51997AC1173

Opinion of the Economic and Social Committee on the 'Proposal for a European Parliament and Council Directive amending Council Directive 93/6/EEC on the capital adequacy of investment firms and credit institutions` () (98/C 19/04)

On 25 June 1997 the European Council decided to consult the Economic and Social Committee, under Articles 57(2) and 198 of the Treaty establishing the European Community, on the above-mentioned proposal.

The Section for Industry, Commerce, Crafts and Services, which was responsible for preparing the Committee's work on the subject, adopted its opinion on 8 October 1997. The rapporteur was Mr Levitt.

At its 349th plenary session (meeting of 29 October 1997) the Economic and Social Committee adopted the following opinion by 106 votes in favour, two against and one abstention.

1. Introduction

1.1. Purpose of the Capital Adequacy Directive and its amendment

1.1.1. Council Directive 93/6/EEC on the capital adequacy of investment firms and credit institutions (CAD) harmonizes capital requirements for credit institutions and investment firms in respect of positions on their trading books which are subject to market risk. Market risk is the risk of losses arising from movements in market prices. The CAD sets out the method to be applied in calculating capital requirements to cover positions in debt securities, interest rates, equities, foreign exchange and financial products which are derived from them. It also defines the nature of the Own Funds which must be held to cover these requirements.

1.1.2. With Council Directive 93/22/EEC in investment services in the securities field, it forms the basis of the coordination of prudential rules for investment business which permits the mutual recognition of authorization by national authorities. These two directives, together with directives relating specifically to credit institutions and insurance companies, form the basis of the single market in financial services.

1.1.3. The CAD was agreed at a time when negotiations aimed at harmonizing capital requirements for market risk were proceeding in wider international fora. It was foreseen at that time that the CAD might need to be amended in consequence (Article 14). In January 1996, the Basle Committee on Banking Supervision published an amendment to the 1988 Capital Accord to incorporate market risks.

1.1.4. The CAD also contained (Article 13) an invitation to the Commission to propose capital requirements for trading positions in commodities and commodities derivatives. The 1996 amendment to the Capital Accord also covers capital requirements in respect of such positions.

1.1.5. The proposal to amend 93/6/EEC is designed to bring the EU capital adequacy regime into line with what has now been agreed by G10 banking supervisors and, subject to the comments in section 3 below, the Commission's proposal follows Basle in its essential respects.

1.1.6. The ESC in its favourable opinion on the Commission's original 1990 proposal for the CAD retained three guiding principles for judging the proposal:

- it should create a broadly level playing field between banking and non-banking firms trading in securities market;

- it should enhance - or at least not impair - the attraction of the EU as a financial centre;

- it should not restrict access to the financial markets, stifle competition, innovation and consumer choice.

1.1.7. In addition, the ESC noted - in relation to the own funds requirements which resulted from the CAD - that these 'should be as consistent as possible, from the technical point of view, with the risk monitoring instruments which already exist in the institutions`.

1.1.8. These guiding criteria remain valid in analysing the current proposal.

1.2. Institutions covered

The proposed directive will apply to investment firms and credit institutions.

1.3. Summary of amendment

1.3.1. The proposal has three elements:

- amendments to allow firms, subject to rigorous safeguards, to use internal statistical models to calculate their capital requirements for market risk rather than the standardized method outlined in Annexes I, II , III and VII. In particular a new Annex VIII is proposed;

- amendments to incorporate commodities risk into the CAD, in particular a new Annex VII. Gold will, however, be treated as foreign exchange and amendments in particular to Annex III of CAD are proposed to allow this;

- an amendment to correct an unintended anomaly in the CAD whereby the issuance of subordinated debt automatically excludes that issuer's equity from being included in a portfolio qualifying for a concessionary 2 % specific risk weighting even if they are highly rated.

1.3.2. The first two of these are the most significant and take up Basle's proposals on market risk. They are analysed in detail below. In keeping with the rationale of the 1993 directive, this proposal allows the market risks as measured by the new Annexes VII and VIII to be covered by the expanded definition of Own Funds created by the 1993 proposal.

2. General remarks

2.1. Why CAD needs to be amended:

2.1.1. Some banks and investment firms already use internal models to achieve a better understanding of the risks to which they are exposed, and to achieve a better basis for the control of risk and the management of the economic capital of the business. Allowing the use of models for the purposes of calculating regulatory capital requirements would encourage firms to refine their understanding of and control over the risks they face to the benefit of all. This is important, and to the benefit of the supervisory authorities, savers and shareholders, because there have been a number of financial scandals in recent times which have revealed inadequate top management understanding and control of market risks.

