The January Effect refers to the tendency of small cap stocks to out-perform large-cap stocks in the month of January. However, the term January Effect is used rather loosely to also refer to stocks generally being strong in the first month of the year, and also to how the direction of the market in January forecasts the market direction of the whole year (this latter effect is also termed the January Barometer). [A previous article explained the multiple January Effects in greater detail.]
Here, we are going to look at a variant of the January Barometer to see if the first five days of the year predict the return for the whole year.
First, we will call this variant of the January Barometer: January Barometer (5D).
The bald figures don’t look encouraging: in the 46 years since 1970, the January Barometer (5D) applied to the FTSE All-Share Index has been right in 26 years (57%). In other words in just over half the years since 1970 the first five days of the year have accurately forecast the full year.
But let’s look at this in more detail and see if we can tease anything out of the figures
The following is a scatter chart that plots the return on the FTSE All-Share Index for the first five days of a year against the return for the full year, for the period 1970-2015.
There is a positive correlation here (given by the positive sloping trend line), however the measure of correlation (R2) is very low.
Summary: the chart shows there is a very low level of correlation between first five-day returns and returns for the full year but it is far from being significant.
However, strictly, the January Barometer only says the direction (i.e. positive or negative returns) can be forecast, not the size of returns. In which case the following chart may be more useful. This plots a binary value for each year:
- 1: if the sign on the full year return was the same as the sign for the return for the first five-days (i.e. either both positive returns or negative returns)
- -1: if the sign on the full year return was different to the sign for the return for the first five-days
In this chart we can see the roughly even split between years when the January Barometer (5D) works and those years when it doesn’t. However, the distribution of years when it works is interesting, as there does appear to be a certain clustering of years when the effect works and when it doesn’t.
For example, in the last 20 years the January Barometer (5D) has been accurate 14 times (a hit rate of 70%). And since 2004 there is this rather odd pattern of not working every fourth year.
US presidential elections
US presidential elections also have a four-year cycle. On the chart presidential years are marked with orange bars.
It can be seen that since 2004 the January Barometer (5D) has worked every year except in years before presidential elections.
And, over the longer term, since 1970 the January Barometer (5D) has only failed in three presidential elections (a success rate of 73%).
The outlook for 2016
Generally, the January Barometer (5D) has a low success rate. However, the effect has been more significant in recent years; plus it has a higher significance in US presidential election years (which 2016 is). In 2016 the market was down in the first five days of the year, and so the January Barometer (5D) would forecast a down year with a 73% probability.