아파트 단지특성이 주택매매거래율에 미치는 영향

Abstract

This paper aims at finding out structural and locational characteristics of apartment complex affect the count rate of housing transaction using Poisson, negative binomial, zero-inflated Poisson, and zero-inflated negative binomial regression model with an offset, the number of houses. Results of the analysis are as follows. First, zero-inflated negative binomial regression model has turned out the best-fitted model. Second, we find out that the number of highest floor number, the ratio of the house below 85m², the number of academy and district heating system have positive effects on the count rate of house transaction. Unlike hypothesis of housing prices, the age of apartment, the construction capability ranking of the constructor of apartment and proximity to the Han River have negative effects. Third, some of the locational characteristics that affect apartment prices do not affect count rate. The park distance, subway station distance, school distance variables do not affect count rate. On the contrary, count rates of the Northwest, Southwest, Southeast, and city center region are lower than the Northeast. This study differs from precedent studies in that it does not deal with regional transaction volume and macroeconomic variables but transaction volumes of apartment complexes and structural and locational characteristics. In terms of methodology, this study introduces the total number of houses as an offset to control the influence to the count rate and zero-inflated model to overcome the problem of zero-inflation. This trial contributes to the methodological development.