IFRS 9 Implementation by EU institutions

The European Banking Authority (EBA) has published its second monitoring report on the application of IFRS 9 by institutions in the European Union (EU). This report focuses on High Default Portfolios (HDP) and aims to encourage further improvements in expected credit loss (ECL) modeling practices among EU institutions, ensuring transparency in the key areas of concern previously identified by the EBA.

Executive summary

The EBA has published its second monitoring report on the application of IFRS 9 by EU institutions. With this publication, the EBA aims to promote further improvements in ECL modeling practices, while ensuring transparency in the key areas of concern previously identified by the EBA. These areas are the approach to assessing significant increases in credit risk, expected credit loss or ECL models, the definition of probability of default (PD), the inclusion of forward-looking information and backtesting.

Main content

• Significant increase in credit risk (SICR) assessment approaches. Staging approaches that are not well designed and implemented by institutions may result in a delayed recognition of significant increases in credit risk and thus may not ensure that ECL is measured with the appropriate time horizon.
• ECL Models. Most institutions in the sample have implemented a PD, loss given default (LGD) or exposure at default (EAD) approach to determine ECL. Post-model adjustments or overlays are generally used by institutions to account for emerging risk factors that are not captured by models in a timely manner and are often considered temporary in nature.
• IFRS 9 PD variability and robustness. Benchmarking data indicate some variability in IFRS 9 12-month PDs across institutions. There are differences in the use of Internal Ratings-based (IRB) models for IFRS 9 estimates in the definition of default, in risk differentiation and in risk quantification.
• Incorporation of forward-looking information (FLI). Relevant areas of variability have been observed in different parts of the FLI incorporation process, such as the definition of the relevant macroeconomic scenarios or the approaches used to incorporate the effects of macroeconomic projections at the risk parameter level and the methods envisaged to account for non-linearity effects in credit loss estimates.
• Backtesting. Backtesting is particularly important for ECL projections given that the use of IFRS 9 models does not require prior supervisory approval. Backtesting ECL lifetime estimates requires not only robust methodologies, tools, policies and effective processes, but also sufficient data sets and actual observations of realized values.

Download the technical note on IFRS 9 Implementation by EU institutions.