When downside correlations exceed upside correlations, portfolio construction should focus on resilience to market turbulence. The paper “The Myth of Diversification”67 introduces full-scale optimization,68,69 which identifies portfolios with superior correlation profiles by maximizing expected utility rather than directly exploiting correlation asymmetries. This approach ensures that assets diversify each other more effectively during downturns and move together during upturns, unlike mean-variance analysis.
Full-scale optimization accommodates any return distributions and investor preferences, thereby identifying the truly optimal portfolio. In contrast, mean-variance analysis assumes normally distributed returns or quadratic utility, providing only an approximation.
For loss-averse investors with a kinked utility function—where utility changes abruptly at a specific wealth or return level—applying both mean-variance and full-scale optimization can be particularly relevant. This approach is essential for investors concerned about breaching critical financial thresholds.