2.1.2. There is also a competitive aspect to the matter. The Amendment to the Basle Capital Accord will take full effect from the end of 1997. From that point, banks from non-EU jurisdictions will be able to use internal models to calculate market risk capital requirements. While the 1993 CAD does not prevent firms from using these models for the purposes of calculating regulatory capital requirements, this is subject to the condition that the result is a capital charge which may be no lower than the charge under the standardized method created by the CAD. Whether or not an institution using internal models will have a lower capital charge than under the 1993 CAD will depend upon the nature of its positions, but it is anticipated that capital requirements be lower in some cases. This reduction is not a lowering of standards, but reflects the recognition by supervisors that properly constructed models (and national authorities will be required to apply strict criteria before approving their use) allow banks and investment firms to be better able to capture, control and reduce the economic and financial risks that they run. This will involve very significant investment by institutions in developing and implementing models and the management information and procedures they necessitate. But the potential benefits - improved understanding and control of risk and the potential for a capital requirement which is more precisely tailored to the real risks being run - are also significant. However, since Member States must apply at least the minimum requirements stipulated by EU directives, many EU firms will not have an incentive to use internal models to calculate capital requirements until such time as the CAD is modified to permit this.

2.1.3. This has two consequences, both of which are significant for the EU and its firms:

- some EU firms may be subject to competitive disadvantage vis-à-vis third country firms by virtue of having to meet higher capital charges where the use of models would imply lower requirements. These represent enforced economic inefficiency;

- EU firms - and EU supervisors - are not able to align internal and regulatory measures of risk. Such an alignment provides incentives for firms to strive constantly to improve their risk measurement and risk management systems, as the ESC recognized in its opinion on the Commission's 1990 proposal. It is important, and in the interests of the soundness of the financial system and of customers, that EU firms remain in the front rank of such developments.

This underlines the need for urgent progress on the proposal.

2.1.4. It is also necessary to amend the CAD to allow positions in commodities and in commodities derivatives to be held in the trading book.

2.1.5. The Commission has limited its proposals to the minimum needed to allow the prompt implementation of internationally agreed standards in the EU so as to ensure as rapid progress through the legislative procedure as possible. This is in the interests both of the competitiveness of the EU and of enhancing control of risks in EU financial institutions and is to be welcomed.

2.2. Future Technical Amendments to the Regulation of Financial Services in the EU - Keeping EU regulation at the forefront of developments

2.2.1. The Commission's proposal is highly technical and a response to progress in international regulatory fora. It should be noted that discussions on other technical issues related to this proposal continue, both among banking supervisors at the BIS, among securities supervisors in IOSCO and between the two bodies.

2.2.2. In particular, discussions continue among banking supervisors on modelling specific risk () and the minimum specific risk charge which is set at 50 % of the charge determined by the standardized method. The Commission proposal contains this provision (Annex VIII point 5), which may have to be dropped or amended in the foreseeable future. The Basle Committee of Banking Supervisors also envisages the possibility of amending its Capital Accord in the light of the experience acquired through back testing e.g. in relation to the multiplication factor.

2.2.3. Means therefore need to be found of ensuring that EU legislation keeps pace, from a technical perspective, with developments both in the markets and in other regulatory fora. Arguably, full directives such as those currently proposed do not readily permit this given the length of the legislative procedure. Nor is the full co-decision procedure required for legislative proposals really necessary for amendments on such technical points of detail. This points to the need for comitological provisions.

2.2.4. The Committee therefore recommends that proposals be devised to make this sort of highly technical regulation subject to comitology as soon as possible. If the Commission does not believe that it is likely to be able to achieve this in the foreseeable future, an alternative approach would be to adopt less prescriptive wording which permitted supervisors greater flexibility in applying agreed international standards, allowing them to incorporate technical changes into national regulations without delay.

3. Specific remarks

3.1. Article 1 points 1-8 and Annex I contain for the most part amendments which are consequential on the decision to bring positions in commodities and commodities derivatives within the scope of the CAD and on the decision to allow firms to use internal models to calculate capital requirements. Annex I also contains mostly consequential amendments. A more detailed guide to the Commission's proposal is included as an Annex to this opinion.

3.1.1. The principal amendments are contained in Article 1 point 6 and in Annex II. Annex II contains the two new Annexes to be added to the CAD, a new Annex VII on commodities risk and a new Annex VIII on internal models. Article 1 point 6 proposes a transitional regime which competent authorities may apply to investment firms which are not yet able to model commodities risk according to the method in Annex VIII, but for whom the standardized method set out in Annex VII is not deemed appropriate.

For ease of understanding, this analysis groups the Commission's proposal under the two principal elements identified above, internal models and commodities.

