Performance evaluation of optimal portfolios during market crises

Abstract

In studies on portfolio selection problems, classical portfolio optimization models have been expected to generate strategies that perform better than the random investments made by people who lack professional investment knowledge. In this paper, to test the efficiency of portfolio optimization approaches, we evaluate the historical performance of the strategies obtained by applying classical portfolio optimization models. To undertake the evaluations, in our empirical analysis, we apply two individual samples that cover the period of the global finan- cial crisis, 2007–2009, and the period of COVID–19 pandemic, respectively. According to the analysis results, we find that minimizing the chosen risk measure/maximizing the chosen performance ratio in the in-sample period does not guarantee the lowest/highest value of the strategy portfolio in the out-of-sample period.