If you are already feeling jaded by the constant procession of so-called experts speculating on who will and will not win the upcoming general election, the good news is you only have another two months to endure. The trouble is, of course, that is also the bad news so let’s try and cheer you up with one further nugget – you might as well ignore anything any pundit has to say on the outcome of the 7 May polls.
Here on The Value Perspective, we are nothing if not even-handed when it comes to our opinion of forecasters. Whether their field is sport, weather or economics, we hold them all in equally low esteem and we see no reason to make an exception of political commentators – especially after reading Expert political judgment: How good is it? How can we know? by US psychologist Philip E Tetlock.
The book was the culmination of a 20-year research project, in which Tetlock and his team tracked 284 different political scientists, journalists and other commentators and the 82,361 predictions on the subject of politics these people made over that period. The conclusion? That their predictions would actually have been beaten by random chance.
Now, presumably the fact a monkey with a pin could have done a better job calling the various election results and other political scenarios is just unlucky for those experts – insofar as how history mapped out over that particular 20-year period. We are not going to suggest their forecasts will always be worse than those of a simian stand-in. Nevertheless, there was still a significant weight of evidence to suggest that, on average and over time, the experts did not really bring very much to the party of perspicacious political prediction.
Furthermore, while one might hope that the closer one comes to a political event, the more accurate the experts might be with their forecasts, again Tetlock unearthed little evidence to indicate that is in fact the case. He did, however, find the various commentators unbowed by their past lack of success and still curiously confident of their ability to forecast future events.
Towards the end of their project, Tetlock and his team revisited a number of the pundits to ascertain their recollections of the predictions they had made at the time. On the whole, they found that – compared with their initial conversations – these people subsequently underestimated the degree to which they had been uncertain about their predictions.
In other words, any chance the forecasters had of learning from their past mistakes had been seriously diminished by their recollection of events and the passing of time. Not only did the experts tend to believe things subsequently were less uncertain than they were, they had also convinced themselves that the outcome that did actually transpire had been more likely to happen.
Unfortunately – and like all good politicians we have hidden the worst piece of news at the end – that does means, if there turns out to be a second general election later this year, you should expect no change in either the identity of the procession of forecasters speculating on the result or the confidence with which they again roll out their predictions.