| Abstract | This paper uses the latest macroeconomic data diagnostic theory and spatial panel data models, has an empirical analysis about the structural changes of China’s provincial GDP statistical data, which contain the data of 30 provinces from 1999 to 2010, and further takes the component decomposition to explore its formation mechanism. The main results are: (1) China’s GDP statistics has significant structural changes, and its information in every year has different impact on the whole information in the sample period. The Cook’s distance curves of GDP growth rate and per-capita GDP growth rate show crest-like non-monotonic trend, and structural changes of China’s GDP statistics are obvious and robust, so it is not fit to use econometrics models with constant parameter vector to analyze China’s GDP statistics, especially there is a greater difference between the significant change points and other sample points. (2) The provincial GDP statistics in China is not in isolation, and has significant spatial effect. For GDP growth rate, the spatial spillover effect among provinces is significantly positive with two kinds of geographic weighting matrix, which is GDP growth rate of one province will increase 0.26%-0.28% if that of the other provinces increase 1%. For per-capita GDP growth rate, the spatial spillover effect among provinces is significantly positive with geographic weighting matrix and economic weighting matrix , which is per-capita GDP growth rate of one province will increase 0.25%-0.38% if that of the other provinces increase 1%.(3) Structural changes of China’s GDP statistics has significant formation mechanism, that is the stochastic component is the main source; cycle component also plays an important part; the determinant component has minimal impact. The above results have very important policy implications for structural changes of China’s GDP statistics and its reliability and accuracy of judgment, and further practical reference for our government to smooth fluctuations in the economy and improve macroeconomic policy-making to achieve stable and rapid growth. |