ENDOGENOUS IDENTIFICATION OF MULTIPLE HOUSING PRICE CENTERS IN METROPOLITAN-AREAS

The monocenter model of urban metropolitan areas has come under increa sing criticism in recent years. Many urban areas are in fact multicent ered. This paper describes a methodology for identifying and ranking t he multiple ''centers'' of a metropolitan area by inferring them from real estate price data. The method generalizes from previous methodolo gies developed to identify the urban center from density distributions and gravity models. It involves applying maximum likelihood methods t o infer the location of each center, testing the relative importance o f each, and testing for the significance of the addition of each incre mental center. It allows the joint identification of the locations of metropolitan centers as well as their (local) price gradients (or elas ticity of price with distance from the centers). It is readily applica ble to GIS tools and data. In addition, there are a number of interest ing issues that may be addressed by assuming that the nuclei of the ci ty are ''unknown'' and must be inferred from the data. It is possible, theoretically, for the pinnacles of commercial and residential land p rice functions to be located in different places. It is possible to de scribe land pricing functions through identifying a series of ''center s'' in metropolitan areas that have irregular topography, or where pri ces follow the contours of transportation arteries. It is possible to identify ''inverse centers'' that act as troughs-rather than pinnacles -of land gradient functions. It is possible for the center or centers to be located in areas outside the traditional central business distri ct, to move over time, and even for pinnacles to lie outside the urban area itself. The methodology is then applied to real estate transacti on prices for Haifa Israel. (C) 1998 Academic Press.