Let l and m be natural numbers with l≥2 and m≥1, let A be a nonempty subset of Euclidean space Rm, and let β be a transition-rate family on l states with control set A and rate bound B. The aggregate fluctuation covariance of β is the function Θ assigning to each (Σ,α) in the probability simplex Δl times the control set A the real matrix Θ(Σ,α) of size l×l with entries
Θγγ(Σ,α)=σ:σ=γ∑(Σσβ(σ,γ,Σ,α)+Σγβ(γ,σ,Σ,α))(γ∈{1,…,l}),
where the sum runs over σ∈{1,…,l} with σ=γ, and
Θγδ(Σ,α)=−Σγβ(γ,δ,Σ,α)−Σδβ(δ,γ,Σ,α)(γ,δ∈{1,…,l}, γ=δ).