A Statistical Debacle
Friday, 26 September 2008

Numbers feel the crunch

At the turn of the year Bulletin carried a detailed piece outlining the difficulties that were being faced by the Office of National Statistics and the worrying impact these might have on the accuracy or otherwise of the information that it produced.

There were numerous challenges: an unpopular relocation to Wales; the loss of a number of senior (and therefore experienced) statisticians; political interference in the timing and presentation of information; and concerns that some of the statistical methodology employed might be flawed.

Events have proven the concerns valid. No sooner had the ONS published figures showing record growth in retail sales in May – when it was clear to everyone that the UK was entering a consumer downturn – than it needed to produce revisions to figures for economic growth which showed that in the three months to June the economy, far from showing a modest growth, had actually shown zero growth for the first time in 15 years.

Add to this the latest monthly retail figures showing month-onmonth growth – which even the British Retail Consortium thought odd – and the news that property transaction figures have had publication delayed because of some as yet undiscovered glitch, and it is becoming clear that the UK's national statistics, like its Home Office, may not be fit for
purpose.

But does it matter for business? Of course it does. The RPI and the CPI are the basis for millions of wage claims, indications of economic output and retail spending swing investment decisions worth tens of billions of pounds and property transactions are watched closely by every institutional investor.

Our "official" statistics are far from dodgy, but there is a growing acceptance that they are far from perfect and that, in a statistical sense, is a disaster.





Digg!Reddit!Del.icio.us!Facebook!StumbleUpon!Newsvine!