Monday, 29 April 2013

Triple dip recession? Dip schmip.

The only kind of triple dip that matters
In my new book, Dice World, one of the subjects I cover is the way that statistics are misused by the likes of politicians and the media. For instance, the way that a politician can get all huffy about a policy having a negative effect on families with just two average earners - without pointing out that having two average earners puts a household in the top 25% of the country. Or the way the media can make a big thing about a statistic by using percentages, where the actual change is negligible. So, for instance, we might hear that pickpocketing has risen by 100%. Outrageous! Sack the chief constable! But if you hear that the number of incidents has gone up by one compared with the previous year (because there was only one instance of pickpocketing last year), the statistic has a rather different flavour.

A rather more subtle statistical misuse comes with all the fuss about whether or not the UK was about to enter a triple dip recession. As it happens we didn't - so the Chancellor could make all sorts of positive political noises about how his policies are correct. If we had, then no doubt the Shadow Chancellor would have jumped in yelling that this showed how bad the government's policies are, and how the country was going to the dogs. Yet either interpretation (and getting all excited about it being a triple dip) is another misuse of the statistics.

To respond this way to these tiny quarterly shifts is similar to sitting on a beach trying to decide whether or not the tide is coming in. A wave splashes onto the beach. 'The tide's coming in!' You cry. Then the sea recedes a little in the post-wave lull. 'No, the tide's going out!' Then the next wave arrives. 'The tide's coming in!' And so on. Very soon you would arrive at a triple tide afternoon.

Any data of this kind has both underlying trends and noise, the noise being random movements that do technically have a reason - but the reason is so complex, a messy mix of factors, that you might as well just regard it as chaotic noise and ignore it. Tiny shifts in GDP from quarter to quarter are just such noise. And should politicians or journalists choose to make much of these 'dips' or 'recoveries' (I was pleased to hear the BBC's Stephanie Flanders avoid this trap on the radio the other day) they are either deluded or trying to mislead us.

No comments:

Post a Comment