The practice of portfolio optimization and diversification has finally caught up with theory—more than 60 years after Nobel-Prize-winning economist Harry Markowitz devised the concept.
The conclusion is based on an analysis, which appears in the European Journal of Operational Research, by Petter Kolm, a professor at New York University’s Courant Institute of Mathematical Sciences, Reha Tütüncü, managing director at Goldman Sachs Asset Management, and Frank Fabozzi, a professor at France’s EDHEC Business School.
Their analysis may be read here.
Markowitz’s paper, “Portfolio Selection,” published in 1952, and his subsequent book, Portfolio Selection: Efficient Diversification (1959), posited that risk in a portfolio depends on the individual variances of the return on different assets as well as covariances of all assets. In other words, the risk of an asset is not the risk of each asset in isolation, but, rather, the contribution of each asset to the risk of the portfolio as a whole.
In effect, this means that risk cannot be totally eliminated from any portfolio. However, Markowitz’s theory offered some guidance for investment managers. Specifically, it outlined how wealth can be optimally invested in assets that differ with regard to their expected return and risk—and, thereby also how risks can be reduced.
Markowitz was awarded the 1990 Nobel Prize in Economics, which he shared with Merton Miller and William Sharpe, in recognition of his work.
However, when investment managers sought to put Markowitz’s theory into practice in subsequent decades, the results were subpar, leading many to question its validity.
Yet, as Kolm and his colleagues describe, the problem was not the theory, but, rather, how to use it in practice.
“By the early 1990s, it became more widely clear why Markowitz’s portfolio optimization didn’t work well in practice: it is very sensitive to small changes and estimation errors in the model inputs,” says Kolm.
Notably, he explains, portfolio managers that don’t use care in estimating the inputs (the expected security returns and covariances) may grossly miscalculate their optimal portfolio allocations —the old “garbage in, garbage out” dynamic.
However, in the past two decades, investment professionals have developed methods to improve their estimation of inputs —and enhanced their ability to make more accurate portfolio allocations.
In their analysis, the authors discuss a number of practical developments in portfolio construction, including the inclusion of transaction costs, investment constraints, and sophisticated multi-period forecasting models.
However, they caution that a better match between Markowitz’s theory and investment practice does not mean risk is fully understood.
“Markowitz’s framework provides one way of measuring risks—but not extreme risks,” Kolm observes. “Extreme forms of risk, also referred to as ‘tail risk’—that’s an area we don’t quite understand how to handle practically.”