Modelling the influence of location on value

Abstract

IntroductionMultiple regression analysis (MRA) is a technique which seeks to link the value of a number of independent variables to a further variable whose value is supposedly dependent on them (i.e. the "dependent" variable). The intention is to produce a model, or equation, which will explain this relationship and hence enable prediction of the dependent variable in cases when this is unknown. Judgements about which model best achieves this aim are based on statistical tests. No attempt is made here to describe the technique fully, since its basis and the problems which may accompany it are well documented elsewhere (e.g.

Hedonic Analysis of Price in the Istanbul Housing Market

“…This can be achieved by three approaches, the fi rst one being the inclusion of the distance or neighbourhood quality variable into the model (So et al, 1997;Watkins, 1998). The second approach is to employ a neighbourhood dummy variable as a proxy for submarkets like in (Rothenberg et al, 1991;Gallimore et al, 1996;Ozus et al, 2007). Finally, the third method is based on estimating a separate equation for each submarket.…”

Combining Geographic Information Systems and Regression Models to Generate Locational Value Residual Surfaces in the Assessment of Residential Property Values

“…Location is an amalgam of several factors that includes a number of spatial elements such as accessibility to shopping, employment, educational and leisure facilities; exposure to adverse environmental effects such as traffic noise and hazard; neighbourhood amenity; perceived levels of neighbourhood security (Gallimore et al, 1996;McCluskey et al, 2000). From these, two key components of location can be isolated, i.e.…”

“…When an overarching model is adopted, such decisions may lead to disparities or inconsistencies where properties adjoin or close to neighbourhood boundaries. A hard edge may be implied at such boundaries, whereas in reality the varying influence of location may operate far more smoothly and the spatial trends occur as opposed to distinct areas of homogeneous property subsets (Gallimore et al, 1996;Mackmin, 1989).…”

“…The residual difference is made up of location as well as data inconsistencies, market variability and model inadequacies. Since location can intuitively and empirically be an influential factor on residential property values, it is possible to estimate locational influence for any point within the area by using the generated residuals (Gallimore et al, 1996).…”

“…Locational influence within an area can be established through an analysis of value residuals (errors) from a location-blind model (Anon, 2003). The residuals or the discrepancies between the actual and estimated property prices using this location-blind model can be regarded as the value of location (Jack, 1989;Gallimore et al, 1996;Figueroa, 1999;McCluskey et al 1999). According to Ward et al (2002), similar improved parcels will sell for different amounts if located in different parts of a town or within a same subdivision.…”