(Originally written Aug 11; posted Sept 5)
I’ve noted a few times that gold rises pretty much during the university year, and falls in the summer months. Being the conscientious and only mildly obsessive-compulsive type, it irked me that I didn’t have (favourably-selected) data at my fingertips to back up my statements.
Fortunately, an American mutual fund manager pulled the data together recently; more fortunately, I stumbled across his editorial piece; and most fortunately of all, when I write these things, it looks like I’m hard at work.

One of the funny things about investments is that they tend to follow loose patterns — they’re just predictable enough that you think you can make enough money on them, and just unpredictable enough to prevent you from doing so.
One of the better-known investment rules — and one which actually works — is “sell in May and go away“. The lesser-known back-half “and stay away until St. Leger Day” (roughly equivalent to Hallowe’en) sadly, is shrouded in relative obscurity, not unlike the three other stanzas for O Canada. …what? You’ve never heard “O Canada! Where pines and maples grow”? Tish, tish…
The funniest part of the “sell in May”/”buy in November” rule is that… it actually works. From the Wikipedia article, the phenomenon has held up in 36 of 37 countries, and has worked in Britain for the past three hundred years. As Wikipedia drily notes,
“According to the efficient-market hypothesis, this is impossible.”
…which pretty much tells you all you need to know about the efficient-market hypothesis! One wonders if ancient encyclopedias concluded their discussion of Magellan’s circumnavigation of the world with “according to the flat earth hypothesis, this is impossible”.
Oh — and bear in mind that a three-hundred year trend might have an off-year that one November you decide to bet everything on “black” at stock-market roulette.
- - - - - -
But back to gold. As explained in the article, and annotated by some charts, gold’s typical rise from August to May is largely a function of demand: weddings in India, Christmas festivities in the West, and Chinese New Year. (The author also mentions Ramadan, but Ramadan shifts back by about ten days every year. As such, over time, it won’t mesh with the others, which stay put in one part of the calendar.)
During this time period, there’s usually a swoon around October and one around March. These tend to be due to the fact that investor “chum” notice the uptick in August-Sept or Jan-Feb, put in some savings, causing temporarily overbought conditions… leaving conditions ripe for market sharks to feed.