Higher Moment CAPM in the German Stock Market
- Abstract
The inclusion of statistic moments of a higher order than mean and variance into the Capital Asset Pricing Model (CAPM) has been the subject of economic discussion since empirical research had shown that the majority of asset returns from several stock markets are not normally distributed. To adequately describe asset returns it seems appropriate to consider skewness and kurtosis of return distributions. Theoretical extensions of the CAPM have been developed that include skewness (three-moment CAPM) and skewness along with kurtosis (four-moment CAPM).
Most empirical investigations that test these models have been examined for the U.S. American stock market. In this master thesis research has been conducted for the German stock market (HDAX). Thereby, weekly asset returns from 2000 to 2015 form the basis of the analysis as other studies commonly use monthly data. It is investigated whether the explanatory power of the CAPM becomes better when higher moments are included. The investigation procedure follows the methodology by Fama and MacBeth (1973). The results of the cross-sectional analysis reveal that no variable has a statistically significant influence on asset returns. The explanatory of all models, even the traditional CAPM, is not satisfying, but becomes slightly better when skewness and especially kurtosis are added.