Measuring correlations based on log returns, rather than raw prices or simple returns, offers several advantages:
- stationarity: Log returns are more stationary, resulting in more meaningful and reliable results - volatility: Log returns give a consistent measure of relative changes of assets with different volatility
Log returns are time-additive and often more stationary than simple returns, making them statistically more reliable for analyses in financial contexts. Additionally, they provide a consistent measure of relative price changes and align more closely with the assumptions of many statistical models, including normal distribution.