Markets or oil, analysts’ predictions are spectacularly wrong too often. But not for want of trying, hopefully

A month after US attacked Iran, analysts predicted crude would average $83 through 2026. Mid-June, the forecast was still bleak: $80-$90 a barrel till Dec.

Yet, Brent has already slipped below $72, close to where it was in Feb.

This is where you throw up your arms and ask, how could they be so spectacularly wrong?

You’ve done it before, when the dot-com bubble burst, when the sub-prime crisis happened, when Trump’s tariffs didn’t plunge US into a recession last year. And you’re none the wiser. Neither are we.

Problem is, the forecast business is a black box. And black boxes, we’re seeing, take forever to crack open and make sense of. What ordinary folk see is that analysts are really smart people. Their suits are impeccably tailored. And the offices they work for have auras that no blunder can tarnish.

Take the US Fed, for example. In March 2007, months before the sub-prime crisis exploded, the Fed’s outlook was sunny: “impact on the broader economy and financial markets of the problems in the sub-prime market seem likely to be contained.”

Or take the Deutsche Bank Securities economist who persistently forecast recession in 2000, not because of the dot-com bubble, but the Y2K bug.

At least he conceded, “I was wrong” and explained why: “I was focusing on the weak links. What I failed to do was to realise there were fewer and fewer of them.”

Yet, other analysts, at that moment, were talking up the dot-com boom, although it was about to fizzle and start a recession.

We can’t say why analysts get it wrong so often, but perhaps there’s some truth in the analysis that they place too much confidence in abstruse mathematical models. What they aren’t guilty of, though, is cooking up stuff.

Don’t believe the wag’s tale about a bored analyst who walks into office after a burger-and-cola lunch, looks at her assignment on forecasting global cheese demand in 2036, sees her $3.68 bill, and pencils in “36.8mn tonnes”.

Another analyst uses this “data” to predict steel demand for cattle pens, a third forecasts dairy software demand, a fourth calculates the impact of both on US GDP, basis which a fifth predicts sales of Taylor Swift concert tickets.

It doesn’t work like that, even though in 2012, the chief of a famous employer of analysts was trawling through company emails for the word “muppet”, after learning his senior staff used it to describe clients.

END OF ARTICLE