From its peak in February 2022 to last August, the seasonally adjusted average MLS® price for the entire country fell by 15.6%. This drop in housing prices was expected by forecasters — and by CMHC — given the anticipated return to more sustainable market conditions than those seen in 2020 and 2021. The negative impact of increasing interest rates required to bring inflation under control also played a role in declining prices.
Based on this decrease in the average price of homes, the tendency would be to believe that property values in Canada have weakened significantly.
It’s important, however, to understand that changes in the average price may also reflect a change in the composition of transactions across the different price ranges.
In fact, this composition effect would account for a little more than half of the decline in the average MLS® price since last February for Canada.
The actual weakening of housing prices would therefore be less significant than suggested by the average price drop. This conclusion may change over the coming months as new data becomes available and the full impact of interest rate increases and the anticipated economic slowdown is felt.
What is a composition effect in the housing market?
In addition to the supply and demand of a market, the characteristics of units being sold also affect the selling prices of homes. The size, location, appearance and other features of a property can affect its demand and price.
The most common home price measure is the average MLS® price, which is calculated from sales made through real estate brokers. It’s calculated by dividing the total value of transactions by their number.
A variation in the proportion of less expensive properties in total sales — due to the type or location of the properties sold — creates a composition effect. If this variation is significant, the change in the average price may primarily reflect a change in the characteristics of the properties sold, rather than an actual variation in property values.
For example, an increase in the share of sales of condominiums, which are generally of lower value than single-family homes, could cause the average home price to fall, even if the value of each property is unchanged.
Some price measures attempt to eliminate these composition effects. One of these measures is the MLS® Home Price Index (HPI). The index is based on a hybrid model that combines the recurring sales approach with the hedonic price method (which captures the impact of various quantitative and qualitative characteristics of homes on their price)1.
Price indexes that eliminate composition effects have decreased less rapidly than the average MLS® price since last February
In Canada, the seasonally adjusted average MLS® price peaked in February 2022 at $786,000. It was $663,000 in August 2022, which represents a 15.6% drop. For the same period, the MLS® Home Price Index decreased by only 7.4% (see figure 1).
The 8-percentage-point difference between the rate of decrease of the average MLS® price and that of the MLS® HPI could be explained by a shift in the composition of sales toward lower-value homes. In other words, more than half of the decrease in the average MLS® price since February would be explained by a composition effect, with the remainder explained by a real weakening of prices.
Figure 1: Average MLS® price and MLS® HPI, indexed to 100 in January 2019
Toronto, Vancouver and Montréal all have seasonally adjusted average MLS® prices that peaked in early 2022. For Montréal and Vancouver, the contribution of the composition effect to lowering their average price was more significant than for Toronto.
After reaching its peak in February 2022, the average price in Vancouver fell 6.5% through August. During this period, the MLS® HPI decreased by only 3.5%. A little over 50% of this decrease in the average price was attributable to a real weakening of prices and the remainder was attributable to a composition effect.
For Montréal, the average MLS® price peaked a little later, in April. Since then, it declined by 5.5%, while the MLS® HPI decreased by only 3.3%. Therefore, about 60% of this decrease in the average price in Montréal would be attributable to a real weakening of prices and 40% to the composition effect2.
For Toronto, the average price decreased by 12.0% from its peak in February to July. During this period, the MLS® HPI decreased by 8.6%, so the actual weakening of prices accounted for just over 70% of this decrease in the average price in Toronto, while the composition effect contributed less.
Where do the observed composition effects come from?
The first months of the COVID-19 pandemic led to a strong increase in housing demand, which resulted in a 12.5% increase in transactions in 2020. In several markets, demand increased especially for single-detached, semi-detached and row homes. This shift reflected an increased preference for larger and generally more expensive homes in response to the pandemic and related telework and homeschooling needs. This contributed to stronger growth in the average price.
In the Toronto and Vancouver metropolitan areas, the share of single-family homes in total transactions is now below pre-pandemic levels (see figure 2). With condominiums now making up a larger share of transactions, their lower values likely contributed to the decrease in the average MLS® price.
An increase in the share of single-family homes in total transactions was not seen in Montréal in 2020. Still, this proportion is below pre-pandemic levels, indicating a slowdown in demand for single-family homes compared to other housing types.
Figure 2: Share of single-family home sales in total transactions
As the impact of rising interest rates and the expected economic slowdown takes full effect on the housing market, the actual weakening of prices should become more pronounced. The gap between the average MLS® price and the MLS® HPI is also expected to tighten.
The composition effect should account for less and less of this variation in the average price, while the evolution of the average MLS® price will be more representative of real home prices.