Neighborhood Effects in Spatial Housing Value Models The Case of the Metropolitan Area of Paris (1999).

In hedonic housing models, the spatial dimension of housing values are traditionally processed by the impact of neighborhood variables and accessibility variables. In this paper we show that spatial effects might remain once neighborhood effects and accessibility have been controlled for. We notably stress on three sides of neighborhood effects: social capital, social status and social externalities and consider the accessibility to the primary economic center as describing the urban spatial trend. Using spatial econometrics specifications of the hedonic equation, we estimate whether spatial effects impact the housing values. Our empirical case concerns the Metropolitan Area (MA) of Paris in France which is divided in 2 636 neighborhood areas. We estimate the housing price distribution from a sample of 21,000 apartments sold in 1999. Our empirical results highlight the lumpy distribution of unit price along the general decreasing spatial trend from the Central Business District once neighborhood effects have been introduced. More precisely, a spatial error model is estimated revealing a positive and significance spatial effects across housing values which extend beyond their neighborhood area. Social capital, social status and social externalities play local role and may positively or negatively impact the housing prices. We showed a positive impact of diversified building patterns but a negative impact of social mixity which is somewhat conflictual but which is in fact in line with many current questions about social segregation and spatial segregation in urban areas.