Copulas and portfolio strategies: an applied risk management perspective

Theo Berger and Martin Missong

Abstract

ABSTRACT

Modeling multivariate return distributions via copula functions is a common approach in financial risk management. However, evidence of the impact of choosing a particular copula function on different portfolio compositions is still lacking, as empirical studies typically analyze only a single portfolio strategy (eg, equally weighted portfolios) at a time. We evaluate copula models for three different portfolio strategies in a twenty-asset/daily return framework with respect to the accuracy of value-at-risk forecasts. From this risk management perspective,(dynamic) t copulas turn out to be a sensible choice for different portfolio strategies, especially when the trading strategy does not exclude highly volatile assets.