Spatial hedonic regression modelling of commercial and office space prices in Singapore

Abstract

In this thesis a hedonic regression model for the office, retail and industrial space prices in Singapore is developed and estimated. The data used for the regression analysis is collected from a commercial and a governmental online platform. Therefore asking and transaction prices are available for the office, retail and industrial market. This study uses a hedonic modelling approach for estimating the price determinants. In addition to the linear ordinary least squares approach, spatial models are used. Different spatial modelling approaches are tested and evaluated to find the best performing model. The influence of the public transport access and the location on the property’s price is investigated in detail. The major price determinants are the property’s tenure conditions, the access by public transport and the floor area. The office and industrial market show some clustering behaviour. The spatial autocorrelations in the data can be removed completely by using spatial hedonic approaches. Spatial error models (SARerr ) performed best and are recommended to use in further studies. The results can be used at the Future Cities Laboratory to model location choice of firms in Singapore by using an agent based transport simulation tool (MATSim).