About a year ago, researchers from Massey University (New Zealand) proved that the difference in return between summer and winter is as old as stock data. They also reported that it has increased in the last decades and can be observed in most stock exchanges all over the world. Similar monthly patterns can be found in the northern and southern hemispheres, which indicates that seasonality has nothing to do with the weather or a legacy of agriculture cycles. Readers wanting more details can read this first.
Such a load of new information was an opportunity to consider new ways to integrate seasonality in investing strategies, and resulted in a chapter of my book. I cannot unveil the content, but I will share a few tricks that can help investors.
According to the traditional version of the Halloween effect, the good stock season is from the 1st of November to the 30th of April. However, October is statistically a good month for at least 50 years, and a very good one for 15 years (in spite of 2000 and 2008).
So the first trick is to include October in the good stock season.
Moreover, January was the worst month for 15 years. February was flat on various time frames from 50 years to three centuries, and is the second worst month for 15 years.
The second trick is to exclude January and February from the good season. It results in a four-season pattern, in opposition to the traditional two-season model. Good seasons are from October to December, and March-April.
The third trick is that it is safer to play seasonal with general indices than with sectors. The superposition of various cycles makes the case of sectors complicated. There are a couple of exceptions. For example, utilities have a steady seasonal pattern, but not better than a general index.
The fourth trick is that some countries are more seasonal than others. For example, I have written here in May 2012 that the German market (EWG) was a good seasonal play. It was confirmed by the academic articles released a few months later with statistics on 108 countries. The USA (DIA, SPY, QQQ), Israel (EIS, ISRA), Brazil (EWZ), Singapore (EWS), Russia (RSX), Turkey (TUR), Norway (NORW), are possible candidates. On the opposite, I would never play seasons with Finland (EFNL), Chile (ECH), Peru (EPU) or Kuwait.
The red line below is the equity curve of investing in DIA since 1999 with a four-season model, going out of the market from the 1st of May to the 30th of September, and from the 1st of January to the last day of February. Dividends and 0.1% transaction costs are included. The blue line is the benchmark (Dow Jones Index).
The seasonal model has a total return of 315% and a maximum drawdown of -28%, vs. a total return of 67% and a drawdown of -54% for the benchmark. The seasonal model improves the return, lowers the risk, and also reduces the exposition at 5 months a year.
The fifth trick is to invest the idle capital during the bad seasons in bonds (or for example TLT, IEF, SHY) or in shorting the stock market (for example selling DIA or buying DOG). If you have doubts about bonds because of the FED policy, and if you are afraid of shorting, cash may be the safest position in bad seasons.
The sixth trick is to combine a seasonal strategy with a momentum strategy. If the trend is consistent with the seasonal statistics, both strategies add their gains. Else, they hedge each other. This kind of auto-hedging portfolio seems to work much better than market timing. A very simple implementation of this model shows an annualized return of 16% and a maximum drawdown of 13% between 2003 and 2013.
Past performance is not a guarantee for the future, but studying the past is still the best way to make a plan.