RE: “Monkey Business. Here's another way to build equity indexes,” Frontlines, April 15.
The study described in this article is flawed in ways that similar previous studies are flawed. The easy way to recognize flawed portfolio simulations is that the results are not real world, but instead of viewing the discrepancies as flaws, the study views them as revelations. In other words, the conclusions are based on a faulty methodology rather than anything real world.
Portfolio simulations do have real value, but at a minimum they need to be constructed to follow two important real-world rules:
Macro-consistency: The aggregation of the simulated portfolios must be the total market. You can construct equal-weighted portfolios to capture common practice, but large companies must be in more portfolios than small companies.
Trading: Real managers trade. Very few buy and hold for 44 years, the period covered by the study.
Properly constructed portfolio simulations are excellent performance evaluation barometers. You can use science to test the hypothesis “performance is good” by creating all of the portfolios the manager might have held, and by contrasting the actual return to the range of returns earned by the hypothetical monkeys. This approach eliminates the biases in peer groups. Pensions & Investments has written about this important application on several occasions, including:
“Consultant offers a way to end classification bias. PIPODS praised, but little demand is seen,” Aug. 18, 2003.
“Classifying Performance: Firms take different approaches to answer hedge fund questions,” Nov. 29, 2004.
PPCA Inc and Target Date Solutions
San Clemente, Calif.