Quantitative Study On The Relative Importance Of Different Sources Of Agglomeration Economies

Industrial agglomeration has not only been considered to be the most important source of strengthening the output efficiency and developing the industrial core competitiveness,but also an important driving force to promote the development of regional economy?The exploration on the microscopic mechanism of agglomeration economies has become an important field of agglomeration economies research. In order to have a quantitative study on the relative importance of different sources of agglomeration economies, we use the firm data from "China's industrial enterprise database" and calculate total factor productivity(TFP) of all the enterprises by the semi-parametric regression method(LP method). Based on the externality theory of Marshall,we quantitatively measured the three different sources of agglomeration economy on four-digit industrial level, namely labor market pooling, input sharing and knowledge spillovers. Finally, we use the dynamic panel system GMM regression method to research the influence of the three cluster sources on TFP in three major administrative areas-the provincial, city and county level. We gave some policy suggestions according to the results of regression, which provide a theoretical basis for the local governments to make suitable industrial policies,the main research results are as follows:First of all, we find that the input sharing and the knowledge spillovers have a positive effect on the city and county level, while the labor market pooling only play a positive role on the city level. Generally speaking, the city level is the most suitable geographical level for the effect of agglomeration economies, and it is also the best level for the local government to make industrial policies.Secondly, we find that the three kinds of the relative importance of different sources of agglomeration economies will differ because of the discrepancy of industrial technology level.knowledge spillovers present positive correlation with TFP in the medium technology industries and high technology industries. Labor market pooling and input sharing generate mainly positive significant role in the low technology industries.Finally, the regression results of different regional market systems shows: the influence of knowledge spillovers in the region with high degree of marketization is significantly positive. Labor market pooling effect plays a positive role in the region with a middle degree of marketization. Input sharing effect has played a positive role in promoting TFP in both the region with high and low degree of marketization. At the same time, the robustness test shows that the result of the paper is credible.