890 Macroeconometric Modeling for Policy
–1.0 2007–0.8–0.6–0.4–0.20.00.20.42008 2009 2010 2011 2012 –.4 2007–.3–.2–.10.1.22008 2009 2010 2011 2012 –1.2 2007–0.8–0.40.00.40.82008 2009 2010 2011 20122007
–4–0481216(^2007) 2008 2009 2010 2011 2012
–10
–5
0
5
10
15
2007 2008 2009 2010 2011 2012
–1.0
–0.5
0.0
0.5
1.0
15
2008 2009 2010 2011 2012
(a) Consumer price inflation (b) Unemployment rate (c) Wage growth rate
(d) Output growth rate (e) Import price inflation (f) Depreciation rate
Changes in percentage points
Figure 17.7 Dynamic multipliers from a permanent reduction in exogenous (foreign) prices
by 5%, 2007(1)–2012(4)
The distance between the two dotted lines represent the 95% confidence intervals. The units on the vertical
axes are percentage points in all panels.
multipliers. The rigidities are not due to “incredibly long” adjustment lags in price-
and wage-setting: domestic inflation is seen to return to its baseline path after two
and a half years. Instead, there is a second wave of effects due to the interaction
between product and labor markets. Hence the experiment demonstrates rather
well the important theoretical point made in section 17.2.6.4, namely that the
steady-state inflation rate does not imply a unique equilibrium rate of unemploy-
ment, since the rate of inflation reaches its steady state long before the multipliers
of the rate of unemployment have died out.
17.4.3 Aspects of optimal policy: the impact of model specification on
optimal monetary policy
As noted above, the version of NAM with an econometrically modeled interest
rate makes no claims of representing optimal interest rate-setting under inflation
targeting or optimal policy response to a shock. Instead, the multipliers of the
last paragraph should be interpreted as counterfactuals: they show the response
that would occur if the interest rate reactedas ifit followed the econometrically
estimated interest rate equation.
However, in a recent paper, Akram and Nymoen (2008) show how optimal mone-
tary policy can be implemented in NAM, and how the predicted economic outcome
depends on the specification of the supply side of the model. For that purpose, they
replace the econometrically modeled interest rate equation with a theoretically
derived interest rate rule due to Akram (2007):
Rt+m=R 0 +( 1 −H)
βε
( 1 −φ)
εt+H(Rt+m− 1 −R 0 ),m=0, 1, 2,...,H. (17.70)