Spatial Econometric Analysis Of Economic Growth Convergence

Convergence of economic growth is the key prediction of the neoclassical growth theory. It explores the problem about whether the economic development level of the backward areas will eventually catch up with the developed areas. It is very meaningful for China which has such obvious regional economic development gap to study the convergence of economic growth. Therefore, from the beginning of the 1990s, this theme has gradually become one of the research hotspots of China’s economic theory. However, study of economic growth convergence in most of the domestic literature usually over-estimated the convergence rate and had empirical estimation error. There are two reasons:the first one is the choice of control variables may have problems; the second one is estimation model did not consider spatial correlation. Therefore, we based on the research results of previous studies on the convergence mechanism and spatial econometric methods, to improve the results of the convergence rate through a way of optimizing variables and considering spatial correlation method.The idea of this paper is as follows:firstly, extending the MRW model to get a regression equation for testing growth convergence, then establishing the general panel data model and spatial panel model by using 1992-2012 provincial segmentation data(T=2), test the convergence and calculate the convergence rate. The ordinary panel model after optimize the selection of control variables are established and the convergence rate is obtained, and the convergence speed results could be compared with the previous documents results based on the traditional method, it shows the importance of careful selection of control variables in the convergence test; the convergence speed obtained by the new established spatial panel model could directly be compared with the convergence speed results of the former ordinary panel model with same control variables, it shows the importance of considering spatial correlation in economic growth’s convergence study and the impact on the estimated results and convergence speed. The main conclusions are:the estimation results of ordinary panel model are good, and the convergence rate is only 8.438%, it has decreased comparing with the previous similar documents, but still higher than the experience of foreign literature. Then, the R2 of a spatial panel model with the same variables and data improved significantly after considering spatial correlation, and the variable which was not fit the expected symbol in the previous model now fits the expectation; and the convergence speed is reduced to 5.65% furthermore, not only close to the empirical value of foreign literature, but also more consistent with the reality that provincial gap in China will be narrow in a long process. Therefore, we believe that it is necessary to consider spatial correlation and use spatial pane! data model in the study of economic growth’s convergence mechanism. Spatial econometric method could correct the errors of traditional method for research on regional economic growth problem to some extent, and make the results more reliable; moreover, considering spatial factors can make the convergence results more reasonable.