ETF pairs trade strategy

An Almost Risk-Free Way For Annual Return Of 50%+

December 15, 2012 | 101 comments | includes: DPK, DRN, DRV, DUST, DZK, EDC, EDZ, ERX, ERY, FAS, FAZ, FXP, GASL, GASX, MIDU, MIDZ, NUGT, RUSL, RUSS, SOXL, SOXS, SPXL, SPXS, TECL, TECS, TNA, TZA, XPP, YANG, YINN

I recently posted a series of articles investigating the performance of leveraged ETFs, titled A Tale of 3 Leveraged ETFs (Part 1, Part 2, and Part 3). I demonstrated there that there is no real reason for investors to stay away from leveraged ETFs, that neither daily compounding nor market timing is essentially detrimental to the performance of a leveraged ETF when compared to the underlying.

However, I also found some strange patterns when examining leveraged ETFs. Consider the following pair of Leveraged ETFs, the Direxion Daily Natural Gas Related Bearish ETF (GASX) and the Direxion Daily Natural Gas Related Bullish ETF (GASL). These are ETFs that are paired, that is while one is a bullish ETF with 3x leverage, the other is bearish with 3x leverage. Over a year, the gains in one should be about the same as the loss in the other. If an investor were to buy the two ETFs in pair, over a year there should be very little upside or downside.

That, however is not the case. Both the ETFs are down, significantly. If an investor were to short both and held for a year, they would get a sizable return. The return is even bigger when you consider daily compounding.

GASX Chart

I constructed a portfolio with $10,000 in cash, and short $10,000 of GASX and $10,000 of GASL. The portfolio initiates on December 9, 2011, and is held to December 9, 2012. Here is what such a portfolio would have returned.

(click to enlarge)

The 1-year gain is 57.3%, but there are times when this portfolio will drop up to 16%. This got me curious, and I decided to try this out for all pairs of leveraged ETFs offered by Direxion over the time period December 9, 2011 to December 9, 2012. The ETFs included are SPXL, SPXS, DRN, DRV, MIDU, MIDZ, XPP, FXP, TECL, TECS, FAS, FAZ, TNA, TZA, ERX, ERY, DZK, DPK, YINN, YANG, EDC, EDZ, SOXL, SOXS, RUSL, RUSS, GASL, GASX, NUGT, and DUST.

Here is the result:

Pair Standard Deviation Total Return Max Risk Correlation

SPXL-SPXS 2.5% 1.0% -10.5% -99.8%

DRN-DRV 2.7% -1.3% -20.4% -99.3%

MIDU-MIDZ 2.8% 4.8% -7.5% -99.7%

XPP-FXP 2.9% 16.8% -1.3% -99.8%

TECL-TECS 3.0% 18.2% -23.6% -99.8%

FAS-FAZ 3.2% -14.2% -25.4% -99.9%

TNA-TZA 3.3% 18.1% -5.6% -99.9%

ERX-ERY 3.5% 30.1% -2.2% -99.9%

DZK-DPK 3.5% 26.7% -0.1% -99.1%

YINN-YANG 3.9% 43.9% -4.5% -92.3%

EDC-EDZ 3.9% 27.6% -7.0% -99.9%

SOXL-SOXS 4.4% 46.2% -10.4% -99.7%

RUSS-RUSL 5.4% 61.8% -26.6% -98.7%

GASL-GASX 5.5% 57.3% -16.0% -98.9%

NUGT-DUST 5.8% 63.1% -55.3% -99.3%

It seems that the more volatile the pair, the more is the excess return end of year! Note that this is not due to a lack of correlation between the pairs, all of them have more or less that same very high correlation. However, with great power comes great responsibility (so said Spiderman), and with great Total Return comes great Max Risk. Many of these portfolios would be down significantly at some point during the year.

(click to enlarge)

If you look at the trend line, the R2 is ridiculously high at 82.5%. This implies as strong a correlation of standard deviation to total return as you can ever imagine.

But the problem of high downside risk needs to be solved. So, I decided to model a synthetic portfolio consisting of the 3 pairs of ETFs with the highest standard deviation, to see what that does to the maximum downside. This is how their relative correlations shake out

RUSS-RUSL GASL-GASX NUGT-DUST

RUSS-RUSL 100.0% 74.5% 52.8%

GASL-GASX 74.5% 100.0% 77.4%

NUGT-DUST 52.8% 77.4%

100.0%

The 3 pairs are somewhat correlated to each other, but not a lot. A little bit of modeling, and it turns out that the magic is in creating a portfolio that is 50% in the RUSL-RUSS combo, 30% in the NUGT-DUST combo, and 20% in the GASL-GASX combo. In other words, the portfolio will have $100,000 in cash, and short $50,000 each of RUSL and RUSS, $30,000 each of NUGT and DUST, and $20,000 each of GASL and GASX.

So how does the risk-return profile look like for this portfolio? Well, it seems that at the end of the year, the portfolio would gain in excess of 60%, and will never fall by more than 2%. This is as close to risk free money as one can get.

(click to enlarge)

So far, so good. But of course no strategy is worth looking at till it is back tested. So, I decided to take the same approach, and test my theory in the time period December 9, 2010 to December 9, 2011. Here's how that looks. The max risk is slightly less than 3%, and the total return is just slightly less than 60%. Still pretty close to risk free money.

(click to enlarge)

I have started a test for this strategy on December 10, 2012, and so far so good. The portfolio is up by about 1.5% this week, and has never been down. It should not be down anyway, as this is a fully hedged strategy. I will keep posting periodic updates on this portfolio.

Disclaimer: This is not meant as investment advice. I do not have a crystal ball. I only have opinions, free at that. Before investing in any of the above-mentioned securities, investors should do their own research, consult their financial advisors, and make their own choice. I am merely stating what I personally plan to do for my own portfolio.