Paperjam.lu

In this contributed opinion piece, Etienne de Callataÿ of Orcadia AM writes that economists should exercise caution in issuing forecasts. Photo: Jan Hanrion/Maison Moderne 

To the good of his ego and the misfortune of his profession, the economist is too often pushed to make forecasts, when his added value lies elsewhere, in the analysis of facts. John Kenneth Galbraith’s line is well known: “There are two kinds of forecasters: those who don’t know, and those who don’t know they don’t know”. A modern, uneconomical version was recently (earlier this year) given by Edgar Morin: “Expect the unexpected”. 

If the unexpected awaits us, then shouldn’t economists refuse to forecast? No, because you have to understand that we are trying to find out what tomorrow might be up to, and this can be useful. Not all forecasts are necessarily invalidated by continual shocks. At the same time, we must hope that economists are questioned about other things and that, when it comes to an exercise in anticipation, they constantly remind themselves of the limits of forecasting and therefore only engage in it with prudence and temperance.

As David Dunning said in 2014, the human being, this “confident idiot”, has a considerable--or annoying--tendency to always want to detect causal relationships and, following that, to share their “intuition” or propose their own little theory. In the face of the unknown, we seek to hold on to something known that resembles the unknown. And this is what leads us to think that “covid-19 is like the flu”. This is called mental model bias. So let us beware of analogies! And beware of extrapolations. They are reassuring, making the future a logical continuation of the present, but at the cost of minimising the probability of shocks.

That’s not all. There is also the adage of Ovid: “Tarde, quae credita laedunt, credimus”, which translates as: “We believe slowly when belief brings pain”. With all our biases, our rationality is put to the test. Behavioural economics finds its breeding ground there. So it is appropriate to evoke overconfidence bias. Olivier Sibony at HEC Paris illustrates this with this survey carried out shortly after the outbreak of the pandemic in France: 59% of people affirmed that chloroquine is effective and 20% declared that it is not, leaving only 21% of people with the wisdom to say they don’t know!

Another extremely common tendency in the forecasting units of economic institutions is imitation bias. Since the worst is to be alone in making an error, and it is not serious to make a mistake in good company, the forecaster follows his colleagues, and this generates inertia in the forecast.

That they did not warn about the 2020 recession is not a disgrace for economists. On the other hand, it is more annoying that they were not good at heralding the inflation of the 1960s and 1970s, the disinflation that followed and the persistence of minimal inflation in recent years. The same is true of the evolution of employment and wages, where the profession has been mistaken over time on the so-called “structural” unemployment rate, compatible with wage moderation.

Originally published in French by Paperjam and translated for Delano