and transfers the
related collateral to FRBNY. FRBNY has the option to sell the collateral to
TALF LLC (see fig. 3) at a price equal to the then outstanding principal
amount of the TALF loan plus accrued but unpaid interest. This process
has several steps:

Process If Borrower Walks
Away from the Loan

•

FRBNY sends a purchase notice to TALF LLC. To purchase surrendered
assets, TALF LLC first uses funds that have been accumulating in the cash
collateral account (as of December 31, 2009, the account contained
approximately $198 million).29

•

When these funds are exhausted, TALF LLC borrows money to purchase
any surrendered assets. Its first source of borrowed funds is Treasury,
which has committed up to $20 billion in TARP funds to the TALF program
for this purpose. To obtain these funds, TALF LLC must submit a request
to Treasury (as the subordinate lender) one business day prior to the
desired funding date. Treasury’s loan rate is equal to the 1-month London
Interbank Offered Rate (LIBOR) rate plus 300 basis points.

•

If the $20 billion TARP loan commitment is fully exhausted, TALF LLC
must ask FRBNY for a loan to purchase any additional surrendered assets.
FRBNY has committed to provide up to $180 billion to TALF LLC for this
purpose. TALF LLC must submit a borrowing request to FRBNY (as senior
lender) one business day prior to the desired funding date. The FRBNY’s

28

TALF loans are nonrecourse, except in cases involving fraud or other breaches of certain
representations.

29
This amount excludes the $84 million (initial $100 million loan from Treasury less the $16
million set aside for administrative expenses), which is held in a separate sub-account.
Therefore, total funds that could potentially be available to purchase surrendered collateral
could be greater, according to Treasury officials, though additional TALF LLC expenses
related to surrendered collateral could also be taken from this amount. For that reason, we
refer to $198 million as the amount in TALF LLC that is dedicated to purchasing
surrendered assets.

Page 15

GAO-10-25 Troubled Asset Relief Program

•

loan rate is equal to the 1-month LIBOR rate plus 100 basis points. The
principal on FRBNY’s loan would be repaid before the principal of the $20
billion TARP loan.

Figure 3: Funding Required if a Borrower Walks Away from a Loan

TARP
Sub
loan

Loan
repaymentb
TALF LLC

TALF
borrower

Custodian
BNYM

Form

FRBNY

+

Form
surrendering
collaterala

ral

Collate

ral

Collate

Cash

Senior loan

Funds for purchasing
collateral, in order of use:
- excess accumulated
interest
- $20 billion sub loan (TARP)
- $180 billion senior loan (FRBNY)

Repayment

Source: GAO.
a

The form goes through the TALF agent and is provided to the custodian and FRBNY
simulataneously.

b

To be supplemented by 90 percent of any remaining TALF LLC assets after meeting all obligations,
as noted in the order of payment shown below.

Alternatively, if a TALF borrower stops payment on a loan and does not
submit a collateral surrender form, under certain circumstances FRBNY
has rights to seize the collateral.

Management of TALF LLC

With the Federal Reserve’s and Treasury’s agreement, TALF LLC may
dispose of assets it has acquired. Agency officials told us that there were
no formal guidelines on when to sell any acquired assets and that the
decision to sell would be made on a case-by-case basis as is necessary
given the differences in types of assets that could be acquired by TALF
LLC. However, they added that factors such as market rates and the nature
of the underlying collateral would likely play a role in any decision to hold
or sell assets purchased by TALF LLC. According to Treasury officials,
because the Federal Reserve’s monetary policy considerations may not
align with Treasury’s investment interests in the TALF program, Treasury
plans to obtain independent advice on the disposition of investments.

Page 16

GAO-10-25 Troubled Asset Relief Program

TALF LLC has certain sources of cash inflows, such as the excess interest
from TALF loans, interest paid from permitted investments, principal and
interest paid on ABS holdings, and proceeds from possible asset sales of
ABS holdings. These inflows are distributed according to the order defined
in an agreement among Treasury, FRBNY, the Bank of New York Mellon,
and TALF LLC. The payment order is as follows:
•

Pay TALF LLC expenses.

•

Fund expense reimbursement account up to $15 million.

•

Repay outstanding principal on any FRBNY senior loans to TALF LLC.

•

Fund the cash collateral account up to the senior loan commitment
(currently $180 billion).

•

Repay outstanding principal on any Treasury loans.

•

Repay FRBNY loan interest.

•

Repay Treasury loan interest.

•

Repay any other secured obligations that may arise that have not been
specified yet by the agencies.

•

Pay Treasury and FRBNY (90 percent and 10 percent, respectively) any
residual amounts but only after the above requirements are satisfied.
As of December 2009, cash inflows have been used to pay TALF LLC
expenses and fund the expense reimbursement and cash collateral
accounts.

TALF Includes
Features That
Mitigate Potential
Losses, but CMBSs
Could Pose Greater
Risks Than Other
Asset Classes

Under some scenarios, TALF borrowers may have economic incentives to
stop payment on their loans and surrender collateral. Treasury and FRBNY
assessments, along with our analyses, however, suggest that a number of
factors reduce the likelihood of this occurring. First, certain TALF features
are designed to protect TARP funds and also limit taxpayer exposure.
Specifically, FRBNY officials told us that risks in TALF are managed based
on four pillars: credit protection, credit ratings, FRBNY due diligence, and
market discipline. FRBNY also sought market perspectives on the level of
returns from TALF-eligible securities that would be required to encourage
market participation while also ensuring proper due diligence by TALF
participants, which in turn reduces the risks to taxpayers. Second, most

Page 17

GAO-10-25 Troubled Asset Relief Program

ABSs issued since the credit crisis began contain features such as
increased levels of subordination and overcollateralization that reduce the
likelihood that TALF borrowers will stop payment on their loans. Third,
Treasury and FRBNY analyses project minimal, if any, likelihood that
TARP funds will be used for TALF-related purchases, and Treasury
currently projects a profit from TALF. While TALF poses minimal risks to
TARP even in adverse market conditions, our analyses showed that
CMBSs held as collateral as of September 2009 potentially pose higher
risks than ABSs and under adverse conditions losses could exceed $500
million. TALF presents a range of other taxpayer risks beyond those
presented to TARP funds. Such risks include the risk that FRBNY might
not identify instances of material noncompliance with program
requirements by TALF participants. However, because of statutory audit
limitations on GAO, this review has been limited to presenting only
descriptive information about FRBNY’s role in TALF and we cannot
evaluate FRBNY’s TALF operations.

Some Scenarios Could
Provide TALF Borrowers
with Economic Incentives
to Stop Payment on Their
Loans

As of January 8, 2010, there have been no collateral surrenders by any
TALF borrowers, but in some instances, a TALF borrower could have an
economic incentive to stop payment on a loan and surrender the
underlying collateral.30 A number of scenarios could result in a borrower
walking away from a loan. For example, the collateral could lose value so
that the loan amount exceeded the value of the collateral. Or, the expected
returns from the collateral could be less than the cost of financing the
loan. Also, interest rates could rise across the board, decreasing the
market value of the collateral or making refinancing more difficult.
A borrower can lose equity in the collateral if the collateral’s value falls
below the outstanding loan balance. As discussed earlier, TALF’s
established haircuts determine the amount of equity borrowers have in
their collateral. This equity represents the amount of money that a TALF
borrower would lose by surrendering the collateral and not repaying the
loan.31 For example, if the ABSs a borrower seeks to use as collateral on a

30

As noted in the background, TARP funds can be used to purchase ABS or CMBS collateral
if not enough excess interest has accumulated in TALF LLC.

31

According to the terms of FRBNY’s Master Loan Security Agreement, TALF borrowers
can choose to surrender collateral in order to stop making payments on a loan. If they do
not voluntarily surrender collateral, FRBNY has the right to take possession of it. In either
case, TALF LLC is likely to purchase the associated collateral.

Page 18

GAO-10-25 Troubled Asset Relief Program

TALF loan were initially valued at $100,000, a haircut of 10 percent would
provide the TALF borrower with a $90,000 loan and require $10,000 of its
own funds to acquire the securities. This $10,000 represents the borrower’s
equity in the securities. If the value of the collateral were to decline—such
that the borrower could sell the securities only for some amount less than
$90,000—and especially if it declined dramatically, the borrower could
decide to cut further losses by stopping payments on the loan and
surrendering the collateral. The borrower is particularly likely to make this
choice if such a situation occurs at the point when the loan matures.
As TALF currently works, TALF borrowers earn the difference between
the ABS’s return and the cost of borrowing from FRBNY on the TALF loan,
multiplied by the inverse of the borrower’s percent equity stake. However,
if the returns on TALF collateral are less than initially anticipated, the
TALF loan costs could exceed them, providing another incentive to stop
making payments. This situation would be most likely to develop if the
underlying loans in the securities defaulted or otherwise failed to meet the
terms of the original loan agreement. In this situation, if the returns were
less than the total cost of the TALF loan—the interest and principal due—
a borrower might stop payment on the loan or surrender the collateral to
FRBNY.
Investors might also choose to surrender collateral at the maturity of their
TALF loan if overall interest rates on credit increased significantly above
levels available at the time the underlying securities were issued. A large
increase in interest rates would lower the market value of the securities,
especially for those with a fixed interest rate, because future cash flows
would be worth less. In this scenario, a borrower could wind up owing
more to FRBNY at maturity than the securities were worth. The borrower
would need to raise funds at higher interest rates to pay back the TALF
loan, but the collateral would be worth less than the new loan, potentially
making it difficult to find a lender. This situation would provide an
economic incentive for the borrower to surrender the collateral and walk
away from the loan. Nevertheless, many of the longer-term securities
pledged as collateral for TALF loans have floating interest rates. Generally,
floating rate securities do not decline in value when interest rates increase
because the interest rates on the securities also increase.
TALF borrowers will not necessarily stop payment on a loan even with
one or more of these economic incentives. For example, a market
participant and a Treasury official told us that TALF borrowers wanted to
avoid “reputation risk” by not walking away from their loans—even if it
might be in their financial interest to do so. By continuing to pay on their

Page 19

GAO-10-25 Troubled Asset Relief Program

loans, even at some loss, borrowers could protect their reputations. Other
market participants we spoke with, however, did not mention concerns
about reputation, though they thought it was unlikely that TALF borrowers
would stop payment on their loans.
Moreover, FRBNY and Treasury officials believe that the most important
disincentive for borrowers to stop payment on a TALF loan and surrender
collateral is maintaining a positive return based on the difference between
the cost of the TALF loan and the return of the underlying ABS collateral.
Even if the value of the collateral declines and removes the borrower’s
equity, these officials stated that borrowers were not likely to walk away
from the loan if they were still receiving positive cash flow from the asset.
While this assumption may be reasonable, it is conceivable that investors
subject to mark-to-market accounting might choose to walk away from an
ABS that was still paying the required scheduled interest and principal
payments but had lost sufficient market value. Investors would be most
likely to walk away if they owed significantly more on the loan than the
current value of the asset so that walking away from the asset would have
positive implications for the borrower’s reported profitability. As of
January 8, 2010, no TALF borrowers had surrendered collateral to FRBNY.
Nevertheless, because the behavior of individual borrowers is difficult to
predict, it remains unclear whether and why borrowers might stop
payment on TALF loans based on their own investment strategies and
other objectives.

TALF Has Several Features
That Likely Protect
Taxpayers

Certain TALF features help protect TALF funds and also taxpayer
exposures through TARP. We discussed these protections with FRBNY
officials but did not evaluate FRBNY compliance with them because of the
limitation of our statutory audit authority. FRBNY officials told us that
risks in TALF are managed based on four pillars: credit protection, credit
ratings, FRBNY due diligence, and market discipline.

•

Credit protection is provided primarily by the borrower’s equity in the
security (set by the haircuts) and a portion of the interest rate on TALF
loans that provides accumulated excess interest in TALF LLC. FRBNY
officials said that haircuts were designed to approximate multiples of
stressed historical impairment rates for ABSs. The size of the haircut is
important, because if it is set too low, the borrower will have less equity in
the collateral and, in the circumstances that we have discussed, could
have an incentive to walk away. The excess interest from TALF loans
accumulated in TALF LLC also protects taxpayers, because this money
would be used before TARP funds to purchase collateral surrendered from

Page 20

GAO-10-25 Troubled Asset Relief Program

unpaid loans. FRBNY officials said that such features were designed to
ensure that the haircuts and excess interest would result in no credit
losses for Treasury or the Federal Reserve.
•

The Federal Reserve requires that TALF collateral be rated AAA or its
equivalent by two of the nationally recognized statistical rating
organizations that it deems eligible to provide credit ratings for TALF,
among other requirements for credit ratings.32 Collateral that is offered
that has a lower rating is not eligible for TALF. The rating requirement
helps ensure that the securities TALF accepts as collateral present
minimal credit risks.

•

FRBNY due diligence serves as another pillar of taxpayer protection. As
discussed earlier, FRBNY, with the support of its collateral monitors,
reviews the credit risks related to individual assets FRBNY might consider
accepting as TALF collateral. In addition, for legacy CMBSs, FRBNY
reserves the right to reject any collateral and has not disclosed its
selection criteria to reduce the likelihood that only the poorest-performing
collateral is put forward for TALF loans.

•

The final pillar, market discipline, includes the TALF borrowers’ due
diligence conducted on the risks related to the underlying collateral, given
their equity in the collateral, as set by the haircut. Such discipline includes
reviewing the prospectuses and understanding how the structure of the
underlying securities impacts its overall risks. Investors purchasing ABSs
would generally review the terms of the security, the anticipated risks, and
the likely return. In addition, TALF borrowers help to facilitate price
discovery.
These design features are intended to help ensure that TALF borrowers
hold equity in the underlying TALF collateral, that such collateral is highly
rated, and that TALF borrowers carry out the same due diligence on TALF
collateral that they would conduct on any other security.

32

See background for additional detail on the rating requirements for TALF collateral.

Page 21

GAO-10-25 Troubled Asset Relief Program

Rates of Return for TALF
Borrowers on TALF
Collateral Have Declined
for Most TALF Collateral
from Their Highs Earlier in
2009

As noted earlier, TALF borrowers earn the difference between the ABS’s
coupon rate of return and the cost of borrowing from FRBNY on the TALF
loan. According to FRBNY officials, they, along with officials from the
Federal Reserve and Treasury, were aware of the importance of striking a
balance between achieving returns on equity that would encourage
program participation in the stressed market conditions when TALF was
announced, while also ensuring that investors had incentives for due
diligence and that the program would be less attractive during times of
less market stress.33 If borrowers saw an opportunity to earn excessive
returns due to a poorly designed program, borrowers might not conduct
appropriate due diligence on the underlying securities, which could put
taxpayers at risk. Moreover, because the loans are nonrecourse much of
the borrowers’ risk is shifted to FRBNY and Treasury, while most of the
earnings potential remains with borrowers. FRBNY gathered information
on the rates of return that would entice potential TALF participants by
surveying market participants about expected returns for TALF-eligible
asset classes.34
To understand the changes in returns on equity over time and the
reasonableness of those returns from the perspective of helping to ensure
that taxpayers were not subsidizing high returns, we analyzed fluctuations
in returns on TALF eligible collateral from March 2009 through September
2009. As seen in figure 4, returns generally decreased for select classes of
TALF-eligible collateral between the first TALF operation in March 2009
and September 2009, with limited exceptions.35

33

For TALF purposes, returns on equity are the earnings on the amount invested in a
particular security or the amount that an investor can expect from an investment.

34

Because of our audit limitations, we did not review the Federal Reserve’s process for
surveying the market on what returns on equity would encourage participation for a
program like TALF.

35
Our analysis focused on auto, credit card, and student loan ABSs, because these
securities make up the majority of the market and are the most commonly traded asset
classes.

Page 22

GAO-10-25 Troubled Asset Relief Program

Figure 4: Overall TALF Borrower Returns at Issuance for Select Asset Classes Accepted as TALF Collateral Generally
Decreased from March through September 2009
Prime auto loans
All

Auto leases

About 1 year

About 2 years

Greater than
2 years

All

A-2

Private
student loans

Credit cards
A-3

A-4

Prime

Subprime

Lapsed Exercise
option option

Percentage
50
43

40

37
32

31

30
23

7

7 6

9
6 6

2

8 7

11
5

11

10
7

-18

-18

-32

-14

18

18

14

7 6 7
5

13

0

-5

M
Ju ay
n
Se e
pt.

