COMITÉ EUROPÉEN DES ASSURANCES

Transcription

1 COMITÉ EUROPÉEN DES ASSURANCES SECRÉTARIAT GÉNÉRAL 3bis, rue de la Chaussée d'antin F Paris Tél. : Fax : DÉLÉGATION À BRUXELLES Square de Meeûs, 29 B 1000 Bruxelles Tél. : Fax : May 2004 CEA COMMENTS ON THE COMMISSION S DOCUMENT FURTHER ISSUES FOR DISCUSSION AND SUGGESTIONS FOR PREPARATORY WORK FOR CEIOPS (MARKT/2502/04) Introduction page 2 General Comments page 4 Contents page 3 CEA Comments page 4 Page 1 of 10

2 Introduction Following its latest comments on Markt/2543/03 of 14 April 2004, the Comité Européen des Assurances (CEA) is pleased to provide its comments on the Commission s document Further issues for discussion and suggestions for preparatory work for. * * * Page 2 of 10

3 Contents Introduction...2 Contents General Comments CEA Comments on Section 2 Asset Management CEA Comments on Section 3 Asset-Liability Management for life and non-life insurance companies CEA Comments on Section 4 Minimum solvency margin and other safety measures CEA comments on Section 5 Supervisory authorities powers CEA comments on Section 6 Solvency control levels...9 * * * Page 3 of 10

4 1. General Comments As a general comment we will point out that the focus in the Solvency II project has so far been on the requirements for single companies and the issues of groups have largely been left aside, as opposed to Basle II which foresees no target capital computation at solo level. In the single market it is extremely important that issues of groups be properly addressed as such and not only as a residual element. Indeed, in the case of groups it is important to shift the focus from the requirements for single companies to a group and financial conglomerate level, not least to take into account the benefits of their structure. This concerns, for example, the powers of the supervisors, their coordination and the calculation of the target capital levels. Lessons from the failure of the insurance groups directive and of its supervision at sub-group level should be drawn in the context of an EU-wide project. Furthermore, an adequate balance should be struck between the directive - in which core principles should be found and drafted in a precise manner to ensure legal certainty - and implementing measures and guidance. Too much emphasis on the latter could indeed jeopardise the aim of maximum harmonisation. Finally, it is essential that the Commission clarify the terminology used. For example, the Minimum Solvency Margin has been used alongside others such as Absolute Minimum Margin or safety net. To avoid confusion, key terms should be defined, notably the Minimum Solvency Margin, the Target Capital Requirement etc. 2. CEA Comments on Section 2 Asset Management Issue 1: Do you have comments on the draft wording of the directive? Would the requirement to cover the capital requirements involve implementing measures or supervisory guidance? Issue 2: Do you agree with the proposals? Do you have any comments or further suggestions? We refer to our comments of 26 February 2004 on the IAIS Guidance Paper on Investment Risk Management, in which we were highlighting the need for a right balance between qualitative principles and quantitative requirements, insisting that the prudent person principle is key in the context of insurers asset management. We stress that restrictive quantitative investment rules are incompatible with this principle and therefore would recommend the approach adopted in Article 18 of the directive on the activities and supervision of institutions for occupational retirement provision. Such an approach is also adopted in the Commission s proposal for a directive on reinsurance. Furthermore, the prudent person principle implies that restrictions for acceptable assets as defined in Articles 23 and 24 of the present Life Assurance Directive are removed (please also refer to our comments under Issue 11 below). Page 4 of 10

5 We accept the introduction of the requirement to have an investment plan, but a general one. The content of the investment plan should be clearly defined before incorporating it in the directive. We stress that it should not be required that the investment plan be as detailed as described in the list of the IAIS Core Principles (ICP 21 paragraph c). Furthermore, the publication of sensitive and/or confidential information should not be required. We do not accept the extension of the asset coverage rules to include the assets covering the Target Capital Requirement (TCR). As a final technical remark, we note that the proposed text includes a contradiction: since it is possible that the actual solvency capital is less than the TCR requirement, the coverage rules should in that case refer to the actual capital, and not to the TCR. 3. CEA Comments on Section 3 Asset-Liability Management for life and nonlife insurance companies Issue 3: Do you have comments on the draft wording of the directive? We generally agree with the draft wording of the directive, which could however be made more precise and subject to the following. The article should provide a definition of Asset Liability Management (ALM), since the meaning of this term varies, notably in the context of international discussions. Furthermore, it is important to define an objective which should be pursued through the requirement to set up ALM. We believe that the asset-liability risk (A/L-risk) should be accounted for in the quantitative and qualitative measures of the target capital requirement, both for life and non-life business. A simplified version of the calculation of the A/L-risk should be included in the standard approach (SA) for calculating the target capital level. More advanced methods can be introduced when going to internal models (IM). This might be completed by providing for the use of stress tests which should be part of a Pillar II supervisory review process. Finally, when talking of ALM in non-life, reference to long-tail or catastrophe insurance classes for which ALM analysis is of particular significance would be more appropriate than significant technical provisions (3.1 Introduction). Regarding the wording of the proposed article, we suggest to replace asset-liability analysis by asset-liability control. Page 5 of 10

