Following Ertur and Kock (2007) and Elhorst, Piras, and Arbia (2010), a spatially-extended neoclassical Solow growth model is developed to explore the spatial characteristics of regional economic growth at the prefecture level over the period 1992–2010 in China. A Spatial Durbin Model with the spatial fixed effects is verified in empirical analysis. The results show that spatial spillovers exist not only among the economic growth of neighbors, but also among the initial economic levels of neighbors. Furthermore, the same analysis is conducted for a new specified sample by controlling spatial heterogeneity, and the results show that spatial spillover effects are enhanced and the velocity of convergence is greater. The empirical analysis verifies that the extended neoclassic growth model is a feasible foundation to explore the spatial characteristics of regional economic growth over the period 1992–2010 in China.
Using a dynamic panel data model and the system GMM, this article examines the relationship between urban–rural income polarization and economic growth at the provincial level in the period 1995 to 2010 in China. The estimated results and significant tests indicate that a certain degree of urban–rural income polarization is beneficial to economic growth at the provincial level in both stages for China as a whole, though the contribution of urban–rural income polarization to economic growth is relatively small. Aggravating urban–rural income polarization has a negative impact on economic growth in China. Based on a cluster analysis of regional economic growth at the provincial level, the authors carried out the same analysis separately for two categories of regions too. The results for those two categories of regions show that positive correlations also exist between urban–rural income polarization and economic growth in both stages, which are very similar to the analysis for the whole of China. In addition, a meaningful finding can be derived that the contribution of consumption growth rate to economic growth rate in the second stage is smaller than that in the first stage obviously.