Report results from the 2025 market risk benchmarking exercise - ASA

General Information

This document is a report published by the European Banking Authority (EBA) in June 2026, referenced EBA/REP/2026/14. It presents the results of the 2025 benchmarking exercise on market risk, specifically on the Alternative Standardized Approach (ASA) of the FRTB (Fundamental Review of the Trading Book). The report mainly covers data collected on the Sensitivities-Based Method (SBM), as well as the Default Risk Charge (DRC) and Residual Risk Add-On (RRAO) components of the ASA. The scope includes detailed analysis of submissions from European banks on several standard portfolios, broken down by asset classes (equities, interest rates, foreign exchange, commodities, credit spreads) and risk components (delta, vega, curvature). The report relies on approximately the first 37 pages provided, out of a total of 82 pages.

Executive Summary

The report addresses the results of the 2025 benchmarking exercise on market risk under the Alternative Standardized Approach (ASA) of the FRTB, focusing on the Sensitivities-Based Method (SBM) and the DRC and RRAO components. This exercise is essential to assess the consistency and quality of regulatory calculations of capital requirements for market risk within the European Union. The main findings are:

- The completeness of SBM OFR (Own Funds Requirements) data submitted by banks is generally satisfactory, with a high number of submissions for most portfolios (p. 10-11).

- Dispersion of SBM OFR results has gradually decreased since 2022, with an interquartile dispersion (IQD) average dropping from 16% in 2022 to 8% in 2025, signaling improved practices and data quality (p. 7-8).

- Dispersion is mainly due to the delta risk component, particularly in equity, commodity, interest rate, and credit spread portfolios (p. 23-24).

- Three major issues affect data comparability and consistency: heterogeneous treatment of foreign exchange risk in non-FX portfolios, errors or divergences in assigning sensitivities to regulatory buckets, and discrepancies in applying the risk aggregation formula (p. 36-40).

- The SBM validation exercise, which checks the correct application of the aggregation formula, shows positive results with a majority of aligned institutions, but participation remains low (around 25%) and some banks show significant deviations (p. 41-43).

- The DRC and RRAO components are calculated relatively consistently, but inconsistencies in submissions (some institutions not reporting all components) increase overall dispersion (p. 49-59).

- Competent authorities emphasize the importance of these exercises to identify divergence areas and improve regulatory harmonization (p. 61-64).

Implicit recommendations are to strengthen supervision on correct instruction application, notably on foreign exchange risk treatment, bucketing, and aggregation, as well as to encourage broader participation in validation exercises to ensure better comparability and data reliability.

Context and Objectives

This report was drafted as part of the fourth annual benchmarking exercise of the SBM method of the ASA under the FRTB, as well as the second exercise including the DRC and RRAO components. The objective is to assess the practical implementation of the ASA methodology by European banks, identify gaps and sources of dispersion in capital requirement calculations for market risk. The benchmarking aims to provide competent authorities with precise elements to supervise and harmonize the application of the FRTB regulatory framework. The scope covers several standard portfolios, broken down by asset classes and risk components, based on data submitted by banks in 2025. The report focuses on the analysis of SBM OFR data, with illustrations of sensitivities, and also examines the DRC and RRAO components. Main limitations lie in the complexity of sensitivity data, partial participation of banks in some components, and divergences in interpreting regulatory instructions.

Summary of Key Points by Theme

Completeness and quality of SBM data:

- The submission rate of SBM OFR data is high and consistent with IMA submissions, mainly covering IR, FX, EQ, CS, and CM portfolios (p. 10-11).

- Dispersion of SBM OFR results, measured by IQD, has steadily decreased since 2022, indicating improved practices (8% in 2025 versus 16% in 2022) (p. 7-8).

Dispersion and sources of variability:

- The delta component is the main contributor to OFR dispersion, notably in equity, commodity, interest rate, and credit spread portfolios (p. 23-24).

- Dispersion is lower for vega and curvature components, except in portfolios designed to test these risks (p. 24).

- The impact of correlation scenarios (high, medium, low) on OFR is limited for most portfolios (p. 21-22).

Sensitivity analysis by asset class:

- Equities: portfolios with low dispersion (e.g., 1107, IQD 1%) show homogeneous sensitivities, while others (e.g., 1019, IQD 7%) reveal divergences linked to indexing approach (p. 26-27).

