The following are in response to Michael Edesess and Kwok Tsui’s article, How Many Monkeys Does it Take to Find a Successful Strategy?, which appeared last week:
Dear Editor,
There are a lot of good points in this article, but as with all such arguments, one demonstrably false element taints the entire work.
A short study of Complexity Theory (aka Chaos Theory) will reveal that no matter how many monkeys are hammering on typewriters, the works of Shakespeare will never appear. Yes, short words and even the occasional sentence that makes sense will appear, but that is about as far as it goes. Beyond that the fundamental chaos of the universe prevents any further random typing from appearing as intelligent discourse.
We do use continuously compounded returns — i.e., log-returns. When constructing MVO-Markowitz portfolios, the Markowitz formulas from his 1952 paper consistently include value and eschew growth-oriented asset classes in domestic stock asset classes. We wind up with a reasonable total return and relatively low variance, just as Markowitz predicted.
We have hitched our wagon to a combination of Markowitz's "Portfolio Selection," reversion to the mean and Pareto's rule. Using mean variance optimization based on historic real returns in 23 identifiable asset classes, while at the same time putting an upper limit (or excluding) asset classes that are significant above their long-term mean 30-year average, seems to have worked remarkably well. Will it work into the future? "Generations" by Strauss and Howe states that if we look back at least 30 years to get the real data, then over the next 30 years there is a high probability that the same behavior patters will emerge, as we tend to repeat our general behavior patterns about every 92 years.
Read "The Panic of 1907," and you will note that the similarities between 1907 and 2008 are eerie. The crisis and panic of 1837 again bear a striking resemblance to what we are seeing today. The generational hypothesis in Strauss and Howe's book suggested to me in the mid 1990s that a serious financial crisis would occur around 2010, and it did. The next is due around 2030.
If a rigorously stated mathematical formula could determine the future of the markets, then computer-driven portfolios would dramatically outperform human-selected portfolios, but that is not the case. Somewhere in the blend of mathematical analysis and human judgment resulting from a study of history there is a middle way that we believe has the greatest potential for success.
The fly in the ointment in the article is that the authors appear to be all-negative, all the time. Those of us who actually build real portfolios to fit individual needs are left with a powerfully expressed argument that effectively states, "Nothing works."
When you throw in a demonstrably false claim in the midst of arguing that others have not been sufficiently rigorous, your argument deflates rather suddenly.
Sincerely yours,
Jeff McClure
The Personal Wealth Coach
Salado, TX
Michael Edesess replies:
In my first draft of the article, I included right the following parenthetical remark after the monkeys and Shakespeare reference: "(Though this paints an evocative picture, the proposition can actually be proven mathematically to be untrue.)" The mathematical proof that I recall is from metamathematics, not chaos theory. However, someone who saw the remark told me: "I suggest you check your facts as it has been done: see here". I think there must be something wrong with what Mr. Jesse Anderson — the person who claimed to have done it — did, but I didn't want to spend time investigating it, so I left the parenthetical remark out.
