Optimal Allocation of High Dimensional Assets through Canonical Vines

The widely used mean-variance criteria is actually not the optimal solution for asset allocation as the joint distribution of asset returns are distributed in asymmetric ways rather than in the assumed normal distribution. It is a computationally challenging task to model the asymmetries and skewness of joint distributions of returns in high dimensional space due to their own complicated structural complexities. This paper proposes to use a new form of canonical vine to produce the complex joint distribution of asset returns. Then, we use the utility function of Constant Relative Risk Aversion to determine the optimal allocation of the assets. The importance of using the asymmetries information is assessed by comparing the performance of a portfolio based on the mean-variance criteria and that of a portfolio based on the new canonical vine. The results show that the investors using the forecasts of these asymmetries can make better portfolio decisions than those who ignore the asymmetries information.