Correction on the paper “Bayesian Estimators for Small Area Models Shrinking Both Means and Variances”
August 31, 2017
The first inequality in (12) is not correct since X′H′1P′2P2H1X = 0. This document gives the corrected evaluation of the integral of
π(θ, γ|D) ∝(θ′Aθ)−(m−p−2)/2
m
∏
i=1
γaiψi(θi− Xi, γ)−(ni/2+ai),
where ψi(θi− Xi, γ) = (Xi− θi)2+ (ni− 1)Si2+ 2biγ.
Define Ω = {θ|θ′Aθ≤ 1} ⊂ Rm. It holds that
∫
Rm×R+
π(θ, γ|D)dθdγ =
∫
Ω×R+
π(θ, γ|D)dθdγ +
∫
Ωc×R+
π(θ, γ|D)dθdγ.
The second term can be evaluated as
∫
Ωc×R+
π(θ, γ|D)dθdγ ≤ C
∫
Ωc×R+
m
∏
i=1
γaiψi(θi− Xi, γ)−(ni/2+ai)dθdγ
≤ C
∫
Rm×R+ m
∏
i=1
γaiψi(θi− Xi, γ)−(ni/2+ai)dθdγ
= C
∫ ∞ 0
m
∏
i=1
{∫ ∞
−∞
γaiψi(θi− Xi, γ)−(ni/2+ai)dθi}dγ,
which corresponds to the last formula in (12), and it is finite. For evaluating the first term, we first note that there exists a (m − p) × m matrix H1 such that A = H′1H1 and H1H′1 = Im−p since A is an idempotent matrix with rank(A) = m − p. By changing the variable as u1 = H1θ and u2 = (um−p+1, . . . , um)′ with ui= θi, it follows that
∫
Ω×R+
π(θ, γ|D)dθdγ ≤ C′
∫
u′1u1≤1
(u′1u1)−(m−p−2)/2du1
∫ ∞ 0
m−p
∏
i=1
γai{(ni− 1)Si2+ 2biγ}−(ni/2+ai)
×
m
∏
i=m−p+1
{∫ ∞
−∞
γaiψi(ui− Xi, γ)−(ni/2+ai)dui }
dγ.
Moreover, it holds that
∫
u′1u1≤1
(u′1u1)−(m−p−2)/2du1 = C′′
∫ 1 0
r−(m−p−2)rm−p−1dr <∞,
thereby the similar evaluation shows that∫Ω×R
+π(θ, γ|D)dθdγ is also finite.
1