CMHC's August Data Shows Fewer Finished Homes Nationally — but the Decline Sat in Quebec and B.C., Not the GTA
Credit: Shutterstock
Canada Mortgage and Housing Corporation reported on September 16, 2026 that the six-month trend in housing starts slipped 1.3% in August to 244,149 units. That is the number most coverage will lead with. For a homeowner planning to list this fall, it is not the most useful one. A second figure in the same release says construction finished on just 17,550 units in August, down 11.2% from July. Starts describe homes that might exist in two or three years. Completions describe homes that are arriving now.
That distinction matters because a finished new home is one of the few things that competes directly with a resale listing for the same buyer in the same season. Fewer completions means, in plain terms, less brand-new inventory landing during the months when fall sellers are trying to attract offers. But the national number hides a regional story that cuts against the obvious reading. The decline did not happen in Ontario. It happened mainly in Quebec and British Columbia, while completions in Ontario and the Toronto area rose sharply.
This piece walks through what CMHC actually measured, where the drop occurred, what it can and cannot tell a GTA seller, and which data releases are worth watching before you commit to a listing date. It is a supply-side reading of one month of data, not a price forecast.
Stay on top of Canadian homeowner news
The August release contains several indicators that are easy to blur together. According to CMHC's August 2026 housing starts release, actual starts in centres of 10,000 or more totalled 17,691 units in August, 2% fewer than the 18,112 recorded in August 2025. The year-to-date total of 149,542 units is running 4% behind last year. The monthly seasonally adjusted annual rate was essentially flat at 229,046 units, compared with 229,360 in July. Deputy Chief Economist Kevin Hughes said modest gains in Quebec and Alberta "only partially offset the decline in other provinces, most notably, Ontario," and added that CMHC expects the downward trend in starts to continue as construction moderates toward year-end.
Hughes was talking about starts. The construction-stage figures, which cover centres of 50,000 or more, tell a separate story, and they are the ones that bear on this fall.
Read the table from the bottom up and it becomes a timeline. Permits turn into starts, starts become units under construction, and those units eventually become completions. Only the completions row describes inventory that a buyer could realistically choose instead of your house this fall.
The definitions are more precise than the headlines suggest. In its survey methodology for starts and completions, CMHC defines a start as the beginning of construction, usually when concrete has been poured for the entire footing. A completion is the point at which all proposed construction work has been performed, although a unit can be counted as complete with up to 10% of the work remaining. The agency also tracks whether completed ownership units are absorbed (sold) or unabsorbed (finished but unsold).
That is the mechanism behind the seller angle. A start in August adds nothing to this fall's market. A completion does: it is either a presold unit whose buyer is now moving in, or an unsold unit that a builder needs to clear. Both shape local demand. The second competes head-on with resale listings.
Two cautions keep this honest. First, one month of completions is noisy; CMHC itself leans on a six-month trend for starts because monthly figures swing hard, and completions are no calmer. Second, a large share of completions are purpose-built rentals and presold condominiums, which do not sit on the market alongside a detached resale home. Fewer completions is a real easing of one competitive pressure. It is not a guarantee of more offers.
Completions, units under construction and the approved-but-not-started count are reported for centres of 50,000 or more. Actual starts are reported for centres of 10,000 or more, and the trend covers all areas. The figures are not directly additive.
The national decline is real, but the regional breakdown in CMHC's tables changes who should read it as good news. Quebec completions fell by more than half. British Columbia's fell by a third. Ontario's went the other way.
This matters for anyone applying the national figure to a GTA listing. In the Toronto census metropolitan area, completions more than doubled in August, from an unusually low July base. Owners in Montréal and Metro Vancouver are the ones who saw a meaningful month-over-month thinning of newly finished supply. Ottawa also slipped, from 843 completions to 591.
The Ontario story in this release is about starts, not completions. Ontario's six-month starts trend fell 7%, and actual August starts in the province were 22% below August 2025. In the Toronto CMA, actual starts dropped 45% year over year, to 1,192 units. That is the part of Hughes' comment that applies to Ontario: less building is beginning now, which shapes supply in 2028 and beyond more than it shapes this fall.
For context on how new-home pricing has been moving in the province, see our earlier coverage of new home prices falling as Alberta and Ontario led the national retreat.
The tempting GTA narrative goes like this: completions are weak, Toronto's starts trend is flat, so less new supply explains why resale prices have stopped falling. The August data does not support the first half of that sentence for Toronto. Completions there rose. What the data does support is a slower, structural version of the argument.
Resale prices in the GTA have levelled off month to month. The Toronto Regional Real Estate Board's August market report shows the MLS Home Price Index composite benchmark was 4.5% lower than a year earlier but essentially flat against July on a seasonally adjusted basis. Sales of 5,057 homes were down 2.1% year over year, and new listings of 12,075 were down 14.1%. TRREB said the number of homes available for sale was "down noticeably" from last year and that sales were arguably limited by less choice in some neighbourhoods.
