Forecasting the time-varying beta of UK and US firms : evidence from GARCH and non-GARCH models

Abstract

This thesis investigates the forecasting ability of four different GARCH models and the Kalman filter method in forecasting the time-varying beta. The four GARCH models applied are bivariate GARCH, BEKK GARCH, GARCH-GJR and GARCH- X; and the Kalman filter approach is the representative of non-GARCH models. The study provides comprehensive comparison analyses on the modelling ability of alternative methods, with an emphasis on their forecasting performance. The study is accomplished by using daily data from UK and US stock market, ranging from January 1989 to December 2003. According to estimation results, GARCH models are successful in capturing the time- varying beta. Moreover, bivariate GARCH and BEKK GARCH outperform, other models in terms of out-of-sample beta forecasts. Kalman filter is found to be less competent in constructing time dependent beta. However, measures of forecast errors overwhelmingly support the Kalman filter approach in terms of out-of-sample return forecasts. Among the GARCH models, GJR model appears to provide somewhat more accurate forecasts than other GARCH model. This study contributes to financial economics research on modelling time-varying beta by providing empirical evidence from UK and US stock markets. These empirical results are helpful for both market participators and academic researchers in their decision making or research development.

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata