Correlation Coefficient

Correlation is a statistical relationship between several variables. We determine the correlation by calculating Pearson's linear correlation coefficient (hereinafter referred to as the correlation coefficient).

The equation for the correlation coefficient (r) is the following:

We calculate correlation coefficients by comparing the following variables:

prices of financial instruments and economic indicators;
changes in economic indicators and price movements of economic instruments;
changes in price dynamics of two or more different economic instruments.

Correlation is calculated both with and without time shift. For example, when calculating a correlation between oil prices and EURUSD rates, we can compare a movement in the oil price for today with the change in the EURUSD rate for tomorrow or the day after tomorrow.

A correlation identified with a time shift allows forecasting future price dynamics and provides an opportunity to make profits.

We use the popular Cheddock scale to evaluate the obtained values of correlation coefficients in the research.

Correlation coefficient	Interpretation
less than 0.1	no relationship
0.1 to 0.3	weak relationship
0.3 to 0.5	moderate relationship
0.5 to 0.7	significant relationship
0.7 to 0.9	strong relationship
0.9 to 0.99	extremely strong relationship
0.99 to 1	functional relationship

Notes

Correlation coefficient is sensitive to price outliers. A single random value can significantly distort the coefficient. So, a large sample of data is needed to ensure that random fluctuations are canceled out.
A correlation does not imply a causal relationship between quantities.
An absence of linear correlation does not mean that there is no relationship. The relationship may be non-linear.