My Conversation With Harry Markowitz: MPT and Behavioral Economics

One of the best things about what I do is getting to talk with very smart, knowledgeable people, who tell me really interesting things. Over the past 30 years, I’ve certainly been blessed with my share of those conversations. But right at the top of the list is a phone conversation I had just this week with Dr. Harry Markowitz, the father of modern portfolio theory. 

I’d heard that talking with Dr. Markowitz (Harry, as he made me call him) was like talking with Albert Einstein, and while I obviously never had that honor (nor, I’m pretty sure, did the person who told me that, as Dr. Einstein died in 1955), I quickly understood the reference. At 85, the Nobel lauriate still has a razor-sharp mind, able to move from subject to subject, reeling off facts, reference and statistics faster than I could type. As you might imagine, I was particularly interested in getting his response to the current criticisms of modern portfolio theory in the wake of the 2007-’08 mortgage meltdown when many MPT-allocated portfolios failed to preserve capital. What I got was what would qualify at least for a master’s thesis on the measurement of risk and the management of investment portfolios. 

As I hope you’re well aware, modern portfolio theory came out of two papers that Dr. Markowitz wrote in 1952 and 1959. His general idea was how to create the best possible diversification strategies by selecting a collection of assets that together have lower risk than they would individually. But it wasn’t until the 1980s, when the stock market began its long bull run, that allocated, fee-paying client portfolios, based on MPT and Brinson, Hood, Beebower’s 1986 paper, became the foundation of today’s independent advisory industry.  

Dr. Markowitz still consults with advisory firms, such as Loring Ward in San Jose, Calif., helping advisors to maximize returns and minimize risk in client portfolios. “Perhaps the most important job of a financial advisor is to get their clients in the right place on the efficient frontier in their portfolios,” he told me. “But their No. 2 job, a very close second, is to create portfolios that their clients are comfortable with. Advisors can create the best portfolios in the world, but they won’t really matter if the clients don’t stay in them. “ 

I know: I was shocked, too. Like many people, over the years, I’d been led to believe that behavioral economics (the study of how investors actually behave in the real world) was at odds with modern portfolio theory. Not so, said Markowitz. “In fact, Hersh Shefrin [professor at Santa Clara University] once told me that I was the ‘father of modern portfolio theory’ and the ‘grandfather of behavioral economics.’”

Dr. Markowitz explained this surprising (at least to me) revelation this way. “There was a big difference in my views about how and why to use mean variance analysis (that’s academic-speak for MPT) in my 1952 paper and in my ’59 paper. In the ’52 paper the only portfolio constraints I was concerned about were to avoid negative returns. But by 1959, I was concerned about [how investors would react to] the portfolio construction:  For the comfort of the clients, we put constraints on ‘risky sounding stuff,’ such as junk bonds, etc. We knew that might dampen returns, but it was better than if they chicken out in a down market.” 

Dr. Markowitz also revealed that his version of modern portfolio theory included increasing standard deviations “to reflect uncertainty,” and “forward-looking estimates of sample frontiers, to make sure things are coming out plausibly. We all have intuitions about relationships between assets. If we use our best estimates of the current risks, and we end up with a portfolio that seems extreme, maybe there’s something wrong with our estimates. The final step is choosing constraints [on certain assets]: we test them, and if we don’t get what we want, then we think about what’s missing in the constraints.”

This is a view of asset allocation that I hadn’t seen before. It further underscores the need for professional advisors to create allocated portfolios that are right for each client. “The risk/return calculations in a portfolio are always a tradeoff,” Dr. Markowitz concluded. “What’s right for one investor might not be right for another.” 

Close single page view Reprints Discuss this story
This is where the comments go.