
Abstract:
In this work we consider the Capital Asset Pricing Model under the multivariate t-distribution with finite second moment. This distribution, which contains the normal distribution, offer a more flexible framework for modeling asset returns. The main goal of this study is to consider the tests of mean-variance efficiency on a given portfolio using the likelihood-ratio, Wald, score and gradient statistics. We provide analytical expressions for the score function and the Fisher information matrix. The results are illustrated by using a set of six markets in Latin American countries. The mean-variance efficiency of the EM index is analyzed using an International Capital Asset Pricing Model. Our main conclusion is that the t-distribution shows a better fit and that the mean-variance efficiency of the EM index cannot be rejected.