Is this a Risky Time - or an Uncertain Time?

A lot of the discussion we, and I’m sure you, have been having over the past 6 months has concerned decision making under uncertainty. The topics vary but are interrelated: What will be the impact of zero interest rates? What should we be saying to pension clients? Will I be safe if the company reopens the office? How do we invest in markets that appear to have lost touch with economic reality? How will advice businesses cope in an environment of significant commercial and regulatory change? Is the demand for personal financial advice growing or shrinking? Should we be expanding our firm at this time or battening down the hatches?

It’s useful to try to think about what characterises these problems. Are they “Risks” in the way investors understand them – reflecting the events of the past and of which we can calculate some probabilities with more or less certainty? Or are they “Uncertainties” where quantification is more likely to be confusing than illuminating? President Obama explained that when told Osama bin Laden might be in a compound in Abbottabad, various members of the CIA were asked to provide estimates of probability of him being there and these ranged from 30 to 90%. Quantification, far from helping, created confusion.

I learned about this way of distinguishing between Risk and Uncertainty in a podcast, Econtalk, a weekly conversation between Russ Roberts, formerly a Stanford economics professor, and the author of some recently published book or article.

A recent conversation was with Mervyn King, ex Governor of the Bank of England, and John Kay, a leading UK economist. Jointly they have written Radical Uncertainty, which tries to address how best to characterise these problems. They introduce the distinction between “Puzzles” and “Mysteries”. The book is primarily about the inherent failure of economic models to be useful in addressing social problems. Mysteries are scenarios where a data set is only one of the elements to consider or where there is not a complete data set or can never be a complete data set. Puzzles have a solution, and generally, one right answer. Computers are good at puzzles.

But mysteries need judgement. Unlike traditional logic which is backward looking, addressing these problems needs forward looking tools. King and Kay say that an answer is usually best arrived at through abductive reasoning or “inference to the best explanation”. What is the simplest, or most likely cause of where we are now?

Mysteries, whether social or commercial, have multiple dimensions and need judgement to resolve. And the best way to capture those dimensions, in my experience, is to ensure that multiple points of view are brought to bear through involving people whose judgement you trust.