3.10.4 Joint-normal Distributions

3.10.4  Joint-normal Distributions

Let X be an n-dimensional random vector with mean vector μ and covariance matrix Σ. Suppose the marginal distribution of each component X_i is normal. Let Y be a random variable defined as a linear polynomial

[3.121]

of X. Using [3.27] and [3.28], we can calculate the mean μ_Y and standard deviation σ_Y of Y. Knowing only that the marginal distributions of the X_i are normal, there is little more we can say about the distribution of Y. However, there is an additional condition we can impose upon X that will cause Y to be normally distributed. That condition is joint-normality.

The definition of joint-normality is almost trivial. A random vector X is said to be joint-normal if every nontrivial linear polynomial Y of X is normal. Joint-normal distributions are sometimes called multivariate normal or multinormal distributions.

We denote the n-dimensional joint-normal distribution with mean vector μ and covariance matrix Σ as N_n(μ,Σ). If Σ is positive definite, it has PDF

[3.122]

where |Σ| is the determinant of Σ. Exhibit 3.20 illustrates a joint-normal distribution in two random variables X₁ and X₂. If we define Y = X₁ + X₂, then Y is normal.

Now let’s illustrate how a random vector may fail to be joint-normal despite each of its components being marginally normal.^[1] Let X be a two-dimensional random vector with components X₁ and X₂. Let X₁ and Z be independent N(0,1) random variables, and define X₂ = |Z|sign(X₁).^[2] In this case, both X₁ and X₂ are N(0,1), but the vector X, whose PDF is illustrated in Exhibit 3.21, is not joint-normal. In this case, the random variable Y = X₁ + X₂ is not normal. Instead, it has the PDF illustrated in Exhibit 3.22.

A random vector is joint-normal with uncorrelated components if and only if the components are independent normal random variables.

A property of joint-normal distributions is the fact that marginal distributions and conditional distributions are either normal (if they are univariate) or joint-normal (if they are multivariate). Specifically, let X ~ N_n(μ,Σ). Select k components. Without loss of generality, suppose these are the first k components X₁, X₂, … X_k. Let X₁ be a k-dimensional vector comprising these components, and let X₂ be an (n – k)-dimensional vector of the remaining components. These partition X, μ and Σ into sub-vectors and sub-matrices as follows

[3.123]

The marginal distribution of X₁ is N_k(μ₁,Σ_1,1) and that of X₂ is N_n–k(μ₂,Σ_2,2). If Σ_2,2 is positive definite, the conditional distribution of X₁ given that X₂ = x₂ is

[3.124]

If X ~ N_n(μ,Σ), b is a constant m × n matrix and a is an m-dimensional constant vector, then

[3.125]

This generalizes property [3.94] of one-dimensional normal distributions.