That is a resale-supply story first. Construction data adds a longer-range layer: CMHC's September 10 housing supply release says that in Toronto the supply gap is "increasingly concentrated in the ownership market, where new construction remains exceptionally weak." Much of what is being built in major markets is purpose-built rental. For an owner selling a detached or semi-detached home, that means the new-build supply competing for the same ownership buyer is thinner than the raw completions count implies.
None of this demonstrates that weak construction caused prices to stabilize. Price stability has several drivers, and the benchmark and average price measure different things. The fair reading is narrower: in the GTA, the new-ownership pipeline is weak, resale listings are down year over year, and prices have stopped sliding month to month. Those facts point in the same direction without proving a cause.
Our coverage of GTA home sales jumping 9.4% in June tracks how that tightening developed earlier in the summer.
New construction is only one source of competition. For most fall sellers, the larger one is the house down the street. The Canadian Real Estate Association's August release shows national new listings rebounded 3.3% month over month in August, reversing three straight declines earlier in the summer. CREA Chair Garry Bhaura said the increase in new supply was broad based across the largest markets and most apparent toward the end of the month, suggesting sellers were getting an early start on the fall season.
The national sales-to-new-listings ratio eased to 49.1% from 51.1% in July, still inside the roughly 45% to 65% band CREA associates with balanced conditions. Inventory held at 4.8 months for a fourth straight month, slightly below the long-term average of five. CREA also noted that prices have been largely unchanged month to month since spring, the longest stretch of price stability since 2024.
TRREB's figures echo that pattern locally: GTA new listings rose from July on a seasonally adjusted basis even as they remain well below last year. In practical terms, a fall seller in the GTA faces less new-build competition than a seller in Montréal or Vancouver might assume from the headline, but more resale competition than they faced in July.
For the national price picture behind those numbers, see our breakdown of Canadian home prices sitting 3% below last year.
The approved-but-not-started count rose 0.7% to 142,423 units in August. In the Toronto CMA it climbed 9.2%, to 19,516. It is tempting to treat that figure as a wall of supply waiting to hit the market the moment financing improves. The research does not support a fast flood.
A CMHC Housing Observer analysis of permits and starts found that building permits generally lead apartment starts by about 6 to 15 months, with longer delays for high-rises: roughly 273 days for Toronto CMA high-rises, versus 92 days for single-detached homes. CMHC also notes that condominium projects generally need about 70% of units presold before bank financing can be secured, and that projects can be paused, repurposed or halted before they ever reach a start.
CMHC's Fall 2026 Housing Supply Report adds that high construction costs, financing challenges, weak presales and affordability constraints continue to keep many condominium projects from reaching the sales thresholds they need to proceed.
Put those timelines together. A permitted high-rise that starts next spring would not typically be completed for years. The backlog is a risk to the long-run competitive picture, not to this fall's. It is a reason to avoid waiting indefinitely for a better market, rather than a reason to rush a listing that isn't ready.
Earlier in the year, we covered how housing starts fell 6% in May as the GTA condo pipeline slowed.
The approved-but-not-started figure counts units with building permits where CMHC has not yet observed construction. CMHC describes it as indicative of future starts, not a commitment that those homes will be built on any schedule.
Slower construction often frees up trades, and some homeowners will read August's data as a signal to book pre-listing or pre-winter work. The evidence here is thinner than the intuition. Statistics Canada's first-quarter 2026 construction price report described subdued demand for new projects alongside persistent skilled-labour shortages in selected trades. Ontario builders reported less bidding and delayed developments. Residential construction costs rose 0.6% in the quarter and 2.8% year over year across 15 metro areas, with Toronto among the smallest quarterly increases at 0.1%.
That data is from early 2026 and concerns new construction, not renovation quotes. It suggests softer demand in parts of the Ontario building sector, but it does not measure how quickly a contractor will return your call this October. If a pre-listing repair is on your list, getting a quote now costs little and tells you more about your local market than national indices can.
August is one data point. The signals that will confirm or overturn it arrive over the next few weeks.
The most useful pattern for a GTA seller would be two things at once: continued weakness in Ontario starts and ownership-oriented construction, and new resale listings rising no faster than sales. If completions in Toronto stay elevated while listings climb, the competitive picture for this fall gets tougher, whatever the national headline says.
Our July housing starts coverage offers a month-earlier baseline for comparing the next release.
When the October data lands, compare your own CMA's completions and approved-but-not-started lines in CMHC's metro table, not just the national total. As August showed, the national direction and your local direction can be opposite.
Stay on top of Canadian homeowner news
About the Author
