Generalised autoregressive models with conditional heteroscedasticity: An application to financial time series modelling

Authors

Shelton Peiris, University of Sydney
David E. Allen, Edith Cowan UniversityFollow
Udara Peiris, University of New South Wales

Document Type

Conference Proceeding

Publisher

University of Woollongong

Faculty

Faculty of Business and Public Management

School

School of Accounting, Finance and Business Economics

RAS ID

3847

Comments

Peiris, S., Allen, D. E., & Peiris, U. (2005). Generalised autoregressive models with conditional heteroscedasticity: An application to financial time series modelling. Proceedings of Workshop on Research Methods: Statistics and Finance. (pp. 75-83). Sydney, Australia. University of Woollongong.

Abstract

This article considers a new class of time series models generated by GAR (generalised autoregressive) models with conditional heteroscedasticity. In particular this class combines the popular GAR models with GARCH ( generalised autoregressive conditional heteroscedastic) models. Some statistical properties of the underlying process are given. An algorithm for parameter estimation based on the MLE procedure is given. A simulation study is presented to compare the finite sample properties of parameter estimates. An example is added to illustrate the theory.