Pass, January: at least two more months have a higher predictive power of stock market profitability than you have.
January has the reputation of being able to forecast the direction of the US market in the next 11 months of a year. This supposed ability is known as “January predictor” and “January barometer”.
You’ll see a lot of references to this indicator in the coming days, now that January is officially on the record as a “down” month – with the S&P 500 SPX,
slipping 1.1%. I also wrote that the January Predictor is based on a weak statistical basis. However, the financial titles will have the supposed negative implications of the January decline for the rest of 2021.
So let me point out a few other ways that the Predictor is not worth following.
What is so special about January?
A good place to start is to remember that January 2020 was also a downward month (down 0.2%) and yet the next 11 months produced a gain well above the average of 18.4% (assuming that dividends have been reinvested).
This is just a data point. Another indication that there is nothing special in January is that the other months have even greater predictive “powers” when forecasting the direction of the stock market in the next 11 months. In fact, since the creation of the S&P 500 in 1954, June has the strongest forecast capacity, followed by February. January is in third place.
Then why not read about a predictor from June or a barometer from February? My assumption is that adherents are less motivated by statistical rigor than by stories and narratives that catch their attention. From a behavioral point of view, the calendar year is a more natural period to focus on than the February-February or June-June periods. But the psychological significance is different from the statistical significance.
The importance of real-time testing
There is another sign that the January indicator is not all that is deciphered: it does not pass the tests in real time.
By this I mean tests performed after it was initially “discovered”. If the January Predictor had succeeded in passing these tests, we would be much more confident that it is not just the result of a data mining exercise in which historical data is tortured long enough to make a model appear.
But he couldn’t. As far as I know, the real-time test of the January Predictor begins in 1973. This is the oldest mention of it on Wall Street, according to an academic study on the subject. Unfortunately, his record is much less impressive. In fact, since 1973, not only is it not significant at the 95% confidence level that statisticians often use to determine whether a model is genuine, but it is not even significant at the 85% level.
We should not be surprised; in fact, the January Predictor is in good company. Consider a study that appeared in the Journal of Financial Studies in May last year. He examined 452 alleged statistical patterns (or “anomalies”) that previous academic research had found to exist. The authors of this recent study failed to reproduce these results in 82% of cases. The remaining 18% proved to be much weaker than initially reported.
There is no correlation between the magnitude of the rise and the return in January over the next 11 months
Another indication that the January Predictor is based on an unshakable statistical basis is that there is no correlation between the strength of the January market and its gain in the next 11 months. If there were such a correlation, we might be able to invent a plausible narrative about investor confidence at the beginning of the year, carried over to the rest of the year.
But there is no such correlation. Because of this absence, to believe in the effectiveness of the January predictor, you will have to believe that an S&P 500 gain of only 0.01 has as much predictive power as a gain of 13.2%. That strains credulity.
By the way, I chose this 13.2% in my illustration because it is the highest January gain for the S&P 500 since its inception in the mid-1950s. It came in 1987. From January 31 of that year to the end of the year In 1987, the S&P 500 lost 9.9%.
To take advantage of a statistical model, you have to follow it religiously for years
Finally, even if the January Predictor relies on a solid statistical basis, you should act on it for many years in a row to rationally try to take advantage of it. A good rule of thumb is that you need a sample of at least 30 before the patterns become significant. In the case of the January Predictor, this means that you should follow it for three decades. Moreover, in those 30 years you would not enter into any other transactions, except to switch to a 100% equity allocation on every 31 January when the stock market grows in January and to a 0% allocation if the market in January is down.
In the absence of patience and discipline, do not do too much to improve your chances over a coin.
The bottom line? From all points of view, you can’t conclude anything from the January stock market decline about where it will stay on December 31st.
Mark Hulbert is a regular contributor to MarketWatch. Its Hulbert balance sheets follow investment newsletters that pay a flat fee to be audited. He can be contacted at [email protected]
More: Get ready for a 10% drop in shares, led by the 3 “Rs”, warns Bank of America
Read also: This surprising stock market synchronization strategy has an academic seal of approval