Level 2 Implementing measures CEA Comments on the Impact Assessment

Transcription

1 Level 2 Implementing measures CEA Comments on the Impact Assessment CEA reference: ECO-SLV Date: 01 February 2011 Referring to: Solvency II Contact person: ECOFIN Department ecofin@cea.eu Pages: 27 The Commission Services are currently developing the level 2 implementing measures that will elaborate the principles that have been set out in the Solvency II Framework Directive. The Commission's proposal for level 2 implementing measures will be accompanied by a full Impact Assessment. The Commission has invited stakeholders to provide their feedback on a document which is a working document of DG Internal Market and Services of the European Commission (the Commission Services), prepared for the purpose of obtaining views of interested parties. This papers contains CEA s feedback on the working document of DG Internal Market and Services of the European Commission (the Commission Services), prepared for the purpose of obtaining views of interested parties on the Impact Assessment for Level 2 implementing measures for Solvency II. We appreciate this opportunity to comment on this document and would like to highlight that the views expressed in this document represent CEA s views at this stage of the project. As our work develops, these views may evolve. Also these views need to be considered in the context of the CEA s other publications, which can be found on the CEA s website: In particular relevant information on CEA s views can also be found in: CEA s paper on Why excessive capital requirements harm consumers, insurers and the economy CEA s paper on How to make Solvency II workable for all CEA aisbl Square de Meeûs, 29, B-1000 BRUSSELS, Belgium Tel: Fax:

2 SECTION 2 Policy Issues Question 1: Technical provisions best estimate risk-free interest rate curve Option 4 Option 5 Use the swap curve Use the government bonds curve Use the swap curve with an adjustment Use the government bonds curve with an adjustment A combination of the previous options The CEA supports (i.e. the use of the swap curve) plus the addition of the illiquidity premium. While, in principle, we would be prepared to accept an adjustment to the swap curve for credit risk (i.e. option 3), in practice this adjustment is expected to be immaterial (a long-term through-the-cycle assessment of credit risk would be expected to be of the order of 10-12bp). Therefore, we question whether it is necessary to make an adjustment to the swap curve. A swap-based risk-free term structure is preferable to government bonds, since the swaps market generally has a much greater supply of matching instruments from a wider range of counterparties. The swaps market does not rely upon a single provider, whereas the government bond market does. Additionally, swap-based discount rates are also used to price traded instruments, so if a different riskfree term structure is used, then technical provisions will be inconsistent with market prices. For example, if the cashflows under an equity put option were valued using a risk-free term structure based on AAArated government bonds, the resulting value would differ from the price observed in the market. With regards to the addition of an illiquidity premium, it is of vital importance that an objective method is used, which is known in advance, so that industry can calculate the illiquidity premium at any point in time using data observed in the financial markets. If insurers do not know in advance with sufficient certainty whether an illiquidity premium will apply, it will be very difficult for insurers to manage their capital planning and asset/liability management strategies and disappointingly, insurers may end up having to assume that the illiquidity premium is always nil in their capital planning. Therefore, the Illiquidity premium should be applied based on a formula set out in the level 2 text - it should not be applied only when Eiopa determines there is a period of stress. Tasking EIOPA with determining whether a market is in stress is a subjective judgement and causes significant uncertainty. A formula (subject to periodic review by Eiopa) should be allowed to operate automatically to determine the level of the illiquidity premium relative to market conditions. Furthermore, there would also undoubtedly be a significant delay between the occurrence of a period of stressed illiquidity and the approval by Eiopa that a stress has occurred. Additionally, an announcement by Eiopa that the markets are in stress will be a very political decision which runs the risk of exacerbating volatility in the market. We note that the illiquidity premium should be fully applied up until the point that extrapolation of the risk-free curve begins and there should be a smooth transition between the illiquidity premium and the extrapolated part of the curve. Furthermore, the illiquidity premium should be added to the forward curve, rather than the QIS5 proposal which was to add it to the risk-free spot curve. On the issue of achieving the listed objectives set out by the Commission, we note: It is important that also the liquidity premium is set in a harmonized manner and that there is consistency between different countries 2 of 27

