The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Examination Report for
Moody's Investor Services, Inc.
("Moody's")
1. Introduction
On August 31,2007, the Staffin the Commission's Office ofCompliance Inspections
and Examinations ("OCIE"), Division of Trading and Markets ("Trading & Markets")
and Office ofEconomic Analysis ("OEA'') (collectively "the Staff') initiated an
examination of Moody's, and two other credit rating agencies. The focus of the
examinations was Moody's activities in rating subprime residential mortgage-backed
securities ("RMBS") and collateralized debt obligations ("CDOs") linked to subprime
RMBS. 1 Specifically, key areas of review included:
"Y the NRSROs' ratings policies, procedures, and practices, including gaining an
understanding of ratings models, assumptions, criteria and protocols;
"Y the adequacy of the disclosure of the ratings process and methodologies used by
theNRSROs;
"Y whether the NRSROs complied with their ratings policies and procedures for
initial ratings and ongoing surveillance;
"Y the efficacy of the NRSROs' conflict of interest procedures; and
"Y whether ratings were unduly influenced by conflicts of interest related to the
NRSROs' role in bringing issues to market and the compensation they receive
from issuers and underwriters.
The examinations also included a review of whether there were any errors in ratings
issued as a result of flaws in ratings models used as a result of a press report indicating
errors in one firm's model? Initial observations as a result of this aspect of the
examinations are included in this report.
Begimring in 2007, delinquency and foreclosure rates for subprime mortgage loans in the United
States dramatically increased, creating turmoil in the markets for RMBS backed by such loans and
CDOs linked to such loans. As the performance of these securities continued to deteriorate, the
three NRSROs most active in rating these instruments downgraded a significant number of their
ratings. The NRSROs' performance in rating these structured finance products raised questions
about the accuracy of their credit ratings generally as well as the integrity of the ratings process as
a whole.
2
See Sam Jones, Gillian Tett, and Paul J. Davies, Moody's Error Gave Top Ratings to Debt
Products, Financial Times, May 20, 2008, at 1.
SEC_OCIE_FCIC_000496
The examination review period generally covered January 2004 through the present. The
firms under examination became subject to the provisions of the Credit Rating Agency
Reform Act of 2006 (the "Rating Agency Act"), which amended the Securities Exchange
Act of 1934 ("Exchange Act"), and the Commission's rules when they registered with the
Commission as NRSROs in September 2007. Although Moody's was not subject to legal
obligations applicable to NRSRO during most of the review period, the Staff nonetheless
sought to make relevant factual findings and observations with respect to the activities of
Moody's in rating subprime RMBS and CDOs during the period, as well as to identify
possible areas for improvement in their practices going forward.
Over 50 Commission Staff participated in the examinations ofMoody's, and two other
credit rating agencies. The examinations included extensive on-site interviews with the
rating agencies' staff, including senior and mid-level managers, initial ratings analysts
and surveillance analysts, internal compliance personnel and auditors, personnel
responsible for building, maintaining and upgrading the ratings models and methods used
in the ratings process, and other relevant rating agency staff
In addition, the Staff reviewed a large quantity of the rating agencies' internal records,
including the written policies, procedures and other such documents related to initial
ratings, the ongoing surveillance of ratings, and the management of conflicts of interest,
and the public disclosures of the procedures and methodologies for determining credit
ratings. The Staff also reviewed deal files for subprime RMBS and CDO ratings, internal
audit reports and records, and other internal records, including a large quantity of email
communications records (the rating agencies' produced over two million emails and
instant messages that were sorted, analyzed and reviewed using software filtering tools).
Finally, the Staff reviewed the rating agencies' public disclosures, filings with the
Commission and other public documents.
2. The Ratings Process
The Rating Agency Act expressly states that the Commission has no authority to regulate
the "the substance of the credit ratings or the procedures and methodologies" by which
any NRSRO determines credit ratings. 3 As part of this examination, however, the Staff
necessarily sought to develop an understanding of the quantitative analysis used to rate
the RMBS and CDOs that have been subject to such dramatic and widespread change.
Moody's rates RMBS and CDO transactions by first assessing the underlying collateral
and then assessing the deal structure. For RMB S collateral assessment, Moody's uses the
Moody's Mortgage Metrics Model ("M3") to quantitatively arrive at initial loss coverage
numbers. 4 Moody's then looks at qualitative factors, such as the originator and servicer,
15 U.S.C. 78o-7(c)(2).
4
Presently, Moody's evaluates over 50 different characteristics of each loan in a pool, through its
M3 model, examples of which are: credit bureau scores, which is an indicator of a borrower's past
credit performance; loan-to-value ratio, which reflects the amount of equity borrowers have in
2
SEC_OCIE_FCIC_000497
and makes adjustments to the initial loss coverage numbers to arrive at the final loss
coverage numbers. 5 For CDOs, Moody's uses its proprietary CDOROM model for
projecting expected loss. For cash flow CDOs, the results of the CDOROM model runs
are processed by Moodys' proprietary CDOEdge model in order to assess the portfolio. 6
In this second stage, Moody's arrives at an estimate of credit enhancement for excess
spread through models which project cash flows for the proposed capital structure.