M
Ju ay
n
Se e
pt.

Ma
Ap r.
r
M il
Ju ay
ne
Ju
Se ly
pt.

Ma
Ap r.
r
M il
Ju ay
ne
Ju
Se ly
pt.

Ma
Ap r.
ril
Ma
Ju y
ne
Ju
ly
Se
pt.
Ma
Ap r.
ri
M l
Ju ay
ne
Ju
l
Se y
pt.
-19
Change in return
(percentage
points)

8
0

-5

-10

10

9

0
-3

15 15
10

14

5

1

1

18

17
11

-15

-11

-16

-8

-7

Ma
y
Ju
ly
Au
g.
Ma
y
Ju
ly
Au
g.

15

13

M
Ju ay
n
Se e
pt.

14

M
Ju ay
n
Se e
pt.

10

19

19

18

Ma
Ap r.
r
M il
Ju ay
ne
Ju
Auly
g
Se .
pt.
Ap
ri
Mal
Ju y
n
Au e
Se g.
pt.

21

20

26

25

-25

-18

Source: GAO analysis of TALF prospectuses, Standard and Poor’s and Moody’s Investors Service reports, and Federal Reserve data.

Note: The percentage point change in return for each asset class is calculated from the first month of
issuance to the most recent month and may not appear to total correctly due to rounding.

Most asset classes demonstrated a significant decline in returns, from a
peak of 43 percent among the first student loan ABS accepted under TALF
to lows of -5 percent for 1-year prime auto tranches36 and 2-year auto
leases, suggesting that if investors had used TALF to finance the purchase
of these auto-related TALF eligible securities they would have locked in

36

Some securitizations—such as ABSs backed by auto loans—are divided into different
segments, or tranches. A tranche is a piece of a securitization that has specified risks and
returns. For additional information on the securitization process and tranches within
securitizations, see appendixes III and IV.

Page 23

GAO-10-25 Troubled Asset Relief Program

losses.37 However, in a sign that health may be returning to the ABS
markets, no TALF borrowings were needed to finance the purchase of
these negative return securities as the issuances were fully funded by nonTALF investors. The most dramatic decreases in returns have been in the
auto loan tranches with longer-term maturities and private student loan
ABSs. The average expected return for TALF borrowers that used prime
auto loan ABSs as TALF loan collateral declined by nearly a fifth after
March 2009. All of the TALF-eligible private student loan ABS transactions
that were completed between May 2009 and August 2009 had a unique
option feature that significantly lessened investor’s expected returns on
equity.38 Additionally, the September 2009 increase in return for subprime
credit cards was primarily due to the unique structure of an issuance by a
large subprime issuer.39
The trend of decreasing returns on equity in the overall market for most
asset classes indicates that the returns required to attract investors have
decreased since TALF’s implementation, as ABS investors perceive the
assets to be less risky. In addition, the trend demonstrates that taxpayer

37

It should be noted that just because an ABS meets the TALF eligibility requirements, it
does not necessarily follow that investors will use TALF to finance their purchase. As part
of the price discovery process that the ABS underwriters engage in as part of the marketing
of a new issue of ABS, they interact with prospective buyers to achieve the lowest interest
rates on the offered bond tranches that will still sell the full offering. As the TALF eligibility
verification process must be completed prior to the issue date, it is possible that bonds will
have sufficient demand to drive the price up or bonds’ coupon rates (since ABSs are issued
at or very close to prices of 100 percent of par) down to a level near or below the TALF
finance rates paid to the Federal Reserve. This is why investors are increasingly unwilling
to use TALF to finance the purchase of TALF eligible securities. In addition, this trend
explains why more issuers are choosing to forego the process of seeking TALF eligibility as
they feel it is no longer necessary to ensure that their bonds will find buyers.
38

This option allowed the issuer to purchase back the bonds after a period of between 30
and 36 months at a significant discount on the face value. For one issuer, the discounts
were 7 percent for May 2009 and 6 percent for both July and August 2009. The August 2009
transaction from a new private student loan issuer had a 10 percent discounted exercise
price. Exercise by the issuer at this discounted price implicitly reduces the amount of
principal that the bondholder receives and negatively impacts the return. The issuer makes
up for this expected shortfall in principal payback with a higher interest rate or “coupon”
on the bond.

39
This subprime credit card issuance was unique in that two tranches had average lives of
nearly 5 years, which is longer than most TALF-eligible credit card ABS collateral.
Investors in ABSs with longer average lives require higher interest rates to compensate for
the greater risk of holding the bond over a longer time period. However, TALF loans for
these ABSs are still 3-year loans with the same interest rate as would be paid by a borrower
with a shorter-lived ABS. To some extent the increased return is compensation for the risk
the borrower is taking given the mismatch on the maturity of the loan and that of the ABS.

Page 24

GAO-10-25 Troubled Asset Relief Program

subsidies to TALF borrowers have not, over the course of the program,
provided what might be considered excessive returns. Although returns
were high in the beginning, they diminished once the ABS markets started
to revive and overall perceived risk began to decrease. Moreover, the trend
toward negative TALF returns on equity have coincided with a drop in
TALF participation by TALF borrowers in those asset classes as the rate
on the securities’ bonds falls below the loan rate charged by the FRBNY.
This occurrence is consistent with FRBNY’s intention for TALF financing
to become less attractive as the ABS market improves and can be viewed
as indicating normalization in the market, since investors can purchase
such ABSs without TALF funding.

New Securitizations Have
Generally Been Structured
with More Credit
Protections since the
Credit Crisis Began

Certain elements of the securities themselves also reduce the risk of loss
to TALF and the taxpayer. Among these features are credit enhancements,
which have increased since the credit crisis began in the second half of
2008. Credit enhancements are features in the structure of a securitization
that protect investors from losses due to defaults on the underlying loans
in the securities. Two main forms of credit enhancement include
subordination, which helps ensure that more highly rated tranches in a
security receive priority of payment, and overcollateralization, which
ensures that funds are available if a borrower stops paying or other credit
problems develop with the underlying loans. For additional details on the
various types of credit enhancement, see appendix V.
Credit enhancements offer several benefits. First, they provide a cushion
against losses, making it less likely that TALF borrowers will decide not to
repay their loans and surrender the collateral because of credit
performance problems in the ABS or CMBS markets. Second, credit
enhancements reduce the probability that TALF LLC will suffer principal
losses on surrendered ABSs and CMBSs from unpaid loans. For this
reason, credit enhancements also provide TALF LLC with an incentive to
decide to hold surrendered collateral to maturity. This reduces potential
losses to TARP funds, which would finance TALF LLC’s purchase of such
ABS and CMBS.
We reviewed the credit enhancements on every TALF-eligible ABS issued
between the program’s initiation in March 2009 and September 2009 and
compared them with credit enhancements on ABSs by the same issuer
between 2006 and 2008 to identify any changes. As shown in figure 5, the
level of enhancement for every TALF-eligible asset class increased.

Page 25

GAO-10-25 Troubled Asset Relief Program

Figure 5: Increases in Levels of Credit Enhancement on ABSs Since the Credit
Crisis Began, March 2009 through September 2009
TALF-eligible ABS

Percentage of credit
enhancement in ABS

Percentage increase (rounded)

Prime auto loans

10.5
7.9

33

Subprime auto loans

44.3
43.6

2

Prime credit cards

18.2
13.6

34

Subprime credit cards

27.5
22.6

22

Private student loans

35.3
8.9

295

Auto leases

25.3
9.7

162

Insurance premium finance

11.0
9.0

19

Motorcycle

23.0
10.3

124

Commercial fleet leases

12.5
12.1

3

Equipment

11.6
8.2

42

Mortgage servicing advances

22.7
18.9

20

Nonauto floor plana

24.0
9.0

167

TALF
Precrisis
Source: GAO analysis of TALF prospectuses and Moody’s Investors Service and Standard and Poor’s reports.
a

Nonauto floor plan loans are made to finance nonautomotive inventory such as recreational vehicles,
boats, and industrial equipment.

In particular, credit enhancements on ABSs backed by private student
loans, nonauto floor plans, auto leases, and motorcycle loans at least
doubled after the credit crisis.40 In other words, from the perspective of
protecting TARP funds, taxpayers would receive more than double the
amount of credit protection for these particular asset classes compared
with these issuers’ ABSs before the financial crisis. This increase was a
combination of market demands and credit rating agencies’ more stringent

40
Although there was a single auto floor plan ABS deal in September 2009, we could not
locate a prior issuance from this issuer to use as a comparison for precrisis levels of credit
enhancement; therefore, this asset class is not included in this analysis.

Page 26

GAO-10-25 Troubled Asset Relief Program

requirements for achieving AAA ratings on many ABSs. TALF agents and a
TALF issuer confirmed that credit enhancements had increased since the
onset of the financial crisis, even for non-TALF issuances, and noted that
requirements from the credit rating agencies had contributed to the
increases.

TALF Poses Minimal Risks
to TARP Even in Adverse
Market Conditions; CMBS
Risks are Potentially
Higher

While TARP funds designated for TALF are exposed to potential loss if
TALF LLC uses them to purchase ABSs or CMBSs used as collateral for
unpaid TALF loans, Treasury and FRBNY analyses suggest that the risks of
loss are minimal. According to the Federal Reserve’s analysis, the
accumulated excess interest from TALF loans will likely cover any ABS or
CMBS purchases for TALF LLC, and Treasury will not need to provide any
TARP funds for such purchases. Accordingly, the total expenditures from
TARP funds would include only the $100 million placed in TALF LLC,
which is in the form of a loan that would be repaid. According to the terms
of the agreement between FRBNY and Treasury, Treasury will receive 90
percent of the monies accumulated in TALF LLC when the program
expires. In particular, if TALF LLC has not purchased any collateral, the
excess interest that has accumulated—along with interest income from
investing such money—will go mainly to Treasury; therefore, Treasury
could potentially gain from its commitment to TALF if losses were
minimal.

Treasury Currently Expects to
Earn a Profit on TALF

Treasury hired a contractor to provide an estimate of potential losses to
TARP funds from TALF.41 According to one analysis conducted by the
contractor, any losses to TARP are likely to be far below the $20 billion
that has been set aside for TALF and in fact are unlikely to exceed about
$190 million. Any assets purchased by TALF LLC would first be paid for
with the excess interest and related interest income that had accumulated
in TALF LLC, which totaled almost $200 million at the end of December

41

Treasury asked the Bank of New York Mellon, as the administrator for TARP, to analyze
the risks that TALF posed to TARP funds. In turn, the bank hired a firm to conduct this
analysis, NSM Structured Credit Solutions, which is now known as RangeMark. The firm’s
responsibilities include determining how the various TALF asset classes perform and the
risks they present to TARP. The contractor made certain assumptions about the
composition of the portfolio and the risks of the various types of collateral, and estimated
losses in an extreme scenario. The results of this analysis contributed to Treasury’s recent
financial statement, which anticipates that its commitment of funds for TALF will actually
provide a return. See GAO-10-301.

Page 27

GAO-10-25 Troubled Asset Relief Program

2009.42 According to the contractor, the analysis of potential loss of TARP
funds did not include this excess interest, so the total estimated losses
could be largely offset by accumulated interest. In fact, the contractor and
Treasury officials said that a subsequent analysis that included projections
for the accumulation of excess interest—and is updated on a quarterly
basis—currently projects more than $1 billion in profit related to
Treasury’s TALF exposure.
To assess the reasonableness of Treasury’s position that TALF may
actually earn money rather than expose the taxpayer to any losses, we
reviewed the Treasury contractor’s model—which is central to estimating
losses for the various asset classes accepted in TALF. We found this model
appeared to incorporate generally reasonable loss assumptions for the
three asset classes that we reviewed and that comprise the largest portion
of the TALF portfolio: credit cards, auto loans, and student loans. That is,
many of the loss estimates were fairly conservative when compared to
recent historical results for these specific assets. For each asset class, the
contractor estimated expected loss percentages based on its own research
and analysis, which was used to generate total-loss estimates. The model
calculates total losses for each TALF borrower on each asset held in the
TALF portfolio. The portfolio includes current TALF collateral and
projections for future TALF collateral that borrowers will use for TALF
loans, based on information the contractor received from Treasury.
Potential Treasury losses from these various scenarios were calculated by
taking the total loss for the TALF borrower and subtracting the equity that
the TALF borrower holds in ABSs. Any difference is considered a loss,
first to TALF LLC and potentially to Treasury (through TARP funds) if
loans extended to TALF LLC for its purchase of TALF collateral are not
repaid. While some of the scenarios generated by the model estimated
losses from asset classes that did not have government guarantees, none
of the possible scenarios estimated losses to Treasury from CMBSs.

Uncertainty in the Commercial
Real Estate Market Could
Create the Potential for Losses
if Conditions Deteriorate

While Treasury has determined that CMBS-related losses are unlikely for a
number of reasons, our analysis shows that if the commercial real estate
markets were to be affected similarly to real estate markets in 2008, the
potential for loss exists under a worst-case scenario. Treasury and its
contractor said that CMBS losses were unlikely for a number of reasons
ranging from the relatively large haircuts (at least 15 percent) required on

42

As stated earlier, according to FRBNY officials the amount increases at approximately $30
million per month, based on the current loan portfolio.

Page 28

GAO-10-25 Troubled Asset Relief Program

CMBS loans to the level of credit enhancement associated with CMBSs
accepted for TALF. However, due in part to the recently weak economy,
commercial real estate continues to undergo price deterioration that
potentially poses risks to the TALF legacy CMBS portfolio and could lead
to increased delinquencies and defaults on commercial real estate
mortgages. For example, the potential risks that CMBSs pose to TALF, and
thus to TARP, are underscored by the fact that 63 percent of TALF’s CMBS
portfolio was underwritten in 2006 and 2007, when underwriting standards
were at their worst. Moreover, as figure 6 shows, commercial real estate
prices have been falling since early 2008, following the deterioration in the
overall U.S. economy, and shortly thereafter CMBS delinquencies began to
rise sharply. The Federal Reserve and Treasury have continued to note
their ongoing concerns about this segment of the market.
Figure 6: Commercial Real Estate Prices and CMBS Delinquencies from January 2004 through October 2009
Moody's/REAL Price Index

Percentage delinquency

200

5

190
Prices peaked
in October 2007.

180

4

170
160

3

150
140

2
As of November 2009, prices
show a drop of 43% since
peak. November was the first
month in which prices
increased since September 2008.

130
120

1

110
100

0
2004

2005

2006

2007

2008

2009

Year
Source: GAO presentation of Moody’s/REAL Commercial Property Price Index and Moody’s CMBS Delinquency Tracker data.

Note: Moody’s/REAL Commercial Property Price Index is a report that shows the change in
commercial real estate asset sales. Moody’s CMBS Delinquency Tracker is a monthly report showing
delinquency rates for commercial mortgage-backed securities.

In addition, prices on CMBSs experienced volatility between May 2008 and
November 2009. Even the highest credit quality AAA legacy CMBSs that
were used as collateral for TALF loans during the third quarter of 2009 had
dropped by one third, on average, between May and November 2008
during the most severe period of the credit crisis. While prices have

Page 29

GAO-10-25 Troubled Asset Relief Program

recovered since the expansion of TALF to include CMBSs in the second
quarter of 2009, market observers project that the sector will continue to
perform poorly into 2010.
In light of the ongoing distress in the commercial real estate market, we
analyzed the prices and values of 99 percent of the CMBS collateral
backing loans made by FRBNY during the third quarter of 2009. We
compared the prices at the time the loans were made with the lowest
prices in November 2008, a period of extreme stress for CMBSs. Our
analysis revealed that if legacy CMBSs accepted as TALF collateral as of
September 2009 reached market values equivalent to November 2008
levels, about 88 percent of such collateral would have fallen to levels at
which the TALF borrower’s equity would be eliminated.43 Moreover, more
than $3.5 billion owed by TALF borrowers—or about 85 percent of the
total value of TALF legacy CMBS loans—would have negative equity in
this scenario. This extreme market stress scenario would result in a loss in
market value on the part of these TALF borrowers of nearly $1.2 billion.
The haircut investment for these borrowers totals $665 million, providing
significant economic incentive to walk away, which would result in a
worst-case loss of about $500 million for TALF.
While this worst-case scenario provides useful information for Treasury to
consider as it monitors risk associated with TALF, we agree with the
Federal Reserve and Treasury that there are a number of factors that
affect whether such losses would be realized even in this adverse scenario.
First, as discussed in the ABS analysis, the accumulated excess interest in
TALF LLC would help offset potential losses. As of December 31, 2009, the
fund had almost $200 million in excess interest, which will continue to
increase every month in the absence of any asset purchases. Second, the
risk that all or a large portion of CMBS assets would be surrendered is
significantly mitigated by an FRBNY requirement that legacy CMBSs

43

We compared the prices on all of the CMBS CUSIPs, with the exception of two, that were
accepted by FRBNY as of September 30, 2009. For this analysis, we did not calculate the
excess interest that would have accumulated in TALF LLC.

