Separate copy of article from Perspectives on Credit ...14 See John P. Hunt, The SEC’s proposed Rating Agency Rules: Unresolved Conflicts, Ber- keley Center for Law, Business and

1

Separate copy of article from

Perspectives on Credit Rating Agencies

Conflicts of Interest and Risk Management Practices in the CRA IndustryROLF H. WEBER & SIMONE BAUMANN

Stockholm Centre for Commercial Law

Juridiska fakulteten

2

Conflicts of Interest and Risk Management Practices in the CRA Industry

283


ROLF H. WEBER & SIMONE BAUMANN

I. IntroductionCredit rating agencies (CRA) exercise different functions; two main purposesof their activities can be identified:1 (i) As special financial “journalists” CRAsupport investors in their market transactions by exercising an informationfunction. This allows investors to overcome informational asymmetries andthus CRA reduce information costs in the market place. (ii) CRA also assumea certification function since they serve as a regulatory tool in financial mar-ket oversight.2

These “main” functions, exercised for different market players, are obvi-ously tied to the risk of conflicting interests. Drastically, it has been said:“The truth is [CRA] are working for Wall Street, and they are going to giveto Wall Street what Wall Street wants”.3 Even if an assessment is not made insuch a clear statement, academics do not assume that conflicts of interest canbe easily avoided in the CRA industry.4 In fact, the rating process is not

1 Giulia Rognoni, Credit Rating Agencies. A Look into Conflicts of Interest, Saarbrücken2011, p. 23.

2 For a general overview see Aline Darbellay, Regulating Rating. The Credit Rating AgencyOligopoly from a Regulatory Perspective, Zurich 2011, pp. 24 et seq.

3 Armen Keteyian, Ratings Agencies to Face Grilling on Goldman (quoting Senator Sand-ers, cited by Darbellay, supra note 2, p.123).

4 Apart from the mentioned study of Rognoni, supra note 1, see Lynn Bai, On RegulatingConflict of Interests in the Credit Rating Industry, University of Cincinnati College ofLaw, Public Law & Legal Theory, Research Paper Series, No. 10-17, 2010, pp. 1 et seq.;Tobias Johansson, Regulating credit rating agencies: The issue of conflicts of interest in therating of structured finance products, Journal of Banking Regulation, Vol. 12, 2010, pp. 1et seq.; Franklin Strier, Rating the Raters: Conflicts of Interest in the Credit Rating Firms,Business and Society Review, Vol. 113, 2008, pp. 533 et seq.; Patrick Bolton/XavierFreixas/Joel Shapiro, The Credit Rating Game, The Journal of Finance, Vol. 67, 2012,pp. 85 et seq.; Aline Darbellay/Frank Partnoy, Credit Rating Agencies and Regulatory

Rolf H. Weber & Simone Baumann

284

exclusively based on hard data, but also includes “qualitative” judgments.5 The risk of having a conflict of interest situation does not appear solely in

the CRA markets. In fact, conflicts of interest generally tend to appear inbusiness life.6 Therefore, subsequently the specific risk situations in the CRAindustry are analyzed and the coherence of the existing rules and regulationswith overarching principles combating conflicts of interest risks are dis-cussed.

II. Conflicts of interest Conflicts of interest can occur on the analyst level (individual level) and onthe CRA level (corporate level).7 Links between the two levels exist; however,the risks for biased behavioral activities are not identical.

A. Analyst level

Looking at the analyst level, mainly five different situations of conflicts ofinterest are to be taken into account, relating to the following activities:

1. Ownership of securities of rated entities

If an analyst and employee of a CRA owns securities or money market instru-ments of issuers or debtors who request a credit rating from the respectiveCRA the risk occurs that ratings are issued which could potentially favor thetrading positions of such person. Even indirect attempts to influence anunduly positive rating (by inducing a colleague to certain assessments) can-not be excluded.8

5 See also Rognoni, supra note 1, pp. 28/29.6 See e.g. for conflict of interest in governance: Anne Peters/Lukas Handschin (eds.), Con-

flict of Interest in Global, Public and Corporate Governance, 2012; Marc A. Rodwin,Conflicts of interest and the future of medicine: the United States, France and Japan,2011; Luc Thévenoz (et al.), Conflicts of interest: Corporate Governance and financialMarkets, Zurich 2007.

7 The following factual description is based on the more detailed explanations to the creditanalyst level of Bai, supra note 4, pp. 7 et seq.

8 See also Bai, supra note 4, pp. 7/8 and Johansson, supra note 4, p. 6.

Reform, University of San Diego Law School, Legal Studies Research Paper Series, No.12-083, 2012, pp. 1, 4/5.


285

2. Employment position or directorship at a rated entity

A conflict of interest is obvious if an analyst of the CRA also has an employ-ment position or a directorship at a rated entity. The risks for biased ratingsis stating obvious that nowadays such kind of double function is obviouslyconsidered as inappropriate.9

3. Business relation beyond the course of business or special personal relationship

A conflict of interest also exists if special business relations exist, for exampleif an analyst borrows money at preferential terms from a rated entity or if spe-cial personal relations between the analyst and an influential employee of therated agency are established.

4. Receipt of gifts from rated entities

Obvious risks exist if an analyst receives gifts from rated entities exceeding an“adequate” value leading to a loss of impartiality in the rating activities.

5. Determination of analysts’ compensation based on rating fees

If analysts are compensated based on the fees they generate from rated enti-ties, an incentive exists to issue high ratings in order to attract more custom-ers; consequently, the rating process might become biased.

