22 Oct 2026 7 min read
Climate and nature-related risks can affect financial institutions through established categories of financial and non-financial risk, including credit, market and operational risk. Accordingly, their treatment does not require the introduction of a separate risk category, but their incorporation into existing risk-management processes and methodologies. The European Central Bank (ECB) approach reflects this principle, with climate and environmental considerations addressed across business strategy, governance, risk appetite, risk management, stress testing and capital adequacy.
The relevance of these risks can also be considered in relation to the environmental dependencies and physical exposures underlying banks’ portfolios. According to the evidence referenced in the presentation, 75% of euro-area corporate lending is directed towards firms that are critically dependent on at least one ecosystem service. In Greece, only 4% of natural-catastrophe losses have been insured, with uninsured losses having implications for collateral and stressed loss-given-default (LGD) estimates.
In May 2026, the ECB updated its November 2022 compendium of good practices for climate and nature risk management. Of the original 26 practices, 18 were updated with information observed since 2022, while five new good practices were introduced. The updated compendium also includes a dedicated chapter on the management of nature-related risks. Further additions concern prudential transition planning, physical risk, reputational and litigation risk, and capital adequacy within the Internal Capital Adequacy Assessment Process (ICAAP).
The update draws on supervisory observations covering the period from 2020 to 2025, while the practices added since 2022 were observed at more than 60 institutions. The resulting framework therefore provides evidence of how climate and nature-related risk-management practices have developed across supervised institutions.
Changes in the distribution of institutions across different levels of maturity are substantial. In 2022, 3% of significant institutions were classified as having leading practices, compared with 56% at the end of 2024. Over the same period, the proportion of institutions with no practices in place decreased from 25% to 5%, while the proportion relying on basic practices declined from 54% to 17%. Emerging practices accounted for 22% at the end of 2024, compared with 18% in 2022.
These developments are relevant to the interpretation of the 2026 compendium. The underlying framework has not been replaced. Rather, the observed practices indicate greater depth in specific areas of risk measurement and management, particularly in relation to physical risk, nature-related risk and risk modelling.
The same development can be observed in climate stress testing. By the end of 2024, 100% of significant institutions had incorporated climate risk into their stress-testing frameworks, compared with 41% in 2022. At the same time, the May 2026 report retains much of the framework established in 2022. Practices relating to the scope of stress testing, scenarios and balance-sheet assumptions remain unchanged, while data-related practices are described as “virtually unchanged”, with an additional emphasis on moving from NUTS 3 regions towards geo-coordinates.
The principal developments instead concern modelling. Transition-risk modelling is extended at counterparty level, physical risk is addressed through a dedicated modelling architecture, and nature-related risks are included in stress-testing practices for the first time. In this respect, the 2026 update suggests a progression from the establishment of stress-testing frameworks towards more detailed quantification of the transmission of climate and nature-related risks into conventional risk parameters.
For transition risk, the practices identified in 2022 remain the methodological foundation. These include sectoral satellite models, carbon-price pass-through and counterparty scorecards. The 2026 update provides additional practices at counterparty level, under which scenario-adjusted client financial information is incorporated into probability-of-default (PD) and rating models. Carbon-related shocks may similarly be translated into changes in equity values, spreads and ratings and, subsequently, into stressed PDs.
The treatment of transition risk is also extended to real-estate collateral. The updated practices include EPC-based LGD methodologies at property level, through which the transition risk of real-estate collateral is reflected in LGD on the basis of the EPC-class change required for alignment with the Energy Performance of Buildings Directive. Additional approaches include the incorporation of cost pass-through into valuations, recovery-rate models and PD-LGD consistency functions.
Physical risk represents one of the principal areas of methodological development. While it was not covered in depth in the 2022 report, the 2026 update sets out practices covering the modelling process from hazard identification through to stressed PD and LGD estimates.
The process begins with hazard-specific indicators at exposure-location level, geo-coordinates at building level and forward-looking hazards under Representative Concentration Pathways (RCPs). Physical effects are subsequently considered through direct and indirect transmission channels. Direct channels include damage to production sites, cash flows and collateral values, while indirect channels include effects transmitted through GDP, sectoral gross value added, inflation and labor productivity.
For credit-risk modelling, physical damages may be incorporated into borrower financial information and subsequently into stressed PDs. In the case of LGD, the practices include collateral haircuts determined according to event probability and severity, as well as the transmission of property-level damage through stressed loan-to-value ratios into stressed LGD.
The requirement for greater geographical granularity is directly related to this methodology. Physical shocks can differ within individual NUTS 3 regions, and the 2026 practices therefore refer to geo-coordinates across asset classes at the highest granularity possible.
Nature-related risk constitutes a new component of both the risk-management and stress-testing material published in May 2026. Within the risk-management compendium, the dedicated nature chapter covers materiality, governance and risk appetite, due diligence and scoring, risk mitigation and capital adequacy.
Within stress testing, the methodologies remain at an earlier stage of development. The report describes nature-risk methodologies as “nascent” and identifies exploratory scenario analyses designed to develop approaches for quantifying their impact, including direct transmission channels. Nature-specific scenarios have also been incorporated into climate stress-testing frameworks to model macroeconomic shocks arising from nature-related risks.
The identified practices distinguish between physical and transition drivers. Physical drivers include resource scarcity and the resulting increases in water and biomass prices. Transition drivers include environmental taxation and regulation and their implications for operating costs. Although the methodologies are characterized as nascent, the practices included in the report had been observed at supervised institutions by the end of 2024.
The May 2026 publications should be considered in the context of the supervisory sequence that began with the ECB Guide on climate-related and environmental risks in November 2020. The subsequent process included the 2022 thematic review and first compendium of good practices, phased remediation deadlines culminating at the end of 2024, and supervisory follow-up activities between 2023 and 2025. The updated compendium and stress-testing report were published in May 2026, while the EBA Guidelines on environmental scenario analysis apply from 1 January 2027.
Importantly, the 2026 good practices are explicitly described as non-binding and illustrative and do not themselves establish supervisory expectations. Nevertheless, their content documents the practices identified through supervisory observations and provides a reference for examining the development of climate and nature-related risk management across institutions.
From this perspective, the principal development in 2026 concerns the level of methodological depth applied to existing risk-management and stress-testing frameworks. The earlier foundations concerning frameworks, scenarios and core data remain largely unchanged, whereas the updated practices provide greater detail on counterparty-level transition modelling, physical-risk modelling and the initial incorporation of nature-related risks into scenario analysis.
Subscribe here to receive our newsletters
The ECB’s July 2026 clarification reinforces existing supervisory expectations, with greater emphasis on how ICAAP and ILAAP are embedded, governed and evidenced in practice.
The EBA’s proposed methodology combines a lighter reporting framework with a more risk-sensitive approach and the first structured integration of climate risk
Targeted amendments following CRR3 preserve the core prudential framework while introducing important clarifications and operational changes.
