Markov Regime Switching Generalized Autoregressive Conditional Heteroskedastic Model and Volatility Modeling for Oil Returns

In conjunction with the recent alternative models, a wide literature has been established for volatility modeling in finance theory. In this study, we examine return volatility of Brent oil returns through generalized autoregressive conditional heteroskedastic (GARCH), exponential GARCH, Glosten-Jagannathan-Runkle GARCH and Markov regime-switching GARCH (MRS-GARCH) models. As a preliminary test concerning the potential regimes, first, we use modified iterative cumulative sum of squares test in order to examine the existence of breaks in the variance of return series. All volatility models are formed under normal, generalized error distribution and Student s t distributions. According to the Akaike information criterion and Bayesian information criterion values, MRS-GARCH model outperforms all other alternative models. Another interesting result is the failure of the models that formed under normal distribution.