S&P Volatility, VIX, and Asymptotic Volatility Estimates

We examine the efficacy with which the CBOE Volatility Index (VIX) predicts future (30 day forward) S&P realized volatility. We find that the VIX’s accuracy is about 63%. An alternative framework that we present, built on asymptotic distribution theory, predicts this volatility with an accuracy of about 91%. The regression adjusted R-square adjudicates for accuracy. Our methodology outperforms the option-based VIX index because (a) the option market does not fully represent the stock market, and (b) our methodology accounts more comprehensively for idiosyncratic risk. Collectively, our findings suggest it is better to employ the model we present than the VIX index if the objective is to predict S&P realized volatility