EcoSta 2021: Start Registration
View Submission - EcoSta2021
A0202
Title: Return predictability from variance differences: A fractionally co-integrated analysis Authors:  Xingzhi Yao - Xi\'an Jiaotong Liverpool University (China) [presenting]
Marwan Izzeldin - Lancaster University Management School (United Kingdom)
Zhenxiong Li - Soochow University (China)
Abstract: Decomposing the realized and implied variances into upside and downside semi-variances, this paper examines the fractional co-integration for each pair of variances is examined. A positive link is revealed between the strength of the co-fractional relation and the return predictability afforded by the corresponding variance differences. Such a linkage is attributable to the common long-memory component in the fractionally co-integrated system, representing the volatility-of-volatility factor driving the variance risk premium. Despite the evidence for the presence of fractional co-integration between the upside and downside realized (implied) semi-variances, their difference is only weakly associated with future returns at low frequencies, and the predictive power dissipates over long horizons. The empirical results are further verified in a simulation study where the issue of the limited number and range of option strikes is alleviated in the construction of implied variances.