Monthly figures are extremely volatile, particularly for housing starts.
Seasonally adjusted annualized data is the correct way to compare monthly data from one month to the next. If I want to compare data from April to May or from May to June, I have to look at the seasonally adjusted annualized data. However, once the data is annualized, even a small shift in the number of actual housing starts results in a big change in the total number. The data is therefore quite volatile from month to month, which essentially undermines the need for a certain degree of stability in the data.
At CMHC, we've been advancing trend measurement of housing starts for several years. This provides a much clearer picture of the overall state of the market. It's more informative to look at the trend in housing starts than to look at month-to-month data, which varies greatly.
Now, let's talk about year-over-year data. For example, when we want to compare the data for April 2026 with that for April 2025 to see where we stand relative to the same period the previous year, what actually happens is that there's a baseline level.
For example, in April of last year—and there's a reason I'm citing this example—housing starts were relatively low compared to the historical average. Obviously, it doesn't take a very high figure in April 2026 to show growth. However, compared to the historical average, what we need to look at is whether growth is—