Spatial Dependence in Apartment Offering Prices in Hamburg

This paper applies spatial econometric techniques to a hedonic apartment price model employing maximum-likelihood techniques. Accounting for spatial dependence of Apartment offering prices in Hamburg, Germany, the empirical analysis uses a semi-logarithmic price equation based on 4,029 offered apartments between 2008 and 2010. Starting with the traditional hedonic OLS-regression, we assess presence of spatial dependence using Lagrange Multiplier test statistics for error and lag dependence. These tests leads us to the spatial Durbin model and a spatial weight matrix based on the 15 nearest neighbors. Estimation results show that apartment prices exhibit a positive relationship with neighboring apartments. In addition to a high spatial autoregressive parameter, the estimated indirect effects (following the methodology of LeSage and Pace [2009]) show significant results. Consequently, a change in a single explanatory variable in a particular apartment not only affects the apartment price itself but also the price of neighboring apartments. Following the estimation results, spatial dependence is present, least- square estimates are biased and spatial hedonic models do explain more of the price variation with significant indirect effects in the spatial Durbin model.