Results from the collateral and cash flow models are reviewed through the rating
committee approval process before a final rating is issued.
M3 was not available to rate subprime securities until December 2006. Prior to that date,
Moody's used a system of"benchmarking" to rate subprime RMBS wherein a subprime
mortgage pool currently being evaluated was compared to several subprime pools
previously rated by Moody's? This process resulted in an initial rating number
after which several "hits" or adjustments could be applied, depending on the pools
characteristics, to arrive at the final loss coverage numbers. 8
As part of the ratings process during this period, Moody's will on occasion revise its
ratings methodology. Many of the changes Moody's made were incremental and did not
affect the overall rating. Moody's transitioned incremental changes over a short period of
time. If a deal was in-house and had been priced, the old methodology would apply, and
the deal would be rated under the former methodology. When a change to the
methodology would affect a rating, Moody's generally published a Request for Comment
notifying the market of the potential change and indicating that it would implement the
change at a later date.
3. Increase in Number and Complexity ofRMBS and CDO Deals
their homes; how fully buyers have documented their income and assets; whether the property will
be owner occupied; and, whether the loan is for purchase or refinance.
Moody's review of the originator, servicer, and master servicer is a significant element of
Moody's rating process as each can greatly influence loss levels depending on their relative
strength or weakness. The originator or servicer can increase or decrease loss levels separately by
up to a total of 20%.
6
For synthetic CDOs, Moody's analysts employ the CDOROM model, a simulation tool designed
to determine the expected loss for each tranche. The CDOROM model runs a minimum of one
million Monte Carlo simulations using the potential asset pools allowed under a CDO's covenants.
For cash CDOs, the CDOROM model is generally employed to generate expected loss figures for
potential collateral pools, which are then run through the firm's CDOEdge model, which applies a
correlated binomial method (incorporating default correlation assumptions) to a proposed deal
structure in order to generate projected payment waterfalls, in order to generate cash flow models
for various scenarios.
The subprime pools had either the same originator or a comparable originator.
For example, if a bucket in a pool had a high percentage of interest only loans a hit would be
applied to the expected loss number for that bucket.
3
SEC_OCIE_FCIC_000498
From 2002 to 2006, the volume of structured finance deals rated by Moody's increased
substantially, as did the revenues Moody's received from rating those deals. The
structured products Moody's was asked to evaluate became increasingly complex, with
many employing derivatives such as credit default swaps to replicate the performance of
mortgage backed securities. Further, the loans made to retail borrowers being securitized
evolved from 30-year fixed rate instruments to newer products such as second lien and
adjustable rate mortgages. The increasing number and complexity of deals may have
compromised various aspects ofMoody's ratings operations for structured finance, as
discussed in greater detail below.
a. Revenue, Deal, and Staffing Levels
From 2002 to 2006 the volume ofRMBS deals rated by Moody's increased by 137%, and
the number ofCDO deals rated by Moody's increased by 700%. Correspondingly, the
revenue Moody's derived from RMBS deals increased from $61.8 million in 2002 to
$168.9 million in 2006 and CDO revenue increased from $11.7 million in 2003 9 to $91.2
million in 2006.
For the RMBS group, contemporaneous staffing increases appear roughly in line with
volume increases (Moody's increased RMBS staffby 114% as volume increased by
137%). 1 For CDOs, however, Moody's staffing increases do not appear to have kept
pace with volume increases (Moody's increased CDO staffby 24% as volume increased
by 700%). 11
°
b. Impact on the Ratings Process
The Staff believes that the deal and staffing levels during the review period may have
impacted various aspects of the ratings process. For instance, several CDO memoranda
reviewed by the Staff indicate that ratings were issued notwithstanding one or more
unresolved issues. For example, the rating committee memorandum for the Costa Bella
CDO, Ltd. deal stated that for one issue involving the collateral manager, "We didn't ha
[sic] time to discuss this in detail at the committee, so they dropped the issue for this deal
due to timing. We will need to revisit in the future." 12 Another potentially unresolved
issue was described as "poorly addressed- needs to be checked in the next deal" 13 and
"WARR- don't ask ©)". 14
9
Moody's was unable to produce CDO revenue for 2002 due to a transition of accounting systems.
