Response to consultation on revised Guidelines for common procedures and methodologies for the supervisory review and evaluation process (SREP) and supervisory stress testing
Q1. What are the respondents’ views on the overall amendments and clarifications made to the revised guidelines (across Titles 2 – 12)?
We welcome the review and appreciate the significant workload that the EBA faced in improving these Guidelines. However, we would like to clarify three key elements. We would like to reiterate that the EBA should not continue requiring that P2G be held only in CET1. Under the CRD, P2G refers broadly to own funds, which include CET1, AT1 and T2. The legislation contains no specifications regarding the quality of own funds required to meet P2G. The only requirement is that the total amount of own funds be met.
Indeed, in contrast to Article 104a (P2R), Article 104b does not stipulate which type of own funds (CET1, AT1 or T2) must be used to satisfy P2G. The current approach therefore introduces unnecessary inconsistency into the capital stacking order. In our view, coherence would be improved by applying to P2G the same default composition used for P2R, which mirrors the P1 requirements (i.e. T1: 75%, of which 75% CET1). Where necessary, competent authorities may prescribe a different composition for both P2R and P2G under the CRD (e.g. requiring CET1 only). However, such deviations should not be the default, given that Pillar 2 measures are intended to be institution-specific. Regarding the entry into force of the Guidelines, we welcome the clear application date of January 2027. However, the final sentence creates legal uncertainty. If competent authorities may introduce selected elements of the revised framework earlier, institutions cannot know in advance which specific aspects might be applied already in the 2026 SREP cycle. Clarification on the scope and limits of such early implementation would therefore be highly appreciated to avoid legal uncertainty and heterogeneity across jurisdictions. We suggest adjusting the text of para. 7 in the following way: “Since the SREP cycle often corresponds to a calendar year, these guidelines are expected to apply from 1 January 2027, following the consultation period and the publication of the final version. However, competent authorities are encouraged to start the reassessment for the possible redetermination of P2R at an earlier stage, where possible.” At the same time, we reiterate the urgency of careful assessment of the interactions between P1R and P2R by the competent authorities. Finally, the effectiveness of the SREP exercise would be significantly enhanced by greater transparency in the communication of results to banks. Clearer explanations of methodologies, assumptions, and drivers of outcomes would allow banks to better understand their individual results, identify concrete areas for improvement, and ensure consistent interpretation across institutions. It is also important that banks understand how supervisors evaluate the remediation actions implemented and how they are incorporated into the supervisory judgment. Improved transparency would also strengthen trust in the process and increase the practical usability of the results for risk management and strategic decision-making. For these reasons, we exhort the EBA to clarify in the text the requirements that supervisors address (e.g., benchmark scope, methodology, data inputs, calibration logic) when communicating SREP results, also considering materiality and proportionality.
Q2. What are the respondents’ views on the integration of ESG risks and factors across the existing SREP elements in the revised guidelines?
NA
Q3. What are the respondents’ views on the enhanced simplification and proportionality aspects?
The new Article 97(4) of CRD VI appears to be reflected in para. 11, para. 14 and para. 15 of the draft Guidelines. However, it remains unclear how the provision laid down in Article 97(4) CRD VI would be implemented in practice throughout the EBA Guidelines, if at all. More specifically, it is suggested that the EBA explicitly address the assessment elements entrusted to it by the legislator. This is particularly the case for the threshold relating to compliance with the requirement that “the vast majority of the subsidiary institutions”, as referred to in point (c) of Article 97(4) CRD VI, meet the relevant conditions.
Accordingly, the EBA should take into account the objective of a correct and effective implementation of the new Article 97(4) in terms of simplification and reduction of supervisory burdens, in particular from the perspective of the “peripheral entities” (i.e., individual banks). The SREP methodology, the scope and depth of supervisory analyses, as well as the intensity and frequency of supervisory engagement, should carefully consider the impact on the group’s peripheral entities. Moreover, all supervisory requests addressed to the “centre” (the “parent” institution) are inevitably subsequently transmitted to the “peripheral entities”.
In order for Article 97(4) CRD VI to be effectively applied, it is necessary for the SREP Guidelines to clearly and verifiably identify the concrete prudential effects deriving from the provision for the institutions concerned. A proportionality rule is not self-executing; it requires regulatory specification enabling the ex-ante identification of access criteria to the specific treatment, its scope of application, and the modalities for reducing supervisory burdens. In the absence of such elements, institutions would not be able to assess whether and to what extent they may benefit from the provision, with the result that the rule remains, in practice, devoid of real effects. In this regard, the EBA should therefore clearly indicate the stages of the SREP process in which Article 97(4) produces effects, for example with respect to:
- initial categorisation;
- scope of required analyses;
- depth of supervisory assessments;
- frequency of the SREP cycle;
- modalities of interaction with affiliated institutions.
To fully reflect the new Article 97(4) CRD, the EBA SREP Guidelines could be amended as follows:
- a) the introduction of an ad hoc category for institutions referred to in the new paragraph 4; or
- b) the introduction of a derogation clause within the criteria governing the allocation of institutions to the intermediate categories 2 and 3.
