Friday 31 January 2014

Economists and predictions

When the latest set of unemployment figures was published, government ministers rushed to claim all the credit for the improvement, whilst opposition politicians tried to find other avenues of attack.  It’s a sight we can expect on a monthly basis from now until the election in 2015 at least, although in any given month when the figures go the other way, the government will blame factors outside their control, whilst the opposition will shout “told you so”.
Will there really be months when the figures go the other way?  I’d bet on it; variation is a normal part of economic outcomes.  Hard and precise prediction, when it comes to economics, is a bit of a mug’s game.
I thought that one of the most revealing comments in the reporting of the latest figures was the almost throwaway comment on the BBC News that many economists were surprised by the figures.  I thought that deserved rather more attention than it got.
It could just be of course that the government figures for unemployment are actually no more reliable than those for hospital waiting lists or crime, both of which have been found to be, shall we say, “wanting” when it comes to the minor matter of accuracy.  It’s more likely though that it reveals an inconvenient fact about economic forecasting – that the term is an oxymoron.
If economic forecasts made by some economists turn out to be right, it’s more likely to be a result of the sheer number of people making predictions and the range of predictions being made, than of some economists being better than others.  It’s a bit like the hypothetical infinite number of monkeys and typewriters producing the complete works of Shakespeare.
Economics, like any other discipline dealing with human behaviour, is of its nature better at analysing the past than at predicting the future.  There are all sorts of reasons for that, not the least of which is that humans hearing any predictions can, and often do, change their behaviour as a result if they actually believe the predictions, thereby invalidating them.
But if there’s one thing that is less reliable than a prediction made by an economist, it’s a prediction made by a politician based on a prediction made by an economist.  It’s unlikely to stop them though…

No comments: