Definition of Pair Correlation
Pair correlation is a statistical measure that describes the historical relationship between the price movements (usually returns) of two currency pairs. In forex, it helps explain whether two pairs tend to rise and fall together, move in opposite directions, or show little consistent relationship.
Correlation is commonly expressed as a number between −1 and +1. A value near +1 means the two pairs’ returns have historically moved in the same direction. A value near −1 means they have tended to move in opposite directions. Values near 0 indicate weak or no linear co-movement.
How it works in forex
To compute pair correlation, you first choose what “returns” mean and which time period to analyze. A simple approach is:
- Pick two currency pairs, A and B.
- Choose a time interval (for example, hourly or daily).
- Convert the pair prices into a return series over time for both A and B.
- Apply the correlation formula to the two return series.
Correlation depends on the exact inputs. If you change the return method (price differences vs. percentage changes), the sampling frequency (intraday vs. daily), or the lookback length (how many observations you include), the reported correlation can change.
A practical interpretation is about co-movement, not direction of an individual trade. Even when two pairs have a positive correlation, the magnitude and timing of their moves can differ.
Adjacent concepts that are often confused
Pair correlation is related to, but not the same as, several nearby ideas:
- Coincidence vs. correlation: correlation measures a statistical pattern, while “similar looking charts” can be subjective.
- Correlation vs. causation: high correlation does not show that one pair drives the other.
- Correlation vs. volatility: two pairs can be uncorrelated yet both be highly volatile, or correlated yet move with different risk.
- Correlation vs. hedging: correlation can matter for hedging intuition, but effectiveness is not determined by correlation alone because execution, costs, and position sizing change outcomes.
Limitations and failure modes
Pair correlation is useful for description, but it has important limitations:
- Instability over time: relationships can weaken or flip sign when market regimes change.
- Choice dependence: results vary with the time window, return definition, and sampling frequency.
- Non-linearity: correlation mainly captures linear co-movement; relationships can be non-linear and still appear weak.
- Outliers: extreme market events can strongly influence the correlation estimate.
- No guarantee of future behavior: historical co-movement does not ensure future co-movement.
These limitations matter because a correlation number can look “stable” under one setting yet be unstable under another.
Verification and what to check next
Because correlation results depend on calculation choices, an independent verification approach is to reproduce the computation with clearly stated assumptions:
- Use the same currency pairs and align their timestamps.
- Specify the return definition (for example, percentage returns vs. simple differences).
- Specify the sampling frequency and the lookback window.
- Recompute correlation across multiple windows to see whether it changes meaningfully.
If correlation varies a lot across windows, treat it as a descriptive statistic rather than a stable property. If it remains consistently positive or negative across reasonable parameter choices, the historical co-movement may be more robust—but it still does not create a forward-looking certainty.