N°24-10: An Averaging Framework for Minimum-Variance Portfolios: Optimal Rules for Combining Portfolio Weights

We propose an averaging framework for combining minimum-variance strategies to either minimize the expected out-of-sample variance or maximize the expected out-of-sample Sharpe ratio. Our framework overcomes the problem of selecting the “best” strategy ex-ante by optimally averaging over portfolio weights. This averaging procedure has an intuitive economic interpretation because it resembles a fund-of-fund approach, where each minimum-variance strategy represents a single fund. In a range of simulations, for a set of well-established strategies, we show that optimally averaging over portfolio weights improves the out-ofsample variance and Sharpe ratio. We confirm the finding of our simulation study on empirical data.