The Use of Copulas in Asset Allocation

This study critically examines the limitations of the mean-variance criterion, developed by Markowitz, in portfolio allocation-particularly when returns deviate from Normality. Since expected utility cannot always be accurately represented under non-Normal return distributions, the author explores whether the loss of optimality in using the mean-variance approach is significant or negligible. Through a comparative analysis of optimal portfolio compositions, the research evaluates the cost of the Markowitz allocation versus strategies based on copula models (Normal, Student-t, Clayton, Gumbel, Frank, mixed, and Canonical Vine copulas). Portfolios of two or more assets and various index combinations are used to assess whether copula-based models enhance investor utility and returns. The study also investigates whether incorporating higher-order moments and co-moments (up to the fourth) can approximate copula-based optimization. This research is particularly relevant for investment fund asset managers and academic researchers seeking advanced tools in financial econometrics.