選擇權隱含波動率預測性之實證研究

Abstract

The predictive power of option-implied volatility and related implementation issues are discussed in this dissertation. While most of prior studies employ the cubic-spline fitting of implied volatility function to compute the model-free implied volatility, we suggest an alternative procedure that have been used in literature to estimate the risk-neutral density from the market prices of options to calculate it. Our empirical findings suggest that it is more effective to compute the model-free implied volatility from specifying a flexible risk-neutral density for the price of the underlying asset at maturity, rather than fitting the implied volatility function with the cubic spline. The model-free implied volatility generated with the former approach contains more information content for future realized volatility. In addition, this study examines whether the volatility information implied in the term structure of VIX can improve the prediction of realized volatility. We first propose several approaches to compile maturity independent proxies of volatility from the VIX term structure and then investigate the information content of these proxies for future realized volatility. The empirical results on the S&P 500 index show that in terms of both in-sample estimation and out-of-sample forecasting, the proxies representing the information on the VIX term structure are more informative than the single VIX with a particular time to maturity. Our empirical results are robust to alternative model specifications and various forms of volatility.