Long / Short Sector ETF Strategy

Today I would like to demonstrate a long/short sector ETF strategy using the Portfolio123 ETF screen backtest tool.

I am going to give a little "teaser" tonight. It isn't a fully developed system at present but demonstratess the potential of a hedged strategy using both long and short sector ETFs (leveraged and normal). First I defined a long ETF universe as follows:

Likewise a short ETF universe:

I am going to run two backtests: (1) long sector ETFs; and (2) short sector ETFs. First the long. Here is the screen setup. It is quite similar to the screen in previous post but this time the screen is looking for the ETF with the strongest yearly rise in each sector.

And now for the results with two holdings rebalanced daily.

Results are not particularly good. The bottom line is that $100 was invested and at the end of one year the trading capital shrunk to $53.00 (including slippage). Now let's see the short ETF screen.

The results for the short backtest. Again the portfolio is rebalanced daily while holding two ETFs.

Now you can see that the results are quite volatile. This is a result of combined use of leveraged ETFs with a long term rule (Close(0)/Close(250)) with short term violent swings. But the bottom line is that the end value of the Short ETF backtest was $215.1, starting from $100 one year earlier. If an investor put equal dollars into each strategy, he/she would start with ($100 + $100) = $200. At the end of one year the investor would have ($53.00 + $215.10) = $268.10. This amounts to a 34% profit over one year. This is a very simple analysis and I hope to expand on this in future posts. I also expect that results can be improved with more sophisticated ranking systems (this feature now available with Portfolio123). I should point out that this strategy works with 2 long and 2 short holdings. But results are not good with larger number of ETFs.

Steve

Comments

[shaun]

Long-short is the way to go in this environment. I am researching long/short stock systems but for the average investor long/short is best done with ETFs. There is a lot of good work being done on the Motley Fool's mechanical investing forum related to a momentum based system called 'Riding the Wave'. I haven't gotten deep into the P123 ETF functionality yet and thank you for getting me interested.

[StockMarketStudent]

Shaun - is there a link you can point me to on Motley fool? When I did a search I came up with all sorts of stuff but no mechanical rules...

Steve

[Jerrod]

Steve, searchfor "Riding the Wave" on the MF Mechanical Investing forum. The key person behind this strategy is is Zeelotes. He gave an overview here: http://tinyurl.com/ccnu66 . Basically, it's a ranking of ETFs based on two moving averages (50 & 235 days in the initial formulation).

Zee had kept a blog at: http://www.zeelotes.org/blog/, complete with downloadable spreadsheets, but it appears that he has stripped it of all content.

I have developed a fancy spreadsheet that implements RTW with a bunch of bells and whistles that were not in the original. [I've mentioned it to you before.] Here's the history starting with $100,000 on 1/1/04 (no friction):

Portval 517,426
CAGR 36.5 %
DDMax 30.8 %
DDAvg 8.3 %
DDRecovery 101 days (time to recover from average drawdown at a rate CAGR
Trades 34
of which 53 % Profitable
AvgHold 57 days
Turnover 6.4 times/year

[StockMarketStudent]

Jerrod - what's the average profit per trade for your version?
Steve

[Jerrod]

Arithmetic average = ( EndValue / BeginValue) ^ (1/Trades) - 1
= (5.174)^(1/34) - 1 = 4.95%

A minor tweak in the model parameters gives virtually the same CAGR, but with only 25 trades (at the expense of somewhat worse DD stats). IN that case,
= (5.161) ^ (1/25) - 1 = 6.78%

BTW, there is a wonderful search engine for the Motley Fool Mechanical Investing forum:
http://www.datahelper.com/mi/search.phtml