Canada's Erroneous Housing Statistics

by: Larry MacDonald

“RESALE HOME PRICES FALL 9.2 PER CENT IN FEBRUARY,” said the headline to press coverage on the latest data released by the Canadian Real Estate Association (CREA). What a huge disservice it is for CREA to disseminate such erroneous statistics.

CREA’s headline measure of house-price change is very misleading because it does not control for changes in the mix of houses over time. CREA’s indicator may have, for example, an average dwelling size of 2,500 sq. ft. in the base period and 3,000 sq. ft. in the observation period. It will also likely have changing proportions for other housing characteristics such as number of bedrooms, region, age, etc. This is not comparing apples to apples.

I’ve been harping on this issue for a few months now – see, for example, Dec. 9, Dec. 11, and Dec. 18 posts. A less biased method is the one now widely followed in the United States: the repeat sale price index (RSPI). According to the RSPI produced for Canada by Teranet Inc. and National Bank, resale house prices are only down about 1%.

Cynics may wonder why CREA continues to publish such dodgy data. Could it have something to do with the fact that real estate agents — like churning stockbrokers — want “to make deals and make them fast,” as Freakonomics authors Levitt and Dubner describe? That is, the perception of tumbling house prices likely makes house sellers more susceptible to admonishments from real-estate agents to lower their asking price on, or after, the listing –which reduces the agent’s listing time. If house prices are not tumbling, homeowners are more likely to keep the agent working on getting a better price.