correlation-analysis

To test if two stocks are cointegrated for pairs trading, apply Engle-Granger and Johansen tests to verify long-run equilibrium, then estimate half-life to assess spread mean-reversion before generating entry signals.

The best way to find correlated assets across an equity universe is performing a correlation scan using Pearson or Spearman metrics to filter candidates, followed by hierarchical clustering on the correlation matrix to reveal sector structures and diversification gaps.

To calculate dynamic hedge ratios for pair trading, apply a Kalman filter to your asset return series to obtain time-varying hedge ratios and spreads, which adapt to changing market conditions better than static beta calculations.

Yes, rolling correlation can analyze market regime shifts by generating rolling correlation series and computing conditional statistics across different market regimes to identify structural breaks in asset co-movements.

Cointegration tests are used instead of correlation for pair trading signals because correlation only measures short-term co-movement direction, while cointegration verifies a long-run statistical equilibrium, ensuring the spread will mean-revert rather than trend apart indefinitely.

Sector clustering works for identifying cross-market diversification gaps by applying hierarchical clustering to a correlation matrix of global market returns, revealing sector structures and lead-lag relationships even across FX-adjusted cross-market linkages.