Empirical tests of input-output forecasts

This paper analyzes previous studies of the accuracy of input-output forecasts as compared with projections derived from alternative forecasting techniques. The problem of constructing appropriate tests of input-output forecasts is discussed. Major tests of the interindustry approach and alternative techniques, such as final demand blowup, GNP blowup and multiple regression, conducted in the past four decades are reviewed and the major findings summarized. It is shown here that, contrary to the belief of some economists, the input-output forecasting model performs as well as and usually better than any of the alternatives considered.