3 It is too early in the IASB process to say whether it will be compatible with this choice. Finally, we note that this is a key issue and it is very important to make sure sufficient consideration is given to the impact of the design of the risk-free rate on the cost for insurers to supply, and thus for policyholders to purchase, insurance products. Particular consideration is needed for the impact on longterm savings products, especially those providing retirement benefits, given the European Commission s aim to ensure all EU citizens can access affordable pension solutions. It would also be important to consider the impact of the design of the risk-free rate on a level playing-field for insurers with other financial services providers. Questions 2 and 3: Technical provisions risk margin Cost-of-Capital rate The level of the Cost of Capital rate should be equal to 6%, as specified in QIS4 The level of the Cost of Capital rate should be lower than 4% The level of the Cost of Capital rate should be higher than 6% The CEA supports (i.e. a 4% Cost-of-Capital rate) which is in line with the previous CRO Forum study 1 on this issue. A 6% Cost-of-Capital rate appears to be too high and has not been sufficiently justified by Ceiops. It is important that the final Cost-of-Capital rate is calculated based on a transparent methodology and that this figure be regularly reviewed and updated, as required by the level 1 text. We note that the Cost-of-Capital rate will have significant impact on the balance sheet. While the difference between 4% and 6% may be less than 1% of technical provisions this could quite typically be more than 10% of the SCR capital requirement. We do not agree with the comment attributed to Deloitte that with regard to life insurance only a small impact on financial requirements is expected for most product types. A cost of capital at 6% would not be in line with rates typically used by market analysts for high standing companies or with spreads on typical corporate bonds rated investment grade. Further, spreads on bonds reflect an asymmetric outcome to risk (in that bondholders receive no additional benefit on a favourable outcome) whereas the outcome to the reference undertaking taking on the insurance obligations, per Article 77(3) of the solvency II Directive, would be symmetrical (i.e. the reference undertaking would retain benefit on favourable outcomes as well as losses on unfavourable outcomes). These observations indicate that 6% is too high. Question 4: Technical provisions risk margin diversification Assume reference undertaking is well-diversified Assume reference undertaking after transfer is a mirror image of insurer transferring the risk Assume reference undertaking is empty before transfer An economic approach requires full recognition of diversification, which is one of the key principles of Solvency II of 27

4 Towers Perrin carried out an analysis of historical transactions of insurance business 2. This analysis showed that in the vast majority of historical cases, insurers transferred their whole portfolio of business, which would support the recognition of full entity level diversification within the risk margin. So, Option 2 reflects the minimum diversification which should be reflected in the risk margin. Furthermore, we note that on top of this, the reference entity would NOT be expected to be an empty shell, so the acquiring insurer would benefit from diversification effects between the acquired business and their existing portfolio. So, would reflect the most appropriate assumption, although we understand that it may be challenging to define this well diversified portfolio. Therefore, the CEA supports as a very minimum. Diversification between the existing undertaking s lines of business should be reflected within the Risk Margin, which should include diversification across entities of a group. Questions 5 and 6: Own funds quantitative limits for SCR and MCR Option 4 Option 5 SCR: min 1/3 T1 (=> max 2/3 T2) and max 1/3 T3 MCR: min 50% T1 SCR: min 50% T1 (=> max 50% T2) and max 25% T3 MCR: min 50% T1 SCR: min 50% T1 (=> max 50% T2) and max 20% T3 MCR: min 80 % T1 SCR: min 50% T1 (=> max 50% T2) and max 15% T3 MCR: min 100 % T1 A combination of the previous options The CEA supports. There is no evidence suggesting the limit structure proposed in the Framework Directive should be changed. Tier 1 should cover at least one third of the SCR and Tier 3 should be eligible to cover up to one third of the SCR as suggested by the Framework Directive. There is equally no justification for arbitrarily setting the limit of Tier 1 capital covering the MCR to 80%. Eligibility limits need to be considered together with the classification criteria proposed for each tier i.e. they constitute a package. Any eligibility limits which would be more conservative than the ones set in the Framework Directive combined with the classification criteria which are currently being discussed under level 2 for each tier would severely hamper the ability of insurers to raise capital at a reasonable price for their policyholders. We have indeed several concerns on the required criteria for Tier 1 set out in the implementing measures: The item must either automatically convert into ordinary share or be written down at a trigger point set at 75% of the SCR. We disagree with this requirement as introducing an automatic trigger is inconsistent with the purpose of the ladder of supervisory intervention. How an undertaking best recovers the level of the SCR should be assessed on a case-by-case basis and should be part of the recovery plan approved by the supervisor. Approval by supervisors is required to redeem or repay the instrument. We disagree with this requirement as it is unjustifiably burdensome to require supervisory approval of repayments and redemptions on a going concern basis, particularly at maturity of 27