Page 30

GAO-10-25 Troubled Asset Relief Program

prepay a portion of any returns in excess of certain limits.44 In short, the
requirement helps ensure that the TALF borrower will retain an equity
interest in the underlying CMBSs. The longer the term of the TALF loan
(assuming that the underlying collateral provides a return) the more equity
the borrower holds, and the less likely the borrower is to surrender the
CMBS collateral. Third, as with ABSs, and as indicated by Treasury
officials, TALF LLC could hold the securities acquired at a discount from
par—instead of selling them—and earn interest income and the equity
forfeited by the borrower as long as the underlying mortgages in the
security continue to perform well. Fourth, the estimated loss represents
less than 1 percent of total TALF loans as of December 2009, and total
TALF CMBS loans represent about 14 percent of all TALF loans. Because
CMBSs is a small portion of the portfolio, it would present a smaller
proportion of total losses. However, if recent TALF borrowing trends hold,
CMBS loans are likely to increase as the percentage of total TALF loans.
Finally, only senior credit-enhanced tranches within each CMBS can be
accepted as collateral. Thus, even if credit losses on the commercial
mortgages underlying the TALF CMBS securitizations are significantly
higher than currently expected in today’s stressed commercial real estate
environment, these senior tranches are unlikely to experience principal or
interest payment interruptions. While losses associated with CMBSs
currently appear unlikely, these securities warrant ongoing scrutiny
because of continuing economic uncertainty and the distressed conditions
in the commercial real estate market.

44
Specifically, the amount of money that FRBNY distributes to TALF borrowers that is in
excess of the interest due on the TALF loan will be limited for all 5-year loans and for 3year loans related to legacy CMBSs. In the first 3 years of a 5-year loan, the limit is 25
percent per year of the haircut amount, 10 percent in the fourth loan year, and 5 percent in
the fifth loan year. Any amount above this excess is applied to the principal on the TALF
loan. For 3-year loans for legacy CMBSs, the limit is 30 percent of the haircut per year. This
requirement ensures that borrowers retain some amount of equity interest in the securities
and reduces the likelihood of nonpayment. More details can be found on the FRBNY Web
site at http://www.newyorkfed.org/markets/talf_faq.html.

Page 31

GAO-10-25 Troubled Asset Relief Program

Treasury Worked with
the Federal Reserve
and FRBNY to
Analyze Risks Related
to TALF but Did Not
Fully Document
Analysis Supporting
Final Decisions

FRBNY, the Federal Reserve, and Treasury worked in a collaborative
manner to design certain elements of TALF, according to these agency
officials. The Federal Reserve led the initial efforts to determine collateral
eligibility and Treasury recommended one asset class and assessed the
risks of others. As part of this work, Treasury hired a contractor to
conduct independent analyses, and the contractor raised concerns about
accepting certain assets as TALF collateral, the size of the haircuts that
were required, and other program terms. While Treasury officials said that
the contractor’s concerns were ultimately resolved, they could not provide
documentation showing how Treasury resolved the contractor’s concerns,
or how the contractor’s analysis informed Treasury’s final decisions.
Treasury also did not document how they reached major decisions that
were made with FRBNY and the Federal Reserve. The lack of an effective
process to make and document decisions may inhibit transparency and
accountability of Treasury’s monitoring of the $20 billion of taxpayer
funds at risk through TARP.

The Federal Reserve and
Treasury Worked Together
to Determine the Eligibility
of Proposed TALF
Collateral and Their
Potential Risks

Treasury, the Federal Reserve, and FRBNY officials with whom we spoke
said that the agencies have a positive working relationship when making
decisions on TALF. For example, Federal Reserve and FRBNY officials
said that they consulted with Treasury to select the types of ABSs to
include in TALF as eligible collateral.45 Under this process, the Federal
Reserve identified all eligible collateral except SBA loan guarantees.
Treasury officials said that one asset class that FRBNY proposed—
insurance premium finance loans—required additional analysis to assess
the risks.46 These loans are not as widely traded or as well understood as
other asset classes. Treasury officials worked with Federal Reserve
officials to better understand the asset class, including its risks and factors
mitigating such risks, and its importance to small businesses. Treasury’s
contractor also reviewed risk information on this class of ABSs. Treasury

45

FRBNY also solicited input from market participants on asset classes that might benefit
from TALF. According to FRBNY officials, this is part of its regular market monitoring
function. Two of these market participants, which are also TALF agents, noted that
officials from the Federal Reserve requested their feedback about TALF before the
program was announced, seeking information such as how to restart securitization
markets, what the latest developments were in those markets, and which asset classes
should be considered.

46

Insurance premium finance loans are used to finance property and casualty insurance
premiums.

Page 32

GAO-10-25 Troubled Asset Relief Program

and Federal Reserve officials determined that this asset class should be
eligible for TALF. According to Treasury officials, other asset classes were
also reviewed by both agencies and not included in the program because
of their risks.
According to Treasury officials, the following criteria were used by
Treasury to evaluate the eligibility of asset classes for inclusion in TALF:
(1) whether including certain asset classes would have a significant or
beneficial effect on the broader economy, small businesses, or consumers;
and (2) whether assistance was needed because of a market failure in
what were otherwise safe asset classes.47 Based on these criteria, Treasury
recommended to the Federal Reserve and FRBNY that TALF accept as
collateral securities backed by SBA loan guarantees. SBA has two loan
guarantee programs—section 7(a) and section 504—that support financing
for small businesses.48 Treasury requested that these securities be included
because they would assist in carrying out TARP’s goals of supporting small
businesses and the risks were deemed to be low because of their
government guarantees.
As part of the asset class selection process, the Federal Reserve and
Treasury each analyzed the potential for loss that the TALF assets
presented. Treasury hired a contractor to, among other things, conduct an
independent analysis of the credit and other risks of the TALF asset
classes and to determine appropriate haircuts for each of the asset classes.
In its initial reports to Treasury, the contractor raised some concern about
accepting certain asset classes for TALF, provided suggestions for
program changes, and, in a few instances, disagreed with haircuts that
FRBNY suggested. According to Treasury officials and the contractor, any
differences were ultimately reconciled.

47

For more information on the approved asset classes and the dates they were announced
as acceptable collateral for TALF, see appendix IV and the background.

48

The 7(a) and 504 programs aim to facilitate the accessibility and affordability of financing
to small businesses. Under the 7(a) program, SBA generally provides lenders with
guarantees on up to 85 percent of the value of loans made to qualifying small businesses in
exchange for fees to help offset the costs of the program. Under the 504 program, which
generally applies to small business real estate and other fixed assets, SBA also provides
certified development companies with a guarantee on up to 40 percent of the financing of
the projects’ costs in exchange for fees, while the small business borrowers and other
lenders provide the remaining 60 percent of the financing with no guarantee. For additional
information, see GAO, Small Business Administration’s Implementation of
Administrative Provisions in the American Recovery and Reinvestment Act,
GAO-10-507R (Washington, DC: Apr. 16, 2009).

Page 33

GAO-10-25 Troubled Asset Relief Program

As an example, the contractor’s report noted that auto floor plan ABSs
faced risks because of the financial problems that the major domestic auto
manufacturers faced, and the nationally recognized statistical rating
organizations were in many cases unwilling to provide AAA ratings on
these securities. The contractor recommended a higher haircut to
encourage TALF borrowers to conduct thorough due diligence for this
asset class. Treasury questioned the contractor’s methodology and asked
the contractor to redo its analysis of auto dealer floor plan ABSs. After
working with Treasury and FRBNY to understand the differences in
methodologies between its analysis and FRBNY’s for this asset class, the
contractor agreed with FRBNY’s estimates on haircuts and even suggested
a haircut lower than FRBNY’s.

Treasury Did Not Fully
Document Its Analysis or
Basis to Support All Major
Agreements with the
Federal Reserve and
FRBNY

Although Treasury told us what its reasons were for not accepting all of
the contractor’s recommendations, Treasury officials were unable to
provide documentation showing when the contractor conducted the
analyses or when or how Treasury made decisions based on these
analyses. Further, no documentation was available showing how major
differences were resolved, including those involving program terms, the
eligibility of asset classes, and differences in haircut estimates among
Treasury, Treasury’s contractor, the Federal Reserve, and FRBNY.
Additionally, Treasury officials could not provide documentation on the
rationale for major program decisions that Treasury, the Federal Reserve,
and FRBNY officials reached. Treasury officials told us that FRBNY, the
Federal Reserve, and Treasury had a positive working relationship when
making decisions on TALF and described the process as “fluid;” therefore,
there was not always documentation of discussions and final outcomes.
Moreover, Treasury officials said that they spoke almost daily to Federal
Reserve and FRBNY officials. In some cases, according to Treasury
officials, FRBNY and the Federal Reserve consulted Treasury, although
technically no consultation was required. Treasury and the Federal
Reserve did not formally document their conversations, but the end result
of those conversations was documented and reflected publicly on the
TALF Web site administered by FRBNY.
Finally, some of the early decisions on TALF were made by Treasury
officials who are no longer at the agency. Without documentation, there
are no records to show, for instance, how certain suggestions made by
these officials about asset classes or program terms were incorporated
into policy choices for TALF.

Page 34

GAO-10-25 Troubled Asset Relief Program

Our Standards for Internal Control in the Federal Government states that
internal control activities help ensure that government management
directives are carried out.49 Such activities are critical to helping ensure
accountability and stewardship for government resources, and include
proper documentation of major decisions. In the context of the Emergency
Economic Stabilization Act of 2008—and the unprecedented size and
scope of government assistance to support the financial sector—
transparency and accountability are of the utmost importance. As an
example, for the largest program in TARP—the Capital Purchase
Program—all major decisions are recorded in meeting minutes that report
who was present, what decisions were made, and when they were made.50
In past TARP reports, we recommended that Treasury increase the
transparency and accountability of TARP, in part by documenting and
reporting certain processes and decisions.51
As we noted in past TARP reports, given the economic environment
surrounding the creation of TARP, and subsequently TALF, during the fall
of 2008, the change in administrations and the lack of staff that Treasury’s
administrative office for TARP—the Office of Financial Stability—faced,
Treasury may have initially had difficulty establishing its decision-making
processes for TALF and recording decisions and important meetings on
TALF program terms. At the time that TALF was created, the Office of
Financial Stability had been in existence for barely a month, and its
strategy and overall staffing needs were not yet in place. The broader
context at the onset of the program—with unprecedented economic
challenges and low, impermanent staffing—may help explain why such
processes were not established and documented when the program was
first established.
However, for TALF decision-making processes and the activities of TALF
LLC to be viewed as credible, Treasury needs to ensure that it has
developed an effective process to document the basis for its decisions. A
year has passed since Treasury began rolling out TARP-related programs,
and other larger programs—such as the Capital Purchase Program—have

49

GAO, Standards for Internal Control in the Federal Government, GAO/AIMD-00.21.3.1
(Washington, D.C.: November 1999).

50

See GAO-09-161.

51

See GAO, Troubled Asset Relief Program: Treasury Actions Needed to Make the Home
Affordable Modification Program More Transparent and Accountable, GAO-09-837
(Washington, D.C.: July 23, 2009) and GAO-10-16.

Page 35

GAO-10-25 Troubled Asset Relief Program

established systems for documenting decisions and the rationale for
decisions. But Treasury’s decisions for TALF still lack a clear process for
tracking how important program decisions are made and why. Without
such documentation, ascertaining what information has been considered
to protect TARP funds committed to TALF is difficult. Further, the lack of
documentation inhibits transparency and accountability.

Treasury’s and
FRBNY’s Indicators
Suggest That Credit
Market Conditions
Have Shown Some
Improvement, but
Treasury Lacks
Performance
Indicators in the
Event that It Must
Purchase TALF Assets

FRBNY, in conjunction with Treasury, monitors TALF by tracking
indicators—such as securitization volumes, changes in pricing, and TALF
loan volumes—by amount and borrower type to identify any possible
impact from TALF. Our analysis of these and other indicators suggests that
market conditions have begun to improve for some TALF-eligible asset
classes, but that others, such as CMBSs, continue to show weakness.
However, any assessment of the effectiveness of an individual program
presents challenges. As we have reported, no indicator can provide a
definitive measure of TALF’s impact because a myriad of programs have
been initiated to stabilize the markets, including actions taken under the
Capital Purchase Program and the Automotive Industry Financing
Program.52 Challenges remain for some of the TALF-eligible asset classes,
and FRBNY and Treasury monitor the performance of TALF loans and
collateral to be aware of all potential risks to TARP funds. However,
according to Treasury officials, Treasury has not yet developed a plan for
tracking assets that might be surrendered to TALF LLC or for publicly
disclosing how up to $20 billion in TARP funds would be monitored.

Treasury, FRBNY, and the
Federal Reserve
Collaborate on Monitoring
Market Indicators

Federal Reserve and Treasury officials said that they collaborated on
monitoring indicators that could help measure TALF’s effectiveness in
improving conditions in the securitization markets and, in turn, its impact
on the availability of credit to households and small businesses. Although
the officials have said they do not have specific benchmarks or targets
they hope to achieve for ABS issuance volumes or volumes of TALF loans,
they are monitoring those indicators. Officials also track interest spreads
for TALF-eligible asset classes, the number of borrowers accessing the
facility, and information about TALF borrowers, such as investor type.
FRBNY collects data on these indicators to monitor TALF’s impact.
According to Treasury officials, they review FRBNY’s metrics related to

52

See GAO-10-16 and GAO, Auto Industry: Summary of Government Efforts and
Automakers’ Restructuring to Date, GAO-09-553 (Washington, DC: Apr. 23, 2009).

Page 36

GAO-10-25 Troubled Asset Relief Program

TALF, including cash flows from TALF loans that it receives monthly from
FRBNY, and speak daily with FRBNY officials.

New Issuances in Various
Asset Classes Increased
after TALF’s First
Subscription in March 2009

To determine the condition of the securitization markets, we also have
been monitoring similar indicators, such as new ABS issuances and
changes in interest rates, types of investors, and spreads—using data from
before and after TALF’s implementation.53 In general, data from the
indicators that we have collected show increases in securitization
volumes, little change in the cost of credit, and declines in perceptions of
risk in certain asset classes since TALF began. ABS issuances in all of the
most liquid TALF-eligible sectors dropped sharply in 2008 from their peak
levels in 2006 and 2007. As figure 7 indicates, new issuance of ABSs had
come to a virtual halt in 2008, significantly reducing a major source of
credit for consumers and businesses.54 While securitization volumes
increased since the end of 2008, these increases have not been sustained
throughout 2009.

53

In this instance, spread refers to the difference between a security’s yield and a
benchmark yield. For ABSs, the customary benchmark yield is interest rate swap yields.
Interest rate swaps are contracts in which one party agrees to pay a fixed interest rate to
another party in exchange for a floating rate.

54

We focused on auto, credit card, and student loan ABSs and CMBSs because these asset
classes represent the largest sectors of the securitization markets and, in the case of
CMBSs, experienced the greatest levels of disruption that could be measured.

