Enhancing the Accountability of Credit
Rating Agencies: The Case for a
Disclosure-Based Approach
Stphane Rousseau*
Credit rating agencies (CRAs) play a vital role in enabling
financial markets to operate efficiently by acting as informational
intermediaries specializing in the appraisal of the creditworthiness of
corporations that issue debt. Despite their importance, however, rating
agencies remain unregulated private institutions. The recent wave of
corporate scandals has led many to call their contribution to market
efficiency into question. In light of such criticism, studies conducted by
lawmakers and regulators sought to further examine the role and
effectiveness of CRAs. Although the studies revealed no particular
wrongdoing, they warned of potential problems that could disrupt the
smooth operation of capital markets. These problems relate to the reliability
and integrity of ratings, as well as to possible anti-competitive practices on
the part of CRAs. These potential problems are worrisome given that
CRAs wield considerable power over issuers and investors. Fundamentally,
the main theme underlying the criticism of CRAs relates to their
accountability towards market participants. In a perfectly functioning
market, the fact that CRAs have such significant power would not elicit
such concern since they would be accountable to both issuers and investors.
The real world departs from this ideal and market failures may lead to a
divergence between, on the one hand, the interests of CRAs, and, on the
other hand, those of issuers and investors.
A review of the legal and institutional environment indicates that
there is a dearth of mechanisms designed to offset these market failures.
Reputation is the primary mechanism that acts to restrain opportunistic
behaviour on the part of rating agencies. Thus, a potential accountability
gap exists, leading to an imbalance between CRAs power, and the
likelihood of holding them responsible for their use of this power.
It is in this context that regulators have examined possible methods
of enhancing the accountability of rating agencies. In light of the recent
flurry of regulatory initiatives, the purpose of this study is to discuss the
attitude that Canadian regulators should adopt in approaching CRA
accountability. The study favours the approach put forward by the
International Organization of Securities Commissions. It proposes
implementing this approach through a disclosure strategy that would
contribute to enhancing the accountability of CRAs, not only by reinforcing
existing reputational pressures that guard against opportunism, but also by
introducing an additional level of regulatory supervision over CRAs that
could hold them responsible for such behaviour.
Les agences de notation de crdit (ANC) jouent un rle crucial
pour lefficience des marchs en agissant titre dintermdiaires
informationnels se spcialisant dans lvaluation de la solvabilit des
entreprises se finanant par voie dendettement. Malgr leur importance,
les agences de notation ne sont cependant pas rglementes. Depuis la
rcente vague de scandales financiers, plusieurs remettent en cause leur
contribution lefficience des marchs. la lumire de ces critiques, les
lgislateurs et les rgulateurs ont men des tudes examinant plus en dtails
le rle et lefficacit des ANC. Bien quelles naient rvl aucun abus, les
tudes mettent en garde contre des problmes qui pourraient nuire au bon
fonctionnement du march. Ces problmes concernent la fiabilit et
lintgrit des notations, de mme que les pratiques des agences qui
pourraient savrer anti-concurrentielles. Ces problmes sont proccupants
tant donn que les ANC exercent une influence considrable sur les
metteurs et les investisseurs. Essentiellement, le thme rcurrent de ces
critiques concerne limputabilit des ANC. Dans un march parfaitement
efficient, le pouvoir des ANC ne soulverait pas dinquitude puisquelles
seraient imputables lgard des metteurs et des investisseurs. Dans la
ralit, des imperfections du march peuvent mener une divergence entre,
dune part, les intrts des ANC et, dautre part, ceux des metteurs et des
investisseurs.
Un examen de lenvironnement lgal et institutionnel rvle
labsence de mcanisme pour corriger les imperfections du march. La
rputation est
le principal mcanisme disciplinaire pour endiguer
lopportunisme des ANC. Ainsi, il existe une lacune pouvant mener un
dsquilibre entre le pouvoir des ANC et la possibilit de les rendre
responsable pour leurs dcisions.
Dans ce contexte, les rgulateurs ont examin des moyens
daccrotre limputabilit des ANC. Lobjectif de la prsente tude est de
discuter de lapproche pouvant tre employe par les rgulateurs au regard
des initiatives de rformes. Ltude favorise lapproche propose par
lOrganisation internationale des commissions de valeurs (OICV). Elle
propose de mettre en uvre la proposition de lOICV par une stratgie
fonde sur la divulgation. Cette stratgie permet daccrotre limputabilit
des ANC non seulement en renforant les sanctions rputationnelles, mais
aussi en ajoutant une supervision rglementaire pouvant les rendre
responsables pour leurs conduites opportunistes.
* Associate Professor and Chair in Business Law and International Trade, Faculty of Law,
Universit de Montral. This paper is derived from a report commissioned by the Capital Markets
Institute of the University of Toronto. The author wishes to thank the personnel of the Capital Markets
Institute for their time, input, and encouragement in preparing this study, including Paul Halpern,
Doug Harris, and Atanaska Novakova. He is also grateful to Paul Bourque, Jerry Goldstein and Robert
Nicholls who made insightful comments on an earlier draft of this work. Finally, the author wants to
thank ric Archambault and Philippe Kattan for their capable research and editorial assistance. The
research for this paper is current as of 1 January 2006.
Stphane Rousseau 2006
To be cited as: (2006) 51 McGill L.J. 617
Mode de rfrence : (2006) 51 R.D. McGill 617
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
618
Introduction
I. Background on the Role and Regulation of Credit Rating
Agencies
A. Credit Rating Agencies and the Operation of Capital Markets
B. An Overview of the Rating Agency Industry
C. Current Regulation of Ratings and Credit Rating Agencies
II. Concerns Over Credit Rating Agencies Accountability
A. Potential Failures in the Credit Rating Market
1. Sources of Market Failures
Imperfect Competition
a.
b. Agency Problems Affecting Credit Rating Agencies
2. A Look at the Empirical Evidence Surrounding Market
Failures
a. The Value of Credit Ratings
i. The Reliability of Ratings
ii. The Timeliness of Rating Changes
iii. The Relevance of Ratings
b. The Potential Conflicts of Interest of Credit
Rating Agencies
c. Abusive Practices of Credit Rating Agencies
i. The Practice of Notching
ii. Unsolicited Ratings
[Vol. 51
620
621
622
624
626
627
627
627
627
629
631
631
631
632
633
634
635
635
636
637
637
640
641
642
642
643
644
644
644
645
646
647
B. Legal and Institutional Constraints Influencing the Behaviour
of Credit Rating Agencies
1. The Role of Reputation in Shaping the Behaviour of Rating
Agencies
Issuer Disclosure
2.
3. The Influence of Securities Regulation on the Transparency
and Dissemination of Ratings
4. The Liability of Credit Rating Agencies
C. Summation: Is There an Accountability Gap?
III. Enhancing the Accountability of Credit Rating Agencies:
The Role of Regulation
A. The Regulation of Credit Rating Agencies: Policy Perspectives
1. The Goals of Securities Regulation
a. Market Efficiency
b.
Investor Protection
2. The Importance of a Cost-Effective Regulatory Regime
B. An Overview of the Current Proposals to Enhance the
Accountability of Credit Rating Agencies
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 619
1. The IOSCO Code of Conduct Fundamentals for Credit
Rating Agencies
a. The Rating Process
b. Credit Rating Agencies Independence and Avoidance
i. The Quality of the Rating Process
ii. Monitoring and Updating
iii. The Integrity of the Rating Process
of Conflicts of Interest
i. General Principles
ii. Procedures and Policies
iii. Analyst and Employee Independence
647
647
647
648
648
649
649
649
650
Conclusion
c. Credit Rating Agencies Responsibilities to the Investing
650
Public and to Issuers
i. Transparency and Timeliness of Ratings Disclosure 650
ii. Treatment of Confidential Information
651
651
d. Disclosure of the Code of Conduct
2. The Securities and Exchange Commissions Proposal to
Define the Concept of Nationally Recognized Statistical
Rating Organization
a.
Issuance of Publicly Available Ratings that Are Current
Assessments of the Creditworthiness of Obligors with
Respect to Specific Securities
b. Recognition of Credibility and Reliability by the Financial
c. Systematic Procedures to Ensure Credible and
Markets
Reliable Ratings
C. A Critical Look at Regulatory Strategies for Implementing
Accountability-Enhancing Rules of Conduct
1. Self-Regulation with a Voluntary Code of Conduct
2. Strategies Regulating both Entry and Conduct
a. Supervised Self-Regulation
b. Registration Model
D. Strengthening Market-based Solutions through a Measured
Regulatory Intervention
1. A Disclosure Strategy to Reinforce Private Accountability
Mechanisms
2. The Implementation of the Disclosure Strategy
651
652
653
653
653
654
656
657
658
660
660
661
664
620
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 51
Introduction
Since their inception at the beginning of the twentieth century, credit rating
agencies (CRAs) have emerged as informational intermediaries specializing in the
appraisal of the creditworthiness of corporations.1 Gradually, these institutions have
become central to the financial markets infrastructure through their role in rectifying
information asymmetries that exist between issuers and investors. At the same time,
CRAs have gained considerable clout over market participants as their assessments of
creditworthiness have come to be viewed as authoritative.2 Despite their importance,
rating agencies remain unregulated private institutions.
Whereas rating agencies have always been subject to periodic criticism, the
recent wave of corporate scandals has led many to call their contribution to market
efficiency into question.3 Commentators have criticized CRAs for failing to play their
role of watchdog. Others have questioned their reliability in general. In light of such
criticism, studies conducted by lawmakers and regulators sought to further examine
the role and effectiveness of CRAs.4 Although the studies revealed no particular
wrongdoing, they warned of potential problems that could disrupt the smooth
operation of capital markets. These studies formed the basis on which the
International Organization of Securities Commissions (IOSCO), the Securities and
1 See generally Timothy J. Sinclair, The New Masters of Capital (Ithaca: Cornell University Press,
2005) at 22-30; Richard Cantor & Frank Packer, The Credit Rating Industry (1995) 5:3 J. Fixed
Income 10.
2 Steven L. Schwarcz, Private Ordering of Public Markets: The Rating Agency Paradox (2002) U.
Ill. L. Rev. 1 at 2.
3 For early criticism, see Francis A. Bottini, An Examination of the Current Status of Rating
Agencies and Proposals for Limited Oversight of Such Agencies (1993) 30 San Diego L. Rev. 579;
Frank Partnoy, The Siskel and Ebert of Financial Markets?: Two Thumbs Down for the Credit Rating
Agencies (1999) 77 Wash. U.L.Q. 619 [Partnoy, Siskel and Ebert].
4 See e.g. International Organization Of Securities Commission, Technical Committee, Report of the
Activities of Credit Rating Agencies (September 2003) [IOSCO Report], online: International
Organization of Securities Commissions
Agencies in the Operation of the Securities Markets (January 2003) [Report on the Role and Function
of CRAs], online: United States Securities and Exchange Commission
Governmental Affairs, Financial Oversight of Enron: The SEC and Private-Sector Watchdogs (8
October 2002) [Watchdogs Report], online: United States Senate
Before the U.S. Senate Committee on Governmental Affairs, 107th Cong. (Washington, D.C.: United
States Government Printing Office, 2002) [Rating the Raters], online: United States Senate
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 621
Exchange Commission (SEC), and the European Commission founded their recent
policy initiatives.5
The main theme underlying such criticism relates to CRAs accountability
towards market participants, by questioning whether the power allotted to rating
agencies is balanced by equally effective mechanisms designed to ensure that they act
in the interests of market participants. In light of the recent flurry of regulatory
initiatives, the purpose of this study is to discuss the attitude that Canadian regulators
should adopt regarding credit rating agencies. First, we offer some background
information on the role and regulation of credit rating agencies. We then move to an
analysis of the concerns raised by various CRA activities. We also examine the legal
and institutional mechanisms designed to direct the behaviour of rating agencies as
well as their effectiveness with respect to minimizing the impact of market failures.
Finally, we identify the initiatives recently put forward by the SEC and the IOSCO.
We argue that while both of these initiatives deal with similar issues, the IOSCO
Code can be more readily transplanted in national regulations than the SEC rule,
which is idiosyncratic to the American system. We then assess the options available to
implement that IOSCO Code. We conclude by recommending that securities
regulators implement the IOSCO Code in Canada using a disclosure-based approach.
I. Background on the Role and Regulation of Credit Rating
Agencies
Since the publication of Moodys Analyses of Railroad Investments in 1909, credit
rating agencies have become central institutions in financial markets.6 They emerged
to rectify some of the information asymmetries that exist in lending relationships.
Throughout the last century, CRAs have adapted to ever evolving financial markets.
Initially focused on railroads, industrial corporations, and financial institutions, CRAs
now rate every type of issuer, both national and international. Ratings cover
traditional fixed-income securities such as bonds, as well as new structured
financial instruments such as asset-backed securities. Increasingly, the opinions of
CRAs carry more importance for market participants. Regulation frequently makes
reference to ratings, giving them a normative dimension in certain instances. Aside
from that normative aspect, market participants rely on ratings not necessarily
5 International Organization of Securities Commissions, Code of Conduct Fundamentals for Credit
Rating Agencies (December 2004), online: International Organization of Securities Commissions
Nationally Recognized Statistical Organization, 78 Fed. Reg. No. 70 (2005) [NRSRO Definition],
online: United States Securities and Exchange Commission
European Commission on Possible Measures Concerning Credit Rating Agencies (CESR/05-1396)
(March 2005) [CESRs Technical Advice], online: Euractive.com
6 See Sinclair, supra note 1 at 22-27.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
622
because the agencies are right, but because they are thought to be an authoritative
source of judgments.7
[Vol. 51
A. Credit Rating Agencies and the Operation of Capital Markets
Credit rating agencies provide an evaluation of the creditworthiness of issuers,
which is essentially an assessment of how likely they are to make timely payments on
their debts in general.8 They also offer ratings of individual debt instruments that
indicate the probability of default or delayed payment with respect to that particular
security. The ratings do not express opinions on whether the particular debt
instruments should be bought or sold. They are only intended to convey information
regarding the relative safety of the securities. Since their primary function is to
evaluate credit risk, CRAs do not assess the economic appeal of investments.9
Individual investors may prefer to purchase less creditworthy instruments as they
receive appropriate compensation for the added risk acceptable. Furthermore, a credit
rating does not express the agencys opinion of the actual value of an issuers equity
securities.
The activities of CRAs can contribute to the efficiency of capital markets by
rectifying some of the information asymmetries that exist between issuers and
investors.10 Inevitably, information asymmetry exists in the debt market because
issuers have superior information regarding their creditworthiness than do investors.11
Consequently, this discrepancy enables issuers to exaggerate their credit quality in
order to get the highest price for their securities, leaving to potential investors the task
of distinguishing between good and bad issues.