5 The maturity date is deemed to be the first opportunity to repay or redeem the instrument. We disagree with this requirement as for many current instruments a first call is standard after 5 years rendering them unable to qualify as Tier 1. No allowance for any incentive to redeem. We disagree with this requirement and instead believe that moderate step-ups should be allowed in Tier 1. This is because the call is optional, and would only be exercised if market conditions allow undertakings to refinance at a lower cost than the new stepped-up coupon. Under stressed conditions this is very unlikely to be the case. The insurance or reinsurance undertaking should have full discretion over payment of dividends on basic own funds. We disagree with this requirement because this is impossible in many countries where corporate law defines that this is the exclusive right of shareholders. Other aspects of the regime would also need to be considered when assessing the impact of limits for tiers. Issues which are currently being discussed under Level 2 such as the contract boundary definition and the liquidity premium for the discount rate are likely to have a great impact on the valuation of technical provisions and as such on the capital available to cover the SCR. For example, if it is decided that the illiquidity premium is zero, except in times of stress at the discretion of Eiopa, we can expect that the cost of own funds will increase and capital raising will become more likely. Finally, the impact of the limits set needs to be considered within the context of pro-cyclicality. Having limits which are too restrictive can increase pro-cyclical reactions to events, if following a stress, not only the total own funds are reduced, but also there is a second order effect due to the fact that the limits are breached. Questions 7 and 8: Procyclicality Pillar II dampener Related to the maximum period of time which supervisors should be able to give undertakings to reestablish the level of eligible own funds covering the SCR in the event of an exceptional fall in financial markets. Option 4 15 months i.e. 6+3 (in normal circumstances) + another 6 (in the event of exceptional market falls) Between 15 and 24 months i.e. 6+3 (in normal circumstances) + another 6 to 15 months (in the event of exceptional market falls) Between 24 and 36 months i.e. 6+3 (in normal circumstances) + another 15 to 27 months (in the event of exceptional market falls) Between 36 and 60 months i.e. 6+3 (in normal circumstances) + another 27 to 51 months (in the event of exceptional market falls) The CEA supports the introduction of the Pillar II dampener. However, we do not support any of the specific options listed by the Commission. We see no need for setting the maximum period in the Implementing Measures, as the length of the extension should be determined with respect to recovery of the financial markets as a whole. Questions 9, 10, 11 and 12: Supervisory reporting content, form and modalities A*. Content of the qualitative aspects of the Report to the Supervisor (RTS) The RTS on every occasion contains complete information on the subjects specified in section of the July 2009 CP 5 of 27

6 Undertakings will provide a full report for the first year and thereafter on a frequency to be established by the supervisory authority, depending on the risk profile of the undertaking. In the intervening years, undertakings will provide information only on those topics (specified in section of the July 2009 CP) where material changes have occurred, or state that no material changes have occurred. B. Frequency All data is provided quarterly Core quantitative data is provided quarterly, while all quantitative reporting templates and all qualitative data are provided annually All data is provided annually unless more frequent submission is required in the Directive C. Level of assurance All quantitative data are externally audited annually Specific quantitative data are externally audited annually, with the remainder unaudited D. Reporting format Standardised reporting formats for all information Free format reporting for all information Quantitative data in a standardised reporting format and qualitative data following a predefined order but in free format A*. B. C. D. Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Reporting requirements should be designed and applied in line with the proportionality principle. The level 2 implementing measures should only require material and/or relevant information to be provided. For this reason, the CEA supports the following Options: A - The CEA does not support the reporting of all information specified in section of the July 2009 CP. For further information, please refer to the CEA response via the following link. The CEA requests that consideration be given as to what information should be reported on a systematic basis and what information could be requested by supervisory authorities on an ad hoc basis. This could reduce the burden of reporting requirements by directing the requests directly to the affected undertaking rather than becoming a general requirement for all of the industry. The RTS has the potential to become incredibly detailed and material changes may not occur on an annual basis. It would be helpful for undertakings to report only material changes to the supervisor in the case of regular supervisory reporting. B -. The default frequency for supervisory reporting of quantitative reporting templates should be annual. The cost to the industry of implementing the necessary IT systems to support systemic reporting of quarterly data would be significant. Additionally, there may be costs associated with 6 of 27

