Response to consultation on Guidelines on methods for calculating contributions to Deposit Guarantee Schemes (DGSs)
Go back
Some further details on a range of parameters for risk assessment, e.g. market risk, interest rate risk, foreign exchange risk, operational risk and risk of weak governance, as well as further guidance on appropriate methodologies for the distribution of risk category weightings could be provided.
To accommodate national industry specificities, the EBA should open the methodology for more national flexibility. As mentioned above the national banking systems and the banking industry vary significantly. We believe that the national DGSs and competent supervisory authorities have a good overview about the structure of their banking industry and could benefit to exercise this expertise with a higher level of discretion, i.e. up to 40%-50%.
Finally, we believe that the guidelines would benefit with a fully worked example, including the use, of the “Bucket” and “Sliding scale” methods.
Either scenario, could lead to a significant financial burden for banks resulting in a substantial pro-cyclical effect. This effect is due to the methodology of the formula that does not involve an extrapolation of the needed funds, but involves only a simple one year view.
More specific illustrative examples (e.g. National Averages) to inform on risk weights application methodologies would be helpful.
There is also a concern that risk indicators are used appropriately by DGS both in consideration of the prevailing economic climate and the stability of individual/multiple credit institutions. There is always the possibility of adverse risk indices resulting in higher contributions that impact competitive advantage with the possibility of compounding impacts on business performance between and within national markets.
Further clarity is required in this area, e.g. NPLs will these be gross" or "net" of provisions as this will be very material in some cases with "net" being more reflective of risk and provision coverage ratio crucial in understanding same - dilutive impacts from adverse weighting arising from high NPLs may be inconsistent with risk where an institution has strong capital and provision coverage ratios (how will this be reflected?).
The risk exposure should generally be reflected through a range of indicators which focus on the quantity and quality of capital, liquidity and risk and which are already available as regulatory ratios. Therefore, due consideration should also be given to total regulatory capital, the capital ratios CET1, RWA/total assets and bail-inable debt. In terms of weighting, it is widely accepted that liquidity was the primary driver and little comfort was taken on the levels of capital by funders. The LCR and NSFR will provide an appropriate view of short and long term funding risks. The interrelationship of these measures needs to also be taken into consideration.
We believe that the proposed indicators for the business model assessment could be complemented by indicators such as funding concentration, the extent of the trading book and the share of RWAs for market risk.
About some of the indicators proposed by EBA, and which are not anchored directly in the banking regulation, some doubts could be expressed. This goes notably for the NPL ratio and the unencumberd assets ratio:
• The disadvantage of the NPL ratio has to do with the fact that this ratio cannot be based on a single uniform definition in accordance with regulation, and consequently it may cause an unlevel playing field between Member States.
• On the other hand, the unencumbered assets ratio may have unintended consequences in that it may be to the detriment of banks which, for their funding, depend to a large extent on retail customer deposits, although this kind of funding certainly is less risky than funding via the (short-term) wholesale market. The use of this ratio may also cause unwanted distortions between Member States.
• Furthermore, the concept of unencumbered assets can include assets that are not encumbered but also assets that not being encumbered will never be potentially encumbered. Hence, it should be more realistic to employ the ratio “available assets for encumbrance/covered deposits.” But again, it is not clear how to measure the encumbrance. The report on asset encumbrance is still in progress and although it tries to achieve comparability among institutions there are some questions that are difficult to assess when defining how encumbered or unencumbered an asset is. Therefore, we suggest that the ratio included in Annex 3 of “own funds and eligible liabilities in excess of MREL” should also be taken into account.
• The retained criteria should not only take into account the fact that an institution may fail, but also the fact that the DGS could be drawn (in resolution or in liquidation). In particular, the RTS 41/2014 on MREL provides that systemic institutions will be required to maintain very significant levels of MREL to make sure that the resolution fund would be drawn (if need be). This implies that the probability of SIBs drawing down the DGS is nil.
• That is why the risk factors should, at the very least, include the amount of own funds and eligible debts owned toward the MREL requirement, and a ratio comparing the MREL and regulatory own funds and consider somehow the shortfall with the 8% threshold provided in Article 44 of BRRD.
Finally, the EBF questions in principle the suitability of the risk category “Potential losses for DGS” as an indicator for the risk of failure for an individual institution. In point 44 EBA mentions explicitly that the risk indicators shall measure the risk (of failure) of the individual member institution, not the fund as a whole. It is clear that the potential loss for the DGS is an important parameter, but this risk is covered by the formula already, i.e. by using covered deposits as basis for the calculation of the yearly contribution. The inclusion of potential losses for the DGS in the Aggregate Risk Weight seems to be methodically questionable."
