EBA欧洲银行-EBA-Report-on-Unsolicited-Credit-Assessments-28Article-138-CRR29_40页_1mb

Summary of the EBA Report on Unsolicited Credit Assessments (Article 138 CRR)

Core Content and Purpose

This report outlines the European Banking Authority's (EBA) analysis and findings regarding the quality of unsolicited credit assessments by External Credit Assessment Institutions (ECAs) under Article 138 of the Capital Requirements Regulation (CRR). The EBA confirmed that for certain ECAs, unsolicited credit assessments do not differ in quality from their solicited counterparts. This decision supports the use of unsolicited ratings in capital requirements calculations, provided they meet the quality criteria. The report serves as a supplement to the EBA's Decision and aims to provide transparent information to stakeholders about the methodology and results of the assessment.

Key Definitions and Scope

• ECAI (External Credit Assessment Institution): Defined as any credit rating agency (CRA) registered or certified under Regulation (EC) No 1060/2009, or a central bank issuing credit ratings exempt from that regulation.
• Unsolicited Credit Assessment: While not defined in the CRR, guidance from the CRA Regulation is applied. The EBA used the definitions provided by ECAs during the assessment.
• Scope: The report focuses on comparing the quality of unsolicited and solicited credit assessments within the same ECAI, not across different ECAs. It also addresses the issue of potential pressure on rated entities from unsolicited ratings.

Main Features of the Assessment

The EBA conducted a comprehensive assessment of unsolicited credit assessments, which included both quantitative and qualitative analyses:

Quantitative Analysis

• Ex-ante Distribution: Compares the distribution of solicited and unsolicited ratings across rating categories.
• Ex-ante Dynamics: Analyzes the time evolution of ratings following a change in solicitation type.
• Ex-post Analysis: Uses the AUROC statistic to evaluate the discriminatory power of the rating systems for solicited and unsolicited ratings.

The EBA relied on the CEREP database for its quantitative analysis, which contains data submitted by ECAs for regulatory purposes. However, the data availability was limited, especially for unsolicited ratings, which affected the depth and reliability of the analysis.

Qualitative Analysis

• Assignment Policy: Examines how ECAs assign and review unsolicited ratings.
• Rating Methodologies: Compares the methodologies used for solicited and unsolicited ratings.
• Data Availability: Assesses how ECAs manage the lack of data for unsolicited ratings.
• Pressure on Rated Entities: Evaluates whether ECAs use unsolicited ratings to exert pressure on rated entities to order services.

Key Findings

Quantitative Findings

• Ex-ante Distribution: For most ECAs, the distribution of solicited and unsolicited ratings did not show significant differences. However, in some cases, unsolicited ratings were slightly more conservative.
• Ex-ante Dynamics: Changes in rating categories following shifts in solicitation type were rare, suggesting no material concerns about rating behavior.
• Ex-post Analysis: The AUROC analysis did not indicate significant differences in discriminatory power between solicited and unsolicited ratings. This suggests that the quality of unsolicited ratings is comparable to solicited ones.

Qualitative Findings

• No material differences were found in the policies and methodologies for assigning and reviewing solicited and unsolicited ratings.
• Data availability for unsolicited ratings was generally lower, but ECAs had procedures in place to ensure no underestimation of risk or difference in quality.
• ECAs implemented measures to prevent unsolicited ratings from exerting pressure on rated entities to order additional services.

ECAIs Covered in the Assessment

• Banque de France and The Economist Intelligence Unit Ltd: These ECAs exclusively assigned unsolicited ratings, and the EBA confirmed their use due to the absence of an 'issuer-pays' model and alignment with Recital 98 of the CRR.
• Other ECAs: For the remaining ECAs that assigned both solicited and unsolicited ratings, the EBA performed assessments based on available data, which was limited and affected the reliability of results.

Limitations and Future Actions

• Data Availability: The limited data in CEREP, especially for unsolicited ratings, hindered the ability to perform granular and statistically sound analyses.
• Publication Consent: Most ECAs provided consent for their data to be published, except Euler Hermes Rating GmbH, whose data was not included in the report.
• Ongoing Monitoring: The EBA will continue to monitor the performance of unsolicited ratings and the alignment of ECAs with the Q&A issued by ESMA. If necessary, it may revisit its conclusions or methodology.

Conclusion

The EBA concluded that there is no evidence of a material difference in quality between solicited and unsolicited credit assessments for the ECAs under review. Therefore, it confirmed that unsolicited ratings can be used in capital requirements calculations under Article 138 of the CRR. The EBA emphasized the importance of ongoing monitoring and potential future reviews of the methodology and findings.