View from City Road: Even actuaries can teach banks something

As they put the finishing touches to the first set of annual results since they began their slow crawl out of recession, the big four banks are pondering ways to avoid being caught next time around.

Recession after recession, the banks have driven gaily into a brick wall of bad debts without seeing it until the moment of impact. Surely there must be a better way to run these giant businesses?

One answer being studied by Barclays, among others, is to adapt the methods of insurance companies to banking. The idea is to use actuarial measurement of risk to improve the way provisions are set aside against bad debts.

At the moment, banks have provisions against two categories of debtors - those who are having trouble paying and those who managers expect may soon get into the same position.

Although they try to smooth the costs of bad debts by putting money aside in the good times, in practice it never really works and the banks end up caught by surprise as the economy turns down and companies get into trouble.

But actuaries would say there is a statistical probability of every loan in the book not being repaid - even those borrowers with unblemished records, where there are no suspicions of problems. So why not set aside a bad debt reserve against all of them, based on actuarial assessment of risk? That is the way insurance companies think. It would smooth the cost of bad debts by setting money aside earlier in the business cycle.

In fact, this is already applied to mass products such as Barclaycard, consumer credit and mortgages. The hard part is to apply it to business accounts, which in Barclays' case number 400,000.

But better computer monitoring systems are being developed for business accounts. It will soon be possible to link these to software that does the actuarial calculations, and assess the risk of loss across the entire loan book, setting aside bad debt reserves accordingly. That might take some of the shock out of the next recession.

But perhaps the most important result of these technical developments, already well advanced abroad at the likes of Bank of America, is that they will make bank managers assess risk more carefully on every loan they make. There is no substitute for that if the banks want to lose less money.