Contribution of Local Agglomeration Economies to Productive Efficiency: Stochastic Frontier Estimation with Establishment-level Data on Japanese Manufactures (Japanese)

There are two types of local industrial agglomeration of economic activity. One is the clustering of like-kind businesses—small and medium-sized companies, plants, and other types of establishments belonging to the same industrial group—in a specific geographic area. The other is the agglomeration of individuals, typically, employees of a large-scale company. The former type of agglomeration relates to localization economies, whereas the latter relates to scale economies arising from the size of operations. It is usually difficult to distinguish the effects of these two types of agglomeration when we use aggregated data, such as those aggregated at the regional level. Also, such regionally aggregated data do not allow us to capture differences in the extent to which each establishment benefits from agglomeration. Most of the previous studies on agglomeration economies and productivity have failed to properly discern the effects of agglomeration at the establishment level. This study focuses on the effects of local agglomeration on productivity, taking into account the spatial distribution of the size of establishments within a certain geographic area. The data set used in this study is comprised of four-digit establishment-level data on Japanese manufactures for 2005. The use of establishment-level data in estimating a frontier production function enables us to identify various agglomeration effects. In this study, we focus on traditional industries and high-tech industries, both of which tend to benefit from localization economies. The estimation model is based upon the stochastic frontier production function approach, in which we examine what type of local agglomeration contributes to the improvement of productive efficiency. The estimation results point to the effectiveness of industrial cluster policy as a means to generate agglomeration economies.