Consistency of Maximum Likelihood Estimators for the Reduced Regime-Switching GARCH Model
2005 (English)Report (Other academic)
The regime-switching GARCH model combines the idea of Markov switching and GARCH model, which also extends Hidden Markov models. The statistical inference for this model, however, is rather difficult because the observations depend on the whole regime history. In this paper, we consider a reduced regime-switching GARCH model, that is, the past regimes are integrated out at every step and observations then depend only on the current regimes. We prove the consistency of maximum likelihood estimators for this model. Simulation studies to illustrate consistency, asymptotic normality of the proposed estimators and a model specification problem are also presented.
Place, publisher, year, edition, pages
2005. , 19 p.
, Research Report, Centre of Biostochastics, SLU, ISSN 1651-8543 ; 2005:2
GARCH model, regime-switching, MLE, consistency, asymptotic normality
Probability Theory and Statistics
IdentifiersURN: urn:nbn:se:umu:diva-65441OAI: oai:DiVA.org:umu-65441DiVA: diva2:604091