VIX Spike Characteristics as Predictors of Short-Term SPY Reversion: Evidence from 82 Events, 2004–2026

Abstract

This study examines whether the cause of a VIX spike predicts the magnitude and consistency of short-term SPY reversion. Using 82 single-day VIX spikes of 5 or more absolute points from 2004 to 2026, each event is classified by cause (Federal Reserve decisions, macroeconomic data releases, geopolitical events, and corporate shocks) and spike tier (Moderate: 5–10 points, Significant: 10–20 points, Severe: 20+ points). Binomial significance testing is applied to SPY reversion rates at 1, 3, 5, and 7 trading days following each spike. Aggregate reversion rates of 54.9% to 57.0% across horizons are not statistically significant (p=0.075). However, Significant-tier spikes (n=18) produce reversion rates of 66.7% to 83.3%, highly significant at p=0.00066. Geopolitical spikes (n=21) exhibit 81.0% reversion at 5 days (p=0.00070), compared to 48.4% for Moderate-tier macroeconomic data spikes — a divergence that persists after controlling for spike magnitude. These findings suggest that cause classification adds meaningful predictive signal beyond VIX level alone.