Spatial Autocorrelation and Real Estate Studies: A Literature Review

Abstract

Spatial autocorrelation is a phenomenon where the values of variable located within certain geographic area show a similar pattern. It is a source of imperfection in house price modelling that employs the popular technique of hedonic regression. Despite its long established concept,it is only recently when spatial autocorrelation has started to gain the attention of real estate studies. However, the evidence has come mainly from the USA. This paper reviews the literature on spatial autocorrelation and real estate studies. It describes some basic aspects of spatial autocorrelation in respect of hedonic price modelling (HPM) for housing markets. The importance of considering spatial autocorrelation and ways of dealing with the phenomenon are outlined. The paper also discusses two main approaches of modelling spatial autocorrelation, namely the spatial weight matrix and the geo-statistical approaches. It stresses the preference of the former in previous real estate studies that involve economic analysis. The paper concludes by highlighting the importance of considering spatial autocorrelation when cross sectional data are used. Evidence from countries including Malaysia would enrich the literature of spatial autocorrelation consideration in real estate studies.