位相市場 外部經濟의 資本化에 관한 硏究

Abstract

This study is to extract the various externalities in the urban housing price determination process. The model is constructed on the basis of modern spatial econometrics since the data for an urban hosing market is, on its nature, a strictly spatially referenced data set. The study utilized Largrange multiplier tests in order to check up two forms of spatial dependence in the traditional linear regression model built up on the basis of a hedonic price function . the results show us both forms of spatial dependence-one is in the spatially dependent variable and the other is in the error term-exist in the traditional linear regression model using the five different forms of the hypothsized spatial weight matrices. First, the dependence in a spatially dependent variable confirmed, the model is reconstructed into the spatillay autoregressive model taking account for the adjacency effect in addition to structural and neighborhood effects. Second, the dependence in the error term is taken care of by the polynomial approximation of a trend surface analysis. This second set of spatial dependence is conceptualized as a pure locational effect presented in the urban housing price determination process. Finally the maximum likelihood estimation is substituted for the bootstrap estimation since the model still suffers from the spatial heterogeneity by way of heteroskedasticity, which should be overcomed for a sound model repersentaion of an urban housing market. And hence, this research succeeds in extracting the three differnt forms of conceptualized externalities in addition to the strict structural efect. They are the adjacency, neighborhood, and locational effect. According to the calculation results of relative importance of these four effects for the different price levels of house, the locational effect is dramatically increasing with the prices of houses, which suggests us that the accessibility to the urban amenities becomes an important factor for higher priced houses.