7 collecting certain data items, for example when the data is not held directly by the undertaking or where subscription to an external fee-paying service would be required. With regards to supplying all data, the scope of this requirement is still under discussion. As a general rule, groups should be given a longer deadline for supervisory reporting as they will first have to collect and process the information from solo undertakings. C- The CEA does not support either of the options outlined under point C. Pillar II of Solvency II will introduce the requirement for undertakings to implement sophisticated internal control and management systems. For the purpose of ensuring that data is of sufficient quality, these functions should provide enough confidence that the procedures upon which data is collected and presented are of a high quality external auditing should not be a requirement. D -. To ensure a harmonised approach it is essential that quantitative data is submitted in a harmonised template form however consideration should be given to the principle of proportionality. It may not be appropriate for all undertakings to complete every data field in the template. Level 2 implementing measures should define the principle of proportionality and indicate, in terms of content, how it should be applied to supervisory reporting. This could then be supplemented with Level 3 Guidelines. For the reporting of qualitative information, flexibility should be allowed for undertakings to report information in a form which best reflects their situation. Question 13, 14 and 15: Public disclosure content, form and modalities Level of detail of SFCR specified in a generic way (brief description of the information to be disclosed in each item of Article 50(1) of the Directive) Level of detail of SFCR identical to the one requested under the RTS (save as non-disclosure allowed for in Article 52) Level of detail of SFCR specified in a concrete way (definition of the minimum content of the information to be disclosed in each item of Article 50(1) of the Directive) Specify where the SFCR will be disclosed and its structure Specify where the SFCR will be disclosed but not its structure The location of the disclosure of the SFCR is left to the undertaking, but its structure is specified With regards to the first box, the CEA supports. A very prescriptive outline of requirements in the Level 2 implementing measures would make it difficult to apply the principle of proportionality. A very detailed level of disclosure would not be appropriate for all undertakings. As a general principle, the CEA believes that public disclosure requirements should be less technical and less detailed than supervisory reporting requirements. The data would be used for different purposes and the level of detail appropriate for each audience will differ. It should not be the case that commercially sensitive data is disclosed to the public. With regards to the second box, the CEA supports. The undertaking should decide where the SFCR is disclosed. There will be overlapping information requirements between the SFCR and the undertaking s annual report and it has been suggested that the SFCR could be presented as an annex to the annual report. This would be more comprehensive for the reader as cross referencing would be easier. 7 of 27

8 Questions 16, 17 and 18: Treatment of holdings in participations and subsidiaries Option 4 Option 5 Apply a differentiated equity stress to all related undertakings Apply a differentiated equity stress to all non-financial and (re)insurance holdings in related undertakings. Apply a different approach to holdings in financial and credit institutions (e.g. deduction/aggregation). Apply a differentiated equity stress to all non-financial holdings in related undertakings. Apply an alternative approach to (re)insurance holdings, which makes use of the additional information available in these cases to determine the holding's contribution to the overall risk profile of the undertaking. Deduct holdings in financial and credit institutions. Apply a differentiated equity stress to all non-financial subsidiaries. Apply a standard equity stress to non-financial participations, which are not subsidiaries. Apply an alternative approach to (re)insurance holdings in subsidiaries and participations, which makes use of the additional information available in these cases to determine the holding's contribution to the overall risk profile of the undertaking. Deduct holdings in financial and credit institutions A combination of options The CEA supports. We do not support the proposal to define strategic participations and to define a differentiated equity stress only for these participations, as tested in QIS5. All participations are strategic by their very nature. In our view, the introducing a definition of strategic participations is inconsistent with the Level 1 text. An investment by an insurance or reinsurance undertaking in more than 20% of the equity of a company is a substantial commitment and not one that is made with the intention of selling the stake. Such investments are intrinsically of a long term nature. We note also that the requirements in Level 1 for the definition of a participation (>20% holding) are higher than in common corporate laws (usually >10%), which reinforces the strategic nature by default for all participations defined under Solvency II. Given that participations which are subject to group supervision are treated properly at group level, only a simple approach needs to be adopted at solo level. The detail required to prove that a participation is strategic seems excessive and disproportionate given the end use - a single shock for all participations is simple to apply and can be applied consistently across undertakings and Member States. Applying a 22% shock to participations should be seen as parallel to the fact that a 22% shock has been considered appropriate for the duration based equity shock, which is a shock applied to equities held for a long time. Furthermore, with regard to the proposals to apply an alternative approach to (re)insurance holdings in subsidiaries and participations, which makes use of the additional information available in these cases, we note again that participations are treated appropriately at group level. We do not support an alternative approach which adds to the complexity of the treatment at solo level, and would be inconsistent with the Level 1 Text, since it could lead to capital requirements for participations that are higher than those resulting from the application of the standard equity shock. 8 of 27