Page 37

GAO-10-25 Troubled Asset Relief Program

Figure 7: Fluctuations in ABS Issuances, First Quarter 2005 through Fourth Quarter 2009
Dollars in billions

Enlarged inset

25

80
20
70

15

60

10

50

5
0

40

Q4
’08

30

TALF
announcement

Q1
’09

Q2

Q3

Q4

Q3

Q4

TALF
subscription

20
10
0
Q1
2005

Q2

Q3

Q4

Q1
2006

Q2

Q3

Q4

Q1
2007

Q2

Q3

Q4

Q1
2008

Q2

Q3

Q4

Q1
2009

Q2

Quarter and year
Credit card
Student loan
Auto
CMBS
Source: GAO analysis of Thomson Reuters IFR Markets data.

After having shown little activity since the last quarter of 2008, issuance of
credit card, auto, and student loan ABSs increased after the initial TALF
subscription in March 2009 (shown as first quarter 2009 in figure 7). The
majority of ABS issuances in the credit card and auto sectors have been
supported by TALF loans. Specifically, of the $46 billion in ABSs issued on
credit card debt in 2009, $29.7 billion, or about 65 percent, have been
eligible for TALF financing. Similarly, about 88 percent—or $44.9 billion—
of the $51.2 billion in ABSs issued on auto loans in 2009 were TALFeligible deals. For more detailed information on securitization volumes,
see appendix VII.
By the third quarter of 2009, credit card and student loan issuances had
declined again in dollar terms, while auto issuances continued to increase.
A number of factors—such as the combined effects of the numerous
stimulus programs, changes in consumer demand for credit, and investor
willingness to invest—also may have contributed to the trend in
securitization volumes in these sectors. Federal Reserve officials

Page 38

GAO-10-25 Troubled Asset Relief Program

suggested that some companies may have been hesitant to issue credit
card ABSs because of uncertainty regarding the continued availability of
the FDIC’s “Safe Harbor” rule in light of new accounting rules effective for
annual financial periods beginning after November 15, 2009.55
Figure 7 also shows that CMBS securitization volumes peaked in mid-2007
and dropped significantly in late 2007 and early 2008 before coming to a
complete halt by the end of 2008. CMBS volumes did not pick up again
until the second quarter of 2009, but those were the result of repackaging
existing securitizations rather than new CMBS issuances. FRBNY officials
noted that the first new CMBS deal to come to the market since summer
2008 was a TALF-eligible, single-issuer deal in the fourth quarter. CMBS
issuances have remained relatively flat nonetheless, and it is unclear if
they will increase in the future. According to FRBNY and Treasury
officials, one possible reason for the small CMBS issuance volumes could
be that the time it takes to complete a CMBS deal is often longer than for
other asset classes.

55

The Federal Deposit Insurance Corporation (FDIC) Rule 360.6 provides that the FDIC will
not use its statutory authority as conservator or receiver to disaffirm or repudiate certain
contracts pertaining to any financial assets transferred by an insured depository institution
to a special purpose entity in connection with a securitization or participation, provided
that such transfer satisfies all conditions for sales accounting treatment under generally
accepted accounting principles (GAAP). For most insured depositary institutions, the 2009
GAAP modifications will be effective for reporting periods after January 1, 2010. However,
in November 2009, FDIC amended Rule 360.6 to extend the safe harbor to assets
transferred in connection with securitizations completed prior to March 31, 2010, so long as
those securitizations complied with the accounting rules as they existed prior to the 2009
GAAP modifications. See, “Defining Safe Harbor Protection for Treatment by the Federal
Deposit Insurance Corporation as Conservator or Receiver of Financial Assets Transferred
by an Insured Depository Institution in Connection with a Securitization or Participation,”
74 Fed. Reg. 59066 (Nov. 17, 2009) (interim rule and request for comments).

Page 39

GAO-10-25 Troubled Asset Relief Program

Interest Rates in Most
Asset Classes Have
Generally Not Decreased
Since TALF Was
Announced

One of TALF’s goals is to increase the availability of credit to consumers
and businesses. TALF assistance to the securitization markets is intended
to result in lower loan rates and increased credit availability to businesses
and individual consumers, including the auto loan, credit card, and student
loan sectors that account for the majority of securitizations. 56 Recent
increased activity in the securitization markets has been accompanied by a
substantial decrease in interest rates for loans originated by auto finance
companies. However, there have been few changes in credit card rates or
interest rates for consumers in auto loans originated by commercial banks.
Consumer interest rates remain flat.
Because auto finance companies rely more heavily on securitizations for
funding than commercial banks, the effects of positive changes in the
securitization markets are more likely to be reflected in their loan rates
than in those of commercial banks. As figure 8 shows, auto loan rates
offered by commercial banks remained fairly steady before and after the
implementation of TALF. The average finance company auto rate has been
consistently below commercial bank auto rates, with the exception of the
fourth quarter of 2008, perhaps reflecting the financial challenges facing
the auto industry at that time.57 Since then, rates at auto finance companies
have declined from an average of 7 percent to approximately 3 percent.
This reduction coincides with the launching of TALF but may also reflect
assistance from the numerous government stimulus programs, especially
those focused on the auto industry. While fixed credit card rates have
remained fairly flat in recent years, variable credit card rates have
increased by approximately one percent since TALF’s inception. FRBNY
officials attributed the elevated credit card rates to increased charge-offs,
which have raised companies’ costs of funds. These rate changes could
also be the result of credit card companies’ efforts to anticipate the
implementation of the remaining part of the Credit Card Accountability

56

Auto finance companies such as Ford Motor Credit Company, GMAC, and Chrysler
Financial provide financing for consumer automotive and dealer purchases. We focused on
interest rates for credit cards and automotive loans and did not evaluate changes in interest
rates for student loans because reliable and consistent data were unavailable for private
student loans. We also did not evaluate changes in interest rates for commercial mortgages,
because these rates depend on factors such as geographic location and property type,
making it difficult to draw broad conclusions about interest rate trends.
57

See GAO-09-553.

Page 40

GAO-10-25 Troubled Asset Relief Program

Responsibility and Disclosure Act of 2009, which will take effect in
February 2010.58
FRBNY officials said that it is possible that without TALF interest rates on
loans to consumers and small businesses would be much higher than they
are now. Issuers of TALF-eligible ABSs have told FRBNY that without
TALF they would have made fewer loans and those loans would have been
at higher rates. Data on private student loan rates are difficult to obtain,
but FRBNY officials said that Sallie Mae has reduced its rates on private
student loans over the past few months.
Figure 8: Trends in Average Finance Rates for Credit Cards, Auto Loans at Banks, and Auto Finance Companies, June 2008
through December 2009
Percentage

TALF
announcement

13

TALF
subscription

Variable card rates
12
11

Fixed card rates

10
9
8
7

Used car rates
New car rates

6
5

Auto finance company rates

4
3
2
6
2008

7

8

9

10

11

12

1
2009

2

3

4

5

6

7

8

9

10

11

12

Month and year
Source: Bankrate.com and the Federal Reserve.

Note: Auto finance company rates from the Federal Reserve are only available through October 31,
2009.

58

Provisions in the law will limit credit card companies’ ability to increase interest rates,
among other things, so some companies have raised interest rates in anticipation of the law
taking effect. See GAO, Credit Cards: Rising Interchange Fees Have Increased Costs for
Merchants, But Options for Reducing Fees Pose Challenges, GAO-10-45 (Washington, DC:
Nov. 19, 2009).

Page 41

GAO-10-25 Troubled Asset Relief Program

TALF Loan Volumes and
the Composition of TALF
Participants Have Changed
since the Program Began

As illustrated in table 3, the volume of TALF loans made since the
inception of the program has fluctuated by month, with loan volumes
peaking in May 2009 and June 2009. As of December 31, 2009, a total of
$61.6 billion in TALF loans had been granted; however, the balance of
loans outstanding at that date was $47.5 billion due to loan prepayments
and principal paydowns. As we reported previously, agency officials
indicated that improvements in securitization and loan markets had made
issuers less dependent on TALF support.59 However, according to FRBNY
officials, other issuers remain more heavily dependent on investor access
to TALF financing. FRBNY officials noted that there have been a number
of prepayments, and market participants also told us that financing under
TALF was now less favorable because better financing terms could be
found in the private sector for certain asset classes. Notably, the first and
only subscription for new-issue CMBSs occurred in November 2009.
FRBNY and Treasury officials stated that the slow new-issue CMBS
activity may be due to the length of time it takes to complete a deal.

Table 3: TALF Loans Disbursed by Asset Class, March 2009 through December 2009
Dollars in millions
Mar.

Apr.

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

Total by
loan type

$1,909

$797

$2,311

$2,946

$2,831

$555

$1,160

$191

$0

$0

$12,699

Credit card

2,804

891

5,515

6,023

1,459

2,554

4,399

224

63

1,529

25,461

Equipment

0

0

446

590

0

0

111

39

57

199

1,441

Floor plan

0

0

0

0

0

1,005

0

887

445

172

2,510

Insurance premium
finance

–

–

–

464

0

0

530

0

0

0

994

Servicing
advances

0

0

0

439

34

108

0

475

0

138

1,193

Student loan

0

0

2,281

227

987

2,445

177

288

85

665

7,155

Small business

0

0

87

29

62

147

162

262

409

275

1,433

Legacy CMBSs

–

–

–

0

636

2148

1,351

1,931

1,330

1,282

8,677

Asset class
Auto

New-issue CMBSs
Total

–

–

–

0

0

0

0

0

72

0

72

$4,713

$1,688

$10,639

$10,717

$6,009

$8,962

$7,890

$4,297

$2,461

$4,259

$61,636

Source: GAO analysis of FRBNY data.

59

See GAO-10-16.

Page 42

GAO-10-25 Troubled Asset Relief Program

Note: Not all numbers will total due to rounding. Until September 2009 the FRBNY’s Web site showed
only TALF loans requested, not loans disbursed. FRBNY provided us with data prior to September
2009. In addition, in discussions with FRBNY we learned that the program changed in April 2009 to
allow for new interest rates on loans secured by ABSs with weighted average lives to maturity of less
than 2 years had resulted in the refinancing of some existing TALF loans. As a result, May 2009
figures may include double counting due to this refinancing.

As shown in figure 9, approximately 75 percent of TALF loans involved the
purchase of ABSs backed by auto loans and leases, credit card receivables,
and student loans. This activity reflects the historical trends in auto, credit
card, and student loan securitizations, which represent the majority of the
ABS markets.
Figure 9: Composition of TALF Loans Disbursed by Asset Class as of December
2009
(Dollars in millions)

2% Insurance premium finance
$994
2% Mortgage servicing advances $1,193
2% Small Business Administration $1,433
2% Equipment
$1,441
4%
12%

Floor plan

$2,510

Student loans

$7,155

41%

14%
21%

CMBS (legacy 14%; new issue 0%) $8,677
$72
Auto

$12,699

Credit cards

$25,461

Source: GAO analysis of FRBNY data.

According to Treasury and Federal Reserve officials, TALF was designed
to encourage broader investor participation in the securitization markets,
with the goal of reviving consumer lending. These officials noted that the
securitization markets stopped functioning in 2008 when many investors
stopped purchasing these securities. The lack of securitization market
activity disrupted a significant source of funding for businesses and
consumers. Gradually, some of these investors have returned to the
markets, but at a slower rate than during past market downturns.
Specifically, Treasury officials noted increasing participation in TALF

Page 43

GAO-10-25 Troubled Asset Relief Program

securitization by asset managers, hedge funds, and traditional institutional
investors such as pension funds and insurance companies. They consider
the return of investors to the securitization markets to be a measure of the
program’s success. Hedge funds traditionally have not invested in ABSs
because of the low returns relative to other opportunities. However,
FRBNY officials believe the access to low-cost financing through TALF
made ABS returns attractive to hedge funds. FRBNY also noted
participation by private investors and banks.

Differences in Prices and
Benchmarks Have
Decreased for Most TALFEligible ABS Collateral

As we have discussed, one method of measuring market participants’
perceived risk of a security is to compare the difference between the
security’s yield and a benchmark yield. The difference is called a spread,
and wide spreads, or large differences, generally indicate that participants
perceive high risk in the market that requires a high rate of return. As
perceived risk declines, differences in such prices decrease, or narrow.
During the fourth quarter of 2008 and first quarter of 2009, spreads likely
reflected high expected costs of selling securities prior to their maturity,
which contributed to low desirability for those securities. Figure 10 shows
the change in spreads in the following TALF-eligible asset classes: auto
loan, credit card, student loan, and CMBS. A trend of widening spreads in
these asset classes began in mid-2007, indicating negative perceptions
about risk. Although there were fluctuations throughout 2008, spreads
began to narrow in early 2009, indicating a perceived decline in risk by
market participants and potentially improved credit market conditions.

Page 44

GAO-10-25 Troubled Asset Relief Program

Figure 10: Trends for AAA ABS and CMBS Spreads from January 2005 through September 2009
Percentage
TALF
announcement

1,500

TALF
subscription

1,200

900

600

300

0

-300
1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9
2005
2006
2007
2008
2009
Month and year
CMBS
Credit cards
FFELP loan
Auto
Private student loans
Source: GAO analysis of spread data from multiple banks.

Treasury Reviews TALFRelated Data from
FRBNY’s Indicators but
Has Not Developed
Indicators to Collect and
Disclose Data on Future
TALF LLC Assets

Treasury reviews the data that FRBNY collects on TALF loan volumes and
borrowers by type, securitization volumes, and changes in pricing.
Treasury officials noted that personnel at both agencies were responsible
for a variety of tasks in tracking TALF-related metrics. We found that
Federal Reserve officials, particularly at FRBNY, typically took the lead in
collecting data and calculating metrics. We also found that Treasury
officials did not have a plan to collect and analyze information related to
assets that might be placed in TALF LLC—assets to which Treasury would
have an exposure. Such information might include the purchase and sale
price of the assets, their current market value, total outstanding loans by
Treasury to TALF LLC for the ABS purchases, and the rationale behind
TALF LLC’s possible future sale of assets. Treasury has not yet developed
such a plan because no TALF collateral has been surrendered thus far, and
Treasury believes it is unlikely that it will have to use TARP funds to
finance TALF LLC’s purchase of surrendered collateral. Moreover,
Treasury does not have a plan to publicly communicate such information

Page 45

GAO-10-25 Troubled Asset Relief Program

in the event that collateral is surrendered and placed there. In previous
TARP reports and in this report, we have discussed the importance of
improving the transparency and accountability of TARP programs. We
have also recommended that Treasury build on existing oversight
procedures to better monitor and report on the use of TARP funds and to
better quantify program results.
Although Treasury is not responsible for implementing or administering
TALF, it has pledged support to TALF LLC with the first $20 billion of
potential loans to allow it to purchase surrendered TALF collateral. As
discussed earlier, commercial real estate continues to show weakness and
could potentially pose greater risks to TARP funds. Without a system for
tracking and reporting on any potential assets such as CMBSs that are
surrendered to TALF LLC, Treasury cannot assure transparent
management of these assets or determine if it is achieving its goals under
CBLI with respect to the use of TARP funds for TALF-related activities.
Further, without properly planning for its role in managing the collateral
should they have to be purchased by TALF LLC, Treasury may not be able
to effectively assess any risks associated with such assets or exercise
appropriately its decision-making responsibilities regarding the potential
sale of any assets.

