Sunday, August 24, 2008

A Look at a Few Market Themes

With increasing globalization, every asset class is related to every other one. This makes it useful for understanding how these relationships shift over time. Many times, movements in one asset class will anticipate movements in others. With an understanding of how these different groups are moving--stocks, bonds, interest rates, and commodities--we can appreciate the themes that are driving markets that day.

Since 2007, on days when the euro (FXE) has risen versus the U.S. dollar, the average price change in the S&P 500 Index (SPY) has been -.15% (107 occasions up, 112 down). When the U.S. dollar has risen versus the euro, the average price change in the S&P 500 Index has been .14% (112 up, 83 down).

On days in which the euro has risen versus the U.S. dollar, the average price change for commodities (DBC) has been .42% (137 occasions up, 82 down). When the dollar has risen versus the euro, the average price change in commodities has been -.22% (91 up, 104 down).

When the euro has risen versus the dollar, the average price change for oil (USO) has been .59% (130 occasions up, 89 down). When the dollar has risen versus the euro, the average price change in oil has been -.32% (88 up, 107 down).

When short-term Treasury notes (SHY) have risen in price (yields falling), the average change in SPY has been -.42% (83 up, 136 down). When Treasury notes have fallen in price (yields rising), the average price change in SPY has been .44% (136 up, 59 down). When short-term notes have risen in price, USO has averaged a gain of .24% (123 up, 96 down). When short-term notes have fallen, USO has averaged a gain of .07% (105 up, 90 down).

These are not permanent relationships that serve as the basis for mechanical trading. Rather, they are correlations that reflect broad themes such as recession concerns, inflation worries, and the like. By understanding how asset classes are moving--and moving relative to each other--we can gain insight into the mindsets of institutional money managers.
.

4 comments:

Brandon Wilhite said...

I appreciate your comment about the correlations not being something we can use as raw signals from which to trade mechanically. Perhaps it could be done with a sophisticated enough algorithm....

Point is, if you follow these markets, there are times when it is fairly obvious the correlations are kicking in. Just because the relationship didn't hold consistently over the last 10 years, doesn't mean it isn't occurring. The more I learn, the more I realize that there's a lot of data that is not on the price chart. How do you account for all of that without a very sophisticated algorithm? (I don't know.) I also find myself moving more and more to a true mixture of technical and fundamental trading. It's one of those things that you hear people talk about and sounds just like lip-service...until you've been around awhile (I guess).

Anyway, thanks for the great post.

BW

Bill aka NO DooDahs! said...

There is increasing economic globalization, and recently many asset classes have shown higher correlation. These are coincidental, however, and not causally related, as the opening sentence implies.

If you check various multi-year return streams over the past 50-75-100 years, you will find similar periods of multi-asset-class correlation ... followed by long periods where they are NOT correlated.

Likewise, this period of correlation will be followed by long periods of non-correlation.

People trying to trade on correlation are forced to constantly re-assess their methods and assumptions, as opposed to those who trade on basics founded upon behavioral finance, whose tenets won't change (because people don't change).

It's just the way markets work ...

nEveR bORn...NevER diED. said...

Dr. Steenbarger. What a nice post! I have been using these correlation from past few months very successfully to trade FX market. That thing really works nicely for me. Again thanks.

Brandon Wilhite said...

Bill,

I have to respectfully disagree with your conclusion that the correlations do not imply a causal relationship. In fact, I would argue that they occur *because* of a common causality. What is this causality? It's the rotation of money from one asset to another, and this rotation is based upon the entire economic, social, and political backdrop of the period(s) in question. That's what I mean by saying that there is a lot of data that is not on the chart.

I know that this is a very hard position to argue, especially for the empirically minded (...and I know why). However, I believe it to be true and would happily marshal several arguments for it.

BW