Alberta built the housing it needed, and finished it as the people stopped coming

2023 · the wave

First, more people arrived than at any time on record

In a single quarter, the third quarter of 2023, net migration reached 58,649 people, of whom 42,135 came from outside Canada. Nothing in sixty years of the record comes close. Alberta has always run on migration, and the shape of the rest of this chart is the shape of the province's economy — people arrive when there is work and leave when there isn't. What happened from 2022 is different in kind, not just degree.

Then it stopped, and it stopped mostly from one direction. Between 2023 and 2025 the international component fell from 144,293 to 19,138, while interprovincial migration — Canadians moving to Alberta from other provinces — went from 44,363 to 22,216, a far gentler landing. In the first quarter of 2026 the international number went below zero, to -648. That is not a fall in immigration: the same quarter recorded 11,139 immigrants arriving. It is the non-permanent-residentNon-permanent residents. People living in Canada on a work permit, study permit, or similar temporary status, plus their families. They are counted in the population and they occupy housing. A federal policy change that shrinks this population shows up as a negative number in the migration accounts even when immigration itself is unchanged. population shrinking by 9,993 — enough to outweigh everyone arriving with permanent status.

Only 3 quarters in the whole record show a negative international number, and before the current one the last was the third quarter of 2020, when the border was closed for the pandemic. Alberta's population is still growing — 5,057,077 people as of the second quarter of 2026 — because a smaller positive number is still positive. These are estimates and they are revised; Statistics Canada changed how it counts non-permanent residents in 2024, inside the window this page depends on.

1948–2026 · the response

Alberta answered with the largest building year it has ever had

In 2025 Alberta started 53,184 homes, beating the previous record of 46,632, set the year before, by 14.1%. Against the low of 2016 it is 135% more building. Behind that number sit 79 years of Alberta housing startsHousing start. A dwelling counted at the moment its foundation is begun — not when the permit is issued, and not when anyone moves in. It is a measure of construction beginning, which is why it says nothing directly about how much housing exists yet. — every year CMHC has counted them — and a province that has built in waves: the late-1970s boom, the collapse through the 1980s, the long climb to the 2006 peak of 40,882 homes, the bust after it.

Adjusted for the province it serves, 2025 works out to 10.6 homes started per thousand Albertans, against an average population of 5,017,119 that year. This was a genuine, large, well-intentioned supply response to a real shortage — the thing housing policy had been asking for.

The last bar is not a crash — it is half a year. 2026 is drawn hollow because it contains only 6 months of data, to June 2026. Compared like for like against the same 6 months of 2025 — 22,077 starts against 27,902 — building is running −20.9%, off the record pace but still above every year before 2024 at this point in the calendar. Alberta building is strongly seasonal, so part-year totals can only be read against the same months of another year.

Demand and supply, same axis

The two lines crossed, and the lag is the whole story

Put the two series on one scale — each indexed so that its own 2010–2019 average equals 100 — and the argument of this page is a single picture. For two decades the lines move together, because builders respond to the market they can see. Then migration goes somewhere the starts line cannot follow: in 2023 demand stood at 432 against its decade norm while starts were at 124.

Building did answer — but it answered late, because it cannot answer any other way. Starts peaked in 2025, 2 years after migration peaked in 2023. By then demand had fallen to 95 on the same index, slightly below its own decade average, while supply had climbed to 187 — the highest reading on the chart. The scissors close the wrong way round. Nobody made a mistake at any single point in that sequence; the mistake is structural, and it is made of construction time.

Two series with different units cannot share an axis without a common baseline, and a single base year would be a poor one for a series as volatile as Alberta migration — so each line is divided by its own average across the 2010s decade. The shapes are the data; the vertical distance between the lines is an artefact of that choice and should not be read as a housing shortfall in units. The 2026 part-year is excluded here, since an incomplete year cannot be compared with complete ones.

What is still coming is visible in the work already committed. In Calgary and Edmonton together, 42,802 dwellings were under construction as of June 2026 — at the completion pace of 2025, about 12 months of finishing work, all of it already begun and paid for. Those two cities completed 43,496 homes in 2025, against 29,087 in 2023; in the first 6 months of 2026 completions are running 5.3% against the same months of 2025. The supply is still arriving.

Calgary and Edmonton only. No Statistics Canada or CMHC table publishes this series for Alberta as a province. These two metropolitan areas hold roughly two-thirds of Alberta's people, which makes them a reasonable proxy and not a provincial total.

2016–2025 · the mix

What got built was not what Alberta usually builds

Alberta is a province of detached houses, and for most of the last forty years its building record looked like it. The record year does not. Apartments and other multi-unit types were 50% of everything started in 2025 — 26,572 of 53,184 units — against 30% in 2016. Single-detached houses fell from 44% of starts to 28%.

Detached-house building did not stall while apartments took off. Single-detached starts rose 51.3% between 2016 and 2025, from 10,010 homes to 15,146 — a substantial expansion by any normal standard. Apartments simply grew far faster, 297.6% over the same nine years, which is what moves a share. The one year detached building actually fell was the last one: −3.5% from 2024 to 2025, while apartments rose another 29.1%.

Even after nine years of growth, 2025's 15,146 single-detached starts remain 39.1% below Alberta's all-time detached peak of 24,850 in 2006 — the mix shifted partly because detached building never fully returned to what it once was.

Apartments taking half a year's starts is not unprecedented: it happened in 1968, 1969 and 1982, and the last time apartments were at least half of Alberta's starts was 1982. What is new is the scale underneath the share.

Both charts are drawn only for the years that carry a dwelling-type breakdown, which begins in 1959. The 11 earlier years — 1948 to 1958 — have a total but no split, and they are absent from these charts rather than drawn as zero. Shares are computed against the sum of the four published types, because in three years of the 1960s the published types do not add exactly to the published total.

1988–2025 · intended market

A province of homeowners started building rental at scale

Dwelling type tells you the shape of a building, not who it was meant for. A separate CMHC series records the intended marketIntended market. What a builder says a dwelling is for at the moment construction starts: purpose-built rental, condominium, or freehold homeowner. It is an intention recorded at the start, not a guarantee about what the unit becomes — projects do convert between rental and condo mid-build. at the moment construction starts: purpose-built rental, condominium, or freehold homeowner. On that measure the change is even sharper than the type mix. In 2016, Alberta started 1,410 purpose-built rentalPurpose-built rental. An apartment building constructed to be rented, and owned as a single building by one landlord. Distinct from a condo building whose individual units are sold to owners, some of whom then rent them out. units — 7% of the three markets combined. In 2025 it started 19,224, or 38%, an increase of 1263%.

For most of the period this series covers, purpose-built rental construction in Alberta was close to a rounding error — the private market built condos and houses, and rental supply came second-hand. What the 2025 figures describe is a different industry: 19,224 rental starts against 10,022 condominium and 21,513 freehold homeowner. It also means a large share of the record supply lands in exactly the segment the vacancy and rent surveys measure — which is why the next two sections move as fast as they do.

They sum to 50,759 against 53,184 starts in 2025, a shortfall of 2,425 units. Two reasons: co-operative and “other market” categories exist in the source cube and are omitted here because Alberta volumes are negligible, and this is a different table with a slightly different universe from the dwelling-type series above.

The two should never be subtracted from one another. Intended market is also an intention recorded at the start — projects do convert between rental and condo mid-build.

2023–2025 · vacancy

The rental market went from nothing available to a normal amount, fast

In 2023, at the height of the arrival wave, 1.4% of Calgary's purpose-built rental stock was empty and available — the market that produces queues at viewings and applications sight-unseen. That figure is the vacancy rateVacancy rate. The share of purpose-built rental units sitting empty and available, measured once a year in an October survey. It is the price of housing expressed as availability rather than as dollars: when it rises, landlords compete for tenants instead of the other way round., the housing shortage expressed as availability rather than dollars. Two years later, in 2025, it was 4.9%, 3.5 times as high. Edmonton went from 2.3% to 3.8% over the same two surveys.

Alberta has been at these levels before and more: Calgary's vacancy reached 7.0% in 2016, after the oil-price crash, and fell as low as 0.5% in 1997. Today's reading is a market that has loosened to something like an ordinary balance, not one that has broken. What is unusual is the speed, and the fact that it happened while the largest supply wave in the province's history was still completing.

There are two vacancy rates, and this page shows both. The solid lines are CMHC's row-and-apartment structures of three units and over, which this page treats as primary. The dashed lines are apartment structures of six units and over — a different population of buildings, reaching back to 1971 where the primary series only starts in 1992.

They disagree: in 2025 the six-plus universe puts Calgary at 5.0% against 4.9%, and at their widest, in 1995, the two were 0.5 percentage points apart. They are different questions and must never be spliced into one line.

Both exclude the secondary rental marketSecondary rental market. Housing rented out that was not built to be rental: condos bought by investors, basement suites, rented single-family houses. CMHC's rent and vacancy surveys exclude it entirely, and in Calgary it is large and growing — which is where much of the new condo supply lands. — rented condos and basement suites are in neither survey.

Calgary and Edmonton only. No Statistics Canada or CMHC table publishes this series for Alberta as a province. These two metropolitan areas hold roughly two-thirds of Alberta's people, which makes them a reasonable proxy and not a provincial total.

2021–2025 · rents

Rents rose hard for three years, then stopped rising

A two-bedroom apartment in Calgary averaged $1,355 a month in 2021 and $1,914 in 2025 — 41.3% in four years. Alberta consumer prices rose 15.4% over the same period, so in constant 2025 dollarsReal and nominal. Nominal is the number as it was written on the cheque. Real restates it in one year's dollars so amounts from different years can be compared. Any rent or price change spanning more than a few years is misleading in nominal terms alone. the increase is 22.5%: still large, and roughly half the headline. Edmonton rose 26.2% nominally and 9.4% in real terms, from $1,270 to $1,603.

The turn is in the final year. Calgary rents rose 1.7% between 2024 and 2025 — from $1,882 to $1,914 — after three years of increases many multiples of that. Edmonton's last year was 4.4%. That deceleration is what new supply arriving into softer demand looks like in the rent series, and it arrives at roughly the same time as the vacancy rise, which is what one would expect if both are measuring the same thing.

CMHC measures rents in occupied units of the existing purpose-built stock, which is not the same as advertised rents. Sitting tenants roll over slowly, so this series lags asking rents by roughly a year at turning points — it is likely to be late rather than early in showing a change. It also excludes the secondary rental market entirely, which in Calgary is large and growing and is where much of the new condo supply lands. Real figures use the Alberta all-items consumer price index, annual average, expressed in 2025 dollars.

Calgary and Edmonton only. No Statistics Canada or CMHC table publishes this series for Alberta as a province. These two metropolitan areas hold roughly two-thirds of Alberta's people, which makes them a reasonable proxy and not a provincial total.

New-home prices

Calgary and Edmonton have spent twenty years going opposite directions

Both cities spiked violently in the 2006–07 boom, and then they parted. Edmonton peaked in October 2007 at 116.7 on the New Housing Price IndexNew Housing Price Index. Statistics Canada's index of what contractors charge for NEW homes — roughly a tenth of housing transactions. It is not the resale price a realtor would quote, and this page has no legally reusable resale series to show instead. and, 19 years later, sits at 106.0 — 9.2% below that peak in nominal terms, before any adjustment for two decades of inflation. Calgary was flat for a decade after the same bust, then climbed from 2021 to a peak of 133.8 in February 2025, and has since eased 2.9% to 129.9. They read as separate housing markets that happen to share a province.

That index is narrower than it sounds. It measures what contractors charge for newly built homes, roughly a tenth of housing transactions. It is not the resale price a realtor would quote, and there is no lawful way for this page to show you that number: the two authoritative resale series, CREA's MLS® HPI and the Teranet–National Bank HPI, both forbid commercial use and derivative works. Nothing on this page is the price a realtor would quote you.

Because the index is set to 100 in December 2016, the levels are themselves percentage changes since then: new homes in Calgary are 29.9% more expensive than in that base month, Edmonton 6.0%, and Canada as a whole 20.6%. Whatever national story about house prices a reader arrives with, Edmonton has not been part of it.

The index is quality-adjustedQuality-adjusted. Corrected so that changes in what is being built don't masquerade as changes in price. Without it, a year of larger houses with better finishes would show up as inflation rather than as a different product., so a year of larger houses does not read as inflation. It also tracks list prices rather than effective prices: a builder who holds the sticker price and adds thirty thousand dollars of upgrades appears here as flat while the real price fell. In a soft market that bias runs one way, and this page cannot correct for it.

House against lot

Is it the building that got expensive, or the ground under it?

The same index splits a new home's price into the building and the serviced lotLand only. The serviced-lot component of a new home's price inside an urban subdivision — the ground, graded, with water, sewer and road already run to it. It is not raw land, not farmland, and not an acreage: it says nothing about what a quarter-section outside town is worth. it stands on, which lets one of the standard arguments about housing costs be checked rather than asserted. In Calgary the two have come apart: the building component stands at 136.3 against 118.7 for the lot — 36.3% and 18.7% respectively since December 2016. Construction cost, not land, is doing most of the work in Calgary's new-build prices. In Edmonton the two move nearly together, at 106.1 and 104.1, and neither has gone far.

“Land only” here is a builder's serviced lot, not a land market. It is the graded ground inside a new urban subdivision with water, sewer and road already run to it — a cost line in a new house's price, not the value of an acreage. No free, commercially licensed price series exists for raw land, farmland or acreages in Alberta. The NHPI 'Land only' component shipped here is a builder's serviced-lot cost inside a new urban subdivision — it says nothing about a quarter-section outside Airdrie. Farm Credit Canada's Farmland Values Report is the only plausible route and its reuse terms have not been verified.

The falsification test

If this page is right, finished homes should be piling up. They are.

Everything above is an argument about timing, and arguments about timing are easy to make and hard to check. There is one series that could straightforwardly prove it wrong: unabsorbed inventoryUnabsorbed inventory. Newly completed dwellings that have not yet been sold or rented. Not homes under construction, and not resale listings — finished new units standing empty, counted at month end., the count of newly completed dwellings that have not yet found a buyer or a renter. If a record supply really is completing into softened demand, finished units should be accumulating. If they were not accumulating, this page would be a story about a lag that the market absorbed without difficulty — and it would have to say so.

They are accumulating. In June 2026, 3,051 newly completed dwellings stood unsold across Calgary and Edmonton — 1,321 and 1,730 respectively — against 2,201 a year earlier, a rise of 38.6%. On the three-month average the page charts, the pile is 186% above its five-year low of 1,137, reached in May 2022. Only 3 months in the whole series back to August 1988 have been higher than the current reading, and all of them are earlier this same year.

Shown as a three-month rolling averageThree-month rolling average. Each month replaced by the average of itself and the two before it. It trades a little timeliness for a lot less noise, which is necessary for a series as small and jumpy as unsold new dwellings., following the source file's own guidance: the raw monthly count is small and volatile enough that single months invite conclusions the data cannot carry. The pile is still modest in absolute terms — 1,240 newly completed units were absorbedAbsorption. A newly completed dwelling finding its first buyer or renter. Absorptions are the outflow from the unsold pile; completions are the inflow. The pile grows when completions run ahead of absorptions. in June 2026 alone, so this is a few months of sales, not a ghost town. The market is still clearing what it finishes, and it is taking longer to do it.

Calgary and Edmonton only. No Statistics Canada or CMHC table publishes this series for Alberta as a province. These two metropolitan areas hold roughly two-thirds of Alberta's people, which makes them a reasonable proxy and not a provincial total.

Reading it back

What we don't yet know

Show the work

Method, sources, and what this page can't do

1. 1.Resale and MLS® prices (unavailable) — There is no legally reusable resale or MLS® price series for Alberta. The two authoritative ones — CREA's MLS® HPI and the Teranet–National Bank HPI — both prohibit commercial use and derivative works. What is shipped instead is the New Housing Price Index, which covers NEW homes only, roughly a tenth of transactions. Neither this page nor any chart on it shows the price a reader would be quoted by a realtor.
2. 2.Raw land, farmland and residential acreages (unavailable) — No free, commercially licensed price series exists for raw land, farmland or acreages in Alberta. The NHPI 'Land only' component shipped here is a builder's serviced-lot cost inside a new urban subdivision — it says nothing about a quarter-section outside Airdrie. Farm Credit Canada's Farmland Values Report is the only plausible route and its reuse terms have not been verified.
3. 3.Commercial and office vacancy (unavailable) — No open commercial vacancy, lease-rate or absorption series exists for Alberta. Every source is a proprietary brokerage report, and they disagree with each other by around ten percentage points on downtown Calgary in the same quarter.
4. 4.Tenure of new apartment supply (partial) — The dwelling-type series in starts.json cannot distinguish rental from condo. starts-by-market.json closes most of this gap using table 34-10-0152, which carries the intended market at time of start for Alberta from 1988. What remains unmeasured is the secondary rental market — rented condos, basement suites and single-family rentals — which CMHC's rent and vacancy surveys exclude entirely and which is large and growing in Calgary.
5. 5.Unabsorbed inventory for Alberta as a province (partial) — No Statistics Canada or CMHC table publishes unabsorbed inventory for Alberta as a whole. inventory.json ships Calgary and Edmonton, which are current to the latest month, plus a seven-centre historical sum that ends in 2022 because the smaller-centre table stopped being published. Neither is a provincial total.
6. 6.Alberta-level completions and units under construction after 2022 (partial) — Every Statistics Canada / CMHC table carrying Alberta-level housing completions stopped publishing after 2022 — the annual table ends in 2022, the quarterly table in October 2022, the monthly tables in December 2022 — and Alberta under-construction stopped in 2002. Starts remain current to the latest month. Because the page's argument is about the LAG between starts and completions, this matters: recent completions are shown for the Calgary and Edmonton census metropolitan areas from table 34-10-0154, roughly two-thirds of the province's population, and must be labelled as two cities rather than as Alberta.
7. 7.Anywhere outside Calgary and Edmonton (partial) — Provincial Statistics Canada series cover all of Alberta, and CMHC's rent survey reaches about thirty Alberta communities, but vacancy, rents and price indexes are effectively a two-city story. A two-city finding should not be called 'Alberta'.

• ·CREA — MLS® Home Price Index and benchmark prices — Terms forbid any commercial use of the content, directly or indirectly, and forbid use of results, graphs or charts derived from it for resale or commercial distribution. terms
• ·Teranet–National Bank House Price Index — Licensed for non-commercial personal use only; derivative works prohibited. terms
• ·Local real estate boards (CREB, REALTORS® Association of Edmonton, AREA) — Republish CREA-family MLS® data under the same commercial restrictions.
• ·Commercial brokerages (CBRE, Avison Young, Colliers, Altus, Barclay Street) — Proprietary. Their downtown Calgary office vacancy figures also disagree with each other by roughly ten percentage points on the same quarter, so no single figure is 'the' number.
• ·Alberta Land Titles / SPIN2 — Pay-per-search under a tariff schedule, no bulk export, no redistribution rights.
• ·Alberta economic dashboard table BldgPermits_34100066 — Serves current data under the name of Statistics Canada table 34-10-0066, which has been inactive since October 2023. Citing that table for this data would be false. Building permits are excluded from this page entirely; starts are the cleaner measure. terms