Risk V/S Uncerainty

In our day-to-day language, and even in finance, people tend to use the terms risk and uncertainty interchangeably. But in the 1920s economist Frank Knight made a distinction that I find quite useful. He argued that risk describes a system where we don’t know the outcome, but we do know what the underlying probability distribution of outcomes looks like. So think of a roulette wheel—when the croupier spins the wheel, you don’t know where the ball will land, but you do know all the possibilities and their associated probabilities. Risk also incorporates the notion of harm—that is, you can lose.

In contrast, uncertainty reflects a situation where you don’t know the outcome, but you also don’t know what the distribution of the underlying systems looks like.  Uncertainty also doesn’t necessarily imply harm, although it often does. So it’s not hard to see that most systems we deal with in the real world are really uncertain, not risky. Uncertainty better describes issues like terrorism or the avian flu, or even markets. Here’s why I’m stressing this distinction: we can model risk using probability calculus. In fact, the statistics of risk are relatively straightforward. In contrast, we can’t model uncertainty easily. And real trouble arises when we model uncertain systems using the mathematical tools of risk. Yet this is precisely what many people do in financial markets and in other domains as well. We’ll come back to this issue of risk or uncertainty quantification in a moment.

Psychologists have demonstrated that  events that are not vivid in our minds get assigned very low probabilities—much lower than the facts warrantI suspect for us to mobilize, as a society, to address risks like global warming or energy constraints we will need one or more 9/11-type events—a tragic incident that reveals what’s really going on.

To summarise, we humans are still not very good at dealing with risk or uncertainty. We are still linear thinkers, we have a nearly-insatiable need to link cause and effect, and we assess probabilities poorly.  However, we do now better understand some of the mechanisms that underlie complex systems, and that  knowledge can be very helpful in preparation for future catastrophic events.