This paper uses the spatial reorganisation of office rent patterns in Berlin as a test case for models of the location variable in rents. Since the late 1970s, a number of empirical studies have been undertaken using multiple regression analysis (MRA) to account for differences in office rents. Many of the variables in the MRA relate to the characteristics of buildings or leases but a critical variable is the location within a city. The paper outlines the strengths and weaknesses of the models used in previous studies to account for the value of location, such as the aerial and street distances from a central point and dummy location variables. A database of 412 office rents in Berlin for the period 1991 to 1997 are used as a test case for location measures. The problems of applying such models to a market like Berlin where dramatic spatial reorganisation has taken place are outlined. New computer-based methods of urban spatial modelling are presented and tested against the Berlin data. These approaches model the configuration of the street network rather than distances from a specific central point and are shown to be much more capable of capturing changes in rent patterns such as the shift in prime rental values to the East that has occurred in Berlin. The use of such models as tools for the prediction of in location trends is discussed.

This item is no longer available.