A Study On Stability Of Conditional Variances For GARCH Models With Application

Abstract

In this paper we study the stability conditions of GARCH models and we find these conditions by using a local linearization technique. In addition we study the stability of conditional variances predictions where these predictions converge to an unconditional variance of the model after finite number of prediction steps. We apply this study to a time series represent the monthly mean temperature in Mosul city for years 1978-2011 and we find that ARCH (1) model is the fit model for prediction.