Correlation describes co-movement#
Correlation summarizes how two series moved over a chosen window. The result changes with timeframe, sampling method, regime and outliers, so it is not a permanent property.
Causation needs a defensible mechanism#
A causal claim needs more than simultaneous movement. Timing, mechanism, alternative explanations and independent evidence must support the relationship.
Why regimes break relationships#
Rates, policy expectations, liquidity and market stress can change how assets interact. A relationship observed in one period may weaken or reverse in another.
Sources and references#
Use these direct research and official references to verify definitions, scope and limitations.
- Correlation — U.S. National Library of Medicine
- Causation — U.S. National Library of Medicine
- No Contagion, Only Interdependence: Measuring Stock Market Co-movements — National Bureau of Economic Research
- Real-Time Price Discovery in Stock, Bond and Foreign Exchange Markets — National Bureau of Economic Research
Key principles#
Window sensitivity
Daily, weekly and intraday samples can produce different correlation readings.
Third-factor risk
Two assets may react to the same macro event without one causing the other.
No predictive shortcut
Historical co-movement does not establish a future price path or personal action.