10
In 2002 RMBS lead analysts were responsible for monitoring their rated transactions. See MISOCIE-RMBS-28799. The tabulation of Moody's RMBS personnel for 2002 does not include the
Group Managing Director or Senior Managing Director and the tabulation for 2006 does not
include the Group Managing Director.
II
In 2002, CDO lead analysts were responsible for monitoring their rated transactions. Senior
Associates, of which there were nine, provided monitoring support.
12
See MIS-OCIE-RMBS-26801.
13
Id.
4
SEC_OCIE_FCIC_000499
The Staff believes that the increase in the number and complexity of deals may have
impacted Moody's subprime RMBS and CDO ratings operations, as is discussed in more
detail below.
The Staff recommends that Moody's periodically evaluate if it has sufficient staff and
resources to manage its volume of business and meet its obligations under the Rating
Agency Act.
Moody's Response: Moody's agrees with the Staffs recommendation and will
periodically evaluate staffing levels.
4. Disclosure of the Rating Process
The requirements of the Rating Agency Act specifically address the importance of
disclosure. An NRSRO is required to disclose publicly the procedures and
methodologies it uses in determining credit ratings. 15 Form NRSRO requires that this
disclosure be a general description; but sufficiently detailed to provide users of credit
ratings with an understanding of the processes employed in determining credit ratings,
including among other things, the quantitative and qualitative models and metrics used to
determine credit ratings. Moody's explained to the Staff that, prior to being registered as
an NRSRO, it disclosed its ratings process during the review period. It appears, however,
that certain significant aspects of the rating processes and the methodologies used to rate
RMBS and CDOs were not always disclosed, or were not fully disclosed, as summarized
below.
New or revised rating methodologies and policies that are deemed material are often first
approved internally and then published as a Request for Comment from the market before
being implemented. 16 In the RMBS group incremental changes to the ratings
methodology or process are approved through RMBS chair meetings. Moody's states
that it uses press releases and web postings to publicly disclose modifications to its rating
methodologies and related practices, procedures and processes. 17 However, Moody's
does not consolidate its methodologies for rating subprime RMBS or CDO transactions in
one location.
As such, the Staff had difficulty locating the disclosure of certain aspects of Moody's
ratings process. Moreover, Moody's does not publish (or publish before implementation)
all incremental changes to its methodology. For example, the Staff found emails where
14
"WARR" stands for weighted average rate of return. See MIS-OCIE-RMBS-26798.
15
Section 4(a)(l)(B)(ii) of the Exchange Act.
16
Moody's Report on the Code of Professional Conduct, April2006.
17
Moody's Code of Professional Conduct, June 2005.
5
SEC_OCIE_FCIC_000500
Moody's made changes to assumptions before the market was notified of the changes. 18
Additionally, the Staff found emails evidencing Moody's analysts utilizing an
unpublished model to assess data. 19
The Staff recommends that Moody's conduct a review of its current disclosures of its
processes and methodologies for rating RMBS and CDOs to assess whether it is fully
disclosing its ratings methodologies and meeting the requirements of the Rating Agency
Act and Form NRSRO. Further, the Staff recommends that Moody's review whether its
policies governing the timing of disclosure of a significant change to a process or
methodology are reasonably designed to comply with these requirements.
Moody's Response: Moody's generally agrees with the Staff's recommendations and is
currently taking steps to improve disclosure of the ratings process, including the drafting
ofunified methodologies.
5. Written Policies and Procedures for Rating RMBS and CDOs
As of September 2007, NRSROs are subject to a requirement to make and retain certain
internal documents relating to its business, including procedures and methodologies used
to determine credit ratings. 20 Prior to this, Moody's ratings policies are described in its
Code of Professional Conduct, Report on the Code of Professional Conduct, Analyst
Handbook- Rating Practices and Procedures, and Moody's Best Practices for the
Conduct of Moody's Structured Finance Rating Committees. While these documents,
taken as a whole, provide a general guideline for an analyst to follow when rating
structured finance products, they were not specific to any type of structured finance
product, such as RMBS or CDOs.
The Staff recommends that Moody's conduct a review to determine whether its written
policies and procedures used to determine credit ratings for RMBS and CDOs are fully
documented in accordance with the requirements of Commission Rule 17g-2.
18
Email chain ending with email from Ariel W eil, Associate Vice President/Analyst, Term RMB S,
Moody's, to Kelly Slicklein, Associate Director, Investor Services Group, Moody's (Nov. 29,
2007,20:08 GMT). See also email chain ending with email from Yuri Yoshizawa, Group
Managing Director, US Derivatives, Moody's, to Yvonne Fu, Team Managing Director, US
Derivatives, Moody's (Apr. 24, 2007, 18:50 GMT). See also email from Ariel Weil, Associate
Vice President/Analyst, Term RMB S, Moody's, to Mark DiRienz, Team Managing Director,
Term ABS, Moody's (Feb. 7, 2007, 20:54 GMT). See also email form Karen Ramallo, Associate
Analyst, Term RMB S, Moody's, to Odile Grisard Boucher, Associate Analyst, Term RMB S,
Moody's (Nov. 15, 2006, 19:10 GMT).
