3.34 Let X1,...,Xm and Y1,...,Yn have joint density 1 m n f x1 ,..., xm...

3.34 Let X1,...,Xm and Y1,...,Yn have joint density 1

σ mτ n f x1

σ ,..., xm

σ ;

y1

τ ,..., yn

τ

, and consider the problem of estimating θ = (τ/σ)

r with loss function L(σ, τ ;

d) =

γ (d/θ). This problem remains invariant under the transformations X

i = aXi, Y 

j = bYj ,

σ = aσ, τ  = bτ , and d = (b/a)

rd

(a, b > 0), and an estimator δ is equivariant under these transformations if δ(ax, by)=(b/a)

r δ(x, y). Generalize Theorems 3.1 and 3.3, Corollary 3.4, and (3.19) to the present situation.