Response to consultation on revised draft Guidelines on common procedures and methodologies for the supervisory review and evaluation process (SREP) and supervisory stress testing

Go back

Question 1: How could the guidelines be further simplified in a way that appropriate focus of assessment is allowed while preserving the comprehensiveness of the assessment and ensuring that all aspects are sufficiently covered?

Initial starting points for a more streamlined SREP approach are mentioned primarily in paragraphs 54, 56 and 58. These are not sufficient, however, particularly as Table 1 requires an “Assessment of all SREP elements (at least)” every three years, even for smaller institutions. It remains unclear which of the numerous individual aspects set out, inter alia, in Titles 5, 6 and 8 (which are expanded even further by references to further EBA guidelines), supervisors are permitted to disregard or consolidate, for example, when assessing category 3 and category 4 institutions. At the very least, clarification should be included that this refers to the main elements of the SREP and not to all of the individual criteria mentioned in the Guidelines.

Paragraph 54 merely refers to Article 97(4a) CRD V. The EBA was, however, mandated in that Article to define criteria on how supervisory authorities could implement adapted (harmonised) procedures for institutions with a similar risk profile. In our view, this is currently not reflected in the Guidelines. The draft version of the Guidelines as a whole does not yet meet the objective of strengthening proportionality set out in the “Risk Reduction Package Roadmaps” of 21 November 2019.

The CRR II not only differentiates between “small and non-complex” institutions, but also makes distinctions based on capital market orientation, which results in additional relief (at least with regard to disclosure) – the distinction could also be made in these Guidelines.

All applicable standards and guidelines that are currently in force should be known to supervisors as well as to institutions. It is therefore appropriate to add cross-references to them, whilst merely repeating contents from other guidelines in the SREP Guidelines (e.g. as in paragraph 170) should be avoided. The SREP GL would be clearer without these repeated references. If the EBA does see a need to provide an overview of the relevant guidelines, these could simply be stated in the form of a list or table at the beginning of the document.

Question 2: Do you think that the proposed overall framework for setting additional own funds requirements appropriately incorporates the ICAAP information and estimates?

No. Based on the current draft, unfortunately only a loose connection between the ICAAP, as the basis for the supervisory analysis, and the final P2R would remain. The idea currently is for the ICAAP to also serve as a quantitative basis for the SREP. From our perspective, this sort of fundamental approach still makes sense, as explained above. This is now to be replaced by wording such as “take into consideration”, “use” or “support” which would create scope for an arbitrary approach for the SREP that goes much too far and leaves the door wide open for supervisors to establish very different approaches to “their” SREP. Consequently, the idea runs contrary to the idea of a consistent European approach to the SREP being adopted by the competent authority and could serve to distort competition. The degrees of freedom created here – presumably, among other things, to open up considerable scope for supervisory benchmark models (paragraph 369(c)) – clearly go too far.

Most banks have been investing in the reliability of their ICAAPs for decades. Strengthening the institutions’ ICAAPs has always been an important goal for the supervisors, which is why the ICAAP should remain the main starting point for the quantification of individual risks and the determination of the P2R. Only overall inadequate and unreliable ICAAPs should be supplemented by supervisory benchmarks. In all other cases, the imprecise nature of benchmarks would lead to overly conservative measures if they were to be introduced. The EBA should therefore describe in detail when ICAAPs are deemed to be generally unreliable and when, as a result, benchmarks are to be introduced, while in all other cases the ICAAP should be the main starting point for the determination of the P2R. As a result, we take the view that the sources of information referred to in paragraph 369 should only be used as supplementary sources; the primary role of the ICAAP should be retained.

Question 3: Do you agree with the proposed clarifications on the assessment of the risk of excessive leverage?

In general, we recommend that the supervisory authorities adopt a pragmatic approach to assessing risks of excessive leverage. In our view, the business models and business activities of the vast majority of institutions do not entail such risks, and imposing a P2R-LR and/or a P2G-LR should ultimately only be necessary in singular cases. Additional information requirements, a separate supervisory leverage ratio stress test or similar measures are only likely to be necessary – if at all – for institutions if the supervisory assessment flags up material indications pointing to risks of excessive leverage. For all other institutions, the requirements relating to the leverage ratio should be implemented as part of a system that is as streamlined as possible so as not to produce any unnecessary additional work for supervisory authorities and institutions alike. The reference in paragraph 397 to the fact that available sources of information should be used should be positioned more prominently within the text and, where appropriate, further information should be added on proportionate implementation options.
In addition, the relevant requirements should make it clearer that a P2R-LR and P2G-LR that is greater than zero does not need to be set by default (e.g. by adding “where necessary / applicable” in paragraphs 394, 398, 404 b, 409 b, 411 and 424). A P2R-LR should only be applied to outlier institutions with a demonstrable risk of excessive leverage, meaning that this can only be the exception and not the rule. In our view, the competent supervisory authorities should also only take action to mitigate the risk of excessive leverage if it is possible to demonstrate that this could have a significant negative impact on the sustainability and viability of the business model and if this risk, which is potentially inherent in the institution’s business model, is not already addressed by other action taken by the institution or under Pillar 1. The competent supervisory authority should adopt a neutral stance vis-à-vis business models as a matter of principle. This should be clarified in paragraphs 392/393.
In order to foster a better understanding of the various capital requirements/ratios (P2R, TSCR, OCR, P2G, P2R-LR, TSLRR, OLRR, P2G-LR), it would be helpful to include a graphic of the two “stacks” based on the previous Figure 6.

