Abstract: From the perspective of total factor productivity, this paper uses the ‘Solo margin rule to select the input variables, and the data envelopment method(DEA)is used to calculate the provincial total factor productivity(TFP)from 1998 to 2017. Furthermore, using the fully modified least squares estimation(FMOLS)method, this paper analyzes the impacts of local government debt expansion on total factor productivity. The empirical results show that although local governments can generate short-term effects of easing financial pressure by innovating various financing platforms, long-term local government debt expansion will have a significant negative effect on total factor productivity. Then, the driving force for the expansion of government debt is more from the ‘weak’ or the uneven development of the central and western regions, or the ‘stronger’ of the better developed eastern countries? The samples are regionalized based on regional heterogeneity considerations. This paper constructs a dynamic panel model with lag period, and further uses the system generalized moment estimation(SYS-GMM)to empirically analyze the degree of “contribution” of local government debt to government debt expansion in different regions. The empirical results show that the government debt expansion in China is more sensitive to the government in the western region than in the eastern and central regions.

Key words: TFP, local government debt expansion, DEA, SYS-GMM