In the absence of CRAs, investors could theoretically attempt to conduct their
own research.12 In practice, it is doubtful that such an initiative would prove cost-
effective for most investors.13 Because of the greater size of their investments and
expertise, some institutional investors may find it economically justifiable to do their
own research of the credit quality of issuers. However, their research remains costly
7 Ibid. at 2.
8 IOSCO Report, supra note 4 at 3; Watchdogs Report, supra note 4 at 98.
9 Amy K. Rhodes, The Role of the SEC in the Regulation of the Rating Agencies: Well-Placed
Reliance or Free-Market Interference? (1996) 20 Seton Hall Legis. J. 293 at 315-316; Schwarcz,
supra note 2 at 6.
10 IOSCO Report, supra note 4 at 3; Gregory Husisian, What Standard of Care Should Govern the
Worlds Shortest Editorials?: An Analysis of Bond Rating Agency Liability (1990) 75 Cornell L.
Rev. 411 at 413; Lawrence J. White, The Credit Rating Industry: An Industrial Organization
Analysis in Richard M. Levich, Giovanni Majnoni & Carmen M. Reinhart, eds., Ratings, Rating
Agencies and the Global Financial System (Boston: Kluwer, 2002) at 43.
11 Arturo Estrella et al., Credit Ratings and Complimentary Sources of Credit Quality Information
(No. 3 – August 2000) [unpublished, archived at Basel Committee on Banking Supervision Working
Papers] at 11, online: Bank for International Settlements
12 Husisian, supra note 10 at 415-419; Rhodes, supra note 9 at 294-295; White, supra note 10, p. 43.
13 Estrella, supra note 11 at 11.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 623
to society because it most likely duplicates the fundamental questions pertaining to
creditworthiness.14 Furthermore, the willingness of institutional investors to assess the
creditworthiness of issuers could be hindered by a public good problem.
The information asymmetry between issuers and investors is troublesome. If left
unchecked, it can lead to an adverse selection problem in that the debt of issuers with
good credit quality will be undervalued, thereby undermining the viability of the
market.15 In such an environment, issuers have an incentive to disclose their credit
quality to investors in order to receive the highest possible price for their issues.
Signaling theory suggests that issuers that have good credit quality can communicate
this information to investors and receive higher market valuation, through actions that
issuers of lower credit quality find too costly to reproduce.16 One possibility would be
for issuers with high credit quality to underprice their issues in order to differentiate
themselves from low credit quality issuers.17 However, this option would not be
optimal, as it would raise the cost of capital for issuers.18 A more effective method
would be for issuers to use outside specialists acting as information intermediaries to
highlight the superior credit quality of their debt issues.19
CRAs acquire and process information with the purpose of ascertaining credit
quality. They research and review information from a variety of sources. Through
their activities, they can make credit assessments at a lower cost than individual
investors, since analysts who have greater expertise in credit rating undertake the
research and review. Indeed, their expertise enables them to gather and analyze
information more cost-effectively. The economies of scale associated with research
and analysis further reduces information costs.20 Moreover, CRAs can disseminate
information rapidly to the market, thereby improving the timeliness of adjustments in
prices. Finally, CRAs eliminate the redundant and therefore wasteful efforts of
investors individually engaging in research activities.
More importantly, they act as certifying agents by offering their reputation to
supplement that of the issuer as a guarantee of quality.21 For prospective investors to
14 See Rhodes, supra note 9 at 295.
15 See generally George A. Akerlof, The Market for Lemons: Quality Uncertainty and the Market
Mechanism (1970) 84 Quarterly Journal of Economics 488.
16 See generally Hayne E. Leland & David H. Pyle, Informational Asymmetries, Financial
Structure, and Financial Intermediation (1977) 32 Journal of Finance 371.
17 IOSCO Report, supra note 4 at 3; See generally Franklin Allen & Gerald R. Faulhaber,
Signaling by Underpricing in the IPO Market (1989) 23 Journal of Financial Economics 303.
18 See W.K.H. Fung & Andrew Rudd, Pricing New Corporate Bond Issues: An Analysis of Issue
Cost and Seasoning Effects 41 Journal of Finance 633 at 642 (study finds no clear evidence of
underpricing).
19 See Ronald J. Gilson & Reinier H. Kraakman, The Mechanisms of Market Efficiency (1984)
70 Va. L. Rev. 549 at 604. See generally Stephen J. Choi, Market Lessons for Gatekeepers (1998)
92 Nw. U.L. Rev. 916 [Choi, Gatekeepers].
20 See Gilson & Kraakman, ibid. at 601.
21 Partnoy, Siskel and Ebert, supra note 3 at 632.
[Vol. 51
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
624
be convinced of the accuracy of certification, the signal conveyed by CRAs must be
credible itself.22 This requires that three conditions be met.23 First, the certifying agent
must have reputational capital at stake, which would be adversely and materially
affected by incorrectly certifying as accurately priced an issue that was actually
overvalued. Second, the value of the agents reputational capital must be greater than
the gains to be made from false certification. Third, it must be costly for issuers to
purchase the services of the certifying agents, and this cost must be an increasing
function of the scope and potential importance of the information asymmetry …24
Actually, CRAs probably meet these three criteria.25 Rating agencies have
reputational capital at stake when they issue ratings. They would likely suffer a
greater loss from falsely certifying the quality of an issue than they would gain in
fees. Finally, the production of ratings is costly.
B. An Overview of the Rating Agency Industry
CRAs are pervasive institutions.26 The Basel Committee on Banking Supervision
estimated that there were over 130 agencies worldwide, with about thirty of them
playing a prominent role in G10 countries.27 Rating agencies may operate at a
national, regional, or even global scale. Some provide ratings, solicited or unsolicited,
on a limited number of issuers while others have the capability of rating all issuers in
a given marketplace using statistical models. Ratings can focus on specific fixed-
income securities, including complex financial instruments issued in structured
finance, as well as on issuers, such as corporations, municipalities, and governments.
Aside from providing ratings, CRAs also offer ancillary services.28 These services
include rating assessment services, whereby they provide an evaluation of the impact
of contemplated corporate action on an issuers rating. Other services include risk
management and consulting services designed to assist financial institutions and other
corporations in their management of credit and operational risk.
Three of the largest CRAs operating on a global scale are based in the United
States. They are Moodys, Standard & Poors (S&P), and Fitch. These three U.S.
22 See Gilson & Kraakman, supra note 19 at 605; Choi, Gatekeepers, supra note 19 at 934ff;
L. Paul Hsueh & David S. Kidwell, Bond Ratings: Are Two Better Than One? (1988) 17:1
Financial Management 46 at 47.
23 Gilson & Kraakman, ibid. at 613-21. In the context of investment bankers and venture capitalists,
see Randolph P. Beatty & Jay R. Ritter, Investment Banking, Reputation, and the Underpricing of
Initial Public Offerings (1986) 15 Journal of Financial Economics 213 at 217; William L. Megginson
& Kathleen A. Weiss, Venture Capitalist Certification in Initial Public Offerings (1991) 46 Journal
of Finance 879 at 881.
24 Megginson & Weiss, ibid.
25 See Husisian, supra note 10 at 426; Rhodes, supra note 9 at 295-96; IOSCO Report, supra note 4
at 3. For a critical view, see Partnoy, Siskel and Ebert, supra note 3 at 703.
26 Sinclair, supra note 1 at 22-30.
27 White, supra note 10 at 44, n. 18.
28 IOSCO Report, supra note 4 at 4; Report on the Role and Function of CRAs, supra note 4 at 42.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 625
rating agencies, which also operate in Canada, are well known, and their activities
have been amply chronicled.29 In Canada, there are two major Canadian CRAs:
Dominion Bond Rating Services (DBRS) and Canadian Bond Rating Services
(CBRS), which has recently been purchased by Standard & Poors.
Traditionally, CRAs earned their revenues from subscriber fees paid by
investors.30 In the early 1970s, CRAs changed their business model and started
charging issuers for their rating services.31 Nowadays, the larger CRAs derive most of
their revenues from the fees charged to issuers.32
All of the most important CRAs organize the rating process around similar
procedures and mechanisms.33 From a structural perspective, CRAs establish a rating
committee to initiate, withdraw or review a rating. A rating committee is generally
formed ad hoc and is composed of a lead analyst, managing directors, and junior
analytical staff.
The rating process begins with a request by an issuer or its underwriter prior to
the offering. The CRA then assigns a lead analyst to that issuer who conducts a
preliminary analysis to prepare the rating. The analyst gathers information from issuer
and non-issuer sources in order to gain a better understanding of the firm and its
environment. Meetings are also held with senior management or government
officials. The analyst then submits a report to the rating committee, which proposes a
recommendation on the creditworthiness of the issuer or the securities. After
discussion, the rating committee issues the credit rating.
Prior to announcing the rating, the CRA notifies the issuer allowing the latter to
review the press release for factual verification and to ensure that no confidential
information is disclosed. Where it disagrees with the rating, the issuer may appeal the
decision by providing new and important information or by pointing out the ratings
reliance on incorrect information or dubious sources.
Lastly, the CRA issues a press release that contains the rating as well as the
rationale justifying it. Subsequent to the issue, the rating agency monitors the issuer
and its securities by reviewing corporate filings, monitoring industry trends, and
maintaining contact with corporate management. When necessary, the CRA will put
the issuer on a watch list to indicate that it is considering reviewing the rating
issued.
29 See Sinclair, supra note 1 at 27-30; Roy C. Smith & Ingo Walter, Rating Agencies: Is There an
Agency Issue? in Richard M. Levich, Giovanni Majnoni & Carmen M. Reinhart, supra note 10 at
293-305.
30 Rhodes, supra note 9 at 308-309; Claire A. Hill, Regulating the Rating Agencies (2004) 82
Wash. U.L.Q. 43 at 50 [Hill, Regulating]; White, supra note 10 at 47.
31 White, ibid.
32 See Estrella, supra note 11 at 25; Smith & Walter, supra note 29 at 292.
33 The following discussion is based on Rhodes, supra note 9 at 309-316; IOSCO Report, supra note
4 at 5; Sinclair, supra note 1 at 30-42; Report on the Role and Function of CRAs, supra note 4 at 25-
27.
626
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 51
C. Current Regulation of Ratings and Credit Rating Agencies
In Canada, many federal and provincial regulatory schemes refer to ratings issued
by CRAs. In a nutshell, the regulatory regimes rely on such ratings to distinguish
investment-grade from speculative securities. For instance, the distinction between
investment-grade and speculative securities serves in prudential regulation of the
banking and investment dealing industries.34 It is also used to identify securities in
which certain types of institutional investors can invest without prior authorization. In
securities regulation, issuers of investment-grade securities benefit from particular
exemptions designed to reduce the regulatory burden to reflect the lower level of risk
of their securities.35
Despite the rather broad use of ratings, there is no principled approach with
respect to CRAs. In fact, the organization and activities of CRAs are not regulated per
se. Whereas regulatory regimes only recognize ratings issued by approved or
recognized rating agencies, these expressions are only defined through a
rudimentary listing of large CRAs.36
In the United States, regulatory schemes use ratings for similar purposes.37 Since
1975, such regulations increasingly require that ratings be issued from a nationally
recognized statistical rating organization (NRSRO) designated by the SEC.38 Thus,
rating agencies that do not have NRSRO status are barred from a significant segment
of the market.39 Despite its importance, the term NRSRO has not been officially
defined, nor have criteria for NRSRO designation been formally adopted.40 Through
the no-action letter process, the SEC staff developed a number of criteria that it
considers pertinent to NRSRO designation.41 Among those criteria, the most
important is that the applicant must be nationally recognized in the United States as
an issuer of credible and reliable ratings by the predominant users of securities
ratings.42 According to experts, the weight attributed to this factor creates a Catch-22
34 See e.g. Money Market Mutual Fund Conditions Regulations, S.O.R./2001-475, ss. 1.1-1.2;
Investment Dealers Association, Rule Book, ss. 100.4E.f-100.4E.g, online: Investment Dealers
Association
03, ss. 3.1, 4.5, 6, online: Canadian Legal Information Institute
35 See Short Form Prospectus Distributions, O.S.C., NI 44-101, 28 O.S.C. Bull. 10385 (23
December 2005), online: Ontario Securities Commission
36 Regulatory regimes typically refer to Canadian Bond Rating Services, Dominion Bond Rating
Services, Moodys, Standard & Poors, Fitch, Duff & Phelps, and Thomson BankWatch.
37 See Report on the Role and Function of CRAs, supra note 4 at 6-8.
38 Cantor & Packer, supra note 1 at 18-19.
39 See White, supra note 10 at 46; Frank Partnoy, The Paradox of Credit Ratings in Richard M.
Levich, Giovanni Majnoni & Carmen M. Reinhart, supra note 10 at 72-78 [Partnoy, Paradox].
40 Hill, Regulating, supra note 30 at 55 [footnotes omitted].
41 See Report on the Role and Function of CRAs, supra note 4 at 9.
42 Ibid.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 627
problem: an agency has to be nationally recognized to be an NRSRO but has to be
an NRSRO to become nationally recognized.43 This problem is exacerbated by the
relative lack of formality and transparency of the recognition process.44 In sum, the
current framework clearly favours existing rating agencies that are already recognized
as NRSROs.45
II. Concerns Over Credit Rating Agencies Accountability
Credit rating agencies provide intermediation services to investors and issuers.
As intermediaries, CRAs wield influential power over both issuers and investors.46
Through their activities they influence the conditions under which issuers will have
access to debt markets, the conditions of their relationships with lenders, and the
structure of their transactions. They can also affect investors portfolio decisions. In a
perfectly functioning market, the fact that CRAs have such significant power would
not be a concern since they would be accountable to both issuers and investors. The
real world departs from this ideal in that market failures may lead to a divergence
between, on the one hand, the interests of CRAs, and on the other hand, the interests
of issuers and investors.
A. Potential Failures in the Credit Rating Market
1. Sources of Market Failures
a.
Imperfect Competition
The credit rating industry is highly concentrated. At the international level, the
IOSCO reports that Moodys, S&P, and Fitch dominate the credit rating business.47
These three agencies are also the dominant players in the U.S.48 In Canada, the rating
industry is also concentrated. The major rating agencies are DBRS and CBRS.
American CRAs also provide ratings for Canadian issuers, especially when their
securities are sold in the U.S.
43 Hill, Regulating, supra note 30 at 55.
44 See Rhodes, supra note 9 at 298-99 and 323-29.
45 Cantor & Packer, supra note 1 at 18; Partnoy, Paradox, supra note 39 at 74.
46 See Dieter Kerwer, Holding Global Regulators Accountable: The Case of Credit Rating
Agencies (Working Paper 11, School of Public Policy, University College London, December 2004)
at 15; T.J. Sinclair, supra note 1 at 22-30.
