Whether average idiosyncratic volatility has recently risen, whether it is a good predictor for aggregate market returns and whether it has a positive relationship with expected returns in the cross-section are still matters of active debate.
Professor of Finance, EDHEC Business School and research fellow at CIREQ and CIRANO
PhD in Finance Candidate, EDHEC Business School and Research Assistant at EDHEC-Risk Institute
Professor of Finance, EDHEC Business School and Scientific Director of EDHEC-Risk Institute
We revisit these questions from a novel perspective, by taking the cross-sectional variance of stock returns as a measure of average idiosyncratic variance. Two key advantages of this measure are its model-free nature and its observability at any frequency, which allows us to present new results on the properties of daily idiosyncratic volatility series. Through central limit arguments, we formally show that the cross-sectional dispersion of stock returns can be regarded as a consistent and asymptotically efficient estimator for idiosyncratic volatility. We empirically confirm that the cross-sectional measure provides a very good proxy for average idiosyncratic risk as implied by standard asset pricing models and that it predicts well aggregate returns, especially at the daily frequency. The predictability power of idiosyncratic risk is further increased when adding a measure of cross-sectional skewness to the cross-sectional variance factor. We finally provide evidence that idiosyncratic risk is a positively rewarded risk factor.
|Research Cluster :||Finance|