How to make business decisions in uncertain times

We live in times of extreme uncertainty, and that makes life hard for the people who are tasked with forecasting the future.

Actuaries, for example, have the task of forecasting trends so that insurers know how much to reserve or what to charge for their policies. Economists are supposed to predict the evolution of interest rates, consumer spending, etc.

Today, both professions admit that they are baffled, but are working hard to adapt.

“Good risk management should mean we are prepared for risks that threaten survival, whether or not we can attach a probability to them,”  comments economist Lord Mervyn King, who is a former governor of the Bank of England.”

“It is extremely dangerous to believe that you can say, ‘this is the precise amount that’s at risk’,” King adds.

Andreas Shakallis,  managing director of Principle I.B. Consultants  and a member of the Board of the Cyprus Association of  Actuaries, is also unhappy with the current models being used in forecasts.

Andreas Shakallis

“We have pretty much been relying on so-called ‘black box’ models for forecasts,” he told the Cyprus Mail in an interview. “This means a model in which certain or all parts of  the processes being modelled are not visible and non-traceable; you have input, then black box phases, then output.”

After the worldwide financial crisis of 2007-2008, insurers were expected to incorporate black box modelling into their forecasts.

“At that time was fresh from financial crisis, a lot of companies, international giants as well as smaller ones, were facing solvency issues, and they thought that they could alleviate their problems by using these internal models. It allowed them to deviate from standard solvency requirements,” Shakallis says.

Since then, techniques using AI for decision-making with black box modelling have come into use at many companies, but the lack of transparency is simply increased using these methods.

But reliance on these models which take past data into account to predict the future, simply shifts responsibility, Shakallis points out. “Future expected outcomes may well not follow past data.These models can be used to gain insight onto problems, but are very suspect in forecasting future events.”

King goes a step further. “It’s another example of regulation being overly precise, and as a result people make up numbers. “I think it is important not to get fooled by the apparently precise nature of black box models, which are usually designed by someone else with fancy distributions but made-up numbers. If the numbers are made up, it doesn’t matter what distribution you have. Regulators accept them because they want to have a number, and people go on fooling themselves,” King insists.

The solution? “Take the example of forecasting mortality in a pandemic. A pandemic will cause you extra mortality or morbidity rates in your portfolio. The best way to deal with this uncertainty is to set aside extra capital reserves to cover the increased loss beyond what the best data would lead you to expect. Then it’s possible to use stress testing techniques to ensure sufficient reserves,” Shakallis explains.

It’s also possible to pass on risk to reinsurers or to use other risk transfer techniques that shift the risk to other entities, he adds.

“The point is, if your calculations based on existing data suggest that you make reserves of an X amount , you should then set aside , depending on your financial condition, a multiple amount of X (e.g. 1.25X or 1.5X), in order  to increase significantly the chances of withstanding uncertainty and unexpected events in the future,” Shakallis notes.  “In times like these, increased caution is more important than trying to justify decisions with a complex modelling system.”

Another expert agrees. “The best judgement possible should be made based on existing data. Then, careful, weighted judgements based on the best knowledge you have should be used to modify them,” says Pantelis Damianou, professor of Mathematics and Statistics at the University of Cyprus.

What about other kinds of decisions, like investments?

In terms of investing, Shakallis also thinks that the use of complex modelling techniques just doesn’t provide satisfactory results. “Assessments of fundamentals, combined with judgements based on reasonable expectations are the best course.”

So what route to take in making business decisions in times of uncertainty?

The experts seem to agree: Make the best judgement possible, and then act with plenty of caution.

“There is really no substitute for decision-makers asking, ‘What is really going on here?’” King says. “What are these numbers about? What are they for? What decision depends on them? Thinking those things through more deeply is worth a thousand numbers coming out of a black box model.”