9 With regards to the treatment of financial and credit institutions, we disagree with a proposal to deduct holdings. We do not see the rationale behind a specific treatment for these participations, which creates an unlevel playing field with banks. It is important that the real risk is reflected in the solvency requirements, and the organisational form should not be material for the capital requirements. For example a large part of asset management companies could be organised as financial divisions within the life insurance company. Requiring full deduction of the insurance companies own funds will give an incentive to organisational change, which is presumably not the intention of the rules. Therefore, we agree with Deloitte s conclusion that the choice of policy option may cause some undertakings to change their corporate structure. We note that with regards to the valuation of participations, we believe that a mark-to-model approach provides a valuation that better reflects the economic balance sheet of the participations and therefore is more in line with the spirit of the Directive. Therefore, if a quoted market value is not available, a markto-model approach should be used, rather than the QIS5 proposal to move straight to the adjusted equity method, which could be very onerous for non-listed participations. Finally, we agree with Deloitte s conclusion that the choice of policy options may cause some undertakings to change their corporate structure. Should the treatment of participations in Solvency II be penalising, groups would have an incentive to restructure, reducing the number of single entities within the group, or transforming them into branches. Question 19: SCR standard formula equity risk Pillar I dampener Option 4 Option 5 Option 6 Less than 3 months Between 3 and 6 months Between 6 and 12 months Exactly 12 months Between 12 and 36 months More than 36 months The CEA supports Option 6 - a symmetric adjustment based on a time period of 36 months as tested in QIS5. A 36 months period would result in a more stable and effective mechanism. Question 20: SCR standard formula loss-absorbing capacity of technical provisions A "one-off adjustment" (based on a "k-factor") is applied to the amount of technical provisions (as tested in QIS2) An approach ("kc-factor" approach) where individual reductions of the SCR capital charge are calculated for each possible risk module and sub-modules of the standard formula, are then deducted from each risk module or sub-module SCR charges, and aggregated using the linear correlation matrices (as the one tested in QIS3 and the more refined one tested in QIS4) An adjustment based on the simulation of a single equivalent scenario (as the alternative method tested in QIS4 see TS.VIII.C.8) The CEA supports option 2. Based on the initial feedback received based on QIS5, the CEA supports a modular approach as the Single Equivalent Scenario does not appear to be a workable solution for the standard formula. 9 of 27

10 Question 21: SCR standard formula diversification effects correlation parameters Use QIS4 correlation parameters across lines of business Use lower than QIS4 correlation parameters across lines of business Use higher than QIS4 correlation parameters across lines of business The CEA supports option 2. This is because we believe that when risks can be considered as independent, the correlation proposed in QIS4 was conservatively set 25%. We believe that this approach resulted in correlation factors which are not in line with the level of calibration required by the Framework Directive. When two risks can be considered to be independent from each other and until there is no evidence suggesting otherwise then the correlation parameter should be set at zero. It is vital that, to the extent that diversification benefits exist, they should be recognised. This is necessary, because companies must maintain proper incentives to spread their risks. Question 22: SCR standard formula diversification effects geographical diversification No recognition of geographical diversification Recognition of geographical diversification as per QIS4 approach (TS.XIII.B; TS.XVI.B default method accounting consolidation) Recognition of geographical diversification using a more granular approach than QIS4 We support option 3. This is because we believe that at least for the EEA a more granular approach is needed. We believe for instance that it is inappropriate to consider France and Germany as a single geographical area with no diversification effects recognised at all between risks. It is vital that, to the extent that diversification benefits exist, they should be recognised. This is necessary, because companies must maintain proper incentives to spread their risks. Different levels of granularity should be allowed depending on how material diversification is and on the degree of reliability of the data available per geographical area. This approach would constitute an incentive to improve the overall quality of data, in particular for undertakings where this effect is material. Undertakings where geographical diversification effects are immaterial, should have the option to use only two geographical areas: home country and rest of the world. Question 23: SCR internal models integration of partial internal models Integration of partial internal models using only coefficients prescribed by supervisory authorities. Integration of partial internal models using techniques provided by supervisory authorities or if these are not possible or there is strong evidence that these are inappropriate dependency structures and parameters provided by the undertaking. Integration of partial internal models using dependency structures and parameters provided by the undertaking or if these are not approved by the supervisory authority - techniques provided by supervisory authorities. The CEA supports option of 27

