Monday, August 15, 2011
Building figures
The Times carried an absolutely stinging lead article on the Business pages on Saturday, reporting ‘Outrage over statisticians who keep getting numbers wrong.’
The latest Statistical Bulletin on Construction Output had been withdrawn within hours because of an ‘arithmetical error’: the figure for growth in Q2 of 2011, which had been given as an (unbelievable) 2.3 % was put back down to 0.5%, the same as was given in the provisional estimates for total GDP which were published on 26 July.
It is impossible to see the original bulletin now for any clue as to what happened, but I feel for all those involved. Only up to a point however – that kind of thing is just too glaring to let slip through, one would have thought.
This however is a particularly difficult time for making reliable early estimates for an industry where patterns of the recent past cannot provide a very robust or reliable guide to what might be the figure when all the returns come in.
Apart from Olympic activity, not a lot seems to be going on. Housebuilding seems to have ground to a halt round this neck of the woods, & there are not many tall cranes around town; oddly shop fitting seems to be thriving still – the old Woolworths has now been fully occupied, a new Pound Store opened in a prime location & a new Primark opens soon.
An absolutely classic example of the everlasting ‘timeliness versus accuracy’ dilemma. ONS has had to undertake a re-assessment of the parameters used for estimating output of businesses which do not send in their returns and dealing with outliers (unusually large returns), to cope with the changed nature of the industry since the credit crisis hit.
Perhaps we should once again consider adopting the Swiss solution – produce one, & only one estimate which is Final.
A year after the event.
