Contributions of local agglomeration to productivity: Stochastic frontier estimations from Japanese manufacturing firm data**This paper is a part of outcomes of the research project ‘Sustainable Regional Economic System’ conducted at RIETI (Research Institute of Economy, Trade & Industry, Japan) in 2010. An earlier version of this paper was presented at the 51st European Congress of the RSAI in Barcelona. The author thanks to the guest editor Andreas Stephan as well as anonymous referees for helpful comments and suggestions which substantially improved this paper

This study focuses on local agglomeration effects on productivity, in which the spatial size†distribution of firms is taken into consideration. The data set is comprised of four†digit firm level data on Japanese manufactures for 2005. The estimation by use of firm data enables us to identify various agglomeration effects. The estimation model is based upon the stochastic frontier (labour) production function approach, in which we examine what type of local agglomeration contributes to improve productive efficiency. The estimation results suggest that for most light industries agglomeration of various sizes of firms is important for productivity rather than the concentration of uniform small size firms.