This is a follow up to Michael’s (MarketSci) excellent post reviewing Jay Kaeppel’s Fidelity sector rotation strategy. http://marketsci.wordpress.com/2010/01/08/kaeppel%e2%80%99s-sector-seasonality-strategy/
I review a lot of different academic articles and there was one that bore a strong relationship and relevance to the sector seasonality rotation strategy. Regardless of whether this is mere coincidence or not, for the current post reader’s may refer to the following article :
The Optimism Cycle:Sell in May (pdf) by Ronald Q. Doeswijk (November,2005)
The author investigates the performance of different sectors in an attempt to improve upon the traditional seasonality strategy of “Sell in May and go away” by keeping continuous exposure with a dynamic sector allocation. In this article a long/short sector seasonality strategy is derived that relies on going long cyclicals/short defensive stocks in the winter starting at the end of October (similar to going long cyclicals like energy for Kaeppel) and long defensive stocks short cyclicals in the summer starting in May (cash for Kaeppel and gold towards the tail end–also similar to a defensive type strategy). The author reports that a market neutral portfolio (long and short) achieved roughly a 7% CAGR with nearly 80% positive years between 1970-2003, and an even higher return in the later part of the sample.
To me the most important question to answer is why monthly/quarterly seasonality (or sector seasonality) should occur and also to derive supporting links for the effect to ensure that it is not the artifact of clever data mining. The author cites the following explantion as “The Optimism Hypothesis” and provides some supporting evidence in kind: