Volatility forecasting of crude oil market: Can the regime switching GARCH model beat the single-regime GARCH models?

By Yue‐Jun Zhang, Ting Yao, Ling‐Yun He, R. Ripple. International Review of Economics and Finance, 2019

80 citations recorded by the source corpus.

GARCH-type models are frequently used to forecast crude oil price volatility, and whether we should consider multiple regimes for the GARCH-type models is of great significance for the forecasting work but does not have a final conclusion yet. To that end, this paper estimates and forecasts crude oil price volatility using three single-regime GARCH (i.e., GARCH, GJR-GARCH and EGARCH) and two regime-switching GARCH (i.e., MMGARCH and MRS-GARCH) models. Furthermore, the Model Confidence Set (MCS) procedure is employed to evaluate the forecasting performance. The in-sample results show that the MRS-GARCH model provides higher estimation accuracy in weekly data. However, the out-of-sample results show the limited significance of considering the regime switching. Overall, our results indicate that the incorporation of regime switching does not perform significantly better than the single-regime GARCH models. The findings are proved to be robust to both daily and weekly data for WTI and Brent over different time horizons.

Subfields: Analytical Chemistry, Economics and Econometrics

View Paper: https://www.semanticscholar.org/paper/079ae03b8792a61d055c714031076701b9883f87