Supply and Demand in Residential Property Forecasts

This paper defines a measure of net housing demand or supply which allows the calculation of a long, high frequency time series, whose principal use is intended to be in a multivariate residential property price forecasting model. The series is easily and inexpensively replicable. We formulate this measure as excess demand. A negative excess demand figure implies a net oversupply. Supply is defined as the available dwelling stock. Demand is calculated from the long term trend in adult dwelling occupancy rates and is thus structural, rather than implicit. We examine the usefulness of our excess demand series in a forecasting model of property index returns over horizons from 6 months to 5 years. It is found that the explanatory power of current excess demand in estimating future property returns increases with forecast horizon.