'So goes January' is a long-established popular myth that says if the first few days of January are down, the month will be too, and so will the year.
There are minor variations of this myth - some claim it's the first day alone that's predictive. Most scoff at that, thinking it's the first three, five, 10 or 17, or some other random and arbitrary number of days (unable to see why if one day isn't predictive, then no other random number will be either).
Some say it's the first week, but then where does the New Year holiday fall into that? And do four days make a week? All of this is nonsense.
There's a whole mess of cognitive errors that occur with this or any other day-, month- or season-related myth (such as 'sell in May', or the supposed 'Santa Claus' or 'summer' rallies).
First, almost never do you hear 'so goes' if January starts strong. People don't say, 'phew! January was positive. No need to worry about the rest of the year'. No - those who like this myth usually use it to prop up bearish biases.
Second, it causes people to reframe. If January is down and the year too, the 'so goes' crowd claims victory (usually loudly). But if January is down and the year finishes up, they never issue a mea culpa.
Instead, they'll claim: 'Of course it doesn't work every year. You must look at the long-term average.' They change the observation period for occurrences that don't fit their pre-set notions.
Years that conform are, to them, hard evidence this works. But years that don't are not treated as evidence it's hooey. This is a perfect example of the human behavioural quality of confirmation bias: seeing what you believe and not seeing what you disbelieve.
Third, it's a silly human quirk to assign meaning to random groups of 30 or 31 days. There's nothing inherently more or less predictive about January, but because of what behaviouralists call 'mental accounting' we assign great import to the start of each year. But capital markets care nothing about mental accounting.
It's actually beyond amazing to me that this bunk endures. Even anecdotally, this one starts falling apart fast. Just think about recent history.
The S&P 500 fell 8 per cent and world stocks fell 9 per cent in January 2009*, but over the year they ended up 26.5 and 30.0 per cent, respectively**. There are plenty of years when January is up, down or sideways, and the year does something entirely different.
We can use the four-box methodology - a good debunkery trick you can use often - to demolish this. The table below (click to enlarge) shows the number of years since 1926 that stocks were up in January and up for the year, as well as down in January and for the year. But it also shows when January is up and the year down, and the reverse.
Is January predictive?
Since 1926, we've had a positive January and then a positive year 48 times - nearly 54 per cent of the time. That shouldn't surprise, since in an overall positive year more months should be up than down. Plus stocks are up more than down over history - roughly two-thirds of the time.
Since 1926, January has been up 63 per cent of the time, and 73 per cent of the years have been up. With no guarantee about the future, historically, stocks have really 'wanted' to be up more than down. With so many positives relative to negatives - about two to one - you get more up years than down years and more up Januarys than down ones.
Just 8 times - 9 per cent - January was up and then the year was down. Rare! So if up-up years happen most, and an up January rarely leads to a down year, does that mean January is predictive? No.
A down January is basically a coin flip, in that historically you get about the same number of up and down years following a down January - 17 to 16.
Seen another way, a flip-flop year - January up in a down year or vice versa - has happened more often (28.1 per cent) than a down January and down year (18 per cent).
One easy way to do debunkery is to ignore any rule that says: 'When this happens, that's automatically bad (or good).' No one thing is ever a definitive sign to either bail on stocks or be bullish: global capital markets are far too complex for one indicator to be so powerful. So goes January is how January goes, and then it's gone.
*Thomson Reuters, S&P 500 total return, MSCI World Index total return with net dividends from 12/31/08 to 1/31/09.