Ok, so it's not as poetic as the Ides of March, but October Mondays may be the most treacherous days of the year for stock market investors. This according to Salis Mehta, who delivers high-level wonk in the blog Statistical Ideas. Mr. Mehta's credentials as an econometrician are as solid as they are impressive. He was director of analytics in the Treasury Department for the TARP program, and is currently a consultant and an adjunct professor of statistics at Georgetown University.
It's established market lore that autumn is the riskiest time of year. In a blog entry entitled "Our Autumn of Discontent," Mr. Mehta had a close look at the autumn months to determine if that lore is simply another urban myth or if it has some validity behind it.
So, what did he find? Seems the tale stands up to analysis. October and surrounding months do, indeed, own the highest frequencies of the worst 1% days in market history. What Mr. Metra did is simple enough. He looked at the worst 1% of single-day losses in the entire history of the Dow Jones Industrial Average. At the time his writing, there were some 29 thousand trading days; the worst 294 of them were scored for month and day of the week. The monthly distributions of that 1% is strikingly skewed toward October. Here's Mr. Mehta's excellent chart.
For the chart above, the green lines show the expected distribution of those 294 worst days if there were no differences among months. The red lines show the actual distributions.
Let's note a couple of small points first. May's somewhat sullied reputation is undeserved as it falls exactly at the expected point, It does, however, skew out a bit relative to its surrounding months, which could be part of the back story behind the air of caution that it brings to investors each spring. A second point of interest is the favored status of January, which looks to be the least threatening of the twelve months. This is more consistent with received wisdom which predicts that the year typically starts on a positive note.
Enough of the small details. Let's look at the big story here: October. That red amoeba pushes out hard in autumn, beginning in September, reaching a peak as ghosts and goblins roam the land, and holding it through turkey time. I didn't have the raw numbers, but it looks like close to half of those 1% days fall in the fall. This year, with all the fear and dread surrounding the Fed and the taper, we may be looking at a truly frightening season of the witch.
Mr. Metra also looked at how days of the week fared. Monday owns the highest number of worst days; Tuesday draws the fewest. Thursdays and Fridays fall almost precisely where expected (there's a nice chart on his blog).
He puts the two together to produce this striking graphic representation. October is coded in Yellow and Monday is in blue. The big green giant hovering over the upper right is October Mondays.
Note too that November is at least as scary as October, maybe more so because you can't even stop holding your breath a bit on Tuesday: Its Wednesdays are just about as bad as October's Mondays.
What to make of all this? Hard to say, but this history converges with signs that a marked correction is in order. With US equity markets bouncing off all-time highs across the board (SPY, MDY, SLY), with the clear trend toward rising interest rates, with an increasingly imminent "tapering" by the Fed, it's hard not to anticipate one of those 1% days looming in the near future.
Consider, too, that 1% of days may sound infrequent but it translates to about 2 trading days a year. The last such day was in November 2011, nearly two years ago, adding to a sense that a big loser is sitting out there ready to pounce.
Might just be a good idea to hold on to a fairly heavy cash position going into the waning days of autumn.