B. CRA level

The conflicts of interest at the CRA level10 seem to be less obvious than atthe analyst level, however, in practice more important and more critical. Tworeasons can be mentioned for this assessment: On the one hand, conflicts ofinterest at the analyst level are quite easily detectable and the CRA as such isinterested to exercise some control in this respect because it needs to safe-guard its reputation (reputation hypothesis)11; on the other hand, rules

9 In the past, the sensitivity of the problem was not recognized: Clifford L. Alexander Jr.,former chairman of Moody’s, served on the board of WorldCom/MCI (receiving ratingsfrom Moody’s) for nineteen years.

10 The following description is based on the more detailed explanations of Bai, supra note4, pp. 9 et seq.

11 See e.g. Steven L. Schwarcz, who stated already at an early point of the CRA discourse, that


286

related to conflicts of interest at the analyst level are less difficult to imple-ment. Since the problems at the CRA level are business-inherent, increasedattention must be paid to the measures which potentially can avoid conflictsof interest.

1. Affiliated underwriter or issuer

A conflict of interest situation is given if a CRA is issuing a rating related tosecurities or market instruments underwritten by an affiliate (as broker ordealer).

2. Ancillary services to rated entities

CRA often provide additional services to rated entities, for example consul-tancy services. The character of this services rendered is usually more or lessancillary to the rating activity.12 Looking from the angle of the CRA theancillary services are a further element of income generating. Therefore, theCRA is interested not to lose the rated entity as customer. Consequently, theCRA might tend to issue a positive rating. Since the wish to increase theincome through ancillary services influences the rating determination proc-ess, the conflict of interest is inherent to this business model.

3. Large subscriber influence

Large rated entities needing several ratings from a CRA might try to use theirinfluence on the activities of the CRA by linking the “results” of the CRAprocess to the continuation of the business relationship. Similarly as in caseof the ancillary services, such situation might induce the CRA to issue“friendly” ratings in favor of the rated entity in order not to risk the contin-uation of the business relationship.

12 For further details see Darbellay, supra note 2, pp. 125/26 and Bai, supra note 4, pp. 9–11.

there is a de facto accountability of rating agencies through reputation: Steven L.Schwarcz, Private Ordering of Public Markets: The Rating Agency Paradox, University ofIllinois Law Review, Vol. 2002, No. 1, p. 15; see also Daniel M. Covitz/Paul Harrison,Testing Conflicts of Interest at Bond Rating Agencies with Marketanticipation: Evidencethat Reputation Incentives Dominate, FEDS Working Paper No. 2003-68, pp. 1–25, 23.


287

4. Issuer-pay business model

The issuer-pay business model is the most critical pillar of the CRA activ-ity:13 Ratings are usually done in conjunction with a securities issue or withthe presentation of financial statements of the rated entity. Since ratings areonly paid at the moment when the rating has been made available, CRA aretempted not to risk the affection of the payment (income) by a rating whichdissatisfies the rated entity. Market incentives to prevent from such tempta-tion are weak.14 Even if no obvious large subscriber influence and no specificaffiliation between CRA and rated entity is given, a certain inherent conflictof interest exists due to the issuer-pay business model.

There has been an intense discourse about risks inherent to this businessmodel.15 The main “defense” argument brought forward by the CRA indus-try is the so-called reputation hypothesis saying that CRA could not affordto issue ratings in a way that pleases the rated entity if the given facts wouldnot be adequately mirrored in the rating since an unduly positive rating turn-ing out to be wrong could endanger the reputation of the CRA in the longrun.16

III. Risk management practices A. Corporate governance

As other entities in the financial markets, CRA have to introduce and complywith basic corporate governance principles, as corporate governance adopts amechanism that helps identifying potential conflicts of interest17. Obviously,some principles such as the application of due diligence or the acting in goodfaith are to be generally recognized.18 The infamous Enron case exemplifies

13 John Patrick Hunt, Credit Rating Agencies and the “Worldwide Credit Crisis”: The Lim-its of Reputation, the Insufficiency of Reform, and a Proposal for Improvement, Colum-bia Business Law Review, 2009, pp. 109–209, p. 155; for further details see Darbellay,supra note 2, pp. 123/24; Bai, supra note 4, pp. 12/13; Johansson, supra note 4, p. 5;Strier, supra note 4, p. 537; Lawrence J. White, The Credit Rating Agencies, The Journalof Economic Perspectives, Vol. 24, 2010, pp. 211–226, p. 214 et seq.

14 See John P. Hunt, The SEC’s proposed Rating Agency Rules: Unresolved Conflicts, Ber-keley Center for Law, Business and Economy White Paper, 2008, pp. 1–5, p. 3.

15 See Rognoni, supra note 1, pp. 14, 21 et seq. with further references. 16 See below sub-section III.C.17 Strier, supra note 4, p. 539.18 The Dodd-Frank Wall Street and Consumer Protection Act for instance also regulates

third party due diligence services (Section 932 (s)(3)(A)(v)).


288

this very well: questions arose about the agencies’ lack of care and diligencein their work, as only four days before Enron declared bankruptcy, majorCRA still rated its debt “investment grade”.19 As far as the CRA are con-cerned, however, additional, more specific elements need to be taken intoaccount.

1. Code of conduct

The risk of conflicts of interest is inherent to the CRA industry. Conse-quently, corporate governance particularly requires addressing the respectiverisks in a code of conduct. The need of such principles has also beenacknowledged by the IOSCO in a Code of Conduct Fundamentals forCRA20. According to IOSCO, CRA (among other market participants)should strive toward high-level objectives that are implemented in individualcodes of conduct so that investor protection, efficiency and transparency areguaranteed and systemic risk is reduced. Irrespective of the IOSCO recom-mendations concerning the CRA independence and avoidance of conflicts ofinterest21 possible rules of such a code could be:

• The functions in a CRA are to be separated in a way that the individualanalysts would not have to follow different duties that potentially lead toa conflict of interest.