3.2. Internal models

3.2.1. Firms will be allowed to use models to calculate capital requirements for interest rate risk, equity risk, foreign exchange risk and commodities risk. The use of such models is subject to prior approval by the competent authorities and the satisfaction of a number of demanding qualitative and quantitative criteria. These qualitative and quantitative criteria follow what has been determined internationally in Basle.

3.2.2. The purpose of the qualitative criteria is to ensure that firms using models have market risk management standards that are conceptually sound and implemented with integrity. These are set out in points 2-4 of Annex VIII. They are tough so as to ensure that the use of internal models leads to no dilution of regulatory standards, but are of necessity expressed in language which leaves a large measure of discretion to the competent authorities. They will need this discretion in order to take into account the nature of individual firms. Particular stress is laid upon the responsibilities of management and directors in ensuring high standards and effective controls. While high standards of controls may be expected of all institutions, the precise nature of the controls will need to vary according to the nature of an institution's business and its geographical spread.

3.2.3. The quantitative criteria set out the minimum statistical parameters for measuring risk. These are set out in points 9-11 of Annex VIII and they follow the requirements of Basle. The 99th percentile one-tailed confidence interval means that the model should calculate a value-at-risk (VAR) value which matches or exceeds the trading outcome 99 times out of 100, that is to say that an institution which uses this technique should not suffer a loss which exceeds the VAR value it has calculated 99 times out of 100.

3.2.4. The capital requirement of a firm which uses internal models will be the greater of the previous day's VAR or the average daily VAR of the previous 60 business days multiplied by the multiplication factor of three. If the back testing programme reveals weaknesses with the model or with the institution's internal controls the multiplication factor may be increased to four by the addition of a plus factor. The multiplication factor of three, together with the possibility of adding a plus factor, means that the approach is very conservative, and that a large reserve of capital will be held against market risks. Financial firms have argued that the approach is much too conservative and that a multiplication factor of three is excessive and will result in excessive capital allocation against market risks. In particular, it has been suggested that the multiplication factor means that greater sophistication is not adequately rewarded: no matter how good a bank's model may be, it is still subject to a minimum factor of three. This is an issue to which Basle has promised to return at a future date, and future changes cannot be ruled out.

3.2.5. It should be noted that volatilities are to be derived over an effective period of at least a year. However, historical volatility may be of limited usefulness in certain circumstances e.g. illiquid, deregulating or emerging markets, or with complex new financial products. In these circumstances it is important that banks use data which adequately reflect the real risks being run and have other controls in place to control exposures to these risk factors. The text might be clarified on this point.

3.2.6. Firms will also be required to provide evidence of the accuracy of their models by conducting regular back testing i.e. comparing the trading outcomes of the businesses covered by the model with the predictions of the model. There are a number of problems with the Commission's proposals on back testing (point 7 of Annex VIII) which could place EU institutions at a disadvantage with their competitors without offering significant compensating benefits in terms of protecting investors or strengthening the financial system.

3.2.7. The specification of the daily back test

This is a key issue. Institutions will be required to back test daily, comparing the value at risk against both actual and hypothetical () trading outcomes. This is costly and burdensome and is stricter than Basle requires. Basle urges banks to develop a capacity to perform back tests using both, but envisages that supervisors will differ in the emphasis that they wish to place on either test.

3.2.8. The Committee believes that the focus of the Commission's attention should be on testing against actual trading outcomes, leaving it to national supervisors how much attention need be paid to testing against hypothetical outcomes.

3.2.9. Back testing against hypothetical outcomes provides a purely statistical test of the integrity of the model. Testing against actual outcomes sheds light on the extent to which the model captures all trading risk to which a bank is exposed. It is a test of how a model performs in live situations.

3.2.10. There is also an issue of practicality. It is not a trivial matter to perform two sets of daily back tests. It should also be noted that the controls ensuring the accuracy of data on actual outcomes have to be much more rigorous as they will be used for management and statutory profit and loss reporting as well as for tax purposes, and that it is very expensive (especially in computing terms) to derive hypothetical outcomes for every aspect of a bank's trading operations on a daily basis. Ultimately, of course, it is actual trading outcomes which matter and which determine the solvency of financial institutions, and for this reason too, testing against actual outcomes should be the principal focus of both the firm's and the supervisor's attention.

3.2.11. It should be noted that the United States in implementing the same Basle provisions has chosen to focus solely on actual trading outcomes.

