For stock market performance, January is cloudy crystal ball
of pithy investment advice, Hirsch also created the strategy often referred to as “Sell in May and go away,” and he identified a year-end seasonal “Santa Claus” rally.
His so-called January barometer posits that a decline in the Standard & Poor’s 500 index in January means that major averages will also end that calendar year with a decline, and vice versa if stocks rise during the month.
Given the global plunge in stock market prices last week — U.S. and European shares continued to drop Thursday in the face of market turmoil in China, the worst start to a new year for the S&P 500 since 2008 — it may be a grim year for stock investors. If the selling keeps up through the end of the month, the almanac suggests investors should sell stocks and wait for the inevitable further decline. If it were only that simple. It’s tempting to dismiss the January barometer as just another statistical coincidence, along with other market predictors that have enjoyed periodic bursts of popularity, such as the Super Bowl indicator (if the winning team is from the old National Football League, now the National Football Conference, stocks will rise); the Chinese zodiac (years of the rabbit and dragon are best for stocks); or the old hemline indicator (skirt lengths shorten during market upswings and fall to the floor during downturns).
But Hirsch’s son, Jeff, who now runs the publishing and investment business his father founded, calculates that the January barometer has been right 87.7 percent of the time since 1950 (ignoring basically flat years) and 75 percent of the time including all years. He said the percentages were roughly the same extending back to 1938. He also examined all other months of the year, and none displayed any predictive capacity. (April was next best but showed only a modest correlation with subsequent stock movements, he said.)
Jeff Hirsch posits that this is not simply coincidence. He said his father’s insight was that new political agendas — the convening of a newly elected Congress, the inauguration of a president, the State of the Union address — emerge in January.
Traders react to them immediately, usually with considerable foresight, and then the economic implications unfold throughout the year, confirming the traders’ January instincts.
“Traders and analysts react to the political agenda, as well as what the new year looks like based on last year and the state of the world economy,” Hirsch told me last week. “I think that’s the root of it.”
But such causation in the stock market is impossible to prove or disprove, unlike, say, the link between smoking and lung cancer.
“You can’t possibly prove this is causal or even predictive, which might be a better concept,” said Rebecca Goldin, a mathematics professor at George Mason University and director of research for the Statistical Assessment Service. “That doesn’t mean it isn’t, but you’d need to know much more.”
She pointed out that a 75 percent or 87 percent correlation sounds impressive, but not if the stock market goes up every year, in which case it is meaningless. In fact, the stock market does tend to rise more often than it falls: It has gained in 21 of the past 30 years, or 70 percent of the time. So simply predicting every year that the stock market would go up would have been nearly as accurate as the 75 percent rate achieved by the January barometer when all years were included.
Moreover, using January to predict the outcome for the full calendar year incorporates January’s results, which further biases the prediction, especially if there’s a significant market move in January.
Given the vast number of variables that might be used to predict stock prices, it’s possible that even a high degree of accuracy of the January barometer is a purely random outcome.
“If you have a truly random variable, and there are, say, 60 million possibilities, it’s impossible not to find some pattern somewhere,” said Nassim Nicholas Taleb, a professor of risk engineering at New York University, author of “The Black Swan” and a former derivatives trader on Wall Street. “You might well find a correlation between changes in your grandmother’s blood pressure and stock prices. But that’s a spurious correlation.”
He added, “It’s surprising we don’t have more weird correlations, given the vast number of possible variables and the large number of markets we have.”
Goldin said that if you flipped enough coins every year, you would most likely find one that predicted stock prices with 100 percent accuracy.
“You could call that a magic coin,” she said. “It has an impressive correlation looking backward. The problem is, it wouldn’t give you any sense of how well it’s going to do going forward.”
That seems to have been an issue lately with the January barometer, whose accuracy in recent years has been slipping. During the past 10 years, it gave false negative indications in 2014, 2010, 2009 and 2005, and a false positive in 2011.
That’s a success rate of just 50 percent, which is what you would expect from a coin toss.
That has not curbed popular demand for Hirsch’s insights. He has been deluged with media requests and has been making television appearances talking about the almanac and the January barometer. Most likely for evolutionary reasons, the human mind looks for patterns, and it is uncomfortable with randomness.
“There’s a reason people love almanacs,” Taleb said. “They purport to find order in random events, like weather or the seasons. People have an irrational belief in them.”