A joint reading of the wording of Article 97(4) and of Article 143(1), point (c), CRD VI does not exclude the possibility of introducing an ad hoc category in addition to the four categories currently set out in the EBA Guidelines. Reference may be made here to the atypical nature of the entities concerned. Achieving differentiated treatment between the central institution and the individual affiliated banks would be fully in line with the above-mentioned European drivers for simplification, most recently also promoted by the ECB. Under this approach, the “parent” institution would remain in its current SREP category, while the affiliated banks would be subject to a less intrusive supervisory treatment, not dissimilar to that applicable to category 3 or 4 institutions.
Alternatively, in light of the indications provided by the new Article 97(4) and of what appears to be the intent of the Level 1 legislator, a derogation clause could be introduced for categories 2 and 3 in order to give concrete effect to the new provision. In particular, an institution otherwise allocable to category 2 could be placed in category 3 if it meets the conditions set out in Article 97(4), unless (of course) it has exhibited significant signs of deterioration in the previous two SREP cycles. Similarly, an institution otherwise allocable to category 3 could be placed in category 4 if it meets the conditions of Article 97(4), subject to the same safeguards applicable to the transition from category 2 to category 3.
Q4. What are the respondents’ views on the introduction of a high-level escalation framework?
The revised SREP Guidelines introduce a high-level escalation framework describing various supervisory actions (dialogue, recommended measures, binding measures, enforcement). However, it remains unclear how competent authorities will determine which escalation step is appropriate in practice, how proportionality influences these decisions, and how consistency across supervisors (and JSTs) and across institution categories will be ensured. More clarity would strengthen supervisory predictability, enhance dialogue quality, and support the risk-based and proportionate application of measures as set out in Title 2.4.
In particular, the framework is presented as flexible and non-sequential, but the criteria for choosing between escalation or de-escalation measures are only broadly defined, which may lead to divergent supervisory practices. We suggest providing further clarification on:
- the decision-making criteria for selecting a specific escalation measure (or the sequence of measures),
- how proportionality should guide escalation choices for institutions with different risk profiles or SREP categories, and
- expectations for consistency and transparency across competent authorities and supervisory teams.
Some illustrative examples could further improve the understanding of such mechanisms.
Q5. Do you consider the coverage and level of detail of this Title appropriate for its intended purpose?
NA
Q6. Do you consider the coverage and level of detail of this Title appropriate for its intended purpose?
NA
Q7. What are the respondents’ views on the updated section 5.7 “ICT systems, risk data aggregation and risk reporting”?
More generally on Title 5, the Guidelines require the competent authorities to assess “appropriate understanding” and “sufficient knowledge and skills” “through regular training” (para. 83(c)) of the institution’s management body. Such expectation appears vague and leaves too much room for subjectivity, and the assessment should be more specifically defined. Expectations are that the assessment will focus on the adequacy of the training. We exhort indeed the EBA to clarify these elements, and fully align the provisions of this Title with the substantial regulation on governance implemented via the already existing Guidelines, currently under review.
Q8. Do you consider the coverage and level of detail of this Title appropriate for its intended purpose?
In table 10 (Supervisory considerations for assigning an operational risk score), we suggest better defining the consideration in relation to inherent risk related to concentration risk towards third-party service. Indeed, the terms "simple," "moderate," and "long/complex" subcontracting chains are vague. We recommend providing clear, quantifiable definitions for these categories.
Q9. Do you agree with the treatment proposed to account for transfer pricing risk in the context of trading book activities? Please elaborate.
NA
Q10. What are the respondents’ views on the integration of the EBA GL on ICT risk assessment under the SREP (EBA/GL/2017/05) and DORA aspects?
NA
Q11. What are the respondents’ views on the introduction of operational resilience (section 6.4.5)?
Members have raised several concerns about the practical implementation of the provision related to operational resilience. Indeed, the draft Guidelines assume the existence of a fully functional and harmonised framework on operational risk management. However, practical experience points to a lack of coherence across Member States in digital and ICT matters, as well as to the absence of a fully harmonised supervisory field. This creates significant challenges for institutions and significant difficulties in translating the EBA Guidelines into practice, notably due to overlapping supervisory requests and an unclear allocation of responsibilities between banks, other regulated entities, and third-party service providers. In addition, national initiatives on defence and ICT resilience further increase divergence among banks across jurisdictions.
These considerations should be carefully incorporated into the Guidelines and, in particular, para. 244 should be amended as follows:
“....Moreover, competent authorities should consider whether the institution has sufficiently mapped the internal and external interconnections and interdependencies that are needed to deliver its critical or important functions. However, the supervisory assessment should take into consideration the existing variety across Member States in addressing operational resilience and the relevance of practical obstacles for banks in addressing certain deficiencies related to third-party providers that may not be in the purview of the institutions”.
Finally, the final Guidelines should clarify that contractual matters will be addressed separately under the Digital Operational Resilience Act (DORA).
Q12. What are respondents’ views on the additional section on CSRBB and the combined score for IRRBB and CSRBB?
