C) the Quality of Ratings and the Rating Process

Financial Markets Policy Directorate

Jörgen Holmquist Director General European Commission DG Market, Unit G3 B-1049 Brussels Belgium

Date Your letter (Reference) Our reference Inquiries 4-9-2008FM-2008/ 2130 MSubjectCredit Rating Agencies Mark Bezemer T +3170-3427130 F +3170-3427901 E [email protected]

Dear Mr Holmquist,

I would like to inform you of the Dutch position on the two recent consultation documents pertaining to Credit Rating Agencies (CRAs). While this letter will address the more general issues, the annex lists substantive and technical comments on the proposed legislation.

The process The Netherlands appreciate the Commission’s proposal to improve regulation on CRAs. In particular, we welcome those measures that aim at preventing conflicts of interest and increasing transparency. As you may have heard, the Netherlands emphasis the better regulation principles and working from a clear evidence base. Unfortunately, in this case a thorough regulatory impact assessment is lacking which makes it difficult for us to judge this proposal. Furthermore, consultation should leave sufficient time to provide a decent assessment. In our view, it is desirable that a full quantitative and qualitative cost-benefit analysis is undertaken before presenting a final proposal which includes assessing the expected impact, costs and benefits on: a) the behaviour of investors b) barriers to entry and competition c) the quality of ratings and the rating process d) the global and European credit rating business e) the effectiveness and cost of supervision Prior to making any decision on the authorisation and supervision structure it might be better to first work out all options into more detail (listing all pros and cons) and to determine the required level of oversight. We also note that the proposed approach risks creating duplication with existing legislation and supervision. In this regard, it would have been good to clarify in what areas the proposed legislation deviates from e.g. the IOSCO-Code, CEBS-guidelines and US-legislation. In our view, we should avoid rushing headlong into new legislation and setting up a supervisory structure that affects both the industry and the economy as a whole for a long time to come. In the current situation, ensuring quality should take precedence over meeting a self-imposed deadline.

Objectives Referring to the three stated objectives, we generally support measures aimed at avoiding and managing conflicts of interest (objective 1) and increasing transparency (objective 3). However, there are some concerns regarding the second objective, i.e. to ‘ensure that CRAs apply a high standard for the quality of the rating methodology and the ratings’. While this outcome is clearly desirable, we are somewhat hesitant whether the present proposal may not inadvertently result in giving a formal or informal public stamp of approval on either the rating or the rating methodology. Such a stamp of approval might increase the reliance on ratings and is difficult to enforce given information asymmetries and the required level of expertise. Moreover, the real possibility of a national competent authority interfering in the rating process because of downgrading of a ‘national champion’ should be avoided. Such interference on a national level would have a harmful effect on investors in all Member States. In this regard, we refer to the CEBS Guidelines on the recognition of External Credit Assessment Institutions1 and the US Credit Rating Reform Act of 20062. To avoid any misunderstandings or moral hazard from supervisors, CRAs or investors, a similar clause in the proposed legislation should clearly state that interference in the substance of credit ratings is not allowed and that the regulatory regime should not undertake any detailed assessment of methodologies or endorse them.

A global credit rating market Given the unique characteristics of the credit rating market in which the impact of faulty credit ratings on investors is unrelated to the jurisdiction where the rating activity took place, one could argue that effective supervision is not a national or European but a global issue. This brings up a number of questions that need to be addressed, such as:

1 ’87 - When assessing an ECAI’s methodology in any of these broad classes or market seg- ments, competent authorities will avoid making a direct judgement as to whether an ECAI’s methodology is objectively correct. They should not be seen as endorsing any particular type of methodology’ (CEBS Guidelines on the recognition of External Credit Assessment Institutions, p. 18). 2 `(2) LIMITATION- […] Notwithstanding any other provision of law, neither the Commission nor any State (or political subdivision thereof) may regulate the substance of credit ratings or the procedures and methodologies by which any nationally recognized statistical rating organization determines credit ratings (2 review of application – (c) Accountability for Ratings Procedures). 2 / 7 a) Is it possible under the proposed legislation to ensure that credit rating activities taking place abroad, which generate ratings that are also used within the EU, adhere to our standards? b) Could strict European legislation affect the location where rating activities take place and could this have an adverse impact on the economy of a Member State or the quality of ratings used in the EU? c) Should a system of recognition based on an equivalency principle (passport- ing) be considered with for instance the US, Japan or Australia? Is the EU proposal consistent with that of major non-EU jurisdictions? d) What are the opportunities to further strengthen supervision at the global level? e) If the subsidiary/branch established in the EU mainly acts as an intermediate for rating operations being performed outside of the EU, how will supervision and enforcement be performed? f) How feasible are enforcement measures for CRAs being established outside of the EU?

