UBC Theses and Dissertations
A method by which to forecast house prices at the community level : a case study White, Douglas Arnet
The premise on which this thesis is developed is that it would be desirable for planners in cities such as Vancouver to be able to forecast house prices at the community level. Given advance warning of house price changes, it may be possible to devise policies which would influence trends and be in the best interests of the community and the city at large. The feasibility of making such estimates is the specific focus of this research. The methodology employed involved the selection of nineteen standard economic indicators as provided on a regular basis by Statistics Canada. These indicators were employed as independent variables in numerous multiple regressions wherein the dependent variables used represented the averages of annual assessed values for a sample of single family dwellings within each of the communities of Shaughnessy, Oakridge and East Hastings Sunrise. Observations for these variables were recorded over the period 1962-1980. The hypotheses tested for each of the three communities asserted that all or a subset of the nineteen independent variables would prove to be significantly related to house prices and would yield usefully accurate forecasts of such. Utilizing same year data for the independent and the dependent variables highly significant relationships were apparent and the hypotheses were supported. The hypotheses were also supported when the independent variables were lagged one year to simulate a forecasting situation. The strengths of this methodology include the ease with which the values of the independent variables may be obtained on a relatively current basis and the speed and relative simplicity of computer processing the data. On the other hand, the accuracy of the forecasts during periods of unusual economic change is open to question and criticism. It is apparent that for a model of this kind to perform effectively it is important that the underlying relationships between the forecasting variables and house prices be monitored perhaps on a monthly or quarterly basis.