• The income of analysts should not depend on the fee generation fromrated entities.

• The acceptance of gifts from the rated entity must be excluded.• In case of personal relations between an analyst and the rated entity the

respective person should not act in this mandate anymore. • Quality control measures have to be implemented.• “Chinese walls” should try to avoid the exchange of sensitive information

between analysts working in different mandates.

19 Arthur R. Pinto, Control and Responsibility of Credit Rating Agencies in the UnitedStates, the American Journal of Comparative Law, Vol. 54, 2006, pp. 341–365, p. 345;Claire A. Hill, Regulating the Rating Agencies, Washington University Law Review, Vol.82, 2004, pp. 43–95, p. 43.

20 IOSCO, Code of Conduct Fundamentals for Credit Rating Agencies, The TechnicalCommittee of the International Organisation of Securities Commissions, December2004, revised in May 2008.

21 IOSCO, supra note 20, 7.


289

2. Potential deficits

Codes are not immune to shortcomings, however. As potential vulnerabili-ties to failures from a legal perspective22 two major elements can be identi-fied:

• The code of conduct does not contain those rules of behavior which aresuitable to combat conflicts of interest situations.

• The code of conduct is not properly implemented and in particular notstrongly supervised; in case of lack of appropriate control, the best ruleson paper do not achieve the desired results.

Non-compliance with legal rules of whatever nature can lead to a corre-sponding responsibility and (financial) liability of the CRA.23

B. Compliance manuals and procedures

Codes of conduct are applied often on a business association level, whereascompliance manuals are binding for employees on the individual companylevel. The implementation of compliance manuals and procedures has be-come a “random” tool in many industries.24 These instruments are to be ap-plied also in the CRA business.

An example for the need of implementing strict compliance manuals isthe antitrust segment; even if the contents of the rules are different in theCRA business, analogies can be drawn: It is a primary issue of the antitrustfield to deal with the risk of undesired behavior or even willful violations ofthe law of board or staff members, too25. Companies assign their legal depart-

22 From the economic perspective see Strier, supra note 4, pp. 539/40.23 The liability issues are discussed in detail in this volume in the contribution of Vibe

Ulfbeck, pp. 309 et seq.24 See e.g. The Coca Cola Company’s anti-corruption compliance program: http://

www.thecoca-colacompany.com/citizenship/anticorruption.html; the Johnson & John-son Healthcare Compliance guidelines for employees: http://www.jnj.com/wps/wcm/connect/13445d804f5568a4a06ea41bb31559c7/the-right-course.pdf?MOD=AJPERES;Novartis Pharmaceuticals Corporation, Compliance: http://www.pharma.us.novar-tis.com/info/corporate-responsibility/business-conduct/compliance.jsp; IKEA Compli-ance Standard IWAY: http://www.ikea.com/ms/en_US/about_ikea/pdf/SCGlobal_IWA-YSTDVers4.pdf.

25 See Rosa M. Abrantes-Metz/Patrick Bajar/Joseph E. Murphy, Antitrust Screening: Mak-ing Compliance Programs Robust, p. 3, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1648948.


290

ments with the task of establishing coherent compliance rules, so thatemployees can refer to the manual and are able to identify critical situations.As it is with CRA, companies in general do not want to risk losing publictrust because of an antitrust offence committed by an employee either.

C. Reputation hypothesis

As mentioned, the reputation hypothesis is brought forward by the CRA indefending their business model by arguing that non-compliance with stand-ards of diligence in the issuing of ratings for what ever reasons would lead toreputational sanctions and to the loss of business in the long run.26 In otherwords, the issuer-pay business model is not considered to be problematicbecause the CRA would have much more to lose by endangering their repu-tation for objectivity than they would have to gain if they favor one singlecostumer.27 Consequently, the adherence of the “reputational capital” makesbelieve that market discipline provides the right incentives to produce highquality ratings.28

In the meantime many scholars29 question whether the aspect of the rep-utation is strong enough to avoid that conflicts of interest would lead to“wrong” ratings30. A first problem has to be seen in the fact that the reputa-tion is based on the perceptions of others and is only indirectly a consequenceof the CRA’s behavior.31 Furthermore, income in the CRA industry is oftenearned through a “fly-by-night” strategy and inaccurate ratings would onlybe discovered after a certain time delay.32 Rating laxity also does not seem tosupport the reputation hypothesis, apart from aspects of pure modeling andpure disclosure.33

Many studies are already available showing that the status of CRA can bedependent on the quality of the analysis to a certain extent, but that the rat-ing market is not flawless and information is not perfectly distributed to all

26 Bai, supra note 4, pp. 14/15.27 Schwarcz, supra note 11, p. 15; Rognoni, supra note 1, p. 33. 28 Darbellay/Partnoy, supra note 4, p. 3; Covitz/Harrison, supra note 11, p. 23; Rognoni,

supra note 1, p. 34. 29 For a sophisticated model see Bolton/Freixas/Shapiro, supra note 4, pp. 91 et seq. 30 Hunt, supra note 13, p. 156 et seq.31 Rognoni, supra note 1, p. 35.32 See also OECD, Competition and Credit Rating Agencies, 5 October 2010, DAF/

COMP(2010)29, p. 48; Rognoni, supra note 1, p. 36.33 Rognoni, supra note 1, pp. 39, 42/43.


291

market participants.34 Since it is difficult to establish what an accurate ratingwould be, the market is also not particularly sensitive to CRA’s failure to pro-duce precise ratings. Furthermore, it has to be taken into account that thenumber of CRA is very limited in the market.35 There has been no real com-petition because the two major companies, S&P and Moody’s, have domi-nated the CRA-business for decades. This also tends to the result that infor-mation available to market participants is limited.36 Therefore, it can beargued that CRA know that a reduction in the quality of their services willnot be followed by a proportionate drop in their profits.37 Consequently, themarket as such does not sufficiently regulate the conflicts of interest problemmeaning that the need for rules of whatever nature is given.

