Essays on Economic Growth and Business Cycle Dynamics
This dissertation consists of five articles. The first article describes a simulated monetary demand-led macro model with different types of interacting agents. As such, it is assigned to the field of agent-based computational economics (ACE), where agents become virtual objects in a computer simulation. The model features long-run growth and generates endogenous business cycles. The second article covers a welfare analysis of macroeconomic stabilization analysis under the ACE paradigm. The third article describes a core-periphery network identification procedure which was applied to data from the monthly ZEW business climate survey. A discrete model and a continuous model were both used to detect the assumed network structure. The fourth article deals with cross-country growth regressions. A large panel data set was analyzed with a functional coefficient model (FCM), that is, regression parameters are modeled heterogeneously in a semiparametric approach. In a one-dimensional FCM initial per capita GDP serves as a status variable for local estimates, in a two-dimensional setting the time dimension provides a second status variable for local estimation. Finally, the fifth article concerns patent analysis. The patent portfolio of publicly financed research organizations (PROs) in Germany is analyzed and compared to US-PRO patents and firm patens. Patent citation data is then used to investigate patterns of knowledge diffusion originating from German PROs.