Use of On-Line Data to Provide Rental Housing Market Mass Appraisals for England

International Geographical Union, Quebec, 28 to 31 August 2018.

This paper reports a mass appraisal exercise for the rental housing market in England using data on individual properties and their environment. Such mass market appraisals are common in the sales market, used primarily for the levying of local property taxes. Within the private rental sector there is less direct pressure for such mass market appraisals, although local property taxes are still usually levied on such properties so there is a need to ensure that such costs are covered through the rental charge. Instead there is the need to place a rental value on a property that reflects current market conditions. A rent too high and the property will remain on the market and not generate any income to the owner and a too low rent will provide a deflated income to the owner. Data available for appraising the rental market are limited from conventional sources. The data used here are derived from a property listings web site. Techniques used are regression, machine learning and a pseudo practitioner based approach. From the regression analysis attributes that increase the rental listing price are: property type; the number of various types of rooms in the property, proximity to central London and proximity to railway stations, being located in more affluent neighbourhoods and being close to local amenities and better performing schools. There is also evidence of some seasonality, and the more popular a property was on the web site (measured through number of views), the higher the rental price. Of the machine learning algorithms used to predict rental price the two tree-based approaches were seen to outperform the regression based approaches. In terms of a simple measure of the median appraisal error, the practitioner based approach is seen to outperform the modelling approaches.