From Latin Quarter to Montmartre: Investigating Parisian Real Estate Prices

Abstract

This paper estimates buyers' preferences for dwelling attributes and neighbourhood characteristics. The collected data allows for the simultaneous consideration of a wide range of intrinsic characteristics, such as surface, floor, etc., and neighbourhood characteristics, including noise, crime, school quality, distance to jobs, etc. The marginal willingness to pay is identified from transaction data under the assumptions of the hedonic model described by Rosen (1974). We use very local fixed effects combined when possible with administrative boundaries as geographical discontinuities to isolate the effect of each amenity. Estimation is achieved by using flexible semi-parametric methods. Characteristics explain more than 90% of the variance of dwelling prices, showing a significant positive marginal willingness to pay for job accessibility and school quality, and a negative marginal willingness to pay for a higher crime rate in the area. By contrast, noise level or public transport accessibility have less influence on housing prices. These results are robust to the inclusion of census tract fixed-effects, which also drastically reduces the spatial correlation of the residual prices.