Evidence of excess volatility based on a new robust volatility ratio

摘要

PurposenThe purpose of this paper is twofold: first, to propose a new robust volatility ratio (RVR) that compares the intraday highlow volatility with that of the intraday openclose volatility estimator; and second, to empirically test the proposed RVR on the cross-sectional (CS) average of the constituent stocks of Indias BSE Sensex and USs Dow Jones Industrial Average index to find the evidence of excess volatility.nDesign/methodology/approachnThe authors model the proposed RVR by assuming the logarithm of the price process to follow the Brownian motion. The authors have theoretically shown that the RVR is unbiased in the case of zero drift parameter. Moreover, the RVR is found to be an even function of the non-zero drift parameter.nFindingsnThe empirical results show that the analysis based on the RVR supports the existence of excess volatility in the CS average of the constituent stocks of Indias BSE Sensex and USs Dow Jones index. In particular, the authors have observed that the CS average of individual constituent stocks of BSE Sensex is found to be more excessively volatile than the USs Dow Jones index during the period of the study from January 2008 to September 2016, based on multiple k-day time window analysis.nPractical implicationsnThe study has implications for the policy makers and practitioners who would like to understand the volatility behavior in the asset returns based on the RVR of this study. In general, the proposed model can be used as a specification tool to find whether the stock prices follow the random walk behavior or excessively volatile.nOriginality/valuenThe authors contribute to the existing volatility literature in finance by proposing a new RVR based on extreme values of asset prices and absolute returns. The authors implement the bootstrap technique on RVR to find the estimates of mean and standard error for multiple k-day time windows. The RVR can capture the excess volatility by comparing two independent volatility estimators. This is possibly the first study to find the CS average of all the constituent stocks of BSE Sensex based on the RVR.