From Markowitz 1.0 to Markowitz 2.0 with a Detour to Postmodern Portfolio Theory and Back

Abstract

In 1994, Larry Siegel and I published an article in this journal in response to an article that claimed that (1) mean-variance optimization (MVO), a vital part of modern portfolio theory (MPT) as developed by Harry Markowitz, is flawed, and (2) postmodern portfolio theory (PMPT) is a superior method for constructing portfolios, thus making MVO obsolete. Although we defended MVO and pointed out the limitations of PMPT, in retrospect, I believe the most important point that we made was that portfolio theory should be more broadly understood to mean any portfolio construction model that presents the investor with a trade-off between reward (however defined) and risk (however defined). Hence, MVO and PMPT are both special cases of portfolio theory. Years later, Sam Savage and I developed such a general framework, in which the user of our model can select a measure of reward and a measure of risk and have a wide choice of return distribution models. This framework, which we dubbed Markowitz 2.0, I believe is a logical progression from MVO, which we dubbed Markowitz 1.0.