47 IOSCO Report, supra note 4 at 8; Smith & Walter, supra note 29 at 294-97.
48 White, supra note 10 at 45; Hill, Regulating, supra note 30 at 44. DBRS has been recognized as
an NRSRO in the U.S. in 2003.
628
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 51
Although it is beyond the scope of this study to offer a detailed assessment of the
level of competition in the credit rating industry, it is worth emphasizing the two
principal barriers to entry that contribute to market concentration.
The first barrier stems from the existing regulatory framework. In the U.S., it is
linked to the regulatory use of the NRSRO concept.49 As mentioned above, issuers
have an added incentive to obtain ratings from agencies that have NRSRO status. An
indirect implication is that new entrants are prevented from acting as raters for a
significant number of issuers. This, in turn, bars emerging agencies from attaining a
level of business whose scope will be sufficiently broad to warrant NRSRO
recognition.
An NRSRO designation may also indirectly affect the Canadian credit rating
industry because of the sheer importance of the U.S. market. Canadian corporations
rely significantly on U.S. bond markets to obtain debt financing.50 Given the
importance of the U.S. market, issuers will prefer rating agencies that have NRSRO
status in order for their debt-instruments to be eligible for purchase by institutional
investors in the U.S.51 In other words, a rating agency that does not have NRSRO
status will be disadvantaged in the Canadian market as well.
Canadian regulations impose barriers to entry of their own, by referring to the
ratings issued by large CRAs that have the status of recognized rating organization.
To develop their business, new agencies need to be recognized by securities
commissions. Recognition is made problematic by the complexity of the criteria and
processes on which commissions base their decision. Furthermore, once recognition
is granted, regulatory instruments need to be amended so as to include the new CRAs
in their list of recognized rating organizations.
The second barrier to entry stems from the market itself, and is based on the
economies of scale and scope, as well as on standardizations that are present in the
rating industry.52 The importance of this market-related barrier to entry would explain
why there were very few rating agencies in the U.S. even before the introduction of
NRSRO regulation in 1975.53 Likewise, the importance of this second barrier to entry
could explain the scarcity of CRAs in Canada where the regulatory framework is not
as restrictive as in the U.S.
49 See Partnoy, Paradox, supra note 39 at 74; Report on the Role and Function of CRAs, supra
note 4 at 36-40; White, ibid. at 46.
50 Walter Engert & Charles Freedman, Financial Developments in Canada: Past Trends and Future
Challenges (2003) Summer, Bank Can. Rev. 3 at 9, online: Bank of Canada
51 See Rhodes, supra note 9 at 341-42; Sinclair, supra note 1 at 44-45; The Multijurisdictional
Disclosure System, O.S.C., NI 71-101 (6 november 1998), online: Ontario Securities Commission
52 White, supra note 10 at 46. See also IOSCO Report, supra note 4 at 9.
53 White, ibid.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 629
The high concentration characterizing the credit rating industry gives rise to two
potential problems. First, the dominant CRAs may be tempted to engage in anti-
competitive behaviour to restrain entry into the market and maintain their position.
Second, they may forgo the quality of the services they provide.54 The existence and
importance of both of these problems is discussed in more detail further below.
b. Agency Problems Affecting Credit Rating Agencies
Credit rating agencies act as conduits between the issuers they rate and
investors.55 Intermediation leads CRAs to act on behalf of both issuers and investors.
From an economic perspective, the relationship that exists between a rating agency
and issuers or investors can thus be qualified as being one of agency.56 The interaction
between agents and their principals gives rise to potential agency problems.57
In the case of CRAs, a first agency problem relates to the investors-CRAs axis.58
Under the current business model, CRAs are paid by issuers to provide ratings. Thus,
the dominant CRAs receive most of their revenue from the issuers that they rate. The
practice of issuers paying for their own ratings creates a potential conflict of interests
for CRAs.59 Under such circumstances, agencies may be tempted to downplay the
credit risk of issuers and to inflate their ratings in order to retain their business. The
practice of charging fees based on the size of offerings also renders CRAs more
vulnerable to pressure by larger issuers.
The second problem relates to the issuers-CRAs axis. The development of
consulting services by CRAs creates another source of potential conflicts of interests.
The rating decisions may be influenced by whether or not an issuer purchases
additional services offered by a CRA.60 Moreover, issuers may feel the need to
subscribe to such services simply out of fear that their failure to do so could
adversely impact their credit rating (or, conversely, with the expectation that
purchasing these services could help their credit rating).61
54 Robert Baldwin & Martin Cave, Understanding Regulation: Theory, Strategy, and Practice
(Oxford: Oxford University Press, 1999) at 251; Partnoy, Siskel and Ebert, supra note 3 at 686; Jean
Tirole, The Theory of Industrial Organization (Cambridge, M.A.: MIT Press, 1988) at 105-115.
55 See generally Daniel F. Spulber, Market Microstructure and Intermediation (1996) 10:3 Journal
of Economic Perspectives 135 at 147-48.
56 Smith & Walter, supra note 29 at 290. See also Jill E. Fisch & Hillary A. Sale, The Securities
Analyst as Agent: Rethinking the Regulation of Analysts (2003) 88 Iowa L. Rev. 1035 at 1080.
57 See generally Michael C. Jensen & William H. Meckling, Theory of the Firm: Managerial
Behavior, Agency Costs, and Ownership Structure (1976) 3 Journal of Financial Economics 305.
58 See Sinclair, supra note 1 at 150-51.
59 IOSCO Report, supra note 4 at 10-11; Report on the Role and Function of CRAs, supra note 4 at
60 Hill, Regulating, supra note 30 at 51.
61 Report on the Role and Function of CRAs, supra note 4 at 43.
41.
630
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 51
One final agency problem exists that may effect both issuers and investors, in
that although the larger CRAs are compensated primarily by issuers, they continue to
offer subscriptions to their information services. Subscribers have access to
substantial information on ratings, as well as direct access to analysts. Although they
do not gain access to information about ratings or rating rationales before it is made
available to the investing public, subscribers may have preferential access to material
information about issuers and credit ratings.62 Preferential subscriber access to
information is made even more problematic by the special treatment CRAs enjoy
under both Regulation FD in the U.S. and Rule 51-201 in Canada. Communications
between issuers and CRAs are exempt from insider trading restrictions. Likewise,
these communications are exempt from selective disclosure prohibitions.63 Because of
these exemptions, CRAs are permitted access to non-public information to conduct
their analysis.
Some contend that allowing issuers to convey non-public information to CRAs
contributes to the informational efficiency of the market.64 Ratings can be viewed as a
mechanism through which issuers communicate inside information to bondholders
without disclosing its substance. For instance, issuers can use ratings to transmit
confidential information, which if disclosed, could compromise the firms competitive
advantage and reduce their market value.65
Still, critics stress that CRAs access to non-public information creates three
specific problems.66 First, CRAs may provide their subscribers with non-public
material information, which threatens to destabilize the level playing field upon
which investors should trade. Second, rating decisions may be accompanied by
greater market volatility as investors speculate as to whether or not non-public
information influenced the analysis. Third and finally, rating agency analysts may
trade on inside information that is the property of issuers.
These three agency problems are exacerbated by the complexity of the CRAs
methodologies and processes.67 During hearings held by the SEC, commentators have
expressed an interest for more detailed information regarding assumptions underlying
the ratings, the ratings criteria, the lists of credit ratings under review, as well as the
information and documents on which rating decisions rely. According to these
observers, a more transparent approach would reduce the uncertainty and market
62 Ibid. at 35.
63 See Disclosure Standards, O.S.C., NP 51-201, 25 O.S.C. Bull. 4492 (12 July 2002), ss. 3.3(2),
3.3(7), online: Ontario Securities Commission
64 Louis H. Ederington, Jess B. Yawitz & Brian E. Roberts, The Informational Content of Bond
Ratings (1987) 10 Journal of Financial Research 211 at 212.
65 See Edmund W. Kitch, The Theory and Practice of Securities Disclosure (1995) 61 Brook. L.
66 IOSCO Report, supra note 4 at 11-14; Report on the Role and Function of CRAs, supra note 4,
67 Report on the Role and Function of CRAs, ibid. at 35-36.
Rev. 763 at 848-55.
pp. 35-36.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 631
volatility that accompany rating changes. Investors would speculate less with respect
to rating changes if they better understood what prompted them.
2. A Look at the Empirical Evidence Surrounding Market Failures
a. The Value of Credit Ratings
In general, ratings provided by CRAs elicit four types of criticism: (1) the
reliability of ratings; (2) the timeliness of rating changes; (3) the CRAs limited
success in properly assessing and accounting for bond covenants in their ratings; and,
(4) finally and more fundamentally, the relevance of ratings.
i. The Reliability of Ratings
Ratings are an evaluation of the creditworthiness of issuers. Following the recent
U.S. scandals, commentators have criticized both the performance of CRAs and the
reliability of their ratings.68 Specifically, they questioned whether rating agencies
analysts conduct sufficiently thorough analyses of the various issuers whose debt they
rate. They also raised concerns regarding the training and qualifications of these
analysts. Others have cast doubt on whether CRAs sufficiently monitor and review
issuers financials, specifying that they too often take the word of issuers officials
rather than seek answers themselves through further probing.
While such criticism is based primarily on anecdotal evidence, a glance at the
existing empirical evidence yields a more nuanced assessment of credit ratings
reliability. A good starting point is to examine the credit ratings own track record.
Ratings are probabilistic statements of the likelihood of issuer default. Therefore, a
basic approach to evaluate the reliability of ratings is to compare them with actual
default statistics. Several studies undertaking such a comparison have found a high
correlation between credit quality as determined by the rating and default rates.69
Another factor that is deemed to be relevant to assessing the success of ratings is
their durability. Some contend that to the extent that they have strong predictive
value, ratings should be more stable and change less frequently. In this respect, a
study conducted by Standard & Poors has shown that ratings were rather stable for
68 See Report on the Role and Function of CRAs, ibid. at 31-32; Watchdogs Report, supra note 4 at
115-25; Claire A. Hill, Rating Agencies Behaving Badly: The Case of Enron (2003) 35 Conn. L.
Rev. 1145; Sinclair, supra note 1 at 149-73.
69 See Charles Adams, Donald J. Mathieson & Garry Schinasi, The International Monetary Fund,
International Capital Markets: Developments, Prospects, and Key Policy Issues (International
Monetary Fund, September 1999) at 137, online:
Cantor & Packer, supra note 1 at 19-22; Louis H. Ederington & Jess B. Yawitz, The Bond Rating
Process in Edward I. Altman, ed., Handbook of Financial Markets and Institutions, 6th ed. (New
York: John Wiley & Sons, 1987) at 23-16 to 23-17.
[Vol. 51
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
632
investment grade issues.70 The significance of this criterion should not be overstated.
On the theoretical side, it is not clear why durability is a positive attribute of ratings.
The timeliness of rating changes, appears more important. On the empirical side, the
results of the S&P study may not be surprising given that investment grade issues
concern securities of corporations with strong fundamentals.
A final measure of reliability concerns absolute risk. As Cantor and Packer note,
ratings ought to provide a reliable guide to absolute credit risk.71 In this respect,
studies indicate that ratings have been less reliable as indicators of credit risk. It
appears that default probabilities associated with their specific letter ratings have
drifted over time.72
ii. The Timeliness of Rating Changes
CRAs maintain surveillance of the issuer or its securities following the rating
through contact with management and access to publicly disclosed information. The
ratings are updated periodically, or on the receipt of material information. The
updated analyses are conducted less thoroughly than at issuance.
Rating agencies have been severely criticized for their performance in the
continual monitoring of assigned ratings. Many have commented on the lethargy of
CRAs when it came to changing their ratings, particularly downgrading.73 Since the
events that spurred such criticism have been amply chronicled, two anecdotal
examples will suffice. CRAs maintained Enrons credit rating at above investment
grade as late as November 28, 2001, only a few days before it filed for bankruptcy.74
More recently, agencies were criticized for keeping the ratings for General Motors
and Ford just above investment grade at a time where the market traded the bonds of
those corporations at spreads equivalent to junk status.75
The anecdotal evidence concerning rating changes questions the contribution of
CRAs to the informational efficiency of the market. This suggests that CRAs do not
invest sufficient effort in monitoring ratings once they are assigned. Thus, the
agencies lag behind the market when it comes to reviewing the creditworthiness of
issuers. When a rating change occurs, the market has already accounted for the
information underlying the change.
70 Leo Brand & Reza Bahar, Corporate Defaults : Will Things Get Worse Before They Get Better?
Standard and Poors CreditWeek (31 January 2001) 15, online: Standard & Poors Japan
note 2 at 13-14; Adams, Mathieson & Schinasi, ibid. at 139.
71 Cantor & Packer, supra note 1 at 19.
72 Ibid.
73 For an overview of the salient cases where CRAs did not get it right, see Sinclair, supra note 1
74 Watchdogs Report, supra note 4 at 108-113. Enron filed for bankruptcy on 2 December 2001.
75 Who rates the raters? The Economist (26 March 2005) 67 at 69.
at 156-72.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 633
Although the anecdotal evidence reveals potentially troubling shortcomings, it
should not be taken as conclusive in light of empirical studies. Early studies have
found that changes in ratings lagged the market by an average of several months.
More recent studies show that downgrades do in fact provide new information to
market participants, since negative returns are observed in equity markets in response
to them.76 Likewise, rating downgrades of asset-backed securities are accompanied by
negative returns and wider spreads.77 These results indicate that rating changes,
particularly downgrades, do not systematically lag the market. This implies that the
monitoring of ratings by CRAs is not inherently defective and can provide new
information to investors.
iii. The Relevance of Ratings
According to Professor Partnoy, ratings have virtually no informational content.
This assertion is justified by the fact that CRAs do not have the expertise nor the
resources to generate information of any real value to the market.78 His opinion is
corroborated by the existence of empirical studies and anecdotal evidence calling the
accuracy of ratings into question.79
For Partnoy, ratings serve primarily to enable favourable regulatory treatment for
issuers. According to this regulatory license view, credit ratings are valuable
because they determine the substantive effect of legal rules: If the applicable
regulation imposes costs, and a favorable rating eliminates or reduces those costs,
then rating agencies will sell regulatory licenses to enable issuers and investors to
reduce their costs.80 In the U.S., it is the use of NRSRO ratings in various regulations
that has led to the emergence of this licensing role for rating agencies.81
While it is true that the U.S. regulations have given substantial power to CRAs,
Partnoy arguably overstates his case. There are several recent empirical studies
indicating that ratings provide new information. In fact, the evidence is serious
enough for academics to conclude that a consensus appears to have been reached
that ratings do convey important information to the market … 82
76 Louis H. Ederington & Jeremy Goh, Bond Rating Agencies and Stock Analysts: Who Knows
What When? (1996) 33 Journal of Financial and Quantitative Analysis 569.
