Optimality of Naive Investment Strategies in Dynamic Mean-Variance Optimization Problems with Multiple Priors
Kyoto University, Graduate School of Economics Discussion Paper Series No. E-16-004
46 Pages Posted: 5 Jul 2016 Last revised: 7 Jul 2016
Date Written: July 4, 2016
We study dynamic mean-variance optimization problems with multiple priors. We introduce two types of multiple priors, the priors for expected returns and the priors for covariances. Our framework suggests that the global minimum-variance portfolio is optimal when the investor strongly doubts the correctness of the estimated expected returns, and the equally weighted portfolio is optimal when the investor strongly doubts the correctness of the estimated covariances. From the back tests, we find that for some data sets, the strategy that invests in the global minimum-variance portfolio or the equally weighted portfolio considering the market condition is more efficient than the other mean-variance efficient portfolios.
Keywords: robust mean-variance optimization, dynamic portfolio selections, naive diversification, global minimum-variance portfolio, mean-variance efficiency
JEL Classification: G11
Suggested Citation: Suggested Citation