The future of macroeconomic forecasting: Understanding the forecasting process

A premise of this paper is that the forecasting process has not been thoroughly investigated in its entirety and that even some of the constituent elements of the process are not well understood. The components of the forecasting process include models, data, the judgment of the forecaster, and their interactions. The paper integrates the research that has considered various aspects of the forecasting process in order to suggest a more complete agenda for improving predictive accuracy. Specifically it was suggested: that past errors be analyzed more carefully; that records that explain the reasons for making adjustments be maintained; and that quality control methods be used to monitor whether or not forecasts are on track.