EBA欧洲银行-2014-09-11-BSG-Opinion-on-EBA-CP-2014-08_10页_375kb

EBA Banking Stakeholder Group Consultation Summary on EBA/CP/2014/08

Core Content

The EBA Banking Stakeholder Group (BSG) has provided detailed comments on the consultation paper EBA/CP/2014/08, which outlines draft Regulatory Technical Standards (RTS) on assessment methodologies for the Advanced Measurement Approach (AMA) for operational risk under Regulation (EU) No 575/2013 (CRR). The BSG evaluates the proposed standards and raises concerns regarding their clarity, implementation, and alignment with existing regulatory frameworks.

Main Views and Key Points

General Comments

• The BSG appreciates the EBA's effort to consolidate the requirements for structuring an AMA for operational risk into a single document, which is beneficial during periods of multiple regulatory proposals.
• However, the document lacks clarity on how new specifications will be phased in and applied, particularly for institutions with pre-existing models accepted by supervisory authorities.
• The BSG emphasizes the importance of distinguishing between models validated before and after the RTS enters into force, to avoid retrospective application.

Article 1(2): Operational Risk Events Related to Model Risk

• Model risk is already included in operational risk under the "product error" category.
• The BSG requests clarification that the current inclusion of model risk is not to be extended and that there is no overlap with EBA/CP/2014/14 on model risk.

Article 2(12): Operational Risk Events Related to Legal Risk

• The definition of legal risk is welcomed, but the BSG asks for clarification that legal risk does not include actions taken by the institution to enforce claims.
• There is a need to clarify the definition of legal risk to avoid including "genuine" goodwill cases that do not arise from operational risk events.

Article 4(2)(a): Operational Risk Events Related to Legal Risk

• The BSG supports the inclusion of out-of-court settlement costs in operational risk.
• However, they caution against including "genuine" goodwill cases that do not stem from an operational risk event.

Article 5(3)(g): Operational Risk Events Related to Market Risk

• The BSG argues that exceeding limits should not be considered an operational risk event unless it results in a net loss.
• They suggest that only intentional unauthorised excess of limits should be considered.

Article 6: Fraud Events in the Credit Area

• Fraud events should be recorded as operational risk losses and supported by own funds.
• The BSG recommends a flat-rate deduction in credit risk to avoid double counting.
• They highlight the need for a threshold to exclude small-scale retail fraud events from loss recording.
• A 5-year transitional period is suggested instead of the proposed 2-year period to allow for proper implementation.

Article 7(1&2): Scope of Operational Risk Loss

• The BSG identifies a contradiction in the inclusion of opportunity costs in the operational risk database.
• They request the deletion of points (d) and (e) in Article 7(1) as they are not directly related to P&L.
• The BSG suggests focusing only on material and significant events, and proposes a threshold for near-misses and gains.

Article 8(1)(d): Recorded Loss Amount

• The BSG questions the use of the outstanding credit amount as the basis for recording operational risk losses in fraud cases.
• They suggest using the reserve amount instead to more accurately reflect the true loss.

Article 8(3): Timing Losses

• The BSG supports the recording of timing losses involving third parties.
• They advocate for the inclusion of excessive income from previous periods as loss mitigation.

Article 11(2)(f): Operational Risk Governance

• The BSG calls for a concrete definition of "operational risk governance" and further explanations on how the effectiveness of governance should be evaluated.

Article 16(4): Internal Loss Data

• The BSG clarifies that internal loss data should not be transferred individually to the accounting system.
• They request clarification that this requirement will not be applied retrospectively.

Article 17(2): External Loss Data

• The BSG asks for clarification on the implementation of data filtering for external loss data and whether it is essential.

Article 21(3): Building the Calculation Data Set

• The BSG supports the extension of the observation period for small data sets.
• They request clarification on how the observation period can be reduced once the data set is deemed sufficient.

Article 23(3): Identification of the Probability Distribution

• The BSG requests the deletion of this section due to its overly prescriptive nature and lack of alignment with data-driven approaches.

Article 24(4): Determination of Aggregated Loss Distributions

• The BSG suggests that right-truncation of loss distributions is acceptable if economically justified, to ensure robustness.

Article 26(1&3): Dependence

• The BSG requests clarification on how empirical data analysis and expert judgment will be combined.
• They question the exclusion of Gaussian or Normal-like copulas and suggest that the dependence structure should be defined more precisely.

Article 33(b): Allocation Processes

• The BSG acknowledges that the "home" regulator's assessment may prevail in international cases.

Article 45(2)(b): Audit and Internal Validation Reviews

• The BSG proposes that reviews should be linked to the severity matrix of the model change policy, rather than being mandatory annually.

Conclusion

The BSG provides a comprehensive critique of the consultation paper, emphasizing the need for clarity, consistency, and a balanced transitional period for implementation. They advocate for a more flexible and data-driven approach to operational risk assessment, while ensuring that existing models and practices are not unduly affected.