Sector vs. index volatility dispersion & implied correlation regime
Dispersion= weighted-average sector realized vol − S&P 500 (SPY) realized vol. Calculated using a 21-day rolling window of annualized log-return standard deviation.
Realized Correlation Proxy is derived from the variance decomposition identity: σ²index = Σ w²ᵢσ²ᵢ + ρ̄ · ((Σ wᵢσᵢ)² − Σ w²ᵢσ²ᵢ). When ρ̄ is low, sectors move independently → stock-picking alpha is higher.
Regime: LOW CORR (below median − 5%) favors single-stock strategies; HIGH CORR (above median + 5%) indicates macro-driven, correlated markets.
Dispersion measures how differently sectors in the S&P 500 are performing from each other. When dispersion is high (low correlation), sectors are moving independently -- this is when stock-picking and sector bets matter most. When dispersion is low (high correlation), everything moves together, usually driven by macro news like Fed decisions or recession fears. In a high-correlation market, individual stock selection adds less value because the "rising/falling tide" dominates.