19
Email chain ending with email from Karen Ramallo-Rodriguez, Associate Analyst, Term RMBS,
Moody's, to Denise Person, Vice President/Senior Criteria Officer, Term RMBS, Moody's (Sept.
24, 2007, 18:26 GMT). Moody's states that it does not publish all criteria changes, particularly
those they consider incremental or non-material.
20
Rule 17g-2 under The Exchange Act. 17 CFR 240.17g-2.
6
SEC_OCIE_FCIC_000501
Moody's Response: Moody's generally agrees with the Staff's recommendation and
will conduct a review to ensure that its policies and procedures are properly documented
in accordance with the Commission's rules.
As a result of a May 20, 2008, Financial Times article detailing a coding error in the
model Moody's utilized to rate constant proportion debt obligations ("CPDOs") 21 the
Staff expanded the scope of its exam to review Moody's policies and procedures for
addressing the discovery of errors in its models and methodologies. 22 The Staff found
that while Moody's does have policies and procedures that emphasize the importance of
providing accurate ratings with integrity, it does not have policies and procedures that
provide guidance on the process that should be followed when errors are discovered in its
models, methodologies, or other aspects of the ratings process.
The Staff recommends that Moody's develop policies and procedures to address the
detection of errors with its models, methodologies, or other aspects of the ratings
process. The Staff also recommends that Moody's develop policies and procedures for
the reporting of discovered errors in its models, methodologies, or other aspects of the
ratings process.
Moody's Response: Moody's responded that it has instituted numerous remedial
measures to address this issue, including implementing new policies and procedures, and
establishing a taskforce to initiate a thorough review of all existing structured finance
models.
6. Integrity and Accuracy of the Information Provided to Moody's
There is no requirement under the federal securities laws that an NRSRO verify the
information contained in RMBS loan portfolios presented to it for rating. Additionally,
NRSROs are not required to insist that issuers perform due diligence, and they are not
required to obtain reports concerning the level of due diligence performed by issuers.
The Staff notes that pursuant to its policies, procedures, and public pronouncements,
Moody's did not engage in any due diligence or otherwise seek to verify the accuracy and
quality of the loan data underlying the RMBS pools it rated during the review period. In
fact, the Code of Ethics for Moody's clearly states that Moody's is under no obligation to
perform, and does not perform, due diligence? 3 Moreover, it states that the assignment
of a rating is not a guarantee of the accuracy, completeness, or timeliness of the
21
CPDOs are a type of credit derivative sold to investors looking for long term exposure to credit
risk on a highly rated note. Investors buy notes issued by a special purpose vehicle ("SPV').
22
The Staff also noted potential conflicts of interest with respect to this issue. These are addressed
more fully below.
23
See Moody's Code of Conduct, p 7, dated June 2005 and updated October 2007 (stating Moody's
has no obligation to perform, and does not perform, due diligence with respect to the accuracy o f
the information it receives or obtains in connection with the rating process).
7
SEC_OCIE_FCIC_000502
information relied on in connection with the rating? 4 Moody's solely performed loss and
cash flow analyses on the data presented to it. Moody's generally did not verify the
integrity and accuracy of such information as, in Moody's view, due diligence duties
belonged to the other parties in the process. Moody's also did not seek representations
from sponsors that due diligence was performed.
Moody's has taken, or announced, measures designed to improve the integrity and
accuracy of the loan data they receive on underlying RMBS pools:
"Y Moody's announced that it was considering enhancements to its RMBS
securitizations that would include the engagement by issuers of independent third
parties to randomly sample, for due diligence, the greater of 10% or 200 loans for
all subprime transactions.
"Y In addition, in an agreement with the New York State Attorney General, Moody's
agreed to develop and publicly disclose due diligence criteria to be performed by
underwriters on all mortgages comprising RMBS, and to review those results
.
.
.
.
25
pnor to 1ssumg ratmgs.