IV. Legislative responses Five years ago, an analyst said that the investment “could be structured bycows and we would rate it”.38 This statement evidences the necessity of a legalframework. This need is independent from possible economic models tryingto avoid the most critical situations of conflicts of interest.

A. Improvement of disclosure

Transparency is an important aspect in financial market regulation. Alreadysome hundred years ago, Justice Brandeis stated that sunlight is said to be thebest of all disinfectants.39 The recent financial crisis has clearly illustrated,that the lack of information played an essential role. Highly rated assets sud-denly had to be downgraded continually40. In the CRA industry, the follow-

34 See the models discussed by Rognoni, supra note 1, pp. 51 et seq. and 67/68.35 The problem of the oligopolistic structure of the CRA industry is discussed in more detail

in this volume by Aline Darbellay in her contribution, pp. 159 et seq.36 Hunt, supra note 13, pp. 167 et seq.37 Rognoni, supra note 1, p. 38. 38 Analyst at one of the main credit rating agencies in an e-mail referring to structure finance

products, April 5, 2007, Securities and Exchange Commission, Summary Report of IssuesIdentified in the Commission Staff ’s Examinations of Select Credit Rating Agencies,2008, p. 12, cited by Bolton/Freixas/Shapiro, supra note 4, p. 85.

39 Louis D. Brandeis, What Publicity Can Do, in: Other People’s Money: And How theBankers Use It, New York 1914, p. 92.

40 Xavier Freixas/Christian Laux, Disclosure, Transparency, and Market Discipline, CFSWorking Paper, No. 2011/11, p. 12.


292

ing transparency requirements are of importance:41

• Disclosure of other services rendered by CRA to the issuer: As mentioned,42

ancillary services can influence the rating process since the CRA is inter-ested in generating additional income; therefore, transparency in respectof these elements is essential.

• Disclosure of default rates for each rating grade: The transparency related todefault rates for given grades may have a preventive effect insofar as itshould improve the accuracy of the ratings; furthermore, ideally the mar-ket would punish errant ratings, whether born of negligence or conflictsof interest.

• Caveats on ratings: As in other market segments specific caveats couldattract the attention of the users by explaining the weaknesses of and theproblems with ratings; relevant reference points could be generallyaccepted standards or reservations in case of specific difficulties.

• Description of rating procedures and methodologies: If an investor is familiarwith the procedures a CRA applies in the rating process, the rating gradebecomes more comprehensible for the investor. Thus, she/he will be ableto better anticipate occurring risks43.

Since the CRA are information intermediaries, they are supposed to addressinformation asymmetries. In a competitive market for information, a CRAthat fails to provide the investors with valuable information would not sur-vive.44 In contrast, as has been shown above, there is limited competition inthe CRA market.45 Therefore, the incentive to provide full disclosure mustbe rooted in a type of binding understanding.

B. Change of CRA-client relationship

Another track for regulation could be to change the CRA-client relationship,in particular by replacing the issuer-pay ratings by another system.46 Obvi-

41 The following description follows Strier, supra note 4, pp. 544/45.42 See above sub-section II.B.2.43 See in this context also: Securities and Exchange Commission, Oversight of Credit Rating

Agencies Registered as Nationally Recognized Statistical Rating Organizations, FinalRule, 18 June 2007, p. 91.

44 See Darbellay, supra note 2, p. 95.45 See above, III.C.46 See also Strier, supra note 4, pp. 545/46.


293

ously, in such a case another obligor for the payments needs to be found. Afirst alternative would be the addressees of the ratings. However, this solutionis hardly practicable since the addressees (i.e. subscribers, investors) are verynumerous and the willingness to pay might be very low, apart from themoney collection and free rider problems.

Another source could be public bodies that issue the ratings, such asnation states or international organizations. In this scenario, the ratingswould have to be considered as public goods; from a theoretical perspective,however, it seems to be hardly justifiable that the financing of CRA shouldbe a public task.

Finally, it could be considered to establish foundations or similar bodiesthat would be financed by sponsors; thereafter, these foundations could makedecisions on the establishment, operation and financing of a CRA. Since themanagement of such CRA would “only” be accountable to the not-for-profitorganization “behind” the CRA, more independence of the ratings could beexpected.47

C. Change of rating system

Several options exist that could help mending the rating system. Amongstothers, the following elements should be considered:48 (i) Establish more inde-pendence from rated companies: Conflicts of interests can be avoided or at leastmitigated if the independence of CRA from issuers is regulated in strictterms; this approach is the present tendency of CRA regulations.49 (ii) Pro-hibit the involvement of CRA in creating new investment products that the CRAis rating: Such kind of provision which intends to clearly distinguish betweenconsultancy services and rating services could help to avoid that the rating isinfluenced by the fee-generating business of the CRA.50 (iii) Hiring of CRA

47 See Roland Berger Strategy Consultants, Blueprint for an European Rating Agency, Octo-ber 2011, http://www.theproject2012.org/wp-content/uploads/2011/10/BLUEPRINT-EU-RATING-AGENCY.pdf; see also the proposal of the Bertelsmann Foundation, Blue-print for INCRA: an International Non-Profit Credit Rating Agency, http://www.bfna.org/sites/default/files/publications/INCRA%20Report.pdf.

48 The following comments structurally deviate from the description given by Strier, supranote 4, pp. 546 et seq.

49 See Hill, supra note 19, p. 75.50 The Technical Committee of IOSCO has discussed this possible advancement: IOSCO,

Media Release, IOSCO addresses the subprime crisis, IOSCO/MR/002/2008, 6 Febru-ary 2008, p. 1.


