ARIMA models, the steady state of economic variables and their estimation

Abstract

This paper presents a procedure to breakdown the forecast function of an ARIMA model in terms of its permenent and transitory components. The result throws some ligth on the structure of ARIMA models and its interpretarion proves useful in economic terms. Indeed, the permanent component of the forecasting function for a base period t ·is a consistent estimate of the equilibrium level or' steady state path at which the corresponding variable is tending at each time t

and the transitory component describes how the approach towards the permanent component tends to arise