This original study examines the potential of a spatiotemporal autoregressive (STAR) approach in modelling transaction prices for the housing market in Paris and its inner suburbs. We use a data set from the Paris Region notary office (ìChambre des notaires díŒle-de-Franceî) which consists of more than 1,000,000 transactions units between the first quarter of 1990 and the end of 2005. We use the exact X -- Y coordinates and transaction date to spatially and temporally sort each transaction. We first choose to use the spatiotemporal autoregressive (STAR) approach proposed by Pace, Barry, Clapp and Rodriguez (1998). We do not find a global significant improvement from the STAR method for the modelling of the Paris Region housing market compared to a standard hedonic estimate. Nevertheless, we find evidence of a strong presence of both spatial and temporal heterogeneity in the model. Hence, we decide to develop a spatial and temporal autoregressive local estimation method. With this approach introduced by Pace and Lesage (1999) in a spatially autoregressive setup, we do no longer need to exogenously specify geographical submarket, nor to impose specified parameter variation function to take spatial heterogeneity in hedonic coefficients into account. It appears that spatial autoregressive effects seem to be much more pronounced in the historical centre of Paris than in its surrounding area. Moreover, these effects which were sizeable and significant for some geographical areas in 1991 have been deeply reduced between 1997 and 2005.