7. Documentation of Significant Steps and Participants in the Rating Process
a. Documentation of Significant Steps in the Ratings Process
An NRSRO is required to retain internal records, including non-public information and
workpapers, used to form the basis of a credit rating it issued (Exchange Act Rule 17g2(b)(2)). Prior to implementation of this requirement, Moody's policy was to retain
records related to the credit analysis and rating process for certain time periods as
identified in its record retention schedule? 6
The Staff reviewed 50 RMBS and 51 CDO deals to determine ifMoody's followed the
policies and procedures for rating RMBS and CDOs and its file maintenance and
recordkeeping policy. Moody's policies and procedures call for a rating committee
memorandum and addendum to be produced for each transaction rated by Moody's. The
information required in the rating committee memorandum and addendum includes the
date the rating committee convened; the names of the rating committee attendees; the
name of the committee chair; the type of rating action and type of instrument under
consideration; the rating recommendation and rationale; the rating committee outcome
and vote tally; quantitative analysis, and supporting materials. 27
24
25
http://www.oag.state.ny.us/press/2008/june/june5a 08.html
26
Moody's record retention policy did not apply to procedural or methodological policies governing
the credit ratings process as a whole. MIS-OCIE-RMBS-28424-28483.
27
See Moody's Best Practices for the Conduct of Moody's Structured Finance Rating Committees,
at 5 (May 2006). See MIS-OCIE-RMBS-312, 397,430. In May of2006 Moody's added the
8
SEC_OCIE_FCIC_000503
For RMBS transactions, most of the ratings memoranda and addenda contained the
minimum required information under Moody's policy. However, the Staff found that the
level of detail provided as to how the committee arrived at its rating levels was
inconsistent. 28 Often boxes on the addendum were checked without explanation despite
fields requiring explanation, contained an inadequate explanation, or in some cases were
not checked at all? 9
The level of documentation of Moody's CDO ratings process varied widely across the
deals reviewed by the Staff Only 15 included all of the four basic required pieces of
documentation: a rating committee memorandum, a rating committee addendum, a
monitoring committee memorandum, and a monitoring committee addendum. For 16 of
the deals reviewed, no monitoring documentation was produced. 30 Additionally, the
rating committee memoranda featured significant variation in the topics covered and
amount of information presented. For instance, over half of the rating committee
memoranda included a section on the collateral manager, however, the level of detail
provided in this section varied greatly. 31 Almost all of the rating committee memoranda
included a modeling assumptions section, however, the level of detail provided varied
greatly. While a small number of memoranda had detailed discussions of the collateral
manager, the majority of memoranda including this section either contained a very brief
summary or essentially included the section only as a placeholder (i.e., the section
included the name of the collateral manager followed by blank fields for items such as
"location," "contact," "key personnel," etc.). The majority of the ratings committee
memoranda also included an "Issues" section. The level of detail in this section also
varied greatly, ranging from detailed listings of rating committee concerns and
issuer/underwriter responses to short lists without any indication of resolutions. To the
extent that they were provided to the Staff, the surveillance and monitoring memoranda,
requirement that the committee chair be identified in all memoranda as well as the committee
outcome. Compare MIS-OCIE-RMBS-312, with MIS-OCIE-RMBS-397, and 430.
28
For instance, one initial rating committee memorandum contained only three sentences that merely
state a base description of the loan pool and there is no way to discern how the rating committee
arrived at its results, GSAMP 2006-Sl, MIS-OCIE-RMBS-28730; others did not discuss the rating
rationale in sufficient detail. See~ Long Beach Mortgage Loan Trust 2006-2, MIS-OCIERMBS-28621-28625.
29
For example, the "Key variable(s) voted upon" sections of the addenda were often left blank or
contained a generic term like "ratings."
30
These numbers are based on the Staffs review of the documentation provided by Moody's.
Approximately three months after the delivery of the majority of the requested transaction
materials, Moody's provided an index that confirmed the Staff's findings as to the extent of
documentation. See MIS-OCIE-RMBS-32168-32171.
31
While a small number of memoranda had detailed discussions of the collateral manager, the
majority of memoranda including this section either contained a very brief summary or essentially
included the section only as a placeholder (i.e., the section included the name of the collateral
manager followed by blank fields for items such as "location," "contact," "key personnel," etc.).
9
SEC_OCIE_FCIC_000504
like the rating committee memoranda, also varied in the amount of information provided,
with sections of some memoranda essentially serving as placeholders. Finally, many of
the rating committee or monitoring memoranda did not contain a narrative discussing the
ratings committee decision or underlying rationale. 32
The Staff's findings with respect to deal files it reviewed are corroborated by other
internal discussion by Moody's. For example, the Derivatives Group was aware that
"delinquencies on adherence to the document retention policy had increased" and sent out
an email to the Derivatives Group noting this issue. 33 Moody's, however, has
subsequently stated that it is in the process of implementing automated committee
memorandum and other document retention procedures which will address these issues.
