Can option implied covariance predict cross-sectional return?

I examine the information content of the option-implied covariance between jumps and the diffusive risks of individual stocks in the cross-sectional variation of future returns. This study is the first investigation that studies the predictive power of implied covariance for expected cross-sectional stock returns. The implied volatility that is derived by the jump-diffusion model of Merton (1976) assumes that there is zero correlation between jumps and diffusive volatility. Although there is no significant relationship between the level of implied covariance and the expected return for a stock, this paper indicates that the difference between realized volatility and implied covariance (RV-ICov) can predict future returns. The results demonstrate a significant and negative association between the expected return and the realized volatility-implied covariance spread in both the portfolio level analysis and the cross-sectional regression study. A trading strategy of buying portfolio with lowest RV-ICov quintile portfolio and selling with highest generates positive and significant returns. This RV-ICov anomaly is robustly persistent after controlling for size, book-to-market value, liquidity and systematic risk proportion.