6 Issue 4: Do you agree with the proposed approach? Do you have any further comments? We would like to make the following comments: ALM is important for life business (although different considerations apply for linked business from traditional business) and for non-life business The sophistication necessary in modelling ALM should always be proportionate to the underlying risk ALM should be reflected in the target capital requirement We would see a role for both implementing measures and guidance, provided however that the aim, definition and main requirements of ALM are set out in the body of the framework directive Issue 5: Do you have any comments or additions? Currency and all embedded options should be included in the nature of liabilities referred to under the first bullet point of Life assurance. Issue 6: In the list above we have not distinguished between different technical approaches or levels of sophistication (e.g. deterministic or stochastic, what risk indicators and assumptions to use etc). Neither have we defined the scope of ALM nor how long a time horizon should be used in the analysis. The organisational aspects of ALM modelling (the role of ALM committee, actuaries etc) have not been addressed either. CEIOPS and other stakeholders are asked to give their opinion if a more harmonised framework and higher level of standardization regarding these issues should be given in implementing measures or supervisory guidance. Life assurance: we are in general agreement with the comments. In particular: ALM has an important role in preparing the investment plan: indeed it is often the most important factor ALM has a clear role in quantitative analysis in Pillar II ALM should (not could as in the text) be used when addressing risk management issues of life assurance including pricing and new business issues ALM will play an important role in internal models on the life side: it is also relevant to cash-flow projections etc. This includes valuation of the cash flows, particularly in with profits and other participating business. The principle of proportionality in approach is important and we agree with the objective that it should be practical enough to be applied in the whole EU and fit in the new overall solvency regime by motivating better risk management. Non-Life insurance: we agree with the analysis that for much of non-life business, ALM will be able to use elementary approaches although for some business a sophisticated approach will be necessary. Page 6 of 10

7 A considerable number of different approaches are appropriate and the framework should permit flexibility. This suggests implementing measures focused on overall standards and amplified with guidance. In general, we favour a maximum degree of harmonisation emerging from an EUwide regulatory framework. Still, within this goal, a number of different approaches are appropriate and the framework should permit the flexibility needed. Similarly, a single EU-wide approach to technical provisions and target capital based on concrete ALM processes will ensure the relevant degree of standardisation. Achieving this would place both regulatory and financial accounts on a consistent disclosure framework. 4. CEA Comments on Section 4 Minimum solvency margin and other safety measures Issue 7: How should the minimum solvency margin be defined, i.e. what kind of formula would be most appropriate? Issue 8: At what level should the requirement be calibrated? The type of the formula and the level of the Minimum Solvency Margin (MSM) depend on what the respective roles of the MSM and the TCR are. We refer here to our reflections on solvency control levels (issues 15-17). In the framework outlined there, it would be logical that the MSM be based on a simple and objective formula in order to have legal certainty (since going below the MSM would trigger severe supervisory intervention). The level of the MSM should be so low, that going below it really constitutes an emergency situation. Issue 9: Do you agree with the proposal and do you have any other comments? Whilst we feel some sympathy for the Commission s preliminary preference for the first option, in practice this will depend on the prudence and diversification benefit built into the standard model and on the severity of the interventions measures available/optional/obligatory to the supervisors when the limits have been breached (see our response to issues 15 and 17). We also note that because the calculation of technical provisions is not harmonised, the present solvency margin calculated on the basis of those provisions results in different prudence levels from country to country. Issue 10: Are there reasons to change the authorised categories of assets (codified life Art. 23)? If so, what changes should be made? The list of authorised categories of assets should be removed as it does not allow insurance companies to benefit from financial markets innovation. Indeed, it is sometimes difficult to put a new financial instrument in one of the categories and therefore to determine whether it is authorised or not. Full flexibility should be allowed, as the TCR will take investment risk into account. Page 7 of 10

8 If a list was to be kept, it should be one of very limited unauthorised categories of assets (e.g. junk bonds), which is more in line with the prudent person principle. Within the assets there should be a substance over form approach. For example, derivatives should be an acceptable asset class in their own right, particularly when the issuer is subject to supervision. Similarly, alternative asset classes including commodities and hedge funds should be permitted. In many cases, these assets will have higher risk profiles than assets which can be held by insurers, which should be addressed through appropriate capital requirements to reflect the risk borne. Where the investment risk is borne by the policyholder (i.e. investment-linked business), the capital requirement would be expected to be low. Issue 11: Regarding the ceilings for asset diversification (Art. 24), a) do we need limits in a future prudent person approach? If the answer is yes, b) how should we define them, and c) should these levels be compulsory (i.e. maximum harmonisation rather than minimum as today)? In view of the application of the prudent person principle, any quantitative limits should largely be refrained from, if concentration risk is taken into account in the calculation of the TCR (please also refer to our comments under issue 1). More work is necessary on the feasibility of the first and second options, or possibly of a compromise between the two (i.e. few, limited quantitative limits in line with the pension funds directive, together with simple indirect charges on unwarranted asset risk concentrations). Issue 12: The available solvency margin is outlined, for example, in Art. 27 of the codified life directive. Do you believe that modifications would be needed? We agree with the Commission s choice of the first alternative, whereby changes are necessary to take the new Solvency II regime as well as international developments into account. Full harmonisation shall be the target, but there is a need not to restrict eligible elements of capital unduly. In addition, there should be adequate consultation of the industry on these matters. 5. CEA comments on Section 5 Supervisory authorities powers Issue 13: Should additional powers be given to supervisors? If yes, what additional powers should be given to supervisors? We praise the Commission s intention to harmonise supervisory practices and to create standards. In our view, the homogeneous exercise of supervisory powers in Europe is an important element in creating a level playing field. Therefore, an EUwide inventory of the current powers of national supervisory authorities seems to be necessary as a basis for further reflections. The end result should be to make management responsible for active risk management, which can best be achieved through making supervisory powers Page 8 of 10

