CMHC’s Starts and Completions Survey provides Canadians with objective, accurate and timely information on actual new residential construction in Canada. Housing starts and completions data are collected monthly as part of this survey. It’s not a survey of construction intentions; it reports on homes for which construction has started and is already underway.
-
Home
-
The Housing Observer
- How CMHC tracks housing starts across Canada
Done!
Link copied
Share via Email
Reliable data on Canada’s residential construction
CMHC’s Starts and Completions Survey provides consistent data on new residential construction activity across Canada.
The survey is carried out through site visits to confirm that new residential units have reached set stages in the construction process. Building permits, issued by most municipalities, are used as an indication of where construction is likely to take place.
In areas where there are no permits, reliance is placed either on local sources or other field enumeration procedures. The survey is conducted in the same period month-to-month for consistency in data reporting. The type of information gathered is categorized into 3 distinctive groups, as follows:
- The date at which the structure reaches important milestones, such as start, under construction and completion.
- The geographic location of the structure (e.g. address, province and municipality).
- Other attributable information, which describes or characterizes the structure (i.e. price, dwelling type and number of units).
This survey has been conducted since the late 1940s, and the resulting information is used for purposes such as:
- determining the contribution of residential construction activity to overall economic output (gross domestic product)
- understanding additions to the housing stock, and
- assessing supply and demand conditions in the housing market
Definition of a “housing start”
A “housing start” is defined as the beginning of construction work on a building. It typically occurs when 100% of the concrete has been poured for the whole of the footing around the residential structure. If there is no basement, it is the equivalent stage in construction.
The rationale for using the concrete pouring over the entire footing around the structure is that it is the actual confirmation that the project has moved beyond the planning and excavation. It’s a firm indication that actual construction has begun.
Differences in market conditions and other location specific circumstances can affect construction timelines. Activity may stall after a site is cleared. It can also pause after a hole is dug to prepare the foundation.
This approach ensures that actual residential construction activity is measured the same way across the country. Consistent application of the methodology also enables accurate comparison of housing starts activity over time.
Why parkade construction isn’t a housing start
Work on constructing a parkade under a structure is not necessarily an indication that residential construction has begun. A housing start is not automatically recorded at time of parkade development. There may be multiple structures on a single parkade, with different towers starting at different times frames and serving different purposes. These towers could include a mix of commercial and residential.
Within the Starts and Completion survey, we are focused on the start of residential units. The commencement of work on a parkade doesn’t meet that definition.
For example, a shared parkade might support 1 office tower and 1 residential tower. If the office tower is first to rise, we would not record a housing start until construction of the residential element begins — even if the office tower is already complete.
Accurately analyzing trends in housing starts data
There are different ways of comparing housing starts over time. Each provides a distinct perspective in analyzing trends in data.
It is recommended some combination — if not all — methods be used to compare data over time to minimize the risk of drawing incorrect conclusions from a single viewpoint.
-
Year-over-Year comparisons of actual data consists of
comparing starts activity during a certain month of the year compared to the
same month a year earlier.
- Comparing activity over the same months helps to account for seasonality and helps to see if activity has increased or decreased in a given month.
- However, the size and timing of large-scale multi-family projects can cause major fluctuations in housing starts data. This includes high-rise condominium and rental structures. These fluctuations can impact the accuracy of trend analysis when using only 2 data points.
-
Year-to-date or moving average comparisons of actual data
compensates for short term or random fluctuations which may hide the true
underlying trends in the data.
- This approach is often referred to as smoothing the data and uses year-to-date or moving averages and then calculating growth rates between 2 years.
-
Seasonally Adjusted Annual Rate (SAAR) provides a more
accurate and reliable comparison of data across different time periods.
- The SAAR is calculated by seasonally adjusting the monthly housing start data to remove the effects of seasonal variations. It then multiplies the adjusted figure by 12 to produce an annual estimate to compare monthly housing start activity with annual levels.
- However, just as for year-over-year comparisons, SAAR numbers can fluctuate considerably from month-to-month due to whether there are large multi-family projects.
- Monthly analysis of SAAR or year-over-year data is not recommended to obtain a view of the trend of housing starts or their likely annual outcome.
- The use of longer-term average of the SAAR (e.g. 6-month moving average) rather than focusing only on the monthly SAAR is recommended.
Example of trend comparisons using data from June 2025 Starts and Completions Survey
Following is a comparison of percent change in housing starts using the above-mentioned calculations. Fluctuations in trends can be masked nationally due to off-setting numbers at the local level. However, differences in trends appear using various calculation methods when looking at Census Metropolitan Areas (CMAs). This is especially true if there is a significant monthly fluctuation in housing starts.
In Vancouver, a surge in multi-family starts during June 2025 compared to June 2024 could lead to the erroneous conclusion that starts are on the rise in that CMA. However, when comparing the actuals and the SAAR over the first 6 months of 2025 to the first 6 months of 2024 reveals that starts activity is lower in 2025 compared to 2024.
In Toronto, all comparison methods show a similar downward trend because monthly starts activity in that CMA has been consistently lower in 2025 compared to 2024.
In Montréal, a comparison of year-over-year actuals or monthly SAARs could lead to the conclusion that overall starts activity in that CMA in 2025 is close to what was observed in 2024. However, a comparison of data over a longer period shows that starts in Montréal are higher in 2025 than in 2024.
It’s also important to note that current trends in housing starts may not always capture real-time market conditions. This is due to the time it takes to move across the following 3 important steps for large scale projects such as high rise-condominiums and rental structures:
- the planning and pre-sale phases of high-rise condominium and rental projects
- the issuance of building permits, and
- the actual start of construction
Starts associated with large scale projects often are reflective of commitments made when market conditions were different than current market conditions.
This is why it’s important to gather market intelligence from industry on current on-the-ground experiences. It helps determine if this information has already been captured in the housing starts data or will be in the future. CMHC staff maintain constant dialogue with industry representatives. This ensures we have the information needed to provide a comprehensive analysis on the state of the housing market.
Condo pre-sales down, however building permits still showing pending housing starts
Condominium developers in Vancouver and Toronto CMAs are reporting significant reductions in pre-sales. However, this does not mean that housing starts will no longer be seen over the coming months in these areas.
CMHC is currently tracking 186 building permits in Vancouver for a total 21, 907 condominium apartment units for which construction has not yet started.
In Toronto, we are currently tracking 37 permits for a total of 6,648 condominium apartment units for which construction has not yet begun.
These permits are likely associated with pre-sales that happened much earlier than the current reduction in sales activity and these “pending” starts may still happen. Starts associated with these permits will be confirmed via ongoing conduct of the Starts and Completions Survey. Beginning in the 4th quarter of 2025, CMHC will also report on the time between the building permit date and the associated housing starts date.
CMHC supports a better understanding of current and expected housing market conditions with regular releases of analysis updating our view on the housing narrative. Read CMHC’s latest forecast of housing starts, the Summer Update: 2025 Housing Market Outlook.
Was this page relevant to your needs?
Date Published: August 18, 2025
