CAPACITY AND ASYMMETRIES IN MONETARY POLICY
21
4.1. Parameter Values. The model is calibrated to match the long-run properties
of the post-war US time-series with the non-stochastic steady state values of the
model. The parameterization follows, in some extent, Christiano, Eichenbaum and
Evans (1998) and Fagnart, Licandro and Portier (1999). The time period is one
quarter. The parameter for preferences and technology are assigned values that
are standard in the business cycle literature. Table 1 summarizes the values of the
calibrated parameters which are described in the sequel. The discount factor is
set at (β) = (1.03)-'25 ; the utility parameter is chosen so that one third of the
time endowment in the steady state corresponds to labor, hence, the consumption
expenditure share in the utility function (7) = 0.35; the relative risk aversion (σ) =
2. Model calibration requires that capital’s share on aggregate income (α) = 0.3485;
the annual depreciation rate of 10% implying a value (<5) = 0.018; the elasticity of
intermediate goods is chosen to obtain a markup ratio of 1.7 and thus, (e) = 8.7364;
the fixed cost that assures zero monopolistic profits is Φ = 0.1057. I deal, in what
follows, with the calibration of the aggregate uncertainty components. As stated
above, I will follow the common practice in the related research by assuming an
AR(I) process for the mean growth rate of money. In particular, the mean growth
rate of money (x) = 0.016, a value that corresponds to the mean quarterly growth
rate of the monetary base in the U.S. as obtained in Cook (1999) for the period
1970:1-1995:1. The persistence of the monetary shock (px) = 0.32 with the standard
deviation of (σ⅛) = 0.0038. _ _ _
The structural parameters that determine the aggregate capacity utilization rate
are two: the variance of the idiosyncratic shock and the degree of substitutability
among intermediate goods. These parameters are chosen in order to reproduce two
different situations, each featuring a different long-run capacity utilization rate. In
this manner, it will be possible to study how different the dynamic properties of
the model under these two different scenarios are. Specifically, the high capacity
economy is characterized by a low variability of the idiosyncratic shock, σ% = 0.25,
and a high value of input substitutability, e = 15. The opposite is true for the low
capacity economy, that is σ( = 1.75 and e = 4.85. The steady state properties of
the fully parameterized model under different scenarios are summarized in Table 2.
4.2. Dynamic Properties. Recall that the main objective of this paper is to
provide a formal theoretical background to the recently documented asymmetric
responses of key macroeconomic variables to unanticipated monetary policy shocks.
In this sense, it is studied whether the level of utilization of the productive capacity
of the economy alters the dynamic properties of the model. To achieve this target,
the equilibrium laws of motion of prices and quantities are approximated using the
undetermined coefficients method described in Christiano (1998). Specifically, the
model is linearized about the non-stochastic steady state and the impulse responses
computed next. The impulse response functions represent the response, over time,
of the elements of the endogenous variables to a pulse in one of the elements of
the vector of stochastic innovations. An important characteristic for a good model
to have is its ability to reproduce real world’s response to simple monetary policy
experiments. This section reports the dynamic responses of selected variables in
the model to a one percent increase in the gross rate of monetary growth in period
3, from the process
(4.1)
Xt = (1 — px) x + pxxt-ι + εxt with σx = 0.0032