IFRS 9 – What did banks disclose?
Laure Guegan, Partner, Ernst & Young et Associés,
Tony Clifford, Partner, Ernst & Young LLP and
Fabio Fabiani, Partner, Ernst & Young LLP
With most banks having published their first set of IFRS 9 financial statements, Laure, Tony and Fabio provided feedback on the key initial findings arising from the EY IFRS 9 disclosure benchmarking assessment. The assessment covered European and UK banks’ expected credit loss (ECL) related disclosures in their most recent annual financial statements. Laure focused on the major findings for European banks, while Tony focused on UK banks. Fabio provided additional input from a UK Taskforce on Expected Credit Losses Disclosures (DECL) perspective.
The key message was that there appears to be significant diversity in how banks have practically applied and disclosed the impact of IFRS 9. This could be due to diversity in the nature of exposures or the approaches and policies followed. It is also possible that some banks were precluded from making some of their intended disclosures due to data availability and quality challenges.
The challenge faced by many users of banks’ financial statements is that the current level of disclosure surrounding key policies and judgements and the level of granularity of ECL related disclosures is, in many cases, not detailed enough to enable users to understand what these differences are and why they exist. This makes comparison within a bank over time and comparison between banks challenging. The key discussion points included:
Disaggregation of stage one and two exposures:
Outside the UK, most banks did not show allocations of gross carrying values between stage one and stage two at a detailed level. Only five banks outside the UK provided separate information for retail and non-retail exposures. The allocation of ECL amounts is often shown at a less detailed level, which does not allow calculation of coverage ratios at a disaggregated level
Relative credit quality of stage one and two assets:
Despite all 18 banks providing analyses of their stage one and two exposures by Probability of Default (PD) or credit rating bands, these analyses were difficult to compare between banks. This was because banks used different numbers of bands and sometimes did not disclose the basis for mapping the rating bands back to probabilities of default. Relative credit quality data would also be more meaningful if disaggregated at product level and accompanied by ECLs or coverage ratios.
ECL reconciliations and the impact of changes in gross carrying value (GCV) on ECLs:
While banks provided ECL reconciliations per class and stage of financial asset, there was limited disclosure of ECL data per product or business area. Also, the main movements were often not commented on. Additionally, many of the banks sampled (eight out of eighteen) did not provide gross carrying value reconciliations for stage one and two assets. Only six banks (including four UK banks) have provided the full detail of gross transfers between stages for both gross carrying amounts and ECL. These movements are key not only to understand the trends, but also to get a sense of how the bank has implemented Significant Increase in Credit Risk (SICR) triggers.
In 2018, there was a general trend of increasing loan book sizes and decreasing coverage ratios. There was also diversity in coverage ratios between banks. There was a significant decrease in the highest coverage ratios (Italy, Spain and France) due to the deleveraging of non-performing loans through write-offs or disposals. There was also a slight overall decrease in stage two proportions due to a variety of reasons, including disposals, refinement of SICR triggers, repayments and new business.
It is crucial to clearly explain and define ECL policy choices. Differences in policy choices present challenges to users seeking comparability if they are not clearly explained. In the case of write-offs, although they are key to understanding ECL movements, many banks provided generic policies without explaining how the definition was applied to different products and regions or whether there were any changes on transition or subsequently. From an analysis of the data, it appears that banks in different jurisdictions have significantly different write-off practices. Without further detail, comparing coverage ratios and ECL movements is challenging.
Only one of the banks provided a sensitivity to changes in SICR definitions, despite lingering concern in the market that IFRS 9’s distinction between 12-month and lifetime ECL’s could result in the impairment number being more volatile than under IAS 39. Two UK banks decomposed stage two exposures by type of SICR trigger. Some banks also disclosed delinquency information for stage two, showing higher levels of days past due (DPD) for retail exposures and potentially some divergence in the reliance on past due data for staging.
Differences in the UK:
UK banks tended to provide more detail than their European counterparts. The enhanced level of detail in the UK has, in part, been driven by DECL and regulatory pressure from the Prudential Regulatory Authority. Despite this, challenges remain for users to understand why differences or movements occur as ECL data was not always presented consistently or accompanied by clear qualitative analysis to explain what caused the moves. The presenters mentioned that this is partly due to management overlays which are used to differing degrees. It can be challenging to disaggregate overlays at a product or geographic region level if they were not calculated with granular disclosures in mind.
In conclusion, users of banks’ ECL disclosures have expressed concern about a lack of comparability. This is, perhaps, normal for first-time adoption of a standard, in that disclosures become more refined and standardized over time as stakeholders become clearer in their expectations. However, the key takeaway is that the current disclosures are not final products and that banks’ finance functions need to retain budget allocations in order to refine the disclosures where necessary.
Sue Lloyd reminded participants that the new disclosures accompanying IFRS 9 (in IFRS 7) are objective based. Instead of being prescriptive about detailed requirements, they lay out an objective (that users can understand the effect of credit risk on the amount, timing and uncertainty of future cash flows) and a non-exhaustive list of disclosures which may help to achieve this objective. Users need to exercise caution to avoid a situation where they comply with individual disclosure requirements, but fail to achieve the overall objective. She mentioned that IFRS 9 is essentially a trial standard in this regard and that future standard setting is likely to take into account how these disclosures are applied.