Research profile | Challenging assumptions about portfolio management

In collaboration with two colleagues from the London Business School, Associate Professor Lorenzo Garlappi recently completed a project that is challenging modern portfolio management theory.

Garlappi's team looked at the performance of a number of optimized portfolios, which are portfolios comprised of a carefully selected assortment of stocks and bonds that are thought to provide high returns for minimal risk. At the same time he examined one naive portfolio, which is a portfolio composed of stocks and bonds that are selected in an ad hoc manner. 

In academic circles it is assumed that the former assists with the creation of a profitable, balanced portfolio. 

However, Garlappi’s research yielded different results.

He and his team questioned current approaches to portfolio creation through optimization. Garlappi theorized that portfolio optimization might not be as effective as its proponents claim due to the dependence on a portfolio manager‘s ability to make correct assumptions about factors that impact a portfolio performance. These influences include inflation rates or expected returns on an investment.

In order to test his hypothesis, Garlappi implemented 15 different portfolio optimization strategies and one naive strategy. In the latter he constructed a portfolio investing an equal amount of money in each of the available assets. After reviewing a significant amount of data over a 20-30 year span, he was not able to statistically prove that the optimized strategies actually outperform the naive strategy. 

To further test the validity of his results, Garlappi simulated future stock performance, leading him to conclude that one would need at least 5,000 months of return data to reap the benefits of most portfolio optimization strategies.

The research is not sitting well with a number of money managers who use their employment of portfolio optimization strategies to justify their fees. “We’ve been scrutinized and criticized as a result of our findings, which is unfortunate.” Garlappi explains. "Our goal was not to criticize the current work being done, but to highlight the severity of estimation errors in implementing portfolio theory in the hope to advance the field of study.”

Learn more about Professor Garlappi and his latest research interests.