Correlation measures how closely two assets' returns move together, on a scale from -1 (perfectly opposite) to +1 (perfectly in sync). Diversification — combining assets with low or negative correlation — is the only way to reduce portfolio risk without sacrificing expected return.
Why It Matters
Diversification is often called the only free lunch in investing, but it only works if the correlation assumptions behind it hold. Portfolio construction, hedge ratios, and risk-budgeting all depend on correlation estimates — which is exactly why the failure of diversification in a crisis (when everything sells off together) is one of the most consistent and damaging patterns in market history.
How It Works in Practice
- 1Calculate the covariance between two assets' return series, then normalize by their individual standard deviations to get a correlation coefficient
- 2Portfolio-level risk depends not just on individual asset volatility but on the full correlation matrix across holdings
- 3Adding a low- or negatively-correlated asset can reduce total portfolio volatility even if that asset is individually riskier than what's already held
- 4Correlations are re-estimated on rolling windows since they are not stable over time
Common Pitfalls
Correlations tend to rise sharply during market stress — 'diversified' portfolios often behave like a single asset exactly when investors most need the diversification to work
Historical correlation is not a guarantee of future correlation, especially across asset classes with structurally different liquidity
Adding more holdings that are all correlated to the same underlying factor (e.g., several equity sectors) creates the illusion of diversification without the actual risk reduction