Multidimensional Green Space Exposure and Housing Prices in High-Density Urban Areas: Spatial Evidence from 3729 Communities in Guangzhou

Abstract

Green space in high-density cities is scarce and shared unevenly, and the question of whether its value affects housing prices and whether that effect holds between communitie is still unclear, as most studies measure one exposure dimension or report a single area-wide estimate. In this study, we examine 3729 residential communities in Baiyun District, Guangzhou, linking a composite green space exposure (GSE) index to second-hand listing prices. The GSE index equally weights three normalized indicators—the green coverage ratio, park accessibility index, and green view index—which describe availability, accessibility, and visibility. We fitted ordinary least squares, geographically weighted regression, and multiscale geographically weighted regression in turn, judging spatial variation through residual spatial autocorrelation and corrected local inference. The three dimensions of the index peak in different parts of the district: the composite GSE runs higher along the eastern foothills and in parts of Jinshazhou and lower across the central and southern built-up areas. High-GSE and high-price areas overlap only partially: the greenest communities are not, in general, the most expensive ones. On average, a 0.01 increase in GSE is associated with an estimated 4.1% price difference at the district scale, whereas the typical local association is much smaller (about 0.4 to 0.6% for the same 0.01 difference), and the global model explains only 19.4% of price variation. Positive local coefficients cover 65.8% of communities, with 25.0% remaining significant after correction, while significantly negative ones cover 0.7%. Greener community planning should therefore account for exposure dimensions and local market responses, not district-wide averages.