A Rebalancing Act: Optimal Implementation of the Minimum-Variance Portfolio

Abstract

Minimum-variance investing has increased in popularity since the global financial crisis, with several studies supporting its superior risk-adjusted returns relative to index funds. This thesis examined the effect of two variables on the performance of the minimum-variance portfolio: the frequency of portfolio rebalancing and the duration of the data set used to predict the future covariance matrix. In addition to simulating the portfolio's performance from 1985 to 2015, the impact of transaction costs were analyzed for each variation of the minimum-variance portfolio. Net of transaction costs, this study finds that less frequent rebalancing using a long duration data set produces the minimum-variance portfolio with the greatest alpha. The conclusion is that an optimal minimum-variance strategy, under realistic cost assumptions, is able to significantly increase risk-adjusted returns for an investor whose capital is concentrated in index funds.