Convergence Of Regional Income Growth Based On Bayesian Spatial Durbin Model

The growth of income is the eternal theme of economic development,the rapid development of Chinese economy, making rapid increase revenues in all regions.However, with the rapid growth of the national economy, the gap of income across regions is gradually increased, affecting the coordination of the overall development of China's economy seriously. The gap of income not only become a hot issue of discussion, but also has become a focus of research in the economic field. As to study of regional economic convergence, it not merely to reflect a good understanding of the overall situation and development trend of China's regional economic development,but also we can get various factors through the analysis, which affecting the growth of the regional economy. Therefore, it can provide a reference for the development of China's economic development of science policy. China began to implement the deepened reform in 2014. If we want to develop a reasonable policy for long-term development of China's economy, and the coordinated development of a long-term strategy, we shoud have a deep understanding of the specific situation of China's economic development, therefore, the study of income growth is very important.To take account for spillover effects across regions, we investigate the revenue growth convergence problem, using the spatial Durbin model. Firstly, we established Bayesian spatial Durbin model, which address uncertain risk of parameters estimate in spatial Durbin model. Secondly, based on the analysis of model statistic structure and the selection of parameters prior, the Metropolis-Hasting within Gibbs sampling method is utilized to estimate model parameters. At last, by using parameter iteration trajectory, the diagnosis of Geweke is provided, and implyes the convergence of Markov chain which obtained by MCMC sampling methods, is steady and convergence.Moreover the parameters of the marginal posterior distribution of kernel density estimation curves are smooth.We apply Bayesian spatial Durbin model to analyse the data of Chinese provinces. The statistics results indicate that the iteration trace of parameters are convergent, and the Metropolis-Hasting within Gibbs sampling method estimate model parameters accurately, showing the effectiveness of Bayesian spatial Durbin model. In addtion China regional income does not exist ?-convergence, but exist conditional ?- convergence. The economic growth has negative effect on percapita income of the initial year, and show that the phenomenon of spatial dependence of the region's economic growth is exist. The development of a regional economic has a negative direct effect and a positive indirect effect to the adjacent regions. Moreover,the spillover effects among all regions income is significant.