Abstract/Summary

House prices are hugely affected by location. This is because of the unique nature of real estate-its durability and immobility. Therefore, accounting for the effect oflocation is necessary in order to estimate a value of housing, and that is why including locational and neighbourhood related characteristics are crucial in a hedonic house price framework. The application of spatial econometric techniques represents an attempt to account for locational aspects of the properties -which cannot be accounted for in a standard hedonic model which is limited to observable locational explanatory variables -and thus improve the estimation of house prices in a hedonic framework. The three empirical papers in this thesis try to contribute to the application of spatial econometrics in housing markets in two aspects. The first contribution is to consider the unidirectional dimension of temporal effect in spatial relations between housing transactions. Unlike normal spatial data based on multidirectional interaction between unities, housing data are usually collected over time due to infrequent transactions. That means that the application of spatial techniques in housing studies needs to be conducted in a spatio-temporal context with consideration ofthe "arrow oftime". The empirical studies in this thesis employ a spatio-temporal weight matrix to deal with this issue; using a Hadarmard product, a spatial weight matrix based on multidirectional effect and a temporal weight matrix based on unidirectional effect are multiplied. The matrix, therefore, captures the spatial relations between housing transactions of different periods weighted for their spatial distance, but controls strictly for the temporal reality simultaneously. The second contribution is to apply the spatio-temporal models in various contexts. The existing literature has focused on confirming the existence of spatial dependence in housing markets and on comparing model performances between spatial and non-spatial models. The three empirical studies in this thesis try to not only provide evidence to support the use of spatial econometrics in housing studies, but also expand the application ofthese techniques to practically explain various phenomena in housing markets.