An Empirical Analysis of Panel data of Agglomeration Economies in the Japanese Manufacturing Industries

本文在Ellison和Glaeser（1997）的基础上，使用Tokunaga and Akune（2005）、Tokunaga，Kageyama，and Akune（2005）的工厂级4位数分类法、面板数据集和新的集聚指数，对1980 ~ 2000年日本制造业生产集聚经济的存在性进行了研究。这是对传统指标如区位商或区位基尼系数的改进。当我们应用Kim（1992）提出的灵活的超越对数生产函数和成本分担公式时，我们发现在就业为基础的集聚（γ_）和增值为基础的集聚（γ_ ^V）的情况下，在没有对同质性进行任何限制的情况下，大约存在1.5%的正集聚效应<EG><EG>。另一方面，我们使用城市2位数分类的面板数据、面板数据集、亨德森（2003）的集聚指数和Ellison and Glaeser（1997）的共集聚指数，对1985 ~ 2000年日本产业的集聚和共集聚效应的存在性进行了检验。利用基于中村和江岛（2004）的Cobb-Douglas生产函数，我们发现日本产业具有正的集聚和共集聚效应以及规模经济。

we examine the existence of agglomeration economies on production in the Japanese manufacturing industries from 1980 to 2000 using plant-level 4-digit sub classification, panel dataset and new agglomeration index in Tokunaga and Akune (2005), and Tokunaga, Kageyama, and Akune (2005), based on Ellison and Glaeser (1997). This is an improvement on the conventional indices such as Location Quotient or Location Gini Coefficient. When we apply a flexible translog production function and cost share equation as suggested by Kim (1992), we find that around 1.5% of positive agglomeration effect exists in absence of any restriction on homotheticity in the case of employment based agglomeration (γ_<EG>) and the case of value added based agglomeration (γ_<EG>^V). On the other hand, we examine the existence of agglomeration and co-agglomeration effect on production in the Japanese industry from 1985 to 2000 using panel data of city-level 2-digit sub classification, panel dataset, agglomeration index in Henderson (2003), and co-agglomeration index in Ellison and Glaeser (1997). Using the Cobb-Douglas production function based on Nakamura and Eshima (2004), we found that a positive agglomeration and co-agglomeration effect and scale economies in the Japanese industry.