294

by independent parties: By doing this, one can avoid that the issuer who isinterested in the rating does have an influence on the choice of the respectiveCRA; as appointment body, an independent private organization or thefinancial market supervisory authority could be taken into account; (iv)Inhibit the use of letter grades in setting standards: Instead of using the presentA, AA, B, etc. model it seems to be worthwhile to consider introducing acompletely new rating system. For example, ratings tools would have to bepurchased from the CRA and the tasks could then be performed by the con-cerned entities themselves.51 This would allow to make sure that the infor-mation is processed and absorbed by an investor in detail and that investorsdo not only rely on a seemingly clear assessment of A, AA, etc.

D. More direct regulation

Several instruments being suitable to concretely intervene into the CRAbusiness could be taken into account as models for more direct regulation.52

For instance, there would be the option of exercising greater oversight: Finan-cial markets supervisory authorities could more thoroughly supervise theactivities of the CRA, set more detailed standards and enforce the regulationsstrictly53.

Also, governments should be allowed to implement carrots and sticks:Regulations could increase the accountability obligations of CRA by settinghigher corporate governance standards. Non-compliance with such stand-ards would lead to responsibility and liability for malfeasance. Governmentalrules also have the possibility to define the level of diligence standards andthereby indirectly the level of risk in case of negligent behavior54.

V. Existing regulations Prior to the outbreak of the financial crisis in 2007, not many regulationsgoverning the behavior of CRA did exist. Financial market supervisoryauthorities believed in reputation mechanisms and market forces. Thisappreciation changed five years ago and has been replaced by hectic efforts to

51 For further details see Strier, supra note 4, p. 548.52 See also Strier, supra note 4, pp. 549/50.53 Supervisiory authorities have been paying more attention to the problem of too lax regu-

lation since the financial crisis, see e.g. SEC, supra note 43.54 The IOSCO Code of Conduct could serve as an archetype, see supra note 20.


295

find appropriate regulations. Hereinafter, the IOSCO Code of Conduct, theEU regulations as well as the US regulations will be discussed.

A. IOSCO Code of conduct

The International Organization of Securities Commissions (IOSCO) hadalready started the efforts of preparing a Code of Conduct governing theCRA in 2003; a detailed draft of a Code of Conduct was issued in 2004, butonly the financial crisis has led to the adoption of the Code of Conduct inMay 2008.55 In a report issued in February 2011 the Technical Committeeof the IOSCO addressed certain recent regulatory initiatives that concernCRA in multiple jurisdictions.56

As other OECD instruments, the Code of Conduct is not legally binding,but OECD Member States are committed to a voluntary compliance inaccordance with the principle “comply or explain”. The major principles ofthe Code of Conduct have been acknowledged by most countries controllinga substantial financial market.57

1. General requirements

The Code of Conduct generally requires CRA to ensure by preventive mea-sures that no conflicts of interest arise (for example by releasing proper organ-izational rules) and by providing for rules which mitigate the effects of anypotential conflicts of interest (for example by disclosing them to the mar-ket).58

Consequently, the provisions of the Code of Conduct are of a preventiveand of a prohibitive nature.

2. Requirements related to employee independence

The Code of Conduct prohibits an analyst from conducting any securities orderivatives dealing that can constitute a conflict of interest as well as fromsoliciting or accepting gifts from a rated entity. The rules are quite detailed

55 See IOSCO, Code of Conduct, supra note 20; Johansson, supra note 4, p. 7.56 IOSCO, Final Report, Regulatory Implementation of the Statement of Principles Regard-

ing the Activities of Credit Rating Agencies, February 2011; see also Roberta S. Karmel,IOSCO’s Response to the Financial Crisis, Research Paper No. 268, 2010, available at:http://ssrn.com/abstract=2025115.

57 Johansson, supra note 4, p. 7.58 IOSCO, Code of Conduct, supra note 20, p. 7, note 2.1; Johansson, supra note 4, p. 7.


296

and should exclude the practically most frequently occurring conflicts ofinterest.59

Furthermore, preventive rules request from the CRA to establish com-pensation arrangements that limit the risk for conflicts of interest. In addi-tion, salaries should not depend on the generated business and internal pro-cedures are to be implemented so that the compliance with the rules can besurveyed.60

3. Requirements related to CRA independence

Prohibitive provisions of the Code of Conduct contain requirements that abusiness relationship with the rated entity must not influence the rating. Italso states that securities and derivatives are not to be dealt by the CRA if itresults in a conflict of interest.61

The preventive rules of the Code of Conduct provide for requirements inrespect of the organizational structure of the CRA. Furthermore, disclosurerequirements should make potential conflicts of interest transparent at anearly stage.62

B. EU regulations

Prior to the financial crisis of 2007 the European Commission consideredthe IOSCO Code of Conduct in conjunction with a yearly assessment ofCRA activities by the Committee of European Securities Regulators (CESR)as sufficient. Only in October 2007, the so-called “ECOFIN Roadmap” wasadopted which required to conduct an oversight of the role of CRA63. Thismandate was given to the Commission, the CESR and the European Securi-ties Market Expert Group.64 The European Parliament and the Councilhave approved a proposed Regulation on CRA on 23 April 2009; the regu-lation has been adopted on 16 September 2009 and amended on 11 May

59 IOSCO, Code of Conduct, supra note 20, p. 8, note 2.13; Johansson, supra note 4, p. 8.60 IOSCO, Code of Conduct, supra note 20, p. 8, note 2.8; Johansson, supra note 4, p. 8.61 IOSCO, Code of Conduct, supra note 20, p. 8, note 2.9.62 IOSCO, Code of Conduct, supra note 20, p. 9, note 2.7.63 Council of the European Union, Press Release 13571/07 (Presse 217), Economic and

Financial Affairs, 9 October 2007, http://www.consilium.europa.eu/uedocs/cms_data/docs/pressdata/en/ecofin/96375.pdf.

