The composition effect is a statistical phenomenon that causes average home prices to be misleading.
Take CREA's average Canadian home price for example.
If there is a shift in the types of homes being sold (e.g., more high-priced single-family homes instead of low-priced condos), that is a change in the composition of home sales.
Such a change can cause average home values to be higher than normal.
This doesn’t mean all home values are rising; it simply reflects that more expensive homes are making up a larger share of sales.
That's why "average" home value changes can be misleading—especially over shorter periods. Over the long run, these distortions average out. Year-over-year comparisons are, therefore, much more reflective of a true average than month-over-month comparisons.
The opposite is also true, of course. For example, if steep interest rates cause a homebuying shift to cheaper condos instead of single-family houses, average prices will be skewed too low.
This is why MLN prefers median price measures whenever available. Unfortunately, CREA doesn't have median data on many regions, including Quebec and Nova Scotia.