Title: Estimating and forecasting long-horizon dollar return skewness
Authors: Kevin Aretz - The University of Manchester (United Kingdom)
Jiayu Jin - The University of Manchester (United Kingdom) [presenting]
Yifan Li - The University of Manchester (United Kingdom)
Abstract: The aim is to develop a parametric estimator of the physical skewness of an assets discrete (i.e. dollar) return over long horizons from the assumption that the assets value can be modelled using a stochastic process from the affine stochastic volatility (ASV) model class. Taking compounding and return dependence effects into account, we demonstrate that our estimator is close to unbiased and efficient, setting it apart from other recent estimators. In a further contrast to those other estimators, it also lends itself naturally to forecasting skewness. Applying our estimator to some representative stock indices, we show that the skewness of long-horizon dollar returns is far less extreme than suggested in the current literature.