MCMC interweaving strategy for estimating stochastic volatility model and its application

Efficient estimation for a stochastic volatility (SV) model has been actively pursued in recent years. In this paper, a new Markov chain Monte Carlo (MCMC) algorithm based on a combination of Kalman filtering and the auxiliary sufficiency interweaving strategy (ASIS) is studied. Compared to other MCMC strategies like Stan algorithm ("Rstan") and the Gibbs algorithm ("R2Winbugs"), it is shown from finite-sample studies that the MCMC interweaving strategy improves both the estimation accuracy and the computation speed. We also applied the final selected algorithm, i.e., KMA5 algorithm, to the return rate of The Chinese CSI 300 Index (the Shanghai and Shenzhen 300 stock index in China), which further verifies the validity and accuracy of the new method.