Gunnar Bårdsen and Ragnar Nymoen 891This rule defines an interest rate path corresponding to a specific policy horizonH.
The response coefficientβε,H≡( 1 −H)( 1 β−εφ)determines by how much the interest
rate must deviate initially from the neutral rateR 0 to counteract the inflationary
effects of a shockεt. A high value of the smoothing parameterHcan be asso-
ciated with a strategy of gradualism in interest rate-setting. Thus two preference
parametersβε,Hand*Hdepend on the policy horizon,H. The last parameter in
the theoretical interest rate equation isφ, which represents the (objective) degree
of persistence in the inflation shock.
In addition to the preferences about policy horizon and gradualism captured
by (17.70), the optimal interest rate response is influenced by the user’s choice
of macroeconomic model. Akram and Nymoen (2008) focus on the labor market
channel, since this part of the transmission mechanism has been the focus of
most model controversy. In one model version, used to derive optimal policy, the
incumbent model of the supply side, with equilibrium correction in both wage-
and price-setting, is used. As explained in section 17.2.6, this model is called the
incomplete model of wage- and price-setting and is referred to as ICM in Figure
17.8. The two other versions are the model with wage and price Phillips curves,
PCM in the figure, and a version with a vertical Phillips curve, PCMr in the figure.
As noted in section 17.2.6.4, among these three specifications, it is the PCMr that
represents the consensus view in modern monetary economics.
Akram and Nymoen show that econometric encompassing tests favor the ICM
model of the supply side, but also that the PCM and the PCMr appear to be well
specified on their own terms, e.g., the residual properties of the two Phillips curve
models do not signal any problems. Hence there is a question about whether the
outcome of the encompassing test has any practical (or “economic”) significance,
or whether this test of model adequacy has only an academic interest. The analysis
suggests that the econometric test result contains valuable information.
Consider, e.g., Figure 17.8, which presents the economic performance of (opti-
mal and sub-optimal) policies employed in response to the supply shock. The left
column of the figure shows that there is a trade-off between price and output sta-
bilization for different ranges of policy horizons. Specifically, in the case of ICM
and PCM there is a trade-off in the range of 0–8 quarters. Policy horizons that are
longer than 8 quarters appear inefficient as both price and output stabilization can
be improved by shortening the policy horizon. The opposite is the case for PCMr. In
this case, the trade-off curve is associated with policy horizons that are longer than
6 quarters, while policy horizons shorter than 6 seem inefficient. In the right-hand
column of the figure, it transpires that the three models recommend substantially
different policy horizons. Even though the efficiency frontiers for ICM and PCM
are defined by almost the same policy horizon, the optimal horizon is 3 quarters
conditional on ICM, but 6 quarters in the case of PCM. In the case of PCMr the
policy horizon is 11 quarters.
Based on the above and several other simulation experiments, Akram and
Nymoen (2008) find that econometric differences bear heavily on (model-based)
policy recommendations and are thus not merely of academic interest. Their anal-
ysis quite strongly suggests that differences in model specifications and even in