The Truth about Intermarket Analysis

Article excerpt

Intermarket analysis works, but only if you keep it simple. While it appears more popular than ever, here's an update on how strategies published eight years ago still work today.

Many of us were introduced to intermarket analysis the study of related markets for clues to their future price direction - by John Murphy's 1991 book Intermarket Technical Analysis (John Wiley & Sons). The book demonstrated the relationships between the various markets and their interrelated connections. It showed how the S&P 500 correlates to the Treasury bond market, how Treasury bonds are related to commodity prices, how commodity prices affect the value of the dollar and how the value of the dollar relates to stock prices.

Taking Murphy's research to another level, I examined mechanical trading systems that were based on these intermarket relationships and many have been published in Futures. I also have validated Murphy's theories and published practical trading systems using intermarket analysis.

Although Murphy's theories proved valid and the systems worked, the entire concept of intermarket analysis has endured some criticism as a form of curve fitting. Unfortunately, the concept has been abused. Some analysts have taken a shotgun style approach, using massive banks of neural networks to data mine spurious relationships that are then billed as a valid trading system.

The bottom line is simple: If a fundamental reason for price relationship cannot be identified, then trading off it is a high-risk proposition.


Intermarket relationships are not always constant. They ebb and flow. As with any type of market analysis, the goal is to isolate an edge over the market. It is unrealistic to expect any form of market analysis to be perfect. As analysts, we are seeking only a small - albeit elusive - statistical edge.

Intermarket relationships often decouple for various fundamental reasons. This decoupling does not mean the relationship is invalid or unreliable. In fact, if the research is done correctly, it can be understood why such a breakdown takes place.

A classic example occurred in 1993 when the Commodity Research Bureau index and the 30-year Treasury bond market rallied through the summer. The relationship between these two markets is normally negatively correlated. In this case, falling interest rates caused T-bond prices to rise so rapidly that it caused an intermarket inversion. Often, two to three sigma moves cause this type of situation. In this particular case, the logic was that significantly lower interest rates stimulated the economy causing inflationary fears and commodity prices to rally at the same time.


The T-bond market is one of the best markets to use for Intermarket analysis. T-bonds are to intermarket analysis what currencies are to trend-following. A typical trading system for an intermarket T-bond system is shown in "What goes around" (left).

This simple model uses a second market that is negatively correlated to T-bonds. This also means the second market is positively correlated to inflation. To demonstrate the effectiveness of this technique, we can take a trading system that was originally published in "Intermarket analysis is fundamentally sound," April 1998, and update the results through today. In that article the data used to design the system were through Dec. 30, 1997.

We used the silver market to predict price changes in the T-bond. The "classic" parameter values were 14 for T-bonds and 26 for silver. This system has performed well since 1998, with the exception of 2003 (see "After the fact," above). In 2003, silver and T-bonds rallied together in other words, they decoupled. This was similar to the problems experienced in 1993 except that market volatility caused the losses to be much greater.

Consider the results on a year-by-year basis. No modifications were made to the original system developed using data through Dec. …