Monday, November 7, 2011

Exploiting Stock Market Anomalies

Much time and effort goes into searching for stock market inefficiencies that can be exploited for profit. Former string theorists work together developing algorithms to comb through historical data looking for persistent patterns. The problem is that once we find a strategy to exploit an anomaly and it becomes widely-known, it stops working. There is one pattern that I bank on, though.

There are those who try to make money from momentum effects and others who believe in “sell in May and go away” until November because stocks have performed poorly in summer. I don’t trust these approaches because they seem like just the sort of thing that would stop working if too many people used them.

If everyone believed in “sell in May and go away” then we could anticipate a big sell-off in May and a rise in November. So the right thing to do would be to sell before May and buy before November. But if too many people did this, the right strategy would change again.

There is one stock market pattern that I bank on, and that is the tendency for people to be conservative. To entice investors, risky investments have to offer significantly higher expected returns than safe investments. This makes sense to a degree, but I think investors are generally too conservative. This makes riskier investments look better to me.

So, I am content to invest in the stock market and hope to get higher returns than I could get with safer investments. I have my limits, though. I don’t use leverage, and I stick to indexes for broad diversification. And I avoid getting drawn into attempts to beat the market with interesting strategies.


  1. Sound choices.

    I hope you are not suggesting that others follow your lead - for the sole reason that you published this post.

    To me it is excellent food for thought.

    Investing is very personal.

  2. I agree. As a software developer, I liken the situation to debugging: as soon as a flaw in the market is found, it is fixed. There's just too much money to be made for this not to occur.

    I think there are two forms of the theory that the stock market gives the best returns in the long run: the weak form and the strong form. The weak form says that historically it has been best, and since we don't have a crystal ball, we do what has worked in the past until we have reason to do otherwise.

    The strong form is this: if there were some asset class that consistently outperformed stocks, then companies that base their business on that asset class would grow faster than the rest of the market. Eventually, they would grow so large that they would dominate the market. From that point onward, the market would track that asset class. Hence, by design, stocks must always give the highest returns in the long run.

    There are flaws in this "strong form" theory (eg. not all companies are publicly traded), but I think the gist of it is close enough to reality to give me confidence that stocks' historical success is systematic, rather than some glitch that will be "debugged" out of existence.

  3. @Mark: I never suggest that others follow me without thinking. I suspect that many people would do well to follow a similar path to mine (perhaps with different percentages of their portfolios in safe investments), but they should think for themselves.

    @Patrick: The debugging analogy works for me, because I've done a fair bit of programming.