By Dr. Maurice Yeates, Paul Du and Dr. Tony Hernandez

This report builds on research from the CSCA that reported on residential property prices in the City of Toronto (CSCA 2012-07). Extending our analysis by geography (to the broader Toronto region) and time period (2006 and 2011/12) we develop a new set of regression models – focusing  on a small set of  housing feature (HF), neighbourhood attribute (NA) and accessibility measure (AM) variables to explain residential property price variation. The resulting base hedonic price models for 2006 and 2011/12 indicate that even in one of the largest metropolises in North America more than three-quarters of the variation in prices is associated with just five independent variables representative of the three hedonic characteristics (HF, NA, AM). We then add a spatial autoregressive (SAR) variable to the base hedonic model, to take into account the impact on price of near-neighbours and the ‘explanatory’ power (R2) of the models increase to 83% for both time periods. Thus, despite being one of the largest metropolises in North America, where poly-centricity generally reigns, in the Toronto CMA mono-centricity prevails even though more than half of the CMA’s population now resides beyond the City. However, there are signs that the largest cities in the outer suburbs are beginning to develop ‘edge centres’, or downtowns, of their own, though they do not currently provide a cohering influence with respect to residential property values over the draw of downtown Toronto.

The authors would like to acknowledge and express thanks to: Yvonne von Jena, Director, Innovation and Marketing and Philip Bermingham, Manager, Data Analytics at Brookfield RPS for providing access to their custom data on property values in the Toronto CMA; and, Jan Kestle, President, Environics Analytics, for access to their Demographics and Projections Data. Any errors or omissions are the responsibility of the authors.