Hedonic price models in a geographic information system: Economic impacts of an urban greenbelt, Seoul, Korea

This study investigates the contribution of Seoul's greenbelt farmland to neighboring property prices and the econometric consequence of omitting the greenbelt characteristic variable(s) in the hedonic price function with the application of a Geographic Information System and spatial econometric methods.; The contribution of Seoul's greenbelt farmland to neighboring property prices was investigated to provide valuable information to Seoul's debate about the use of greenbelt farmland for urban development. When a portion of a greenbelt is developed, the external benefits provided by the greenbelt such as air cleansing, flood control, balanced microclimate, recreation, and scenery will be lost. Therefore, it is important to estimate the monetary value of the benefits so that the economic value of the loss can be assessed. Four different functional forms of hedonic price functions were applied to measure the economic value of Seoul's greenbelt farmland. The results from the estimation of these forms show that the typical household's marginal willingness to pay for a 0.1% decrease in the area of greenbelt farmland within a 1.5km buffer zone is between 11,503.45 and 32,352.5 Korean won ({dollar}10–{dollar}28). This result shows that the greenbelt farmland in Seoul has a negative economic impact due to its adverse externalities.; The greenbelt characteristic variable(s) represents the composition of unique amenity generating land-use types in a greenbelt of interest. The consequence of omitting the greenbelt characteristic variable(s) was investigated because the literature review showed that most of the hedonic studies done for the area abutting a greenbelt that consists of diverse land-use types have not considered the greenbelt characteristics and have not noticed the econometric problem of omitting the characteristic variable(s). For this particular data and study area, it was found that there exists substantial bias in the size of the estimated hedonic coefficients if we omit the necessary greenbelt characteristic variable(s). This result shows us the importance of incorporating a greenbelt characteristic variable(s) in hedonic modeling.