Exploring Kaeppel’s Sector Seasonality (Breaking Out Individual Sectors)
This is additional drilldown into Kaeppel’s Sector Seasonality Strategy. A refresher on the strategy’s performance from 1987 (in red):
[logarithmically-scaled, growth of $10,000]
Kaeppel’s strategy rotates between three stock sectors (technology, energy, and gold) plus cash, based on yearly seasonality, using Fidelity Select Sector Funds or similar products.
In this post I’m going to break out the historical performance of each of those three sectors in the particular months the strategy trades.
Technology Sector
Kaeppel’s strategy calls for going long the technology sector in November, December, and January. I’m going to use the Nasdaq Composite as a proxy for the Fidelity sector fund FSPTX because more data is available, and correlation between the two is reasonably high.
Historical results from 1971 (frictionless):
[logarithmically-scaled, growth of $10,000]
Note: volatility-adjusted return = average monthly return / std. dev. of returns
Energy Sector
The strategy originally called for going long between February and May, but I just don’t understand why February was included. The month has historically been pretty blah for energies.
In the test below (my only modification to Kaeppel’s strategy), I’m only long March, April, and May. I’ve used the Amex Oil Index as a proxy for the Fidelity fund FSENX.
Historical results from 1984 (frictionless):
[logarithmically-scaled, growth of $10,000]
Gold Sector
The strategy calls for going long the gold sector only in September. I’m using the PHLX Gold/Silver Index as a proxy for the Fidelity sector fund FSAGX.
Historical results from 1984 (frictionless):
[logarithmically-scaled, growth of $10,000]
Side note: no relevance to this particular strategy, but the gold sector has performed very poorly in October. Just something to keep in the back pocket.
Thoughts…
With the exception of the energy sector in February (mentioned above), all of the seasonality plays performed reasonably well over these extended backtests. My favorite of the three is the energies play for its consistency, and my least favorite is gold for the fact that most gains came from just a handful of years.
As I’ve warned in both of my previous posts here and here (as well as CXO’s here) this strategy is by its nature curve-fit, but given its consistency historically and its performance out-of-sample, I think it should at least be a background consideration for traders of the longer-term variety.
I’ve been thinking about adding an addendum to the State of the Market for strategies that aren’t really appropriate for the main report, but are still worth keeping on the radar. This one would be perfect for inclusion. More to follow.
[click for a summary of all recent posts about Kaeppel's strategy]
Happy Trading,
ms
. . . . .
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Stock Market Sectors, Time-based, Trading Strategies | 2 Comments
The energy sector results are extraordinary. The XOI is a price only index and doesn’t reinvest dividends from my recollection and that will more than offset the transaction costs from two trades per year. Given the traditional high yields on energy stocks, you may have to throw in an extra percent per year.
RE to Highgamma: I agree, exceptionally smooth.
In my original test of the whole strategy, I used the Fidelity funds, so the whole strategy test is accurate in terms of total return (assuming you traded directly w/ Fidelity and incurred no per-transaction cost).
I used proxies for this post (like XOI) because of the add’l data available, and you’re right, the average return numbers are going to be a little off (to the low side). In this post I wanted folks to focus more on the consistency of the play rather than actual returns.
Thanks for the comments,
michael