According to article 519 of the CRR the EBA shall assess whether the revised IAS 19 in conjunction with the deduction of net pension assets and changes in net pension liabilities lead to undue volatility of institutions’ own funds. The report appears focused on the deduction of assets and we believe that the EBA should also assess the impact when actuarial losses will result in a net pension liability (according to IAS 19) and hence whether changes in net liabilities lead to volatility in institutions’ own funds.
Since it is not only defined benefit pension funds that has an impact on own funds the ESBG would also ask the EBA to consider all other post-employment schemes covered by IAS 19 including medical care for retirees. An overall impact should include all post-employment schemes and not only defined benefit pension funds.
We believe that the analysis would have been improved by the use of a longer time series. This would have captured any behavioural changes over a cycle thus capturing periods of both economic peaks and troughs. Theoretical examples which would allow for the isolation of individual factors would be appropriate when evaluating the impacts over time. In the appendix we have performed some simulations based on historical Swedish market data to show the difference in outcome compared to the results of the EBA draft study.
As stated under question 1 the report focuses on the deduction of assets. We are of the opinion that the EBA should also assess whether changes in net liabilities lead to volatility in institutions’ own funds. Furthermore, we support the approach of using the full application of the prudential requirements rather than the transitional when calculating the impact on own funds.
Additionally we believe that it would be prudent to update Article 26 § (d) of the CRR to state the following: “accumulated other comprehensive income including actuarial gains and losses.”
We think that the main, immediate, impact on own funds will be experienced by entities which applied the corridor approach. This approach was removed when IAS 19 was amended and all actuarial gains and losses are now accounted for in other comprehensive income.
Both changes to actuarial assumptions and changes to the terms and conditions of pension agree-ments will have an impact on net pension assets or liabilities. Changes in different market parame-ters will also influence net pension assets and liabilities. That said we believe that only market parameters should be the focus when evaluating the cyclical behaviour of defined pension obligations in a going-concern entity. Most actuarial assumptions are long term estimates i.e. estimates for the remaining time of service of the individual employee while market parameters are current parameters rather than estimates.
We agree that the main impact of the amendments to IAS 19 comes from the immediate recognition of actuarial gains or losses in other comprehensive income and the immediate recognition of all past service costs.
We do not agree with the statement on page 18, §38 of the Discussion Paper which says that there is no expected impact on own funds from the changes discussed in §38. All actuarial changes will have an impact on own funds. The amendments to IAS 19 will accelerate these and therefore increase volatility in own funds. If they should be considered in a simulation is another question. Please refer to our answer to question 4 for more details regarding our views on this.
We do not agree with the analysis, as outlined in our answers to questions 4 and 5. We would ask that the EBA and the European Commission, when moving forward with this project take into consideration the following points:
• Consider the impact on both net pension assets and net pension liabilities;
• Use long time series to capture possible pro-cyclical behaviours of defined benefits schemes;
• Use a theoretical sample, the outcome will otherwise become heavily dependent on the present financial position of the institutions covered by the stress test; and
• Focus on market parameters like plan assets and the discount rate. Other parameters will normally not be highly correlated with the market parameters since the market parameters are current values while other parameters are long-term assumptions covering the average expected value during the remaining time of service for the employees.
We agree that the institutions that suffered the main initial impact are those who applied the corri-dor approach and had significant amounts of unrecognised actuarial gains and losses and historical service costs.
We however believe that the methodology and thus the analytical outcome could be much im-proved. The EBA has collected quantitative information for a 3-year period, 2010-2012 and the outcome of their study is shaped by the specific economic situation during the three year period in combination with the fact that the majority of the selected banks held excess values in their pension systems.
ESBG believes that the analysis would be improved if it took into consideration the fact that the IAS 19 calculation is heavily pro-cyclical. Therefore we believe it would be more appropriate to consider a much longer time period in order to capture both periods of economic peaks and troughs. We further believe that the analysis would be improved by the use of a theoretical example designed in such a way that actuarial losses will result in a net pension liability. One such example can be found in Appendix 1 to this position paper.
We agree that the discount rate and the inflation rate are significant actuarial assumptions and we also believe that salary increases and turnover rates are significant. There is no difference between the original and the amended version of IAS 19 in regards to the discount rates, but the removal of the corridor approach has introduced a significant element of volatility into CET1 capital. Other impacts stem from the corporate bond market and the appreciation of what is a deep liquid market within the euro-zone.
In appendix 1 to this position you will find an analysis based on historical changes to the discount rate in Sweden over a period of 27 years. The analysis highlights possible extreme volatility in the CET1 due, solely to changes in the discount rate. We believe that if possible volatility in plan assets due to changes in market values would have been included the volatility would be significantly higher.
The analysis highlights that high interest rate environments may cause higher volatility than low rate environments, but the analysis also shows extreme volatilities in CET1 due to interest rate fluctuations in a low rate environment. The analysis further highlights that severe negative value changes in pension liabilities may coincide with generally stressed market conditions highlighting the pro-cyclical behaviour of IAS 19 without filtering mechanisms in the CRR. A possible way of avoiding these fluctuations would be to filter out all actuarial gains and losses stemming from changes in actuarial rates from CET1. Finally the analysis highlight that a lot of the changes are very temporary in nature. Deferring those temporary changes could therefore significantly reduce pro-cyclicality in the capital situation of banks, and consequently the supply of credit to the real economy under normal economic conditions.