Here’s a little circle that seems to need squaring. Banks, pushed by regulators busy trying to avoid a repeat of the last credit crisis, have begun accounting for credit losses on the basis of what might be expected over the entire life of loans.
But now, in the middle of another crisis, regulators are telling banks that they should not be kitchen-sinking credit loss reserves for fear of cutting off the flow of credit to economies that desperately need all the help they can get.
The problem is that you can’t do one without doing the other. As Bank of America CFO Paul Donofrio said in April to an analyst who wanted to know if the bank was likely to see credit loss provisions get worse as the year went on: “If we thought we were going to have to add more reserve build in the future, we would have put it into this quarter. That’s how the rules work.”
And that’s the point. Building up provisions now rather than scrambling to account for losses when they happened was the purpose of moving to expectations of lifetime losses.
But it’s tricky to have banks be prepared without letting them prepare. Kitchen-sinking is baked in.
Regulators can’t wriggle out of the blame for this now with a desperate nod and a wink to banks to ignore rules lenders always said were pro-cyclical.
In June 2016 the Financial Accounting Standards Board in the US brought out its snappily titled ‘Accounting Standards Update (ASU) 2016-13’, which introduced the current expected credit losses (CECL) model. Anyone outside the US who is following International Financial Reporting Standards has IFRS 9 to contend with.
Those lenders reporting in both jurisdictions have the delight of preparing accounts that conform to both.
The two approaches emphasize the concept of expected losses over the lifetime of a contract, but they differ in some ways too. IFRS 9, for instance, has an initial requirement to assess expected losses in just the first 12 months, before switching to lifetime losses upon a significant increase in credit risk.
CECL doesn’t do this, meaning that it avoids the sudden and belated jump in provisions at 12 months.
In the current crisis, the difference between their treatments of undrawn commitments may be more important, given how quickly undrawn has changed to drawn over the last month or so.
CECL simply ignores the undrawn part of a facility and treats the drawn part as an amortizing loan.
IFRS 9, meanwhile, allows lenders to make their own assessment of likely future usage. That sounds like it might be a more lenient approach, but in fact the CECL is much more generous here since it fails to capture the fact that a drawdown might happen at a time of stress.
After all, someone who has lost their job might suddenly spend more on credit. Equally, a company worried about cashflow might suddenly draw down its revolver. In both instances there might be extreme correlations between different borrowers’ situations, which was precisely the problem in 2007 and 2008.
By pushing the provisioning cliff edge back to the start of a loan, rather than the point at which it becomes stressed, both approaches introduce more pro-cyclicality rather than less.
That’s because a vast chunk of new provisions must suddenly be considered for entire loan books at the first sign that conditions are set to worsen. And because such provisioning itself will necessarily stymie banks’ ability to lend.
CECL and IFRS 9 might have differences, but both share the same critical weakness. They were designed to tackle the last crisis, not the one that’s happening now.