ABSTRACT
This dissertation studies the following topics in the business cycle literature: the
persistence properties of business cycle models, the estimation of dynamic stochastic
general equilibrium models and the roles of structural shocks. These topics are studied in
order to take business cycle models to data and see their plausibility.
Firstly, I investigate the persistence properties of an RBC model with
consumption habit and capital adjustment cost. I use spectral analysis due to its advantage
of providing a good summary of the temporal behavior of economic variables. It turns out
that introducing capital adjustment cost lowers the heights of the model spectra at all
frequencies due to lower variances. Consumption habit moves the peak of consumption
spectrum towards too low frequency. The benchmark model whether calibrated or
estimated, could not generate any significant peaks at 4-5 year business cycle frequencies.
Secondly, a version of New Keynesian models is investigated that features Calvostyle
sticky price/wage with various rigidities and structural shocks similar to Christiano,
Eichenbaum and Evans (2005) and Smets and Wouters (2003). The model is estimated by
the MCMC algorithm of Bayesian methodology with a particular effort to lower the price
stickiness parameter estimate to be consistent with microeconomic evidence. For this
purpose, I introduce an assumption of the endogenous elasticity of demand to the model
of the above authors as an additional source of real rigidity, which successfully lowers

the duration of average price to about 2.2 quarters. The estimated model also mimics well
the spectral peaks shown in the U.S. aggregates. However, the model has a difficulty in
explaining inflation inertia. The dilemma is that it is hard to capture both the inertial
behavior and the volatility of inflation at the same time with the model. Finally, impulse
response and variance decomposition analyses show that the price markup and the labor
supply shocks turn out to be important in accounting for the variations in the real
variables. The key thing is that these shocks produce the right correlations among the
endogenous variables.