economy, their lumpy investments generate stochastic synchronization which results in
a considerable aggregate fluctuation.
After establishing the analytical results on the propagation distribution in a partial
equilibrium setup, we simulate a general equilibrium calibrated by a finely disaggre-
gated sectoral data to examine under what conditions the model generates the right
magnitude of fluctuations. As argued by Thomas (2002), the general equilibrium effects
via wage and interest rate dampens the fluctuation effects due to the (S,s) behavior.
We first identify this dampening effect by using approximated price functionals, and
then simulate the exact general equilibrium path. Simulations show that the right mag-
nitude of fluctuations is obtained when the intertemporal substitution of consumption
as well as leisure is large. We also show that the autocorrelation and correlation struc-
ture of the production and demand components matches the empirical business cycle
patterns.
The rest of the paper is organized as follows. The next section presents the model of
investment propagation and the analytical propositions in a partial equilibrium setup.
Section 3 numerically examines the quantitative properties of the propagation and the
business cycle fluctuations by explicitly incorporating the consumer’s behavior. Section
4 concludes the paper.
2 Model of Investment Propagation
In this section we focus on the inter-industrial equilibrium relations in the product
market provided with the other prices, namely wage and interest rate. The product
market consists of N monopolists and a representative household. Each monopolist j
produces a differentiated good Yj , using capital Kj and labor hj .
Let us specify the production technology by a constant returns to scale Cobb-
Douglas function: Yj,t = Kjα,t(Athj,t)1-α, where At is a labor-augmenting technology
parameter which grows at rate g . We consider a balanced growth path where Yj,t ,
Kj,t , and consumption Ct grows at rate g and ht stays constant. Let us normalize the
variables by a growth factor At as yj,t ≡ Yj,t/At , kt ≡ Kj,t/At , ct ≡ Ct/At , ij,t ≡ Ij,t/At ,
etc. Then the production function is written in the normalized terms as:
yj,t = kjα,thj1,-t α. (1)
The capital is accumulated over time as:
gkj,t+1 = (1 - δj)kj,t + ij,t (2)