The Skewed Perception of the Distribution of Stock Returns

Abstract: This thesis investigates the sensibility of the often used simplifications of how stock returns behave in financial models by studying Swedish stock returns using data from 1979 to 2012. The data is tested for normality by using Jarque-Bera test in several steps and exogenous factors are examined for significant impact on the skewness and kurtosis of the stock returns using a non-parametric test developed for this particular purpose. The results clearly show that stock returns are not normally distributed overall. Furthermore, both skewness and kurtosis are significantly affected by exogenous factors. There is a "June effect", where the kurtosis is significantly higher during June than other months. The implication of these findings is that financial models lack in dimension by not incorporating skewness and kurtosis. It is evident that the underlying assumptions of financial models are too simplified in some cases. The thesis brings attention to the need to better understand how skewness and kurtosis varies across stocks and affect investor utility, which clearly is a subject not fully understood as of today.