11 Undertakings should be allowed to use their own integration methods validated by supervisors. It is important that the process is designed so that it does not become too burdensome. The benefit of a partial internal model should always be greater than the cost of integrating the model. The range of approaches prescribed by supervisory authorities, and from which undertakings will have to choose, will never be wide enough to include all approaches that may potentially be appropriate for certain undertakings depending on their risk profile. On the other hand, it should be noted that the burden on undertakings would be all the more important if the prescribed list of methods contains a too wide range of sophisticated methods of which undertakings would be required to demonstrate their inappropriateness before being able to use their own integration technique. would be very burdensome and likely to cause undue delays in the approval process which is already heavy enough. is contrary to the purpose of using internal models. The aim of partial models should be to facilitate and encourage more sophisticated risk management, which should include how companies integrate the internal and standard formula components in their partial models. Question 24: SCR standard formula underwriting risk (other than catastrophe risk) arising from non-life insurance obligations Simulation of the impact of a pre-defined shock on the financial position of the (re)insurance undertaking (Scenario based approach) Closed formula calibrated to a VaR at the 99.5% confidence level over a one-year period (Factor based approach) In general we support option 2. A factor based approach presents the following advantages: it is generally easier to calibrate it allows for an easier use of USPs it is easier to integrate with internal model outcomes for which simulations have been used. It is vital however that suitable method is retained to capture the risk mitigating effects of reinsurance. For example, standard adjustment factors for different ranges of excess points should be provided based on market data. However, we should note that if the proposed factor based approach is not suitable, it should be possible to use USPs, which could be in the form of a scenario based approach. This may be a preferable solution to capture the risk mitigating effects of reinsurance or if no data is available to be able to calibrate LoBs using factor based approaches. It is vital that all methods are developed in cooperation with the industry. Question 25: SCR standard formula underwriting risk (other than catastrophe risk) arising from life insurance obligations Simulation of the impact of a pre-defined shock on the financial position of the (re)insurance undertaking (Scenario based approach) Closed formula calibrated to a VaR at the 99.5% confidence level over a one-year period (Factor based approach) 11 of 27

12 The CEA supports option 1. Question 26: SCR standard formula underwriting catastrophe risks arising from insurance obligations Simulation of the impact of a pre-defined shock on the financial position of the (re)insurance undertaking (Scenario based approach) Closed formula calibrated to a VaR at the 99.5% confidence level over a one-year period (Factor based approach) A combination of the previous options For life underwriting CAT risk we support option 1. However, given the fact that the exposures to CAT risk differ significantly from one undertaking to another, the use of undertaking-specific scenarios for CAT risk should be allowed subject to supervisory approval where they better capture the risk profile of undertakings. Undertaking s Specific Scenarios are of particular importance for international groups and reinsurers when standardised scenarios are not available for third countries. It is important that international groups and reinsurers are not penalised compared to other international competitors not applying Solvency II. Question 27: SCR internal models use test As a minimum requirement, the internal model is to be used at the topmost organisational level of the undertaking. The model is to be used, for instance: in setting the risk strategy; allocating risk capital; and/or taking strategic business decisions. The internal model is to be used at all levels of organisation. The areas or processes in which the undertaking has to make use of its internal model are comprehensive and mandatory for all undertakings and include, as an example, the pricing of individual insurance contracts We agree to the Commission s view to apply option 2 on a modified basis, i.e. requiring from undertakings to demonstrate that their internal is fully integrated in the risk-management system, by using it throughout the organization, without imposing a list of mandatory uses for the internal model. A mandatory list would be contrary to the principle based approach needed for Solvency II. In any case, it cannot be required that decisions are always in line with the recommendations derived from a internal model because such recommendations are just some amongst many others that need to be considered for business decisions. It should be expected that undertakings will have a number of tools used to support decisions made within in the business. 12 of 27

13 Question 28: SCR internal models statistical quality standards Option 4 Firms should check the quality and source of all data (internal, external) as well as expert judgement. Firms should agree the use of internal and external data and expert judgement with the supervisor on a case-by-case basis. Undertakings establish their own policy on data quality in line with general supervisory principles. The policy specifies the respective data sources (internal, external) and use of expert judgements, as well as the methods used and the responsibilities for validating the data and expert judgements. Furthermore, the interrelation between data and expert judgement must be addressed. The policy as well as major changes to it, are subject to supervisory approval. Internal as well as external data and the use of expert judgement must be reviewed by an independent third party. Expert judgement may be used in all areas. The use of expert judgement must be well-justified, explained and documented. In particular, when data is available, expert judgement must be reconciled with the data. Internal as well as external data and the use of expert judgement must be reviewed by an independent third party. The use of expert judgement should be kept to a minimum and is only allowed when data is unavailable. It must be well-justified, explained and documented. The CEA supports option 2. Insurance data are never perfect and realistically never will be. Professional judgement is an essential element in the collection and interpretation of data. Questions 29 and 30: Capital add-ons Supervisors would take the decision on whether or not to apply a capital add-on on the basis of harmonized criteria established at level 2 Supervisors would take the decision on whether or not to apply a capital add-on on the basis of harmonized criteria established at level 2. A harmonised reference value of [5%-15% 2] of the overall SCR is established at level 2. This reference value serves as a presumption that the deviatio is significant. Supervisors would only consider deviations that exceed this quantitative threshold A harmonized reference value of [5%-15%3] of the overall SCR is determined at level 2. This reference value serves as a rebuttable presumption that the deviation is significant. Supervisors may decide to depart from it (on both ways) based on the application of harmonized criteria established at level 2 The CEA supports option 2. Capital Add-ons should be used as a last resort measure and as such, it is important for both supervisors and undertakings to understand exactly when such a requirement might apply. It would also be useful to develop a qualitative assessment which would help the supervisor to better understand the situation of the undertaking. Assessment of a significant deviation from the standard formula, internal model, and the system governance, should be proportionate to the nature, scale and complexity of their business, and in particular to the risks inherent in that business. In terms of remedying deficiencies in governance systems, the CEA would support a flexible approach since the time required may depend on the nature, scale and complexity of the undertaking. It may be required to investigate interdependencies between the different elements of the governance system. Capital Add-ons should be used as a temporary measure and once an undertaking has corrected the deficiency, it should be possible to consult with their supervisor about its removal. The unnecessary 13 of 27