77 John M. Ammer & Nathanael Clinton, Good News Is No News? The Impact of Credit Rating
Changes on the Pricing of Asset-Backed Securities Federal Reserve Board: International Finance
Discussion Paper No. 809 (July 2004), online: The Federal Reserve Board
78 Partnoy, Siskel and Ebert, supra note 3 at 651-53.
79 Ibid. at 647, 658-659.
80 Ibid. at 682.
81 Ibid.
82 Jeff Jewell & Miles Livingston, A Comparison of Bond Ratings from Moodys S&P and Fitch
IBCA (1999) 8:4 Financial Markets, Institutions & Instruments 1 at 3. See also Soku Byoun & Yoon
[Vol. 51
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
634
More specifically, Professor Hill has convincingly argued that the two rating
norms that exist in the U.S., and pursuant to which issuers purchase ratings from
Moodys and S&P, cannot be reconciled easily with the regulatory license view.83 If
issuers are merely buying a license when they solicit a rating, why pay more by
purchasing two ratings where the regulation only requires one? Furthermore, why
does the market react differently to debt instruments carrying two ratings as opposed
to only one? As Hill emphasizes, if issuers were mainly buying the regulatory
treatment … they would not pay much more than was necessary to obtain that
treatment.84
Thus, it is doubtful that ratings convey absolutely no pertinent information to the
market. Rating agencies may have some advantages over investors that enable them
to produce valuable information about the fundamentals of issuers. CRAs may also
have more time than investors to analyze information about the creditworthiness of
issuers. They have also developed specialization in ranking issuers by relative credit
quality. Moreover, CRAs may have greater access to information than investors, as
issuers may be more inclined to disclose confidential or inside information to them.
Besides, ratings may be valuable for investors in that they reflect the issuers own
perception of the creditworthiness of the debt in question.85 Finally, ratings in
themselves may constitute valuable information in that they tend to be self-fulfilling
prophecies.86 A rating conveys information about the entitys borrowing costs as well
as the marketability of the debt that is issued.
b. The Potential Conflicts of Interest of Credit Rating Agencies
Although concerns are frequently expressed concerning the potential conflicts of
interest of CRAs, there appears to be little evidence demonstrating the existence of
abuses.87 In the inquiries following Enrons demise, there was no allegation that
Moodys decision not to downgrade Enron had been based on improper influence.
Following public hearings, the SEC noted that most hearing participants agreed that,
for the most part, the rating agencies had effectively managed their potential
conflicts of interest.88
S. Shin, Unsolicited Credit Ratings: Theory and Empirical Analysis (October 2002), online: Social
Science Research Network
83 Hill, Regulating, supra note 30 at 66-67.
84 Ibid. at 66.
85 See Hseuh & Kidwell, supra note 22. See also H. Kent Baker & Sattar A. Mansi, Assessing
Credit Rating Agencies by Bond Issuers and Institutional Investors (2002) 29 Journal of Business
Finance & Accounting 1367 at 1376-77.
86 Ibid. at 74. See also Rating the Raters, supra note 4 at 44 (Jonathan Macey).
87 White, supra note 10 at 50.
88 Report on the Role and Function of CRAs, supra note 4 at 23.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 635
At the empirical level, a recent study showed that the conflicts of interests of
CRAs do not influence their actions significantly.89 The study examined one
mechanism in particular through which CRAs may act in the interest of issuers, and
that is in the context of a rating downgrade. Delaying the negative news conveyed in
downgrades can benefit issuers in several ways. If CRAs are conflicted they will
attempt to delay downgrades where such changes involve issuers that are important
clients. Likewise, conflicted agencies will delay costly downgrades such as those that
create fallen angels. The study demonstrated that the market anticipated about
seventy-five per cent of the downgrades. However, it found no evidence consistent
with rating agencies acting in the interests of issuers due to a conflict of interest.90 On
the whole, it does not appear that the potential conflicts of interest to which CRAs are
susceptible have materialized into any real misbehaviour on their part.
c. Abusive Practices of Credit Rating Agencies
i. The Practice of Notching
Notching occurs where a rating agency bases its rating on ratings already
assigned by other agencies.91 CRAs often resort to notching when rating collateral
debt obligations (CDO). They use this practice to assess those underlying securities
that they themselves did not rate. Putting methodological questions aside, notching is
not a practice that is problematic per se.
In the U.S., several market participants argue that Moodys and S&P can employ
notching to prevent it from rating certain structured finance markets.92 According to
Fitch, these other agencies frequently engage in lowering their ratings on, or
refusing to rate, securities issued by certain asset pools (e.g., collaterized debt
obligations), unless a substantial portion of the assets within those pools were also
rated by them.93 In other words, Moodys and S&P used notching as an
89 Daniel M. Covitz & Paul Harrison, Testing Conflicts of Interest at Bond Rating Agencies with
Market Anticipation: Evidence that Reputation Incentives Dominate (December 2003) [unpublished,
archived at the Federal Reserve Board, Washington, D.C.], online: Social Science Research Network
90 See also Alexander W. Butler & Kimberly J. Rodgers, Relationship Rating : How Do Bond
Rating Agencies Process Information? European Finance Association 2003 Annual Conference Paper
No. 491 (27 June 2003), online: Social Science Research Network
91 See Mark Adelson, NERA Study of Structured Finance RatingsMarket Implications Nomura
Fixed Income Research (6 November 2003) at 1, online: Securitization.net
92 Report on the Role and Function of CRAs, supra note 4 at 24.
93 Ibid. [footnotes omitted].
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
636
anticompetitive strategy.94 Although these are serious allegations, a recent study
conducted by the National Economic Research Associates on the practice of notching
proved disappointingly inconclusive, with evidence pointing in both directions with
respect to the abusive character of notching.95
[Vol. 51
In Canada, DBRS acknowledges that when rating an issuer, it may rely on the
ratings assigned by other agencies.96 It can adjust such ratings to reflect what it
considers to be the credit quality of the issuer. Such adjustment can occur where there
is a significant variance between the other major rating agencies. DBRS emphasizes
that it does not believe that automatically notching the ratings from other agencies
serves any intrinsic analytical purposes.97 Nevertheless, notching has yet to elicit any
specific criticism from Canadian market and industry participants.
ii. Unsolicited Ratings
CRAs can issue unsolicited ratings based primarily on public information.98
Unsolicited ratings are not uncommon but are controversial.99 Critics stress that CRAs
use them as a means of increasing their market share:100 By giving borrowers a low,
unsolicited rating, the big agencies may force unwilling issuers to pay for their
services in the hope of getting a better one.101 They point to the case of Moodys who
frequently assigned unsolicited ratings of bonds and debt-instruments that were
substantially lower than the ratings issued by other agencies.102 In addition,
94 Greenberg Quinlan Rosner Research Inc., Most Structured Finance Senior Executives Oppose
Notching (26 March 2002) at 8, online: Greenberg Quinlan Rosner Research Inc.
95 Andrew Carron et al., National Economic Research Associates, Credit Ratings for Structured
Products: A Review of Analytical Methodologies, Credit Assessment Accuracy, and Issuer Selectivity
among the Credit Rating Agencies (National Economic Research Associates, 6 November 2003),
online: National Economic Research Associates
96 Mark Adams, Manroop Jhooty & Jireh Wong, Dominion Bond Rating Services, Collateral Debt
Obligations Methodology: Industry StudySecuritization (Dominion Bond Rating Services, June
2004) at 5, online: Dominion Bond Rating Services
97 Ibid.
98 Moodys asserts that issuers do participate to unsolicited ratings. See Moodys Investor Service,
Moodys Special Comment, Designation of Unsolicited Ratings in Which the Issuer Has Not
Participated (November 1999) at 3, online: Moodys
supra note 29 at 311.
99 See Hill, Regulating, supra note 30 at 51-52.
100 Bottini, supra note 3 at 598; Report on the Role and Function of CRAs, supra note 4 at 24;
IOSCO Report, supra note 4 at 15.
101 Credit-Rating Agencies: AAArgh! The Economist (6 April 1996) 80.
102 Bottini, supra note 3 at 598; Sinclair, supra note 1 at 152-54; Jefferson County School District v.
Moodys Investors Servs., 988 F. Supp. 1341 (D. Colo. 1997) at para. 14; 25 Media L. Rep. 2351
[Jefferson].
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 637
unsolicited ratings allegedly distort the pricing mechanism since they tend to be less
accurate given the more limited amount of information on which they rely.103
Although these allegations are troubling, it remains unclear whether unsolicited
ratings amount to an abusive and harmful practice. Aside from the Moodys case,
there is little evidence to support the claim that unsolicited ratings are substantially
lower than solicited ratings.104 Interestingly, some have even gone so far as to claim
that unsolicited ratings may actually amount to a positive contribution as far as
market participants are concerned. First, unsolicited ratings could convey information
to the market about credit risk through a signaling effect.105 According to this view,
bad firms will choose not to signal their credit quality by refraining from
purchasing a rating. The inferior quality of those firms creditworthiness would then
be
ratings may
counterbalance the rating shopping bias that can arise when agencies sell favourable
ratings to gain or maintain their market share.106 Finally, unsolicited ratings can also
be seen as a mechanism to facilitate potential entry by new competitors.107 By issuing
unsolicited ratings, would-be entrants could build their reputations and gradually
establish themselves as credible alternatives to already established CRAs.
ratings. Second, unsolicited
the unsolicited
revealed by
These views are paramount in clarifying the true contribution of unsolicited
ratings. Still, in the absence of convincing empirical evidence, they cannot be taken to
mean that unsolicited ratings actually do have a positive impact on the market. Given
the Moodys case, there continues to be a risk that this practice may lead to abuse.
B. Legal and Institutional Constraints Influencing the Behaviour of
Credit Rating Agencies
1. The Role of Reputation in Shaping the Behaviour of Rating
Agencies
Issuers are only willing to pay for the service of rating agencies if it reduces their
cost of capital. Therefore, investors must consider that the agencys opinion
diminishes
issuers
creditworthiness. The value of this certification depends on the CRAs reputation with
respect to accuracy, independence, and integrity.108 If a CRA has a reputation for
erratic or biased analysis, investors will discount the value of the ratings assigned. If
information asymmetries
it accurately certifies
that
in
103 Schwarcz, supra note 2 at 16-17.
104 Ibid. See Byoun & Shin, supra note 82; David M. Ellis, Different Sides of the Same Story:
Investors and Issuers Views of Rating Agencies (1998) 7:4 Journal of Fixed Income 35 at 40-41.
105 Byoun & Shin, ibid.
106 Smith & Walter, supra note 29 at 312.
107 IOSCO Report, supra note 4 at 15.
108 See Hill, Regulating, supra note 30 at 50-51; IOSCO Report, ibid. at 3; Schwarcz, supra note 2
at 14; Smith & Walter, supra note 29 at 310-11.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
638
investors doubt the accuracy or independence of the ratings of a particular CRA,
issuers will seek a more credible agency to signal their creditworthiness.
[Vol. 51
Because of its value, the reputation of CRAs provides an economic incentive to
behave diligently and ethically, even in the absence of regulation. To build and
maintain their reputation, CRAs should be expected to put a concerted effort into
providing high quality services. In this respect, it is interesting to note that the larger
CRAs voluntarily disclose information about their rating methodologies and
processes.109 In its survey of CRAs, the IOSCO Technical Committee remarks that the
press release issued when a rating change occurs will usually provide information
about the various assumptions underlying the change.110 The effectiveness of
reputation as an incentive is debatable according to Partnoy who notes that the ability
of agencies to generate valuable information is hindered by factors such as the high
turnover level of their staff, and the relatively modest salaries paid to their analysts.111
As far as worries over CRA independence are concerned, since the fees derived
from a given issuer form a relatively small portion of their total revenues, CRAs
should not be willing to risk damaging their reputation to retain a particular firm as a
client.112 Industry practices indicate that reputation does in fact influence the
behaviour of CRAs. The reputational concerns raised by conflicts of interest are
reflected in the practices of rating agencies.113
Reputational concerns can also curb the potential for abuses stemming from
notching and unsolicited ratings.114 A CRA that disseminates false or misleading
information through such practices risks damaging its reputation.115 The case of
Moodys exemplifies this point. After having been chastised for giving low
unsolicited ratings, Moodys decided voluntarily to identify such ratings to investors
in order to help to dispel misconceptions, and increase the credibility and utility of
109 See generally IOSCO Report, ibid. at 10; Standard & Poors, Corporate Ratings Criteria 2006,
online: Standard & Poors
Code of Conduct, s. 2.1, online: Standard & Poors
Bond Rating Services, General Overview: Basic Approach to Rating Industrial Companies, online:
Dominion Bond Rating Services
110 IOSCO Report, ibid.
111 Partnoy, Paradox, supra note 39 at 72.
112 Smith & Walter, supra note 29 at 310-11. See also Stephen J. Choi, A Framework for the
Regulation of Securities Market Intermediaries (2004) 1 Berkeley Bus. L.J. 45 at 51-52 [Choi,
Framework].
113 Rating Agencies: Exclusion Zone The Economic (8 February 2003) 65: We may be
incompetent but were not dishonest. Standard and Poors, Rating Services Code of Conduct, s. 3.
114 Schwarcz, supra note 2 at 17-18.
115 See generally U.S., Hearings on the Current Role and Function of the Credit Rating Agencies in
the Operation of the Securities Markets, Securities and Exchange Commission (Washington, D.C.:
United States Government Printing Office, 2002), online: United States Securities and Exchange
Commission
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 639
[its] ratings in the capital markets.116 The willingness of CRAs to preserve their
reputation may explain why they tend to disclose the unsolicited character of the
rating in the press release accompanying it, even though they are not legally required
to do so.117
Competition is important for the enforcement of reputational sanctions that shape
the conduct of CRAs. In a competitive market, information about prices charged, the
level of service provided, and performance, tends to be more visible. Furthermore,
there are more alternatives with which to compare services offered.
to
investors. Second, other
Although the CRA market can be qualified as oligopolistic, there are several
mechanisms that facilitate the development of reputational sanctions. First, ratings are
publicly disclosed and accessible
information
intermediaries exist, such as financial analysts, in-house rating analysts, and niche
players, which also provide investors with data on the creditworthiness of issuers.118
Third, a CRA may render an unsolicited rating on any issuer that has already been
rated by a biased or incompetent agency.119 Finally, in the U.S. at least, issuers
typically seek ratings from the two major CRAs, which then compete to demonstrate
their analytical abilities. These mechanisms enable investors to monitor, compare, and
assess the performance of rating agencies.
