This final report presents the proposed amendments to the regulatory technical standards concerning material changes to internal ratings-based (IRB) models under the CRR regulation. The amendments aim to simplify the approval process for model changes by reducing the number of changes classified as material while maintaining an appropriate information flow to supervisors. The goal is to enhance supervisory…
- Title: Final report on draft RTS on IRB material model changes
- Author/organization: European Banking Authority (EBA)
- Date: March 30, 2026
- Type: regulatory standard (Regulatory Technical Standards - RTS)
- Scope: amendments to the delegated technical standards (Commission Delegated Regulation (EU) No 529/2014) relating to the assessment of materiality of extensions and changes to internal models under the Internal Ratings Based (IRB) approach for the calculation of risk-weighted exposure amounts (RWEA) in credit.
- Target population: credit institutions using the IRB approach in the European Union.
- Sector: banking sector, prudential supervision.
- Period: regulatory framework updated in 2026, incorporating CRR3 amendments (Regulation (EU) 2024/1623).
The document addresses the revision of the delegated technical standards (RTS) governing the assessment of materiality of changes and extensions to IRB internal models used by banks to calculate credit capital requirements (RWEA). This revision is mandated by Article 143(5) of the CRR, in connection with amendments introduced by CRR3.
The main issue is to improve the efficiency of the approval process for IRB model changes, which, after more than ten years of experience, has shown limitations: supervisory authority overload, excessive delays, uncertainty on implementation dates, and hindrance to model improvement and effective use.
Key findings are:
- The current framework distinguishes material changes (requiring prior approval) from non-material changes (requiring ex ante or ex post notification).
- The high volume of material changes has saturated supervisory resources.
- Regulatory changes from CRR3 require removal of references to IRB models for equity exposures and the AMA approach for operational risk.
Conclusions are that the new version of the RTS should:
- Align the framework with CRR3 by removing obsolete references.
- Reduce the number of changes considered material by strengthening quantitative thresholds (e.g., a decrease of 1.5% or 15% of RWEA) to trigger materiality.
- Maintain a restricted list of qualitative criteria qualifying a change as material (e.g., complete model redevelopment, fundamental change in default definition).
- Clarify the scope of changes subject to approval or notification, notably excluding purely technical alignments related to CRR3 that do not impact model performance.
- Introduce strengthened documentation requirements to better justify the nature and impact of changes.
Recommendations are:
- Apply the new criteria to classify changes and extensions.
- Bundle similar modifications to avoid dilution of quantitative impacts.
- Submit the amended RTS to the European Commission for approval and publication.
Thus, this report aims to streamline IRB model supervision, reduce implementation delays for changes, and enable more efficient allocation of control resources while ensuring the robustness of models used for capital requirement calculations (p. 3-4).
- The European regulatory framework requires banks to obtain approval from competent authorities to use the IRB approach for calculating RWEA and for any material extension or change to their internal models (Article 143 CRR).
- Commission Delegated Regulation (EU) No 529/2014 (CDR on MC) specifies qualitative and quantitative criteria to assess materiality of changes and extensions.
- Since the initial publication in 2014, supervisory experience has shown that the high number of material changes overloads authorities, causing delays and uncertainty in model implementation.
- CRR3 modifies the regulatory framework, notably removing the IRB approach for equity exposures and the AMA approach for operational risk.
- The main objective is to revise the RTS to align the framework with CRR3, reduce the number of changes classified as material, and improve the efficiency and clarity of the approval and notification process.
- The scope excludes purely technical changes related to CRR3 implementation that do not affect model performance.
- The revision covers qualitative criteria, quantitative thresholds, notification modalities, documentation requirements, and IT aspects related to implementation (p. 5-16).
1. Regulatory framework and scope:
- The IRB approach requires prior approval for material changes and notification for other changes.
- The scope excludes purely technical alignments related to application data updates and new originations of exposures already covered by an existing system.
- Changes affecting the methodology for assignment to exposure classes remain within scope.
