Canadian real estate.
A fragile recovery.

Canada’s housing recovery is uneven and vulnerable: less resale competition in parts of Ontario and British Columbia is meeting softer employment, higher fixed-rate funding costs and very different local supply pipelines.

Since the August edition, the story has become less comfortable for a simple recovery call. August hiring weakened, the national resale sales advance paused, U.S. policy tightened and Canadian bond yields rose. At the same time, revised population estimates changed an important part of the demand narrative. This report follows the established sequence from work and productivity through output, prices and financing, then tests the implications against local housing evidence.

The question is not whether every market rises or falls together. It is whether a particular household can carry a particular home, and whether the competing supply in that market is expanding or contracting. The charts, tables and local comparisons below are built to answer that more useful question.

Select a market or series, change the time window, inspect a value with a pointer or keyboard, or open the accessible data table. Each chart includes CSV and PNG exports and its own embeddable frame. The same observations feed the slideshow and report, so a published September chart stays a September chart.

August employment fell by 42,000 after July’s increase. The unemployment rate remained at 6.4%, but the employment and participation rates each slipped by 0.1 percentage point. Average hourly wages rose 2.0% from a year earlier. These are different signals: a lower number of people working, fewer people participating and slower growth in the pay attached to each hour. An unchanged unemployment rate cannot stand in for all three.

For housing, the mechanism is straightforward. Employment supports qualification; expected job security supports the willingness to take on a long commitment. A household can still qualify for a mortgage while choosing to wait. That is why I watch sales and listings alongside jobs, rather than assuming every labour-market change must immediately show up as a price change. One noisy monthly survey is not proof of a recession, but August makes a confident, nationwide rebound harder to argue.

Local employment matters more than the national label. An energy employer, an export manufacturer, a university and a federal department expose their surrounding housing markets to different risks. The provincial chart is a first screen. For an individual client, the next questions are their employer, occupation, second household income and cash buffer. A province with relatively low unemployment can still contain a vulnerable neighbourhood or a weak property segment.

Statistics Canada’s business survey puts reported AI use at 19.2% in the second quarter of 2026, compared with 6.1% two years earlier. Separately, business-sector labour productivity increased 1.0% in Q2 as output rose and hours worked edged down. It would be a mistake to join those two charts and label the productivity gain an AI effect. They measure different populations, use different methods and do not establish that causal link.

The relevant housing channels run in both directions. Lower costs for research, administration and coordination can improve business margins and expand output. Substitution can put pressure on particular tasks and occupations. Investment in computing, power and buildings can add to demand before the productivity benefits arrive. The Bank of Canada’s September discussion of AI treats this as a supply-and-demand question, which is the right frame for the market call.

For a real estate business, the useful test is whether a repeatable workflow produces a correct, reviewable result at a lower total cost. A sourced monthly report is one example: collect the releases, reconcile the numbers, render the charts and adapt the explanation for a local audience. That is why this edition exposes its chart data and remix guide. Readers can use the same evidence in Homies, ChatGPT or Claude, with the capabilities of their chosen tools determining which finished assets they can produce. HomieBench provides separate context on AI workflow evaluation; its published outlook should not be confused with measured housing-market evidence.

June real GDP by industry increased 0.3%, with service-producing activity growing while goods production slipped. The separate expenditure accounts showed real GDP rising 0.8% in the second quarter, equivalent to a 3.3% annualized pace. Annualized growth is the hypothetical result if one quarter’s rate continued for four quarters; it is not the economy’s observed year-over-year growth rate.

The July advance estimate was flat. The official July industry GDP release is scheduled for September 29, the recording date and one day after this edition’s September 28 cut-off. This report keeps the advance estimate clearly labelled and does not insert an unreleased official figure. The first update to watch before presenting is therefore GDP, including revisions to earlier months.

The composition of growth matters to real estate. Residential construction affects trades and the future supply pipeline. Resale transactions generate commissions and related spending but do not constitute newly built dwellings. Manufacturing and resource output support different regional incomes. Stronger aggregate GDP can coexist with weak pre-construction sales or a strained household sector. The sector chart lets readers inspect those differences rather than treating one national growth number as a forecast for every property.

There is also a vintage issue in per-person comparisons. The August quarterly GDP release preceded Statistics Canada’s September population revisions. Its published per-capita calculation and the newly revised population series should not be combined silently. This edition preserves the original release context and uses the revised population data in the demographic chapter.

August CPI was 3.0% above a year earlier. Gasoline rose 22.8%, while CPI-median was 2.0% and CPI-trim 1.9%. Rent CPI increased 2.8%; mortgage interest costs were slightly lower than a year earlier. That combination does not describe a uniform acceleration across the economy. The component charts show where the pressure sits and how different shelter costs are moving.

The practical transmission runs through the household budget. Money spent on fuel, food or utilities cannot also fund a down payment or a larger mortgage payment. A supply-driven price increase can therefore raise measured inflation while reducing the amount of other spending a household can sustain. Whether that becomes persistent inflation depends on wider pricing behaviour, wages, expectations and demand. A single energy print does not settle the rate outlook.

Rent needs particularly careful interpretation. CPI measures rents paid across the occupied stock. A landlord advertising a vacant unit faces a different market. Asking rents can soften while many sitting tenants still pay more than a year earlier. Mortgage-interest CPI is different again: it measures the cost of interest across the mortgage stock, not the rate offered to a buyer today. These distinctions keep a statistically correct chart from becoming a misleading client conversation.

