U.S. equity indexes constructed randomly by monkeys — or more precisely 10 million virtual monkeys — would have outperformed the equivalent cap-weighted index, according to a joint study by City University London's Cass Business School and Aon Hewitt.
Based on data between 1968 and 2011, researchers programmed what they dubbed “the stock-picking abilities of a monkey” to randomly construct portfolios from 1,000 stocks, repeated 10 million times over each of the 43 years. “Our initial surprise was that all of these alternative equity indexes outperformed the market-cap-weighted index,” which currently tracks trillions in assets, said Nick Motson, lecturer in finance at Cass and co-author of “Heuristic and Optimised Weighting Schemes.”
During the same period, however, most of the virtual monkeys failed to consistently outperform smart beta indexes, particularly low-volatility and fundamentally weighted indexes, suggesting smart-beta strategies did not achieve superior risk-adjusted returns by luck alone. Mr. Motson also emphasized the study “in no way is saying that our monkeys outperformed active managers.”
“Compared to active managers, who are constrained in all sorts of ways and who are fired if they (consistently) underperform a benchmark, our monkeys don't care about job security and therefore are not constrained by what stocks they pick,” Mr. Motson said. However, Cass is considering further studies that put constraints on the monkeys to see the effects on performance. — THAO HUA
This article originally appeared in the April 15, 2013 print issue as, "Here's another way to build equity indexes".