3.2.12. The Commission's proposal requires further clarification. The point is technical, but important in ensuring the coherence of back testing as a regulatory tool. Annex VIII paragraph 7 suggests that the value-at-risk number calculated by means of the model - which uses a 10 day holding period - should be compared with the change in the value of the portfolio. If this is left unchanged this would require firms to compare the capital requirement generated by their model to cater for a ten-day price move with a one-day profit or loss figure. This would produce misleading results and not be an adequate test of the models being used.

3.2.13. Basle's paper 'The Supervisory Framework for the use of "Back testing" in conjunction with the internal models approach to Market Risk Capital Requirements` makes it clear (p. 3) that because daily trading outcomes are being used 'the back testing framework described here involves the use of risk measures calibrated to a one-day holding period`. For the avoidance of any doubt, and to ensure the coherence of back testing programmes, the Commission should follow Basle in making it clear that for the purposes of back testing a one day holding period should be used.

3.2.14. Unclarity about the plus factor

Where back testing reveals shortcomings in a firm's internal model which are not so serious as to call into question the basic integrity of the model but which are nevertheless material, the competent authorities may impose a 'plus factor`. The plus factor is a number between 0 and 1 which is added to the multiplication factor so that the multiplication factor is, in effect, raised from three to four.

3.2.15. It is not clear from the Commission's text which 'overshootings` or exceptions to use the language of Basle 'count` for purposes of setting the plus factor. Consistent with the arguments deployed above, the Committee believes that only exceptions which arise from back testing against actual outcomes should count as overshootings for the purposes of Table 5.

3.2.16. The Commission admits that in an exceptional situation the authorities may waive the requirement to add a plus factor. But 'exceptional situation` is too high - and too unclear - a test. Basle is significantly less rigid and, particularly where a bank has between five and nine exceptions within the 250 working day period, calls upon supervisors to exercise judgement in interpreting back testing results. The Basle Committee notes that there are a number of reasons why even high quality models may produce results with between five and nine exceptions over such a period and notes that 'increases in the plus factor are not meant to be purely automatic`. The Commission's proposal reflects the intentions of Basle in respect of models which generate 10 or more exceptions over the 250 observations.

3.2.17. In the interests of ensuring that European banks operate on the same basis as institutions outside the EU, the proposal should be amended to better reflect Basle which states that back testing results which show between 5 and 9 overshootings 'should generally be presumed to imply an increase in the multiplication factor unless the bank can demonstrate that such an increase is not warranted`.

3.2.18. The requirement to report overshootings 'immediately` with reasons is not reasonable. It is reasonable - under normal circumstances - to expect institutions to report to the competent authorities the following business day that there may be the possibility of an overshooting. But an institution will need time to discover where the problem lies and whether the overshooting is a genuine one or whether it arises because, for example, the trading outcome was subject to error and has had to be restated. It would be reasonable to expect an institution to produce a detailed report for the supervisor within a week.

3.2.19. Modelling specific risk

Point 5 of Annex VIII provides that where an institution does not model specific risk, it is subject to the specific risk charges specified according to the standardized method of the CAD. However, where an institution does model specific risk accurately, it is nevertheless subject to a minimum specific risk charge of 50 % of the standard charge.

3.2.20. This is pointlessly burdensome and creates the wrong incentives for institutions. The standardized specific risk charge makes no allowance for diversification, which is a well-accepted method of reducing risk and, in particular, of reducing specific risk. It is expensive in terms both of the capital requirements generated and the computer systems required to calculate capital requirements under two unrelated methods. It also fails to give institutions adequate benefit of improvements to their model, and thus diminishes the incentives to improve specific risk modelling. To that extent it also weakens the incentives to improve risk control in financial institutions.

3.2.21. Although it is also a feature of Basle, it is under review and may well be changed in the near future. National supervisors should therefore have the discretion to give full recognition of specific risk models where they are satisfied of the accuracy of those models.

3.3. Commodities and commodities derivatives

The Commission's proposal expands the definition of the trading book contained in the CAD to embrace positions in commodities and commodities derivatives. This is essential if their treatment is to fall under the terms of this directive.

3.3.1. The proposal follows the Amendment to the Basle Capital Accord in allowing firms to opt between three possible approaches depending on the degree of sophistication of the firm and the extent of its involvement in commodities trading. Firms may either use internal models (see below), the maturity ladder approach (Annex VII points 13-16) or a highly penal simplified approach which is only appropriate for those doing de minimis levels of business (Annex VII point 17).

3.3.2. The maturity ladder approach follows Basle in making no distinction between categories of commodities. Gold is treated differently, being regarded as foreign exchange. This is because the volatility of gold is more in line with foreign currencies and banks manage positions in gold in a similar manner to foreign currency positions. Points 3(a)-3(c) of Annex I are consequential to this.

