The writer is a former investment banker in New York and former head of the Chilean wealth fund investment committee.
The merits of portfolio diversification were probably first acknowledged by the simple rule proposed in the Babylonian Talmud: one-third in real estate, one-third in merchandise (working capital) and the remaining third in liquid assets.
However, a rigorous mathematical argument in favour of diversification was only articulated by Harry Markowitz’s renowned paper, “Portfolio Selection”, which appeared in March 1952 in the Journal of Finance. As we approach the 70th anniversary of its publication, it seems fitting to assess its influence.
Markowitz is universally recognised as the father of modern finance. In fact, before him, portfolio management did not exist as a discipline. Investment decisions were driven mostly by ad hoc rules and gut feeling rather than sound quantitative analyses. The fact that Markowitz is still alive is a testimony of how young this discipline is.
Did Markowitz get everything right? Yes and no. His big contribution was to provide a quantitative framework to analyse the merits of a group of investments (or a portfolio) as a whole. This framework allowed investors to assess the degree and benefits of diversification of a given portfolio. And it formally set out the idea of a risk-return trade-off; in other words, investors wishing to obtain higher returns must be willing to bear more risk.
These concepts have survived well the test of time. In fact, they are still the bedrock on which much of modern finance rests.
A powerful offspring of these two concepts is the idea of the efficient frontier — that investors should aim for the sweet spot of return and risk in portfolio construction. The efficient frontier idea has not only passed the test of time well, it has become — albeit with many modifications — the guiding principle for all serious investors.
Markowitz’s framework, nevertheless, exhibited two weaknesses. One was its reliance on a mathematical construct known as the correlation (of returns) matrix. In essence, this describes the extent to which any two assets move together.
After all these years, financial analysts still do not agree on which is the best way to determine it. Worse yet, small changes in the correlation values result in major differences in the conclusions one derives. In practical terms, structuring an efficient portfolio based on this approach does not work.
Yet, it is another issue with Markowitz’s formulation that has been more problematic — he confused risk and uncertainty. Risk is the possibility that things might go wrong. Uncertainty is not knowing what the future might bring.
By choosing the standard deviation of returns as a proxy for risk he made a conceptual mistake. Standard deviation — a basic statistical metric — focuses on dispersion. It does not distinguish between good and bad scenarios, between getting a return higher than you expected and a return lower than you expected. That is, the standard deviation captures uncertainty, not risk. This conceptual faux pas sent finance and economics down the wrong path for many years.
Although this shortcoming was identified about 30 years ago with the introduction of risk metrics focused on financial losses such as the VaR (Value-at-Risk), the academic community has been slow to embrace these metrics. Practitioners, however, have moved much faster. No serious asset manager relies on the standard deviation of returns or Markowitz’s correlation matrix for anything.
Would this empirical failure amount to an indictment of Markowitz’s ideas? Certainly not. Perhaps they could be forgiven as youthful indiscretions; remember, finance is a young discipline. In sum, we could say that Markowitz’s ideas have been a practical failure (difficult to implement in reality) but a theoretical success (a set of solid principles to guide investment decisions).
This might sound like a harsh judgement; it is not. True, the numerical tools suggested by Markowitz have been progressively replaced by better algorithms, although to some extent this is still work in progress. However, the efficient frontier, the risk-return trade-off, and the merits of diversification have been the lighting rod behind pretty much everything that has happened in finance since 1952.
Great ideas are often marred by implementation difficulties. Think democracy. Markowitz simply had a bunch of great ideas. Financial practitioners should be thankful for that. And the fact that many people are still working hard at implementing them is a testimony of their timeless relevance.