Not all heritages generate a premium: heterogeneous effects of cultural heritage as infrastructure on housing prices

In heritage-led urban regeneration discussions, there is an assumption that culture brings in a housing premium, but the evidence remains limited because different heritage types, tourism levels, and local context are not examined together. In Nanjing’s main urban area, there are 34,076 cleaned second-hand housing transactions and heritage spatial data, with a community-level analytical sample of 2,653 observations. To find the generalized average relationship, OLS is used; MGWR is used to identify variation across multiple scales; and SHAP is used to detect nonlinear effects. Tangible heritage density is always connected with more expensive houses, but in MGWR there was no local significance for the association between proximity to intangible heritage and housing prices. By contrast, distance to A-level scenic attractions has a positive local coefficient, meaning that houses near A-level scenic attractions may have lower prices in some areas. The SHAP plots further show nonlinear threshold effects for intangible heritage density and scenic-spot accessibility. In short, there is no single universal effect of heritage on housing prices, and the effect depends on heritage type, location, and tourism intensity.