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.
Caelra Capital response:
We agree that NACE level 2 granularity represents a significant improvement over the current framework and is appropriate for transition risk identification. However, we submit that NACE level 2 granularity alone is insufficient to identify exposures subject to water stress as a physical risk, for the following reasons:
Water stress risk within NACE sectors is heterogeneous. Within NACE C20 (Manufacture of chemicals and chemical products), a facility drawing cooling and process water from the Rhine in Germany faces materially different water risk from a comparable facility in Scandinavia. Both are classified identically at NACE level 2 and even at more granular levels. The risk differential is not captured by economic activity classification alone — it requires the combination of NACE activity with geographic water stress data.
Similarly, within NACE D35 (Electricity, gas, steam and air conditioning supply), thermal power generation facilities — which are critically dependent on cooling water availability —face documented operational risk from river temperature limits and low flow events. The2022 reduction in French nuclear output due to river temperature constraints is an example of a water stress event that impaired credit-relevant operational continuity in an asset class already included in national net-zero frameworks for emissions risk. Water was the unmodelled second channel.
We recommend:
The EBA consider explicit guidance that, for water-intensive sectors —including but not limited to NACE C20, C21, C10-C12, D35, A01, and C26 — water stress indicators should be incorporated as a named sub-dimension under physical risk classification, complementing NACE economic activity classification with geographic water stress overlays.
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.
Caelra Capital response:
We note that EPC and energy consumption metrics address transition risk and energy efficiency. While we do not object to their inclusion, we submit that an analogous water-specific metric — water consumption intensity or water stress exposure rating — would beat least as material for physical risk identification in water-intensive sectors.
Water consumption data at facility level is increasingly available through corporate sustainability reporting (CSRD Article 29e requires water consumption and water intensity disclosure), and through industrial water permit registries at Member State level. A water consumption intensity subdimension, analogous to energy consumption buckets, would allow regulators to identify high water-consuming counterparties within already-granular NACE classifications.
The CSRD disclosure framework creates the data infrastructure for this approach. Banks are already receiving water consumption data from large corporate counterparties through CSRD reporting chains. The prudential framework should be designed to make use of this data rather than allowing it to remain siloed in sustainability disclosures.
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.
Caelra Capital response:
We strongly support the move to LAU-level geographic granularity. For water stress specifically, this is essential. Water availability is a hyper-local phenomenon determined by watershed boundaries, aquifer extent, and river basin dynamics — none of which correspond to NUTS administrative boundaries.
We note three specific points:
First, groundwater depletion — unlike surface water events — is not captured by flood or drought event frequency at NUTS or LAU level. It is a slow-moving, cumulative physical risk that impairs long-run asset and collateral values. Agricultural land in the Po Valley, the Guadalquivir basin, and parts of Eastern Europe is already experiencing groundwater depletion that does not appear in current collateral valuation methodologies. LAU-level granularity, combined with groundwater stress data (available from the GRACE-FO satellite mission and national hydrogeological surveys), would enable a materially more accurate identification of physical risk in real estate and agricultural loan collateral.
Second, the proposed framework's use of geographic location primarily in combination with "type of collateral" (paragraph 11 of the consultation paper) should be extended to also capture the location of the counterparty's operations. A chemical manufacturer located in a water-stressed LAU faces physical risk to operational continuity regardless of where its loan collateral is located. The combination of NACE economic activity with LAU-level water stress data captures this transmission channel.
Third, we note that WRI Aqueduct provides basin-level water stress scores that can be mapped to LAU boundaries. WWF Water Risk Filter provides sector-specific water risk assessments at comparable geographic granularity. These are established, publicly available tools that could serve as harmonised data sources for water stress identification under the amended Guidelines, reducing the data gap concerns raised in paragraph 9 of the consultation paper.
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?
Caelra Capital response:
We strongly support this flexibility. The water stress transmission mechanism is precisely the type of risk that requires multi-dimensional identification — it manifests at the intersection of economic activity (water-intensive sectors), geographic location (water-stressed basins), and type of counterparty (operational exposures vs. collateral exposures). No single dimension captures it adequately.
We would specifically recommend that the final Guidelines include an illustrative example— alongside the existing examples of "Mining of coal and lignite" and "Manufacture of coke and refined petroleum products" — that demonstrates how a water stress SyRB measure might be designed using the combination of NACE C20 or D35 with LAU-level water stress geographic classification. This would signal to national authorities that water stress is a legitimate and analytically tractable physical risk dimension within the amended framework.
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.
Caelra Capital response:
We support the strengthened information sharing provisions. We note that for water stress specifically, harmonised data sources already exist at European and global level — WRI Aqueduct, WWF Water Risk Filter, GRACE-FO groundwater data, and national hydrogeological surveys — and their use as reference data sources for water stress SyRB measures should be explicitly recognised in EBA technical guidance.
We further note that the European Environment Agency (EEA) publishes river basin management plans under the Water Framework Directive that contain basin-level water stress assessments for all EU Member States. These are official, harmonised, and updated on six-year cycles. They represent an underutilised resource for banking supervisors seeking to identify geographically differentiated water stress exposures.
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?
The ECB's 2023 report on nature-related financial risks identified water-related risks as the most material nature-related risk for the euro area economy. The proposed amendments to EBA/GL/2020/13 represent an important step toward a more granular and risk-sensitive macroprudential framework. We submit that explicitly incorporating water stress as a named physical risk variable — through the combination of NACE sector classification with geographic water stress overlays at LAU level — would materially strengthen the framework's ability to identify and address one of the most consequential but currently underspecified sources of systemic physical risk in European bank loan books.
Caelra Capital is available to provide further technical input on water risk measurement methodologies and data sources if this would be useful to the EBA's finalisation process.