Financial Services Advisory

Summary Report on the 2022 Credit Risk Benchmarking Exercise

ContributorName(contributor, true)
By:
insight featured image
This report details the findings of the supervisory benchmarking exercise conducted in 2022 in line with Article 78 of the Capital Requirements Directive (CRD).
Contents

Subscribe to our mailing list

Receive the latest insights, news and more direct to your inbox.

It also covers the regulatory and implementing technical standards (RTS and ITS) that establish the extent, methods, and categories for evaluating internal models for Credit Risk and Market Risk.
The benchmarking exercise is an essential supervisory tool to enhance the quality of internal models and to monitor the variability of RWAs for institutions applying the IRB approach.

 

Market Risk Scope

  • The primary objective of the exercise is to assess the level of variability observed in risk-weighted assets (RWA) for market risk produced by banks’ internal models.
  • A stress test was performed on a sample of 41 European banks from 13 jurisdictions. The relevant institutions submitted data for 81 instruments recombined into 62 market portfolios across all major asset classes.
  • This analytical part of the exercise provided the competent authorities (CAs) with a detailed list of outliers to be reviewed.

Main findings

  • The report measures variability in terms of the interquartile dispersion (IQD) and the coefficient of variation (CV) observed within each benchmark portfolio.
  • The 2022 results show a small reduction in the dispersion of the initial market valuation (IMV) versus the 2021 exercise regarding the FX asset class.
  • Regarding the single risk measures - across all asset classes (except for credit spread), the overall variability for value at risk (VaR) is lower than the observed variability for stressed VaR.
  • For the larger banks with significant trading activities, the benchmarking portfolios are generally relevant to their actual trading book.

Competent Authority's (CA's) assessments based on supervisory benchmarks

The CA's reviews confirmed the existence of some areas that require follow-up actions on the part of specific institutions whose internal models were labelled as outliers in this benchmarking exercise.

Therefore, CAs must pay close attention to the minority of cases in which the over/underestimations are highlighted as outliers, to closely monitor these institutions and put in place additional efforts to reduce these gaps in future exercises.

Stress Value at Risk distribution

Stress Value at Risk distribution graph

[Source: 2022 Market Risk Benchmarking Exercise]

Overview of the results reflect portfolios 1 to 59. Broken down by portfolio, the figures show the average sVaR distribution over the 10-day submission period for each bank. Dispersion of sVaR stands at 28%, a slightly higher level of dispersion than VaR mainly due to FX an IR portfolios.

Read the full publication

Credit Risk Scope

  • Institutions which apply the IRB approach calculate their own funds requirements based on a set of parameters by which they partially (under FIRB) or completely (under AIRB) estimate themselves.
  • Article 78 of the CRD provides the statutory basis for the monitoring and assessment of RWAs which determine the own funds requirements, that result from the application of the institutions’ own estimates for pre‐defined benchmarking portfolios.
  • The annual benchmarking exercise aims to monitor the variability of the RWAs for institutions applying the IRB.

Main highlights

  • Support actions taken at EU and Member States’ level in response to the Covid-19 pandemic played an important role in explaining the trend of decreasing default rates and average PD estimates.
  • The observed decreasing average PD is likely triggered by recalibration procedures, supporting measures such as moratoria, and different origination policies.
  • Basel III (Basel IV) reforms are expected to further reduce the RWA variability across institutions that apply the IRB approach. In addition, the harmonisation of IRB models across European Institutions in accordance with the IRB roadmap shall also reduce unwarranted RWA variability.
  • There were no anticipations or signals of the current economic outlook and the energy crisis as at 31st of December 2021.

Next steps for supervisors and regulators

  • Supervisors and institutions risk controllers should be vigilant to ensure that the long‐run average default rates used for (re‐)calibration of PD estimates reflect the likely range of variability of default rates relevant to a considered type of exposures.
  • EBA has put forward a regulatory review of the IRB approach by setting out and completing several guidelines and technical standards, which are aimed at limiting unjustified variability by harmonizing practice.

Read the full publication

Methodology guide to the report

Publication provides a detailed overview of methodology and technical specifications followed by the EBA in the credit risk review.

Read the publication

 

State of compliance with the GL on PD and LGD for material IRB exposure

State of compliance with the GL on PD and LGD for material IRB exposure graph

[Source: EBA/REP/2023/09]

Analysing the development of average PDs for benchmarking portfolios as of 31.12.2020 and 31.12.2021 The impact of the IRB roadmap is yet to be fully realised in the Benchmarking exercise. As shown in the below figure, the supervisory process (model approval process) is still a work in progress for the majority of models within the exercise. It can be surmised that the impact of the IRB roadmap will be further accounted for in future benchmarking exercises.

Credit Risk Mitigation Impact 2020 - 2021

Credit Risk Mitigation Impact 2020 - 2021 graph

[Source: 2021 Credit Risk Benchmarking exercise]

The increased outflows observed in CORP, COSP, SMEC and SMOT as of December 2021 indicate an increased use of RW substitution in these portfolios compared to December 2020, indicating an increasing trend in the use of COVID-related measures throughout 2021.