Title: Modeling VIX Index dynamics with Bayesian methods and intraday data
Authors: Milan Ficura - University of Economics in Prague (Czech Republic) [presenting]
Jiri Witzany - University of Economics in Prague (Czech Republic)
Abstract: A methodology for the Bayesian modelling of the VIX Index dynamics is presented. The proposed model captures most of the empirically observed properties of the VIX Index time series, including stochastic volatility (of the volatility), self-exciting jumps, asymmetry of the returns, as well as the long-memory mean-reversion of the VIX index towards its long-term levels. In order to facilitate inference on the model, high-frequency power-variation estimators of the VIX volatility and the S\& P500 volatility are used as additional sources of information. MCMC algorithm is proposed for the estimation of the model parameters and the latent state variables in the in-sample period, while a particle filter is used to provide out-of-sample forecasts. The ability of the model to capture VIX Index dynamics and approximate its empirical probability distribution under different market conditions is assessed on the historical VIX index time series with generally positive results.