As prescribed in the draft Guidelines, banks often face JST comments based on “benchmarking with peers”. We exhort the EBA to promote a homogeneous approach and more transparency about the criteria and the results used by the ECB to identify peers. In fact, JST communications lack clarity in defining the buckets and banks often do not understand the basis for the identification of the peers. It could be illustrated, for instance, whether the peer group is defined on the basis of size, geographic spread, business model (and type of activities), or even segment of market activity – or a combination of these. This is particularly important for Cooperative banks, given the specifications of their business model, legal structure, and organisational features. These clarifications and additional specifications should apply to each benchmarking practices: from IRRBB/CSRBB to credit risk.
Q13. What are the respondents’ views on the proposed assessment of the interaction between Pillar 1 and Pillar 2 requirements and on the proposed approach for operationalizing concerning cases where an institution becomes bound by the output floor?
Members express strong reservations regarding the inclusion of Overall Recovery Capacity (ORC) in the SREP score. These concerns stem from several substantive and practical considerations. First, incorporating ORC into the SREP assessment risks creating redundancies or double counting, particularly in relation to Pillar 3 (Capital) and Pillar 4 (Liquidity), which already capture key dimensions of loss-absorbing and funding capacity. Second, ORC is inherently highly context-specific, as it depends on whether the scenario considered is idiosyncratic, market-wide, or a combination of both. As a result, its inclusion would introduce a significant degree of subjectivity into the overall SREP score, potentially undermining its consistency and robustness. While para. 50 (page 18) states that the inclusion of ORC does not “imply mechanistic or automatic correlation between the ORC scores and SREP scores for capital or liquidity adequacy”, successive references seem to point to a rather substantial link (see also reference to “weak” liquidity ORC score in para. 426).
Moreover, integrating ORC into the SREP framework would require explicit consideration of different time horizons (typically six months for liquidity-related ORC and up to eighteen months for capital-related ORC). This would add further layers of complexity to the SREP process at a time when supervisors and institutions alike are engaged in efforts to simplify and streamline supervisory assessments. Finally, given its strong dependence on institution-specific circumstances, a direct reference to the ORC risks blurring the methodological comparability across institutions, which is a core objective of the SREP framework.
In addition to these concerns, member banks request further clarification on several key aspects of the proposed approach. In particular, greater transparency is needed regarding the criteria that would be applied in the assessment of ORC, including the respective roles of qualitative judgment and quantitative metrics. Clarification is also sought on whether certain recovery options could be excluded from the assessment due to implementation constraints, such as feasibility, timelines, or costs, and whether supervisors envisage a form of preference hierarchy among recovery options or the application of additional haircuts. Furthermore, banks would welcome guidance on the potential implications of ORC integration for valuation models.
Finally, member banks request access to the ORC calculation methodology embedded in the Standardised Reporting Template. Transparency in this regard is considered essential to ensure that ORC assessments are based on objective and verifiable criteria and to enable meaningful, informed discussions between banks and their supervisors. Regarding Title 7.1, para. 297, we stress the importance of competent authorities assessing the interactions between P1R and P2R, where material impacts on an institution’s capital profile are expected due to relevant changes to the regulatory framework for determining P1R. In particular, it is not obvious how supervisors will balance:
- model deficiencies under internal models,
- the Output Floor (OF) constraining the overall capital requirement, and
- the proportionality principle when an institution becomes bound by the OF.
The Guidelines might provide more examples illustrating when a model deficiency leads to additional P2R despite the OF, and when P2R should remain unchanged due to the OF acting as the binding constraint. Clearer articulation is needed to avoid double-counting risks, ensure consistency across JSTs, and allow institutions to anticipate supervisory responses when transitioning into OF-bound status.
Therefore, we reiterate the feedback of the industry already shared with EBA on para. 297, which must be adjusted as following:
“When a material impact on institution’s capital profile is or may be expected due to relevant changes to the regulatory framework for determining P1R or to its implementation on the specific institution, competent authorities must assess such impact in terms of interaction with the P2R. Such assessment may result in redetermining the quantity or composition of the P2R to make sure the institution’s overall own funds requirements are in line with Article 104a(1) of CRD, in particular that P2R cover risks or elements of risks that are not covered or not sufficiently covered by the P1R. To perform such an assessment, competent authorities may increase the frequency of the SREP assessment as set out in the SREP engagement model".
Q14. What are the respondents’ views on the merger with the ‘SREP liquidity assessment’ and the merger of the scores into a combined liquidity and funding adequacy score?
Para. 372 provides for coverage of all material legal entities: we subscribe to the expectation as well as to entities’ responsibility for the implementations of risk management, but the EBA should also clarify that for consolidated groups, the findings as well as the remediation should be managed by the consolidating entity.
Q15. What are the respondents’ views in relation to enhanced communication aspects?
NA
Q16. Do you consider the coverage and level of detail of this Title appropriate for its intended purpose?
NA
Q17. Do you consider the coverage and level of detail of this Title appropriate for its intended purpose?
NA
Q18. Do respondents consider the guidance for the assessment of third-country branches appropriate and sufficiently clear?
NA