9 converge towards internationally established best practices, rather than on accumulating existing practices. Thus, there should be no general extension of supervisory powers: powers of intervention should be clearly defined and should depend on the solvency situation of the individual company. Clear and workable supervisory objectives should be laid down not least to ensure consistent enforcement across the EU. The responsibilities of supervisors should not interfere with company management, irrespective of the situation of the company, as the new solvency architecture implies a control proportional to the overall solvency of the company. For instance, off-site inspections should be seen as the general rule and on site inspection as the exception as long as the company covers its capital requirement. In the context where a TCR is seen as a soft target (please refer to our comments on issues 15 and 17 below), the relationship between supervisor and supervised has to evolve towards one of ongoing co-operation. Issue 14: Should the supervisors powers (currently held or increased) be reflected in a new general article of the directive in addition to the current existing ones mentioning supervisory powers? Supervisory powers and the framework in which they will be exercised should be reflected in the directive to increase the harmonisation of supervisory standards. Transparency of supervisory action is also desirable as it should encourage a standard approach to be adopted. Finally, in the paper the focus is on the requirements for single companies. The issue of groups in a single market is extremely important and should be addressed more profoundly. Indeed, in the case of groups it is important to shift the focus from the requirements for single companies to a group and financial conglomerate level. As mentioned in our general comments, this is essential in order to take into account the benefits of the groups structure and to avoid the flaws which characterise the insurance groups directive. This further reinforces the aim of convergence centred on established best practices. 6. CEA comments on Section 6 Solvency control levels Issue 15: Do you agree with the outlined approach? Issue 16: The Commission seeks CEIOPS advice on the number and the definition of the solvency control levels. First of all we will stress that CEA sees the TCR as a group concept. In other words, in the case of insurance groups, the group level TCR should be the requirement on which the supervision will be based. Indeed, this is the only way to take into account the group diversification effects (see also our response to Issue 14). While we understand the need for setting control levels between the target and the minimum level, we do not agree with the proposal of setting control levels over and above the TCR. There should be no supervisory interventions above the TCR level in addition to the normal disclosure and checking of the reliability of the information and Page 9 of 10

10 company management. Otherwise, this would ignore the fact that supervision should intensify the more a company s financial strength diminishes (rather than the opposite). Furthermore, the TCR should be closer to the economic capital and should serve as an early warning type limit. It should be acceptable that the company may occasionally go below the TCR without being declared unhealthy by supervisors or market analysts. In other words, the TCR should be seen as a soft target and not as a hard limit. One reason for this is that if the companies were expected always to hold more capital than the TCR, this would represent a double capital requirement in the sense that companies would then consider the TCR (instead of the MSM) as the reference value for minimum capital and therefore, in practice, would hold capital perhaps about twice the TCR. The severity of possible supervisory actions below the TCR should increase when the actual capital decreases and goes towards the MSM. Perhaps there should be a midlevel (defined in Pillar II?) between the TCR and the MSM, which would serve as a checking-point for more serious supervisory measures. As in previous CEA comments 1, we refer to the World Bank s Supervisory Ladder 2 (Annex) for a reflection on the different levels and the corresponding supervisors possibilities to intervene. However, between TCR and MSM the supervisory interventions should mainly be optional (the supervisor should have the power but not an obligation to use its powers). The level of the MSM should be set so low that going below the MSM should always trigger automatic intervention, though this intervention should not be automatic license withdrawal. As such, the proposal that in the case of dropping below the MSM the principlesbased regulations should be switched to more rules-based regulations seems appropriate in order to protect the interests of policyholders, presupposing that the level of the MSM is calibrated to correspond to a real emergency situation. Finally, we do not favour the possibility of setting control levels for each of the different elements of available capital as this would create unnecessary complication. Issue 17: The Commission seeks CEIOPS advice on the consequences (actions and/or sanctions) that should be attached to each level. This raises serious worries because of potential conflicts of interest, as supervisors should not be asked to determine the modalities of their powers of action and sanction. * * * 1 CEA comments on the European Commission note Considerations on the Design of a Future Prudential Supervisory System, 20 th February, In its study Re-engineering Insurance Supervision. Page 10 of 10