Sept. 16, 2020
We are reaching a time of the year when markets typically exhibit seasonal weakness based on research that focused primarily on the U.S. markets. But is there seasonal behaviour in Canadian securities as well? And if so, has seasonality changed over the years?
Using Canadian data for the period 1957-2018 and sub periods, I examined the following questions. Is there a January Effect in stock returns in Canada? How much truth does the popular expression “Sell in May and go away” have? Are there seasonal patterns not only in stocks, but also in government bonds? Are these two seasonalities interlinked? If such patterns exist, what drives them and are there profitable opportunities arising from such behaviour of financial securities?
I proxied equities by the Canadian Financial Markets Research Centre (CFMRC) equally weighted (EW) stock index and the risk-free securities by the long-term government of Canada (GOC) bond index.
In the accompanying tables, we see that the EW stock index shows high average returns in the month of January (and in the months around January) and low returns thereafter, particularly in the May to October period. Both sub-periods exhibit similar behaviour, although the strength of January has weakened somewhat in the second sub-period. The GOC bond returns exhibit strong returns only in the August to December period. Moreover, while stocks seem to do better than GOC bonds in January and at the beginning of the year, the GOC bond index is outperforming the EW stock index in the August to November period, especially in the second sub-period, when the GOC bond market was more liquid and better developed.
There is a January effect in Canada and, more importantly for this time of the year, a “sell in May and go away” effect. In fact, the average annual rate of return over the past six decades (1957-2018) would have been 17 per cent had investors gone long in the equally weighted index in November-April and gotten out of risky securities altogether in the May-October semi-annual period and, over that period, invested instead heavily (and exclusively) in government of Canada bonds. It would have been 18 per cent in the 1988-2018 sub-period.
In more recent years, that is, since just before the Great Recession, the above examined seasonal patterns have moderated, but not disappeared. For example, the aforementioned strategy would have generated an 11.5 per cent return from 2006 to 2018.
What drives the seasonal behaviour of stocks and government bonds? In my opinion, it is portfolio rebalancing and gamesmanship by professional portfolio managers.
Portfolio managers exhibit herd mentality. They are safe when their portfolios look pretty much like everyone else’s who invest with the same mandate, as no one loses his/her job because of average performance or holding the same securities as the rest of the peer group. Herding becomes more pronounced at two points in the year. First, at the beginning of the year when portfolio managers take riskier positions than their benchmarks in an attempt to outperform and affect their Christmas bonus. And second, toward the end of the year when portfolio managers a) window-dress to spruce up their portfolios by selling stocks that are obscure and have fallen in price and buying up other stocks (and other securities, such as government bonds) that have done well and are visible and in the public eye, and b) lock in good performance by selling risky stocks (which they bought at the beginning of the year) and move into lower risk securities, such as GOC bonds to affect that Christmas bonus.
Window-dressing and remuneration-motivated portfolio rebalancing, exacerbated by herding, affects prices and returns of financial securities throughout the year in a predictable way. Risky stocks are bid up at the beginning of the year (down toward year-end), whereas less risky securities, such as GOC bonds are bid up toward year-end (and down at beginning of the year). The pattern repeats annually mimicking window dressing and/or the annual performance evaluation cycle of portfolio managers.
November and/or December returns tend to also be significantly positive for the EW stock index. It is quite possible that some arbitrage is taking place by those investors not bound by the constraints or conflicts that portfolio managers face. The fact the January strength has diminished over time for the EW stock index provides some evidence that arbitrage has indeed taken place. In addition, it is possible that some risk-taking behaviour is followed by “desperate," so to speak, portfolio managers who have lagged their benchmarks and are trying to catch up by investing in extremely risky stocks.
Nevertheless, the documented seasonal behaviour is difficult for the markets to fully eliminate for two reasons. First, it is related to window dressing and/or remuneration-motivated portfolio rebalancing by professional portfolio managers who pursue their own interest year in and year out. Second, seasonality is not consistently observed every year. Unless we have a unified theory to help us anticipate seasonal behaviour on a consistent basis, market participants cannot fully arbitrage the seasonal behaviour of financial securities.
George Athanassakos is a professor of finance and holds the Ben Graham Chair in value investing at the Ivey Business School, University of Western Ontario.
This Globe and Mail article was legally licensed by AdvisorStream.