Capital Ideas^™ 
Investment insights from Capital Group

June 22 was a sad day in the asset management world with the passing away of Harry Markowitz. The Nobel Prize-winning American economist was best known as the father of Modern Portfolio Theory, which says that the performance of an individual stock is not as important as the performance and composition of an investor's entire portfolio.

In the asset management world, he is mostly remembered for inventing mean-variance optimization. But this tool was just one component in his broader vision of finance and investing. That vision took shape more than seventy years ago, in two seminal papers he published in 1952, his annus mirabilis.

The paper we remember was titled Portfolio Selection which proposed that the tradeoff between mean and variance should be the basis for portfolio construction. While the tools he described have been widely adopted, many of his underlying ideas and insights remain under-appreciated.

Some of these are technical. Much of the paper is a close analysis of what efficient portfolios look like, and how the efficient portfolio evolves as risk increases – the basis of his famous critical line algorithm. It’s a wonderful case study of understanding what a model is actually saying, rather than simply feeding inputs into a black box. And in the process, he showed how to explain a sophisticated algorithm intuitively, without needing to resort to “complicated mathematical statements,” setting a standard for exposition ever since.

Other ideas are more fundamental. He pointed out that investors need to seek what he said was the “right kind” of diversification, for the “right reason.” Simply owning lots of securities is not enough. Even diversifying across many industries is not enough. One must diversify across industries with different economic characteristics. That is, to build a genuinely diversified portfolio, you need to understand the economic fundamentals of a business.

Markowitz also had a practical view of how to estimate expected return (µ)_i and risk (σ)_ij, the key inputs to mean-variance optimization. He said investors should “combine statistical techniques and the judgment of practical men.” Using historical data was not enough. Instead, he advocated a “probabilistic reformulation of security analysis” – in other words, a fundamental investment approach to analyzing firms and selecting securities.

The other paper Markowitz published in 1952 was The Utility of Wealth. This is less well remembered, but it is no less important. At the time, economists had started actively applying utility theory to explain decision-making under uncertainty – and in particular, investing. But it was not clear what real people’s “utility functions” looked like, thus, it was difficult to understand how individuals weigh the tradeoff between risk and return.

Markowitz tested economists’ theoretical arguments against real-world observation, experiment and a priori reasoning about people’s behavior. He showed that some key assumptions being made by theorists at the time were wrong. Yet practitioners continue to make many of those assumptions today.

In the process, he introduced some useful concepts, like “customary wealth” the idea that investors care about the level of wealth they are used to. Most models look only at an investor’s current financial position: their present wealth. In reality, it matters how they got there, and what they were used to before.

He made other uncomfortable observations. In particular, it’s commonly assumed – both in 1952 and today – that investors’ utility functions are convex: that they always bend downwards, for all levels of wealth. Mathematically, that’s a convenient assumption.

But as Markowitz pointed out, this amounts to assuming that risk-taking is due to the “fun of participation”. This is a strange assumption for a fiduciary to make. Yet implicitly, it’s still pervasive in our industry.

Seventy-year-old papers might seem dated. But the classics still have important lessons for us. Harry Markowitz did more than invent sophisticated tools. He had a vision of how to use them in the real world, for the benefit of real people, alongside insights from traditional security analysis.