Response to consultation on revised Guidelines for common procedures and methodologies for the supervisory review and evaluation process (SREP) and supervisory stress testing
Q12. What are respondents’ views on the additional section on CSRBB and the combined score for IRRBB and CSRBB?
We welcome the European Banking Authority’s initiative to further integrate IRRBB and CSRBB into the Supervisory Review and Evaluation Process (SREP) and would like to share the industry’s feedback structured around three key points.
First, the structural interaction between EVE and NII sensitivities means that some configurations in Table 12 are unlikely to arise under standard balance‑sheet positioning across the rate cycle. Clarifying how the criteria should be applied would enhance consistency and prevent unintended penalisation of prudent ALM practices.
Second, we strongly support the fact that the SOT NII is not used directly in the final SREP assessment, which is appropriate given its methodological shortcomings and sensitivity to assumptions. We share EBA’s view that the SOT NII gives incomplete view and hence we strongly support that it’s used only for the preliminary assessment and supported by complementary measures. However, although the SOT NII is presented as an initial benchmark, complemented by the additional economic dimensions introduced in IRRBB Heatmap (EBA/REP/2025/04) and incorporated into the SREP proposal, its binding legal status elevates it far beyond the complementary measures. This raises a risk that in practice the legally binding threshold will play a dominant role, while the non‑binding dimensions designed to provide a broader economic perspective may play a secondary role. Aligning the legal standing of these components would give better regulatory certainty, ensure proportionality, harmonization and level playing field across the EU.
Finally, the EBA IRRBB & CSRBB Heatmaps shows that CSRBB practices remain highly heterogeneous across institutions. Maintaining the current Pillar 2 framework-built on institution‑specific methodologies supported by supervisory review offers a proportionate and economically coherent basis for assessing CSRBB.
1. Structural challenges in the IRRBB scoring framework
We appreciate the comprehensiveness of the SREP framework for IRRBB and, in particular, the clarified considerations for assigning scores in Table 12 of the Consultation Paper. However, we would like to draw attention to a structural aspect of the inherent risk assessment that in our view warrants further reflection. The table doesn’t state whether any or all of the criteria must be met for a given score. Assuming that all criteria must be met and taking “very low” (Score 1) as an example, it requires an institution simultaneously exhibits very low sensitivity to interest rate changes in both economic value (EVE) and net interest income (NII). While this objective is understandable in isolation, it is generally not aligned with the manner in which well managed institutions position their balance sheets over the interest rate cycle. At different points in the interest rate cycle the institutions might decide to hold higher or lower EVE sensitivity, depending on their business strategy, risk appetite and outlook for the future. NII sensitivities behave inversely due to the asymmetric effect of deposit floors and the time structure of repricing. As a consequence, the configuration in which both EVE and NII sensitivities are simultaneously very low is rare and typically occurs only in narrowly defined interest rate environments, rather than as a sign of generally superior IRRBB management. On the other hand, considering “high” (score 4) as an example, for the aforementioned reasons, taking excessive risk in one of the IRRBB measures usually results in low risk in the other one, therefore no institution would fall under score 4.
Further clarification on how to interpret inherent risk considerations outlined in Table 12, given this structural trade‑off, would help ensure that institutions and supervisors apply the scoring framework in a manner that reflects the economic realities of IRRBB, avoids inadvertent penalisation of sound balance‑sheet management, and preserves the intended comparability and proportionality of the SREP assessment.
Moreover, Table 12 does not reflect the broader set of earnings‑related aspects referenced in paragraph 261. To ensure that the inherent‑risk assessment captures the full spectrum of factors relevant to earnings sensitivity, we propose refining the criterion to read: “The sensitivity of earnings to changes in interest rates is not material/very low, or the analysis of the additional dimensions mentioned in paragraph 261 indicates that the level of risk is not material/very low.” Analogous refinements could also be considered for other risk scores
This adjustment would allow supervisors to form an assessment that is consistent with the multidimensional nature of earnings risk described in the Guidelines, while preserving the proportionality and intent of the SREP framework.
2. Alignment with the broader SREP methodology
We welcome the fact that Table 12 does not incorporate the SOT NII threshold directly into the inherent risk assessment. This design choice reflects the Consultation Paper’s broader objective to ensure that IRRBB assessments are grounded in supervisory judgment across multiple complementary dimensions rather than anchored mechanically to a single quantitative indicator. We also appreciate that the SOT NII is described as an initial benchmark and not a definitive or binding determinant within the SREP process.This structure is consistent with the revised SREP principles emphasised throughout Titles 2 and 6, including proportionality, avoidance of mechanistic escalation, and the expectation that supervisors consider an institution’s strategy, business model, and risk‑appetite framework in their assessment of IRRBB.
