Forecasting Petroleum Futures Markets Volatility with GARCH and Markov Regime-Switching GARCH Models

Subject Areas : Financial Knowledge of Securities Analysis

مرتضی بکی حسکوئی ¹ (ندارد)
فاطمه خواجوند ² (مسئول مکاتبات)

Received: 2014-03-11 Accepted : 2014-05-05 Published : 2014-10-12

Keywords: Markov Regime-Switching GARCH , GARCH models, volatility, Forecasting, Forecast Evaluation, Fat-tailed Distributions,

Abstract :

In this paper we compare a set of different standard GARCH models with a group ofMarkov Regime-Switching GARCH (MRS-GARCH) in terms of their ability to forecastthe petroleum futures markets volatility at horizons that range from one day to onemonth. To take into account the excessive persistence usually found in GARCH modelsthat implies too smooth and too high volatility forecasts, MRS-GARCH models, wherethe parameters are allowed to switch between a low and a high volatility regime, areanalyzed. Both gaussian and fat-tailed conditional distributions for the residuals areassumed, and the degrees of freedom can also be state-dependent to capture possibletime-varying kurtosis. The forecasting performances of the competing models areevaluated with statistical loss functions. Under statistical losses, we use both tests ofequal predictive ability of the Diebold-Mariano-type and test of superior predictiveability, such as White􀀀s Reality Check and Hansen􀀀s SPA test. The empirical analysisdemonstrates that MRS-GARCH models do really outperform all standard GARCHmodels in forecasting volatility at shorter horizons according to a broad set of statisticalloss functions. At longer horizons standard asymmetric GARCH models fare the best.All this tests reject the presence of a better model than the MRS-GARCH-t in thisresearch