Tomonari NOMURA
A Study on Interregional Trade in the Spatial Computable General Equilibrium Model
Supervisor: Shoji MATSUMOTO, Satoshi TSUCHIYA, Kazushi SANO
Computable general equilibrium model (CGE model) extended into multiple regions, or, spatial CGE model, inevitably involves a problem of describing trading goods/commodities between regions. This study aims at summarizing advantages and problems of some typical approaches of formulating interregional trades in SCGE model, and at discussing another formulation.
The most widely-used approach is to use the constant elasticity of substitution (CES) function with Armington¡Çs assumption. Although substitution elasticity on region of purchase is sometimes such a sensitive parameter that it effects big changes in the spatial computable general equilibrium analyses, few systematic study of estimating the elasticity has been done for domestic regions in Japan. Thus, in this paper, we firstly review a method of estimating interregional trade elasticity in Japan.
Another typical approach is to employ Harker model. A great advantage of this model is to make SCGE analysis possible without a perfect interregional input-output table. On the other hand, it has a problem of ¡Èhomogeneous of degree zero¡É in the whole SCGE equation system. Reproducibility with actual interregional trade data is not good. In this paper, we propose Multiplicative Competitive Interaction (MCI) model to solve these problems of Harker model.
Reproducing the current status of data, it is found that MCI model shows better performance than Harker formulation. In addition, MCI solves a problem of ¡Èhomogeneous of degree zero¡É. Consequently, we conclude the MCI formulation can be a promising method for SCGE model.
Please use "previous botton" to return the previous page