ditions for determinacy of the difference system are
(Case I)
a> 0.5 and 1 <μ< min∣ΓC, ΓC, ΓC |;
(42)
(Case II)
0.5 > a > -----
2-δ
2(1+β)+
1-δ-
(μ - 1)Л1
Λ2 (2μ(1 — a) — 1)
and
(43)
[Λ2 - 2α(1 -2a)(2-Λ2)] <0;
where Γ1C ≡
__1__ γc =
2(1-a), 1 2 ≡
(l-β)Λ2+Λ1α(2a-1)
Λια(2a-1)+(1-β)Λ2 2(1-a)
, Γ3C ≡
2(1 + e)A.2+A.1 [Λ-2+(2a-1)α(2-λ2 )]
4(1+β)(1-a)Λ2+Λ1 [Λ2+(2a-1)α(2-Λ2)] '
Suppose α = 0.36, β = 0.99, δ = 0.025 and λ = 7.66. Given the assigned parameter
values conditions (41) and (43) of Propositions 5 and 6 are violated if a < 0.4716 and thus
(first-order) indeterminacy arises ∀μ > 1. If a > 0.5 then under domestic price inflation
targeting the open-economy introduces no additional requirements for determinacy. This
follows by direct comparison of the upper bounds on μ given by conditions (39) and (40):
Γ1A < Γ1B and Γ2A < Γ2B. However if a > 0.5 and consumer price inflation is targeted then
comparing (39) with (42) yields ΓC < ΓA and ΓC < ΓA. Since ∂Γf/∂a > 0 for i = 1, 2, 3,
the inflation coefficient μ is constrained by these upper bounds, all of which are increasing
with respect to a. Thus the range of indeterminacy is potentially greater the higher the
degree of trade openness (i.e. the lower is a).
As discussed by Kurozumi and Van Zandweghe (2007) in a closed economy an active
forward-looking policy makes inflation expectations self-fulfilling entirely because of the
cost channel of monetary policy. However in the open-economy indeterminacy is more
severe because of the additional impact the trade channel has on inflation. Under an active
forward-looking policy, the increase in the real interest rate results in a future expected
deterioration in the terms of trade (Tt+1 increases relative to Tt). Thus the trade effect
puts upward
Etπt+1 = E⅛ + (1 - a) (Etft+1 - ft)
pressure on inflation, the effect of which is stronger the higher the degree of trade openness
(a I).
23