Conclusions

TALF is one of several programs created by the Federal Reserve to help
address the recent crisis in the financial sector. Specifically, this program
was designed to restart securitization markets, a critical part of financial
markets. Given the myriad of programs initiated to stabilize the financial
system and increase credit availability, it is difficult to attribute
improvement in markets to any one program. Nevertheless, according to a
variety of indicators, TALF appears to be contributing to measured
improvements in the securitization markets. As of December 31, 2009,
$61.6 billion in loans were made through TALF and TALF LLC had
received $100 million of the $20 billion in TARP funds committed to the
program. In addition to the $20 billion, funds provided by FRBNY to
operate TALF could expose additional risks. However, because we are
statutorily prohibited from auditing the Federal Reserve’s monetary policy
activities, we believe our ability to completely assess and report on
taxpayers’ exposure to the entire program or the Federal Reserve’s
management of the program is limited.
Although the government has taken a number of steps to mitigate the risk
of loss from TALF, in the long term risks remain. For example, while
analyses by the Federal Reserve and a Treasury contractor that were

Page 46

GAO-10-25 Troubled Asset Relief Program

based on predictions of market performance and other factors estimated
that a loss of a substantial portion of the $20 billion TARP commitment
would be unlikely based on current conditions in the securitization
markets, we found that until the TALF borrowers repay their loans, TALF
still presents risks. While we acknowledge that overall market conditions
have generally improved since 2008, some asset classes—specifically
CMBS—are still performing poorly and may continue to perform badly for
the foreseeable future. Moreover, markets remain fragile and predicting
how the overall ABS markets will perform in the future and how
borrowers might respond to new declines in the markets is difficult. A
return to 2008 conditions could have adverse impacts on the program,
such as significantly reducing the value of TALF collateral, providing an
economic incentive for borrowers to walk away from their loans, and
requiring TARP funds be used to buy TALF collateral. However, several
TALF program features make this less likely.
Treasury, which worked with FRBNY and the Federal Reserve on certain
decisions related to TALF, was not able to provide documentation on how
these decisions were made. As we noted in past TARP reports, Treasury
has yet to develop systems to ensure the transparency and accountability
for TARP activities by implementing a strong, transparent strategic
framework with appropriate oversight mechanisms. Among other things,
these mechanisms would ensure accountability by tracking why decisions
are made, and whether goals are being achieved. Documenting the basis
for decisions is an important part of the decision-making process.
Moreover, documenting the rationale for major program decisions would
help ensure that the program objectives are being met and that it is
functioning as intended. Unless Treasury documents the rationale for
major program decisions that it made with the Federal Reserve, it cannot
demonstrate accountability for meeting the goals of TALF and could
unnecessarily place TARP funds at risk.
Believing it is highly unlikely that it will have to use TARP funds to finance
ABS purchases by TALF LLC, Treasury has not taken steps to develop a
set of metrics or a plan for tracking and reporting on the performance of
the collateral that could be placed in TALF LLC. While TARP funds may
never be used to finance purchases of ABS or CMBS used as TALF
collateral, Treasury should at least be prepared for the possibility. Without
a plan for collecting and analyzing such data, Treasury would have to
develop one as it is financing or after it has financed collateral purchases
by TALF LLC and risks being ill prepared to make informed decisions on
whether TALF LLC should keep collateral until the securities mature or
sell them. Unlike many other programs that were developed and

Page 47

GAO-10-25 Troubled Asset Relief Program

implemented in the midst of the crisis, Treasury has an opportunity to be
strategic by developing a plan in the event that its role in TALF is
triggered. In addition, without a plan Treasury cannot measure TALF’s
success in meeting its goals under CBLI with respect to any assets that are
placed in TALF LLC. Finally, without a plan for communicating the
findings that result from tracking and analyzing such metrics, the public
will not be aware of how the assets are managed and financed,
undermining Treasury’s efforts to be fully transparent about TARP
activities.

Matter for
Congressional
Consideration

To enable GAO to audit TARP support for TALF most effectively, we
recommend that Congress provide GAO with audit authority over all
Federal Reserve operational and administrative actions taken with respect
to TALF, together with appropriate access authority.

Recommendations for
Executive Action

To improve transparency of decision making on the use of TARP funds for
TALF and to ensure adequate monitoring of risks related to TALF
collateral, we recommend that the Secretary of the Treasury direct the
Office of Financial Stability to take the following actions:
1. Given the distressed conditions in the commercial real estate market,
as part of its ongoing monitoring of TALF collateral, continue to give
greater attention to reviewing risks posed by CMBSs.
2. Strengthen the process for making major program decisions for TALF
and document how it arrives at final decisions with the Federal
Reserve and FRBNY. Such decisions should include how Treasury
considers expert and contractor recommendations and resolves those
recommendations that differ from those of the Federal Reserve and
FRBNY.
3. Conduct a review of what data to track and metrics to disclose to the
public in the event that TALF LLC purchases surrendered assets from
FRBNY. Such data and metrics should relate to the purchase,
management, and sale of assets in TALF LLC that potentially impact
TARP funds. Metrics related to TALF LLC could include periodic
reports on the date and purchase price of assets; fluctuations in the
market value of assets held; the date, price, and rationale when assets
are sold; and the total amount of loans outstanding to Treasury.

Page 48

GAO-10-25 Troubled Asset Relief Program

Agency Comments
and Our Evaluation

We provided a draft of this report to Treasury for its review and comment.
We also provided the draft report to the Federal Reserve to verify the
factual information they provided to us about TALF. Treasury and the
Federal Reserve provided written comments that we have reprinted in
appendixes VIII and IX, respectively. Treasury and the Federal Reserve
also provided technical comments that we have incorporated as appropriate.
In their response Treasury welcomed our recognition that TALF
contributed to improvements in the securitization markets but believed
that the draft report understated the success of the program. In so doing
Treasury reiterated several points that were already underscored in the
draft report. For example, as discussed in the draft report and Treasury’s
response, we acknowledged that securitization volumes in markets had
come to a complete halt in 2008, but increased after TALF’s first
subscription in March 2009. Moreover, we also noted that recent TALF
subscription levels for the majority of eligible asset classes have tapered
off, which is an indication that investors’ perception of risk has decreased.
As we have noted in our previous TARP reports, any assessment of the
effectiveness of TALF is complicated by the fact that a variety of programs
have been established by the Federal Reserve, Treasury, and others to
stabilize the markets—making it virtually impossible to definitively single
out and measure TALF’s impact.
Treasury also stated that it disagreed with our methodology related to
potential losses for CMBSs. As we discussed in the report, the adverse
scenario analysis of the TALF CMBS portfolio was not intended to project
an expected loss amount for this portfolio but to help assess the possible
range of losses in TALF. We used a stress scenario and selected loss
assumptions that were similar to those the Federal Reserve imposed on
the 19 bank holding companies that participated in the 2009 stress test.
Treasury states that it would take a 65 percent loss on underlying
commercial real estate prices to experience losses. While we agree that
this is an unlikely event, commercial real estate prices have already fallen
by an average of 43 percent since prices peaked. Combined with the fact
that CMBSs have much longer time horizons than other TALF ABS asset
classes and hence greater uncertainty of outcomes, we continue to believe
that CMBS warrants ongoing attention.
Treasury also noted that it appreciates the recommendations GAO makes
in the report to strengthen the documentation of decisions Treasury made
concerning changes to the program. Treasury stated it is committed to
ensuring that not only TALF, but TARP as a whole, is administered in a
way that protects the taxpayer. We believe that development of a sound

Page 49

GAO-10-25 Troubled Asset Relief Program

decision-making process that includes steps for formal approval and
documentation of the basis of the final decisions at an appropriate
management level will improve transparency and accountability of the
TALF program. As we noted in past TARP reports and most recently in the
October 2009 report, Treasury has yet to develop systems to ensure the
transparency and accountability for TARP activities by implementing a
strong, transparent strategic framework with the appropriate oversight
mechanisms. Among other things, these mechanisms would ensure
accountability by tracking why decisions are made and whether goals are
being achieved.
Finally, regarding our recommendation that Treasury review what data to
track and metrics to disclose to the public in the event that TALF LLC
purchases surrendered assets from FRBNY, Treasury noted that it will
continue to enhance its existing reporting on its investments in TALF that
strikes an appropriate balance between its goal of transparency and the
need to avoid compromising either the competitive positions of investors
or Treasury’s ability to recover funds for taxpayers. We believe that having
a plan in place for tracking and reporting on the performance of any
collateral that could be placed in TALF LLC will help Treasury strike that
balance.
In its comments, the Federal Reserve did not agree with our
recommendation that Congress consider providing GAO with authority to
audit the Federal Reserve’s TALF operational and administrative actions
because it disagreed that there are limitations on GAO’s authority to audit
these Federal Reserve activities. The Federal Reserve also noted that it
fully cooperated in GAO’s conduct of this audit and provided us access to
records and personnel.
The Federal Reserve did cooperate and voluntarily provided all access we
requested in this audit of Treasury. We appreciate this cooperation, which
enabled us to factually describe the TALF program and to evaluate
Treasury’s involvement in it. However, we believe the express statutory
prohibition in 31 U.S.C. § 714(b) on GAO auditing the Federal Reserve’s
monetary policy and discount window activities, which the Federal
Reserve believes include TALF’s operation and administration, prohibits
us from auditing the Federal Reserve’s TALF activities, even from the
perspective of TARP. We limited the scope and conduct of this audit
accordingly, and thus did not request access to information to audit the

Page 50

GAO-10-25 Troubled Asset Relief Program

Federal Reserve’s performance of these activities.60 Further, the Federal
Reserve’s decision to voluntarily provide requested access in this instance,
while helpful, does not create GAO authority for access to information the
agency may not volunteer, nor GAO authority to audit the Federal
Reserve’s TALF operational activities or other performance. In our view,
our lack of authority to audit the Federal Reserve’s actions limited our
ability to fully assess the risk to taxpayer funds presented by TALF.
Accordingly, we continue to believe that Congress should provide GAO
with authority to audit the Federal Reserve’s operation and administration
of the TALF program. Our detailed response to the Federal Reserve’s
comments on these issues is contained in appendix X.

We are sending copies of this report to the Congressional Oversight Panel,
Financial Stability Oversight Board, Special Inspector General for TARP,
interested congressional committees and members, Treasury, the federal
banking regulators, and others. The report also is available at no charge on
the GAO Web site at http://www.gao.gov.

60

A performance audit of the Federal Reserve’s TALF operational and administrative
activities would, for example, have involved evaluating the sufficiency of how certain TALF
program terms were arrived at, such as whether the haircuts or the amount of equity the
TALF borrower holds in the collateral protected TALF from losses. We also would have
evaluated FRBNY’s system of internal controls and the role of TALF participants in
certifying and validating compliance with certain program requirements. In addition, we
would have interviewed some of the entities that helped administer TALF to validate
agency information and to better understand their roles and interactions with the Federal
Reserve and FRBNY. These would have included entities such as the TALF agents, Bank of
New York Mellon, Trepp, or CW Capital.

Page 51

GAO-10-25 Troubled Asset Relief Program

If you or your staffs have any questions about this report, please contact
Orice Williams Brown at williamso@gao.gov or (202) 512-8678. Contact
points for our Offices of Congressional Relations and Public Affairs may
be found on the last page of this report. GAO staff who made major
contributions to this report are listed in appendix XI.

Gene L. Dodaro
Acting Comptroller General
of the United States

Page 52

GAO-10-25 Troubled Asset Relief Program

List of Committees
The Honorable Daniel K. Inouye
Chairman
The Honorable Thad Cochran
Vice Chairman
Committee on Appropriations
United States Senate
The Honorable Christopher J. Dodd
Chairman
The Honorable Richard C. Shelby
Ranking Member
Committee on Banking, Housing,
and Urban Affairs
United States Senate
The Honorable Kent Conrad
Chairman
The Honorable Judd Gregg
Ranking Member
Committee on the Budget
United States Senate
The Honorable Max Baucus
Chairman
The Honorable Charles E. Grassley
Ranking Member
Committee on Finance
United States Senate
The Honorable David R. Obey
Chairman
The Honorable Jerry Lewis
Ranking Member
Committee on Appropriations
House of Representatives

Page 53

GAO-10-25 Troubled Asset Relief Program

The Honorable John M. Spratt, Jr.
Chairman
The Honorable Paul Ryan
Ranking Member
Committee on the Budget
House of Representatives
The Honorable Barney Frank
Chairman
The Honorable Spencer Bachus
Ranking Member
Committee on Financial Services
House of Representatives
The Honorable Charles B. Rangel
Chairman
The Honorable Dave Camp
Ranking Member
Committee on Ways and Means
House of Representatives

Page 54

GAO-10-25 Troubled Asset Relief Program

Appendix I: Objectives, Scope, and
Methodology

Appendix I: Objectives, Scope, and
Methodology
The objectives of this report are to (1) analyze the risks that the Term
Asset-Backed Securities Loan Facility (TALF) presents to Troubled Asset
Relief Program (TARP) funds and therefore to taxpayers, (2) evaluate how
the Department of the Treasury (Treasury) analyzed the risk of TALF
assets and used this information in making decisions on TALF with the
Board of Governors of the Federal Reserve System (Federal Reserve) and
the Federal Reserve Bank of New York (FRBNY), and (3) assess the
condition of securitization and credit markets before and after TALF’s
implementation based on indicators tracked by Treasury and FRBNY.
GAO has statutory limitations on auditing certain functions of the Federal
Reserve.1 Because of these limitations, the evaluative content of this report
is limited to Treasury’s role of safeguarding TARP funds related to TALF
and we did not review or evaluate any monetary policy actions taken by
the Federal Reserve or FRBNY with respect to TALF. We collected
information on Federal Reserve practices related to TALF, but did not
audit those practices. Specifically, we did not evaluate the sufficiency of
how certain TALF program terms, such as haircuts and interest rates, were
arrived at.2 In addition, we did not assess FRBNY’s system of internal
control or the role of TALF participants such as agents, borrowers, and
auditors in certifying and validating compliance with certain TALF terms.
Finally, we did not validate the comments or background information
provided to us by Federal Reserve and FRBNY officials about TALF.

1

Section 714 of Title 31 of the U.S. Code limits GAO’s authority to audit certain Federal
Reserve activities. Specifically, GAO audits of the Federal Reserve and Federal Reserve
banks “may not include deliberations, decisions, or actions on monetary policy matters,
including discount window operations, reserves of member banks, securities credit,
interest on deposits and open market operations. . . , or transactions made under the
direction of the Federal Open Market Committee” 31 U.S.C. § 714 (b)(2)-(3). This
prohibition limits GAO’s ability to audit the Federal Reserve’s actions taken with respect to
TALF. The Helping Families Save Their Homes Act of 2009, enacted on May 20, 2009,
amended Section 714 to provide GAO authority to audit Federal Reserve Board actions
taken under Section 13(3) of the Federal Reserve Act with respect to a single and specific
partnership or corporation. See Pub. L. No. 111-22, 123 Stat. 1632, 1662-63. Among other
things, this amendment provides GAO with authority to audit Federal Reserve actions
taken with respect to three entities also assisted under TARP—Citigroup, Inc., American
International Group, Inc., and Bank of America Corporation—but does not provide GAO
with authority to audit Federal Reserve monetary policy actions taken with respect to
TALF generally.

2
Haircuts set the amount of equity the TALF borrower holds in the collateral and is a
percentage assigned by the FRBNY. Haircuts vary by FRBNY’s assessment of market risks
for each sector and subsector.

Page 55

GAO-10-25 Troubled Asset Relief Program

Appendix I: Objectives, Scope, and
Methodology

To address the first objective, we first reviewed publicly available
documentation on the Web sites of the Federal Reserve and FRBNY. We
also interviewed Treasury, FRBNY, and Federal Reserve officials to
understand how TALF fits in to Treasury’s Financial Stability Plan and
how risks to the taxpayer were reduced in TALF’s design. Next, we
assessed how Treasury reviewed the risks of the various asset classes
considered for TALF eligibility by collecting and analyzing reports that
Treasury requested through a contractor, Bank of New York Mellon, which
in turn subcontracted the work to NSM Structured Credit Solutions, which
has since been acquired and is now known as RangeMark. We also
reviewed the subcontractor’s methodology for assessing the likelihood of
loss to TARP funds and interviewed the subcontractor, contractor, and
Treasury officials about the assumptions in the loss model. We also
reviewed other factors that have an impact on the risk to TARP funds and
taxpayers, including the return on equity for TALF borrowers, credit
enhancement of TALF securities, and the risks of asset-backed securities
(ABS) and commercial mortgage-backed securities (CMBS).
•

To assess the changes in return on equity (ROE), we analyzed the returns
based on information collected from prospectuses for TALF-eligible ABSs
on credit cards, auto loans, auto leases, and private student loans issued
between March and September 2009. Some of these prospectuses were
provided by the Federal Reserve. We also used information collected from
reports from Moody’s Investors Service and Standard & Poor’s. We
calculated returns for fixed-rate bonds by using the tranche-level3 interest
rate paid to the FRBNY. For floating-rate bonds, we used a spread
between the interest rate paid to the FRBNY and an index, such as the
London Interbank Offered Rate. This is the “coupon” variable in the
equation below:
ROE = Coupon – (1 – Haircut%)* Rate _ on _ loan _ paid _ to _ FRBNY
Haircut%

•

To assess the levels of credit enhancement for TALF securities, we
analyzed information collected from prospectuses related to public and
private offerings of TALF-eligible securities issued between March and
September 2009, along with related reports from Standard & Poor’s and

3
Some securitizations—such as ABSs backed by auto loans—are divided into different
segments, or tranches. A tranche is a piece of a securitization that has specified risks and
returns.

