Title: Quantiled moments by Cornish-Fisher expansion and its applications
Authors: Ningning Zhang - The University of Hong Kong (Hong Kong) [presenting]
Ke Zhu - University of Hong Kong (Hong Kong)
Abstract: The conditional moments play an important role in many financial applications. However, some parametric models for studying the conditional moments may exit model mis-specification problems and computation burden. To avoid these problems, a novel simple method is proposed to learn the conditional mean, variance, skewness, and kurtosis by using the classical Cornish-Fisher expansion. Our method provides an easy-to-implement non-parametric way to estimate the so-called quantiled moments, based on a sequence of estimated conditional quantiles. Some regression-based Wald tests are proposed to check the validity of our quantiled moments. Simulations show that the quantiled moments could be good proxies for their unobserved counterparts, and they exhibit robust performances across the choices of quantile estimation method and quantile level. As two important applications, the quantiled moments unveil unknown news impact functions and interactive effects among the conditional moments.