Estimating coastal premiums for apartment prices: Towards a new multilevel modelling approach

Abstract

This article attempts to assess ‘coastal premiums’ for apartment prices in Corsica, France. The coastal premium should consist of two parts, those related to the view of the Mediterranean Sea, and those associated with living close to beaches for access purposes. Views, being a qualitative and subjective variable, are difficult to measure and quantify. Further, apartments are located in neighbourhoods and then are situated in districts or at more aggregated levels. This induces complex spatial interactions. To deal with these, we first employ a viewshed analysis implemented in geographical information systems (GIS) to generate an objective, continuous measurement. We then develop a Bayesian spatial/spatiotemporal multilevel model, which integrates Leroux’s conditional autoregressive process with multilevel modelling for conducting a hedonic analysis in the presence of multiple-scale housing data. Since the developed model pertains to latent Gaussian models, estimation is carried out by integrated nested Laplace approximations. The results demonstrate that homebuyers have a higher marginal willingness to pay for larger views of the sea, and proximity to beaches produces positive impacts on apartment sale prices. Further, strong spatial spillovers emerge among high-level units, but heterogeneity dominates among apartments. Our findings illustrate the importance of coasts to homebuyers. Further, the estimated value permits planners and policymakers to assess the trade-off between developing and preserving land along the coast.