Abstract
Incorporating spatial econometric tools in hedonic pricing (HP) models for environmental valuation has become the standard approach in the literature. The effect of house prices on other house prices is taken into account and usually measured by distance or contiguity in spatial weight matrices. Disaggregate house sale datasets are composed from observations each at a specific location and time. Nevertheless, the symmetric spatial weight matrices commonly employed in HP studies ignore the temporal dimension in disaggregate house sale data. Thus, not only are previous house sales taken to affect subsequent house prices, but so do future house sales. However, information does not travel backwards in time; hence, there is a clear theoretical impossibility of actual future prices affecting current/past prices. Estimates derived from HP models where spatial dependence is incorrectly specified or ignored will exhibit inaccuracies. This article proposes an alternative specification of spatial weights in HP that includes spatial effects on each sale price only from preceding house sales. The temporal aspect of spatial effects is then developed further by specifying a time-decay rate to capture the diminishing effect over time of preceding sale prices to succeeding house prices. This novel specification of spatial weight matrices is shown to have a significant effect on the estimates of house price depreciation from aircraft noise. Monetary values of aircraft noise externality are successfully derived from the HP models for Athens Airport.