Social

Predictions Gone Awry

When you want to know what’s going to happen in the markets or in our increasingly complex future, it’s best to go to the experts to get the answer. Right?

A recent article in The Atlantic suggests that this may actually be the worst possible strategy. It doesn’t delve into the well-known problems that experts have had in predicting market twists and turns or even the movements of interest rates, much less the next recession. Instead, it looks at research which suggests that so-called “experts,” with decades of experience and deep knowledge in a particular subject, may be the worst sources of accurate predictions about any aspect of the future—and particularly in their own field.

One study looked at more than 80,000 predictions, many of them about the Soviet Union during the Cold War. It found that where experts declared that future events were impossible or nearly so, 15% actually occurred nonetheless. When they declared events to be a sure thing, 25% of them failed to happen. When faced with their results, many of the experts doubled down on their predictions instead of admitting flaws in their judgment.

Another study compiled a decade of yearly predictions about the value of the dollar against the euro, made by experts at 22 large international banks. Every year, the economist at every bank predicted what the exchange rate would be at the end of that year. Over ten years, the banking experts missed every single change of direction in the exchange rate, and in six of those ten years, the true exchange rate fell outside the entire range of all 22 forecasts.

The article cites a third study which shows that the most accurate forecasters are not subject matter experts, but people with wide-ranging interests and unusually expansive reading habits—but no particular relevant background. They viewed their peers’ views as learning opportunities rather than seeing the conversations as an opportunity to convince people of their rightness. The best forecasters appear to view their ideas as hypotheses in need of testing, and learn from their mistakes.

Please Note: This article is used with permission from a newsletter to which KFG subscribes. It is for our clients only and may not be republished.

Print Friendly, PDF & Email
Comments are closed.