Response to consultation on amending Guidelines on the appropriate subsets of exposures in the application of the systemic risk buffer
Q1. Do you agree that the proposed use of more granular economic activity classifications (including NACE level 2 or more granular levels where necessary) is appropriate and suffi-cient to enable authorities to effectively target exposures subject to climate transition risk while limiting unintended consequences for transition financing? If not, please explain and suggest alternative approaches or safeguards.
The use of NACE classification is general practice. However, we do not consider such classifications to be sufficient on their own to effectively target exposures subject to climate transition risk while limiting unintended consequences for transition financing.
Even within the same NACE category, risk profiles can differ materially depending on factors such as technology mix, energy sources, transition plans, geographic location and time horizon. This is particularly evident in sectors such as energy, manufacturing and transport, where transition pathways vary significantly within a single NACE code and risks penalizing companies actively transitioning while failing to distinguish laggards.
In addition, NACE codes such may not effectively identify exposures subject to climate transition risk, since in some cases NACE classifications may not fully or accurately capture company’s activities. Large companies in particular often operate across several business lines with different risks. Overreliance on NACE classifications therefore risks masking heterogeneity and misidentifying actual risk.
The climate-risk exposure embedded in a bank’s credit portfolio is inherently bank-specific, reflecting the institution’s own client-engagement and selection practices. Accordingly, the Pillar 2 framework is the appropriate tool to capture these bank-specific portfolio characteristics-something an overly generic SyRB would inevitably overlook.
Also, transition risks are strongly influenced by a dynamic policy environment, which static classifications cannot adequately capture.
Moreover, granular classifications also significantly increase data demands, while reliable, comparable and forward-looking data, especially for SMEs, remain limited. Also, for institutions with predominantly retail portfolios, the practical relevance of NACE classifications is more limited, as a significant share of climate-related exposures currently arises from residential mortgage lending rather than corporate activities.
It is therefore important that:
•Enhanced granularity remains proportionate for institutions whose portfolios are primarily SMEs and retail in nature.
•Authorities carefully assess unintended consequences for transition financing, particularly where sectors undergoing transformation may temporarily exhibit higher risk profiles while contributing to long-term climate objectives. Any measures undermining financing of companies’ transition should avoid systemic penalization of transitional sectors that are actively reducing emissions.
•Reliable proxies may be used ( subject to minimum quality criteria and a clear review process governing their selection, validation and periodic updating), thereby strengthening methodological consistency across jurisdictions and enhancing the comparability of results.
•Public, centralized and interoperable databases are created.
Effective targeting of exposures subject to climate transition risk would require an approach that would consider the transition and investment plans of corporates. The methodology the ECB uses for its climate factors is interesting in this regard and could be further considered by EBA as it adjusts assets by an uncertainty score, which is composed of three elements:
1. a sector-specific stressor: a uniform “market factor” derived from the expected shortfall in the adverse scenario of the Eurosystem climate stress test, which applies to all assets issued by firms within a specific sector;
2. an issuer-specific exposure: a measure of an issuer’s exposure to transition-related uncertainties, based on the methodology developed for the tilting of the Corporate Sector Purchase Programme;
3. an asset-specific vulnerability: an assessment of how sensitive an asset’s market price is to unexpected future climate shocks, taking into account its residual maturity.
Q2. Do you consider that introducing an additional subdimension related to Energy Perfor-mance Certificates (EPCs) or energy consumption buckets within the risk profile would be appropriate to better capture climate-related risks? If so, please comment on its relevance and potential implementation challenges, as well as data availability and possible proxies that relevant authorities could consider.
At this stage, there are important implementation challenges:
• Significant data quality and availability issues across Member States, including outdated EPCs and missing updated labels after renovations (clients often only seek an EPC when they need it for a loan, and do not renew it afterwards). Many certificates are outdated, particularly for older buildings.
• Non-uniformity across EU Member States, which makes EPCs unsuitable for comparative or prudential purposes. Methodologies and quality standards vary across jurisdictions.
• Retail portfolios require scalable, automated integration of large datasets.
• Risk mitigating factors already present (e.g., low LTV ratios) make EPC based adjustments in risk profiles disproportionate. EPC values are already reflected in collateral valuations. Using EPCs again within a SyRB sub‑dimension would constitute double‑counting and questionable supervisory insight.
• The EPC class alone might not adequately reflect risk across European countries. Carbon prices likely affect households with fossil fuel dependency disproportionately.
