Automated valuation model: an application to the public housing resale market in Singapore

Purpose - This paper aims to construct an appropriate automated valuation model to value Housing and Development Board resale flats in Singapore. The paper also aims to test the accuracy of the model by comparing the values generated with actual valuations performed by a property firm in Singapore. In addition, it seeks to examine whether models for the sub-markets of Housing and Development Board resale flats based on location or type of flat are more "sufficiently accurate" than the general model. Design/methodology/approach - Using transacted data of 1,483 HDB resale flats, a hedonic price model is used to estimate housing price. The variables adopted include floor area of the housing unit, floor level of the housing unit, age, distance from central business district and distance from the mass rapid transit station. Findings - The study found that the general model provides sufficient accuracy when producing valuations. The models based on sub-markets, namely, "location" and "type of flats" produced reasonable levels of accuracy, although more variables could be added to the "type of flats" model to improve its reliability. Research limitations/implications - The research is limited to a few locations in Singapore. Future studies can include data from all over the island to provide better coverage. Practical implications - The automated valuation model could bring time and cost savings, which could result in higher profit margin for property firms. Thus, valuers could spend more time on complex valuations and issues. The model can also be modified to fit other property markets with appropriate characteristics (for example, high volume transactions). Originality/value - This paper represents an initial attempt to apply the automated valuation model in the valuation of Housing and Development Board resale flats.