Optimal Scale and Interregional Allocation of Public Investment in a model bwith Spatial Agglomeration and Endogenous Growth(in Japanese)

The purpose of this paper is to examine the optimal scale and interregional allocation of public investment in an economy with urbanization employing the frameworks of new economic geography models such as Krugman (1991). In our model, Dixit and Stiglitz (1977) types of differentiated goods and transportation costs in trading these goods are introduced, as is in the standard new economic geography models. In our model, however, the transportation costs are needed in trading all the other goods, as well as the differentiated goods. Moreover, the sizes of the transportation costs depend on the scale and interregional allocation of the public investments, which are financed by tax. Then, we can discuss the optimal scale and interregional allocation of the public investment and evaluate the public investment policies in Japan. The larger the household's preference for the differentiated goods is, and the more intensively the differentiated goods sector uses the differentiated goods themselves as intermediate inputs, the more likely it is that the households and differentiated goods firms agglomerate in one region and the urbanization occurs. In the case that the differentiated goods sector uses themselves as intermediate inputs intensively, shifting the weight of the allocation of the public investment from rural to urban area enhances the productivity of the differentiated goods sector concentrated in the urban area. Consequently, such a reallocation can improve the economic welfare of not only the residents in the urban area but also the ones in the rural area who import the differentiated goods from the urban area. Our numerical analyses show that the public investment should be larger in the urban area than in the rural area, with respect to per capita. Based on this result, we can conclude that in Japan it is preferable to increase the share of public investment to the urban areas.