However, despite this intended proportionality, a significant asymmetry remains between the binding legal nature of the SOT NII, established in CRD and the associated RTS, and the non‑binding supervisory dimensions proposed by EBA in the IRRBB Heatmap (EBA/REP/2025/04). Institutions publicly disclose SOT results under hard law, whereas the additional assessment dimensions (market‑value effects, administrative and operating expenses, net fees and commissions, embedded gains/losses) do not carry equivalent normative force. As a result, there is a risk that SOT NII will have a de facto leading role, notwithstanding the EBA’s intention to avoid automaticity. This is particularly concerning in the context of SOT NII limitations outlined below. SOT NII binding legal status amplifies their impact.
2.1. Limited suitability of SOT NII as a peer benchmark
Firstly, the idea of IRRBB SOT was to develop a single regulatory benchmark across EU. As recognized by EBA itself in the IRRBB Heatmap (EBA/REP/2025/04), institutions apply materially different modelling approaches to non maturing deposit (NMD) margins. Only 38% use constant margins; the remaining 62% employ internal behavioural modelling. This divergence is explicitly accepted in the current framework and inevitably leads to non comparable SOT NII outputs, given the significant sensitivity of NII projections to NMD assumptions. Because SOT NII outcomes depend heavily on these modelling choices, using a single quantitative threshold as a cross institution benchmark risks overstating or understating the underlying risk across jurisdictions.
Furthermore, we would like to note that the Basel Committee’s 2015 review demonstrated the inherent difficulty of developing a single, objective and standardised measure for IRRBB. Although the Committee explored the possibility of a Pillar 1 treatment, it ultimately recognised that the heterogeneous nature of IRRBB-arising from differences in balance‑sheet structures, behavioural modelling, product characteristics and interest‑rate environments-makes a uniform benchmark inherently inadequate. The decision to retain IRRBB within Pillar 2 therefore reflects the structural complexity of the risk and recognises that the heterogeneity of IRRBB precludes the use of any uniform benchmark as a meaningful assessment tool.
2.2. SOT NII does not function as a meaningful test for sufficient earnings
The Consultation Paper emphasises that supervisory assessments should reflect institutions’ business model viability and earnings sustainability (Title 4), as well as their capacity to withstand adverse developments. In this context, SOT NII-defined as a decline in NII relative to Tier 1-does not measure whether institution’s post shock earnings are sufficient to sustain normal business operations. Empirical evidence from across EU markets demonstrates that the degree of NII volatility that banks can absorb varies widely. For example, based on sector level data: • Some jurisdictions (e.g., Poland, Lithuania, Spain, Czechia) can withstand NII declines far exceeding 5% of Tier 1 before reaching break-even profitability. • Others (e.g., France, Germany) reach the same point at significantly lower levels. Moreover, even after a full 5% Tier 1 NII shock, institutions in multiple jurisdictions maintain post shock ROE above the EU average, demonstrating the lack of conceptual linkage between a SOT breach and earnings insufficiency. The metric therefore does not satisfy the stated regulatory purpose of signalling under performance relative to a bank’s operating cost base. Given that SREP explicitly requires a holistic assessment of viability-including profitability, sustainability, and the ability to continue operations-the mechanistic SOT NII threshold provides limited supervisory insight for those purposes.
2.3. Consequences of likely automaticity in SOT NII application and risk of distorted IRRBB management
The revised SREP Guidelines emphasise (Title 2 and Title 6) that supervisory tools should promote effective risk management, avoid mechanistic escalation, and prevent undue cliff effects. Aforementioned legal asymmetry poses a risk that SOT NII will work as a de facto binding limit. Based on industry feedback (published e.g. by Risk.net) this is already the case in some jurisdictions. Market participants further reinforce this dynamic by monitoring SOT disclosures under Pillar 3.
This dynamic may create several unintended consequences:
- Incentives to over hedge in low rate environments. Due to the deposit floor effect, the magnitude of negative NII shocks increases when market rates approach zero, even if interest rate risk positions remain unchanged. To avoid breaching the SOT threshold, banks are incentivised to extend the duration of assets beyond economically justified levels, thereby increasing exposure to long term interest rate shocks risk, contrary to prudent IRRBB practices.