Page 56

GAO-10-25 Troubled Asset Relief Program

Appendix I: Objectives, Scope, and
Methodology

Moody’s Investors Service. For each security, we compared the level of
credit enhancement for the TALF issuance with that issuer’s most recent
securitization prior to TALF, which ranged from 2004 through 2008.
•

To understand the recent activity in CMBS markets, we collected
information from Moody’s Investors Service on commercial real estate
prices (Moody’s/REAL Commercial Property Price Index) and on CMBS
delinquency (Moody’s CMBS Delinquency Tracker). We determined that
the data was reliable for our purposes of demonstrating recent trends in
the commercial mortgage sector. In addition, we collected CMBS price
performance data from Thomson Reuters DataScope and determined that
the information on the price, yield, and performance of securities was
reliable for our purposes of understanding trends in CMBS prices and
vintages for the TALF portfolio.

•

We interviewed a range of market participants and market observers about
the taxpayer protections and other features of TALF, to include three
dealers that also serve as TALF agents; three issuers (one for credit cards,
one for auto loans, and one for student loans) and an SBA securities
dealer; three industry associations representing the CMBS market, the
hedge fund industry, and small and regional banks; a buy-side investment
firm with interest in TALF; two large auditing firms that provide auditor
attestations for TALF; an attorney with securitization market expertise;
two TALF-qualified credit rating agencies, or nationally recognized
statistical rating organizations; an academic in banking and securitization
at the Massachusetts Institute of Technology; an analyst from Brookings
Institution; an analyst from a student loan firm; and a representative of a
consumer advocacy organization.

•

For details on our methodology for assessing adverse scenario losses from
TALF ABSs and CMBSs, see appendix II.
To address the second objective on Treasury’s analysis of the risk
associated with TALF assets and how that analysis was used to make
decisions with the Federal Reserve and FRBNY related to TALF, we
analyzed reports from a subcontractor with Treasury—NSM Structured
Credit Solutions—that provided assessments of various risks of TALF to
TARP funds. In analyzing these reports, we reviewed the asset class risk
assessments, the recommendations made to change TALF program terms,
and the suggested haircuts for each asset class. We also interviewed
Treasury’s contractor and the subcontractor for clarification on the
reports and to understand Treasury’s interaction with both. In addition, we
interviewed Treasury officials about their role in reviewing and shaping
the terms of TALF, how they considered the analysis and

Page 57

GAO-10-25 Troubled Asset Relief Program

Appendix I: Objectives, Scope, and
Methodology

recommendations of the subcontractor, how they decided to include
certain asset classes, and how they came to agree on haircuts and other
program terms with the FRBNY and Federal Reserve. We also interviewed
officials from FRBNY and the Federal Reserve about the reasons for
differences in haircuts and other TALF program terms and how they were
resolved with Treasury and the subcontractor.
To address the third objective on changes in the securitization and credit
markets before and after TALF was created, and to understand how
Treasury tracks the impact of TALF and its potential risks to TARP funds,
we collected and analyzed information from a variety of data sources
relevant to the ABS, CMBS, and credit markets. Specifically:
•

To review changes in securitization markets for ABSs backed by auto
loans, credit cards, student loans, and commercial mortgages, we
collected data from Thomson Reuters IFR Markets, a database that
collects information on activity in the securitization markets. To analyze
changes in interest rates for auto loans and credit cards, we reviewed
quarterly data from the Federal Reserve’s G.19 Consumer Credit Release, a
widely used data source, as well as weekly data provided by
BankRate.com. We selected the auto, credit card, student loan, and CMBS
asset classes because they were the most widely traded in securitization
markets and the latter had recently experienced significant trading and
price volatility. Because reliable interest rate data for private student loans
and commercial mortgages were more difficult to obtain, we collected and
analyzed data only on auto loans and credit cards. We validated the
securitization and interest rate information against reports and data
provided by credit rating agencies, issuers, and dealers. We determined
that the data sources were sufficient for our purposes of demonstrating
trends in the markets before and after TALF was created.

•

To report on the amount of TALF loans settled, we accessed data publicly
available on the FRBNY Web site and also information provided to us from
FRBNY for periods when FRBNY did not publicly report the settled loan
amounts, but only the requested amounts. Because of the limitations on
our audit authority, we did not review the internal systems that generated
this information.

•

To analyze spreads for ABSs backed by auto loans, credit cards, student
loans, and CMBSs, we analyzed dealer-provided data from three dealer
banks. Because this spread information is not available from one data
provider, we determined that collecting data from three dealers—and

Page 58

GAO-10-25 Troubled Asset Relief Program

Appendix I: Objectives, Scope, and
Methodology

ensuring that the numbers were within an acceptable range of each
other—would ensure the reliability of such data for our purposes.
•

To determine what information Treasury collects to assess TALF’s impact
on securitization and credit markets and the risks TALF poses to TARP
funds, we interviewed officials from the Treasury about what data they
collect and received reports that Treasury’s subcontractor provided on the
various risks of TALF activities. We also interviewed Federal Reserve and
FRBNY officials about what type of data they collect related to TALF’s
impact on the securitization and credit markets.

Page 59

GAO-10-25 Troubled Asset Relief Program

Appendix II: Methodology for Market Value
Analysis of ABSs and CMBSs

Appendix II: Methodology for Market Value
Analysis of ABSs and CMBSs
To understand the possible range of losses to Troubled Asset Relief
Program (TARP) funds from the Term Asset-Backed Securities Loan
Facility (TALF), we conducted an analysis based on extreme market value
losses, similar to those experienced in the asset-backed securities (ABS)
and commercial mortgage-backed securities (CMBS) markets in
November 2008. This provides an alternative approach to the Department
of the Treasury (Treasury) subcontractor’s analysis, and provides an
estimate of how large losses potentially could be in the event that the
markets returned to their November 2008 lows. Selecting November 2008
as the market low point is generally consistent with the approach used by
the Board of Governors of the Federal Reserve System (Federal Reserve)
in its “stress tests” for determining the capital that large bank holding
companies must maintain. Our scenario provides a more-adverse than
expected loss estimate for our sample of ABSs and all of the CMBS loans
remaining as of September 30, 2009.
Our first analysis focused on ABSs. We conducted a market value analysis
on a sample of the three largest asset classes—ABSs backed by credit
cards, auto loans, and student loans—because they make up the majority
of TALF’s portfolio. Of the $42.5 billion in TALF loans backed by credit
card, auto loan, and student loan ABSs that had been disbursed as of
September 30, 2009, we took a sample of $16.5 billion, or 39 percent. The
sample was selected to broadly match the makeup of these asset classes in
this subset of the TALF portfolio (see table 4).
Table 4: GAO Sample Selection for ABS Extreme Market Value Loss Analysis
Dollar in billions
Percent of Percent of TALF Portfolio
sample
(as of Sept. 30, 2009)

Asset class

Collateral

Credit cards

$9.35

57%

56%

4.54

27

30

Auto loans
Student loans
Total

2.64

16

14

$16.53

100%

100%

Source: GAO.

Within each asset class the sample was selected to include ABSs that gave
the largest sample size on a TALF loan dollar basis; hence larger deals
predominate. In addition, TALF loans were spread across 6 of the 7 TALF
ABS subscription months between March and September. Nevertheless, it
is a nonprobability sample and is not necessarily generalizable to all TALF
deals.

Page 60

GAO-10-25 Troubled Asset Relief Program

Appendix II: Methodology for Market Value
Analysis of ABSs and CMBSs

We modeled TALF collateral cash flows using assumptions consistent with
the FRBNY’s assumptions on the rate at which ABS principal is paid back
to investors. Then we calculated the discount rate that brought the price
back to par as of the issuance date. To this discount rate, we added the
incremental yield (or spreads) that would be required to deplete the
borrowers’ entire equity investment and any excess interest that had built
up in TALF LLC as of September 30, 2009, for that specific tranche.1 These
incremental spreads were compared with the widest levels seen in
November 2008 for the appropriate asset class and expected average life.
November 2008 spreads were obtained from Wall Street ABS-dealer
weekly price data that are published for the more widely traded ABS
classes. If the spread seen in November 2008 was greater than that
required to deplete the borrower’s equity and TALF’s excess interest, a
stress loss was calculated. No loss was assumed if the required spread
widening was less than the extremes of November 2008.
Our analysis makes the following assumptions: (1) excess interest has
accumulated as of September 30, 2009 at the tranche level of each TALF
security; (2) borrowers will surrender their TALF ABS collateral to the
FRBNY and stop paying the TALF loan when the ABS market value falls
below the TALF loan balance; (3) TALF will mark-to-market the
surrendered collateral, ignoring any recovery that Treasury might make if
the ABS collateral fully pays all cash flows over the life of the securities;
and (4) the change in market value is strictly based on mark-to-market,
with no assumption about the underlying credit performance of the ABS.
For the separate analysis on the risks that legacy CMBS collateral may
pose to TARP funds, we compared the prices on the 139 legacy CMBS
CUSIPs that were accepted by FRBNY as of September 30, 2009, with the
exception of 2 for which no price information was available.2 These prices
were then compared with the lowest prices that, on average, most CMBS
across the TALF portfolio reached in November 2008. This CMBS analysis
did not include consideration of the excess interest accumulated in TALF
LLC but otherwise made the same assumptions noted above for the ABS
analysis. As discussed earlier, the $198 million of excess interest that had
accumulated in the cash collateral account as of December 31, 2009,
would be available to absorb the first losses bourn on surrendered

1

A tranche is a piece of a securitization that has specified risks and returns.

2

CUSIP stands for the Committee on Uniform Security Identification Procedures. CUSIP
numbers are alphanumeric identifiers assigned to individual securities.

Page 61

GAO-10-25 Troubled Asset Relief Program

Appendix II: Methodology for Market Value
Analysis of ABSs and CMBSs

collateral prior to any outlay by Treasury, and this amount is expected to
increase over time. In conducting this analysis we utilized certain data
from Thompson Reuters Datascope and the Federal Reserve.

Page 62

GAO-10-25 Troubled Asset Relief Program

Appendix III: The Securitization Process
Explained

Appendix III: The Securitization Process
Explained
Securitization is a process that packages relatively illiquid individual
financial assets—such as loans, leases, or receivables—and converts them
into interest-bearing, asset-backed securities (ABS) that are marketable to
capital market investors. As outlined below, the market participants in
securitization—borrowers, consumer and small business lenders,
investment banks or pool assemblers, credit rating agencies, and
investors—each derive specific benefits from the transaction. For
example, borrowers might gain access to loanable funds with more
favorable terms, such as longer repayment periods and lower interest
rates, than may otherwise be available. Similarly, securitization offers
consumer and small business lenders a funding source for making new
loans, improving balance sheet and capital management, and diversifying
fee or income streams. Securitization also allows the cash flows from asset
pools to be structured to satisfy the maturity, risk, and return preferences
of investors.
The degree to which participants receive these benefits depends, in large
part, on how efficiently the markets for securitized assets are functioning.
With accurate and more comprehensive performance data regarding
financial assets, capital markets can more easily profile the risk of a pool
of similar assets. This risk can be divided and sold to investors who are
willing to purchase it at an acceptable risk-adjusted return, sometimes
called the “investor-required yield.” As the markets for securitized asset
classes grow in volume and liquidity, and as the performance and risk
characteristics of those assets become better understood, investorrequired yields on particular ABSs and transaction costs of securitizing
those assets may decline. Declining investor-required yields and
transaction costs can lower the cost of financing for consumer and small
business lenders and ultimately borrowers. Conversely, with inadequate
performance data, and low volumes of similar financial assets, these
benefits may not sufficiently materialize for securitization to be a viable
financing arrangement for consumer and small business lenders or
borrowers.

Page 63

GAO-10-25 Troubled Asset Relief Program

Appendix III: The Securitization Process
Explained

Figure 11: The Securitization Process

Investors

Investors purchase securities with cash flows that have desirable risk-return and
maturity characteristics. A single pool can often contain multiple classes, or "tranches,"
of securities.
The issuing trust holds the pool of assets which are insulated from the performance
and credit of the underwriter and originating lenders. The issuing trust sells securities
to investors.

Issuing
trusts

NRSROs

Underwriters

Nationally recognized statistical rating organizations (NRSRO), or credit rating
agencies, assess the performance and expected losses (credit quality) of the pool of
assets, including internal and external credit enhancements, and provide a credit rating
on the securities to be sold.
The underwriter structures assets within the ABS trust and facilitates the sale of
securities to investors. Underwriters, as part of structuring, stratify the credit and
payment positions of cash flows generated from the pool into different classes of
securities, or tranches, based on investor preferences.
For the Term Asset-Backed Securities Loan Facility (TALF), auditors either provide an
attestation or use agreed upon procedures to certify certain characteristics of TALF
collateral.

Auditors

Lenders
(e.g., banks and
auto finance
companies)

Consumer borrowers

Lenders originate loans that conform to underwriting criteria acceptable to a pool of
loans and ultimately sell the loan into the pool. Lenders may fund credit enhancements
to support the credit quality of a pool of loans and service the loan by collecting
payments for distribution to the issuing trust, which in turn remits payments to
bondholders.
Borrowers finance consumer spending using credit cards, auto and student loans, and
then provide specified repayments of principal and interest to lenders.
Source: GAO.

Page 64

GAO-10-25 Troubled Asset Relief Program

Appendix IV: Descriptions of Asset-Backed
Securities

Appendix IV: Descriptions of Asset-Backed
Securities
Auto Loan and Lease

Asset-backed securities (ABS) for the auto market compose the largest
share of ABS issuances. Auto securitizations are collateralized with a fixed
pool of loans. In most cases, these transactions are divided into at least
four senior segments, or tranches, which have the same payment priority
in the event of default but different priorities for principal repayment (with
the exception of the shortest pay securities or A1 tranche designed to be
marketable to money market funds, which take priority). Tranches are
structured so that all scheduled principal amortization and prepayments of
principal are paid back first to the tranche with the lowest interest rate.
This tranche is generally designated the A1 tranche. Once the principal on
the first tranche is paid off, subsequent principal is paid to the A2, A3, and
A4 bondholders. The sizes of the tranches are designed so that the
expected average life on these securities is generally consistent within
each tranche—for instance, the A1 average life is usually 3 months, the A2
average life 1 year, the A3 average life 2 years, and the A4 average life
approximately 3 years or more. In some deals, there are also “subordinate
tranches,” or tranches that receive ratings below AAA. In many cases, the
issuer retains subordinate tranches rather than selling them to the public.
Auto ABSs include the following subasset classes: prime auto loans,
subprime auto loans, auto leases, and motorcycle loans.

Credit Card

Credit card ABSs tend to use a master trust structure through which a
credit card issuer collateralizes a series of ABS issuances with receivables
from a large pool of credit card accounts. This pool is not a static set of
account balances but absorbs new receivables as they are created. New
issuance can be used to support an increase in the size of the receivables
collateralizing the securitizations. Investments in credit card ABSs are
usually divided into senior and subordinated, or junior, tranches where the
investors in the senior tranches are paid first.

Student Loan

Student loan ABSs can be collateralized with either federally guaranteed
Federal Family Education Loan Program (FFELP) loans or consumer
loans that are not part of a government guarantee program. Student loan
ABSs tend to have longer terms to maturity than other ABS classes due to
the longer repayment terms and the fact that students do not tend to pay
any principal or interest until at least 6 months after they graduate, thus
lenders might not receive cash flows on a student loan for years after the
initial cash disbursement.

Page 65

GAO-10-25 Troubled Asset Relief Program

Appendix IV: Descriptions of Asset-Backed
Securities

Insurance Premium
Finance

Insurance premium finance ABSs are collateralized with loans made to
businesses to finance their property and casualty insurance coverage. The
typical commercial insurance policy requires a down payment, with equal
monthly payments, typically over a time frame shorter than the term of the
insurance policy, which in effect creates overcollateralization. When a
policy is cancelled, refunds of unearned premiums will be used for making
payments to the securitization trust for the remaining term of the loan to
protect the ABS holders.

