News or Noise? an Analysis of Gnp Revisions

This paper studies the nature of the errors in preliminary GNP data,It first documents that these errors are large. For example, suppose theprelimimary estimate indicates that real GNP did not change over therecent quarter; then one can be only 80 percent confident that the finalestimate (annual rate) will be in the range from -2.8 percent to +2.8percent. The paper also documents that the revisions in GNP data are notforecastable, This finding implies that the preliminary estimates are theefficient given available information. Hence, the Bureau of EconomicAnalysis appears to follow efficient statistical procedures, in making itspreliminary estimates