Despite its role, reputation remains a noisy indicator. Investors are only privy to
the efforts of rating agencies indirectly through the default rate of the debt-
instruments that are rated. Thus, investors may attribute the same reputational effect
to debt-instruments that fail for different reasons such as fraud, bad luck or inaccurate
rating. In this respect, the fact that rating processes remain opaque further complicates
the task for investors.120 Also, reputation may not work effectively in periods of crisis:
rating agencies can intensify their mutual observations, thus producing similar
ratings in order to avoid being the only one wrong.121 In addition, the risk that a
particular agency should decide to milk its reputation, by lowering quality while
continuing to charge premium prices, limits the effectiveness of reputation as a
control mechanism over time.122 Finally, imperfect competition also reduces the
effectiveness of reputational pressures. Even if an agency suffers a loss of reputation,
new entrants may not be able to displace it because of the barriers to entry. Hence, a
loss of reputation may not necessarily translate into a loss of market share for an
116 Moodys Investors Service, supra note 98 at 3; Schwarcz, supra note 2 at 17.
117 Smith & Walter, supra note 29 at 311.
118 See Partnoy, Siskel and Ebert, supra note 3 at 650-51; Husisian, supra note 10 at 425-26;
Rhodes, supra note 9 at 316.
119 Hill, Regulating, supra note 30 at 76.
120 Partnoy, Siskel and Ebert, supra note 3 at 651.
121 Dieter Kerwer, Standardising as Governance: The Case of Credit Rating Agencies (Bonn:
Max-Planck Institute Collective Goods Preprint No. 2001/3, March 2001) at 25, online: Max Planck
Institute for Research on Collective Goods
Governance].
122 IOSCO Report, supra note 4 at 16.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
640
agency, thereby mitigating the impact of reputational pressures on established
agencies.
[Vol. 51
2.
Issuer Disclosure
For CRAs to make accurate assessments of the creditworthiness of debt-
instruments, they require access to information about issuers at the moment of the
offerings as well as subsequently.123 The primary source of firm-specific information
for CRAs is the disclosure documents released by issuers. Agencies may also engage
in discussion with the management of issuers to gain additional access to non-public
information that is relevant to their analysis. In this respect, communications between
issuers and CRAs benefit from special regulatory treatment.124
Although issuers are subject to mandatory disclosure provisions, the quality and
the timeliness of the information provided are not assured.125 In the U.S.,
commentators have remarked before the SEC that several deficiencies exist with
respect to issuers disclosure of short-term credit facilities and ratings triggers in
material contracts.126 Furthermore, mandatory and voluntary disclosure is not always
accurate. However, it is not currently a practice of CRAs to verify or audit the
information disclosed by issuers. This lack of inquisitiveness on the part of CRAs has
been criticized by the U.S. Senate Committee on Governmental Affairs investigating
the Enron failure: their monitoring and review of [Enrons] finances fell far below
the careful efforts one would have expected from organizations whose ratings hold so
much importance.127
To be fair, this behaviour on the part of CRAs may be economically justifiable.
Rating agencies cannot investigate issuers on an ongoing basis as thoroughly as they
did during the initial rating process since the costs of doing so would be very high.128
In this respect, it appears reasonable for rating agencies to rely on other third-party
123 Ibid. at 12.
124 See Securities Act, R.S.O. 1990, c. S.5, s. 76(2) [Securities Act]; Disclosure Standards, supra
note 63, s. 3.3.(2); Borden Ladner Gervais, Securities Law and Practice, 3rd ed. (Toronto: Carswell,
1988) at 18.2.8.
125 See e.g. Ontario Securities Commission Staff Notice, 51-708, Continuous Disclosure Review
Program ReportAugust 2002 (August 2002), online: Ontario Securities Commission
Commission, Staff Notice, 52-713, Report on Staffs Review of Interim Financial Statements and
Interim Managements Discussion and AnalysisFebruary 2002 (February 2002) , online: Ontario
Securities Commission
DisclosureInterim Report, (1996) 19 O.S.C. Bull. 1 at 40-41.
126 Report on the Role and Function of CRAs, supra note 4 at 30.
127 Watchdogs Report, supra note 4 at 115.
128 Hill, Regulating, supra note 30 at 70.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 641
certifying mechanisms to ensure the disclosure of accurate information.129 Besides,
although CRAs may downgrade or withdraw an existing rating where they detect that
issuers are not fully transparent and are withholding important information, this
sanction has limited effectiveness.130
3. The Influence of Securities Regulation on the Transparency and
Dissemination of Ratings
Securities regulation does not impose any direct obligations on CRAs. Issuers
must disclose ratings obtained from CRAs in their prospectus.131 However, the extent
of the mandated disclosure is rather limited as the prospectus need only provide
information on the rating obtained, the name of the rating organization, as well as
details on the meaning of the rating.132 In the secondary market, pursuant to National
Policy 51-201, a rating change consists primarily of material information and should
lead issuers to establish a material change report as well as issue a press release.133
This does not mean that CRAs operate in a total regulatory vacuum. It is
interesting to note the general remarks formulated by securities regulators in National
Policy 51-201 when setting out the contours of the selective disclosure exemptions.134
The policy states that CRAs benefit from such an exemption since their ratings are
either provided confidentially to issuers or disclosed to a wide public audience.
According to the policy, the objective of the rating process is to provide a widely
available publication of the rating. In contrast, regulators note that securities analysts
reports are primarily aimed at the firms clients. These remarks suggest that the scope
of the disclosure of ratings is a factor that directly influences the regulatory treatment
of communications between CRAs and issuers. This treatment also provides an
additional incentive for CRAs to publicly disclose and disseminate their ratings.
More specifically, CRAs must respect the provisions of securities regulation that
deal with insider trading.135 When they obtain material non-public information from
issuers, CRAs may not communicate or trade on this inside information. This
prohibition extends to their analysts and other employees. Besides, as seen below,
CRAs are subject to the liability regimes enacted by securities regulation.
129 See Marcel Kahan, The Qualified Case Against Mandatory Terms in Bonds (1995) 89 Nw.
130 IOSCO Report, supra note 4 at 12-13; Report on the Role and Function of CRAs, supra note 4 at
U.L. Rev. 565 at 590-93.
31-32.
131 See Prospectus Disclosure Requirements, O.S.C., NI 41-101, 23 O.S.C. Bull. (Supp.) 761 (15
December 2000), Item 10.8 [Information Required in a Prospectus], online: Ontario Securities
Commission
132 Ibid.
133 Disclosure Standards, supra note 63, s. 4.3.
134 Ibid., s. 3.3(7).
135 Securities Act, supra note 124, s. 76.
642
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 51
4. The Liability of Credit Rating Agencies
Although CRAs are not specifically regulated, their activities remain subject to
the general civil liability regimes that apply to other participants in the securities
market. Theoretically, liability can positively influence the accuracy of CRAs
analyses.136 CRAs can be held liable towards investors for the ratings they disclose
where they contain misrepresentations. In the primary market, the rating of debt
instruments by a CRA in the context of an offering triggers disclosure obligations for
issuers. Issuers must disclose the rating obtained in their prospectus, the name of the
rating organization, as well as details on the meaning of the rating.137 In theory, such
disclosure of information about the rating granted should activate the application of
the statutory liability regime enacted by securities legislation. However, the liability
regime only applies to those persons whose consent has been filed pursuant to a
requirement of the regulations. Exceptionally, in the case of ratings, issuers need not
obtain the written consent of the rating agency for the disclosure of this
information.138 Therefore, the statutory civil liability regime will not apply to ratings
disclosed in the prospectus.
Thus, CRAs are only subject to the general common law (or civil law) liability
regime in place in the primary market. The situation is the same in the secondary
market where the new statutory civil liability regime does not cover ratings. Since the
application of the general liability regimes proves very difficult for investors, liability
is not a significant constraint affecting the behaviour of rating agencies in Canada.
For different reasons, liability also plays a limited role in disciplining rating agencies
operating in the American markets. In the U.S., the publications of CRAs have
traditionally been afforded the protection of the First Amendment.139 CRAs are
therefore not liable for negligent misrepresentations. Only if their conduct is reckless
can they be held liable for misrepresentations.
C. Summation: Is There an Accountability Gap?
Given their role in capital markets, CRAs wield power over issuers and
investors. As standards of creditworthiness, ratings have a coercive impact on issuers
given the regulatory use of ratings. Irrespective of regulation, ratings also affect
issuers as they can determine the conditions under which they may access debt
markets, the conditions of their relationships with lenders, and the structure of their
transactions. For investors, ratings are a screening tool that influences the
composition of their portfolios as well as their investment decisions.
However, despite the fact that rating agencies have become increasingly
influential in global financial markets, it is very hard to hold them accountable for
136 Husisian, supra note 10 at 430.
137 See Information Required in a Prospectus, supra note 131, Item 10.8.
138 See ibid., s. 13.4(4).
139 Husisian, supra note 10 at 446ff; Jefferson, supra note 102 at para. 13.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 643
their action[s] … 140 This is because the exit and voice options that are available to
market participants to hold rating agencies accountable work imperfectly. As for exit,
it is difficult for issuers and investors to do away with CRAs given the current
regulatory use of ratings. As for voice, the only real existing mechanism appears to be
reputation.141
In other words, there appears to be an accountability gap, which constitutes an
imbalance between the power of CRAs and the possibility of holding them
responsible.142 This accountability gap is worrisome. For CRAs, the accountability
gap may affect their credibility in the marketplace.143 For market participants, it is of
particular concern given the role that CRAs play in capital markets: Investors and
markets generally are hurt if they give ratings more credence that is warranted; they
also may be hurt by the volatility caused by precipitous upgrading and
downgrading.144
III. Enhancing the Accountability of Credit Rating Agencies: The
Role of Regulation
The accountability gap affecting credit rating agencies preoccupies policy makers
and regulators. This has led to an active effort to find solutions to bridge the
accountability gap. In December 2004, the IOSCO published a Code of Conduct
Fundamentals for Credit Rating Agencies, which aims to ensure the quality and
integrity of ratings.145 According to the IOSCO, the implementation of the codes
principles should be left to market pressures. In April 2005, the SEC proposed
adopting a rule that would clearly define the conditions that an entity must satisfy in
order to obtain an NRSRO designation.146 In Europe, the Committee of European
Securities Regulators, which provided counsel to the European Commission
regarding possible measures concerning CRAs, remarked that the IOSCO Code
strikes a balance between the different interests of the rating process; those of the
agencies themselves, those of the issuers and those of investors.147 It favoured the
approach of the IOSCO and advised the Commission that the implementation of the
Code be left to market pressures for the time being. In a recent speech, Internal
Market Commissioner McCreevy indicated that the Commission would follow this
140 Kerwer, Standardising as Governance, supra note 121 at 3. See also Watchdogs Report, supra
note 4 at 122 (the U.S. Senate Committee Staff concluded that credit analysts do not view themselves
accountable for their actions).
141 CESRs Technical Advice, supra note 5 at 51.
142 Kerwer, Standardising as Governance, supra note 121 at 20. See also David Millon, Who
Caused the Enron Debacle? (2003) 60 Wash. & Lee L. Rev. 309 at 324.
143 See Clark C. Havighurst, The Place of Private Accrediting Among the Instruments of
Government (1994) 57:4 Law & Contemp. Probs. 1 at 4.
144 Hill, Regulating, supra note 30 at 84.
145 IOSCO Code, supra note 5.
146 NRSRO Definition, supra note 5.
147 CESRs Technical Advice, supra note 5 at 49.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
644
advice.148 In light of these initiatives, this part examines which approach of enhancing
the accountability of CRAs would best suit the Canadian regulatory framework.
[Vol. 51
A. The Regulation of Credit Rating Agencies: Policy Perspectives
1. The Goals of Securities Regulation
Credit rating agencies can be qualified as information intermediaries. By
reducing information asymmetries between issuers and investors, they perform a
function that contributes to an efficient capital market. Given this function, CRAs are
institutions that fall under the scope of securities regulation. Any regulatory effort
concerning rating agencies should espouse the twin goals underlying securities
regulation: efficiency and investor protection.
a. Market Efficiency
Securities regulation aims to foster greater market efficiency.149 Generally
speaking, this entails encouraging the most rational allocation of capital resources.
First, the goal of market efficiency requires that regulation promote informational
efficiency by reducing information asymmetries between issuers and investors.
Second, efficiency requires that transaction costs be kept low so as to ensure the
continued use of capital markets.
In the context of CRAs, the goal of market efficiency requires that regulation
seek to prevent the market failures that can affect rating agencies activities and
processes. Regulation would thus improve the accuracy and credibility of credit
ratings, and thereby contribute to the informational efficiency of the market.
At the same time, regulatory interventions should factor in the need to maintain
low transaction costs for participants in the market for debt instruments. In this
regard, there has been little concern voiced over the level of fees charged by rating
agencies in the United States or Canada.150 Without dismissing the possibility that
CRAs may be able to extract supra competitive fees, in the absence of hard data,
more attention should be directed towards those costs associated with regulation.
148 Rating Agencies let off the hookfor now EurActiv (11 April 2005) online: EurActiv.com
149 Securities Act, supra note 124, s. 1.1. See David Johnston & Kathleen Doyle Rockwell,
Canadian Securities Regulation, 3d ed. (Toronto: LexisNexis Butterworths, 2003) at 2-6.
150 Schwarcz, supra note 2 at 12-13.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 645
b.
Investor Protection
Securities regulation also purports to protect investors from fraud and other forms
of exploitation in order to preserve public confidence in the market.151 To some
extent, the goal of investor protection runs opposite to the goal of market efficiency.
For this reason, there has been considerable debate about the appropriate emphasis of
each of these goals in the context of securities regulation. Without presuming to
resolve this debate it seems that investor protection should carry less weight than
efficiency when discussing regulatory interventions aimed at credit rating agencies.
The market for corporate bonds tends to be dominated by institutional investors
or sophisticated investors.152 Furthermore, it appears that the ownership of bonds of
specific issues is concentrated with a small number of investors owning a high
proportion of the bonds of a single issue. These characteristics of the bond market are
highly relevant when discussing the goal of investor protection.153 Institutional
investors possess the expertise to analyze the information disclosed by issuers and to
verify its accuracy. In this respect, since institutional investors tend to participate in
many different offerings, they tend to have significant experience when it comes to
analyzing the value of issuers. Institutions also possess the resources to dig deeper
when researching issuers, enabling them to uncover undisclosed information that can
affect credit risk. In sum, institutional investors are not defenceless in credit markets.
