Title: On the prediction of the cross-section of returns in the long run from realized skewness and kurtosis
Authors: Jozef Barunik - UTIA AV CR vvi (Czech Republic)
Josef Kurka - UTIA AV CR, v.v.i. (Czech Republic) [presenting]
Abstract: Large stream of literature has been dealing with factor-investing lately, however many of the supposedly significant asset pricing factors display poor out-of-sample performance. Therefore, we believe that it is important to incline to theory-based factors such as moments of returns distribution. Most popular moment-based factor is volatility, but there is a theoretical intuition and also empirical evidence, that investors are largely interested also in what happens in tails of distribution, i.e. in skewness and kurtosis. Moreover, the traditional asset pricing framework is working with over-the-frequencies aggregated information. If volatility, skewness and kurtosis are the right factors predicting the cross-section of returns, we can still get additional valuable insight by including information about different investment horizons of investors. We achieve this by using multiresolution analysis to construct short-term, medium-term and long-term measures of realized volatility, skewness and kurtosis. Then we construct an asset pricing model comprising these factors which is estimated using Fama-Macbeth type of regression.