Introduction of the project
This research relaxes the conventional assumption of time-invariant spatial autoregressive coefficient and extend spatial modeling to analyze dynamic of international system. We develop a dynamic multilevel Bayesian model based on TSCS data structural characteristics to model the overtime change of structural connectivity as well as accommodating the most considered methodological issues of serial correlation heterogeneity in TSCS data analysis. Because the proposed model is a temporal-spatio dynamic model, we carefully discuss the conditions of two-dimensional stationarity for various specifications of the spatial autoregressive coefficient. We then develop simulation algorithms to estimate the proposed model with the Markov Chain Monte Carlo. Finally, we apply the model and MCMC algorithm to analyze the change of connectivity of the world trade system based on the World Trade Organization over more than 40 years. The empirical illustration demonstrates that the spatial modeling with time-varying spatial coefficient would provide opportunities to the researcher to empirical learn more about the evolution of international system.
Achievements
