If you're looking for it, you'll see the Gambler's Fallacy pop up in random places.
The Economist posted a not-so-interesting graph relating a year’s stock market returns to its Chinese Zodiac sign. Although a random number generator making up returns for clusters of years since 1900 and ranking the outcomes would produce a similar outcome to the chart, let’s play devil’s advocate and see if there is something to be said about these numbers.
Business Cycles
Gambler's Fallacy: Returns by Year in the Chinese Calendar (The Economist)
Economic theory, news headlines and politics often revolve around the theory of business cycles, loosely defined as periods of monetary, investment and GDP expansions followed by contractions. If there is a trend of ups and downs in an economy’s growth, it would be highly likely that a few of the years would be hit more heavily by the highs or lows because of the cyclical nature of that growth. For example, the years of the dragon coincided with the years 1976 and 1988 which happened to be two solid years of returns for the S&P 500. The year of the snake (next year’s zodiac sign) was hit hard in 1977 and 2001. Any gain in the year of the dragon might be counteracted by the year of the snake just due to randomness in the data.
The Gambler's Fallacy and Election Years
Some studies have been made showing a correlation between the stock market and the point in a presidential cycle. Returns in the final year of a president’s term tend to be greater, which would fall on the years of the dragon, monkey and rat. It's possible (assuming no political manipulation) that these spurious relationships based off of noisy data are just pointing to the same random out-performance. Election years, much like zodiological signs, have no more correlation with the stock market than what I roll in a game of craps... but the two noisy indicators allow relationships to be imagined.
So what does this all mean? As has been stated many times by behavioral finance and behavioral science pundits/critics/skeptics, there will always be a correlation between past events and past results, even with random data. The chance that there would be no trend with so few data points is near zero. The chance that this trend will hold up over time is still completely random, so it is also not correct to say that it needs to start regressing to the mean and bet against these outcomes (known as the gambler’s fallacy). In other words, this article is just another reminder that there is no way to cheat stock market gains with goofy theories.
Cheers,
Cameron Daniels
