ICAEW.com works better with JavaScript enabled.

IFRS 9: the dog that barked too much?

Author: ICAEW

Published: 28 Sep 2021

When IAS 39 was replaced by IFRS 9 on 1 January 2018, it was perhaps little lamented, writes Steven Hall, Partner, Financial Services Risk Consulting, KPMG and Nicholas Mead, Director, KPMG. In a financial crisis that had manifold causes, some laid at least part of the blame for the 2008 meltdown on the ‘incurred loss’ provision model in IAS39 – inadequate, too little too late, or simply the dog that didn’t bark.

As vaccines offer a glimmer of light at the end of a different crisis, social rather than financial, but with equally profound and long-lasting economic effects, it is worthwhile evaluating how its successor has fared 1. Has IFRS 9 delivered security and stability for the banking system during COVID-19 or instead has it done the opposite, increasing uncertainty and volatility? Put another way – have we instead created the dog that barks too much? And if we have, how are banks taming it?

Expected credit loss design

The design of the expected credit loss (ECL) model at the core of IFRS 9 meant that loan impairments increase as crises emerge (as banks ‘expect’ events to deteriorate from the situation when the lending originated) rather than only when the events have actually happened. 

Critically, IFRS 9 is not meant to increase the overall amount of ECLs that are accounted for, but simply to bring forward when they are recognised. However, as with so many things in life, timing is everything. 

Impairments reduce equity and therefore, all other things being equal, reduce banks’ common equity tier 1 (CET1) capital and therefore their capital ratios.  By accelerating the recognition of impairments, the introduction of IFRS 9 has reduced banks’ CET1 ratios as they go into stressed scenarios, potentially reducing their ability to lend and choking off credit to the wider economy. One of the chief criticisms, and unresolved dilemmas, of IFRS 9 is its pro-cyclicality. 

Industry reaction

Arguably, this was all foreseeable at the outset.  This may account for the degree of trepidation with which banks and regulators ‘turned on’ IFRS 9 on 1 January 2018. Across the industry, banks, policymakers and regulators expected a sharp up-tick in impairments. Many bodies, including the IASB and EBA issued accounting guidance on how to interpret IFRS 9 and how to deal with payment holidays for example to try to reduce the impact of stage 1 to stage 2 movement. 

More significantly, both international and EU regulators pre-empted the introduction of IFRS 9 with a set of ‘transitional provisions’ which allowed banks to add-back the increased impairments to their CET1 capital base and gradually phase them in over time. A steady descent rather than a cliff edge.  

So, the industry was braced, the regulators prepared and then… nothing, or at least, not much. The economic environment in January 2018 was relatively benign and whilst there was an increase in impairments, it was far from the doomsday scenario that some had feared. Many banks across the EU did not even use the transitional reliefs available to them. 

However, this rosy picture belies the problems that banks have been grappling with ever since implementation. These broadly fall into two categories – modelling and forecasting. On the modelling side, the dilemma is that credit models which function very well in business-as-usual economic circumstances, function less well under stressed conditions. Put another way, banks find it difficult to measure ECLs accurately in the circumstances when accurate measurement is most needed – in a crisis. This makes the level of impairments under stress uncertain and prone to significant levels of management judgement. This in turn leads to the second problem – forecasting. The introduction of the ECL impairment model makes it hard to predict impairment levels and therefore makes capital planning both under BAU and stress more challenging. In particular, banks struggle with the so-called ‘volatility’ of impairments as exposures move from stage 1 to stage 2, and the impairment moves from one year to lifetime ECL. 

Covid reaction

These issues, of course, came to a head in March 2020 with the emerging global pandemic. As the virus swept the world and the economic clouds darkened, banks impairments increased sharply (the impairment charge for the large UK banks increased between 3-5 times comparing Q1 2020 vs Q4 2019) and, as the 2018 transitional provisions were reducing, CET1 ratios dropped. Banks were prepared to tighten their belts and reign in lending to protect their capital positions.

