Seasonal effects take a part in my search for probabilistic strategies. I learned that it may be dangerous to look for seasonal patterns in sector ETFs. Indeed when we dive in the sector rotation world, we have to cope with the superposition of cycles of different and sometimes variable lengths. That's why in the seasonal theme I have a strong preference for strategies on general indexes.
However, a particular sector is very interesting because that is where theory meets practice.
The longest economic cycles has been identified and documented by Kondratiev and Schumpeter. Basically these cycles are similar, they mainly differ in their explanations. Kondratiev sees the phenomenon as a consequence of a macroeconomic lifecycle, Schumpeter as a consequence of a technological innovation lifecycle.
The current wave, starting in the '90s, is lead by information technology. So the major schumpeterian ETF is IYW, that holds stocks of IT companies. In 5 to 10 years information technology should be replaced with the next schumpeterian sector: green energy, biotech, nanotech or whatever.
It appears that IYW is also the major sector ETF that gives the best historical return on 11 years for a seasonal swing - being long the "good months" and short the "bad months." You may find in a precedent article a definition of good months (March, April, October, November, December) and bad ones (the rest of the year).
I have reported another case of correlation between leadership and seasonal volatility amplification. That was for a German index (EWG) compared with other European indexes.
Let's have a look at some numbers:
A long position on IYW during "good months" (reinvesting gains and dividend or subtracting losses) returned +355% since April 1st, 2001. This means multiplying a capital by 4.5:
Click to enlarge.
A short position on IYW during "bad months" returned 85%. This means multiplying a capital by 1.8:
Honestly, the short side chart is quite ugly: Gains were made only in strong bear markets (2001-2002 and 2008); the rest of the time the short position on bad month is flat or with a small loss. Fortunately (or unfortunately) we know that strong bear markets historically occur once or twice in a decade.
Combining both, we obtain a multiplication by 8.1, which gives an average annualized return very close to 20%. That is more than 14% above the S&P 500 benchmark (SPY) and slightly better than gold (GLD) on the same period.
- The leader index in Schumpeter's wave is the best in a seasonal swing for an statistically significant period.
- This phenomenon may be used as a filter in a top-down approach for seasonal investment strategies.
- It is one more clue that leadership and volatility are correlated.
My conclusions stop here. I don't consider this IYW swing a good investment strategy. There are much better ones in the seasonal theme.