In November last year, the economy grew by 0.1 per cent, according to the Office for National Statistics (ONS) last week. The news, in the midst of many a miserable headline, was greeted with great excitement. Most City economists had predicted a fall by around 0.2 per cent.
Cue animated performances on the TV news channels, with presenters waving their arms at complicated graphs. And the news lifted the stock market because it reduced the chances of a technical recession.
Of course, in the current circumstances we should be grateful for anything which has even the palest tinge of rose. But the overall economic picture is still bleak. Comparing the September to November period, output fell by 0.3 per cent. Even more starkly, UK output is still slightly below what it was just before the pandemic, almost three years ago now.
Far too much weight is being placed on miniscule differences in economic numbers as they appear.
For most of the post-war period we had to be content with new GDP estimates appearing every quarter. It is only over the past few years that the ONS has started to publish monthly estimates. Before the decision to implement this change was taken, the ONS noted that the proposal to change publication timings was partly a response to two major reviews on the subject. These concluded it was “important for the ONS to strike a balance between the timeliness of GDP estimates to aid policy makers and the available data – and therefore the potential for revisions.”
The final phrase is significant. Estimates of how much the economy has grown in any particular month or quarter, and even year, are almost invariably revised as more information becomes available to the national accounts statisticians.
The economy cannot be put onto a set of scales and weighed. Its size has to be estimated from a variety of sources which vary enormously in their timeliness and reliability.
For example, the PAYE returns which employers submit give a pretty good idea of how many employees there are and how much they are being paid. But what about the self-employed? Their tax returns are only filed with a lag of many months. So any estimate of how much people in work have earned may be revised as more information comes in.
These revisions can be large, completely dwarfing the 0.1 per cent figure which caused so much excitement.
As with many aspects of economic data, the United States leads the way on making the size of revisions to GDP accessible.
We can look back, for example, at the fourth quarter of 2008, when the impact of the financial crisis really started to take hold. At first, the Americans thought that GDP had fallen by 3.8 per cent at an annual rate in that quarter. That was pretty bad – but nowhere near as bad as the 8.5 per cent drop which they finally settled on.
The recovery after the crisis was also substantially overestimated. The first estimates of growth between September 2009 and March 2010 put it at over 12 per cent. Now, a much more modest 7 per cent has been settled on.
None of this is intended as criticism of the national accounts statisticians, as they face a difficult and challenging task. What’s concerning is the lack of awareness of the political classes when it comes to the uncertainties inherent in these numbers.
Ultimately, the potential for erroneous decisions being taken on the basis of the monthly ONS estimates seriously outweighs the benefits of publishing them.
