News and noise in G-7 GDP announcements

Revisions to GDP announcements are known to be quite large in all G-7 countries; quarterly growth rate revisions are regularly more than a full percentage point at an annualized rate. We examine the predictability of these revisions using standard statistical tests of whether the preliminary announcement is a rational forecast of the subsequently revised data. Previous work suggests that U.S. GDP revisions are largely unpredictable, as would be the case if the revisions reflect news not available at the time that the preliminary number is produced. We find that the degree of predictability varies throughout the G-7. Although we find little predictability in U.S. revisions, the data revisions for several foreign countries are highly predictable. We also perform a simple real-time forecasting exercise showing that for several countries, the predictability of data revision could be used to generate improved preliminary data.