I ask because, in one sense, it is almost certain that GDP didn't fall in Q4 at all. In fact, it probably rose very sharply. I'm referring to real GDP before seasonal adjustment. Since records began in 1955, this has always risen in Q4. In the last 10 years, it has grown by an average of 2.5 per cent in Q4. In 2010, for example, it grew by 2.5 per cent, even though it fell by 0.4 per cent after seasonal adjustment.
Early figures suggest we're on course for a repeat. Industrial production, before seasonal adjustment, was 6.4 per cent higher in October than it was on average in Q3.
This matters, because the numbers before seasonal adjustment correspond to people's actual experience. During a cold snap, nobody (or at least, nobody outside Newcastle) says "it's quite warm after adjusting for seasonal variations, so I'll go out in just a T-shirt." Why treat GDP differently?
But we do. In Q1, GDP almost always falls; it's done so in 38 of the last 40 years. And yet policy-makers have never fretted much about these regular annual recessions, even though they have generally been very deep: for example in the last 10 years GDP fell by an average of 2.1 per cent in each Q1, as must as it fell during the worst seasonally-adjusted quarter of the 2008-09 recession.
Why then are these different recessions treated so differently by politicians and the media? Why is so much attention paid to small seasonally-adjusted recessions and almost none to large seasonal recessions?
There are three reasons: one bad, one possibly better, and one good.
The bad reason is that seasonal fluctuations are unavoidable - they are due largely to gearing up for Christmas and winding down after it, and not since Cromwell has a politician considered abolishing Christmas – whereas seasonally-adjusted recessions are considered avoidable.
I say this is a bad reason because it's not clear that seasonally-adjusted recessions are any more avoidable than seasonal ones. Even before the current crisis, IMF research had found that forecasters had almost always failed to foresee recessions, which suggests that they cannot in practice be eliminated, at least by policy-makers trapped within a "predict-and-control" mentality.
A slightly better reason is that seasonal fluctuations cancel themselves out, with Q1's drop being reversed later in the year, whilst non-seasonal recessions might lead to long-lasting slower growth.
I say this is slightly better because although there is evidence that deep recessions do have long-lasting effects, there's also evidence that seasonal fluctuations can sometimes lead to nastier downturns. Yi Wen, an economist at the St Louis Fed, has estimated that half of the variability of US output that looks like business cycle variation in fact arises from seasonal changes in output.
The good reason is that risk-pooling works better during seasonal recessions than during seasonally-adjusted ones. When GDP falls in Q1, the workers laid off are very often those only marginally attached to the labour market, such as students or home-makers who had taken temporary jobs. The typical response to such slowdowns is that people work less hard than they did in the run-up to Christmas. And, as Chicago University's Robert Lucas has shown, if risks are pooled well then fluctuations in GDP aren't important.
By contrast, seasonally-adjusted recessions are accompanied by a failure of risk-pooling. The problem with these is not that everyone suffers a bearable one or two per cent drop in incomes, but that a minority suffers a massive drop as businesses close and people lose their jobs. And our addle-brained politicians have always thought that the solution to this problem is the futile effort to prevent recessions, rather than the more technically feasible option of improving risk-pooling.
You might reply that there's another reason to care about seasonally-adjusted recessions. Surely, the probability that GDP fell in Q4 is proof that George Osborne's austerity policies are damaging the economy. Not quite. The best evidence for this is partly theoretical: the mechanisms through which expansionary fiscal policy were supposed to work are absent; Mr Osborne was insufficiently heedful of risks; and the idea of a structural deficit is just hooey. But it also lies in the broader fact that GDP has grown very little since mid-2010. To judge policy by whether quarterly GDP is plus a little or minus a little is to make a fetish of statistics without understanding them.
