The economics department of the International Monetary Fund (IMF) is up to its old tricks again – changing its mind.
Just three months ago, the team cut its forecast for UK GDP growth in 2013 to 0.7 per cent from 1 per cent, sparking charges that George Osborne should alter his policies. Last week, the IMF forecast was lifted to 0.9 per cent, accompanied by a great deal of comment that the chancellor seems to be on the right track.
The IMF itself said that “the UK economy is moving from rescue to recovery”. The psychological impact of statements such as this can be considerable. Macroeconomic policy in the modern world is much more about influencing sentiment than it is about the traditional policy tools in the economics textbooks.
But from a scientific perspective, how significant are the changes to the IMF’s UK growth projections for this year? First it was 1 per cent, then 0.7, and now 0.9. The numbers are obviously different. But at the same time, they are effectively the same. A depressing feature of economic commentary is the importance placed on trivially small changes such as these.
We see similar reactions when the Office for National Statistics (ONS) changes its estimates of GDP growth in the recent past, even by just one tenth of 1 per cent. In reality, the potential measurement errors around any single measurement of annual GDP growth are as much as a whole percentage point either side. An estimate of 1 per cent could in reality mean anything between zero and 2 per cent. We see this in the size of the historical revisions which are made to growth rates by the ONS over time. Admittedly, revisions of a full percentage point are unusual, but they are by no means unheard of. Revisions on the scale of the changes to the IMF forecasts, from 0.7 to 0.9 per cent for example, happen all the time. This is not a result of incompetence at the ONS. It is an inherent part of how estimates of the size of the economy are put together.
We cannot put the economy on a pair of scales and weigh it. Instead, national statistical bodies around the world gather information from a wide variety of sources to estimate what is going on. Income tax and VAT data tell us a lot about personal incomes and spending. But self-employment income, for example, is only reported to HMRC with a considerable time lag. So in the first instance it has to be estimated. None of this detracts from the fact that it is time to call it a day on economics at the IMF. Its forecasts are usually wrong, often ludicrously so, as this column has documented previously. IMF models give wildly different estimates of the effects of increased public expenditure, for instance, depending on which version you look at.
Withholding the UK’s contribution to IMF economics would only be a small cut in public spending, but it would be perfectly formed.
As published in City Am on Wednesday 17th July 2013