New Methods for Using Monthly Data to Improve Forecast Accuracy

Publisher Website
Google Scholar
Abstract
Econometric forecasters continually seek ways to increase forecast accuracy. As new data are released, the residuals of forecasting models are examined for evidence of structural change and equations are modified if necessary. Several of the participants in the Model Comparison Seminar have recently investigated alternative methods for using monthly data in a systematic way to adjust forecasts produced by quarterly models. These initial studies are reviewed in this chapter and some illustrative results are presented. It begins with a review of some of the implications of temporal aggregation for the specification and estimation of models and their use in economic forecasting. This review is intended to provide motivation for the use of high-frequency (monthly) data in forecasting economic aggregates, as well as to indicate some of the difficulties that are involved. The chapter concludes with a presentation of some illustrative results obtained using the Michigan Quarterly Econometric Model of the United States economy.

This publication has 0 references indexed in Scilit: