Canada Approved 143,929 Homes That Nobody Can Afford to Build
In September, CMHC counted 143,929 homes in Canada that have a building permit and no construction started, up 1.1% from August. Over the same month, starts in centres of 10,000 or more fell 11% year over year, to 19,984 units from 22,375.
The standard prescription for the housing shortage is to approve more, faster. Cut the rezoning delays, legalize density, and supply follows. For years that was the right call, and it still matters. But a growing pile of permitted, unbuilt homes says the binding constraint has moved. What decides whether a home gets built in 2026 is whether the numbers work on the day the shovel would go in. If you develop, invest, lend or write policy in British Columbia, that shift changes which projects you back, which data you trust, and where you push.
Approvals reform did its job
British Columbia went further than most provinces. The Housing Supply Act sets targets municipalities are expected to hit, and the small-scale multi-unit rules allow 3 units on small lots, 4 on standard lots and 6 near frequent transit. Abbotsford and the rest of the Fraser Valley carry real weight in those targets as the Lower Mainland's growth corridor.
Those changes removed a genuine delay. Projects that once spent a year or more in rezoning now reach permit stage far sooner. The supply of buildable sites has grown, and that is a lasting gain.
A permit is an option, priced on start day
A permit gives the owner the right to build. Nobody is obliged to use it. A builder exercises that right when the expected value of the finished units clears construction cost, financing cost, and the return that justifies the risk, with enough left over to satisfy the lender.
Take an illustrative fourplex. Say the all-in cost to build, carry and sell comes to $3.2 million and the four units are expected to sell for $3.4 million combined. That $200,000 cushion is about 6%. A lender looking for a wider margin, or a modest rise in trade costs, or a slower sales pace, erases the deal on paper. The permit stays in the drawer.
That is the squeeze across the market. CREA reported the national average sale price up just 0.6% year over year in August 2026, and September sales down 6.9%. Labour stays expensive, lending criteria for construction are tighter, and coverage tests are harder to pass. Flat values against firm costs keep the gap closed, and adding approvals to that market adds options nobody can afford to exercise.
Trades matter too. A permit doesn't frame walls; a carpenter does, and B.C. has been short of skilled trades for years. Some unstarted projects are waiting on crews as much as on capital.
Watch starts, and the cost side
For anyone measuring progress, permits overstate it. Canada recorded 259,028 starts back in 2025, against CMHC's estimate that the country needs 3.5 million additional units by 2030. Starts are the number that eventually becomes inventory. The policy levers with the most room left sit on the cost side: development charges and fees, the cost and terms of construction financing, and certainty about timelines so builders can price risk.
For developers and investors, test every project against start-date conditions, using today's costs and today's lending terms. A pro forma built at approval can be badly out of date by the time a crew is available. Anyone buying pre-construction in the Fraser Valley should ask what happens if the start slips a year, because more projects are slipping.
There's also a timing effect to plan for. When financing loosens or values firm, a large share of these held permits becomes viable at once. Trades, materials and construction lenders will all be in demand together, and the projects with financing lined up early will move first.
The question worth asking of any site, any policy and any pipeline report is the same: at current costs and current lending terms, would this start next month?
In September, CMHC counted 143,929 homes in Canada that have a building permit and no construction started, up 1.1% from August. Over the same month, starts in centres of 10,000 or more fell 11% year over year, to 19,984 units from 22,375.
The standard prescription for the housing shortage is to approve more, faster. Cut the rezoning delays, legalize density, and supply follows. For years that was the right call, and it still matters. But a growing pile of permitted, unbuilt homes says the binding constraint has moved. What decides whether a home gets built in 2026 is whether the numbers work on the day the shovel would go in. If you develop, invest, lend or write policy in British Columbia, that shift changes which projects you back, which data you trust, and where you push.
Approvals reform did its job
British Columbia went further than most provinces. The Housing Supply Act sets targets municipalities are expected to hit, and the small-scale multi-unit rules allow 3 units on small lots, 4 on standard lots and 6 near frequent transit. Abbotsford and the rest of the Fraser Valley carry real weight in those targets as the Lower Mainland's growth corridor.
Those changes removed a genuine delay. Projects that once spent a year or more in rezoning now reach permit stage far sooner. The supply of buildable sites has grown, and that is a lasting gain.
A permit is an option, priced on start day
A permit gives the owner the right to build. Nobody is obliged to use it. A builder exercises that right when the expected value of the finished units clears construction cost, financing cost, and the return that justifies the risk, with enough left over to satisfy the lender.
Take an illustrative fourplex. Say the all-in cost to build, carry and sell comes to $3.2 million and the four units are expected to sell for $3.4 million combined. That $200,000 cushion is about 6%. A lender looking for a wider margin, or a modest rise in trade costs, or a slower sales pace, erases the deal on paper. The permit stays in the drawer.
That is the squeeze across the market. CREA reported the national average sale price up just 0.6% year over year in August 2026, and September sales down 6.9%. Labour stays expensive, lending criteria for construction are tighter, and coverage tests are harder to pass. Flat values against firm costs keep the gap closed, and adding approvals to that market adds options nobody can afford to exercise.
Trades matter too. A permit doesn't frame walls; a carpenter does, and B.C. has been short of skilled trades for years. Some unstarted projects are waiting on crews as much as on capital.
Watch starts, and the cost side
For anyone measuring progress, permits overstate it. Canada recorded 259,028 starts back in 2025, against CMHC's estimate that the country needs 3.5 million additional units by 2030. Starts are the number that eventually becomes inventory. The policy levers with the most room left sit on the cost side: development charges and fees, the cost and terms of construction financing, and certainty about timelines so builders can price risk.
For developers and investors, test every project against start-date conditions, using today's costs and today's lending terms. A pro forma built at approval can be badly out of date by the time a crew is available. Anyone buying pre-construction in the Fraser Valley should ask what happens if the start slips a year, because more projects are slipping.
There's also a timing effect to plan for. When financing loosens or values firm, a large share of these held permits becomes viable at once. Trades, materials and construction lenders will all be in demand together, and the projects with financing lined up early will move first.
The question worth asking of any site, any policy and any pipeline report is the same: at current costs and current lending terms, would this start next month?
Read Next
Why a Stress Test Floor Change Barely Moves Your Pre-Approval
Assume a 5.25% Stress Test and You'll Shop for a Home You Can't Qualify For
Contract Rate or the 5.25% Floor: Which One Decides What You Can Buy
Rate Shopping Won't Fix a Qualification, Penalty or Cash Flow Problem