STUART THEOBALD: Banks go back to the future as discretion makes a comeback

Once upon a time, banks could make up their earnings. Well, not quite, but bankers had a lot of discretion.

This column was first published in Business Day. 

Once upon a time, banks could make up their earnings. Well, not quite, but bankers had a lot of discretion. With banks’ earnings season now upon us, it is tempting to reminisce. For various reasons, discretion has made something of a comeback.

There are many studies that show banks engaged in earnings smoothing. This was the practice of managing just when a bank would recognise losses or gains to keep earnings and dividends growing steadily year after year. That was thought (and probably did) build confidence among shareholders and depositors alike.

One academic study on SA showed that banks tended to smooth earnings when they were undercapitalised. Globally there are standout examples such as the 1987 case of Citicorp, then the largest bank in America, which dramatically increased its loan loss provisions by $3bn in 1987 so that it could achieve its earnings target in 1988. Here, smoothing was much more subtle.

There are several ways banks did it — loan provisioning was the most obvious, with banks’ managers determining that more should be put aside for bad loans in good years, so that less could be put aside in bad years, keeping credit costs smooth. The use of derivatives was another way to do it, with banks using flexibility on how and when to recognise gains and losses on interest rate and exchange rate hedges.

The impact of this practice ranged from benign to outright shams. The benign reading was that bank managers have the best insight into the performance of their business, so giving them discretion to manage their reported earnings over time leads to good outcomes for shareholders, who want a nice smooth earnings and dividends profile after all.

The sham reading was that managers are rewarded on the short-term performance of the business, and discretion just meant ramping up earnings so they can book bonuses. The latter reading became more prevalent in the 1990s as performance-based remuneration, in which managers made most of their money from short-term outcomes, became dominant. The rise of modern portfolio management also led to demands that financial accounts be completely transparent to leave it up to investors how to respond to bank performance.

The accounting regulators moved in. IAS 39 (International Accounting Standards) was the first strike in 2003 (hitting SA a year later), which forced bankers to provide for loan losses based on “objective” evidence, largely the historic performance of loans. This was, however, only backward looking. When the global financial crisis hit, we realised with a shock that it wasn’t such a good idea for banks to be provisioning by only looking in the rear-view mirror. That left them free to take huge risks on instruments that may have poor performance in the future but allowed them to book growth now.

And so, we came to a new standard: IFRS9 (International Financial Reporting Standard) which came into being in 2018. This requires bankers to book loan loss provisions on a forward-looking basis, so they must anticipate what the likely losses on loans will be and take a cost up front for this.

But the Covid-19 crisis showed up a problem with this: it is hugely procyclical. When the crisis broke, banks had to take large losses immediately to reflect the obvious deterioration in the credit outlook. That had the potential to impair capital which meant banks’ risk appetite plummeted, further weakening the outlook.

Banks and their regulators have been on a careful dance since. Following international precedent, the SA Reserve Bank last year told banks to go easy on provisioning for loans that they had to restructure for clients affected by lockdowns.

Lockdowns, the Bank said, were not a change in fundamental creditworthiness, but just an interregnum, and therefore banks didn’t have to throw the kitchen sink at their provisions.

But banks also faced a problem in that even their forward-looking models weren’t equipped to deal with the economic shock represented by Covid-19. These models still rested on historic data to guide expectations about the future. So, the concept of “judgmental overlays” was introduced, which saw banks putting aside billions in additional provisions last year to cover the bleak outlook that wasn’t quite being captured by the models. Management discretion was back.

Fast forward a year and we can test how good these predictions were. Large global banks have been reporting an enormous jump in profits, largely because the judgmental overlays they took have turned out to be too conservative. So that means they can write those costs back into the profit line and report enormous growth.

Locally, the focus is on how restructured loans are doing, as well as how those judgmental overlays are looking in retrospect. My view is that the restructured loans were underprovided, but the judgmental overlays were probably overdone.

Banks won’t want to look too profitable — there is going to be a big year-on-year jump in profits anyway — so the judgmental overlays won’t be released yet. They will be banked for when things have calmed down and the income statement can do with some support. We’ve gone back to the future.

Theobald is chair of research-led consulting firm Intellidex.