The following is an explanation published in a historical edition of the Moore
Research Center Report which explains the use of our correlation charts. Any
questions regarding these charts can be directed to Research Director Nick
Colley or EditorinChief Jerry
Toepke.
Of course, one would expect gold and silver futures to move in tandem, to
correlate closely, with each other. They're both precious metals, aren't they?
But silver is also a by product of copper mining. Would they tend to trade
togetheror contrarily?
Similarly, one might expect price activity in Euros and Tbonds to be similar.
Conversely, the Swiss Franc and the Dollar Index should have opposite reactions
to the same market input.
A primary rationale behind the continuing bull market in stocks has been
declining interest rates. How closely, in fact, have Tbond and SP500 futures
tracked each other on a daily basis?
For years traders have made a very nonfundamental connection between the silver
and soybean markets. How closely have they traded?
Heating oil and crude oilyes. But live cattle and the JYen???
Some brief explanation is required. The singular relationship under
consideration is the frequency of duplicated up/down daily closings. E.g., if
Market A closed higher on Day 1, did also Market B? (In that respect, the study
is qualitative, not quantitative, i.e., the amount is irrelevant.)
Without going into further details of leastsquares and scatterplots, the
precise statistical terminology that describes each relationship is a sample
coefficient of correlation, a number greater than 1 and less than +1. Thus, if,
every day over the sample period, each of two markets duplicated the other's
higher or lower closing, they would have a coefficient of correlation equal to 1/1,
or +1. Conversely, two markets that always closed contrarily to each other
would have a coefficient equal to 1. A 0 value indicates no correlation
whatsoever. For convenience, however, all values in the spread sheet have been
stated in terms of percentages of +1 or 1.
Remember, the amount of change is irrelevant, which can account for differences
in trend.