14 application of capital add-ons would strain an undertaking s capital resources in terms of required versus available capital. It could also have reputational consequences to the undertaking. Questions 31, 32 and 33: Actuarial function A) The standards to be applied by the function The function should use technical standards developed by CEIOPS on Level 3 The function should rely on technical standards that are widely accepted in the industry and the profession The function should rely on European technical standards to be developed and endorsed by a body of representatives of different stakeholders, including CEIOPS B) The scope of the tasks of the actuarial function It should be left to undertakings to decide on the scope of these tasks individually The general scope of the tasks should be prescribed on Level 2 to some extent C) The reporting of the actuarial function Require annual reporting with definition on Level 2 of its structure and content Require annual reporting but leave the decision on the details up to the undertakings The CEA supports option 2. The actuarial function should rely on technical standards widely accepted by the industry and the profession. In the rather extreme case where such guidance for the actuarial function will generate significant distortions, we believe that the actuarial guidelines should be constructed by interaction of the stakeholders (The European Commission, the European Parliament, CEIOPS, industry, profession). An open process is essential to avoid any one party following ill conceived ideas. There is also a need to avoid unnecessarily subjecting the work of actuaries to multiple standards and guidelines from CEIOPS, the profession, auditors, IASB. The scope of the tasks of the actuarial function should be left to the undertakings to decide (option 1). This would allow for the scope to set in accordance with the principle of proportionality. We support the proposal that the actuarial function should, at least annually, produce a written report to the administrative, management or supervisory body. Decision on the details of the report should be left to the undertakings (option 2). Guidelines should relate solely to technical issues. Professional and ethical matters are more appropriately dealt with by the actuarial profession itself. Questions 34, 35 and 36: Supervisory co-operation and co-ordination A) Membership of branches to the College Level 2 measures should include binding quantitative thresholds for the determination of significant branches Level 2 measures should include indicative thresholds (quantitative and/or qualitative) for the determination of significant branches 14 of 27

15 B) Frequency of college meetings Level 2 measures shall establish a minimum frequency Frequency of meetings and contacts between supervisors shall depend on the risk-based assessment made by the college, but should take place at least annually In order to support the process of supervisory convergence, the CEA believes that binding commitments in terms of: the decision making process; the frequency of meetings; and how information will be exchanged, should be specified under Level 2 implementing measures. The CEA supports the Commission Services suggested approach. The CEA is concerned that the membership of colleges may become too large which could also impact upon efficiency, on the other hand, it is important that supervisors with a vested interest are involved in the College. Thus, supervisors of significant branches should be included in the College. The gross written premium of a branch could be used as an indication of significance. Where a branch exceeds 10% of the gross written premium of the group, the supervisor of this branch should be able to join the College. On the frequency of meetings, the CEA supports option 2. In order to ensure that colleges become efficient fora, the minimum information to be exchanged envisaged by the Commission should be set at level 2 However it is crucial that, for any information to be exchanged beyond the minimum information envisaged by the Commission there is some discretion allowed to enable supervisors to decide what information should be shared within the college. Only information which is necessary for the college to obtain an overall oversight of the European group and key regulatory issues should be shared within the college. 15 of 27

