StockWatch
How January Goes, so Goes the Year?
As this is written, it seems almost certain that the stock market will post a loss for the entire month of January.
And, that’s bad news indeed, according to followers of the January Indicator, which holds that the first month of the year has the ability to forecast the stock market’s direction for the rest of the year.
But, just how strong is the statistical foundation for the January Indicator?
By way of an answer, consider the data in the following table, which is based on the Dow Jones Industrial Average back to the late 1800s, when this benchmark was created.
When the DJIA in Jan… |
% of time DJIA gained between Jan. 31 and Dec. 31 |
% of time DJIA declined between Jan. 31 and Dec. 31 |
Average DJIA gain between Jan. 31 and Dec. 31. |
...Gained |
72% |
28% |
7.81% |
...Lost |
51% |
49% |
3.70% |
The data presented in the table certainly appear to provide compelling evidence in favor of the January Indicator. But, before you conclude that the first month of the year is endowed with some special forecasting ability, you should know that there are other months that appear to have just as good a record, if not better, of foretelling the stock market’s direction over the subsequent 11 months. April and November stand out in this regard, for example.
This is just a reflection of the stock market’s tendency to move in trends: A declining month for the stock market means that there is an above-average chance that the market will continue to decline for the next 11 months -- regardless of whether that month is January, August or any of the other months of the calendar. If you say your belief in the January Indicator is based on the evidence, you should also be a believer in an April Indicator and a November Indicator.
There’s another reason that the statistical case for the January Indicator is weaker than it otherwise appears on the surface: The strength of that statistical case depends crucially on whether the decade of the 1930s is included in the analysis. If that decade is omitted, the statistical case in favor of January’s forecasting ability grows significantly while the case in favor of the other months simultaneously declines.
Credit for this insight goes to an academic study called “The Other January Effect,” which was conducted several years ago by three finance professors: Michael J. Cooper of the University of Utah, John J. McConnell of Purdue University and Alexei V. Ovtchinnikov of Virginia Polytechnic.
Is there any reason why the decade of the 1930s should be ignored when calculating the January Indicator’s track record? The most plausible argument I have come across is that the capital gains tax was not inaugurated until the early 1940s. That tax, the argument goes, might give January a different significance in the years since 1940 than it had in the 1930s.
But, Professor Cooper, in an interview, said that he does not place much weight on this argument. January had a good forecasting record in the decades prior to the 1930s, in addition to the decades since then. And, in those earlier decades, there was no capital gains tax. So, he argued, its adoption in the 1940s can’t be the reason why the January Indicator was a failure during the 1930s.
Are there other rationales that are more compelling for why the decade of the 1930s should be ignored, which, in turn, would increase our confidence in betting on the January Indicator? I am not aware of any, and Professor Cooper says that he and his co-researchers aren’t aware of any, either.
The bottom line? The statistical basis for the January Indicator ranges from marginal to quite strong, depending on the importance you place on the decade of the 1930s.
And, you need to decide whether that decade should be ignored before deciding whether to bet your portfolio on the notion that, as January goes, so goes the year.
Who said forecasting the stock market is easy?Mark can be contacted via email at mhulbert@marketwatch.com.
