This paper investigates the bias and dispersion in official revisions of eight national accounting growth rates. The growth in GDP, consumption expenditure and personal disposable income by households has been subject to significant upward revisions and bias, especially after 1994. No significant bias was found in the revisions to the other national accounting aggregates. The official revisions are subject to a high degree of dispersion. Based on the 1984-2003 period, there is a 30 per cent probability that the “final” growth rate in GDE deviates by more than 5 percentage points from the first release growth rate. For most magnitudes, other than exports and imports, the dispersion in South Africa’s official revisions is similar to that of a sample of OECD countries. Using two examples, it is shown that the vintage of the data has a profound impact on the magnitude and significance of regression results based on such data.
Policy Paper 10