The process is adequately covered and allows for sufficient flexibility. There are however two elements on which further specifications would be useful:
- Paragraph 444 & table 13: there is no indication of the relative weight of the different risks which are parts of the overall score, nor how competent authorities should assess their relative weight. For example: which overall score should a bank get with high risk business model and governance, and no risk on capital and liquidity. It would be useful to provide additional details, such as:
o Risks which, if individually evaluated as high risk, would set the overall score as high risk
o Risks which, combined as high and/or medium risk, would set the overall score as high risk
o Risks with potential offsetting effects (e.g. high risk business model with no risk governance)
o An indicative relative weight or a ranking of each individual risk in the overall score (this should remain flexible as it may vary between banks, countries, internal or external factors, supervisory judgment. For example it can be provided by weight range to be more flexible)
- The evaluation process of ICAAP is shortly described in paragraph 5.6.2 (together with ILAAP, description therefore limited to common evaluation areas) and paragraph 321. If we compare this to other parts of the guidelines, where better known processes have been described with many more details (e.g. credit risk is detailed through 16 pages, going into details such as credit limits and non-performing exposures), we may consider unlikely that these short paragraphs will be sufficient to guarantee that different authorities would provide a similarly adequate evaluation of the ICAAP process. It would be useful for the guidelines to provide additional details, such as:
o The presence of ICAAP evaluation under normal vs. stressed conditions, and the respective importance given to these two separate exercises
o How the equal treatment of similar risks is ensured, despite potential dissimilar treatment in pillar I (e.g. credit risk for IRB vs standard counterparties)
o How to deal with the continuity vs change of methodologies over time (e.g. ensuring that a change of ICAAP methodology is due to changed economic conditions or methodological improvement, and not set to avoid reaching a risk appetite trigger)
o How correlated risks should be dealt with (e.g. interest rate risk influencing credit risk which in turn influences interest rate gap), as well as correlated stress tests (e.g. market stress test impact on fair value for credit risk)
Yes, the use of a proportionate approach is appropriate. However, it would be useful for EBA to also ensure with competent authorities that the required frequencies can be met for each category of institutions, without impact on the quality of the review.
- For example, a full annual review of all SREP elements for category 1 institutions may lead to a lighter review, focused only on higher risks, because of lack of resources to perform a thorough review.
o In paragraph 35, EBA demands that “competent authorities update at least annually the assessments of all individual SREP elements, and assess at least the most material individual risks to capital and risks to liquidity and funding”. If a previous review showed that a material risk is fully under control and reliably evaluated by the institution, it could be better for the competent authority to be granted the ability to perform a lighter review of the institution’s results for this risk the next year, in order to free resources for a deeper review of lesser risks which may become high if left unattended.
- Since there is a difference in both size and risk exposure between categories 3 and 4, the SREP monitoring requirements could better reflect this difference. For example, the monitoring of indicators may be reviewed by the competent authority on a less than quarterly basis for level 4 institutions (“limited scope of activities with non-significant market shares”), as some indicators may not be available quarterly for these and competent authorities resources may be more usefully focused elsewhere.
We believe that there are a few elements on which additional details could bring improvement to title 4:
- In the preliminary assessment and identification of areas of focus (paragraphs 4.2 and 4.3), it is not offered to competent authorities to take into account the evolution and trends of activities and geographies, in addition to their point in time value, for the selection of the areas of focus. As currently written, the guidelines consider the analysis of trends only after selection of the areas of focus. In some circumstances, this could lead to leaving unidentified risks behind.
o For example, a minor activity with a growth rate 10 times higher than average may be worth looking at, although it may not be a key activity at the time of the SREP, nor considered a change in strategic plans or business model due for the same reason
As the SREP guidelines aim at promoting an early detection of raising risks, it could be useful to leave more flexibility in the selection process, in order to give to competent authorities the ability to select less visible areas of focus.
