It’s easy for investors to ‘discover’ a pattern where none exists
Take my word for it: April is the best month of the calendar for stocks. Or it isn’t.
Yet some exuberant bulls on Wall Street are telling their clients that the U.S. stock market’s seasonal tendencies are strongest in the month of April. While I have no doubt that the historical data can be tortured to make April look like the best month for equities, the data mining required to do so means that the conclusion has no statistical significance.
You still might want to bet that the stock market will perform well this month. But you need to support your bullishness on other grounds besides the calendar.
One of the tweets I saw supporting April’s claim to being the best month was based on data back to 1950. But why pick 1950? Had the tweeter expanded the lookback period to 1920, April would not have been in first place. Unless there is some theoretical explanation for why April should be a particularly strong month for stocks just since 1950, April’s inconsistent ranking is a dead giveaway of shameless data mining.
| Period | Month that is in first place for average DJIA return | April’s rank |
| All years since DJIA created in 1896 | July | 3rd |
| 1896-1920 | March | 9th |
| 1921-1945 | August | 7th |
| 1946-1970 | December | 5th |
| 1971-1995 | January | 3rd |
| 1996-2022 | April | 1st |
Another illustration of this inconsistency is provided in the table above. I constructed it by segregating all historical data for the Dow Jones Industrial Average DJIA, +0.40% into five equal-sized groups of approximately 25 years each. Notice that, over the longest period for which monthly Dow data is available (since its creation in 1896), July is in first place for average return, and April is in third place. And in only one of the quintiles is April in first place. This table is the very picture of a random pattern.
April showers
This review of April’s place in stock market history is worthwhile because it reminds us of how easy it is for investors to “discover” a pattern where none exists.
Psychologists have long been aware of this tendency, even if most investors are not. A good summary of the research is provided in a study entitled “‘Superstitious’ Investors,” which was conducted by Jessica Wachter, a professor of financial management at the Wharton School, and Hongye Guo, a doctoral candidate in finance at that institution. All of us, whether when we’re investing or engaged in any other activity, have great difficulty truly contemplating randomness. We find patterns in random data even when we receive specific training on how to differentiate what is random and what is not.
It’s hardly a flattering picture, but the researchers draw an analogy between investors’ behavior and that of the pigeons in the famous experiments conducted by famed psychologist B. F. Skinner many decades ago. “In Skinner’s study, hungry pigeons were presented food at regular intervals. Most of the pigeons developed bizarre habits of behavior, the reason for which is that they happened to have displayed that specific behavior when the food [previously] was offered.” Like investors, the pigeons have the “tendency to create structure out of randomness.”
This is why it’s so important for investors to base their investment decisions on objective statistics rather than their subjective intuitions. Even when you’re aware of your tendency to find patterns in randomness, and are trying to avoid doing so, you still are vulnerable.
Mark Hulbert is a regular contributor to MarketWatch. His Hulbert Ratings tracks investment newsletters that pay a flat fee to be audited. He can be reached at [email protected]
