What CMHC's July construction pipeline says about who you will be competing with when you sell
Credit: Shutterstock
Canada Mortgage and Housing Corporation released its July construction data on August 18, and the headline number is blunt. Actual housing starts in centres of 10,000 or more fell 19% from a year earlier. Most coverage stopped there, or led instead with the smaller 5% decline in the seasonally adjusted annual rate — a different number, measuring a different thing, over a different geography.
Neither framing answers the question a homeowner actually has. Housing starts get filed as a builder story, a proxy for cranes and permits and construction payrolls. But if you already own, a starts number is something else entirely. It is a rough forecast of who will be listed against you when you eventually sell.
That forecast has two halves, and they run in opposite directions. What follows walks the July release stage by stage, defines the metrics that keep getting conflated in coverage, and separates the supply that is already locked in from the supply that may never arrive.
Stay on top of Canadian homeowner news
The headline decline is real, and it is steep. Actual monthly housing starts came in at 18,834 units in centres of 10,000 or more, down from 23,155 units in July 2025, according to CMHC. That is the 19% in the headline.
Then there is the proportion that headline strips out. Year-to-date starts through July totalled 131,851 units, down 4% from the same stretch of 2025. A single month fell off a cliff; seven months of activity slipped modestly. Both statements are true, and only one of them describes a trend.
The smoothed measures land between the two. The six-month trend measure sat at 247,377 units, down half a percentage point from June. The total seasonally adjusted annual rate for all areas fell 5%, to 229,074 units from 240,773. Rural starts were estimated at an annual rate of 11,274 units.
CMHC publishes all three because the multi-unit segment is lumpy by nature. A single large apartment project breaking ground — or not — moves a monthly figure in ways that say nothing about the market's direction. The agency's own guidance is that analyzing the monthly rate alone can be misleading, precisely because multi-unit construction drives the market and swings hard from month to month.
That caution matters here. A 19% drop is a headline; it is not yet a verdict. Readers who watched May's 6% decline to a 258,010-unit pace will recognize the pattern of a single month overstating a slower underlying drift.
Most confusion about housing data comes from treating the construction pipeline as one number. It is not. It is a sequence of stages, each measuring a different moment in a project's life, and each reaching the market at a different time.
The under-construction and completions figures cover centres of 50,000 or more, a narrower geography than the starts count. That distinction matters when comparing the numbers directly, and it is routinely lost in summary coverage.
Read as a sequence, the table tells a cleaner story than any single line in it. Money and permits move through these stages in order, and the delay between stages is measured in quarters and years, not weeks. A project approved today is not competing with you at any point in the near future. A project completing today already is.
The permit stage is the earliest signal available, which is why building permit data tends to move before starts confirm it.
Here is where the July release splits in half.
The near-term picture is one of abundance, not scarcity. There were 373,091 units under construction in July, up slightly from June, and completions rose 8.1% to 19,773 units. Those homes are being finished right now. Nothing that happened to July's starts number changes when they land, because the concrete is already poured and the financing is already committed. CMHC's Housing Observer puts the construction period for multi-unit housing at one to two years from start to completion, inside a total timeline of seven to eight years from project conception to occupancy.
Run that clock forward and the arithmetic is straightforward. Units that broke ground in July arrive in 2027 and 2028. Units already well into construction arrive sooner. For anyone thinking about selling in the next eighteen months, the competitive supply is effectively fixed, and it is still growing.
The longer-term picture is where July's decline actually bites. Fewer starts now means fewer completions in 2028 and 2029. Tania Bourassa-Ochoa, CMHC's Deputy Chief Economist, framed both halves in a single sentence: although the pipeline of homes under construction remains substantial and completions are increasing, fewer new projects are being started in many markets, notably in Vancouver, Calgary and Toronto.
The supply that will compete with your listing in 2027 is already built or being built. The supply that would have competed with you in 2029 is the part that is thinning. Any read on "easing competition" is a statement about the back half of this decade, not about next spring.
The 141,480 figure is the most interesting number in the release and the most frequently misread. It counts units that have received a building permit but where CMHC's survey has observed no construction activity — units the agency describes as potentially indicative of future housing starts.
