EMPIRICAL DISTRIBUTIONS OF STOCK RETURNS AND APPLICATIONS IN VALUE AT RISK

摘要

This paper examines the best distribution for stock returns. Normal distribution is the basic assumption we assume when understanding and analyzing different kinds of models. However, this does not always work when describing stock returns distribution because they always have fat tails that cannot be explained. We used several databases and distributions to find out the most suitable distribution for stock returns by examining them through A-D test and Value at Risk application. The result is that Generalized Pareto Distribution is the most suitable one statistically.