This isn’t a happy time to be involved with the production of national economic statistics. In the past month the Office for National Statistics has been assailed from every side for its inflation measures, and its professional competence questioned when it reported a decline in GDP for the final quarter of last year.
 
Tomorrow, revised estimates of that quarter are due to appear. If the revisions are large, and paint a less gloomy picture than the first estimates, ONS will face further embarrassment. Bernard Jenkin MP, Chair of the Select Committee on Administration, has bet publicly that they will be revised upwards.
 
On Conservative Home, he commented: “As Chairman of the Commons Select Committee which oversees governmentt statistics, I BET the bleak growth figures for the last quarter of 2010 (negative 0.5 per cent) will be revised UPWARDS. I will be inviting my committee to look into the quality of GDP data. The UK can do without such self-induced false pessimism.”
 
Tomorrow’s data will show whether Mr Jenkin is right. GDP figures often are revised, though nobody has ever been able to show that this reflects systematic bias rather than the intrinsic uncertainties of measurement.
 
The attacks on its inflation measures are potentially more damaging to ONS. Criticism from the left claiming that neither the Retail Price Index (RPI) or the measure favoured by the Treasury, the Consumer Price Index (CPI) truly reflect changes in the cost of living as experienced by the poor are not new. But this long-voiced criticism gained academic support last month with the publication of a paper in the Statistical Journal of the International Association for Official Statistics that concluded these measures are misleading.
 
Then the Bank of England, possibly seeking somebody else to blame for continuing high rates of inflation, added to the ONS’ woes by saying in its February Inflation Report (page 39) that ONS’ persistent undercounting of the price of clothing and footwear between 1997 and 2010, now corrected, were part cause of the recent rise in CPI inflation.
 
So not only are the indices a poor way of measuring inflation, they’ve also been poorly constructed, if both these criticisms are right.
 
Defenders of ONS have been thinner on the ground. Allan Monks, an economist with JPMorgan, said he thought the Bank had exaggerated the effects of the clothing price issue, and Karen Dunnell, the former national statistician, defended ONS’ integrity in a letter to The Guardian. She said that it was the Government which chose what measure of inflation to use: the ONS published three – the RPI with or without mortgage costs, and the CPI – and its releases showed clearly the time trends in each of them and the current significantly higher level of the RPI.
 
“Which indices are used by the Bank of England as inflation targets and by government to uprate benefits and pensions etc is entirely a matter for the government of the day” Dame Karen said. “I do not think it is government statisticians who are misleading the public about the current level of inflation.” 
 
The opposite conclusion is reached in the Statistical Journal paper, written by the late Michael Ward (a former Principal Economist at the International Economics Department of the World Bank), Derek Blades (consultant in economic statistics and former head of National Accounts Division of the OECD) and Carol Carson (who lists no affiliation but whom I guess to be the Carol S. Carson who was Director of the Statistics Department of the IMF between 1996 and 2004).  
 
They argue that ONS has failed to satisfy the public’s need “for a clear and transparent measure of price inflation”. Instead, it has given more weight to questionable economic theory. “The end result is that politicians may make bad decisions because they are using bad statistics and the general public loses faith because of the gap they see between their own daily shopping experience and the official measure of inflation” the team concludes.
 
They distinguish between a pure price measure and a cost of living index. A national cost of living index should be weighted to reflect the impact of price changes on the median or, preferably, modal household so that the index then reflects more correctly the impact of any price change on the majority of households, they argue.
 
The UK Government’s preferred measure, CPI, is “little more than a political convenience” to aid comparability with other EU countries, leaving out major costs such as housing “on the curious ground that it was not possible to reach agreement among the EU member countries on how to measure them”. While true that it may be difficult to find a comparable way of measuring housing costs in a group of countries where housing tenure is so different, “there is no reason why housing costs should not be included in a CPI designed to measure inflation in the United Kingdom.”
 
Many people would agree with that, but progress towards including owner occupiers’ housing costs into the CPI seems exasperatingly slow. The CPI has been the UK Government’s preferred measure of inflation since December 2003, when it replaced RPIX (the RPI excluding mortgage payments). Yet seven years later it still fails to reflect a significant element in house owners’ experience of inflation.
 
Why this should be is partly the technical difficulties of devising a way of doing it that does not offend against the purity of CPI as a measure of consumption, and partly (I suspect) inertia. People do not “consume” their homes, so how do you incorporate those costs?
 
Inevitably – this being the UK – there is an expert committee to advise on the issue, the Consumer Prices Advisory Committee. Its lengthy annual report for 2010, published last November, is largely concerned with the technicalities of this issue. It singles out two methods (the net acquisitions and rental equivalence approaches) and urges ONS to carry forward development work on both. It says this work should be a “high priority” for ONS and sufficient resources should be made available to complete the work in the next two years. 
 
That would make it a round ten years between the CPI being adopted as the preferred measure of inflation and its improvement so that it can more properly do the job. Small wonder that many outsiders remain dissatisfied and why it has been important that the ONS continue to publish the RPI, even if the proliferation of indices may add to public confusion.
 
Last word to the authors of the Statistical Journal article. “Politicians and their civil service advisers should not be allowed to get away with deciding what they wish to define as their ‘preferred’ measure of inflation. In the UK, this ‘preferred’ index, the CPI, is a harmonised – some would say bastardised – index that falls between two stools as an inflation measure and a cost of living index.”
 
And this measure, don’t forget, is now being used to uprate many pensions and benefits, with long-term effects that will disadvantage millions of people.