In our view, it is necessary to add this issue as a fourth pillar under the heading ‘the way forward’.

Options Our initial observations on the structure of authorisation and supervision are: a) Option 1 - National authorisation and supervision process with strong coordination through CESR with compulsory involvement of other competent authorities Important advantages of this option are its fast implementation and the fact that the authorisation process could benefit from the experience and expertise generated during supervision activities in countries where other CRAs are already active. Coordination through CESR is essential in order to help guarantee that all competent authorities are adequately involved and to contribute to a consistent and coherent level of supervision between the different Member States.

Although we recognize that the proposed home-host supervision addresses some concerns associated with a decentralised system, disadvantages include the fact that a) economies of scale are unlikely to be maximised 3; and b) in practice it is unlikely to result in a consistent and coherent level of supervision as the level of supervisory activities will vary across countries4. More importantly, this option would not fully take the unique aspects of credit ratings in consideration. Unlike for instance banking or insurance, the impact of faulty credit ratings is unrelated to the location where the sale or rating activity took place. Investors using and relying on ratings are dispersed

3 It is more difficult to accumulate the required high level of expertise and experience on a national level especially in countries where few CRAs are active. 4 This risk would be mitigated if CESR would have binding powers. 3 / 7 throughout the EU and could actually be concentrated in another member state. The impact of poor supervision would therefore be felt on a European rather than a national level. The key question is whether national supervision of CRAs would adequately incorporate this European aspect. The proposed strong coordination and involvement of other competent authorities might in practice be insufficient. For instance, Article 23a which allows refusal to answer to a request for cooperation from another supervisory authority in case ‘[…] such an investigation, on-site inspection, supervisory activity or exchange of information might adversely affect the sovereignty, security or public policy of the Member State addressed’ may provide too much scope to refuse cooperation in cases where the legitimate interests of other Member States are affected. This option should at least include a) a clear mechanism in which affected European stakeholders enjoy equal access and responsiveness from a certain supervisor; and b) an effective mechanism to resolve supervisory conflicts of interest. b) Option 2 - Community authorisation via an Agency combined with supervision by national regulators This option appears to be a variation of option 1. In principle, the permanent and most important activities (supervision) would still take place at the national level. Providing community authorisation power to CESR5 which is already performing a central, coordinating role would in our view have more value added than setting up an entirely new agency which could lead to unnecessary duplication and delays. This would of course necessitate additional staff and resources. This option still involves the duplication of national expertise without fully utilising the main advantages of centralisation: a) concentrating expertise and capacity; b) ensuring a homogenous level of supervision; and c) bringing supervision on a level that corresponds with the area of impact.

The main advantages of this option would be that the decision to withdraw authorisations is taken at the level that corresponds with the likely impact of CRAs and that disagreements between national competent authorities can be effectively resolved at the appropriate level. The agency could also be a good point of entry for complaints from investors in other Member States.

This option resolves some of the supervision problems associated with option 1 and additional advantages could be generated if CESR is allowed to gradually expand its responsibilities and its decision powers would be binding. In this regard, the possibility to delegate national tasks and responsibilities may be particularly useful for countries with a limited number of CRAs. At the same time, such evolution would lead to a situation in which supervision on CRAs takes place both at CESR and some national

5 We refer to the paragraph on linking the supervisory and authorisation process with the existing banking supervision: it appears logical to structurally involve CEBS in this single authorisation process. 4 / 7 regulators, which might be undesirable. c) Central authorisation and supervision via an Agency Centralising and concentrating both supervision and authorisation within CEBS/CESR, a real one-stop-shop, would have the advantages of: a) economies of scale that result in a higher average level of supervision; b) a consistent and coherent level of supervision throughout the EU; c) supervision at the level where the impact of the rating activity takes place; and d) a single point of entry for CRAs. We recognise that fully concentrating authorisation and supervision of CRAs at the European level still may seem ambitious at this point in time, but believe it would be preferable to include this option in the discussion. This way, the main advantages and disadvantages of all main options could be identified and evaluated. While this third option might in the end turn out not to be desirable at this moment, it could be still a structure that would make sense in the future and hereby affect the decision between the two remaining options.

