joint distributions

A joint distribution describes two or more random quantities together, capturing not just how each behaves on its own but how they move in relation to each other. Knowing the distribution of a policyholder's medical claims and, separately, their disability claims is not enough — you also need to know whether the two tend to happen together. The joint distribution is the full picture of two variables at once.

For two variables it gives the probability (or density) of every combination of values: P(X = x and Y = y) for discrete cases, a joint density f(x, y) for continuous ones. From the joint distribution you can recover each variable on its own — its marginal distribution — by summing or integrating over the other. You can also ask conditional questions: given Y took a certain value, how is X distributed? A vital special case is independence: X and Y are independent exactly when their joint distribution factors into the product of the two marginals, meaning one carries no information about the other.

Joint distributions matter to actuaries because real risks rarely live in isolation. The losses on different policies, the lifetimes of a husband and wife in a joint-life annuity, claim frequency and claim severity, asset returns and liability values — all may be linked, and pricing or reserving as if they were independent can be badly wrong. The joint distribution is the honest foundation for measuring that linkage; covariance, correlation, and modern copula methods are all ways of summarizing or modeling the dependence structure it contains.