16 SECTION 3 Impact on Insurance markets and products Please note that the CEA has not produced any more specific or systematic survey of the impact on the insurance markets of the implementation of Solvency II. The following answers are based on input from our members on their assessment of possible impact. We agree that the introduction of a robust economic risk-based approach will be beneficial to the European insurance industry. However, due attention should be paid, at this stage, to the objective to make the implementation of the regime workable for all insurers. The feedback that the CEA is increasingly receiving from its members, particularly in the light of QIS5, is that the current proposals will be too burdensome, too complex and too expensive for them to carry out. It is important to ensure that complexity is reduced. The impact on markets and products will, among other factors, depend on the design of the risk-free interest curve including the illiquidity premium, the design and calibration of the SCR and MCR and the definition of own funds elements. Impact on products Question 37: Do you anticipate that the Commission Services' suggested approach for level 2 implementing measures would result in an increase or decrease in insurance prices? (Please provide details of the types of product or groups of policy holders affected, the magnitude of the increase or decrease expected and whether the change results from change in the value of technical provisions or capital requirements) The impact on insurance prices will vary from one product to another. The main driver in this respect is likely to be the changes in calculation of technical provisions. For life products the level of the impact will depend on the outcome of the finalisation of the Implementing Measures, in particular the risk free rate, the liquidity premium, the contract boundaries and the treatment of future profits in future premiums. Particular consideration is needed for the impact on long-term savings products, especially those providing retirement benefits, given the European Commission s aim to ensure all EU citizens can access affordable pension solutions. We believe that the current measures will lead to an increase in annuity prices resulting from a significant increase in the technical provisions and capital requirements. This will result in lower demand for traditional annuity products and we may observe a shift to products that place risks with the policyholders. Similarly, we should equally expect that future products incorporate lower guaranteed rates if none at all. 16 of 27

17 If insurers are incentivised by regulations to invest heavily in low-return assets, they would need to raise life and non-life premium rates to compensate for lower investment returns. Earning lower investment returns would also reduce the returns that life insurers could offer their policyholders.. It should also be noted that an inappropriate treatment of future profits in future premiums could equally undermine the role of insurers as providers of long term retirement and savings products. Indeed, the classification of future profits in future premiums in any tier other than tier 1 is likely to result in insurers shifting from products with recurring premiums towards single premium product for which risks are not double counted between the treatment of own funds and the solvency capital requirements. Regarding non life products, Motor TPL and other long-tail LOBs - especially because of the high reserve risk calibration - might end up with excessive capital requirements, paving the way to price increases. This would be particularly concerning in case of compulsory insurance covers. Question 38: Would the Commission Services' suggested approach for level 2 implementing measures result in a reduction of cross-subsidisation between different lines of business or groups of policy holders? (Please provide details of which lines of business or groups of policy holders will be most affected and the reasons for this) In general we expect that there will be a reduction in cross-subsidisation at an industry level. However, for some markets there is a significant likelihood of implicit or explicit cross-subsidization between in-force business and new contracts. Due to the long duration nature of life insurance contracts companies will be bound to their obligation from previously written contracts, often over decades. In many cases companies are bound to features, such as high interest rate guarantees. If these contracts do not earn sufficient return on the (Solvency II) required capital, companies will be tempted to subsidize these contracts by increased prices on new business. In some cases national insurance law regulations even may enforce such a cross-subsidisation, e.g. by demanding to provide uniform policyholder participation rates over the full portfolio of contracts, independent of differences in the riskiness of different sub-portfolios. Question 39: Would the Commission Services' suggested approach for level 2 implementing measures stimulate product innovation? (Please provide examples of the type of product innovation that is expected and details of the lines of business that this product innovation will relate to) 17 of 27

18 In general, there may be a trend towards products where asset risk is transferred to policyholders. Increased product innovation may take place in some markers to replace traditional annuity products with for example income drawdown products. In other markets, on the other hand the impact may be that some products such as products which offer long term guarantees will be simply withdrawn from the market leading to less competition and in turn to reduced product innovation. Question 40 Would the Commission Services' suggested approach for level 2 implementing measures result in a withdrawal of certain products from the market? (Please provide reasons and examples of products that may be withdrawn) If the design or calibration of the system is incorrect leading to higher than economically justified capital requirements then companies may respond by withdrawing certain capital intensive products or passing the additional costs to policyholders. Some insurers may withdraw from guaranteed products and from the annuity market. Also, Products where the number of policyholders is too small to allow for diversification may also become less attractive. Motor TPL and other long-tail LOBs - especially because of the high reserve risk calibration - might turn out to be highly capital intensive. In case the implementation of price increases is impossible, companies may reduce their exposure or withdraw from the market. This would be particularly worrying in case of compulsory insurance covers. Impact on markets Question 41 Would the Commission Services' suggested approach for level 2 implementing measures promote particular types of insurance business model (e.g. specialisation vs. diversification, joint-stock companies vs. mutuals, branches vs. subsidiaries, groups vs. single legal entities)? (Please provide reasons and examples) The proposed measures will promote diversified business models. Companies would be encouraged to enter business lines and new asset classes to maximise their diversification benefits. Whilst we consider this to be appropriate as diversified insurers operate a lower risk business model, it needs to be noted that companies will have to build up the adequate capabilities to control these strategic changes which will initially come at a cost. 18 of 27