They are also not entirely dependent on CRAs to assess debt instruments.154 Since they
can form critical judgments of the ratings in light of their own analysis and research,
institutional investors can contribute to moderating the conflicts of interest affecting
CRAs.155
The dominance of institutional investors in bond markets also carries implications
for retail investors. Investors can protect themselves by investing their funds through
institutional investors to benefit from the informational advantage that the latter
possess. Generally speaking, retail investors can benefit from the efforts of
institutional investors, which are instrumental in debt instruments valuation. Thus,
retail investors are not defenceless either in credit markets. Therefore, it is doubtful
that the goal of investor protection should drive regulatory interventions targeted at
151 Securities Act, supra note 124, s. 1.1; See Johnston & Rockwell, supra note 149 at 2-4.
152 See IOSCO Report, supra note 4 at 4; Bank of Canada, Submission on the Revised CSA
Alternative Trading System (ATS) Proposal, Legislative Comment at 5, online: Ontario Securities
Commission
Bank of Commerce, [1998] 44 B.L.R. (2d) 192 (Ont. Ct. J.) at para. 34.
153 See Kahan, supra note 129 at 583-86.
154 Baker & Mansi, supra note 85 at 1395, online: Social Science Research Network
155 See Alexander Ljungqvist et al., Conflicts of Interest in Sell-side Research and The Moderating
Role of Institutional Investors (12 September 2005) [unpublished, archived at Salomon Center, Stern
School of Business, New York University], online: Social Science Research Network
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
646
credit rating agencies. Policy should be more concerned with ensuring that the market
is efficient.156
[Vol. 51
2. The Importance of a Cost-Effective Regulatory Regime
When discussing policy options, it is important to remember that government
failures can also exist. It is possible to distinguish two types of government failures.157
The first arises where the government is guided by the public interest. Even under
such circumstances, regulation generates costs that must be taken into account when
assessing the utility of government intervention. One category of costs can be
qualified as administrative. These are generated by the government agency charged
with the formulation of the rules and standards of conduct, the monitoring of
behaviour and the enforcement of compliance. A second category is the compliance
costs borne by market participants. These costs depend on the degree of precision and
flexibility of regulation. They are compounded by the lack of harmonization of
national regulatory regimes that apply to market participants, such as CRAs, that
operate in multiple jurisdictions.
A second type of government failure surfaces where the assumption concerning
the competence of government officials and the objective guiding their interventions
are relaxed. In the real world, government officials can be incompetent, plagued by
informational problems or affected by psychological biases, just as any other market
actor. Where this is the case, the costs of regulation will be greater and the benefits
smaller. Even more worrisome, government officials may not be guided by the public
interest when shaping a regulatory regime. As public choice theorists argue, they may
be pursuing the private goals of concentrated interest groups which captured them in
one way or another. This may lead to excessive or insufficient regulation.
These two types of failures have undesirable implications when making policy
choices. The first is the need to consider that regulation generates costs that may
offset any efficiency gains sought. Even where there is a market failure, government
intervention may not always yield a superior outcome. Any case for regulatory
intervention will thus have to demonstrate not only that a market imperfection exists,
but that its impact would be reduced by the proposed policy measure or reform in a
cost-effective manner. Thus, the analysis of policy options concerning CRAs should
always take cost-effectiveness into consideration.158 The second implication involves
considering that the goal of efficiency may be compromised when private interests
groups capture the regulator. Where such capture occurs, the risk that regulation be
156 CESRs Technical Advice, supra note 5 at 52.
157 Peter P. Swire, Markets, Self-Regulation, and Government Enforcement in the Protection of
Personal Information (1999) [unpublished], online: Social Science Research Network
158 See William P. Albrecht et al., Regulatory Regimes: The Interdependence of Rules and
Regulatory Structure in Andrew W. Lo, ed., The Industrial Organization and Regulation of the
Securities Industry (Chicago: University of Chicago Press, 1996) at 11.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 647
biased against entry and competition surfaces.159 In light of this risk, Mayer cautions
that [t]he scope of regulation should therefore be limited to areas where there is a
clear case of market failure.160
From this perspective, enhancing the accountability of CRAs requires balancing
the failings of both market and regulatory mechanisms. Following the framework
proposed by Choi, this involves evaluating the regulatory options in light of market-
based incentives that exist to address the various problematic issues affecting CRAs.161
Thus, [w]here the market has an incentive to correct for any failings, less
interventionist regulation is required.162
B. An Overview of the Current Proposals to Enhance the
Accountability of Credit Rating Agencies
1. The IOSCO Code of Conduct Fundamentals for Credit Rating
Agencies
At the international level, the IOSCO Technical Committee published a statement
of principles that laid out high-level objectives that rating agencies, regulators,
issuers and other market participants should strive toward … 163 Subsequently, the
Committee proposed a Code of Conduct Fundamentals that provides guidelines with
respect to those principles. Pursuant to the Code, discretion involving the
implementation of the fundamentals is left to CRAs and is subject to market
pressures.
a. The Rating Process
i. The Quality of the Rating Process
The Code sets forth a series of provisions to ensure the quality of the opinions
expressed by CRAs. Although the concept of quality is not defined, it appears to
refer to the accuracy of the ratings. The Code states that CRAs should seek to avoid
disclosing analyses or reports that contain misrepresentations as to the general
159 See Luigi Zingales, The Costs and Benefits of Financial Markets Regulation, Law Working
Paper No. 21/2004 (April 2004) [unpublished, archived at the University of Chicago European
Corporate Governance Institute], online: Social Science Research Network
160 Colin Mayer, The Regulation of Financial Services: Lessons from the U.K. for 1992 in
Matthew Bishop, John Kay & Colin Mayer, The Regulatory Challenge (Oxford: Oxford University
Press, 1995) at 148.
161 Choi, Framework, supra note 112 at 71.
162 Ibid.
163 IOSCO Code, supra note 5 at 1.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
648
creditworthiness of an issuer or security. To attain this objective, the provisions stress
the importance of methodologies and processes, people, and resources.
[Vol. 51
With respect to methodologies and processes, the Code emphasizes the need for
CRAs to conduct a thorough analysis of all public and non-public information
known and believed to be relevant. CRAs are also encouraged to use rating
methodologies that are rigorous, systematic, and, where possible, result in ratings that
can be subjected to some form of objective validation based on historical
experience.164
Recognizing the central role of analysts, the Code dictates that CRAs should rely
on analysts who, individually or collectively have appropriate knowledge and
experience in developing a rating opinion.165 The analysts should be grouped in
teams to promote continuity and avoid bias in the rating process. Furthermore, the
analysts should use the same methodologies established by the CRAs that employ
them.
Finally, the Code recommends CRAs ensure that sufficient resources are devoted
to ratings in order to carry out high quality credit assessments.166 Any decision to
rate or continue rating should be preceded by an assessment of whether adequate
personnel possessing sufficient skills sets can be involved to make a proper rating
assessment.
ii. Monitoring and Updating
The Code provides that CRAs should monitor the ratings on an ongoing basis. To
monitor ratings, CRAs should regularly review the issuers creditworthiness. They
should initiate a review of the rating upon receipt of any information that might
reasonably be expected to result in a rating action. Based on the results of such
review, CRAs should update the rating in a timely fashion. Where a CRA
discontinues rating an issuer or debt instrument, it should make the decision public. If
the discontinued rating remains published, the CRA should indicate the date the
rating was last updated and the fact that it is no longer being updated.
iii. The Integrity of the Rating Process
The Code contains several provisions that purport to reinforce the integrity of the
rating process. In general, the Code provides that the CRAs and their employees
should comply with applicable laws and regulations, and deal fairly and honestly with
issuers, investors, and other market participants. More specifically, the CRAs and
their employees should not give issuers any assurances or guarantees of a particular
164 Ibid. at 4.
165 Ibid.
166 Ibid.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 649
rating prior to the actual assessment. Furthermore, the Code explicitly states that a
CRAs analysts should be held to high standards of integrity.
The Code provides that a CRA should institute policies and procedures that
clearly identify a specific person responsible for ensuring compliance with the code
of conduct and with applicable laws and regulations. Furthermore, the Code imposes
a whistleblowing duty upon every CRA employee to report illegal or unethical
conduct.
b. Credit Rating Agencies Independence and Avoidance of
Conflicts of Interest
i. General Principles
The Code urges the CRA and its analysts to use care and professional judgment
in order to maintain both the substance and appearance of independence and to
promote objectivity. It states that the determination of a rating should be based solely
on factors relevant to the credit assessment. The existence or potential existence of a
business relationship between CRA and issuer, or the absence of such a relationship,
should not affect the rating process. To avoid such a situation, the Code provides that
CRAs should separate their credit rating business and CRA analysts from any other
business they conduct.
ii. Procedures and Policies
CRAs should adopt internal procedures and mechanisms to identify and
eliminate, or manage and disclose, any actual or potential conflicts of interest that
may affect rating opinions. Where the CRAs elect to manage the conflicts of interest
through disclosure, they should ensure that the disclosure of actual or potential
conflicts of interests is complete, timely, clear, concise, specific, and figures
prominently in the disclosure documents. In either case, the CRAs should disclose
these internal procedures and mechanisms to the public.
The Code concedes that certain conflicts of interests cannot be managed
effectively. It enjoins CRAs and their staff to refrain from trading securities or
derivatives presenting conflicts of interest with their ratings activities. Likewise, the
Code states that CRAs should proceed with caution when rating a government issuer
that is also involved in the oversight of rating. Specifically, a CRA should ensure that
it uses different employees to conduct the rating than those involved in the oversight
issues with the government issuer.
The Code also contains specific provisions dealing with compensation of CRAs,
which can raise conflict of interest issues. It provides that CRAs should disclose the
general nature of their compensation arrangements with issuers and other rated
entities. They should also disclose compensation already received from issuers,
including compensation which is unrelated to the rating service, making sure to
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
650
highlight the proportion of total earnings such compensation represents in comparison
to fees received strictly from rating services.
[Vol. 51
iii. Analyst and Employee Independence
Pursuant to the Code, the CRA should structure their compensation arrangements
in such a way as to eliminate or manage effectively all actual or potential conflicts of
interests in order to ensure the independence of its analysts and employees. CRAs
should explicitly state and ensure that the employees and analysts compensation or
evaluation will not depend on the amount of revenue derived from the issuers they rate.
The Code also provides specific prohibitions in order to ensure the independence
of analysts and employees. Whenever an analyst becomes involved in a personal
relationship that creates the potential for any real or apparent conflicts of interest, the
analyst should be required to disclose such a relationship to the appropriate officer of
the CRA, in accordance with its compliance policies.
c. Credit Rating Agencies Responsibilities to the Investing
Public and to Issuers
i. Transparency and Timeliness of Ratings Disclosure
The Code enjoins CRAs to strive for transparency with respect to their ratings
and rating processes. It requires that the ratings decisions, or the subsequent decision
to discontinue a rating, be dispensed in a timely manner, on a non-selective basis, and
free of charge. When issuing a rating, agencies should explain the key elements
underlying the rating decision. Furthermore, where feasible and appropriate, CRAs
should advise issuers of all critical information and principal considerations upon
which rating will be based prior to issuing or revising a rating, in order to afford the
issuer the opportunity to provide the CRA with any factual clarifications that will be
necessary for it to produce an accurate rating. CRAs should also disclose whether or
not ratings are solicited at the request of the issuer, and whether or not the issuer
participated in the rating process.
Moreover, CRAs should publicly disclose their policies for distributing ratings
and reports. They should also publish sufficient information about their procedures,
methodologies and assumptions so that outside parties can understand how the rating
was derived. When these procedures, methodologies or assumptions change, the
CRAs should publicly disclose all of the modifications.
Finally, the Code requires that CRAs publish sufficient information about the
historical default rates of its ratings categories, and whether the default rates of these
categories have changed over time, so that interested parties can understand the
historical performance of each category, as well as see if and how ratings categories
have changed. This requirement seeks to promote transparency and enable the market
to assess the performance of the ratings by drawing comparisons from ratings issued
by different CRAs.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 651
ii. Treatment of Confidential Information
CRAs should establish procedures and mechanisms to protect the confidential
nature of the information it receives from issuers. Unless permitted, the CRA should
not disclose confidential information in press releases or in other communications
with market participants.
The CRAs should only use the confidential information obtained for purposes
relating to their rating activities. Their employees should not share confidential
information with employees of any affiliated entities that are not CRAs, nor share
such confidential information with colleagues within the CRA unless necessary. In
addition, the employees should not selectively disclose any non-public information
about rating opinions or future rating actions.
The employees of CRAs should be prohibited from engaging in transactions
involving securities when they possess confidential information concerning the issuer
of such securities. They should also refrain from using or sharing confidential
information for purposes relating to trading securities, or for any other purpose
unrelated to the conduct of the CRAs business.
CRAs should ensure the confidentiality of the information gathered by ensuring
that their employees take all reasonable measures to protect it. The agencies should
also make sure that their employees familiarize themselves with internal securities
trading policies.
d. Disclosure of the Code of Conduct
The IOSCO Code recommends that CRAs publicly disclose their codes of
conduct, making sure to highlight how they measure up to the provisions contained in
the IOSCO Code. Thus, CRAs should describe to what extent the provisions of their
own codes of conduct are consistent with those of the IOSCO Code. CRAs should
also describe how they intend to implement and enforce its code of conduct, as well
as disclose any changes to its provisions or implementation on a timely basis.
2. The Securities and Exchange Commissions Proposal to Define
the Concept of Nationally Recognized Statistical Rating
Organization
Following hearings and reports by lawmakers, the SEC conducted its own study
on the role and the function of CRAs.167 It published a Concept Release where it
discussed various issues relating to credit rating agencies, including whether credit
ratings should continue to be used for regulatory purposes and, if so, which process to
167 Sarbanes-Oxley Act of 2002, 15 U.S.C. tit. 7 702(b), online: FindLaw
Function of CRAs, supra note 4.
[Vol. 51
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
652
use in determining whose credit ratings should be used, and the level of oversight to
apply to credit rating agencies.168 The ensuing discussions generated by the Concept
Release led the SEC to propose a rule that would define the conditions that agencies
must satisfy in order to obtain an NRSRO designation.169 In addition to the proposed
rule, the SEC is also considering the development of a voluntary framework for
oversight of CRAs that would enable it to assess whether the latter continue to meet
the NRSRO designation criteria. The development of this framework is necessary
because the SEC lacks formal jurisdictional authority over CRAs in this respect.
The rule proposed by the SEC seeks to bring a greater degree of clarity and
transparency to the NRSRO designation process, while ensuring that agencies
recognized as NRSROs use systematic procedures to address concerns raised by their
activities. Specifically, the rule provides that a rating agency must satisfy the three
conditions to be designated as an NRSRO, at which point, the NRSRO designation
would be approved by the SEC staff through no-action letters as is presently the case.
a.
Issuance of Publicly Available Ratings that Are Current
Assessments of
the Creditworthiness of Obligors with
Respect to Specific Securities
The organization should issue ratings that are publicly available.170 The public
accessibility criterion would refer only to the rating itself and not to other information
such as the ratings rationale. Thus, the organization would continue to benefit from
the possibility of offering subscriptions to newsletters providing more details about its
ratings.