3.3.3. Differences between the volatility characteristics of other commodities (e.g. between precious and base metals) are not taken into account. The Basle Committee notes (p. 29) that it has adopted this approach for simplicity, despite the differences in volatility between different commodities given the fact that banks normally run rather small open positions in commodities. It also states (p. 28) that both the simplified approach and the maturity ladder approach 'are appropriate only for banks which, in relative terms, conduct only a limited amount of commodities business. Major traders would be expected over time to adopt a models approach ...`.

3.3.4. The CAD, however, has a broader scope than Basle since it also applies to investment firms. Some investment firms conduct a significant amount of commodities business relative to their size. The Basle approach therefore requires some modification.

3.3.5. The Commission has determined that these investment firms may experience difficulty in applying models immediately upon entry into force of the directive, and that they will need time to develop such models and the appropriate risk management systems that go with them. It is therefore proposed (Article 1 point 6) that Member States should, until end-December 1999 be empowered to use the maturity ladder approach described in Annex VII but to set spread, carry and outright rates to be applied to positions in commodities which are different from those set out in points 13 and 16(ii) and (iii).

3.3.6. Member States' discretion is considerably limited by the provisions of Article 1 point 6 which, inter alia, require the competent authorities which take advantage of these provisions to provide empirical justification of their actions.

3.3.7. Overall, with the caveat outlined below, the Commission's approach to this issue seems well-balanced. It is desirable that firms should have the incentive to develop appropriate risk management and modelling, but also that legitimate business needs be taken into account. It is also right that any differential treatment as between banks and investment firms be removed after the transitional period.

3.3.8. Nevertheless, there are very real difficulties in modelling commodity risk on account of the varying liquidity of the market and the fact that commodity prices may move in sudden 'jumps` rather than gradually. The time that will be needed to devise the appropriate models and satisfy the supervisors of their validity should not be underestimated. This means that a two year transitional period is barely generous. Given the limited competitive distortions between banks and investment firms that would be caused - banks which used models would still have lower capital charges than investment firms using some intermediate regime - it would be possible safely to envisage a lengthier transition or at least a further review at the end of the two year period.

4. Conclusions

4.1. The ESC welcomes and approves the Commission's proposal and urges the Council and the Parliament to consider it with all speed. This is necessary in the interests of protecting the competitive position of EU firms operating in international and highly competitive markets; it will also enhance the prudential control of banks and investment firms which is self-evidently desirable from a public policy perspective and is in the interests of the financial stability of the EU financial system and of the customers of financial institutions.

4.2. There are a number of points on which further attention and some amendments may be required.

4.2.1. Consideration needs to be given as to how highly technical amendments to financial services legislation is effected so as to ensure that EU legislation remains current in a fast-moving international environment. The ESC recommends that greater use be made of comitology in this area.

4.2.2. The Commission's proposals on the back testing of internal models require some amendment to bring them closer to the intentions of international regulators and to avoid putting EU institutions at a competitive disadvantage through excessive, prescriptive requirements. In particular, financial institutions should not have to backrest daily on two different bases, and it should be clear which basis forms the basis on which regulators will apply increased capital charges. In line with Basle, national regulators should also have greater flexibility than the Commission proposes in applying increased capital charges via the plus factor to banks which have had between 5 and 9 overshootings within a 250 working day period.

4.2.3. The ESC recommends that the principal focus of back testing should be against actual outcomes in order to capture how models perform in live situations.

4.2.4. National supervisors should have the discretion to give full recognition to specific risk models where they are satisfied as to the accuracy of those models.

4.2.5. The approach of providing transitional arrangements for investment firms which have a substantial involvement in commodities, but which are not yet able to model this risk is the correct one. The ESC considers, however, that a two year transitional period may not be sufficient and may need to be reviewed.

Brussels, 29 October 1997.

The President of the Economic and Social Committee

Tom JENKINS

() OJ C 240, 6. 8. 1997, p. 24.

() Market risk is composed of two elements: general and specific. General market risk captures the exposure to movement in the market as a whole. Specific risk is the risk that an individual instrument moves by more or less than the general market in day-to-day trading and includes event risk - the risk that an individual instrument moves precipitously relative to the general market.

() 'Hypothetical outcome` in this context means the hypothetical change in the value of a firm's portfolio between close of business on successive days assuming no change in the portfolio. (Whatever the outcome on the substance of the argument developed here, the drafting of the second sub-paragraph of paragraph 7 of Annex VIII needs to be made substantially clearer on this point.).

APPENDIX to the opinion of the Economic and Social Committee

This appendix sets out how the Commission's proposal amends the 1993 directive.

>TABLE>