Seasonal patterns have been documented for decades, but scientific explanations have failed despite three centuries of evidence (read here). The easiest part may be to explain why November and December are among the best months of the year. There is a combination of the commercial and psychological impact of Halloween and Christmas, accounting optimization, and "window dressing" by some fund managers.
Regardless of the possible reasons (and regardless of the probabilities for the current year), here are the facts: investing in DIA (SPDR Dow Jones Industrial Average) only in November and December from 11/1/1999 to 12/31/2012 would have returned 68.8% (in red) compared with 71.7% for the Dow Jones index (in blue).
The blue line doesn't take into account the monthly dividend. But a 0.2 % monthly yield paid twice a year for the red line doesn't make a big difference. Investing in the Dow Jones index (excluding dividend) only two months a year for 14 years would have given almost the same annualized return as holding it all year long. Moreover, it would have reduced the maximum drawdown from -54 % to -21 % and the standard deviation from 23 % to 8.5 %.
Indexes, countries and sectors have not been created equal for the November-December seasonal pattern. I wrote last year that the Nasdaq 100 (ETF: QQQ) and the German stock market (ETF: EWG) were good candidates for a seasonal trade (article here).
Here is a table with the returns of investing in these ETFs just two months a year in November and December for 14 years:
Return w/dividend %
Max Drawdown %
With an average return above 7% in two months, a 22% drawdown, and only two losing years, the winner is EWG. The next chart shows the equity curve of holding EWG in November and December for 14 years (in red), compared with holding SPY all year long (in blue). This time, all dividends are included.
The total return is 174.8 % and the standard deviation is 12.6 % for EWG two months a year.
They are respectively 86.9 % and 23.9 % for SPY.
Even if seasonal patterns are powerful on the long term, pure seasonal strategies may have long drawdowns. It is risky to play a probabilistic game when you can play it only once a year, even with academic publications showing that this game works for three centuries. Yet, seasonal patterns may be valuable components of a core portfolio. For example, a seasonal strategy can improve the risk-adjusted return of an ETF portfolio when it is mixed with a momentum strategy. Both add their gains when the market is consistent with seasonal statistics, and they hedge each other when it is not.
You may want to click here to read more about our methodology and portfolios.