Ultimately, the Staff found that Moody's failed to retain or document certain significant
steps in the rating process, which made it difficult for the Staff to assess compliance with
its rating policies and procedures, and to identify the factors that were considered in
developing a particular rating. This lack of documentation would similarly make it
difficult for the Moody's internal compliance staff or internal audit staff to assess
compliance with the firm's policies and procedures.
b. Documentation of Participants in the Ratings Process
An NRSRO is also required to make and retain records of the identity of any credit
analyst that participated in determining the rating and any person that approved the rating
before it was issued (Exchange Act Rule 17g-2). This requirement is intended to assist
the Commission in monitoring whether the NRSRO is following its procedures and
methodologies for determining credit ratings and whether the NRSRO is complying with
procedures designed to prevent the misuse of material nonpublic information by
identifying the persons with the best information as to how the credit rating was
determined. Prior to this, Moody's policy required that the rating committee attendees
and rating rationale be a part of the rating committee memorandum. 34
For the subprime RMBS and CDO transactions reviewed by the Staff, the Staff found
that, at times, in both the initial ratings memoranda and addenda the vote tally was
incomplete with either a generic "agreed with levels" type comment in the field for "Key
variable(s) voted on" accompanying the vote tally or no indication of the vote tally at
32
In many cases, the sole documentation of the decision-making process is in the addenda section
labeled "RC Outcome," with checkboxes for "RC Recommendation accepted" and "RC Outcome
differed from Recommendation." In several cases where the latter box was checked, no
explanatory narrative was provided.
33
Email from Gus Harris, Senior Managing Director, New Products Group, Moody's, to
'SFG/Derivatives- US' listserv, Moody's (May 18, 2007, 22:16 GMT).
34
See Moody's Best Practices for the Conduct of Moody's Structured Finance Rating Committees,
at 5 (May 2006). See MIS-OCIE-RMBS-312, 397, 430.
10
SEC_OCIE_FCIC_000505
all? 5 The Staff also found that like the initial rating committee memoranda, the majority
of surveillance memoranda failed to record the voting results.
The Staff recommends that Moody's conduct a review of its current policies and practices
for documenting the credit rating process to review whether they are reasonably
designed to ensure compliance with Rule 17g-2 and to address weaknesses in the policies
or in adherence to existing policies that result in gaps in recording the voting in the
credit rating process.
Moody's Response: Moody's generally agrees with the Staff's recommendation and
will continue to monitor to ensure compliance with its recordkeeping requirements.
8. Surveillance Practices
Under the Rating Agency Act, an NRSRO is required to disclose publicly the procedures
and methodologies it uses in determining credit ratings. In addition, Section 4( d) of the
Rating Agency Act states that an NRSRO must maintain adequate financial and
managerial resources to produce credit ratings with integrity.
Moody's does not have written policies and procedures for the surveillance of subprime
RMBS and CDO bonds, although it publishes criteria that describe the methodologies
under which such bonds are monitored. 36 For both RMBS and CDO Moody's uses
automated surveillance tools that on a monthly basis flag for review securities whose
performance indicates that their current credit rating may not be consistent with the
current estimated expected loss? 7 Aside from its monthly outlier screening, Moody's
also regularly performs ratings sweeps by issuer and/or origination year, where Moody's
looks at each outstanding deal individually.
Once a rated instrument is selected based on the automated surveillance tools, a Moody's
surveillance analyst will further investigate the status of the transaction and present
findings to a ratings committee. If the rating committee believes that a rating may need
to be adjusted, then the securities are placed on review for a potential downgrade or
upgrade. 38
It appears that Moody's regularly performed RMBS and CDO surveillance during the
exam time period. However, while Moody's publishes criteria that describe its
35
See~
36
From 2003-2007, Moody's released three comprehensive publications that detail how Moody's
monitors RMBS transactions.
37
For RMBS the surveillance process is based on a review of collateral performance, for CDO it is
based on the ratings of the individual assets comprising the collateral pool.
38
For CDO Moody's follows a very similar process; however, surveillance analyst analyze different
metrics.
Long Beach Mortgage Loan Trust 2006-2, MIS-OCIE-RMBS-28621-28631.
11
SEC_OCIE_FCIC_000506
methodology for the surveillance ofRMBS and CDO bonds, Moody's does not have
internal written procedures documenting the steps staff should undertake for surveillance
ofRMBS and CDO bonds.
The Staff recommends that Moody's develop RMBS and CDO surveillance policies and
procedures. The Staff also recommends that Moody's conduct a review to determine if
adequate resources are devoted to surveillance of outstanding RMBS and CDO ratings.
Moody's Response: Moody's agrees with the Staff's recommendations and informed
the Staff that it is implementing, or has implemented, global procedural and structural
changes that address the Staff's recommendations. Among these policies and changes is
the building out of its compliance function to facilitate surveillance policy development.