- Stabilisation on unstable funding The SOT NII framework also risks generating prudentially undesirable incentives with respect to behaviourally unstable funding. Because margin compression arises on both stable and unstable deposits in low rate environments, institutions may be implicitly encouraged to hedge or otherwise “stabilise” balances that are inherently volatile, solely to avoid breaching the supervisory threshold. Such an outcome would conflict with fundamental IRRBB and liquidity risk principles, which require that short term, non stable balances remain treated as transient and not form the basis for long term structural positioning. Incentivising institutions to lock in exposures on unstable funds increases vulnerability to abrupt deposit migrations, amplifies procyclical losses when rates normalise, and may replicate risk patterns observed in past banking failures driven by excessive reliance on non-core funding. In this regard, a supervisory metric that incentivises stabilisation of behaviourally unstable balances introduces systemic risk externalities and stands at odds with the intended objectives of the IRRBB framework, which seeks to avoid precisely such forms of misaligned risk taking.
2.4. Recalibrated BCBS shocks may disproportionally impact IRRBB SOT results
Basel Committee’s recalibrated IRRBB shocks will substantially impact SOT results of some institutions, particularly for non‑EUR currencies. For example:
- Shocks for PLN and HUF increase by 60%, from 250 bps to 400 bps.
- In jurisdictions with historically higher interest rate volatility, the recalibration dramatically raises NII sensitivity, pushing entire sectors into SOT breach territory.
We acknowledge that EBA stated in the IRRBB Heatmap that the impact of recalibrated shocks scenarios will be further considered, drawing on QIS and dialogue with competent authorities. In our view, shock recalibration in accordance with the revised BCBS methodology will further deteriorate the goal of a single benchmark as shock levels will materially differ across jurisdictions whereas other relevant economic dimensions are not accounted for in SOT NII calculations.
2.5. Conclusions and recommendations
In light of the considerations outlined above, we agree that including the SOT NII as an initial benchmark within the IRRBB SREP assessment-supplemented by broader supervisory dimensions-can provide a more informative and balanced view of institutions’ interest rate risk profiles. However, the current framework assigns materially different legal weight to these components: the SOT NII threshold is embedded in binding Level 1 and Level 2 legislation (CRD and RTS), whereas the complementary assessment dimensions stem from non‑binding guidelines. This asymmetry may result in the SOT NII operating as a de facto binding constraint, irrespective of the EBA’s stated intention to avoid mechanistic outcomes and to promote a holistic supervisory judgment.
To enable the framework to function as intended, the legal hierarchy between the SOT NII and the complementary dimensions would need to be re‑balanced. This could be achieved either by elevating the additional dimensions to an equivalent legal status-ensuring that institutions and supervisors treat them with comparable authority-or by demoting the SOT NII threshold to a guideline‑level tool, at par with other dimensions. Either approach would reduce the current dominance of the SOT NII, restoring alignment with the SREP principles of proportionality, risk sensitivity, and supervisory judgment.
If disproportionate legal hierarchies persist, then it’s of a paramount importance to make it explicitly clear in the SREP guidelines that an institution can still have Risk score of 1, even if it’s a SOT NII outlier, but complementary measures indicate that the level of interest income sensitivity risk is not excessive, i.e. SOT NII breach is a ‘false positive’ indication. In this context, we propose refining paragraph 261 as follows: “Where the institution is identified as an outlier by the supervisory outlier test on net interest income, competent authorities shouldn’t conclude risk score solely on the result of the test; instead, they should analyse additional dimensions to complement the assessment, when deemed appropriate (…) Institutions can still achieve inherent risk score of 1, even if they are SOT NII outliers, if complementary IRRBB measures indicates that their risks to NII are not excessive and do not endanger continuation of normal business operations"
3. CSRBB: Maintaining a Proportionate and Institution‑Specific Pillar 2 Framework
The Heatmap confirms substantial divergence in the way banks define the CSRBB perimeter, construct spread scenarios and assess the impact of spread movements, with heterogeneity described as significant across the sector. This variation reflects structural differences that cannot be reconciled through a uniform measurement standard. Against this background, the SREP framework rightly avoids prescribing detailed modelling rules and instead focuses on ensuring that institutions embed CSRBB into their governance and risk‑management practices in a manner consistent with their balance‑sheet profiles and market environments.
Given this diversity, we support the decision not to include CSRBB in the inherent risk scoring matrix in Table 12. Introducing a centralised scoring system would imply a level of comparability that the current state of practice cannot support. Because scenario design, calibration choices and perimeter definitions naturally differ across institutions, any attempt to force these results into predefined scoring categories could lead to misleading conclusions. The current approach avoids this risk and is fully aligned with the SREP principles of proportionality and institution‑specific assessment.
