Crystal balls

Predictions are rubbish

Horoscopes are incredibly popular, quite literally unbelievably so. Financial and political predictions should be treated the same. Forecasts are usually either wild guesses or extrapolations of past trends.

In a series of experiments, behavioural economists have collected convincing evidence pointing to the human tendency to believe that patterns of the past will prevail in future. We think we are looking ahead, but in fact we are staring into history. The same fallacy is hard-wired into state-of-the-art models used to forecast economic growth. Moreover, they assume that any genuine, exogenous shock - be it a drop in oil prices or the bursting of a property bubble - will only have a short-term impact. It is assumed that economies will digest disruptions quickly and revert to long term trends. This would be fine if we lived in a mechanistic, predetermined world where the relevant probability distributions were known beforehand.

But unexpected events - occurrences that Nassim Nicholas Taleb famously dubbed “black swans” - are neither predetermined nor predictable.

Just think of the three most important events of the past 25 years: the fall of the Berlin Wall, 9/11 and the financial meltdown. They were huge surprises which have had a fundamental and long lasting impact on political, social and financial affairs.

If they had been anticipated, meanwhile, they might not have happened because economic and political agents would have adjusted their behaviour in advance.

Even trends that are visible, well in advance, present headaches. The mobile internet is a case in point. At the turn of the century, experts more or less accurately predicted a rapid rise in wireless communications.

Yet, still, telecom companies spent too much on 3G licences and the technology took much longer to take off than was expected. And outfits such as Apple, Google and Facebook - companies one step removed from the golden ether - have drawn the largest tranches of benefit.

There is nothing wrong with making predictions. Indeed, if you refuse to look into the future you will almost certainly be less well able to cope with present dangers.

Like horoscopes, predictions are also entertaining. Ultimately, however, they are akin to a chocolate fire guard. Enticing but, well, a bit rubbish.