Lies, damn lies and statistics…

In this context, may I make a suggestion that would apply to any circumstance where there was a major revision by the ONS to an earlier published figure.

In this case, the doubling of the second quarter 2003 GDP growth rate from an initial estimate of 0.3% to a revised 0.6% would surely count as a major revision. In these circumstances, it would be helpful to have an arithmetic reconciliation between, for example, the original 0.3% figure and the revised 0.6% figure.

In the private sector, where a finance director provides the board with a different final profit figure to their first estimate, they would be expected to provide an arithmetic reconciliation between the two. I submit that the ONS should be no different, notwithstanding the fact that I do appreciate national statistics are not based on double-entry bookkeeping.

Related reading