I had been intending to write a post on the potential time bomb for bank capital embedded in IFSR9 but Adrian Docherty has saved me the trouble. He recently released an update on IFRS9 and CECL titled Much Ado About Nothing or Après Moi. Le Deluge?
This post is fairly technical so feel free to stop here if you are not a bank capital nerd. However, if you happen to read someone saying that IFRS 9 solves one of the big problems encountered by banks during the GFC then be very sceptical. Adrian (and I) believe that is very far from the truth. For those not discouraged by the technical warning, please read on.
The short version of Adrian’s note is:
- The one-off transition impact of the new standard is immaterial and the market has largely ignored it
- Market apathy will persist until stressed provisions are observed
- The dangers of ECL provisioning (procyclical volatility, complexity and subjectivity) have been confirmed by the authorities …
- … but criticism of IFRS 9 is politically incorrect since the “correct” narrative is that earlier loan loss provisioning fulfils the G20 mandate to address the problem encountered during the GFC
- Regulatory adaption has been limited to transition rules, which are not a solution. We need a fundamentally revised Basel regime – “Basel V” – in which lifetime ECL provisions somehow offset regulatory capital requirements.
Adrian quotes at length from Bank of England (BoE) commentary on IFRS 9. He notes that their policy intention is that the loss absorbing capacity of the banking system is not impacted by the change in accounting standards but he takes issue with the way that they have chosen to implement this policy approach. He also calls out the problem with the BoE instruction that banks should assume “perfect foresight” in their stress test calculations.
Adrian also offers a very useful deconstruction of what the European Systemic Risk Board had to say in a report they published in July 2017 . He has created a table in which he sets out what the report says on one column and what they mean in another (see page 8 of Adrian’s note).
This extract from Adrian’s note calls into question whether the solution developed is actually what the G20 asked for …
“In official documents, the authorities still cling to the assertion that ECL provisioning is good for financial stability “if soundly implemented” or “if properly applied”. They claim that the new standard “means that provisions for potential credit losses will be made in a timely way”. But what they want is contrarian, anti-cyclical ECL provisioning. This is simply not possible, in part because of human psychology but, more importantly, because the standard requires justifiable projections based on objective, consensual evidence.
Surely the authorities know they are wrong? Their arguments don’t stack up.
They hide behind repeated statements that the G20 instructed them to deliver ECL provisioning, whereas a re-read of the actual instructions clearly shows that a procyclical, subjective and complex regime was not what was asked for.
It just doesn’t add up.”
There is of course no going back at this point, so Adrian (rightly I think) argues that the solution lies in a change to banking regulation to make Basel compatible with ECL provisioning. I will quote Adrian at length here
“So the real target is to change banking regulation, to make Basel compatible with ECL provisioning. Doing this properly would constitute a genuine “Basel V”. Yes, the markets would still need to grapple with complex and misleading IFRS 9 numbers to assess performance. But if the solvency calculation could somehow adjust properly for ECL provisions, then solvency would be stronger and less volatile.
And, in an existential way, solvency is what really matters – it’s the sina qua non of a bank. Regulatory solvency drives the ability of a bank to grow the business and distribute capital. Accounting profit matters less than the generation of genuinely surplus solvency capital resources.
Basel V should remove or resolve the double count between lifetime ECL provisions and one-year unexpected loss (UL) capital resources. There are many different ways of doing this, for example:
A. Treat “excess provisions” (the difference between one-year ECL and lifetime ECL for Stage 2 loans) as CET1
B. Incorporate expected future margin as a positive asset, offsetting the impact of expected future credit losses
C. Reduce capital requirements by the amount of “excess provisions” (again, the difference between one-year ECL and lifetime ECL for Stage 2 loans) maybe with a floor at zero
D. Reduce minimum regulatory solvency ratios for banks with ECL provisioning (say, replacing the Basel 8% minimum capital ratio requirement to 4%)
All of these seem unpalatable at first sight! To get the right answer, there is a need to conduct a fundamental rethink. Sadly, there is no evidence that this process has started. The last time that there was good thinking on the nature of capital from Basel was some 17 years ago. It’s worth re-reading old papers to remind oneself of the interaction between expected loss, unexpected loss and income. The Basel capital construct needs to be rebuilt to take into account the drastically different meaning of the new, post-IFRS 9 accounting equity number.”
Hopefully this post will encourage you to read Adrian’s note and to recognise that IFRS 9 is not the cycle mitigating saviour of banking it is represented to be. The core problem is not so much with IFRS9 itself (though its complexity and subjectivity are issues) but more that bank capital requirements are not constructed in a way that compensates for the inherent cyclicality of the banking industry. The ideas that Adrian has listed above are potentially part of the solution as is revisiting the way that the Counter cyclical Capital Buffer is intended to operate.
From the Outside
From the outside 
3 thoughts on “The answer is more loan loss provisions, what was the question?”