MarketSci’s Take on the January Indicator
As a self-described geek, nothing annoys me more than analysis like…
Since 1931, when the month of January was positive the S&P 500 went on to close even higher 82% of the time for a median gain of +11.4%. But when January was negative it went on to close higher only 52% of the time for a paltry gain of +0.4%.Wowzers, look how strong the January Indicator is.
That statement, though true (they’re my calculations), does nothing to tell me how consistent this observation has been over time, or whether its effectiveness is waxing or waning. The graph below however, does.
The graph is using a 20-year rolling average to look at median returns for the year following a positive (blue) or negative (red) January. In other words, the results for 1950 only include 1931 to 1950, the ones marked 1960 include 1941 to 1960, etc.
Here we see that the January Indicator has been (surprisingly) consistent over the last 79 years.
The next graph is also using a 20-year rolling average, but rather than looking at median returns, we’re looking at the percentage of years that went on to close even higher after a positive (blue) or negative (red) January.
And just for good measure, volatility-adjusted average returns (annualized return divided by annualized standard deviation) following a positive/negative January.
The indicator has been far from decisive, but still pretty consistent.
Is the January Indicator Real?
I think there’s a wee bit of truth to the January Indicator. No other month has been consistent in predicting returns over the following 11-months, and there’s actually some logic in the rule (asset managers positioning themselves in January based on expectations for the remainder of the year).
Of course as active traders, this is all shop talk and polite dinner table conversation. We’re trading in much, much shorter timeframes, and if I might be so bold, I don’t think any strategy with such a long-term horizon is going to keep pace with a robust short-term strategy.
[click for a summary of all recent posts about January Indicators]
Happy Trading,
ms
. . . . .
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Time-based, Trading Strategies | 3 Comments
You say “No other month has been consistent in predicting returns over the following 11-months.” I’d be interested in seeing that data. My first thought was that to the extent that the market trends on a monthly basis any month is likely to “predict” the months around it.
RE to Blue: long time not talk sir – good to hear from you.
If one approached the problem in terms of just averages (like my much despised first paragraph) then yes, many months have been predictive of future 11-month returns. I’ve seen quite a few analysis like that in the blogosphere. But if someone looked at it over time, all of those months went through long periods of success and failure (implying w/ the small sample size that they’re just noise). Side note: I’m not convinced the “Jan Effect” is anything but noise either, but I’m more sold on it than the other 11 possibilities.
If the subject generates some interest, I’ll post the other 11 graphs.
Hope all is well,
michael