Forecasting Using First-Available Versus Fully Revised Economic Time-Series Data

First-reported monthly and quarterly time-series data on nine macroeconomic variables from 1960-1993 are given. Features of this so-called "unrevised" or "first-reported data" are discussed, and the data is compared with standard "fully revised" data using Granger causality tests. For the purposes of real-time forecasting, as well as comparing professional forecasts with traditional econometric forecasts, the use of unrevised (or, even better, "real-time") data has a number of advantages over the use of fully revised data.