2. Qualitative materiality criteria:
- Reduction of qualitative criteria considered material, now limited to fundamental changes:
a. Fundamental modification of the default definition (notably for retail exposures under Article 178 CRR).
b. Fundamental redevelopment of the model (change in data typology, switch between discrete and continuous models, major modification of the functional relationship between inputs and outputs, fundamental change in rating order or distribution).
c. Fundamental change in risk quantification (change in estimation method at grade or portfolio level, change in estimation of recession impact).
- Other changes previously material are reclassified as ex ante or ex post notifications.
3. Quantitative materiality criteria:
- Strengthened thresholds to trigger materiality:
- Decrease ≥ 1.5% of consolidated or institution-level RWEA.
- Decrease ≥ 15% of RWEA associated with the rating system scope.
- For scope extensions, a threshold of increase ≥ 15% of RWEA associated with the new scope is introduced.
- Scope reductions do not trigger quantitative thresholds but are covered by other rules (e.g., Article 149 CRR).
4. Notification and approval process:
- Material changes require prior approval.
- Non-material changes require ex ante notification (at least two months before) or ex post notification.
- Some changes related to validation, exposure class assignment methodology, or organizational processes are now under ex post notification to reduce burden.
- Removal of specific derogations for slotting and purchased receivables, which now follow the general framework.
5. Documentation and IT requirements:
- Strengthening of documentation requirements for changes and extensions, including:
- Detailed description and justification of the change.
- Confirmation of internal approval.
- Qualitative and quantitative materiality assessment.
- Impact on exposure distribution and classification.
- Assessment of data representativeness for extensions.
- Technical documentation and validation reports.
- Clarification of IT requirements, notably the ability to calculate capital requirements and report according to COREP from the approval date (p. 6-16, 23-36).
6. Bundling of changes:
- Similar modifications affecting the same system must be aggregated for materiality assessment.
- An exemption allows splitting changes if implementation spans more than one year, with an implementation plan submitted for notification.
- Changes affecting multiple systems must be assessed separately for each system.
7. Alignment with CRR3:
- Removal of references to IRB models for equity exposures and the AMA approach.
- Adaptation of qualitative and quantitative criteria to reflect new regulatory requirements.
- Clarification of scope extension cases and required approvals depending on whether it concerns extensions of an existing system or development of a new system (p. 5-8, 18-22).
8. Impact and objectives:
- Reduction of material changes to lighten supervisory burden.
- Acceleration of change implementation by institutions.
- Maintenance of adequate prudential level through clear and robust criteria.
- Improvement of transparency and quality of information provided to authorities (p. 37-39).
- Established facts:
- The current framework generates a high number of material changes, saturating supervisory resources.
- Approval delays slow down model improvements implementation.
- CRR3 modifies the regulatory framework, requiring RTS updates.
- Assumptions:
- Greater reliance on quantitative thresholds will reduce material changes without compromising supervision.
- Reducing qualitative material criteria to fundamental changes is sufficient to ensure robustness.
- Interpretations:
- Grouping similar changes avoids impact dilution and improves clarity.
- Clear distinction between approval, ex ante, and ex post notification optimizes administrative burden.
- Uncertainties:
- Precise impact on implementation delays remains to be measured after application.
- Adaptation of institutions to new documentation requirements may initially increase workload.
- Lessons learned:
- A more pragmatic and risk-focused framework is needed to balance efficiency and prudence.
- Supervision must have appropriate tools for targeted control of the most significant changes (p. 3-16, 37-39).
- The report concludes on the necessity to adopt the amended RTS to:
- Align the regulatory framework with CRR3.
- Reduce the number of material changes by strengthening quantitative criteria and limiting qualitative criteria to fundamental changes.
- Clarify the scope of changes subject to approval or notification.
- Strengthen documentation requirements to improve transparency and supervision.
- Enable effective bundling of changes to avoid fragmented assessments.
- The report recommends submitting these amended RTS to the European Commission for approval and publication in the Official Journal of the European Union.
- A complementary revision of Delegated Regulation 2022/439 is also planned to strengthen risk-based supervisory tools.
- These measures aim to improve supervisory efficiency, reduce implementation delays for changes, and ensure the robustness of IRB models used in capital requirement calculations (p. 3-4, 37-39).
Synthesis note written from the full document by DataSAI Academy. This note comes from the scientific library of the DataSAI Academy, open to all.