Forecasting VIX using realized EGARCH model with dynamic jumps

This paper proposes a realized EGARCH model with dynamic jumps (REGARCH-Jump) incorporating high-frequency information and time-varying jump intensity to forecast VIX. We derive the risk-neutral dynamic of the REGARCH-Jump model as well as the corresponding model-implied VIX formula. Our empirical results show that the REGARCH-Jump model outperforms a range of competing models in terms of VIX forecasting performance.