But wait, the Fed, the BoE & the ECB have all published on this (or the US equivalent). And they are all trying to find a way to tweak a rule which is not really in their remit (it’s an accounting rule.) So maybe there *is* actually something important?
I’ll explain why first with the theory and then a STUNNING example. Bear with me through some accounting stuff, before we get to the juicy numbers.
Under IFRS 9, there are three types of loans: Stage 1 (all good), stage 3 (defaulted) and the tricky Stage 2, which does not depend on the credit quality of a loan but whether that credit quality has deteriorated.
You can have a Stage 1 loan rated CCC if you originated it at CCC and a A-rated loan in stage 2 if you originated it at AA. I know, it is very weird, but it’s how it works.
The HUGE difference between S2 and S1, is that in S1, you book a provision for the Expected Loss (EL) over a year whereas in S2 you take EL over the loan’s *entire lifetime*. The difference can be huge for some long-dated loans, like…. Mortgages!
When the macro becomes shitty, IFRS 9 has a double blade effect:
1) A-year EL increase, but that was already taken into account in bank regulations so it’s not a significant change.
2) a bunch of loans go to Stage 2 and therefore losses have to be calculated over a much longer horizon (and for a low probability of default (PD) a 5-year PD looks a lot like 5 times the 1y PD… Ouch.
If you want to model this, it’s really complicated: you need credit rating transition matrices, model the portfolio over the next year with a default intensity and then use a life-time transition matrix for the share of loans that went to S2,
Complicated. That’s what I’m trying to do, as precisely as I can, for EU banks, but ofc I will keep the numbers for myself, but I’ll still give you an example. This time it’s not our good friend Deutsche Bank, but Lloyds, a strong UK bank. And honestly I think you'll be stunned.
Meet Lloyds’s HUGE mortgage book: around 300bn GBP, average (1y) default proba of 1.2%, 230bn in the least risky part (average PD 0.29%). 1.2% NPL rate with 40% coverage. Honestly, this looks like a good, strong mortgage book.
Even if you stress it, (like PD x3, coverage x2, etc) you’ll find very reasonable losses. Nothing to worry about. In fact, that’s what was done in the 2016 EBA stress test. The cost of risk on that book in the adverse scenario was 0.2%. (€600M of annual provisions.) Peanuts.
Fast forward to the 2018 EBA stress test. OMFG. WHAT JUST HAPPENED: The estimated losses in the first year are €8,600m !!! THAT’S 14 times the losses in the 2016 scenario!!!
I’ll tell you what just happened: IFRS 9. And long maturity mortgages going to Stage 2.
I know, I know, you don’t believe me and think I’m making this stuff up. Surely, there is another explanation and an accounting rule (even a stupid one) can’t create 8bn of losses.
But it turns out, the EBA published a very interesting and geeky capital template (I told you banks are transparent and publish a lot of data) where we can see that the impact of IFRS 9 on Lloyds for the 2018 stress test was…. €9,800m. MONSTRUOUS.
So maybe now you know why central banks want to tweak this as much as possible, and why they insisted banks use transitioning rules – which will basically “erase” 70% of those losses from capital *in 2020*
Because when that number was just in a stress test, nobody really cared, it was just a way to make banks hold more capital. But when the stress materialises in real life and it affects the banks’ lending capacity, nobody is happy about that rule anymore!
What will the market do with those transitioning arrangements, that is another story… and as always this was not investment advice, especially not on Lloyds (one reason among many is that Lloyds knows very well about this & has planned for it.)
