Are Relative Performance Measures Useless?

Abstract

Each year, hundreds of billions of dollars are moved from one investment manager to another. Frequently, these decisions are based on relative performance measurements over periods of 5 to 10 years. The authors use a Monte Carlo model to test the logic of such an approach by constructing managers that have (1) positive skill, (2) no skill, and (3) negative skill versus their benchmarks. The authors show that the typical positive-skill manager suffers a longest underperformance period (defined as the longest month-to-month period during which the returns for the simulated fund trail those of the benchmark) of approximately 10 years during 15-year simulations. Conversely, managers with negative skill enjoy longest outperformance periods of a similar length. The authors also simulate a series of very skilled managers over a 100-year period and show that, on average, there is a 25-year period in which they underperformed. These findings are confirmed by empirical analysis of the realized performance of 5,500 actively managed funds with 15-year track records from the United States, Canada, the United Kingdom, the Eurozone, Europe ex-euro, and Developed Asia ex-Japan markets. The authors conclude that the standard time frames for evaluating the relative performance of investment managers are insufficient to evaluate the skill of a manager with any degree of confidence. The authors cannot state with confidence that relative performance measures are useless, but the authors can state that they are so when they are employed in the customary fashion.

TOPICS: Manager selection, performance measurement, emerging, simulations