Forecasts and actuals: The trade-off between timeliness and accuracy

Abstract

Abstract The process of estimating GNP starts with forecasts made years before a quarter has begun and continues for years after it has ended as preliminary estimates are repeatedly revised. Earlier estimates can be more useful to decisionmakers even though later estimates, based on more information, tend to be more reliable. This article documents the trade-off between the timeliness and accuracy of alternative estimates. It concludes that the preliminary estimate of nominal and real GNP, available about 15 days after a quarter has ended, is overrated in that it is less accurate than later estimates and no more reliable than the flash estimate made 15 days before the quarter has ended.