Asymptotic Properties of the Maximum Likelihood Estimator of the Mixture Autoregressive Model with Applications to Financial Risk

• Activity Number: 173
• Type: Topic Contributed
• Date/Time: Monday, August 10, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Risk Analysis
• Author(s): Mary Akinyemi* and Georgi N. Boshankov
• Companies: University of Lagos and University of Manchester
• Keywords: Mixture autoregressive models ; maximum likelihood estimator (MLE) ; financial risk ; Asymptotic Properties

Abstract

Mixture autoregressive models provide a flexible framework for modelling time series. These models capture conditional heterogeneity, multi-modality, skewness, kurtosis and heavy tails using only standard distributions as building blocks. We show that the maximum likelihood estimator (MLE) of this class of models is consistent and asymptotically normal. We also give applications to estimation of financial risk.