Saturday, October 20, 2007

Dow Industrials vs. Russell 2000: A Long-Short Illustration

Here's a chart of the Dow Jones Industrial Average (DIA) denominated in shares of the Russell 2000 Index (IWM). The charts we're accustomed to denominate shares in dollars. When we use a second market as the denominator, we create a new instrument that reflects the relative strength of the numerator with respect to the denominator.

In other words, what we're looking at is the price performance of a holder of a long-short position (long DIA, short IWM). Note that this looks like a pretty nice bottom pattern on a chart; DIA has been in an uptrend vs. IWM for over a year now.

When you trade a long-short portfolio, the relationships between markets become your trading instrument. When those markets are highly correlated, your portfolio is thus hedged against general market risk. A long-short portfolio is an excellent way to exploit market themes without having to time or crystal ball general market direction.

Moreover, your new instrument--the relationship between markets--has its own historical price patterns that you can identify and exploit, no less than an individual stock or index. More on this to come shortly.

RELATED POSTS:

2007 Performance By Style

2007 Performance By Style and Region
.

3 comments:

Brandon Wilhite said...

Pairs trading :) It's amazingly hard to find good information on this style of trading (there are only 3 books I know of). Trading currencies has helped me to think more in this manner, since in currencies buying one side always means selling the other. So when I see the S&P go down, I always think..."those people have sold their S&P and bought dollars." Of course, then the dollars could flow somewhere else...

An idea I've had recently is this: The various indexes are all determined by someone. So those trading instruments, such as S&P or Russell 2000 are manufactured. The extension of my thought is "What if we could create/find our own indexes?" In other words, if we could create our own trading instrument wouldn't this be an advantage? So take the pairs trade to the next logical step, and instead of just one instrument vs. another maybe search for a set of instruments in various ratios. I realize curve-fitting could become a problem here, but I think the idea has some promise. What do you think?

BW

Brett Steenbarger, Ph.D. said...

Hi Brandon,

Yes, this is a very promising idea and one that Pete Steidlmayer has been pursuing. The idea is to create synthetic indices that have better trading properties than the established alternatives.

Brett

Brandon Wilhite said...

That's exactly the idea. I'll have to look up his work. Thanks.

BW