That phrasing is deliberate, and it cuts against the intuitive reading. This is a queue, not a graveyard. It also grew last month, rising 3% from June. A backlog getting bigger is genuinely ambiguous: it can mean projects are stacking up because conditions are poor, or that a wave of approvals is working its way toward shovels.
Timing tells you which reading to lean toward. The lag between permit and start runs roughly two to ten months for single-detached homes and nine to fifteen months for mid- to high-rise apartments. On normal timing, most of that 141,480 breaks ground between late 2026 and late 2027 — and completes well after that.
The mechanism that stalls them is financing, and CMHC has documented it plainly: when completed units do not sell, lenders restrict credit, and developers delay or cancel new projects because many rely on high presale thresholds to secure financing, according to its Spring 2026 Housing Supply Report. The same report counted 6,134 unsold condominium apartments at completion across seven major markets in 2025, up from 5,019 a year earlier, and recorded what it called record condominium project cancellations.
So the cancellation dynamic is real. It is also specific — a condominium story concentrated in a handful of large markets, not a national arithmetic identity. The presale threshold sits at the centre of it, which is why pressure on the 70% presale requirement has become its own storyline.
National numbers are an average of markets that are not behaving alike. In July, Vancouver starts fell 42% year over year. Toronto fell 10%. Montreal rose 3%.
A 45-point spread in a single month is not noise around a national trend. It is evidence that the national trend is a statistical artifact of three very different markets, and that an owner in Montreal and an owner in Vancouver are reading the same release about entirely different futures.
The divergence is expected to persist. CMHC's Summer 2026 outlook forecasts national starts declining from 241,400 in 2026 to 223,400 in 2027 and 211,900 in 2028, and says historically low levels of construction will be most visible in Ontario and British Columbia, particularly in the condominium market. Within that, Toronto apartment starts are forecast to hold at roughly 20,000 to 22,200 units annually through 2028, even after July's 10% actual decline.
That last detail is worth sitting with. Toronto's one-month drop and Toronto's multi-year forecast disagree — which is exactly the situation the trend measure exists to handle. Vancouver's trajectory, by contrast, is forecast to keep falling. Owners who followed February's split, when Toronto dropped 28% while Vancouver surged 60%, have already seen how quickly these positions trade places.
For owners, the honest summary is a timing statement rather than a price call.
If you are selling in the next year or two, this release offers no relief. Completions are rising, hundreds of thousands of units are mid-build, and new inventory keeps arriving — particularly in condominium-heavy submarkets where finished, unsold units are already accumulating. The competitive picture for that window is set.
If your horizon runs to the back half of the decade, the thinning pipeline is a genuine tailwind, strongest in the markets where starts are falling hardest. That is a conditional forecast, not a promise; it depends on the approved-but-unstarted backlog converting slowly or not at all, and on demand not shifting underneath it. The broader freeze in transaction activity — 55% of homeowners now say they are staying put indefinitely — cuts in the same direction by keeping resale listings scarce.
There is a third possibility worth flagging carefully, because it is frequently asserted without evidence: that falling starts free up trades and make renovation quotes more competitive. The mechanism is plausible, but it has not shown up in the data yet. Construction employment actually rose by 16,000 jobs, or 1.0%, in July and was little changed year over year, Statistics Canada reported. Starts fell 19% in the same month that construction hiring went up. If capacity does loosen, it will lag the construction slowdown rather than track it — and the labour data is the place to watch for it, not the starts release.
The next CMHC housing starts and construction release lands September 16 at 8:15 a.m. ET. Three things in it will confirm or complicate everything above.
Housing starts are a slow signal, and CMHC's own guidance is that they are likely to remain subdued over the coming months. For owners, the useful discipline is to stop asking what the pipeline says about the market and start asking what it says about a specific year — because the answer for 2027 and the answer for 2029 are not the same answer. Provincial targets face the same arithmetic, as Ontario's revised forecast of 64,800 starts illustrates.
Stay on top of Canadian homeowner news
About the Author
Angela Nightingale is the Senior Editor at Homeowner.ca with two decades of experience in digital publishing and content strategy. She has owned two homes, taken on her share of DIY projects, and learned what most guides fail to mention. She writes from the belief that the best home guidance comes from people who have lived through the decisions — and her goal is always to leave readers feeling confident, not overwhelmed.