We also believe that maintaining flexibility for institutions to design methodologies appropriate for their business models-combined with supervisory ability to review and challenge those approaches-creates an effective balance. This structure reflects the economic reality of CSRBB and avoids distortions that could arise from an attempt to harmonise rules for risk that is heterogeneous in nature.
Overall, the proposed CSRBB framework, supported by the Heatmap and integrated into the governance‑focused SREP components, provides a coherent and proportionate basis for supervisory assessment. It preserves the necessary discretion for institutions and supervisors while remaining consistent with the nature of CSRBB as a Pillar 2 risk.
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?
We welcome the revised liquidity and funding framework in Title 8 and appreciate the EBA’s efforts to enhance consistency and supervisory convergence. However, we would kindly request further clarification regarding the interaction between the supervisory liquidity stress testing framework (paragraph 386) and the LCR framework and its associated liquidity buffer.
Interaction between supervisory liquidity stress testing and LCR buffer (paragraph 386)
Paragraph 386 refers to supervisory liquidity stress testing as a key input to the SREP liquidity assessment. At the same time, the Guidelines do not explicitly specify whether the LCR liquidity buffer may be used in supervisory liquidity stress tests without automatically [KK1] [JR2] worsening the institution’s inherent liquidity risk score.
This uncertainty arises because:
- Title 8 merges the assessment of inherent liquidity risk and liquidity adequacy into a single “liquidity and funding adequacy score” (paragraph 31).
- Paragraph 50 clarifies that inherent risk and controls should be reflected in the risk score, but it does not define how temporary depletion of HQLA during supervisory stress should be treated.
Alignment with the intended regulatory role of the LCR buffer (paragraph 417(d))
We would like to highlight paragraph 417(d), which clearly states that the LCR buffer is designed to be used under stress, even if such use results in LCR values below 100%. The paragraph specifies that:
“the LCR liquidity buffer is designed to be used in case of stress, even if that leads to LCR values below 100%, […] and is part of the expected management of liquidity risk under stress.”
This statement reflects the core Basel and EU regulatory intent of the LCR framework.
Therefore, clarity is needed on how this principle is applied in the SREP scoring methodology.
Interaction Between Stress Test Outcomes, Table 19, and SREP Scoring
Because Title 8 merges inherent liquidity risk and liquidity adequacy into a single score, it is essential to clarify how the requirement to maintain liquidity buffers “comfortably above supervisory quantitative requirements” (as stated in Table 19 for Grade 1) interacts with the use of supervisory liquidity stress tests. In particular, it remains unclear whether this criterion applies only to the institution’s baseline liquidity position or also to post‑stress outcomes. If interpreted to apply after the supervisory stress scenario, institutions would effectively need to maintain significantly elevated pre‑stress LCR levels—potentially above 200% when stress severity is comparable to LCR assumptions—to ensure buffers remain “comfortably above” requirements after severe outflows. This would create a de facto minimum LCR far above the regulatory requirement.
To ensure scoring consistency across EU jurisdictions, in our view it is therefore necessary to clarify:
• whether the “comfortably above requirements” criterion refers solely to business‑as‑usual conditions;
• whether an LCR below 100% in a supervisory liquidity stress automatically precludes a Score 1 assessment; and
5.4. Harmonisation and level playing field across the EU
We respectfully note that without explicit guidance, supervisory practices may diverge, which could lead to inconsistent scoring practices under Section 8.5 and lack comparability of SREP outcomes across institutions with similar liquidity profiles. Providing explicit clarification would support supervisory harmonisation, reduce uncertainty, and strengthen the coherent application of EBA Guidelines throughout the EU.
Proposed clarifications
To support harmonised supervisory practice, we respectfully propose the following additions to the Guidelines:
Suggested addition to paragraph 386 (or immediately thereafter):
“For the avoidance of doubt, the supervisory liquidity stress test should take into account the regulatory design of the Liquidity Coverage Ratio (LCR) buffer as described in paragraph 417(d). In particular, the use of the LCR liquidity buffer under stress conditions — even if this results in temporary LCR values below 100% — should not by itself be considered an indication of excessive inherent liquidity risk, provided that the institution demonstrates sound liquidity risk management, appropriate escalation mechanisms, and adequate counterbalancing capacity.”
Suggested clarification to the scoring guidance in Table 19 of Section 8.5 or its accompanying text:
“For the avoidance of doubt, the Grade 1 criterion that liquidity buffers remain ‘comfortably above supervisory quantitative requirements’ shall be assessed with reference to the institution’s baseline liquidity position and forward‑looking projections under normal operating conditions. It shall not be interpreted as requiring liquidity buffers to remain comfortably above requirements within supervisory stress scenarios.”