The EBF advocates an evolutionary approach to harmonising the risk based methodology over time while the EBA and DGS gain experience and empirical evidence to support a common approach that would limit inconsistencies across the European Union.
We think that it must be the competent authority who decides the appropriate method for its national banking system also with regard to calibration of buckets for risk indicators, or minimum and maximum values for a sliding scale approach. More generally, we suggest that risk indicators buckets or sliding scale approach result into linear impact and minimise any cliff effects.
Question 2: Do you consider the level of detail of these draft Guidelines to be appropriate?
On the whole EBF considers the level of detail of the Guidelines as appropriate. And while we agree with the core risk indicators in principle, we note there is no statistical proof that the core risk indicators model risk accurately. This may warrant some flexibility until more supporting evidence for these core risk indicators will be established.Some further details on a range of parameters for risk assessment, e.g. market risk, interest rate risk, foreign exchange risk, operational risk and risk of weak governance, as well as further guidance on appropriate methodologies for the distribution of risk category weightings could be provided.
To accommodate national industry specificities, the EBA should open the methodology for more national flexibility. As mentioned above the national banking systems and the banking industry vary significantly. We believe that the national DGSs and competent supervisory authorities have a good overview about the structure of their banking industry and could benefit to exercise this expertise with a higher level of discretion, i.e. up to 40%-50%.
Finally, we believe that the guidelines would benefit with a fully worked example, including the use, of the “Bucket” and “Sliding scale” methods.
Question 3: Is the proposed formula for calculating contributions to DGS sufficiently clear and transparent?
EBF considers the proposed formula for calculating contribution to DGS sufficiently transparent. Not completely clear, however, is the implementation of the adjustment coefficient µ into the annual contribution process with regard to responsibilities (who will determine a relevant economic downturn), to indicators/indicator thresholds (definition of economic downturn) and to the target level. For example, regarding the functioning of the methodology after year 10, if deposits for the whole financial system increase substantially, it means that a contribution of 0.8% of the new deposits will have to be made, which can entail a high impact. Also, in the event when a payout occurs in year 8 or 9 where you need 90% of the financial means of the DGS at this point of time (below the 0.8% threshold) the DGS is forced to collect contributions of almost 0.8 % of covered deposits in one or two years. Will the adjustment coefficient take that impact into consideration?Either scenario, could lead to a significant financial burden for banks resulting in a substantial pro-cyclical effect. This effect is due to the methodology of the formula that does not involve an extrapolation of the needed funds, but involves only a simple one year view.
More specific illustrative examples (e.g. National Averages) to inform on risk weights application methodologies would be helpful.
Question 4: Considering the need for sufficient risk differentiation and consistency across the EU, do you agree on the minimum risk interval (75%-150%) proposed in these Guidelines?
EBF considers the risk interval as appropriate to mitigate moral hazard. We however encourage that this range is subjected to an empirical analysis of a range of credit institutions and is applied in a consistent manner across the EU. This will help to enhance transparency on the range of application. Flexibility should be given to the competent authority to set the national range within EBA’s boundaries on the basis of its own assessment of the banking sector risk profile. However, it is worth noting that the key variable of the formula must remain the share of covered deposits since, if DGS contributions are finally required in insolvency, the extent of the compensations provided will directly depend on the amount of covered deposits for each bank. A too wide interval could create a significant deviation from the contribution calculated according to the key variable, the share of covered deposits.Question 5: Do you agree with the core risk indicators proposed in these Guidelines? If not, please specify your reasons and suggest alternative indicators that can be applied to institutions in all Member States. Do you foresee any unintended consequences that could stem from the suggested indicators?
While agreeing with the core risk indicators on the whole, there is no statistical evidence of the discriminating power of the risk model. The EBF again notes that currently there is no evidence that the risk model (e.g. weights) set by the Guidelines measures bank risk appropriately.There is also a concern that risk indicators are used appropriately by DGS both in consideration of the prevailing economic climate and the stability of individual/multiple credit institutions. There is always the possibility of adverse risk indices resulting in higher contributions that impact competitive advantage with the possibility of compounding impacts on business performance between and within national markets.
Further clarity is required in this area, e.g. NPLs will these be gross" or "net" of provisions as this will be very material in some cases with "net" being more reflective of risk and provision coverage ratio crucial in understanding same - dilutive impacts from adverse weighting arising from high NPLs may be inconsistent with risk where an institution has strong capital and provision coverage ratios (how will this be reflected?).
