On Wednesday we held a webinar to discuss the statistical arbitrage trading parameter "Beta". We had good turnout, and a lot of you requested the replay (at the bottom of this post). We had some great questions, and I wasn't able to answer them all during the webinar. One of them in particular I wanted to cover here, because it was quite thoughtful and insightful. In answering it, I thought about the many interesting ways to look at spread trading.
The question was about my selection of spread. I talked about FDX-UPS, and you can see a chart of the spread
here. This spread is not cointegrating, so the question asked why I had chosen it in a discussion of statistical arbitrage spreads. I like that spread because the companies are so similar, and I make money trading it when it is statistically cheap or expensive. I do not worry about the cointegration, because for me the company relationship and the fundamentals are my spring force. I use the other statistical parameters such as zscore, half-life, and beta in my trading.
The person who asked the question suggested that the 1 year chart showed positive autocorrelation (which to me means it is trending higher). Betting on mean reversion is fundamentally different from betting on momentum (trends). Both views can be expressed using spreads. There is no way to know for sure which way the spread will go in the future. For this reason, some traders (including me) look at additional factors such as fundamentals or technicals to help predict the future performance.
What do you think of FDX-UPS? Will it continue to trend higher, or revert back to the mean?
Article originally appeared on (http://biggercapital.squarespace.com/).
See website for complete article licensing information.