This was exactly what regulators and policymakers could not afford to happen – a financing crunch at precisely the time when businesses needed financing to survive. Among the many regulatory interventions in the crisis was the so -called EU COVID ‘quick fix’. This allowed banks to add-back 100% of the increase in stage 1 and stage 2 impairments since 1 January 2020 for a period of 2 years. This helped to solve the immediate problem but raises broader questions. It is, of course, not unusual for regulators to intervene to mitigate unforeseen and unintended consequences of new rules. It is, however, much more uncommon for regulators to intervene to mask and mitigate impacts that were not only foreseeable but intended. IFRS 9 did not break down during COVID-19 but worked, more or less, exactly as it was designed to do. As it turned out we just could not live with the prudential regulatory consequences.

Now it would be wrong to say that IFRS 9 was the only part of the framework that worked as intended but that regulators nevertheless felt they had to ‘fix’ during the crisis. There was also a volatility in risk weighted asset (“RWAs”) for those firms using internal rating models, as the macro-economic climate adversely affected probabilities of default on which these risk weights are based. The difference being that because the issue of RWA volatility lies in the regulatory framework itself, it is much easier for regulators to define an enduring solution rather than a temporary fix. The PRA’s push for banks to use ‘hybrid’ models with less cyclicality for UK mortgage exposures should be seen in exactly this light – a lasting policy change to reduce pro-cyclicality. 

Accounting and capital frameworks

The problem with IFRS 9 is more intractable – how to address the fact that the regulatory capital framework is built on the accounting framework but the two have different objectives and therefore what works for one does not necessarily work for the other. The accounting standard setters, justifiably, argue that IFRS 9 ‘does what it says on the tin’– it gives a better and more realistic picture of potential losses to investors as stresses emerge and therefore does what financial statements are meant to do. Furthermore, it has to cater for all types of entities and cannot just be designed around the specific needs of the banking industry. However, the regulatory authorities (with their prudential banking hat on), justifiably, argue that whilst this may be true, banks are the most significantly impacted and in the short term IFRS 9 exacerbates the impact of stresses on bank’s capital positions, potentially impairs investor confidence and inhibits lending to the wider economy.  

There is no easy answer to this problem. The only solution available to regulators in this circumstance is to mitigate and manage the impacts of IFRS 9 when crises arrive. However, this solution is far from perfect. It creates an unwelcome dislocation, between a bank’s equity position and its regulatory capital position. This has one of two consequences. On the one hand it can make banks’ financial disclosures more opaque and less intelligible to investors therefore counter-acting the desired accounting impact. On the other hand, investors may simply look through the transitional provisions to evaluate the ‘true’ CET1 position therefore undoing the desired regulatory objective. Thus we find ourselves in an unwelcome and ultimately undesirable state of Nash equilibrium between accounting stand setters and regulators. 

And so this allows us to return to the question – has IFRS 9 improved financial stability or just exacerbated economic uncertainty? The answer is, well, neither. Although IFRS 9 has sharply accelerated provisions – the wider impacts on both financial stability and the macro-economy have been masked, or (if we allow ourselves to extend our canine metaphor) muzzled, by transitional regulations. However, this leaves banks with all the complexities of IFRS 9 without delivering the wider macro benefits. In the absence of changes to IFRS 9, there is not an easy solution to this. However, the absence of straightforward answers creates space for more radical ideas. Perhaps it is time to ask whether there should be a more clear and complete separation of the regulatory and accounting frameworks rather than the current ‘halfway house’.  However, with significant regulatory change from Basel 3.1 already around the corner, neither banks nor regulators appear to have the appetite for more wholesale regulatory change. And so, being left with the barking dog, banks and regulators need to continue in their efforts to live with it. 

1 Note as part of its standard practice the IASB is performing a post-implementation review of IFRS9 and the impairment project is due to start in Q4 2021.