The Gross Domestic Product (GDP) of the United Kingdom for any year is an estimated figure. After its first publication it is regularly revised in the light of new information. Revisions continue to be made for as long as 12 years after the event. There is therefore no true GDP figure, merely the most recent estimate which can be more than 6% different from the first estimate. This fact carries implications for users of GDP data, whether they use it as general background information or as an input to techniques. This paper shows that the size and pattern of the errors can be predicted via time series and regression models and that, as a result, GDP data can be adjusted to make allowance for the errors.