- Interest rates: simple portfolios (e.g., 2205, IQD 2%) present homogeneous delta sensitivities, but more complex portfolios (e.g., 2013, IQD 41%) suffer divergences related to bucketing and foreign exchange risk treatment (p. 28-29).

- Foreign exchange: strong consistency (IQD 2%), notable improvement compared to 2024 (p. 30-31).

- Commodities: significant improvement in delta sensitivity consistency (e.g., 4401, IQD 7% versus 57% in 2024), but bucketing issues persist (p. 32-33).

- Credit spreads: high dispersion (e.g., 5516, IQD 42%) mainly due to divergences in foreign exchange risk treatment and bucketing (p. 34-35).

Identified issues and supervisory recommendations:

- Heterogeneous treatment of foreign exchange risk in non-FX portfolios, due to divergent interpretations of instruction "kk" in annex 2, causing artificial dispersion inflation (p. 36-38).

- Errors in assigning sensitivities to regulatory buckets, a significant source of dispersion, especially in certain specific instruments (p. 38-40).

- Suspected inconsistencies in applying the risk aggregation formula, despite generally positive results in the SBM validation exercise (p. 40-43).

SBM validation exercise:

- Limited participation (around 25% of banks), but mostly consistent results with IQD close to zero for most portfolios, especially on delta risks (p. 41-43).

- Some significant deviations persist, not captured by IQD, located at distribution tails (p. 42).

DRC and RRAO components:

- Calculations generally consistent, but uneven bank participation on these components increases overall ASA OFR dispersion (p. 49-59).

- Improved submission consistency compared to 2024, but divergences remain (p. 40, 49-59).

Role of competent authorities:

- Detailed analyses identify portfolios and components to monitor for harmonizing practices (p. 5, 61-64).

- Need to clarify instructions, notably on foreign exchange risk and bucketing, and encourage compliance with validation requirements (p. 36-41, 61-64).

Main Findings and Lessons Learned

Established facts:

- SBM OFR data collection is complete and covers the majority of submitted portfolios (p. 10-11).

- Dispersion of SBM OFR steadily decreases, with an average IQD of 8% in 2025 versus 16% in 2022 (p. 7-8).

- The delta component is the main source of dispersion, while vega and curvature contribute less (p. 23-24).

- Foreign exchange risk treatment in non-FX portfolios is heterogeneous, causing artificial dispersion inflation (p. 36-38).

- Bucketing errors and divergences in applying the aggregation formula are additional dispersion sources (p. 38-40).

- The SBM validation exercise shows a majority of aligned banks, but participation remains low (25%) (p. 41-43).

- The DRC and RRAO components are calculated relatively consistently, but uneven participation increases overall dispersion (p. 49-59).

Assumptions and interpretations:

- Residual dispersion is attributed to regulatory interpretation differences, notably on foreign exchange risk and bucketing (p. 36-38).

- The limited impact of correlation scenarios on median OFR suggests scenario differences are not a major dispersion factor (p. 21-22).

- Low participation in the validation exercise limits conclusions on widespread application of the aggregation formula (p. 42).

Uncertainties:

- The real impact of identified divergences on banks’ actual portfolios remains to be confirmed, as the exercise uses hypothetical portfolios (p. 21-22).

- Partial bank participation in some components and exercises limits result representativeness (p. 41-43).

- Some extreme deviations are not captured by statistical measures used (IQD), requiring increased vigilance (p. 42).

Conclusions and Recommendations

The report concludes that the 2025 benchmarking exercise on the FRTB ASA, particularly the SBM method, has improved understanding and consistency of regulatory capital requirement calculations for market risk. The progressive decrease in SBM OFR dispersion reflects better practice homogeneity. However, significant divergences remain, mainly related to heterogeneous interpretation of foreign exchange risk treatment in non-FX portfolios, errors in sensitivity bucketing, and discrepancies in applying the aggregation formula. The SBM validation exercise, although positive, suffers from low participation, limiting the scope of conclusions. The DRC and RRAO components are generally consistent, but uneven participation increases overall dispersion. The report implicitly recommends competent authorities strengthen supervision on these key points, encourage broader participation in validation exercises, and clarify regulatory instructions, notably on foreign exchange risk and bucketing, to improve harmonization and reliability of ASA calculations within the European Union.