IFRS 9: Issues for preparers
The implementation of new reporting standard IFRS 9 from 1 January 2018 is a key priority for the banking industry. In July 2017, ICAEW's Financial Services Faculty brought together key stakeholders from the investor and analyst communities so that they might understand the respective challenges faced by banks in preparing IFRS 9 expected credit loss provisions. Here we report on the discussions around the issues for preparers.
Preparers made it clear they are still working hard in 2017 to finalise their internal models in order to generate the IFRS 9 numbers. This is proving difficult as existing expected loss models were designed to produce numbers for regulatory and internal reporting purposes. The numbers required for IFRS 9 are subtly, but crucially different.
This means that while banks feel they do understand their expected credit losses, the standard’s requirements mean more work. Most banks are still working very hard to get the models in place to generate the numbers necessary. Due to the vast number of product lines, even small changes add to effort, time and costs.
The standard, and the increased use of models it promotes, is in direct contrast to Basel and their approach to bank regulatory capital. Banking regulators are increasingly trying to limit the use of models and standardise approaches. Further, US standards (Current Expected Credit Loss model - CECL) are different in that they book lifetime losses from day one.
Under IFRS 9 there is a smaller provisions for new assets or those that have not seen a significant increase in credit risk. For these ‘Stage 1 assets’ banks need to hold provisions for 12 months of expected losses. This nuance may make it appear as if UK and EU banks are under-providing and attract negative press comment.
Similarly, the standard also promotes movements from Stage 1 (new loans) to Stage 2 (significant increase in credit risk), and back again, which will prompt a change in behaviour from the banks. They will intervene earlier to avoid cliff-edge effects to their P&L as provisions increase materially as assets move from Stage 1 to 2 (provisions change from 12 month expected loss to lifetime expected loss).
Worries over consistency
The analyst community voiced concerns about the time it would take to bed the standard down, recognised at around two years. There will be a lack of consistency during this time which will make it difficult to make sound investment decisions. It was noted that it may take a credit cycle or a recession for the standard to be fully understood, tweaks to be ironed out and consistency to occur. In the first round preparer bandwidth had been focused on developing the first numbers and not on fine-tuning disclosures.
There is an expectation that there may be a ‘pull to par’ – meaning that after a few years the forward looking element will find a middle ground. However, it was noted that this had not been seen in the case of IAS 39 so market discipline, on its own, may not be relied on to deliver optimal outcomes and disclosures.
Challenge of pro-cyclical attitudes
Banking is a very cyclical industry, and this standard may exaggerate results in good and bad periods in a pro-cyclical manner. Expected loss necessarily means taking a forward-looking approach but human nature is to be over-optimistic in good times and overly-pessimistic in bad times. We expect this could lead to increased volatility.
The intent of the standard is to drive management to use all information to build their unbiased and probability-weighted estimate of expected credit losses by evaluating a range of possible outcomes. However, the number of judgments increases exponentially as a consequence. It was noted that auditors should use their extended audit report to comment on these judgments and flag to investors, issues they perceive as material.
It is not straight forward for management to simply explain their best estimate of expected loss because of the way the standard has been written. As above, it requires the estimate of expected credit losses to reflect an ‘unbiased and probability-weighted’ view. By way of illustration, this might mean a bank disclosing that a dice would land on 3.5 (the unbiased probability weighted outcome) even though this is not what management ‘expects’. Investors expressed concern that the expected loss number would not reflect what management thought the losses would be in a base case.
Negative impact on consumers
The standard would increase provisions very quickly for banks with a growing balance sheet. They would start with low provisions from Stage 1 assets, but as the book seasons they would move to Stage 2 and provisions would increase rapidly. This poses a threat to challenger banks, who are generally smaller but who are seeking to increase their market share. Ultimately the consumer would suffer as a consequence through elevated pricing and reduced competition.