Sanjin Bogdan, Head of IFRS at Aryza
Last month, The European Banking Authority (EBA) released its Risk Assessment Report on the implementation of the International Financial Reporting Standard (IFRS 9) within EU institutions. This report, with a specific focus on high default portfolios (HDPs), aims to drive enhancements in the Expected Credit Loss (ECL) model practices across EU institutions. The primary goal is to enhance transparency by highlighting key areas of concern identified by the EBA.
Since the initial implementation of IFRS 9, institutions have made notable strides in incorporating their ECL impairment models, despite operating in a challenging environment. Benchmarking analyses reveal a diversity of practices, attributable to the principle-based nature of the standard. This diversity may, in part, account for the observed variability in the final ECL figures for HDPs among different institutions. Certain practices continue to be a source of prudential concern, prompting the expectation that EU institutions will promptly address these issues, meaning it is imperative for financial institutions to not only address these challenges but also proactively implement robust solutions.
Challenges in SICR Assessment
Despite the persistent macroeconomic uncertainties, there is still a continued lack of collective Significant Increase in Credit Risk (SICR) assessment, violating IFRS 9 requirements. Prudential concerns remain on practices employed to determine SICR thresholds that are not always in line with the main objectives of the impairment model of IFRS 9, in particular the concept of ‘significance’ as envisaged in the Standard.
Financial institutions must adopt modelling approaches, enabling fast deployment and update of Probability of Default (PD) models. Solutions with automatised model updates with a flexible SICR rule setup can facilitate an effective testing and analysis process, ensuring adherence to significance criteria and adequate SICR can be deployed. Simulation environments for historical portfolio testing further strengthen the SICR back testing process, especially in the domain of the definition of historical ratings, where due to potential discrepancy between historical and current ratings, preferable options for grouping should be based on default characteristics of the observed clients. An advanced option would be the application of current model methods to historic portfolios to reach consistency and comparability of historic and current ratings, where again robust IT solutions for such exercises are essential.
Addressing Variability in ECL Outcomes
Delays in transferring to Stage 2 contribute to increased variability in Expected Credit Loss (ECL) outcomes. Overreliance on specific triggers for staging assessment poses a risk, with a threefold increase in the annualised lifetime PD suggested as a backstop indicator.
To address this, financial institutions should explore and implement software allowing specific rule setups for different portfolio segments, enabling tailored treatment for stage allocation, model coverage, and provision processes. Back testing of Stage 2 triggers and detailed analysis of hard factors such as 30 DPD vs SICR, alongside other factors, can be extremely beneficial to detect the sensitivity of SICR and any other stage allocation mechanism. A solution which can enable such historic observations and simulations would be beneficial, especially in the creation of audit evidence and confirmation of the stage allocation process while avoiding overreliance on single-stage triggers.
Effective Governance of Overlays
The extensive use of overlays with varied calibration practices highlights the need for institutions to follow sound methodological approaches and governance. IFRS 9 software should support overlay deployment at transactional and portfolio levels, with tracking mechanisms and approval processes for overlays ensuring governance and methodological soundness. Having an overview of deployed management overlays in a systematic manner can provide insights into underlying risk developments over time even if overlays were used as risk mitigation mechanisms. Eventually banks via the above-mentioned approach would gain insights into the effectiveness of used overlays and could estimate the potential portfolio effects of their termination.
Tailoring IFRS 9 Models
Some institutions still lack targeted IFRS 9 models for specific portfolios, resulting in a uniform application of loan loss provisions by applying the same level of provisions as those used for other portfolios where IFRS 9 models have been applied. To address this, institutions should adopt software supporting the development of models for any segment, ensuring tracking of portfolio performance and offering a back test function of selected exposures to validate the model replacement process of the bank. In the case of a low default portfolio grouping, it can be supported by historical observations in performance such as observations of the Stage 2 dynamics of the portfolio and any other common behaviours. The essential element, in this case, is the historic simulation environment which enables cross-segment observations.
Incorporating Climate and Sustainability Risks
Few institutions incorporate climate and sustainability-related risks into their ECL models, even though overlays have become an integral part of the framework. This consideration reinforces the need for institutions to follow a structured approach when overlays are used for loss provisioning purposes.
Software integration of ESG (Environmental, Social, and Governance) characteristics allows specific treatment within provision processes, being that specific models are needed for such exposure, unique staging, or that management overlays have been deployed. A solution which can develop and deploy models including the ESG risk characteristics would be the first step towards ESG integration into the IFRS 9 process. In the first instance, such ESG characteristics should be used to monitor the performance of underlying exposures vs other portfolios and eventually be used as a correction factor to model values.
Embracing Forward-Looking Information
Divergent practices impact the modest effect of forward-looking information, hindering the reflection of IFRS 9 figures’ point-in-time and forward-looking nature. To address this, financial institutions should deploy a modelling solution which enables rapid macroeconomic scenario updates and tests the effects on provision and staging, ensuring automated model updates and life-long model vector processes to test the effects of changes efficiently.
Enhancing Back Testing Effectiveness
Limited effective use of back testing for the periodic review of the IFRS 9 model raises supervisory concerns due to the testing results not producing actions and model improvements. Financial institutions should invest in software enabling detailed back testing processes to track a model’s performance on a transactional level over time. This facilitates historical performance analysis, recalculations using alternative models, and flexible deployment of back testing processes. An essential element should be the continuity of back testing and timely follow-ups to identified shortcomings, instead of annual efforts by institutions.
The lack of follow-up actions on back testing results raises prudential concerns. Implementing software automating processes to escalate identified deviations and automatically prompt feedback or updates to deployed scenarios, can ensure adequacy in reported ECL figures.
The evolving financial landscape demands a strategic response to the challenges highlighted in the EBA’s Risk Assessment Report. By embracing innovative solutions and leveraging advanced software capabilities, financial institutions can not only navigate these challenges with IFRS 9 but also position themselves for sustained success in an ever-changing environment. Going forward, legacy IFRS 9 systems solely based on risk provision calculations are proving not to be up to the task or meeting the expectations of regulators. What is becoming evident is that IFRS 9 solutions should be comprehensive systems which can enable sophisticated analytical work, modelling functions and historical simulations. Such systems not only provide additional efficiency within the cost of risk system but are essential to address internal and regulatory expectations when it comes to prudent risk management.