Academic journal article The Journal of Real Estate Research

Housing Price Gradients in a Region with One Dominating Center

Academic journal article The Journal of Real Estate Research

Housing Price Gradients in a Region with One Dominating Center

Article excerpt


This paper primarily focuses on predicting housing price gradients in a Norwegian region with one dominating center. Spatial separation is represented by a function of the traveling distance from the city center in a traditional hedonic regression equation. Several functions are tested, and some alternatives provide a satisfying goodness-of-fit, consistent coefficient estimates, and intuitively reasonable predictions of housing price gradients. Still, not all commonly used functions are recommended. The findings also indicate that the strength of spatial autocorrelation is reduced when the hedonic function is properly specified.

The main ambition of this study is to estimate a housing price gradient for a region in the southern part of Western Norway. This region has a dominating city (Stavanger), and the study also tests for the appropriateness of the monocentric city model in this kind of area. The basic idea underlying this model is represented by a steadily declining unit price for houses with an increasing distance from the central business district (CBD). For a presentation of the modeling framework, comparative static results, and interesting extensions, see Anas, Arnott, and Small (1998).

Many empirical studies have aimed at finding rent gradients, land value gradients, and/or housing price gradients. In a few studies, the variable indicating the access to work came out with an insignificant sign, and occasionally a counter-intuitive sign was reported [see for instance Bartik and Smith (1987) for a review]. Such results can, for example, be explained by the fact that the area under study in some cases involves a restricted urban area rather than a housing market area. Another reason for such results is that modern metropolitan areas tend to be multicentric. Both Richardson (1988) and Heikkila, Gordon, Kim, Peiser, and Richardson (1989) state that the main reason for insignificant or counter-intuitive results stems from a misspecified hedonic price function. This is demonstrated in Waddell, Berry, and Hock (1993), who find that the impact of distance to the CBD is significant even when access to multiple employment centers and other nodes are accounted for. Adair, McGreal, Smyth, Cooper, and Ryley (2000), on the other hand, claim that transport accessibility has limited explanatory power in modern segregated and segmented cities, and they recommend that studies focusing on the effect of spatial separation on housing prices are performed in homogenous markets. According to McMillen (2004), however, the basic insights of the model also apply to complex polycentric cities, and he claims that the decline in the explanatory power of the model is a misunderstanding of the empirical evidence.

Most empirical studies on spatial variation in house prices consider complex metropolitan areas around large cities. Stavanger is the fourth largest city in Norway, but the population in the Stavanger municipality is only approximately 115,000. Still, the city is very dominating in the region, the structure is relatively monocentric, and the area is appropriate for studying housing price gradients. Alternative functional specifications of the relationship between the housing price and the distance from the CBD are considered here. Since housing price predictions are rather sensitive to the choice of functional form, it is important to find the form that best fits the data in the relevant type of area.

Housing price gradients represent an important input to studies covering a wide range of regional and urban policy issues. One example is that investments in transportation infrastructure might cause reductions in traveling times to the CBD and capitalization through property values. The findings, for instance, in studies related to constructing new roads (tunnels/bridges), changing speed limits or in analyzing investments inducing increased capacity and reduced queues on existing links. In some cases, the only available information on spatial characteristics might be the distance from the CBD. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.