Commercial
Mortgage-Backed
Securities

Cash flows on commercial mortgage-backed securities (CMBS) are
generally backed by principal and interest payments on a pool of
commercial mortgage loans. Most commercial mortgage loans are
structured with a 30-year amortization term, but CMBS terms are generally
shorter than the corresponding amortization terms. However, recent years
have seen an increase in the number of loans with interest-only periods
during which the mortgagee pays no principal on the mortgage.
Commercial mortgages are made on a wide variety of different property
types, including rental apartment buildings, industrial properties, office
buildings, hotels, healthcare related properties, and retail properties such
as shopping malls, strip malls, and freestanding outlets. CMBSs are highly
structured and frequently have more than 20 tranches in their capital
structure. The coupon payment generally is positively correlated with both
the expected average life of the tranche and the risk that the bondholder
will not receive the entire principal amount. Also differentiating CMBSs
from other asset classes is their sensitivity to the underlying commercial
real estate prices and the cash flow generated from the commercial
properties backing the mortgages.

Commercial Fleet
Leases

Unlike other ABS classes for which investors own direct stakes in the trust
assets, commercial fleet lease ABSs are collateralized with special units of
beneficial interest (SUBI) in open-ended leases and fleet management
receivables on a pool of vehicle leases mainly for commercial trucks,
trailers, and equipment. The leases are made on a per-vehicle basis to large
corporate customers with fleets that may have more than 5,000 vehicle
leases with the issuer. Open-ended leases require the lessee to reimburse
any loss in a vehicle’s residual value to the lessor. Commercial fleet ABSs
usually are structured out of a master trust with the ability to issue
numerous term securities. Collateral in the master trust can be replenished
with new or renewed leases as older contracts prepay and expire. The
lease SUBI entitles the ABS holders to receive the monthly lease
payments. This SUBI also includes beneficial interest in all the vehicles

Page 66

GAO-10-25 Troubled Asset Relief Program

Appendix IV: Descriptions of Asset-Backed
Securities

that are being leased or are in the process of being leased but have not
completed the process. The fleet management receivables SUBI includes
beneficial interest in the receipt of management and other fees that the
lessees pay to the lessor.

Mortgage Servicing
Advances

Mortgage servicing advance ABSs are collateralized with receivables owed
to the servicer for servicing advances made by the servicer to and on
behalf of the residential mortgage-backed securitization (RMBS) trusts.
There are three types of advances: principal and interest, which cover
these types of payments on delinquent loans; escrow advances, which
cover expenses related to maintaining ownership of a mortgaged property,
including property taxes and insurance premiums; and corporate
advances, which are costs for the process of foreclosure, including
attorney and other professional fees and expenses related to maintaining a
repossessed home. As there is no interest paid to the RMBS trusts when
advances are paid back out of either the proceeds of the liquidation of a
repossessed property (loan-level servicing advances) or broader pool level
cash flows (pool-level servicing advances), a discount factor is applied.
The discount factor reflects the estimated time frame for repaying the
loan. As a result of this discount factor, the issuer receives less than the
face value of the servicing advance at the time of securitization.

Dealer Floor Plans

Floor plan ABSs are collateralized by loans made to finance either
automobiles or nonautomotive durable goods. Nonautomotive floor plan
inventory includes, among other things, recreational vehicles, boats,
motorcycles, industrial equipment and farm equipment, appliances, and
electronics. Automotive dealer floor plan arrangements tend to be
between a single financial entity and a dealer network. Nonauto dealers
can have multiple floor plan arrangements with several capital providers.
Financing could be in the form of revolving or nonrevolving lines of credit.
Once a floor plan agreement is in place, dealers place orders for inventory
from the manufacturer and specify that a lender will provide the financing.
The loan is repaid by proceeds from inventory sales, or the dealer can
arrange to repay the loan in monthly installments.

Equipment

Equipment ABSs are collateralized with retail installment sale contracts,
loans and leases secured by new and used agricultural equipment,
construction equipment, industrial equipment, office equipment, copiers,
computer equipment, telecommunications equipment, and medical
equipment, among other things.

Page 67

GAO-10-25 Troubled Asset Relief Program

Appendix IV: Descriptions of Asset-Backed
Securities

SBA-Loan Backed
ABS

The Small Business Administration (SBA) provides guarantees on loans
made to small businesses. The most common SBA loan programs are 7(a)
and 504. In the 7(a) program, SBA guarantees up to 85 percent of the loan
amount made by participating lenders. 7(a) loans are usually made for
general business purposes, including working capital, equipment, furniture
and fixtures, and land and buildings. 504 program loans are typically longterm, fixed-rate loans for the purpose of expanding or modernizing a small
business. When pooled together for securitization purposes, underlying
loans must have similar terms and features—for example, similar maturity
dates.

Page 68

GAO-10-25 Troubled Asset Relief Program

Appendix V: Credit Enhancement

Appendix V: Credit Enhancement

Credit enhancements are features in the structuring of a securitization that
protect investors in the securitization from losses due to defaults on the
underlying loans. The following are some methods of credit enhancement
that have been used on asset-backed securities (ABS) eligible for the Term
Asset-Backed Securities Loan Facility (TALF).
Subordination: This feature is a method of prioritizing cash flows from
the underlying loan collateral. The senior tranches within a securitization
get priority over subordinate, or junior, tranches in the event of a default
on the underlying collateral.1 All TALF-eligible securitizations must have a
AAA rating from at least two TALF-eligible nationally recognized statistical
rating organizations, and all of the AAA-rated ABSs have first priority for
cash flows. While most AAA tranches or bonds within an ABS have the
same priority in the event of a default, the sequential nature of the
principal paydown for certain classes of ABSs (for example, auto loans)
leads to higher risk of default for the tranches with weighted average lives
that extend further into the future. This higher risk requires the issuer to
pay a higher interest rate or coupon on longer tranches. Any losses are
applied to the most subordinate tranche first.
Overcollateralization: When the total face value on the loan collateral
underlying an ABS is greater than the face value of the bonds, the
securitization is said to be overcollateralized. These assets are maintained
on the balance sheet of the issuer and are the first to absorb credit losses
on the collateral.
Reserve account: This is a cash account set up at the origination of an
ABS. This account is accessed when the cash flows from the collateralized
loan assets are insufficient to cover the contractual payments on the
bonds, including servicing and other fees.
Excess spread: Excess spread refers to the funds leftover after payments
to bondholders and other contractual obligations have been met. This can
be used to make up for insufficient cash flows if the underlying borrowers
are delinquent or default on the loan.

1
Some securitizations—such as ABSs backed by auto loans—are divided into different
segments, or tranches. A tranche is a piece of a securitization that has specified risks and
returns.

Page 69

GAO-10-25 Troubled Asset Relief Program

Appendix V: Credit Enhancement

Yield-supplement overcollateralization: This feature applies to
securitizations with assets in the underlying pool that are paying interest
that is below the coupon rate on the bonds. For example, borrowers
frequently pay very low interest rates on loans within certain auto loan
securitizations. These loans are often extended with advantageous
borrower terms as part of a sales promotion. Generally, the issuer will set
up a yield-spread overcollateralization account to make up the difference
over some portion of the life of the securitization. The initial balance in
this account is set as the present value of the shortfall on those loans for
the life of the loans.
Mortgage servicing advance discounts: A form of enhancement that is
implicit in the discounted price at which the securitization trust purchases
the servicing advance receivables from the mortgage servicing company
issuing the security. The servicing advances are segregated into several
classifications based on whether the servicing advances are treated at the
pool- or loan-level in order of repayment, whether the underlying mortgage
is located in a state that has a judicial or nonjudicial foreclosure regime,
and the type of cash flow for which the servicer is advancing payment. The
servicing advance is classified into one of three classes: principal and
interest advance, escrow advances, and corporate advances. Principal and
interest advances are made by mortgage servicers to holders of residential
mortgage-backed securities for mortgages whose underlying borrowers
are delinquent on their monthly payments. Escrow advances are used to
pay the property taxes, insurance premiums, or other property-related
expenses that the borrowers should have paid. Corporate advance costs,
usually in the form of attorneys’ and other professional fees, are also
accounted for in the event that the servicer incurs them while foreclosing
on and liquidating repossessed real estate.
The advance discount percentage is calculated based on assumptions
about the length of time it will take to repay that particular type of
advance and the risk that it might not be paid back. Pool-level servicing
advances have the lowest discount percentage, because the advances can
be repaid to the servicer out of the entire pool’s available funds, including
principal and interest payments received for nondelinquent mortgages.
Loan-level servicing advances are not repaid until the borrower repays all
the money advanced or from the proceeds of the sale of the repossessed
property (the likely scenario in a default). Servicing advances on
mortgages secured with properties in judicial foreclosure states have
higher discount rates than those in nonjudicial states, because judicial
foreclosures take longer. Principal and interest servicing advances are

Page 70

GAO-10-25 Troubled Asset Relief Program

Appendix V: Credit Enhancement

viewed as the safest instruments and have lower discount rates than
escrow, which in turn has slightly lower discount factors than corporate
advances.

Page 71

GAO-10-25 Troubled Asset Relief Program

Appendix VI: Additional Information on TALF
Compliance

Appendix VI: Additional Information on TALF
Compliance
According to Federal Reserve Bank of New York (FRBNY) officials,
FRBNY has in place a number of compliance measures to (1) ensure that
borrowers and collateral are eligible for the Term Asset-Backed Securities
Loan Facility (TALF); (2) protect FRBNY from fraudulent activity;
(3) reduce the risk of fraud and address conflicts of interest; (4) ensure
that agents have adequate compliance regimes; and (5) build multiple
layers of compliance where possible.
TALF has a certification regime in place for a number of TALF
participants. TALF agents and sponsors must certify that they are
complying with certain TALF requirements, and TALF agents review the
eligibility of TALF borrowers. According to FRBNY officials, TALF agents
are the first line of defense against fraudulent participants in TALF, as they
conduct “Know Your Customer” reviews of potential TALF borrowers.
FRBNY noted that it had antifraud measures in place and receives
referrals for those investors that TALF agents raised concerns about.
Moreover, FRBNY officials stated that they have developed an inspection
program to conduct on-site reviews of TALF Agent’s “Know Your
Customer” programs and files. The entire process is under the
management of FRBNY, without the Department of the Treasury’s
(Treasury) participation.
Sponsors and issuers must include in any offering document a certification
required by FRBNY. Borrowers must also provide representations to the
TALF agent, who conducts the review of the borrower. According to
FRBNY officials, because the issuers and sponsors include certification to
TALF eligibility and acknowledge certain responsibilities related to the
TALF collateral in the offering documents, any material
misrepresentations would be covered under relevant securities laws. In
addition, the TALF agents and borrowers also make certain
representations on their eligibility and the eligibility of the collateral.
Though this is a certification and self-disclosure regime, FRBNY officials
told us they have established additional measures to detect and address
noncompliance. First, FRBNY has a 24-hour fraud hotline. Second, it has
hired a law firm to assist in assessing fraud risks associated with the
program. Third, it is cooperating with other government and law
enforcement agencies to gather additional information on potential TALF
participants.

Page 72

GAO-10-25 Troubled Asset Relief Program

Appendix VI: Additional Information on TALF
Compliance

In addition to certifications, FRBNY requires auditor attestations for nonCMBS collateral, which state that the Report on Management Compliance
fairly states compliance with certain TALF program criteria specified by
FRBNY. For CMBS collateral, agreed upon procedures (AUP) are required
to provide more detailed specifications on what to review.1 FRBNY has
published broad guidelines to the auditors for carrying out these
responsibilities, which are paid for by the issuers of TALF-eligible
securities. In addition, FRBNY requires that the attestation and AUP
processes follow standards issued by the Public Company Accounting
Oversight Board and the American Institute of Certified Public
Accountants. Most of the information that the auditors review is provided
by the issuers and is not verified independently, according to auditors we
spoke with. FRBNY officials added that loan-level testing is required and
this includes a review of original loan files or electronic versions thereof.
According to Treasury officials, Treasury provided some input into the
design of the auditor attestations and AUPs but primarily leaves oversight
of this function to the Board of Governors of the Federal Reserve System
and FRBNY, which are responsible for designing and implementing TALF.
Treasury does not review these documents; however, should the assets be
placed to TALF LLC, Treasury may review them.

1
According to the guidance provided by the American Institute of Certified Public
Accountants, Public Company Accounting Oversight Board, and discussions with two
auditors involved in the process, attestations are generally considered a higher form of
assurance than AUPs. This is because in an attestation the auditor must attest to all of its
statements and design a methodology that provides reasonable assurance of the accuracy
of its attestation. In the case of an AUP, the process is stipulated ahead of time and the
auditor simply follows it, providing a statement as to the outcome of the process.

Page 73

GAO-10-25 Troubled Asset Relief Program

Appendix VII: New Securitization Volumes
Have Increased since the Inception of the
TALF Program

Appendix VII: New Securitization Volumes
Have Increased since the Inception of the
TALF Program
Since the Term Asset-Backed Securities Loan Facility’s (TALF) March
2009 inception, securitization volumes have increased in some TALFeligible sectors. For auto loan securitization, new issuance dropped off
significantly in the third quarter of 2008, bottoming in the fourth quarter
(see table 5). By 2009 issuance began to pick up, especially in the second
and third quarters, when most were TALF-eligible. There were 45
issuances through December 2, 2009, a marked contrast to the peak of 85
in 2005.
Table 5: Auto Loan Securitization Volume, 2005 through Fourth Quarter 2009
Dollars in millions
Number of non- TALF Issuance
TALF
Volume
securitizations
(dollars)

Non-TALF
issuance
volume
(dollars)

Deals

Total issuance
volume

Number of TALF
securitizations

2005

85

$76,912

–

2006

78

88,114

–

–

–

–

2007

69

71,015

–

–

–

–

2008

32

29,113

–

–

–

–

Year of issuance

–

–

–

2009

45

51,192

35

10

$44,941

$6,251

Q1 2008

10

10,265

–

–

–

–

Q2 2008

14

14,022

–

–

–

–

Q3 2008

5

3,226

–

–

–

–

Q4 2008

3

1,600

–

–

–

–

Q1 2009

5

7,583

3

2

5,213

2,370

Q2 2009

11

14,952

10

1

13,445

1,507

Q3 2009

14

19,294

12

2

18,394

900

Q4 2009

15

9,363

10

5

7,889

1,474

Source: GAO analysis of Thomson Reuters IFR Markets data.

Credit card securitization volumes show a similar pattern (see table 6).
The peak in credit card securitizations occurred in 2007, with 112
issuances, a sharp contrast to 2009 when only 35 securitizations were
issued through December 2, 2009. The majority of credit card ABSs
issuances in 2009—about 65 percent—have been TALF supported.

Page 74

GAO-10-25 Troubled Asset Relief Program

Appendix VII: New Securitization Volumes
Have Increased since the Inception of the
TALF Program

Table 6: Credit Card Receivables Securitization Volume, 2005 through Fourth Quarter 2009
Dollars in millions
Year of
issuance

Number of nonTALF
securitizations

TALF
issuance
volume

Non-TALF
issuance
volume

Deals

Dollar volume

Number of TALF
securitizations

2005

80

$53,019

–

–

–

–

2006

93

60,374

–

–

–

–

2007

112

94,539

–

–

–

–

2008

67

67,319

–

–

–

–

2009

35

45,988

23

12

$29,713

$16,275

Q1 2008

25

28,785

–

–

–

–

Q2 2008

24

28,180

–

–

–

–

Q3 2008

18

10,354

–

–

–

–

Q4 2008

0

0

–

–

–

–

Q1 2009

2

6,500

1

1

3,000

3,500

Q2 2009

15

20,385

8

7

13,835

6,550

Q3 2009

15

15,500

12

3

10,775

4,725

Q4 2009

3

3,603

2

1

2,103

1,500

Source: GAO analysis of Thomson Reuters IFR Markets data.

Student loan securitization volumes show similar patterns to the auto and
credit card asset classes, with a marked low of no new deals in the last
quarter of 2008 (see table 7). Lenders may be tightening their lending
standards, potentially resulting in fewer loans and reducing the need to
access the securitization markets as frequently as in the past. Although
both Federal Family Education Loan Program (FFELP) and private loan
securitizations are TALF eligible, to date no FFELP deals have been
underwritten to TALF eligibility standards. There have been five TALF
private student loan securitizations since the program’s inception.