• EPC is a European metric, not available in other jurisdictions which is an important consideration for European banks operating in other jurisdictions.It would be very challenging to take this additional dimension in consideration for assets located in countries where EPC are not mandatory and/or market prices do not apprehend such dimension.
Accordingly, it is recommended that:
• At the moment, EPC use in SyRB is avoided.
• Minimum harmonized quality standards are established across Member States.
• Clarification is provided on how differences between energy classes would translate into capital implications, including in cases where EPCs do not reflect the current condition of the property.
• The framework explicitly avoids discouraging financing for renovation and energy-efficiency improvements, which are central to climate transition objectives. An ill-conceived SyRB targeting energy-intensive assets could prevent the EU and its member states from achieving their greenhouse gas reduction targets.
Q3. Do you consider the proposed extension of geographical granularity (including the use of LAU level) appropriate for identifying exposures subject to climate physical risks? Please comment on the relevance of the proposal and potential implementation challenges, as well as data availability and possible proxies that relevant authorities could consider.
Extending geographical granularity to the LAU/municipality level may be conceptually attractive for capturing the local nature of certain physical hazards. For mortgage portfolios, physical climate risk often directly affects real estate collateral values, making geographical precision particularly relevant. However, it presents significant limitations for widespread prudential application.
While the use of LAU level (municipalities) is consistent with the structure of the current sectoral SyRB framework, LAU granularity is too crude to meaningfully capture physical climate risks:
• Climate data and many available hazard maps do not consistently offer robust resolution at the municipal level (provincial or regional scales being more common), which can create a sense of false precision.
• LAU boundaries may not fully capture micro-geographical variations (e.g. flood-prone zones within the same municipality).
• Physical climate risk mapping in practice typically relies on much finer spatial units than LAU. Effects would differ per sector and the granularity would not match the manner in which these risks are currently captured in maps, e.g. flood maps. LAU level application may therefore distort the true risk profile, because the areas affected by physical hazards often cut across or sit within only part of a municipality. A more risk sensitive and credible approach requires higher resolution spatial data, harmonized across jurisdictions and aligned with actual hazard footprints.
• Access to harmonised and reliable physical risk maps differs across Member States.
•Material gaps in geolocation persist across several portfolios, especially in Corporates and Project Finance:
- Corporates: Identifying productive assets usually requires the acquisition and consolidation of heterogeneous external sources (commercial and/or open-source) with uneven coverage.
- Project Finance: additional difficulty regarding linear assets and obtaining detailed layouts or locations.
- Stock: incomplete postal addresses are frequent.
- Retail portfolios: involves large volumes of granular exposures requiring operationally feasible data solutions.
• Volume of data and its processing require long and complex internal technological developments (GIS infrastructure, integration, governance, and controls
• LAU is a European standard that cannot be extrapolated to other geographies, which is critical for European banks with global portfolios. Therefore, the use of globally comparable administrative levels (e.g., GAUL/ADM) should be considered. It is essential to ensure that criteria are stable, auditable, replicable, and comparable - a condition that is particularly necessary if the output is used to calibrate a SyRB.
It would be beneficial to:
• Promote harmonized European physical risk databases or at least promote comparable methodologies in assessing the physical risk levels across banks. The standardization and availability of a common climate/hazard dataset at the banking (or supervisory) level, will allow entities to conduct their analysis using the same database, thereby reducing methodological dispersion and uncertainty. Standardization should include, at a minimum: (i) hazard, (ii) exposure, (iii) vulnerability/damage, (iv) outcome metric, and (v) homogeneous reporting.
• Use proxies:
- Permit the use of recognized public or insurance-based risk maps as appropriate proxies
- For certain risks and counterparties granular spatial information is not able to capture the risk (e.g. business interruption) as risk is emerging through the supply chain dependencies. A form of an upscaled proxy is necessary to adequately cover specifically economic losses from business interruption
• Recognize levels of adaptation (lowest entry point for water of an asset) to stimulate adaptation investments
• Base the location on the available geocoding of the portfolio and aggregate consistently to a global standard level (e.g., ADM2/GAUL) when lower granularity is not feasible.
• Assess whether the use of postal code level data is sufficiently detailed in EU Member States to truly extract the exposure to climate risks. There should be consistency in geographical mapping in the various countries, not LAU level in one country and postal code in another country
• Consider the different effects of acute (with possible severe occurrence in one or more locations simultaneously) and chronic impacts (can be anticipated and mitigated)
Q4. Do you agree with the proposed flexibility to combine different dimensions (e.g. type of counterparty, economic activity, geographic area, type of collateral) when defining subsets of sectoral exposures for SyRB purposes? In your view, does this flexibility sufficiently sup-port risk sensitivity while preserving transparency and comparability across jurisdictions?
