Forecasting Power of Five CBOE Volatility Indexes for the Price Interval

Young Cheol Jung

This paper evaluates the forecasting power of five CBOE Volatility Indexes (VIX; VXN; OVX; EVZ; GVZ) for the price interval. For the period of June 2008- March 2017, this paper builds a series of one-month ahead price intervals of underlying asset with a band of ± 1 or 2 times of its daily volatility index. Then assesses whether each volatility index is a good estimator for the price interval through comparing its performance to that of the ex-post realized volatility and normal distribution benchmark. Further this paper tests whether the expected shortfall (ES) beyond the price interval forecast of the volatility index is lower than ES of the realized volatility and the benchmark generated by Monte Carlo Simulation. Based upon the test results, this paper suggests that VIX and VXN are robust estimators for the future price interval and can be used in estimating the Value-at-Risk (VaR). VIX and VXN performed well particularly in predicting the upper bound of returns due to the feature of the asymmetric negative return/volatility index relation that exists in the stock markets. The evaluation for OVX, EVZ and GVZ is reserved because their performance was inconsistent and inferior to VIX and VXN.

DOI : https://doi.org/10.21102/irbrp.2018.03.141.12