Correlation

A measure from −1 to +1 of how much two assets move together. Key to avoiding doubled-up risk when you open several seemingly different positions.

Correlation measures the degree to which two assets move in the same direction. The Pearson coefficient ranges from −1 to +1. A value of +1 means they always move together and proportionally; 0, that there is no linear relationship; −1, that they move in opposite directions perfectly. It is computed on returns (percentage changes), not on price levels, which can produce spurious correlations.

In trading it is crucial because opening two positions in correlated instruments is, in part, opening the same position twice. EUR/USD and GBP/USD usually show a high positive correlation, often above 0.7 although it changes over time; EUR/USD and USD/CHF tend to correlate negatively because both revolve around the dollar. Buying EUR/USD and selling USD/CHF amounts to a double bet against the dollar.

For two positions with the same risk r and correlation ρ, the joint volatility of losses is r × the square root of 2(1 + ρ). With r = 1% per position: if ρ = 1, combined risk is 2%; with ρ = 0.8, it is 1% × the square root of 3.6 ≈ 1.90%; with ρ = 0, ≈ 1.41%; with ρ = −0.8, ≈ 0.63%. In other words, two highly correlated 1% positions behave almost like one 2% position, and this formula is a statistical approximation, not a cap on maximum loss.

Correlation is not stable. It moves toward +1 or −1 in moments of stress, when assets are driven by a common factor such as an inflation print or a flight to the dollar, and weakens in quiet periods. So the diversification that seems to exist in calm times can vanish exactly when you need it most. Compute correlation over different windows, for example 30, 90 and 250 days, and distrust short samples.

Distinguish correlation from causation: two series can move together by coincidence or because both respond to a third factor. And low correlation does not imply independence, since the measure only captures linear relationships. Do not confuse correlation between assets with correlation between the results of your own strategies, which is also worth monitoring.

Frequent errors: trading four dollar pairs at once without noticing they make nearly the same bet, counting each position as independent risk, and trusting a correlation calculated from little data. A typical case: longs on EUR/USD, GBP/USD, AUD/USD and NZD/USD, which are really one single short-dollar position.

In risk management, add up the risk of same-direction correlated positions and apply an overall limit, for example 2% or 3% of the account, instead of only a per-trade limit. Before opening a new position, ask whether it increases or reduces your exposure to one and the same factor.

More terms