In our view, centralising authorisation and supervision at this level might not create a precedent given the unique characteristics of the CRA market:  The area of impact of poor credit ratings is largely unrelated to the area where the rating activity takes place: investors using credit ratings are dispersed throughout the EU  Relatively few CRAs are subject to oversight  Limited entry of new CRAs  The main CRAs are concentrated in a few countries  It is difficult to maintain a homogeneous level of expertise and capacity in each member state as the number of CRAs in many countries is limited  Oversight at a national level makes it difficult to identify best practices and weaknesses in the credit rating market

In conclusion, it is too early to take a definitive decision on the desirable structure of authorisation and supervision. All three options need to be further analysed before making a decision.

Linking the supervisory and authorisation process with the existing banking supervision The proposal remains silent on the existing banking authorisation and supervision of CRAs. In our view, two important issues need to be addressed. First, we should avoid ending up with a regime that is based on two distinct and partially overlapping procedures for authorisation and supervision that would reduce the overall quality of supervision, add to its overall cost and create unnecessary administrative burdens. Second, there should be a sound division of labour and/ or cooperative structure for both authorisation and supervision. With regard to the authorisation process, it appears logical to structurally involve

5 / 7 CEBS in a single authorisation process. This issue should be adequately addressed for supervision as well. Would possibly ending up with over 30 national authorities be the most effective supervision structure for a unique market that is dominated by a small number of CRAs with activities that have an impact in the entire EU?

Entry barriers In several areas, the proposed legislation risks creating entry barriers in a market which unfortunately is already dominated by very few dominant CRAs. In our view, it is desirable to amend these articles and/ or introduce a lighter regime for CRAs smaller than a certain size. Moreover, article 3-3 which requires that ‘CRAs whose headquarters are located outside the Community shall establish a subsidiary or branch in the Community for the purposes of performing the rating activity’ might result in a barrier to entry and limit competition if it prevents small foreign companies from generating ratings that are used in the European Community. We refer to the annex for other examples.

Flexible and principle based regulation As pointed out in the annex, part of the proposed legislation could gain in effectiveness by making the articles more flexible and principle based (instead of rule based) with detailed rules being adopted at a lower level. This would make it more robust for future developments and insights and allow for fine-tuning.

The way forward6 In our view, the following three options are not mutually exclusive and should be further explored and specified.

1. Require regulated and sophisticated investors to rely more on own risk analysis, especially for (relatively) large investments We are in favour of creating suitable conditions that enable investors to rely more on their own risk analysis, although we should be careful not to overregulate. First, it is not entirely clear if the Commission envisages a full disclosure of CRAs’ methodologies and models to create the necessary conditions for investors to carry out their own risk assessments. If this would be the case, such a measure would come at great cost by creating strong disincentives for investment in these methodologies and models in the future. Second, it may not be desirable to require regulated and sophisticated investors to rely too much on their own risk analysis when large diversified portfolios that mitigate risks are concerned. For relatively small investments the costs of a comprehensive internal risk analysis could outweigh the benefits, both for the firm as for the economy as a whole. An appropriate balance needs to be maintained.

2. Require that all published ratings include ‘health-warnings’ informing of the specific risks associated with investment in these classes

6 As described on p. 5 of the consultation document. 6 / 7 We would welcome specific information on precisely what is referred to.

3. Examine regulatory references to CRA ratings and revisit them as necessary This could be a very useful exercise to avoid undue reliance of investors on the ratings. We are inclined to share the tentative view of DG MARKET services that a one-size-fits-all approach need not necessarily be followed, as ratings are used in different contexts, with varying intensity and for different purposes.

In conclusion, I would like to express my support for finding the most efficient and effective way to improve regulation of the credit rating market in order to remove one of the several root causes behind the financial crisis. I would be most grateful if the Commission could add a comprehensive impact assessment and a cost-benefit analysis of the proposal, which can only enhance the quality of discussion and may in the end contribute to a speedy positive conclusion. Obviously I remain at your disposal for further information and cooperation.

Yours sincerely,

Johan Barnard

Deputy Director Financial Markets Policy

Enclosed Other substantive and technical comments from the Netherlands on the EC-proposals pertaining to CRAs