The risk exposure should generally be reflected through a range of indicators which focus on the quantity and quality of capital, liquidity and risk and which are already available as regulatory ratios. Therefore, due consideration should also be given to total regulatory capital, the capital ratios CET1, RWA/total assets and bail-inable debt. In terms of weighting, it is widely accepted that liquidity was the primary driver and little comfort was taken on the levels of capital by funders. The LCR and NSFR will provide an appropriate view of short and long term funding risks. The interrelationship of these measures needs to also be taken into consideration.
We believe that the proposed indicators for the business model assessment could be complemented by indicators such as funding concentration, the extent of the trading book and the share of RWAs for market risk.
About some of the indicators proposed by EBA, and which are not anchored directly in the banking regulation, some doubts could be expressed. This goes notably for the NPL ratio and the unencumberd assets ratio:
• The disadvantage of the NPL ratio has to do with the fact that this ratio cannot be based on a single uniform definition in accordance with regulation, and consequently it may cause an unlevel playing field between Member States.
• On the other hand, the unencumbered assets ratio may have unintended consequences in that it may be to the detriment of banks which, for their funding, depend to a large extent on retail customer deposits, although this kind of funding certainly is less risky than funding via the (short-term) wholesale market. The use of this ratio may also cause unwanted distortions between Member States.
• Furthermore, the concept of unencumbered assets can include assets that are not encumbered but also assets that not being encumbered will never be potentially encumbered. Hence, it should be more realistic to employ the ratio “available assets for encumbrance/covered deposits.” But again, it is not clear how to measure the encumbrance. The report on asset encumbrance is still in progress and although it tries to achieve comparability among institutions there are some questions that are difficult to assess when defining how encumbered or unencumbered an asset is. Therefore, we suggest that the ratio included in Annex 3 of “own funds and eligible liabilities in excess of MREL” should also be taken into account.
• The retained criteria should not only take into account the fact that an institution may fail, but also the fact that the DGS could be drawn (in resolution or in liquidation). In particular, the RTS 41/2014 on MREL provides that systemic institutions will be required to maintain very significant levels of MREL to make sure that the resolution fund would be drawn (if need be). This implies that the probability of SIBs drawing down the DGS is nil.
• That is why the risk factors should, at the very least, include the amount of own funds and eligible debts owned toward the MREL requirement, and a ratio comparing the MREL and regulatory own funds and consider somehow the shortfall with the 8% threshold provided in Article 44 of BRRD.
Finally, the EBF questions in principle the suitability of the risk category “Potential losses for DGS” as an indicator for the risk of failure for an individual institution. In point 44 EBA mentions explicitly that the risk indicators shall measure the risk (of failure) of the individual member institution, not the fund as a whole. It is clear that the potential loss for the DGS is an important parameter, but this risk is covered by the formula already, i.e. by using covered deposits as basis for the calculation of the yearly contribution. The inclusion of potential losses for the DGS in the Aggregate Risk Weight seems to be methodically questionable."
Question 6: Do you agree with the option to use either capital coverage ratio or Common Equity Tier 1 ratio as a measure of capital? Would you favour one of these indicators rather than the other, and why?
The measure of capital should generally be reflected through a range of indicators which focus on the quantity and quality of capital in relation to risk and total assets. Therefore, due consideration should also be given to total regulatory capital, the capital ratios CET1, RWA/total assets and bail-inable debt. The interrelationship of these measures needs to also be taken into consideration.Question 7: Are there any particular types of institutions for which the core risk indicators specified in these Guidelines are not available due to the legal characteristics or supervisory regime of these institutions? Please describe the reasons why these core indicators are not available.
On the whole the core risk indicators are taken from regulatory reporting and therefore we expect them to be available for all institutions. Some indicators may not be available for particular types of institutions (e.g. financial institutions, investment firms) on the basis of applicable prudential requirements.Question 8: Do you think that more guidance, or specific thresholds, should be provided in these Guidelines with regard to calibration of buckets for risk indicators, or minimum and maximum values for a sliding scale approach?
DGSs with risk based contribution experience do not need further guidance. DGSs that have no or only little experience with risk based contribution models could have demand for further guidance.The EBF advocates an evolutionary approach to harmonising the risk based methodology over time while the EBA and DGS gain experience and empirical evidence to support a common approach that would limit inconsistencies across the European Union.
We think that it must be the competent authority who decides the appropriate method for its national banking system also with regard to calibration of buckets for risk indicators, or minimum and maximum values for a sliding scale approach. More generally, we suggest that risk indicators buckets or sliding scale approach result into linear impact and minimise any cliff effects.