The characteristics of industrial agglomeration based on micro-geographic data

Our research goal is to test the spatial agglomeration according to industries and firms sizes at the city level, which is based on a unique plant-level data set of Hangzhou. In the paper, we employ a new method based on the distance and firm point data to explore industrial agglomeration in the city. The result from this method shows great differences with that from the method dealing with the surface data based on administration boundary such as the Moran's I and Getis-Ord Gi*. On the basis of firm's data provided by Hangzhou Trade and Industry Bureau, a complete process has been finished from making spatial data. We only use text firm address information to process spatial data, then to construct counterfactuals. Finally,the results are interpreted in this research. We select nine representative industries to reveal the discrepancy of agglomeration characteristics among industries. The finding shows that the spatial agglomeration of knowledge- intensive industries is significant, while most of enterprises from traditional labor and capital- intensive industries are approximately to the random distribution in urban areas. Specifically, the spatial agglomeration degree of finance,information service and high- tech and equipment manufacturing is obviously higher than the average degree of service and manufacturing industries; on the contrary, the agglomeration degree of consumer services and manufacturing industries, such as retail and food processing,fails to pass the counterfactuals. Although the degree of agglomeration of textile and apparel and heavy industry is higher than the counterfactuals in the range of 15~40 km, such a large distance means most of enterprises are dispersed in the suburbs. It is worth noting that most of business service are dispersed in industrial space rather than clustered at a small scale as the producer service should be. This unusual result probably means that business service is under development in Hangzhou. Besides, we analyze the further impact of establishment size on industrial agglomeration. Generally, the spatial agglomeration of manufacturing industries has been driven by the larger establishments, whereas service industries are mixed. While the spatial agglomeration of finance and business is also driven by small establishments, the agglomeration of large retails is more important than that of small ones. In the field of information service, it seems that industrial spatial agglomeration is driven by neither large nor small enterprises. Actually, the contribution from the agglomeration of a large number of medium- sized enterprises in the range between 0~3 km is dominant for service cluster. To manufacturing industries, it is clear that small enterprises dominate the spatial agglomeration,but the agglomeration of large ones is also important at a certain scale.