9. Management of Conflicts of Interest
a. "Issuer Pay Model" /Fee Discussions
Moody's uses the "issuer pays" model, in which the sponsor or other entity that issues the
security is also seeking the rating. Under Exchange Act Rule 17g-5(b)(l), it is a conflict
of interest for an NRSRO being paid by issuers or underwriters to determine credit
ratings with respect to securities they issue or underwrite. Section 15E(h) of the
Exchange Act requires an NRSRO to establish, maintain, and enforce policies and
procedures reasonably designed to address and manage conflicts of interest. Such
policies and procedures are intended to maintain the integrity of the NRSRO' s judgment.
Avoiding a conflict of interest prevents an NRSRO from being influenced to issue a more
favorable credit rating in order to obtain or retain business of the issuer or underwriter? 9
In order to manage this conflict of interest, in October of 2007 Moody's established a
policy to restrict analysts and their immediate managers from participating in fee
discussion with issuers. 40 Moody's has also organized its rating group as a separate
organization within a larger company. 41 However, Moody's does not actively monitor
employees' compliance with the prohibition against analysts from participating in fee
discussions.
The Staff found multiple communications that indicate that analysts are aware of the
firm's fee schedules, and actual (negotiated) fees. There does not appear to be any
internal effort to shield analysts from emails and other communications that discuss fees
and revenue from individual issuers. In some instances, analysts discuss fees for a rating.
For instance, one analyst wrote to his manager "in the past it took us 2- 3 months to rate
one [a type of deal], so I assume fees should be much higher than for typical
39
See Release No. 34-55857 and Exchange Act Rule 17g-5.
40
Moody's Code of Professional Conduct (October 2007).
41
In 2007, Moody's Corporation effected a separation at the corporate level between its credit rating
business, Moody's, and its non-ratings product and service business, Moody's Analytics.
12
SEC_OCIE_FCIC_000507
reperforming deal." 42 Another analyst wrote to his manager asking about whether the
firm would be charging a fee for a particular service, and what the fee schedule would
be. 43 In addition, there were indications that analysts were involved in fee discussions
with employees of the firm's billing department. 44 The Staff is concerned that analysts
could be influenced in their ratings by their awareness of the amount of fees charged to
Issuers.
b. Business considerations in the Ratings Process
As a result of a May 20, 2008, Financial Times article detailing a coding error in the
model Moody's utilized to rate CPDOs, Moody's began a review by outside counsel
surrounding the issue. As a result of that investigation, Moody's reported to the Staff that
a European CPDO rating/surveillance committee had knowledge that Moody's had issued
ratings on certain CPDO securities in 2006 using a model that contained a coding error. 45
The coding error resulted in most of those securities receiving a rating several notches
higher than if the model had not contained the coding error. In January of 2007, a CPDO
committee first became aware that the ratings were several notches higher than they
should have been. Despite this fact, the committee agreed to continue to maintain the
higher unwarranted ratings for several months until the securities were eventually
downgraded for performance reasons. Members of the committee, all analysts or
analytical managers, considered the rating agency's reputation when deciding not to
downgrade the securities and make the coding error public.
Moody's recently informed the Staff that as a result of these findings, it has implemented,
or plans to implement several global procedural and structural changes that address the
issues identified. The Staff is still reviewing the facts related to the CPDO ratings.
The Staff recommends that Moody's continue to review its practices, policies and
procedures to further mitigate and manage the "issuer pays" conflict of interest. In
particular, the Staff recommended that Moody's consider steps that would insulate or
prevent the possibility that considerations of market share and other business interests
could influence ratings or ratings criteria.
42
Email from Gulmira Karaguisiyeva, Analyst, Term RMBS, Moody's, to David Teicher, Team
Managing Director, TermRMBS, Moody's (May 9, 2007, 13:46 GMT).
43
Email from Zhiqin (James) Huang, Analyst, Term RMBS, Moody's, to Mark DiRienz, Team
Managing Director, Term ABS, Moody's (May 7, 2006, 13:38 GMT).
44
Email from Gulmira Karaguisiyeva, Analyst, Term RMBS, Moody's, to Joy Mayo, Manager,
Middle Office, Moody's (Aug. 23, 2007, 23:10 GMT).
45
The Staff met with Moody's and outside counsel conducting the investigation on June 27, 2008, to
discuss the CPDO coding error and the progress of the investigation. Moody's represented that
the investigation into this matter will be completed in mid-July 2008.
