An article in The Times said that many economists were saying that the GDP preliminary estimate published on Friday 23 October was wrong and suggesting that this figure was liable to be revised. The Director-General wrote to point out that the figure was based on large amounts of data and that the revisions record was good.
In response to criticism in Saturday's edition of The Times ('Traders sneer at useless economists as data shows surprise fall in GDP'), I should point out that the Office for National Statistics' preliminary estimate of GDP is based on information requested from nearly 40,000 businesses relating to their turnover in the first two months of the quarter. The remaining month includes data from a further 20,000 returns.
There is always a trade-off between the timeliness of statistical estimates, and the prospect of revisions to these estimates as additional information becomes available. However, historically, revisions to the preliminary estimate have been on average no more than 0.2 percentage points when compared with the more complete data available three months after the end of the quarter (published in the UK’s Quarterly National Accounts Statistical Bulletin). The preliminary estimate is also, on average, a good predictor of later estimates, that is, it is neither systematically pessimistic nor too optimistic.
ONS is also completely transparent about the quality of all of the published estimates of GDP, and publishes extensive information on its website to help users to understand the reliability of the statistics.
In passing, the British Retail Consortium (BRC) has for many years published an estimate of retail spending based on a survey of its members. This is not something BRC has "resorted to publishing" in recent months.
Office for National Statistics