#Paragraph 393
We do not agree with the vast majority of the proposed examples. Letters a. to d. of paragraph 393 mention a number of “aspects” that are to be taken into account when assessing the “risk of excessive leverage”, even though these aspects have no legal basis. In the relevant Article 4(1) no. 94 CRR, this risk is defined relatively narrowly as follows:
’risk of excessive leverage’ means the risk resulting from an institution's vulnerability due to leverage or contingent leverage that may require unintended corrective measures to its business plan, including distressed selling of assets which might result in losses or in valuation adjustments to its remaining assets’

This definition is based on the central assumption that the observed institution-specific leverage ratio includes the risk that corrective action will be required due to a potential breach of the minimum leverage ratio requirement of 3%, which could lead to additional (fire sale) losses. Situations in which the institution operates close to the lower limit or has to tolerate a very volatile leverage ratio due to the nature of the business model can therefore be identified as causes of the “risk of excessive leverage”. These situations are covered by c. and d. In our opinion, however, neither the three aspects mentioned under a. nor the aspects in point b. can be summarised under the level 1 definition set out above, meaning that they have to be removed.
The content of a. and b. is also partly unclear/incomprehensible, for example:

• Paragraph 393(a)(i) It remains unclear how supervisors will decide whether the products concerned are optimisations or not. The minimum requirement for the leverage ratio was only introduced with CRR 2 with effect from 28 June 2021. Daily SFT value reporting has also been introduced. Before requesting SREP measures, this future reporting should be monitored and evaluated; if any anomalies emerge, the supervisory authority should ensure transparency which would then serve as a starting point for dialogue between the supervisory authority and the bank.

• Paragraph 393(a)(iii) There is no definition of what “highly exposed” means.

• Paragraph 393(b) Non-compliance with the leverage ratio should not be addressed within the P2R-LR, but needs to be addressed by other regulatory measures. Otherwise, compliant exclusions as defined in the CRR should not be undermined by the SREP GL.

Furthermore, Article 429a CRR explicitly provides for exemptions for authorised exposure. We fear that the exceptions provided for by the legislator will be undermined by these Guidelines. The use of those exceptions cannot be regarded as evidence of a risk of excessive leverage per se. In this context, it must only be verified that the prerequisites for making use of an exception are met. The leverage ratio is a non-risk sensitive measure by nature and the SREP Guidelines should not try to make it risk sensitive by introducing adjustments that were deliberately not taken into account when establishing the leverage ratio.

Competent authorities should perform a test run in the upcoming SREP cycle to double check whether the defined approach can be applied proportionately. Setting binding P2R-LRs should only follow as a second step.

Question 4: Do you think that the assessment of dimensions and indicators described in this explanatory box would also be relevant for the assessment of the risk of excessive leverage? Are there any other elements / indicators that you are using in the assessment of this risk?

No, further indicators do not seem to be appropriate. As a result, we welcome the fact that the examples discussed in the explanatory box are not included in the draft SREP Guidelines, as none of them can be summarised under the authoritative definition of “risk of excessive leverage” provided in Article 4(1) no. 94 CRR. In our opinion, a link to “leverage risk” based on the CRR definition cannot be meaningfully drawn in these examples. What is more, the aspects mentioned are already taken into account elsewhere.

Question 5: Can you provide examples of situations which in your view might require CET1 instead of other capital instruments to cover potential losses in relation to P2R and P2R-LR?

The proportion of CET1 for the P2R specified in Article 104a CRD V and the requirements for the P2R-LR comply with the requirements set out in Article 92 CRR. In our opinion, more stringent requirements (i.e. higher CET1 ratios) are generally not necessary and should be limited to justified individual cases, as provided for in the wording of the Directive. We do not believe that there are any generally applicable examples of situations or that there is any need for more detailed regulations.

Question 6: Would you consider the introduction of a standardised template for the communication to the supervised institution of the outcome of the SREP to be beneficial?

The competent authorities should be encouraged to provide a structured template for the communication of P2R, P2R-LR, P2G and P2G-LR. The main driver of the structure of such templates should be transparency. Institutions need detailed information on the basis/methods for determining the Pillar 2 capital requirements imposed so they can address them appropriately. That is, institutions must be informed which risk types, deficiencies, benchmarks, etc. contribute to the given determination of P2R, etc. and to what extent. In this context, a list of all items/risks examined (including references) would increase comparability and traceability for institutions. Overall, benchmarking across institutions, where performed, and benchmarking results should be made more transparent.
Setting standardisation objectives aside, however, it is crucial to maintain the individual assessment and diversity of the institutions and business models. Since these can differ from country to country, and in order to allow NCAs to adopt a proportionate approach for the smaller institutions they supervise, no standardised EU-wide template should be imposed as a requirement and the transparency requirements should be tightened up instead. Best practice examples could also be provided where appropriate to work towards greater convergence and comparability of the SREP throughout the EU.

