VIX Option Pricing With Detected Jumps

ABSTRACT This article examines VIX option pricing using a direct modeling approach that emphasizes the dynamics of the VIX by incorporating identified jumps. Using high‐frequency intraday VIX data, we isolate jumps from continuous movements, and integrate realized jump variation and bipower variation, respectively, into the dynamics of conditional variance and jump intensity. We subsequently derive a closed‐form pricing formula for VIX options and conduct a comprehensive evaluation of the model's pricing accuracy. Empirical results indicate that the jump‐based model consistently outperforms models based on conventional realized variance and the classic Heston‐Nandi GARCH framework, both in‐sample and out‐of‐sample. Our findings demonstrate that decomposing VIX variation into jump and continuous components significantly improves VIX option pricing performance.