Stock Market Correlations

Includes: CME, DIA, QQQ, SPY
by: John Lounsbury

Damien Hoffman at the Wall Street Cheat Sheet has this interesting correlation table for various stock market indices:

Click here for larger image.

I have gone to the CME (Chicago Mercantile Exchange) web site (the source of the chart) and can not determine the basis of correlation. I will assume these are the correlations of daily returns over the two month interval.

These correlations of daily returns may have some value for short-term traders (holding periods of one to several days), but for investors trying to reduce risk in a portfolio over months at a time, the correlations of weekly and monthly returns are more appropriate.

For any time period, daily through monthly, the correlations can shift significantly over time. Investors trying to balance risk must continuously test correlations to avoid drift to higher risk. For an example, look at the chart from the CME (here) for January only:

Click here for larger image.

Four examples of correlation drift are shown in the following table:

Data: Chicago Mercantile Exchange

The significance of these examples of shifting correlations is increased when you recognize that the data that produced the January correlations is included in the data that produced the two month correlations. The February correlations by themselves (not shown) must be significantly different from the January correlations to produce the effects shown.

Using correlations may be useful in trading and establishing portfolio risk balance, but only if it is recognized that correlations can be fickle. They must be checked constantly. If you set and forget correlations, they might put your portfolio in a coma.

Disclosure: No positions in indexes or stocks mentioned at present.