S&P volatility, VIX, and asymptotic volatility estimates

• Published in: Finance research letters Vol. 51; p. 103392
• Main Authors: Bonaparte, Yosef, Chatrath, Arjun, Christie-David, Rohan
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.01.2023
• Subjects: Asymptotic distribution theory; Realized volatility; VIX; G12; VIX; Realized volatility; Asymptotic distribution theory; G10
• ISSN: 1544-6123; 1544-6131
• DOI: 10.1016/j.frl.2022.103392

•The efficacy of the VIX as a predictor of 30-day forward S&P 500 volatility is between 20% and 25%, depending on sampling period.•Our framework, built on asymptotic distribution theory (AVE), predicts this volatility with an accuracy between 90% to 95%.•The adjusted R-square adjudicates for accuracy. Other goodness-of-fit measures corroborate this evidence.•Why is this? We suggest that this outcome is underpinned by the fact that (a) options price behaviors do not adequately reflect stock market volatility patterns, and (b) our methodology accounts more comprehensively for idiosyncratic risk.•What about implications? The evidence from this study has important implications. One, the VIX, which relies on current prices of options will provide poor and perhaps misleading guidance of future volatility. Two, standard statistical frameworks like the one we employ outperforms the VIX, and three, market participants should regard the VIX as a measure of current sentiment rather than as a volatility-forecasting tool. We examine the efficacy of the VIX as a predictor of 30-day forward S&P 500 volatility. We find that its accuracy hovers between 20% and 25%, depending on sampling period. An alternative framework, built on asymptotic distribution theory (AVE), predicts this volatility with an accuracy between 90% to 95%. The adjusted R-square adjudicates for accuracy. Other goodness-of-fit measures corroborate this evidence. We suggest that this outcome is underpinned by the fact that (a) options price behaviors do not adequately reflect stock market volatility patterns, and (b) our methodology accounts more comprehensively for idiosyncratic risk.