64 See Johansson, supra note 4, pp. 8/9; Rognoni, supra note 1, pp. 74/75.


297

2011.65 Meanwhile, a Commission delegated regulation has been publishedabout the first regulatory technical standards on CRA.66

The new EU rules introduced a registration process of CRA with theCESR as the coordinating authority. This registration is one of the principalprerequisites for CRA in order to be allowed to issue credit ratings.67

The registration process is regulated in a detailed way on the basis of atwo-level system: The application must be sent to the CESR but the actualreview and final approval is done by the national supervisor of the concernedmember state, the CESR may only ask for a reconsideration which, however,is not mandatory.68

This fragmented system is not fully appropriate in global markets. There-fore, the de Larosière Group has proposed a general strengthening of the so-called level 3 committees.69 The report identified severely limited resourcesavailable and suggested a significant increase in financial means to enable thelevel 3 committees to fulfill their heavy workload.70

1. General requirements

The rules related to the avoidance of conflicts of interest are quite detailedand mainly of preventive nature. The CRA are required to establish a super-visory board being mandated to ensure the independence of the rating pro-cess and the avoidance of conflicts of interest.71

The most recent amendment of the Regulation EC/1060/2009 high-lights the importance of transparency of information underlying the ratings

65 See EU press release IP/09/629; Regulation (EC) No 1060/2009 of the European Parlia-ment and of the Council of 16 September 2009 on credit rating agencies; Regulation(EU) No 513/2011 of the European Parliament and of the Council of 11 May 2011amending Regulation EC/1060/2009.

66 Commission delegated Regulation EU/447/2012 of 21 March 2012 supplementing Reg-ulation (EC) No. 1060/2009 of the European Parliament and of the Council on creditrating agencies by laying down regulatory technical standards for the assessment of com-pliance of credit rating methodologies.

67 Art. 14 et seq. Regulation EC/1060/2009; Johansson, supra note 4, p. 9.68 Art. 16 (7) Regulation EC/1060/2009; Johansson, supra note 4, p. 9.69 The high-level group on financial supervision in the EU, chaired by Jacques de Larosière,

Report of 25 February 2009, http://ec.europa.eu/internal_market/finances/docs/de_larosiere_report_en.pdf; Johansson, supra note 4, p. 9.

70 De Larosière report, supra note 69, recommendation no. 161.71 Johansson, supra note 4, p. 9.


298

of all financial instruments. The Regulation expects that this will lead tohigher competition and enhancement of the quality of the ratings.72

2. Requirements related to employee independence

The CRA Regulation is quite extensive in respect of the requirements ofemployee independence. An analyst may not be involved in the rating of thesame rated company for more than a few years. Furthermore, an analyst maynot assume a key management position at a rated entity within a period ofsix months from the time of the rating.73

The amended CRA Regulation substantiates in point 19 of Annex III (I.)the duties of a CRA in regard of independence and avoidance of conflicts ofinterest.74 It is considered an infringement of said article 6 (2) of the Regu-lation if the CRA “is not identifying, eliminating or managing and disclos-ing, clearly or prominently, any actual or potential conflicts of interest thatmay influence the analyses or judgments of its rating analysts, employees, orany other natural person” that are assigned with issuing or approving creditratings.

3. Requirements related to CRA independence

The CRA Regulation encompasses quite strict rules on disclosure, for exam-ple related to the executed activities, to the important clients, to rated entitiesgenerating more than five 5% of the CRA’s revenues.75

The amended CRA Regulation demands that a compliance functiondepartment be established which operates independently.76 Besides, the CRAis obliged to implement an appropriate and effective organization or admin-istration. The company must be capable of identifying, eliminating or man-aging and disclosing any conflicts of interest as well as recording significantthreats to the independence of the credit rating activities.77

72 Regulation EC/513/2011, supra note 65, Consideration (7); see also ibid. Annex III (I.)point 4.

73 Regulation EC/1060/2009, supra note 65, Consideration (26, 33), Art. 6; Johansson,supra note 4, p. 9.

74 See Regulation EC/1060/2009, supra note 65, Art. 6 (2); Regulation EC/813/2011,Annex III (I.).

75 Johansson, supra note 4, p. 9.76 Regulation EC/513/2012, supra note 65, Annex III (I.) point 13.77 Regulation EC/513/2012, supra note 65, Annex III (I.) point 15.


299

C. US regulations

The CRA are regulated and supervised in the United States since 200678.The regulations encompass registration requirements for those CRA whoseratings are used for regulatory purposes. Approved CRA are called “NationalRecognized Statistical Rating Organizations” (NRSRO).79 Before, CRAhave defended themselves successfully when facing legal proceedings as theyargued that their ratings were opinions under the First Amendment of theUS Constitution.80

The implementation of the Dodd-Frank Act of 2010 has increased thepowers of the supervisory authority SEC81. The SEC has implemented sev-eral organizational changes since, including the establishment of the Officeof Credit Ratings as a new specialist group within SEC supervising CRA.82

The most recent revision of the regulation of CRA aims for an enhancementof disclosure to investors and the integrity of the rating process. It addressesthe conflict of interest and highlights the importance of transparency andaccountability of CRA.83 In addition, the SEC has compiled a proposal fornew rules that would establish professional standards for credit analysts andrules regarding ratings symbols.84

1. General regulation

Prior to the Dodd-Frank Act the foundation of the rules on conflicts of inter-est were already stated in the Securities Exchange Act 1934: Section 15E(h)(1) states that nationally recognized statistical rating organizations“should establish, maintain, and enforce written policies and procedures rea-

78 Following the Credit Rating Agency Reform Act of 2006, enacted 29 September 2006;see Section 2, Findings (6).

79 For further details see Bai, supra note 4, pp. 41 et seq.; Johansson, supra note 4, p. 9. 80 See Darbellay, supra note 2, p. 68; Marilyn B. Cane/Adam Shamir/Tomas Jodar, Below

Investment Grade and above the Law: A Past, Present and Future Look at the Accounta-bility of Credit Rating Agencies, in: NSU Shepard Broad Law Center, Research Paper No.11-005, 2011, p. 24.

