Lies, damned lies and changing GDP figures
The official figures showing the recession is not yet over were a surprise. But what if they are revised to show the economy was actually growing after all? It would change not only our outlook but our faith in all statistics.
If government data can change so radically they turn recession into growth, why should we believe unemployment or inflation statistics? The public would not only be confused, but entitled to regard figures as flexible tools of propaganda.
Speedy statistics are good, and to produce an estimate for the change in gross domestic product just 23 days after the quarter ends is admirable. But if the revision a month later shows the opposite result, were does that leave the credibility of such information – or if the further revision after another month changes the conclusion, where are we?
Britain’s GDP was estimated to have shrunk by 0.4 per cent in the third quarter, according to the widely-publicised first estimate, making this the longest recession since figures were first compiled half a century ago. But the previous quarter’s UK fall in the GDP was revised from 0.7 to 0.6 per cent after three months and the American figure was slashed from 1.0 to 0.7 per cent on revision.
Over the past five years the average revision in Britain’s GDP initial statistic is 0.03 per cent after one month and 0.08 per cent after another month. After three years, the figures are revised by an average 0.21 per cent.
However those averages of errors disguise some large changes in individual figures and individual sectors. Revisions of the UK’s second-quarter GDP stats cut the first estimate for the construction industry of a 2.2 per cent fall to just 0.8 per cent and halved the initial fall in manufacturing output to just 1 per cent.
That the third-quarter UK figures seem so at odds with the view from the shopfloor suggests the new estimate due on 25 November will show less of a fall or possibly even a rise in GDP. December’s updated figures could show the same trend. Certainly the rough rule is that when the economy is going down, revised figures tend to be worst that the original estimate but when the economy is strengthening the revisions are upward – as the second quarter showed.
But – especially with a general election looming – what will the public make of revisions that can turn slump into growth at a stroke. Why will they believe unemployment or inflation statistics? Why believe claims that bank lending is rising? Or population figures? Statisticians who change their mind undermine their own credibility.