Question 7: What are your views on the guidance for setting P2G and P2G-LR? Is it sufficiently clear?

#Title 7.7.1
When it comes to P2G-LR it is rather unclear how the static balance sheet assumption as given in the EU-wide stress test could form a proper basis for the exploration of a leverage ratio under stress, which would already be misleading from a theoretical point of view.

There is no clear guidance for setting a P2G-LR. We suspect that the LR-related risk is overestimated and that, in consequence, the P2G-LR is set too high. Paragraph 423 states that the “level of P2G-LR should protect against the breach of TSLRR in the adverse scenario”. On the other hand, paragraph 429 sets out that the P2G should cover at least the maximum stress impact. Regardless of how far the starting point was above the minimum requirements or how far the minimum requirements were exceeded in the stress scenario, the outcome would be very different.

We ask that it be clarified that there is generally no room for (non-institution specific) minimum (floor) P2G and P2G-LR.

In principle, we welcome the methodological freedom afforded to the competent supervisory authorities in “translating” supervisory stress test results into an own funds recommendation (P2G), as it allows the competent supervisory authority to give adequate consideration to special national features of the banking sector or of certain institutions (e.g. deposit protection schemes). In substantive terms, we believe that the danger is that a risk will be counted several times. We believe that the aspects below should be added to the text in principle:

1) Static balance sheet assumption (paragraph 430): The use of the static balance sheet assumption (SBA) in supervisory stress testing means that risks can sometimes be systematically overestimated for certain institutions and business models. In this respect, we explicitly welcome the option to adjust the own funds recommendation (P2G) in paragraph 430 (2nd sentence) to avoid a misleading semblance of precision, or to avoid the stress test method/results being dependent on mechanistic aspects. In addition, competent authorities should adopt a neutral stance vis-à-vis business models as a matter of principle (see also answer to question 3) and should not put any business model at a systematic disadvantage due to the use of the SBA. Paragraph 430 should be supplemented to this effect.

2) Climate risk (paragraph 433): At present, there is still no established market practice for the systematic identification and assessment of climate risk. In our opinion, merely expanding the list of the possible risks that increase the potential losses associated with an adverse stress scenario does not help to refine risk identification and assessment for climate risk, but rather creates greater uncertainty in the assessment process for supervisors and institutions alike. As a result, we currently take a critical view of the addition of “climate risk” in paragraph 433.

3) Transparency vis-à-vis institutions (paragraph 436): In general, it should be transparent for institutions which of the factors set out in paragraph 432 is decisive for the purposes of determining the own funds recommendation (P2G) for each institution. Paragraph 436 should be supplemented accordingly.

#Paragraph 437
In our view, keeping the provision introduced in 2018, stating that the P2G is to be met using CET1, is not permissible. The wording of Article 104b CRD V adopted in 2019 only focuses on own funds overall. Working papers on the CRD review reveal that the majority of member states had favoured the introduction of a “soft” recommendation with regard to the P2G. As a result, an expectation to meet P2G with CET1 only within the SREP Guidelines is not consistent with the intention of the EU legislator and is lacking any legal basis. This tightened requirement should be deleted and the wording should be based on own funds in line with the CRD V.

Question 8: What are your views on possible disclosures, which may be attached to P2G and/or ranges of buckets in case they are identified?

It is not clear from the question which type of disclosure is meant here. As a result, we will address several potential aspects:

The disclosure of buckets or “ranges of P2G add-ons” by the supervisory authority could make sense in principle as a way of communicating to the markets. Such publications by the supervisory authorities should, however, be limited to large, capital market-oriented institutions. In any case, we do not encourage that individual results of single institutions be published.

Our view is that extending the disclosure requirements for institutions with regard to individual P2G would clearly not make sense, which is why we reject such a proposal. The disclosure requirements have only just been revised extensively and in full as part of the CRR II in conjunction with Implementing Regulation 2021/637, meaning that they reflect the current needs from a Pillar 3 perspective. Small and non-complex, as well as non-capital market-oriented other institutions were granted considerable relief by law as part of this process, as the disclosed data is generally not accessed (cf. EP, 2016/0360A(COD) of 11 December 2017). With this in mind, we also consider the disclosure regulations in the CRR II to be entirely sufficient.

Question 9: What are your views on the capital instruments potentially used to cover losses in relation to P2G-LR? Please provide the rationale or specific examples for your views.

The structure of the leverage capital requirement as a Tier 1 requirement should also be applied for the purposes of the P2G-LR.

From a conceptual point of view, we fully agree that using Tier 1 to meet the P2G-LR upholds consistency within the leverage ratio stack based on Tier 1, and would leave the calculation coherent and straightforward. As a result, we ask for clarification in paragraph 437 that the P2G-LR is expected to be met using Tier 1 capital.

Upload files

Name of the organization

Die Deutsche Kreditwirtschaft - German Banking Industry Committee