13
SEC_OCIE_FCIC_000508
Moody's Response: Moody's agrees with the Staff's recommendation and will continue
to review its practices and procedures to further mitigate and manage the "issuer pays"
conflict of interest.
c. Analyst Compensation
Moody's has a policy that generally provides that an analyst may not be compensated or
evaluated based upon the amount of revenue that the rating agency derives from issuers
or issues that the analyst rates or with which the analyst regularly interacts. 46 While
Moody's does not compensate its analysts based on the deals they rate or the ratings
provided, like all employees, the amount of an analyst's bonus is tied to the overall
success of the company.
d. Securities Transactions by Employees
Moody's has adopted a policy to prohibit employees and their immediate family
members from owning any security that is subject to a credit rating of a team on which
such employees are members to guard against potential insider trading. 47 Furthermore,
Moody's has implemented procedures to monitor employees' ownership of securities of
issuers or obligors rated by groups within the company with whom an employee is not
affiliated. 48 Managers are required to review all trades ofMoody's employees that report
directly to them and raise all potential conflicts of interest or violations with the Legal
department or Ratings Compliance department. 49 The Staff found Moody's employee
securities transaction program to be adequate.
The Staff has recommended that Moody's conduct a review of its policies and procedures
for managing the securities ownership conflict of interest to determine whether they are
reasonably designed to ensure that its employees' personal trading is appropriate and
does not violate Rule 17g-5.
Moody's Response: Moody's agrees with the Staff's recommendation and will continue
to review its policies and procedures for securities trading and ownership to ensure
compliance with Rule 17g-5.
10. Internal Audit
46
See Section 2.11 of the Moody's Code of Professional Conduct.
47
See Moody's Revised Securities Trading Policy and Reporting Procedures (November
2005).
48
Moody's employees must report all the securities holdings of the employee and/or members of
his/her immediate family, upon initiation of employment and periodically thereafter, as well as
records of securities transactions that the employee and immediate family members engage in. An
employee's compliance with the transaction reporting requirement can be achieved by ensuring
that his/her employer receives duplicate copes of brokerage statements and trade confirmations.
49
Section 3 of Appendix A to the Revised Securities Trading Policy and Reporting Procedures
(November 2005).
14
SEC_OCIE_FCIC_000509
Historically, Moody's employed an outside firm, KPMG, to perform its audits. In
addition, at least since 2006, Moody's has performed internal audits to evaluate ratings
group's compliance with its best practices, its electronic storage requirements, securities
trading restrictions, and the Moody's Code of Conduct. 50 The auditors evaluate the
adequacy of implementation of internal controls designed to address these areas. The
auditors perform a risk assessment to determine where to perform their audits based upon
a number of factors, including the type of debt and geographic location of the group
within Moody's responsible for rating an issue. For example, RMBS securities in Europe
may be audited in one year, in the U.S. in the next, and in Asia in the year after that,
meaning that a specific group in a particular geographic area is audited once every three
years.
During the period reviewed by the Staff, Moody's conducted three internal audits related
to the RMBS and CDO rating process. 5 1 The internal audit reviews discovered, among
other things, certain non-compliance with document retention policies, over-reliance
upon individuals for technological expertise to test software, and lack of adherence to the
rating committee guidelines (inquiries about committee member conflicts, no
documentation of majority vote, lack of rating committee memo, missing rating letter,
empty electronic document management system ["EDMS"] folder, two models not filed
on EDMS or CDO Edge, and a lack of documented re-prioritization of deal monitoring).
In the Staff's opinion, the most significant finding arising out of the internal audit
performed in 2006 for the U.S. Derivatives Team in Structured Finance revealed that
derivative models are not formally reviewed and/or validated by management before they
are posted for general use. For one deal, the analyst used a banker's proprietary model,
without any form of review of the model to determine whether it was reliable or
consistent with Moody's methodology. The auditors recommended that management
implement a review process to periodically assess the integrity of the models used in
support of the rating. Moody's was unable to demonstrate evidence of its management's
follow-up on the recommendations of the auditors. The Staff believes that Moody's
should be able to provide records of such follow-up as part of an examination of the
internal audit record.
The Staffrecommends that Moody's review whether its internal audit functions,
particularly in the RMBS and CDO ratings areas, are adequate, and whether it provides
for proper management follow-up.
Moody's Response: Moody's agrees with the Staff's recommendation and will review
the adequacy of its internal audit functions and will develop procedures that address
management follow-up.
11. Conclusion
50
Moody's provided no records of any audits performed prior to 2006.
51
See 2006 U.S. Derivatives Monitoring Internal Audits; 2006 U.S. RMBS Internal Audit; and 2006
U.S. Derivatives Team Structured Finance Internal Audit.
15
SEC_OCIE_FCIC_000510
The Staff intends to send a deficiency letter to Moody's outlining its findings and
recommendations. The Staff will request that Moody's provide a written response within
30 days outlining any remedial action planned or already taken to address the findings
and recommendations in the letter. Moody's will be asked to include in its response a
timetable for implementing the proposed remedial action. The letter will also request that
Moody's send OCIE a written confirmation in 12 months detailing the status of
implementation of each remedial action.
16
SEC_OCIE_FCIC_000511
@FedFRASER