81 Darbellay, supra note 2, p. 61.82 Office of Credit Ratings, see: http://www.sec.gov/about/offices/ocr.shtml.83 Darbellay/Partnoy, supra note 4, p. 11; see also the proposals submitted by Bai, supra note

4, pp. 49 et seq; Rognoni, supra note 1, p. 77.84 See Pending Action Chart on SEC website: http://www.sec.gov/spotlight/dodd-frank/

dfactivity-upcoming.shtml.


300

sonably designed … to address and manage any conflicts of interest”. Section15E (h)(2) empowers the SEC to issue final rules on the basis of this generalrequirement (in accordance with Section 15E (n)(1)(A)).85

2. Regulation concerning employee independence

The prohibitive and preventive rules concerning the analysts are ratherdetailed: Direct ownership interest in the rated entity is forbidden; gifts of anaggregate value of more than USD 25 may not be accepted86. Situations inwhich the independence of an analyst could be endangered must be disclosed(for example business relationships with the rated entity, special personalcontacts).87

3. Regulation related to CRA independence

A CRA must not rate an entity if this entity is providing the CRA with rev-enues equal to or exceeding 10% of total net revenues, if the CRA has directownership interest in the rated entity or is otherwise associated with it or ifthe CRA provides advice on the structure of the transaction.88

VI. Global approach? The globalization of financial markets makes national regulations problem-atic; this assessment applying to financial institutions generally is particularlyrelevant for the CRA since the number of rating entities is very small (inprinciple, only three CRA dominate the market).89 Therefore, the questionmust be discussed whether it would not be possible to develop criteria whichcould have a global range90.

85 Johansson, supra note 4, p. 10; for Final Rules see Code of Federal Regulation (CFR),Title 17: Commodity and Securities Exchange, Part 240 – General Rules and Regulations,Securities Exchange Act of 1934, Nationally Recognized Statistical Rating Agencies,§ 240.17g-5; the Dodd-Frank Act also provides regulation to prevent conflicts of interest,see: Freixas/Laux, supra note 4, 34; Darbellay/Partnoy, supra note 4, pp. 13 et seq.

86 See rule in 17 CFR § 240.17g-5 (c)(7); see Cane/Shamir/Tomas, supra note 80, p. 19 et seq.87 Johansson, supra note 4, p. 10.88 See 17 CFR § 240.17g-5 (c)(1); Johansson, supra note 4, p. 10.89 See the contribution in this volume written by Aline Darbellay, pp. 159 et seq.90 See also Hill, supra note 19, p. 87, who argues that globalization offers cause for opti-

mism.


301

A. Regulation based on guidelines

The present regulations related to conflicts of interest are not only geograph-ically narrow, but also very much situation-driven. The more detailed therules are formulated, the less likely a cross-border harmonization is takingplace.91 Rigid rules are said to lead to gaps, inconsistencies and are likely toproduce “creative compliance”.92

A new approach could rely more on principles and guideline-based regu-lations and less on a specific rule-based regime. This does not imply that thisargumentation is favoring the pre-crisis-situation, when CRA were stillexempt from regulation. It is rather a vote in favor of less rigid regulation andtowards a more stringent self-regulating approach. A principles-based frame-work has the advantage of a higher degree of flexibility, of more cost-effec-tiveness, of higher efficiency and of a better suitability for an oligopolisticmarket such as the CRA industry.93

1. Requirements of a new framework

The incitement for regulating CRA should be based on the assumption thatthe avoidance of conflicts of interest contribute to higher market confidence.If a principles-based approach is chosen, the CRA themselves would have thepossibility to decide how best to align their business objectives and processeswith the regulatory outcomes. The flexibility of such an approach shouldlead to a more effective oversight; whereas regulators usually are laggingbehind new financial innovations, CRA as market participants are closer tosuch developments.94

Furthermore, a principles-based approach allows adjustments of therequirements for individual market entries. This would be important inorder to overcome the present oligopolistic structure.95

91 Johansson, supra note 4, p. 11.92 See Julia Black/Martyn Hopper/Christa Band, Making a success of Principles-based reg-

ulation, in: Law and Financial Markets Review 2007, pp. 191–206, p. 193.93 The Principles-based regulation is not a new idea: Principles were introduced in financial

services regulation in the UK in 1990 already: See Black/Hopper/Band, supra note 92,p. 191.

94 Johansson, supra note 4, p. 11.95 Black/Hopper/Band, supra note 92, p. 193; see also Darbellay, supra note 2, pp. 154 et

seq.


