Cécile Sourbès speaks with Gabriel Bernardino, chairman at the European Insurance and Occupational Pensions Authority (EIOPA), which launched the first quantitative impact study (QIS) for the revised IORP Directive in mid-October.
Cécile Sourbès: What are you expecting from this QIS? And how confident are you that it will enable you to assess the risk borne by occupational pensions?
Gabriel Bernardino: In our advice to the European Commission, we already recognised the importance of performing a QIS. In the area of valuation and capital requirements, we even made our advice conditional on the QIS outcomes.
We stressed that the adoption of the holistic balance sheet in practice needed to be further investigated with respect to the feasibility of developing of a methodology for the quantification of the security and benefit adjustment mechanisms and the effectiveness in terms of costs and benefits of such a methodology. Further information is also needed on the feasibility in practice of a common level of security and its effectiveness in terms of costs and benefits, given the diversity of IORPs in the different member states.
So we expect that the results of the QIS will provide a better overview on these issues and help us to understand whether EIOPA needs to offer further views on this matter.
CS: On what type of questions might a second QIS focus?
GB: There are several issues where further work is necessary, and, in our introduction to the draft technical specifications for the QIS for occupational pensions, we mentioned those issues. Most notable are the questions on supervisory responses – including recovery periods – and sponsor support.
Further study is also needed on more technically granular issues such as multi-employer IORPs, the expenses borne by employers, pension protection schemes, minimum capital requirements, the segmentation of pension obligations, discretionary benefits and proportionality.
CS: APG has said EIOPA should introduce a new solvency measure that considers the 'closed fund' aspect of a DB pension scheme. It also said it would be more relevant to consider valuing the HBS in a projected benefit obligations framework. Do you think APG's proposals are viable?
GB: First of all, I did not find that APG was so negative towards the holistic balance sheet concept. Second, in my opinion, APG values the HBS – including sponsor support – and, hence, shows that it can be implemented in practice. So this is already a good reason to praise their work because they don't just say 'no', but make constructive suggestions on how to strengthen the HBS approach.
As regards their suggestions themselves, I can say that supervisory responses indeed were not included, but this will need to be discussed after the QIS. EIOPA does not prescribe Accrued Benefit Obligations (ABOs), but formulates conditions in which cases ABOs or Projected Benefit Obligations (PBO) should be applied.
And finally, in my view, the supervision of valuation models is very important, in order to achieve consistency at EU level.
CS: Could it be something EIOPA might want to look into in future and study in greater detail?
GB: Yes, we will seriously take into consideration the constructive suggestions made by APG.
CS: With respect to the confidence level, the Commission said in its letter to EIOPA for the launch of the first QIS that it was "important" that QIS participants submit results based on a VAR at the 99.5%, 97.5% and 95% confidence levels. Do you welcome these alternatives?
GB: Actually these confidence levels were proposed by EIOPA, and we are glad that the Commission shares our approach. In our proposal, we were based on the most recent estimates for Solvency II in the area of stresses and correlations relating to the risks. In the course of the QIS, the capital requirement under the 99.5% capital requirement will be adjusted by the input spreadsheet using a common method to reflect the lower confidence levels.
CS: Some industry figures have voiced concerns about the 'wisdom' of using a Value-at-Risk (VaR) calculation methodology to calculate the minimum confidence level, since this valuation has traditionally been used by banks and insurers.
GB: Our advice to the European Commission contains an elaborate discussion on the use of the one-year Value-at-Risk approach. In particular, it considers the use of a multi-year time horizon, such as three or five years. A longer reference period may be more suitable for IORPs, since they often have very long-term pension obligations. However, we identified several drawbacks of a multi-year VaR. We reached the conclusion that it would be preferable to take into account the long-term nature of pension obligations through the modelling of equity risk, discount rates and recovery periods.
CS: Is there any other calculation method that could be used?
GB: The one-year Value-at-Risk approach is not ideal, but, at this moment, I am not aware of any superior alternatives.
CS: Regarding reporting obligations, the ECB stressed that such requirements could contribute "significantly" to the information required by the European System of Central Banks (ESCB) under a 'steady-state approach'. However, insurers have expressed concerns over those obligations, which has led EIOPA to review some of the proposals in July this year. Would the reporting obligations be the same under IORP II and in Solvency II?
GB: No, the reporting obligations cannot be the same for insurance companies and IORPs because we do realise the differences between these two kinds of products.
But, for example, the SII principle of having an economic valuation of assets and liabilities should be taken by the IORPs because this can considerably improve their risk management.
CS: Might EIOPA consider launching a QIS on such issues in the future?
GB: The development of a risk-based supervisory regime for IORPs will have different phases. For sure, the elements related to reporting to supervisors will be tested and assessed in due time.
CS: Considering the option that was given to insurers to implement their own internal model, could EIOPA and Brussels give the same flexibility to pension funds under IORP II in the future? For instance, the HBS were introduced within the revised IORP Directive, might pension funds have the option of modelling it to better match their liabilities?
GB: We do explore the potential use of internal models in our QIS. Particularly, our qualitative questionnaire contains questions for IORPs on internal models. I hope the results of the QIS provide us with a clearer picture of whether this option is feasible.