Essays on Empirical Macroeconomics

Abstract

The nonlinear time series analysis has been applied widely in many empirical macroeconomic studies and deepened our understanding of the economy. One of the traditional issues of most importance is how to describe the monetary policy conducted by the central bank. A monetary policy reaction function is a hypothetical function connecting the central bank's monetary policy instrument with the central bank's objectives, e.g., the stability of the output and inflation. The first two chapters in this dissertation add to the literature of the nonlinear monetary policy reaction function through the lens of the Taylor rule using U.S. data. On the other hand, understanding the macroeconomic dynamics has long been the main motivation of using nonlinear time series analysis. The state space model is one of the most useful techniques which can be applied to a wide range of empirical problems. The third chapter in this dissertation is concerned with a class of the state space models whose disturbances follow stochastic volatility processes. For reasons which will become clear as this dissertation proceeds, I adopt Bayesian approaches, which are especially useful for comparing several competing empirical models and for estimating models that are nonlinear in latent variables, to deal with all issues under consideration.