302

2. Particular regulatory elements

Reality shows that regulations should not deal with “the rating” as a wholesince a unique legislative process cannot cover the different requirements ofthe manifold rating activities. Moreover, adjustments of the applicable prin-ciples must be possible as far as the size of the CRA is concerned, but also inrelation to the type of products rated by the CRA.96

Some provisions contained in the CRA Directive of the EU do not seemto comply with a principles-based approach. In view of the fact that experi-ence and knowledge is very important for an analyst, the requirement of ana-lyst rotation is doubtful in view of the efficiency and the cost-effectivenessobjective. Such a rotation mechanism is most likely burdensome to imple-ment and is improbable to impact the quality of ratings positively.97

Furthermore, the limitation for analysts to be employed by the ratedentity might reduce the attractiveness of the CRA in hiring competent per-sonnel, because this provision could restrict the career advancement oppor-tunities. Therefore, such prohibitive rule must go along with a certain incen-tive, e.g. severance payment.98

The requirement to establish a supervisory board composed of memberswith certain experience and knowledge seems to be rigid, but underlines theimportance of independence. Even if it could be argued that the concurrentCRA should have the authority to decide on organizational matters andstructures, the strict reliance on independence has its merits.99

B. Global co-operation

As mentioned, the globalization of financial markets calls for global co-oper-ation. The CRA business has a cross-border effect by nature; therefore, a“global rule book” and harmonized supervisory standards would lead tohigher regulatory efficiency in the regulation and supervision of the CRAindustry and minimize the risk for regulatory arbitrage. Furthermore, har-

96 Johansson, supra note 4, p. 11.97 Kristina St. Charles, Regulatory Imperialism: The Worldwide Export of European Regu-

latory Principles on Credit Rating Agencies, in: Minnesota Journal of International Law,Vol. 19, 2010, pp. 399–451, p. 430.

98 St. Charles, supra note 97, p. 431; Johansson, supra note 4, p. 12.99 Insofar different the approach of Johansson, supra note 4, p. 12.


303

monized standards across jurisdictions would also reduce the costs of imple-mentation.100

1. Industry-related organization

Such kind of “global rule book” encompassing harmonized supervisorystandards cannot easily be developed by national or regional legislators. Anindustry-related organization would be better suited to assume this function.An organization that could perform this role seems to be the Global Finan-cial Markets Association (GFMA).101

Rules issued by an industry-related organization are of a self-regulatorynature. To date, legal doctrine has dealt with the merits (and the weaknesses)of self-regulation in the financial market to a far extent102. Without goinginto details of this discussion the summary conclusion can be drawn that theflexibility, efficiency and cost-effectiveness of self-regulation prevail in com-parison with national and regional regulation if most (if not all) market par-ticipants are included in the self-regulatory regime and consider compliancewith the respective “private” rules as important in view of market reputation(“name and shame” factor).

As an international standard-setting body, an almost institution-likemechanism that enhances compliance should be established and ensure themembers’ compliance with the principles.103

2. Public supervision

Even in case of an elaborated self-regulatory regime containing principles-based requirements for the avoidance of conflicts of interest, some kind of“final” public supervision is still required for various reasons: The self-regu-latory framework possibly does not address sensitive issues (such as the intro-

100 Johansson, supra note 4, p. 12.101 GFMA does support the establishment of cross-border regulation, see e.g.: http://

www.gfma.org/initiatives/cross-border-resolution/cross-border-resolution/.102 See for instance the critical examination and debate of the benefits and deficits of self-reg-

ulation in the financial industry: Steven L. Schwarcz, Financial Industry Self-Regulation:Aspiration and Reality, in: University of Pennsylvania Law Review, Vol. 159, 2011, 293–302; Rolf H. Weber, Overcoming the Hard Law/Soft Law Dichotomy in Times of (Finan-cial) Crises, in: Journal of Governance and Regulation, Vol. 1, 2011, pp. 8–14.

103 Chris Brummer, Why Soft Law Dominates International Finance – And not Trade, in:Journal of International Economic Law, Vol. 13, 2010, pp. 623–643, p. 640.


304

duction and enforcement of sanctions for non-compliance with duties andthe removal of high ranking officers), the enforcement mechanisms apartfrom reputation do not seem to be satisfactory or “black sheep” have a neg-ative influence on the rating market. In these situations, effective publicsupervision is necessary.

The most suitable organization seems to be the Financial Stability Board(FSB). In order not to overload the FSB it could be considered to establishan advisory and cooperative committee under the umbrella of the FSB. Thecommittee would consist of experts of the CRA industry and could be sup-ported by a Regional Consultative Group.104

VII. OutlookThe failure of CRA to deliver accurate assessments of companies played acrucial role in the forefront of the financial crisis. Governments around theworld agree that CRA play a systemically relevant role and that regulationmust not be left to the CRA themselves. As has been shown, reputation alonedoes not serve as a stringent correcting tool and does not provide incentiveenough so that the conflicts of interest in the CRA industry would not occur.The quest for a reasonable regulatory framework is unavoidable. At the sametime, it should be considered that basic “rules of thumb” are more likely tohelp achieving the desired results than sophisticated models that consumethe time of thousands of persons in the finance industry for implementationand adaption.

As a reaction to this failure, financial regulators in various jurisdictionstended to regulate CRA very rigidly. According to the authors’ opinion, thisapproach might not prove successful. Instead, there should be a harmoniza-tion of commonly agreed guidelines, i.e. a set of principles that are applicableinternationally. These guidelines could be implemented by a SRO (Self-reg-ulating Organization).

Thereby, the compliance of CRA could be enhanced, as principles wouldallow the individual CRA to elaborate a tailor-made business model withinthe given framework. A set of principles possibly leads to greater adherenceto the rules, as the CRA and the SRO would work in symbiotic exchange. In

104 The FSB is mandated with several tasks that work towards sustainable and strong financialsystems and the stability of financial markets; for further details see: http://www.financial-stabilityboard.org/about/mandate.htm.


305

addition, a principle-based regulation is more cost-effective, because the setof rules can be adapted to new regulatory demands faster. Following theassumptions above, the conclusion can be drawn that public supervisionshould be established.

Separate copy of article from Perspectives on Credit ...14 See John P. Hunt, The SEC’s proposed Rating Agency Rules: Unresolved Conflicts, Ber- keley Center for Law, Business and

Documents