A0667
Title: A spatiotemporal autoregressive factor model of the global business cycle
Authors: Tomohiro Ando - Melbourne Business School (Australia)
Matthew Greenwood-Nimmo - University of Melbourne (Australia)
Yongcheol Shin - University of York (United Kingdom)
Chaowen Zheng - University of Southampton (United Kingdom)
Matthew Greenwood-Nimmo - University of Melbourne (Australia) [presenting]
Abstract: To study the synchronicity of business cycles across countries, a new heterogeneous parameter panel data model is developed, in which the global business cycle is characterized as a spatiotemporal autoregressive process with a common factor error structure. A modified quasi-maximum likelihood approach is developed to estimate the model parameters in the presence of parameter heterogeneity and endogeneity. It is proven that the estimators are consistent and asymptotically normally distributed, and Monte Carlo simulations are used to show that their finite-sample performance is satisfactory. A framework is then developed for network analysis based on forecast error variance decomposition and diffusion multipliers. These tools are applied to analyze business cycle synchronization among 79 countries over the 50-year period 1970-2019.
