Normal versus Student in Measuring Value at Risk. An Empirical Bayesian Overview

It is well known that financial returns exhibit positive kurtosis and flat tails. The Student model has been proposed in the literature as most adequate in treatment of financial problems than the Normal model. One of those problems is measuring Risk in a given portfolio using the Value At Risk (VAR). We apply Empirical-Bayesian (EB) techniques to the Normal and Student models to obtain VAR. The resultings VAR are easy to calculate and that from Student model has a pretty interpretation in terms of kurtosis. Both VAR were applied to a conjunct of diary observations of returns in six international indexes during the last decade. The results shows that the Student model is better than the Normal, but also shows that exists characteristics other than kurtosis in the financial returns that affect the goodness of results.