Ratings publicly disseminated by the organization would have to focus on the
creditworthiness of the issuer with respect to specific securities or obligations. This is
because the regulatory use of the term NRSRO primarily relates to credit ratings on
specific securities or obligations.171 It would therefore be insufficient for the
organization to only provide ratings on the general creditworthiness of issuers.
Finally, the organization would have to provide ratings that are current, i.e., that
reflect its opinion of the creditworthiness of the securities at the time of the rating up
until it is amended or withdrawn. In order to meet this requirement, the organization
would have to implement and adhere to procedures destined to ensure that ratings are
reviewed and updated whenever material changes occur.
168 U.S. Securities and Exchange Commission, Concept Release, 33-8236, Rating Agencies and the
Use of Credit Ratings under the Federal Securities Laws (10 July 2003), online: United States
Securities and Exchange Commission
169 NRSRO Definition, supra note 5.
170 See ibid. at 24-25.
171 Ibid. at 25.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 653
b. Recognition of Credibility and Reliability by the Financial
Markets
The NRSRO status would continue to be granted to organizations that have a
wide recognition in the marketplace. Thus, the definition of NRSRO would contain a
requirement whereby the organization would be required to demonstrate that it is
generally recognized in the marketplace as an issuer of credible and reliable ratings
by the majority of participants that make use of securities ratings.172 This criterion
raises the question as to whether niche players would be able to meet the proposed
NRSRO requirements. In its proposed rule, the Commission remarks that an
organization that has gained recognition for a limited sector of the debt market or a
limited geographic area could meet this criterion. The SEC does not preclude an
organization that relies primarily on quantitative statistical models from being
recognized as an NRSRO.
c. Systematic Procedures to Ensure Credible and Reliable
Ratings
the
The third component of the proposed NRSRO definition would require that the
organization implement systematic procedures to ensure credible and reliable ratings.
The SEC proposes eight factors that would be important in its assessment of whether
an organization meets this component of the definition: (1)
and
training of a firms rating analysts; (2) the average number of issues covered by
analysts; (3) the information sources reviewed and relied upon by the credit rating
agency and how the integrity of information utilized in the ratings process is verified;
(4) the extent of contacts with the management of issuers, including access to senior
level management and other appropriate parties; (5) the organizational structure of the
credit rating agency; (6) how the credit rating agency identifies and manages or
proscribes conflicts of interest affecting its ratings business; (7) how the credit rating
agency monitors and enforces compliance with its procedures designed to prohibit the
misuse of material, nonpublic information; and (8) the financial resources of the
credit rating agency.
experience
Albeit through a different approach, these factors deal with issues that are similar
to those addressed in the IOSCO Code.
C. A Critical Look at Regulatory Strategies for Implementing
Accountability-Enhancing Rules of Conduct
The above initiatives are characterized by two main dimensions. The first relates
to the substantive rules of conduct. In this respect, a comparison of the IOSCO and
172 Ibid.
[Vol. 51
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
654
SEC initiatives indicates that both propose rules of conduct that deal with similar
concerns. The second dimension relates to the regulatory strategy favoured to
implement the rules of conduct. In this respect, the SEC and the IOSCO differ in their
respective approaches. The SEC proposes to maintain the use of the NRSRO concept
while clarifying the definition of this concept so as to ensure that it imposes clearer
and arguably more demanding requirements with respect to agencies organization
and conduct. The IOSCO favours a self-regulatory approach based on a detailed
model code of conduct. In Canada, David Brown, Chair of the OSC at the time,
endorsed the statement of principles contained in the IOSCO Report.173
The IOSCO and SEC initiatives offer a good indication of the issues that
Canadian regulators may have to consider dealing with in order to enhance the
accountability of CRAs. Assuming that this is the case, the question to address
becomes which strategy should be used to implement these substantive rules of
conduct?
1. Self-Regulation with a Voluntary Code of Conduct
A first strategy would be to follow the approach suggested by the IOSCO
pursuant to which CRAs should self-regulate through the adoption of voluntary codes
of conduct.174 A voluntary code of conduct could be developed by each individual
rating agency or by an industry association. CRAs could even establish a voluntary
industry association at a national, regional, or international level with the mandate to
adopt a model industry code of conduct.175
With respect to content, CRAs could use the IOSCO model. In the case of an
industry-wide code, adherence to the code would take the form of agreements
between CRAs, which would
the
association.176 The contracts would provide for sanctions for breach of the code. In
the case of an individual firm-specific code of conduct, it would be incumbent on the
imply a contractual obligation
towards
173 Ontario Securities Commission, News Release Communiqu, OSC Chair Endorses
International Standards for Analysts and Credit Rating Agencies (26 September 2003), online:
Ontario Securities Commission
174 See generally Margot Priest, The Privatization of Regulation: Five Models of Self-Regulation
(1998) 29 Ottawa L. Rev. 233 at 245-51; Industry Canada, Voluntary CodesA Guide for their
Development and their Use (Ottawa: Office of Consumer Affairs, March 1998) [Voluntary Codes],
online: Industry Canada
175 Giorgos Katiforis, European Parliament Committee on Economic and Monetary Affairs, Report
on Role and Methods of Ratings Agencies (A5-0040/2004) (Committee on Economic and Monetary
Affairs, 29 January 2004) at 7, online: European Parliament
176 Industry Canada, An Evaluation Framework for Voluntary Codes (Ottawa: Office of Consumer
Affairs, 2000), online: Industry Canada
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 655
board of directors to adopt the IOSCO Code as is, or with some modifications, and to
ensure the enforcement of its provisions.
From an efficiency perspective, the use of a code of conduct would not involve
significant administrative costs. The drafting of the code would be done by regulated
entities themselves, rather than by the government. CRAs would thus mobilize their
own expertise and resources towards the development of the code, thereby reducing
drafting costs. Of course, they would also have the luxury of relying on the IOSCO
Code in order to reduce drafting costs.
Given the small number of industry players at the regional and the international
levels, a global industry code of conduct would be less expensive to produce through
negotiations between leading agencies, rather than by the governments involved,
because a consensus could be easier to obtain amongst industry members. Thus, the
development of an industry code of conduct could be a relatively simple way of
attaining the essential goal of harmonization.
For CRAs, adopting a voluntary code of conduct model could yield some
benefits. This model would avoid the difficult question of setting requirements for the
recognition of organizations as credit rating agencies. The CRAs would have the
discretion of adopting the IOSCO Code off-the-rack or developing their own code
of conduct. This would prevent the creation of additional barriers to entry and would
reduce the risk of lowering competition. Because of its non-statutory nature, a code of
conduct would be more flexible than regulation and could therefore adapt more
rapidly and inexpensively to reflect changes in the industry. Moreover, adopting the
IOSCO Code would assist rating agencies in bonding their credibility vis–vis
market participants thereby attenuating the potential agency problems outlined above.
However, this approach would not remove the barriers to entry that can result from
the current regulatory use of ratings.
To be sure, the participation of CRAs in drafting an industry code could have
some drawbacks. In the case of an industry code, the large agencies could use their
clout to exercise control over the essence of the code, thereby emphasizing
requirements that may prove ill-suited for smaller rating agencies. Likewise, the large
CRAs could take advantage of their power to include anticompetitive provisions in
the industry code.177 Although agencies would be free not to adhere to such an ill-
suited code, failure to do so could be perceived negatively by market participants who
consider the code to be an industry standard. Under similar circumstances, the
industry code of conduct could become a barrier to entry. This drawback would
however be avoided by using the IOSCO Code as the model for the industry code.
From an accountability perspective, the use of an industry code would give rise to
several challenges. The first relates to the openness of the codes development. If left
solely in the hands of CRAs, the drafting of the code may leave little place for
177 See generally Swire, supra note 157; Voluntary Codes, supra note 174 at 6.
[Vol. 51
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
656
involvement by stakeholders.178 This may limit the effectiveness of the code in
addressing the concerns raised by the activities of CRAs with respect to the public. To
alleviate this problem, the code would have to be developed using a process that
provides for some involvement of market participants as well as other members of the
public. In this respect, the proposition of the IOSCO appears particularly interesting
in that the model code is the product of consultations with regulators and market
participants.
The second and most important challenge of this option relates to compliance and
enforcement.179 Reputation would be the main incentive leading CRAs to adopt a
code of conduct modelled on the IOSCO Code, and to respect its provisions.180
However, the voluntary code of conduct option does not provide any solution to the
limits of reputational pressures. Furthermore, if the degree of disclosure is left to the
discretion of CRAs, there is a possibility that the market would not be properly
informed on howand in which measurethe CRAs are implementing the IOSCO
Code.181 It is true that a code would operate within the existing legal and regulatory
landscapes. The principles set forth in the code would influence the construction of
the legal duties and liabilities imposed upon CRAs, thereby extending the effects of
the code in favour of third parties.182 Reliance on the existing legal regime to ensure
the enforcement of the provisions of the code remains a questionable approach. In the
absence of effective compliance and enforcement mechanisms, it remains unclear
whether CRAs could be held accountable to market participants.183 Ultimately, the
adoption of a code of conduct by CRAs could amount to nothing more than mere
window dressing.184
2. Strategies Regulating both Entry and Conduct
Rather than follow the IOSCO approach, Canadian regulators could devise a
strategy modeled on the U.S. approach to regulate both entry and conduct within the
CRA industry. This strategy could be implemented through supervised self-regulation
or a registration model.
178 Priest, supra note 174; Voluntary Codes, ibid. at 6-7.
179 See Harvey L. Pitt & Karl A. Groskaufmanis, Minimizing Corporate Civil and Criminal
Liability: A Second Look at Corporate Codes Of Conduct (1990) 78 Geo. L.J. 1559 at 1604-605.
180 CESRs Technical Advice, supra note 5 at 42.
181 Ibid. at 43.
182 See e.g. Priest, supra note 174; Voluntary Codes, supra note 174 at 27.
183 CESRs Technical Advice, supra note 5 at 43.
184 Raymonde Crte, Limplantation des codes dthique dans le milieu des affaires qubcois et
canadien Quelques rflexions sur les pratiques actuelles in Dveloppements rcents en droit
commercial (Cowansville, Qc.: Yvon Blais, 1998) 135; See Pitt & Groskaufmanis, supra note 179 at
1604-605.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 657
a. Supervised Self-Regulation
To establish a system of supervised self-regulation CRAs would form an
association that would file an application with the securities commission for
recognition as a self-regulatory organization (SRO).185 Once recognized, the SRO
would be responsible for regulating the operations, the standards of practice, and the
business conduct of its members. Membership would be mandatory in order for
agencies to gain access to the status of approved credit rating organization under
the various legal regimes. The SRO would set the membership requirements and
establish rules governing credit rating, which would be binding on the members
through contractual arrangements. In this respect, the SRO could use the IOSCO
Code as a model. The SRO would also have the ability to monitor behaviour of its
members and impose sanctions. The regulatory activities of the SRO would be
subject to continuous oversight by securities commissions.
Supervised self-regulation has developed primarily in the securities industry. The
Securities Act already establishes a framework to deal with SROs.186 In a nutshell, the
Act provides that SROs may seek recognition from the Commission. Where an SRO
is recognized, it becomes subject to the oversight of the Commission. Presently,
recognition is only mandatory for stock exchanges. SROs that are not recognized may
nevertheless set standards of conduct that will apply to their members on a voluntary
basis. These SROs are not subject to the oversight of the securities commission.
From an efficiency perspective, reliance on supervised self-regulation could have
some beneficial impact. Recognition by the SRO could increase the perceived status
of newer CRAs.187 Moreover, supervised self-regulation could provide an additional
mechanism to foster loyalty amongst rating agencies towards issuers and investors,
thereby attenuating agency problems. Recognition would signal to the market that the
SRO considers that the rating agency has proper mechanisms in place to ensure rating
accuracy and integrity. It would also indicate that the agency has accepted to be
subject to the ongoing monitoring and sanctions of the SRO.
Whether these efficiency gains justify the adoption of an SRO model is
debatable. Firstly, the SRO model would imply the adoption of recognition criteria
for CRAs that could rapidly become barriers to entry. For instance, the adoption in
Canada of any of the conditions embedded in the proposed definition of NRSRO
would likely have a negative impact for would be entrants. The SRO model would
also involve higher compliance costs for CRAs than in a pure self-regulation
185 On supervised self-regulation, see Peter Dey & Stanley Makuch, Government Supervision of
Self-Regulatory Organizations in the Canadian Securities Industry in Proposals for a Securities
Market Law for Canada, vol. 3 (Ottawa: Minister of Supply and Services Canada, 1979) 1399;
Johnston & Rockwell, supra note 149 at 279ff; David Mullan & Antonella Ceddia, The Impact on
Public Law of Privatization, Deregulation, Outsourcing, and Downsizing: A Canadian Perspective
(2003) 10 Ind. J. Global Legal Stud. 199 at 216-18; Priest, supra note 174.
186 Securities Act, supra note 124, ss. 21.1-21.11.
187 CESRs Technical Advice, supra note 5 at 45.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
658
regime.188 Although the rules of conduct would be similar under both regimes if the
model of the IOSCO Code were to be followed, supervised self-regulation would
impose mandatory rules of conduct to rating agencies. Each rating agency would
have to follow the rules of conduct enacted, whether or not they are justified and cost-
effective. Furthermore, CRAs would be subject to potentially costly regulatory
oversight absent under the self-regulation model.
[Vol. 51
Second, in order for the efficiency gains to materialize, the SRO would have to
avoid being captured by the large CRAs. There is a risk that the SRO would actually
wind up being accountable primarily to its members. Specifically, the SRO could face
a conflict of interest because of its dual role as regulatory body acting in the public
interest, on the one hand, and as protector and promoter of the private interests of its
members on the other hand. This conflict of interests could extend to the rulemaking
process, by leading the SRO to adopt entry requirements that are in the interest of
incumbents for instance, making it difficult for newer firms to satisfy them. With
respect to enforcement, the potential conflict of interests of the SRO could lead it to
be less severe toward its members.189
In a system of supervised self-regulation, securities commissions would have
oversight powers over the SRO, enabling them to review and approve its bylaws, and
to hear appeals from decisions rendered by the SRO. They would also have the power
to compel an SRO to retain an auditor to conduct compliance reviews. The ability of
the commission to oversee the SRO would provide an accountability mechanism for
the latter. Still, the extent of such supervisory power is limited by the commissions
own experience and expertise, which may prevent them from adequately evaluating
the SROs decisions and initiatives. Furthermore, commissions could themselves be
captured by the SROs members.
