FRM Glossary · Part I · Quantitative

Copula

A function that joins univariate marginal distributions into a joint distribution — the standard tool for modelling asset dependence.

In more detail

Sklar's theorem: any joint distribution can be written F(x_1, …, x_n) = C(F_1(x_1), …, F_n(x_n)). The Gaussian copula assumes normal dependence and was at the centre of the 2008 financial crisis because it understates tail dependence. t- and Clayton copulas capture stronger co-crash behaviour and are standard in credit-portfolio modelling.