A Markov-Switching Range-Based Volatility Model with Applications in Volatility Adjusted VAR Estimation

We propose a more flexible range-based volatility model which can capture volatility process better than conventional GARCH approach. Considering the regime switching process is appropriate for dealing the structure change embedded in the time series data. Range-based volatility CARR model with Markov-switching structure can assist us to describe the effect for exogenous shock to market data. After the data fitting and VaR estimation, we conclude that the range-based volatility method is better than the return-based GARCH model in volatility fitting. In particular, incorporating the possibility of regime switching into volatility process can boost the efficiency for VaR estimation. We also present an empirical application for demonstrating our model could characterize the unexpected switching of volatility process. Furthermore, comparing with non-regime switching volatility model, our model outperforms other alternatives on the estimation of volatility adjusted historical VaR.