Forecasting with Dynamic Regression Models

This study presents the concepts and practice of interpreting single equation dynamic regression models. Emphasis is placed on possible dynamic patterns, such as distributed lag responses of the output series to the input series and the auto-correlation patterns of the regression disturbance.
This volume presents the basic concepts and practice of building, using and interpreting single equation dynamic regression models (also called transfer function and intervention models). The book is a companion volume to "Forecasting with Univariate Box-Jenkins Models", published in 1983. The emphasis of the book is on applications. It pulls together time series in the Box-Jenkins tradition that are important for the informed practice of single equation regression forecasting. Special attention is given to possible dynamic patterns - distributed lag responses of the output series to the input series, and the auto- correlation patterns of the regression disturbance.