Climate related risks are multi factor and context specific, therefore we recognize that the combination of various data points is key in understanding climate risks (such as type of counterparty, economic activity, geographic area and type of collateral). Allowing competent authorities to combine relevant dimensions may therefore help to better reflect underlying risk drivers, enhance risk sensitivity and avoid overly broad or blunt sectoral measures that are insufficiently risk based.
However, it should be recognized that, in particular for retail portfolios, certain dimensions (e.g. economic activity of natural persons) may have limited relevance to climate risk and excessive combinatorial complexity may increase operational burden without proportionate gains in risk differentiation.
While we support flexibility, we would like to stress that flexibility should not come at the expense of transparency, proportionality and cross jurisdictional comparability. Excessive discretion or overly granular national approaches risk:
•Reducing the comparability of capital requirements across Member States.
•Increasing complexity and compliance burden, particularly for cross border banking groups.
For example, the subdimension “approach to calculate risk-weighted exposure amount for credit risk” (Standardized vs IRB), may support implementation clarity, however a heterogeneous application across jurisdictions could affect comparability and reciprocity.
In our view, flexibility can only be effective if exercised within a clear and harmonized framework, supported by adequate safeguards. In particular:
•The rationale and methodology for combining dimensions should be clearly articulated and grounded in robust risk evidence.
•Operational complexity should remain aligned with actual risk materiality
•Granularity should remain proportionate, avoiding narrowly defined exposure subsets based on weak or proxy driven indicators. Proportionality considerations should be explicitly recognized.
•Data availability, quality and consistency across jurisdictions should be a key precondition, to prevent mechanistic or unstable outcomes.
•Transparency towards institutions and markets is essential, including clear communication of how different dimensions interact in the risk assessment (e.g. clarification on which dimensions are most relevant for natural persons). Transparency and cross-border comparability must be preserved.
•Measures should be regularly reviewed, allowing for adjustment as data quality and risk understanding evolve.
As already mentioned, the introduction of additional complexity in the SyRB design should be carefully assessed against the broader EU level strategic direction, namely the current EU level policy focus on simplification and reduced regulatory complexity. Introducing highly granular or multi layered ESG related sub dimensions into the SyRB framework may be difficult to reconcile with these broader objectives.
Even with greater granularity, a SyRB framework cannot assess the particular characteristics of banks’ portfolios, their climate-risk exposures, or the way banks integrate this risk into their risk-management and loan-origination policies. As noted earlier, only the Pillar 2 framework is capable of capturing these nuances.
Q5. Do you consider the strengthened provisions on information sharing and the use of harmonised data sources adequate to facilitate the assessment and reciprocation of SyRB measures across Member States? Please indicate any remaining obstacles to effective reciprocity and how they could be addressed.
As a principle, strengthened provisions on information sharing and prioritization of harmonized data sources are supported, as they are essential to facilitate reciprocity across Member States.
However, to date, the Member States continue to use different methodologies for physical risk mapping, different reporting templates, scope of application, definitions, and data platforms. Data quality also varies. This fragmentation directly jeopardizes consistent identification of sectoral or geographical subsets necessary for SyRB reciprocity.
Climate and environmental risk data (e.g., physical hazard maps, EPC frameworks, transition risk indicators) remain non-uniform and non-comparable, meaning authorities could reach different conclusions when assessing the same exposures.
To enhance effective reciprocity, consideration could be given to:
• Developing interoperable physical risk databases.
• Standardizing templates for defining subsets of exposures.
• Ensuring clear disclosure of methodological assumptions where non-harmonised data are used. This should include an objective description of the key technical characteristics of the data employed (such as their source, version and level of detail),in order to enhance the transparency of the analyses and facilitate their replication across jurisdictions.
• Providing practical guidance to maintain operational feasibility of reciprocity assessments.
• Establishing a clear ESRB methodology for application of the reciprocity principle that would consider that exposures already surcharged locally should not be included in the calculation that determines that exposures affected by the SyRB in a group are sufficiently material to trigger reciprocity. The CRD should be aligned with the ESRB framework and clarify the calculation of the reciprocity activation threshold. In addition, the ESRB methodology should clarify that reciprocity should not be applied at a level of consolidation that does not exceed the materiality threshold set by the competent authorities. More specifically, the calibration of the materiality threshold for the reciprocity of a sectoral SyRB should reflect the approach set out in Recommendation ESRB/2017/4 (amending ESRB/2015/2) and the ESRB Handbook on operationalising macroprudential policy, whereby the threshold is assessed at entity level and complemented by a relative metric linked to the relevant exposures. Exposures that have already triggered the SyRB at local level should therefore be excluded from the threshold calculation, as their inclusion at consolidated level leads to multiple counting effects (local application, threshold assessment and subsequent consolidated application), resulting in an over-calibration of capital requirements not aligned with the objective of Article 133 CRD. Moreover, reciprocity is intended to mitigate cross-border spillovers (both inward and outward), and therefore the materiality threshold should focus exclusively on exposures that give rise to such spillovers. In this context, exposures already subject to a SyRB at local level—typically originated and held via subsidiaries—should be excluded from the threshold calculation, as they do not create cross-border arbitrage risks. Including these exposures would not contribute to the objective of the measure and would instead lead to double counting and potential competitive distortions across Member States
Q6. Do you have any additional comments on these draft Guidelines amending EBA/GL/2020/13 on the appropriate subsets of sectoral exposures to which competent or designated authorities may apply a systemic risk buffer?
Authorities should refrain from applying macro prudential measures to address climate risks since such measures would be:
• crude in nature, as the effects of climate change are likely not to be distributed generally over sectors and geographies and across banks portfolios,
• create wrong incentives,
• undermine transition finance and
• undermine the development of risk management methodologies.
Please see our reasoning elaborated below:
Risk based approach
To ensure that capital to cover climate risk is risk-based, pillar 2 is most suitable at this point. We invite the EBA and the ECB to make P2R calculations more transparent in the revision of the SREP Guidelines to facilitate institutionalized learning.
Financial inclusion, regional equity and impact on transition financing
Authorities should assess potential regressive effects and ensure that the SyRB calibration does not unintentionally contribute to financial exclusion. A sectoral or concentration SyRB presents several downsides. It could severely impact the transition for those sectors most needed. These are often the sectors where the financing of the transition is the most relevant. Hence, implementing a capital buffer for exactly these sectors would limit the room for banks to help these sectors in their transition. The Fit for 55 climate scenario analysis performed by the EBA also showed that energy intensive sectors would require the highest level of financing to meet EU greenhouse gas reduction objectives. A climate-related SyRB targeting these sectors would thus likely make the EU reduction targets more difficult to achieve and would directly contradict one of the Union’s key objectives in environmental policies.
Applying a SyRB to geographically defined high-risk areas may indirectly affect access to credit for households located in vulnerable regions. This may have distributional implications, particularly where households have limited capacity to relocate. Higher capital requirements applied to certain geographical areas may also influence property valuations, loan-to-value ratios and credit supply dynamics. Macroprudential assessments should therefore carefully consider whether such measures could inadvertently amplify systemic risks in real estate markets, including second-round effects.
Also, to avoid disincentives to renovation and transition investments, the SyRB framework should distinguish between:
• Properties currently exposed to higher physical or transition risk.
• Properties subject to credible mitigation or renovation plans.
• Renovated or energy-upgraded properties.
Capital calibration should avoid discouraging financing for energy efficiency and climate resilience improvements, which are essential for achieving climate objectives
Double counting
Introducing a SyRB for climate related risks would result in double counting. In recent years, European regulators, supervisors and banks have focused on strengthening the micro prudential and accounting framework to address risks stemming from climate change. The integration of ESG risk drivers into the ICAAP has increasingly become standard practice, and climate related risks are already addressed under the EBA Guidelines on ESG risk management and will be very shortly addressed in the proposed amended EBA SREP Guidelines. Risks that are identified but not adequately managed or covered by an institution should be captured under Pillar 2. The ESG materiality assessment serves as the basis for identifying such risks, and insufficient risk identification should lead to supervisory feedback, findings or Pillar 2 measures, not to the application of a SyRB.
In case the supervisor decides to implement macroprudential buffers for climate, it should be able to formulate clearly what risks are not yet covered in the loan loss provisions and/or in the risk weighting of exposures or under the micro-prudential measures under Pillar 2. Transparency, clear communication of scope and calibration rationale, and preservation of cross-jurisdiction comparability remain essential.