Optimizing Full-Scale Optimization for Asymmetric Dependence

Abstract

Asymmetric dependence in equities markets have been shown to have detrimental effects on portfolio diversification as assets within the portfolio exhibit greater correlations during market downturns compared to market upturns. By applying the Clayton canonical vine copula (CVC) to model asymmetric dependence, we produce a measure of systemic risk across a portfolio of assets. In addition, we use the Clayton CVC to produce estimates of expected returns in an application to higher-moment portfolio optimization and find evidence of an improvement in performance across a range of risk-adjusted return measures and the indices of acceptability.