Meaning and calculation
Return correlation measures the linear relationship between two series. In the usual calculation, it is related to their covariance and standard deviations. Comparisons require the same dates, frequency, currency and treatment of distributions. Sharpe’s materials show the role of correlation in portfolio variance, but do not treat it as a permanent property of asset names.[104](#source-cv-covariance)
An illustrative two-asset model
In a hypothetical two-asset model, changing the specified correlation while holding the weights and individual volatilities constant changes the overall dispersion. This is a mathematical consequence of the model, not proof that a relationship measured over a past window will persist during a crisis.
Historical measure, not protection
Zero correlation does not guarantee independence of all outcomes, and negative correlation does not promise that the second asset will gain every time the first declines. Nonlinear risks, rare events and the ability to sell require separate analysis. Infrequently updated prices can distort the estimate. Keep the period, sources, missing-data rules and limitation with the metric: it is a measured historical relationship, not contractual protection for a portfolio.