Inference for Stochastic Volatility Models with Semi-Markov Regime Switching

• Activity Number: 84
• Type: Contributed
• Date/Time: Monday, August 9, 2004 : 8:30 AM to 10:20 AM
• Sponsor: Business and Economics Statistics Section
• Author(s): Zhaohui Liu*+ and Nalini Ravishanker
• Companies: University of Connecticut and University of Connecticut
• Keywords: stochastic volatility ; Bayesian inference ; financial time series ; hidden semi-Markov process

Abstract

We discuss stochastic volatility models with hidden semi-Markov regime switches. In particular, the models analyzed include the univariate stochastic volatility model for financial return time series, and a bivariate model for the return and transaction volumes. In addition to the AR structure of the volatility process, it is assumed that the unobserved volatility follows a semi-Markov regime switching process, with the duration time at each state assuming different distributions. Under this framework, the mean and variance as well as duration times of the volatility at different states could vary. These assumptions afford a richer model which is expected to better describe the underlying volatility process due to its influence from different economic forces. Statistical inference for such models using the sampling based Bayesian approach will be described. Model selection and prediction will be described and illustrated.