ONS censured for arithmetical error

As if the Office for National Statistics wasn’t already in some trouble with the building industry (see Straight Statistics here and here) it has compounded its difficulties by some dodgy adding up.

No sooner had it published its construction estimates for the second quarter of 2011 than it was forced to issue a correction. The industry hadn’t grown by 2.3 per cent in the quarter, but by a much more modest 0.5 per cent, it admitted on 12 August.

How the mistake occurred is not clear. The ONS says it was “an arithmetical error in the final stages of the Output in Construction estimates”. But we’ll find out eventually, because the UK Statistics Authority has called for a report on the incident and has today announced its own independent inquiry into construction statistics, which it intends to publish. 

This will be a blow to ONS, which robustly defended itself against charges that its building statistics were questionable. It took over responsibility for the statistics from the former Department for Business, Enterprise and Regulatory Reform in 2008, but its figures on output have been questioned by construction industry leaders.

It has recently recalculated data for the industry back to 1955, which is was previously unable to do because it was tied to GDP data in Government Blue Books that couldn’t be changed. The revision showed that the industry was bigger than we thought in the peak year of 2007, by £6.5 billion, and that the fall between early 2008 and early 2010 was sharper that we previously thought, 16.0 per cent rather than 13.7 per cent. See Brian Green’s Brickonomics blog for a specialist analysis.

The UKSA has also raised concerns about the way ONS commented on the GDP data published on July 26, which showed anaemic growth of 0.2 per cent in the second quarter of 2011. The commentary listed various reasons for this, including the Royal Wedding, the Japanese earthquake and tsunami, warm weather and the sale of Olympic tickets that took money out of people’s pockets without a balancing output, since the tickets remain undelivered.

Sir Michael Scholar says in a letter to Stephen Penneck, Director General of ONS, that he was concerned that some commentators saw this as ONS making excuses for the Coalition Government. There’s no grounds for this, says Sir Michael, but he goes on to suggest that if ONS statisticians are going to cite special factors, they should do so every time, regardless of whether their effect is to increase or diminish GDP. And the actual effect of the factors should possibly not be quantified, as they were for Q2 2011.