We’ve investigated the best-day-of-the-week and the best-day-of-the-month, so I thought my series wouldn’t be complete without taking a look at the best-month-of-the-year.
For the test I simply computed the average return for each month of the year across every liquid stock listed on the JSE since 2003 till the present day. Before I conduct any testing I like to take an educated guess at the possible characteristics of my test results. If my testing confirms my estimation, then we probably have something worthwhile pursuing. In this particular case however, I couldn’t arrive at an economically sound reasoning for one month to outperform any other. I did have a bunch of ideas though:
1) The trader’s old age adage “Sell in May and go away”. The strategy states that the performance of stock markets during May through October are inferior when compared with November through April. If there is any substance to this claim then we’d expect May through October to generate below average returns.
2) The Santa Claus rally, which we tested here and found to be massively significant, may result in outsized returns in December. People’s spirits are generally lifted during the festive period which may overflow into the markets.
3) What about tax reporting months? I’m no tax guru, but perhaps clients looking to lock in losses for the year to write against taxable income may have some impact. This would result in selling, and with the majority of individuals reporting by the end of Feb, with provisional during August, perhaps we should cast our eye to those months.
4) The January effect, which states that the first five days of January provide abnormal returns. The reason is assigned to market participants in general selling the market during December to raise cash for their holidays. This has the effect of temporarily depressing equities, which are in turn picked up by bargain hunters, pushing the markets higher into January.
With those ideas in mind I proceeded to test the best-month-of-the-year hypothesis and present the results for you below.
The first theory, “Sell in May and go away”, appears to be partially confirmed by my testing; May shows the worst monthly gain on average and the subsequent months tend to underperform. However, September and October provide above average returns which refute the theory. So in SA it seems the adage should be revised to “Sell in May and go away, but return in June for a month of work, and then resume vacation, but hurry back for September and October”. Hmm….not the sort of uninterrupted vacation I had in mind, but interesting nonetheless.
Santa’s effect on investors appears to be rather temporary, lasting only briefly though the Christmas period, and having no significant effect on December returns. In fact, December ranks among the worst performing months, which may confirm part of the January effect, but does little for us here. It seems Santa has his work cut out.
I really can’t comment on the tax reporting, other than to say I’m not particularly fond of it. If any tax wizzies out there see a potential pattern I’d be very interested to hear your thoughts, just pop me a mail.
This now brings me to my final thought, the “January effect”. There are two parts to this theory 1) selling in December to raise cash resulting in inferior returns for the month and 2) the outperformance in January. The first part definitely provides some evidence of the theory – December ranks among the worst performing months. However, January does not generate above average returns, which may provide evidence contrary to the theory. But, one possible explanation for this may be linked to the “January effect” lasting only until the early part of the month, and perhaps further selling in January to finance vacation dampens the overall effect on the month.
Ok so a mixed bag of results I know. Perhaps a pattern jumps out at you? Let me know and I’ll share your thoughts in a future post.