In any case, supervision of the SRO by the securities commissions would entail
expenses that would raise administrative costs. When calculating these costs, it is
necessary to take into account the expenses incurred by securities commissions to
oversee the SROs activities. Thus, the savings generated by self-regulation may
rapidly evaporate.
b. Registration Model
Registration is an approach frequently employed to regulate securities market
professionals.190 Under a registration system, rating agencies would have to register
with the competent securities commission. Registration conditions would be
188 Ibid. at 46.
189 See Ren Sorell, Supervision of Self-Regulatory Organizations in Ontarios Securities Market
in Securities Regulations: Issues and Perspectives: Papers Presented at the Queens Annual Business
Law Symposium, 1994 (Scarborough, Ont.: Carswell, 1995) at 178, 182-83; Fisch & Sale, supra note
56 at 1096.
190 Mark R. Gillen, Securities Regulation in Canada, 2d ed. (Toronto: Carswell, 1998) at 432-44.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 659
established in rules of the commission, which would have broad discretion in
deciding whether or not to grant registration. Registration would be subject to an
annual renewal process.
As registrants, CRAs would become subject to regulatory requirements
developed by the securities commission aimed at addressing the concerns raised by
their activities. The commission could establish oversight mechanisms that would
provide it with information on the registrants, and enable it to conduct inspections
and examinations. To sanction non-compliance with regards to requirements, the
commission would have the ability to use its power to reprimand, suspend, cancel or
restrict registration with respect to a particular CRA.
From an efficiency perspective, a registration system could be beneficial for new
or smaller agencies in that it would provide them with regulatory recognition of their
expertise and qualifications.191 By granting them regulatory approval, the system
would provide issuers with more alternatives, thereby boosting competition in the
industry. Moreover, putting securities commissions in charge of the regulation and
supervision of CRAs could enhance their accountability to the public thereby
providing an additional bonding mechanism.
The registration system is not flawless. A first concern is that the beneficial
impact of the system on competition is dependant on the registration system being
finely tuned both in terms of the recognition criteria and the ongoing rules of conduct
it imposes. A flawed framework would create significant barriers to entry for
emerging firms given the mandatory nature of the registration system.192 For instance,
the adoption of the proposed NRSRO recognition criteria could have a negative
impact in Canada.
The enactment of a finely-tuned registration system could prove costly. It would
involve administrative costs in the form of expenses associated with gathering
expertise and drafting of rules. Admittedly, these costs would be lower if the
securities commissions agreed to base their registration system on the IOSCO Code.
Still, the commissions would also have to incur costs to consult industry players
during the rulemaking process. In this respect, proponents of public choice theory
will argue that regulators will have to expand resources in order to ensure that they
are proceeding in the public interest.193 A powerful interest group such as the CRAs
could, to some extent, capture the regulators and impose registration requirements
that restrict entry by new players. Moreover, drafting costs would be magnified by the
necessity to develop a regulatory framework through a joint process involving both
191 CESRs Technical Advice, supra note 5 at 40.
192 Ibid. at 45.
193 Mary G. Condon, Making Disclosure: Ideas and Interests in Ontario Securities Regulation
(Toronto: University of Toronto Press, 1998); Patrick Moyer, The Regulation of Corporate Law by
Securities Regulators: A Comparison of Ontario and the United States (1997) 55 U. T. Fac. L. Rev.
43 at 66.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
660
Canadian and foreign securities commissions, in order to ensure harmonization, with
an uncertain result.
[Vol. 51
A second concern relates to accountability. A government regulator may be less
successful than an SRO in the development of effective ethical norms.194 If this is
indeed the case, registrant compliance with respect to rules of conduct could be
hampered and could increase costs for the regulator to ensure enforcement. Others
question whether the securities commissions have the expertise to review and assess
the procedures used by CRAs to ensure an effective enforcement of the regulatory
framework.195 If expertise is lacking, the rules imposed by the registration system may
have little impact on the conduct of CRAs activities.
D. Strengthening Market-based Solutions Through a Measured
Regulatory Intervention
None of the main regulatory strategies reviewed above are entirely satisfactory.
This does not imply that the status quo is the preferred option. Rather, it suggests that
a more customized approach to regulating CRAs is needed, one that acknowledges
that the market failures exist alongside incentives among … market participants to
ameliorate those failures … 196 In such a setting, regulators would be able to
contribute by assisting the market with limited intervention. Regulators could seek to
enhance the accountability of CRAs by assisting them to bond their credibility toward
investors. As suggested below, a disclosure strategy appears well-suited to attain this
goal.
1. A Disclosure Strategy
Mechanisms
to Reinforce Private Accountability
Regulation could seek to reinforce the effectiveness of reputation by imposing
disclosure obligations on rating agencies. Specifically, regulators should require that
rating agencies disclose their codes of conduct to the public and indicate to what
extent their provisions are consistent with the IOSCO Code. Where the provisions of
a CRAs code of conduct would deviate from the Codes provisions, the agency
would have to explain where and why these deviations exist, and whether such
deviations nevertheless achieve the Codes objectives. In addition, rating agencies
would be required to describe what mechanisms are in place to ensure the
enforcement of the provisions of their codes. Finally, any material changes to the
provisions of an agencys code should be disclosed publicly.
194 See Dey & Makuch, supra note 185 at 1436-37; Joel Seligman, Cautious Evolution or
Perennial Irresolution: Stock Market Self-Regulation During the First Seventy Years of the Securities
and Exchange Commission (2004) 59 Bus. Law. 1347 at 1361-62.
195 See Rhodes, supra note 9 at 358.
196 Choi, Framework, supra note 112 at 71.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 661
Generally speaking, disclosure would be an effective mechanism to assuage
concerns over CRAs lack of accountability.197 Mandatory disclosure about codes of
conduct may have the desired effect of exerting additional disciplinary pressures on
rating agencies. Disclosure would assist investors in ascertaining the reputation of
CRAs by permitting them to identify those that have adequate mechanisms into place
to protect against abuses. Investors assessments would be reflected in overall
appreciation of the credibility of the ratings. Where investors would doubt the
accuracy or independence of ratings, issuers would tend to avoid the latter and seek a
more credible agency or an alternate mechanism to signal their creditworthiness. In
this respect, the disclosure of CRAs codes of conduct could lead other information
intermediaries to rate agencies themselves thereby magnifying reputational
pressures.
This proposal must be considered in light of two important caveats. The first is
that in order for disclosure to reinforce reputational pressures, issuers must have the
ability to use other raters or types of signaling mechanisms. Otherwise, issuers will
remain trapped with the same rating agencies. This caveat underscores the need to
review the regulatory use of ratings in order to allow entry by new or small rating
agencies as well as to ensure the availability of alternate rating mechanisms.
The second caveat involves the impact of the disclosure strategy in itself on
competition. It must be acknowledged that disclosure obligations would have little
positive effect on entry. New rating agencies would derive some reputational gains
from disclosing their codes of conduct in a comparative perspective. Still, in the
absence of a track record, the reputational gains would likely remain marginal since
investors and issuers are more concerned (in the effectiveness of the agency) with the
assessment of credit risk.
2. The Implementation of the Disclosure Strategy
Who should be in charge of regulating credit rating agencies? The functional
approach to regulation advocated by Merton is instructive in complementing this
question.198 This theory proposes regulating financial market participants on the basis
of the economic functions they provide. While different regulatory regimes refer to
ratings, CRAs are essentially informational intermediaries. Applying a functional
approach to regulation, the activities of CRAs should fall under the jurisdiction of
securities commissions. The latter are in charge of ensuring the application of
securities regulation, which purports to reduce information asymmetries between
issuers and investors. Thus, securities commissions have expertise with respect to the
critical function occupied by CRAs. Furthermore, securities commissions already
197 See Warren M.H. Grover & James C. Baillie, Disclosure Requirements in Proposals for a
Securities Market Law for Canada, supra note 185, 349 at 385.
198 Robert C. Merton, A Functional Perspective of Financial Intermediation (1997) 24:2 Financial
Management 23.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
662
regulate or oversee other capital markets informational intermediaries. Granting
jurisdiction to securities commissions would thus improve regulatory efficiency.
Centralizing regulatory oversight in the hand of one type of regulator would also
contribute towards limiting compliance costs.
[Vol. 51
Securities commissions could take formal jurisdiction over rating agencies
through the concept of market participant.199 Specifically, the commissions could use
their rule-making power to adopt a rule providing that any organization performing
securities rating activities should be considered a market participant.200 Each
organization would have to register with the relevant securities commission and
comply with the disclosure regime outlined above. The registration process would not
impose entry requirements. It would serve as a tool for providing securities
commissions with basic information concerning organizations qualified as market
participants. It would also enable the commissions to establish contact with those
organizations subject to the disclosure regime.
In parallel, it is suggested that securities commissions amend their rules and
policies so as to ensure that the concept of approved rating organization refers to any
organization registered with the securities commissions under the proposed rule.
Likewise, federal and provincial authorities referring to specific rating agencies in
their legislation and regulations should consider making similar changes.
The regulatory change proposed would enhance the accountability of rating
agencies. It would grant oversight powers to securities commissions with respect to
CRAs. As market participants, rating agencies would have to keep such books,
records and other documents as are necessary for the proper recording of their
business transactions and financial affairs.201 They would have the obligation to make
these books, records, and documents available to commissions when necessary.
Furthermore, rating agencies would have to keep their codes of conduct on file with
the commissions as well as those documents identifying the codes conformity with
respect to the IOSCO Code. To ensure enforcement of these obligations, the
commissions would have the power to conduct compliance reviews. More
importantly, as market participants, CRAs would be subject to the commissions
power to make decisions in the publics interests.202 Finally, the designation of CRAs
as market participants would reinforce the disciplinary effect flowing from the threat
of additional regulatory interventions.
Second, the proposed regulatory change would remove at least one potential
barrier to entry. Indeed, the designation of specific rating agencies as approved
rating organizations is not without significance. It bolsters the reputation of these
agencies by giving them regulatory imprimatur. In addition, it creates a potential
199 Securities Act, supra note 124, s. 1.
200 Ibid., s. 143(1).
201 Ibid., ss. 19-20.
202 Ibid., s. 127.
2006] S. ROUSSEAU THE ACCOUNTABILITY OF CREDIT RATING AGENCIES 663
barrier to entry for new firms that cannot act as rating agencies so long as they are not
recognized by regulators and so long as the regulation has not been amended. The
fuzziness surrounding the recognition process and criteria exacerbates the problem
facing new entrants.
Nonetheless, the impact of this modification on the credit rating industry should
not be overstated. The NRSRO designation enacted by U.S. regulation and legislation
exercises a strong influence on the Canadian rating agency industry as well. Rating
agencies that have the NRSRO status have considerable clout in Canada.203 Canadian
issuers making debt offerings will tend to prefer dealing with a rating agency that has
the NRSRO status since it will facilitate access to the U.S. market. Nevertheless, the
proposed modification would be relevant for new rating agencies that target a specific
niche and that may not attract issuers seeking to tap the U.S. market.
Would these potential benefits be offset by the costs of this new regime? From
the perspective of rating agencies, the disclosure obligations would not impose
significant compliance costs.204 Rating agencies would have to incur the costs of
drafting a document presenting their codes of conduct in a comparative perspective
with the IOSCO Code. Drafting costs should be relatively low given the scope of the
disclosure obligations. As for the costs of disseminating the document, they should
also remain minimal. Regulation should only require that agencies file the disclosure
document with securities commissions on an annual basis, and that CRAs should only
be required to make this document available to the public in an electronic format,
which would spare them considerable printing costs. Similarly, the obligation to keep
books and records would impose negligible additional expense to rating agencies,
which are already equipped with comparable internal control mechanisms. Thus, it is
unlikely that the compliance costs associated with the production of the proposed
regime would create a barrier to entry for new rating agencies.
The regime would not generate significant administrative costs either. The
disclosure obligations would be enacted in the rule applying to organizations
performing rating activities. In this respect, drafting the rule should not entail
extensive analysis or consultation by regulators. As mentioned previously, the
registration of rating agencies would not involve the enactment of recognition
criteria. The proposed comply or explain approach is already used by securities
regulators with respect to corporate governance disclosure. Furthermore, disclosure
requirements would not be overly extensive or complicated. In terms of enforcement,
the number of interventions by securities commissions would be relatively limited by
the small number of rating agencies, the frequency of filings, and the relatively
neutral character of the information disclosed.
203 See supra note 51 and accompanying text.
204 See Zingales, supra note 159 at 3, 19.
[Vol. 51
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
664
Conclusion
Credit rating agencies play a central role in financial markets. Their views on
creditworthiness
participants.
Creditworthiness influences the conditions under which issuers access debt markets.
It also plays a significant part in investors portfolio decisions. Whether or not they
are considered to be accurate or relevant, ratings shape market participants
conduct.205 Put plainly, CRAs are economic agents who wield power over issuers and
investors.
for market
implications
have
normative
In a perfectly functioning market, the fact that CRAs have such power would not
be a cause for concern. Specifically, issuers and investors would have the means to
ensure that CRAs interests are aligned with their own. As seen above, the real world
departs from this ideal, and failures in the market may lead to a disconnect between
the interests of CRAs on the one hand, and those of issuers and investors on the other.
The review of the legal and institutional environment indicates that there is a dearth
of mechanisms designed to ensure accountability. In fact, it would appear that
reputation is the primary, if not the only, mechanism that acts to restrain opportunistic
behaviour on the part of rating agencies. Thus, there is a potential accountability gap,
that is, an imbalance between CRAs vast power and the likelihood of holding them
responsible for their use of this power.
This accountability gap has preoccupied regulators ever since the recent wave of
corporate scandals. Although studies conducted revealed no wrongdoing on the part
of CRAs, they warned of potential problems that could result if reputation were to fail
in reigning them in. As a result, regulators have examined possible methods of
enhancing the accountability of rating agencies: the IOSCO published the Code of
Conduct Fundamentals for Credit Rating Agencies, and the SEC proposed a rule that
would define the conditions an entity must satisfy in order to obtain an NRSRO
designation. While both these initiatives deal with similar issues, the IOSCO Code is
better suited to implementation in national regulations than the SEC rule, which is
idiosyncratic to the American system. To the extent that the IOSCO Code sets forth
principles that should be adopted in order to shape the conduct of CRAs, the question
of how it should be implemented by Canadian regulators remains.
This paper argues that the implementation of the IOSCO Code should be done
through a comply or explain disclosure strategy. Securities commissions would be
charged with overseeing the disclosure strategy and would retain jurisdiction over
CRAs by assigning them the status of market participant. The proposed approach
would contribute to enhancing the accountability of CRAs not only by reinforcing
reputational pressures that guard against opportunism, but also by introducing an
additional level of regulatory supervision over them that could hold them responsible
for such behaviour.
205 See Sinclair, supra note 1 at 52.
