Be careful when using correlation to trade


Many traders use correlation to form their market outlook and trade the markets. I don’t. I use correlations between markets to form my discretionary outlook, but I never trade based on these correlations. I only look at correlations between assets with a passing interest. Here’s why.

If one market’s outlook is wrong, the others will be wrong

Traders that use correlation will say “if market X goes up, market Y will go down, so market Z will go down. I think market X will go up, so I’m bearish and shorting markets Y and Z”. They say this because markets X are Y are inversely correlated while markets Y and Z are positively correlated.
Here’s the problem.
If your outlook for any one of these correlated markets is wrong, all the other market outlooks will be wrong as well.
So if 3 markets are correlated right now, then your market outlook is either right for all 3 of them or wrong for all 3 of them. And as long as your outlook is wrong for 1 market, your outlook for the other markets will be wrong as well.

It’s better to understand why a correlation exists. Otherwise the correlation can break at any time.

It’s better to understand the market’s fundamentals that are DRIVING these correlations. That way you can predict when the correlations will break (i.e. when the fundamentals change).
Remember: correlation is not causation. Correlation is sometimes a sign of causation, but not always. Correlations exist sometimes for purely random reasons. E.g. the U.S. stock market and chicken breast sizes have had a positive correlation over the past 100 years. Do rising chicken breast sizes = a rising stock market? No. There is no causation.
Correlations are useful when 2 markets are being driven by the same fundamental reasons. E.g. there is an inverse correlation between the U.S. stock market and U.S. dollar right now. That’s because both of these markets are being driven by the same fundamentals: rising inflation. Historically, rising inflation causes the stock market to rise and the U.S. Dollar to fall.
This correlation will be valid AS LONG AS the fundamental driver is still valid. Hence, you can use one market’s short term outlook to predict the other market’s short term outlook.
If a correlation exists purely due to randomness, then you can disregard the correlation because it can break at ANY TIME. You cannot use a correlation that can break at any time to trade. You won’t know if the correlation is broken until it has already been broken for a long time.
Here is an example of a correlation that can break at any time.

On the other hand, you can use a correlation to trade if it exists for a fundamental reason.

Correlations can conflict with each other

Sometimes 2 correlations will conflict with each other.
Let’s assume that markets X and Z are positively correlated with each other. Let’s assume that markets Y and Z are positively correlated with each other.
Let’s assume that you’re bullish on market X and bearish on market Y.

  1. You should be bullish on market Z based on the first correlation.
  2. You should be bearish on market Z based on the second correlation.

Which correlation is more important than the other? Which correlation should override the other? It’s hard to say.

Correlations are rarely extremely strong

2 markets are rarely EXTREMELY positively correlated or EXTREMELY inversely correlated. Most of the time there is a moderate or weak correlation. Moderate and weak correlations aren’t very useful. A market can swing back and forth from “no correlation” to being “moderately correlated”.

Conclusion

You can only use correlations as a supplement to your market outlook. It cannot be the main reason for your market outlook. Correlations should be of secondary importance in the list of factors that you consider.
I prefer to look at the market’s own fundamentals and technicals when trading.

1 comment add yours

  1. Divergences
    Many clues can be gained by watching the S&P market in relationship to other markets or indexes. There is very often a leading/lagging relationship with the following: the NASDAQ, the Dow Jones Industrial Average, the Transportation Index, the bond market, and even the OEX’s. These types of observations are another form of tape reading. How is the market acting? Is it holding together as the bonds are making new lows? What does this tell you? Is the S&P failing to confirm a new afternoon low made by another index? This may be a sign of relative strength.

Leave a Comment