When Santa fails to call.

4th Jan '19 by Gromley

7 comments 1824 reads

This time of year tends to lend it self to a reflective mood and thoughts of whether the year just gone can gives any lessons for the one to come.

One such piece was Ed Croft’s thought provoking review of 100 years of market history. I mentioned in the commentary to that article that I’d been looking at something similar, but was struggling for a robust data source; so first off I should proffer my thanks to Nick Ray for helping me in that. (Stockopedia is a marvellous community).

(Just to note that I am using the FTSE ALL share as representative of “the market” in this piece, other indices may give different results and it is certainly likely that markets in different countries will differ).

The title probably gives away that I am looking at the failure of the Santa Rally to deliver in 2018.

The first question to ask is of course :

Is the Santa Rally even "a thing"?

Well at first sight the fact that the market has gone up in December 77% of the time over the last 39 years, doesn’t sound that impressive given that markets tend to go up most of the time.

In actual fact, though it is quite significant in the proper context.

December has in fact been the second best performing calendar month over the period only slightly behind April. The 77% 'success rate' shows that the market has only failed to rise in December 9 times (including this latest instant) in the last 39 years.

As an aside, that I may come back to in the future, this chart also gives strong support to the old adage that one should “sell in May and go away, don’t come back ‘til St Ledger day(mid September)”.

But we are in a down market so it’s unsurprising that December was down surely?

Perhaps or perhaps not.

Somewhat unscientifically, I divided the market in to upswings and downswings as per this graph

What this reveals is that December has fallen during a down-phase on eight occasions (including now 2018), but on four of those occasions the Santa Rally has held and the market still rose in December.
In fact one of those instances (1980) was during a down leg that only ran from late November to Mid-January - so one could make a…

Unlock with Email

Unlock with Google

Already have an account?
Login here

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

median %

1.4

1.0

1.3

1.8

0.5

-0.7

1.2

0.9

0.2

2.0

1.0

1.7

mean %

1.1

0.6

2.3

0.2

-0.7

1.0

0.8

-1.1

0.7

0.8

1.8

Martin5760

4th Jan '19

Really interesting and useful points made about how the markets perform in December and expectations - thanks for taking the trouble. The term 'santa rally' - I have seen quite a few commentators considering it to begin around the 10th of the month rather than the whole month and in the US starting the fifth trading day before the holiday and for it to run into the early part of the new year. Anyone else have an opinion on a definition ?

ezlifeme

Well Executed and researched piece as always Gromley. Many thanks for agitating my grey matter!

Gromley

Yes, I agree I'd always associated with the latter part of December and in fact Wikipedia describes it as being about the final week of trading before the New Year.

One of the early articles I read recently though quoted the stat. on the number times December has been a winner and that seemed a powerful enough thing to run with.

mmarkkj777

I do believe in Santa (rally). I was discussing this with some friends and the general feeling was that the rally usually starts around 10 trading days before Xmas.

I think Brexit put paid to it this year, although the trading days since Xmas have not been bad so far, so maybe he was just late!!!

timarr

5th Jan '19

The research, such as it is, suggests the Santa rally is a real thing:

Statistical evidence from this study supports the existence of the so-called Santa Claus rally. Daily returns during the holiday period showed both a higher mean and more desirable risk metrics than returns for the rest of the year. These findings are significant using both parametric and non-parametric tests.

See: Yes Virginia, There Is A Santa Claus Rally: Statistical Evidence Supports Higher Returns Globally

All empirical of course, even the longest sequence of data points, for the S&P 500, only dates back to 1950. Without any understanding of the underlying mechanism there's no real way of knowing how robust this is.

However, if I was interested enough to analyse it I'd want to look at daily volumes: you'd have to suspect that in thin markets with little institutional involvement the natural inclination of private investors towards a touch of irrationality might be brought to the fore.

AnonymousUser226237

Thank you, really insightful.

Nick Ray

As Gromley points out there are only 39 data points for each month. Since Gromley and I both have the same data set and it is relatively easy to ask R to calculate the returns on the same basis, I couldn't resist having a look at all the data points to get a flavour for how big the spread of results is.

We note immediately that October and September can be a bit scary. December looks broadly similar to other months except maybe a little more compact.

So to help the process of understanding the data, let's add a boxplot:

The middle "waist" marks the median, so 50% of points are above that line and 50% are below it for each month.

The top and bottom of the box marks the 25% and 75% percentiles. So half the points are within the space between them and half are outside.

The "whiskers" extend over what would cover about 95% of all points if the distribution is in fact Normal with outliers beyond that marked in red. December has no outliers above or below. So maybe it is "lucky" or maybe it is quiet.

The notch either side of the median line extends out to show a roughly 95% confidence interval for comparing medians. If the notches for different months do not overlap then we have some grounds for suspecting that they may be significantly different. But the notches do overlap for almost all months so any effects are very small.

Gromley actually looked at means rather than medians. The means have been pulled by the outlier points away from the median - for example look at Sep and Oct.

	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep	Oct	Nov	Dec
median %	1.4	1.0	1.3	1.8	0.5	-0.7	1.2	0.9	0.2	2.0	1.0	1.7
mean %	1.1	1.1	0.6	2.3	0.2	-0.7	1.0	0.8	-1.1	0.7	0.8	1.8

Maybe you should sell in May, go away until St Ledger's day, with a quick dabble in July, and watch your back in October!

When Santa fails to call.

Is the Santa Rally even "a thing"?

But we are in a down market so it’s unsurprising that December was down surely?

Unlock the rest of this article with a 14 day trial

About the Author

Gromley

7 comments

Unlock the metrics for all these stocks and more...

Unlock the metrics for all these stocks and more...