For Ramos to avoid blame, she must tell credible story on Absa blunder

SA’s big four banks are fairly homogeneous. Sure, they have their quirks. Absa and Standard Bank have a retail bias compared with Firstrand and Nedbank. Some try to carve out marketing positions by being the greenest bank (Nedbank), or engaged with social media (FNB). But the marketing gloss does little to hide that they are basically commoditised businesses.

The problem with commoditised businesses is that it is too easy to compare them. When the products are essentially the same, a difference in performance comes down to the decisions of management. They cannot blame the market. They cannot blame the product suite, or particular assets. The heat comes down to their own decisions.

So in a week in which Absa announced a 6% decline in profits and Nedbank announced its profits will be up to 26% higher, investors have to ponder the stark difference in performance between the two banks.

The difference in profits comes down to this: Absa has been steadily underprovisioning for bad loans in its home loan books. Nedbank has been steadily over provisioning. Why? Was it incompetence or an effort to massage the numbers? Neither answer provides much comfort.

A bank has to decide how much of every loan it writes it should put in a kitty to cover the loans that inevitably go bad. Bankers like to make out that there is something objective about doing this. They have what they call a model, which takes a lot of data such as current loan repayments and macroeconomic indicators and spits out a number that is deemed the appropriate guide to provisions. Everyone including management, auditors and regulators see the number derived from the model, and sleep soundly at night thinking the provisions made by the banks are “objective”, and therefore, reliable.

But every bank’s model is different, and every bank believes theirs is best. Imagine four cosmologists, each with a different model of the solar system. You’d pause before deciding any one of the models was the right one. What the banks do is similar: each has its own view of reality, yet we happily accept each one is right.

Provisions are taken as expenses when you make them, but if it turns out you’ve provisioned too much, you get to release the provisions and add them to the income statement. So provisions are a useful way to smooth earnings. Managers get bonuses for meeting targets, not exceeding them. So keep the fat for when you need a bit of help in future.

It turns out that Absa’s model underdetermined the provisions it should have held. I suspect Nedbank’s did the opposite; we’ll see when it publishes its results on Wednesday. Absa’s model told the bank it should hold provisions of 17% against its legal book — that part of its home loans book that it is in legal proceedings over to recover its money.

Nedbank held provisions of 28% against its equivalent.

The big question is why? There are some obvious reasons. Nedbank has had a problem in its loan book stemming from 2008 when it splurged on new loans at exactly the wrong point of the cycle. But it still seems to have overprovisioned. It looks as if the bank was just being too conservative. But why did Absa underprovision?

What makes things look pretty bad for Absa is that just six months ago it thought it could cut provisions from a 19% coverage ratio to 17%, which helped earnings. But now it has taken more than R1bn in extra provisions to push that number to 23%.

Many have assumed that Absa cut provisions earlier because it was under pressure to deliver bigger earnings to parent Barclays. But CEO Maria Ramos has said often that she feels much more pressure from minority shareholders. Anyway, in the context of Barclays’ balance sheet, the difference Absa’s move made would have been immaterial, and shortlived. So if we leave aside manipulation, we are left with only one alternative: incompetence.

And who should we pin that label to? Ms Ramos may be able to take cover from the rapid turnover at senior level the bank has seen. There is a previous generation that can be blamed for the design of the model. She has also said the underprovisioning came to light because of a spike in writeoffs, although this should have been predictable.

The bank has suspended its head of collections, though it has refused to name him or her, saying investigations are under way.

Shareholders need answers: if it was incompetence, they need to know who did it, and need to know that it won’t happen again. For Ms Ramos to avoid the blame herself, she has to tell a credible story. One we’ve yet to hear.