- In the areas for strategy analysis, paragraph 68, it could be useful to isolate point e. (deltas between past plans and achievements) from the other elements. Since this control evaluates the ability of an institution to provide realistic forecasts, it would be better if it was considered separately, as an adjustment factor for the evaluation of the other elements (projected financial performance, economic assumptions, etc.).
o For example, if a growth rate has been in the past forecasts systematically overestimated by an average of x%, the competent authority could then evaluate the institutions present projections after an x% adjustment.
This adjustment factor could then be used for other areas of evaluation, such as the evaluation of the institutions projected financial performance described in the assessment of sustainability in paragraph 73.
- In the assessment of business model viability, paragraph 71, the guidelines currently don’t indicate whether the underlying elements of the overall ROE should be considered. More generally, precisions could be provided on whether the strategy and business model analysis are conducted solely on an overall basis, or considered for each area of focus. An analysis of the underlying elements of ROE would for example consider:
o The weight of the different components (activities, geographies). In particular, it could assess the ROE of growing vs. decreasing activities / geographies, in the recent past and according to the strategic plan, in order to get a better view on future ROE evolution.
o The interactions between activities. For example, an institution may need to keep low ROE activities as a necessary step towards higher ROE activities.
o The influence of a change in strategic plan. For instance, it may be acceptable to have a temporary low ROE when targeting a new market / activity, while this may not be acceptable on the medium term.
The breakdown of risk categories and sub-categories provides an appropriate coverage of the risk framework. However, the proportion in which these risks are covered may be unbalanced. There is indeed a risk that the guidelines would provide too many details on risks which are already well known and controlled by most banks (credit risk, market risk), while other risks which may have a higher influence on the institutions sustainability (e.g. legal risk, reputational risk) are only briefly described.
o For example, the credit risk part of the guidelines mentions assessing counterparty credit limits, while the operational risk part does not mention assessing the early warning system (compliance and legal risk control) or the whistleblowing process.
While it may be difficult to give very detailed instructions on how to evaluate low frequency/high impact risks, giving too many details on well-known risks will have the drawback to focus proportionally more attention and resources on these details, and leave less for other as important but less detailed risks.
While it is useful to indicate an own fund requirement under the form of a standard ratio, it is not very relevant, when requiring own funds for a specific risk, to express them through a ratio not linked to this risk (i.e. on TREA basis). On the other hand, since the guidelines aim at ensuring a comparable approach between institutions, it may be efficient to use a standard ratio of additional own funds as a % of TREA, as an indicative ratio at the date the requirement is set.
- For example, if additional own funds are required solely to cover the outsourcing risk on a portfolio in run-off in a given country, all the rest being adequately covered by article 92 of Regulation 575/2013, it would be appropriate to express the regulatory requirement for additional own funds under a format like:
o “As of date and until otherwise directed institution is required to hold additional own funds in excess of the requirement set out in article 92 of Regulation 575/2013. These own funds are set to x% of risk exposure in portfolio xxx, of which x% are to be covered by CET1 and x% by T1. As of date, these additional own funds represent x% of TREA, leading to a TSCR of x% of TREA”.
- Example in case of own funds for multiple risks:
o “As of date and until otherwise directed institution is required to hold additional own funds in excess of the requirement set out in article 92 of Regulation 575/2013. These own funds are set as follows :
x% of risk exposure in portfolio xxx to cover unexpected losses due to operational risk, of which x% are to be covered by CET1 and x% by T1
x% of cumulated interest rate gap to cover concerns over ALM governance, of which x% are to be covered by CET1 and x% by T1
As of date, these additional own funds represent x% of TREA, leading to a TSCR of x% of TREA”.
It may indeed be appropriate to grant additional transition periods to competent authorities, as having an experience of SREP processes before the implementation of the final guidelines will help them to get a better view on what is feasible and efficient, before having to actually make demands on banks own funds. This will allow regulatory requirements to be better calibrated.
Besides, transitional arrangements do not prevent competent authorities from anticipating these requirements when needed. There is therefore no negative side to the transition.