Page 75

GAO-10-25 Troubled Asset Relief Program

Appendix VII: New Securitization Volumes
Have Increased since the Inception of the
TALF Program

Table 7: Student Loan Securitization Volume, 2005 through Fourth Quarter 2009
Dollars in millions
Total
student
loan deals

Total dollar
volume

FFELP
deals

2005

42

$46,277

–

–

–

–

–

–

2006

43

70,058

–

–

–

–

–

–

2007

28

50,672

23

5

$42,663

$8,009

–

–

2008

22

29,427

21

1

29,302

125

–

–

2009

20

21,247

13

7

12,350

8,897

$7,367

$1,530

Year of
issuance

Private
Private dollar
deals FFELP volume
volume

Private Private NonTALF dollar TALF dollar
volume
volume

Q1 2008

6

8,400

6

0

8,400

0

–

–

Q2 2008

10

14,624

10

0

14,624

0

–

–

Q3 2008

6

6,403

5

1

6,278

125

–

–

Q4 2008

0

0

0

0

0

0

–

–

Q1 2009

2

2,047

1

1

547

1,500

0

1,500

Q2 2009

5

8,315

4

1

5,722

2,593

2,593

0

Q3 2009

5

6,124

1

4

1,910

4,214

4,184

30

Q4 2009

8

4,761

7

1

4,171

590

590

0

Source: GAO analysis of Thomson Reuters IFR Markets data.

Note: Thomson Reuters IFR Markets data does not break out FFELP and private securitizations until
2007.

This report discussed the severe disruption in the commercial real estate
sector following the economic downturn. Table 8 shows that commercial
mortgage-backed securities (CMBS) volumes peaked in 2006 with 97
issuances before dropping dramatically to just 7 by 2008. These sharp
declines in part motivated the inclusion of CMBSs as a TALF-eligible asset
class. Three new-issue deals have been offered since TALF was expanded
to CMBSs, and only one used TALF for financing. Other CMBS deals in
2009 were repackaging of existing securitizations. Part of the sluggish
activity in the CMBS sector could be attributed to the length of time it
takes to put together a deal, which officials have noted is considerably
longer than for the other asset classes. There could be other reasons as
well. As we discussed in this report, the CMBS sector continues to show
signs of volatility resulting from sharp declines in commercial real estate
prices and increases in CMBS delinquency rates.

Page 76

GAO-10-25 Troubled Asset Relief Program

Appendix VII: New Securitization Volumes
Have Increased since the Inception of the
TALF Program

Table 8: CMBS Securitization Volume, 2005 Through Fourth Quarter 2009
Dollars in millions
Year of issuance

Number of deals

Par value

TALF

Non-TALF

2005

80

$152,271

–

–

2006

97

201,419

–

–

2007

73

199,925

–

–

2008

7

9,482

–

–

2009

6

2,500

$400

$2,100

Q1 2007

17

45,959

–

–

Q2 2007

25

74,360

–

–

Q3 2007

19

54,236

–

–

Q4 2007

12

25,370

–

–

Q1 2008

3

4,387

–

–

Q2 2008

2

1,955

–

–

Q3 2008

2

3,140

–

–

Q4 2008

0

0

–

–

Q1 2009

0

0

–

–

Q2 2009

2

343

–

343

Q3 2009

0

0

–

–

Q4 2009

4

2,157

400

1,757

Source: GAO analysis of Thomson Reuters IFR Markets data.

Page 77

GAO-10-25 Troubled Asset Relief Program

Appendix VIII: Comments from the
Department of the Treasury

Appendix VIII: Comments from the
Department of the Treasury

Page 78

GAO-10-25 Troubled Asset Relief Program

Appendix VIII: Comments from the
Department of the Treasury

Page 79

GAO-10-25 Troubled Asset Relief Program

Appendix IX: Comments from the Board of
Governors of the Federal Reserve System

Appendix IX: Comments from the Board of
Governors of the Federal Reserve System

Page 80

GAO-10-25 Troubled Asset Relief Program

Appendix IX: Comments from the Board of
Governors of the Federal Reserve System

Page 81

GAO-10-25 Troubled Asset Relief Program

Appendix IX: Comments from the Board of
Governors of the Federal Reserve System

Page 82

GAO-10-25 Troubled Asset Relief Program

Appendix IX: Comments from the Board of
Governors of the Federal Reserve System

Page 83

GAO-10-25 Troubled Asset Relief Program

Appendix IX: Comments from the Board of
Governors of the Federal Reserve System

Page 84

GAO-10-25 Troubled Asset Relief Program

Appendix IX: Comments from the Board of
Governors of the Federal Reserve System

Page 85

GAO-10-25 Troubled Asset Relief Program

Appendix IX: Comments from the Board of
Governors of the Federal Reserve System

Page 86

GAO-10-25 Troubled Asset Relief Program

Appendix X: Analysis of Legal Comments
Submitted by the Federal Reserve

Appendix X: Analysis of Legal Comments
Submitted by the Federal Reserve
As noted, in its comments, the Board of Governors of the Federal Reserve
System (Federal Reserve) did not agree with our recommendation that
Congress consider providing GAO with authority to audit the Federal
Reserve’s operational and administrative actions because it disagreed that
there are limitations on GAO’s authority to audit these Federal Reserve
activities.
However, we believe the express statutory prohibition in 31 U.S.C. §
714(b) on GAO auditing the Federal Reserve’s monetary policy and
discount window lending activities, which the Federal Reserve believes
includes the Term Asset-Backed Securities Loan Facility’s (TALF)
operation and administration, prohibits us from auditing the Federal
Reserve’s TALF activities, even from the perspective of the Troubled Asset
Relief Program (TARP). We limited the scope and conduct of this audit
accordingly, and thus did not request access to information to audit the
Federal Reserve’s performance of these activities.1 Further, the Federal
Reserve’s decision to voluntarily provide requested access in this instance,
while helpful, does not create GAO authority to audit the Federal
Reserve’s TALF operational activities or other performance. In our view,
our lack of authority to audit the Federal Reserve’s actions limited our
ability to fully assess the risk to taxpayer funds presented by TALF.
The basis of the Federal Reserve’s view that GAO has authority to audit its
TALF operational and administrative actions is a May 2009 amendment to
the Emergency Economic Stabilization Act of 2008 (EESA), the statute
authorizing TARP. While GAO vigorously pursues its audit and access
authority in all appropriate circumstances, the amendment referenced by
the Federal Reserve, section 601 of the May 2009 amendments, provides
no authority for GAO to audit the Federal Reserve. 2 Rather, responding to
congressional concern that GAO lacked authority to obtain access to

1

A performance audit of the Federal Reserve’s TALF operational and administrative
activities would, for example, have involved evaluating the process by which certain TALF
program terms were arrived at, such as whether the haircuts or the amount of equity the
TALF borrower holds in the collateral protected TALF from losses. We also would have
evaluated FRBNY’s system of internal controls and the role of TALF participants in
certifying and validating compliance with certain program requirements. In addition, we
would have interviewed some of the entities that helped administer TALF to validate
agency information and to better understand their roles and interactions with the Federal
Reserve and FRBNY. These would have included entities such as the TALF agents, Bank of
New York Mellon, Trepp, or CW Capital.
2

See section 601 of the Helping Families Save Their Homes Act of 2009, Pub. L. No. 111-22,
123 Stat. 1632, 1659 (May 20, 2009), codified at 12 U.S.C. § 5226.

Page 87

GAO-10-25 Troubled Asset Relief Program

Appendix X: Analysis of Legal Comments
Submitted by the Federal Reserve

information directly from banks and other firms receiving TARP funds, the
amendment provided GAO with such access to enable us to more
effectively review Treasury’s actions under our existing TARP audit
authority. As the Federal Reserve correctly noted in its comments, section
601 included access to records of any entity “that is established by a
Federal reserve bank and receives funding from the TARP,” thus covering
records of TALF LLC, the special purpose vehicle created by the Federal
Reserve Bank of New York to which Treasury has committed up to $20
billion of TARP funds. But GAO’s authority to obtain access to records of
TALF LLC does not provide GAO access to other TALF program
information, nor GAO authority to audit the Federal Reserve’s operation
or administration of TALF.
As support for its view that the May 2009 amendment in section 601
authorized GAO to audit the Federal Reserve’s TALF performance, the
agency quoted a portion of remarks made by Senator Grassley, a lead
sponsor of the amendment. As Senator Grassley noted, however, he was
describing the Federal Reserve’s position. The Senator provided material
for the record stating in part, “According to Federal Reserve staff, . . .
[Senate] amendment No. 1020 [section 601] would expand GAO’s authority
to oversee TARP, including the joint Federal Reserve-Treasury Term
Asset-Backed Securities Loan Facility (TALF) . . ..”3 As noted, however, the
Federal Reserve was only partially correct: section 601 provided GAO
authority to access records of the Treasury-funded TALF LLC in order to
audit Treasury, but not authority to audit and evaluate the Federal
Reserve’s TALF actions.
That section 601 did not authorize GAO to audit the Federal Reserve is
confirmed by the legislative history of a second amendment authored by
Senator Grassley, section 801 of the May 2009 amendments. An earlier
version of section 801 would have modified GAO’s Federal Reserve
authority under 31 U.S.C. § 714 to authorize audit of all Federal Reserve
emergency actions under section 13(3) of the Federal Reserve Act, as well
as Federal Reserve TARP-related actions, thus giving GAO TALF audit
authority. Changes to section 801 shortly before the committee vote
eliminated GAO’s TALF audit authority, however, leaving GAO authorized
only to audit Federal Reserve actions taken under section 13(3) “with
respect to a single and specific partnership or corporation”—that is, AIG,

3

155 Cong. Rec. S5181 (daily ed. May 6, 2009).

Page 88

GAO-10-25 Troubled Asset Relief Program

Appendix X: Analysis of Legal Comments
Submitted by the Federal Reserve

Citigroup, Bank of America, and Bear Stearns. This language was enacted
as a new subsection (e) to 31 U.S.C. § 714.4
Finally, the Federal Reserve commented that because of TALF’s unique
“hybrid” nature—it serves objectives of both monetary policy and TARP—
GAO has “ample authority” to audit TALF operations, Treasury’s
participation in them, and the Federal Reserve’s administration of TALF
“on behalf of” Treasury, all from the perspective of TARP. In this regard,
the Federal Reserve noted that in practice, it obtains Treasury’s input and
agreement on many aspects of TALF. However, the view that GAO can
separately audit the Federal Reserve’s TALF performance as long as the
audit is limited to TARP objectives, without violating the statutory
prohibition against GAO auditing Federal Reserve monetary policy
actions, is not supported by either the language of 31 U.S.C. § 714, its
original legislative history, or the amendments Congress enacted to it in
May 2009.
Section 714(b)(2) prohibits GAO from auditing the Federal Reserve with
respect to “deliberations, decisions, or actions on monetary policy matters,
including discount window operations . . ..” In the Federal Reserve’s view,
all of its TALF activities are an exercise of its monetary policy and
discount window lending authority, thus prohibiting GAO from auditing
these actions at least from the perspective of monetary policy. While the
Federal Reserve stated it believes the prohibition does not apply to nonmonetary policy aspects of “hybrid” Federal Reserve actions, permitting
GAO to audit the agency from the perspective of TARP, for example, the
debate leading to enactment of § 714(b) in the Banking Agency Audit Act
of 1978 indicates Congress did not intend to create such an exception.
Representative Ashley, the author of the final version of the § 714(b)
prohibitions, made clear that the monetary policy prohibition extends to
all discount window lending actions, even those serving multiple
objectives such as regulation/supervision and monetary policy. Rep.
Ashley explained that all discount window lending is necessarily

4
Section 801 (Senate Amendment 1021) as introduced would have removed all restrictions
on GAO’s authority to audit the Federal Reserve in 31 U.S.C. § 714(b), including the
prohibition against auditing monetary policy actions. 155 Cong. Rec. S4991 (daily ed. April
30, 2009). As detailed by Senator Grassley, a substitute amendment then narrowed GAO’s
audit authority to all Federal Reserve section 13(3) actions and TARP-related actions, 155
Cong. Rec. S5118-19 (daily ed. May 5, 2009), and the final version narrowed GAO’s authority
even further, to Federal Reserve section 13(3) actions “with respect to a single and specific
partnership or corporation.” 155 Cong. Rec. S5173-74, S5181-82 (daily ed. May 6, 2009); sec.
801, Pub. L. No. 111-22, 123 Stat. 1662-63 (May 20, 2009).

Page 89

GAO-10-25 Troubled Asset Relief Program

Appendix X: Analysis of Legal Comments
Submitted by the Federal Reserve

“inextricably bound up in monetary policy” and is intended to be covered
by the prohibition.5 Under this reading, GAO is prohibited from auditing
the Federal Reserve’s TALF discount window lending activities even from
the perspective of TARP.
The Federal Reserve’s position also conflicts with Congress’ enactment of
§ 714(e) in May 2009, noted above, authorizing GAO to audit Federal
Reserve section 13(3) actions with respect to a single and specific
partnership or corporation. If GAO already could audit TARP aspects of
“hybrid” Treasury and Federal Reserve activity, such additional authority
would have been unnecessary regarding AIG, for example, because both
Treasury and the Federal Reserve already were providing assistance to
AIG. Yet GAO was required, as the Federal Reserve then agreed, to await
enactment of additional authority in order to audit this joint assistance.6
In light of these statutory restrictions on GAO’s authority to audit the
Federal Reserve’s TALF activities, we continue to believe that Congress
should provide GAO with authority to audit the Federal Reserve’s
operation and administration of the TALF program.

5

123 Cong. Rec. H11015-20 (daily ed. Oct. 14, 1977).

6

See GAO, Troubled Asset Relief Program: Status of Government Assistance Provided to
AIG, GAO-09-975 (Washington, D.C.: Sept. 21, 2009), p. 2.

Page 90

GAO-10-25 Troubled Asset Relief Program

Appendix XI: GAO Contact and Staff
Acknowledgments

Appendix XI: GAO Contact and Staff
Acknowledgments
GAO Contact

Orice Williams Brown, (202) 512-8678 or williamso@gao.gov

Staff
Acknowledgments

In addition to the contacts named above, Karen Tremba (Assistant
Director), Angela Burriesci, Emily Chalmers, Rudy Chatlos, Joe Cisewski,
Rachel DeMarcus, Mike Hoffman, Robert Lee, Sarah McGrath, Marc
Molino, Tim Mooney, Omyra Ramsingh, Susan Sawtelle, and Cynthia
Taylor made important contributions to this report.

(250491)

Page 91

GAO-10-25 Troubled Asset Relief Program

GAO’s Mission

The Government Accountability Office, the audit, evaluation, and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO’s
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of
GAO Reports and
Testimony

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO’s Web site (www.gao.gov). Each weekday afternoon, GAO
posts on its Web site newly released reports, testimony, and
correspondence. To have GAO e-mail you a list of newly posted products,
go to www.gao.gov and select “E-mail Updates.”

Order by Phone

The price of each GAO publication reflects GAO’s actual cost of
production and distribution and depends on the number of pages in the
publication and whether the publication is printed in color or black and
white. Pricing and ordering information is posted on GAO’s Web site,
http://www.gao.gov/ordering.htm.
Place orders by calling (202) 512-6000, toll free (866) 801-7077, or
TDD (202) 512-2537.
Orders may be paid for using American Express, Discover Card,
MasterCard, Visa, check, or money order. Call for additional information.

To Report Fraud,
Waste, and Abuse in
Federal Programs

Contact:

Congressional
Relations

Ralph Dawn, Managing Director, dawnr@gao.gov, (202) 512-4400
U.S. Government Accountability Office, 441 G Street NW, Room 7125
Washington, DC 20548

Public Affairs

Chuck Young, Managing Director, youngc1@gao.gov, (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, DC 20548

Web site: www.gao.gov/fraudnet/fraudnet.htm
E-mail: fraudnet@gao.gov
Automated answering system: (800) 424-5454 or (202) 512-7470

Please Print on Recycled Paper