Spatial econometrics and spatial data pooled over time : toward and adapted modeling approach

In this paper, we address the possible problem related to using strictly spatial modeling techniques for spatial data pooled over time. For these data, such as real estate, the spatial dimension is present, but subject to constraints related to temporal dimension. Three empirical examples are presented to investigate the impact of neglecting the temporal dimension in spatial analysis and to show how such an approach overestimates the pattem of spatial dependence, along with the estimated spatial autoregressive coefficient. If generalized to all other empirical applications, this conclusion may have important considerations if one tries to measure the effect of extrinsic amenities on house prices