The Bank of Canada left its policy rate at 2.25% on September 2. By September 25, the five-year Government of Canada benchmark yield was 3.65%, 51 basis points above July 15. The two- and ten-year yields were 3.35% and 3.94%. These are daily market observations, not mortgage offers. They show why a household waiting for an unchanged policy rate to deliver a cheaper fixed mortgage may be disappointed.

Variable-rate borrowing is closely connected to lenders’ prime rates. Fixed mortgage pricing also depends on term funding, competition, credit and operating costs. Government bond yields are a useful reference, but there is no fixed one-for-one pass-through from a five-year yield to every mortgage product. A lender’s actual quote, conditions and expiry still govern the household decision.

The Federal Reserve raised its target range by 25 basis points to 3.75–4.00% on September 16, while the Bank of Canada had held earlier in the month. That is an observed policy divergence, not a forecast of how long it will last. It adds context for cross-border funding and currency pressure without implying that the Bank of Canada must mechanically follow the Fed.

The Canadian dollar averaged US$0.7058 on September 28. Currency weakness can raise the Canadian-dollar cost of imported goods and equipment. U.S. rates, global term premiums, energy prices and trade conditions can all influence Canadian financing, even when domestic employment would argue for relief. Exporters may benefit from a weaker currency while import-dependent firms and consumers pay more. That is one reason the same macro shock can help one local economy and squeeze another.

My practical reading is to separate rate speculation from payment resilience. Buyers need to know whether the property works at a confirmed financing rate, with room for renewals and ordinary ownership costs. Sellers need to understand what today’s payment buys, rather than relying only on the price achieved when borrowing was cheaper. The next Bank of Canada decision is October 28; its outcome is not assumed in this report.

The household debt-service ratio eased to 14.52% in the second quarter because disposable income grew faster than required debt payments. Payments themselves still increased. This is a share of aggregate income, so it says little about a household whose employment income fell, whose renewal is approaching or whose revolving debt is expensive. Lower aggregate stress and acute individual stress can coexist.

Consumer insolvencies totalled 145,681 in the twelve months ending July, up 5.4% from the previous twelve months. Proposals and bankruptcies are distinct legal outcomes. Neither is a count of mortgage foreclosures, forced listings or homeowners; housing tenure and the reason for filing are not established by this chart. The rolling annual view helps keep one seasonal monthly move in perspective.

The Bank of Canada’s May Financial Stability Report estimated that pandemic five-year fixed-payment loans representing about 12% of outstanding mortgages would renew over the following year, with an average payment increase around 15%. Its refinancing scenarios also identified greater vulnerability in Toronto than nationally. Those are dated model estimates, not September delinquency observations. The underlying mechanism is still relevant: falling values can reduce the equity available to refinance just when a household would like more flexibility.

For the market outlook, I would become more concerned if weakening employment were followed by a persistent increase in motivated listings and longer selling times. A financial buffer can absorb a payment shock; a sustained loss of income is harder to bridge. This edition excludes an outdated February arrears feed from the current dashboard instead of relabelling it as a September reading.

Statistics Canada’s September 23 release estimated 41,798,407 people on July 1, up 0.5% over the year and 0.2% over the quarter. The release revised earlier population estimates using updated treatment of non-permanent residents. That is a material change for a market report: earlier-vintage claims about nationwide contraction should not be carried into this edition as though the data had not changed.

Temporary-resident numbers still fell. The July stock was 2,779,774, 5.3% below a year earlier. In the second quarter, immigration of 99,148 exceeded the net decline in non-permanent residents. It is entirely possible for one component to fall while the total population rises. Stocks, quarterly flows and annual growth rates are kept separate in the charts.

Migration also redistributes demand within Canada. Alberta gained 5,893 people through interprovincial migration in Q2, while Ontario lost 7,063. These are net moves, not the gross number arriving or leaving. For housing, the next step is to compare those flows with completions, vacancy and the type of accommodation people need. Provincial growth alone does not tell us which municipality, price range or tenure absorbs it.

Population is a demand input rather than a complete housing forecast. Household formation, age, income, students, roommate arrangements and the existing dwelling stock all matter. A slower inflow can change leasing conditions quickly in a student or newcomer submarket, while a family-oriented ownership market follows a different timetable. The revised series should change the evidence, not invite a new oversimplification.

CREA reported that seasonally adjusted home sales fell 0.7% in August. Actual sales were 6.9% below August 2025. The national average sale price was $668,219, up 0.6% year over year, while the MLS® Home Price Index was flat month over month on a seasonally adjusted basis and down 3.0% year over year on an actual basis. The average reflects the mix of homes sold; the HPI is designed to track benchmark price changes. Neither should be substituted for the other.

The national seasonally adjusted sales-to-new-listings ratio was 49.1%, down from 51.1% in July, and months of inventory stood at 4.8. Those figures suggest room for negotiation nationally, but they do not establish the balance in every board area. The local explorer deliberately uses actual monthly ratios and labels them accordingly. Seasonal local readings should be compared with their own history, not assigned a buyer-or-seller label mechanically from a national threshold.

My base case is an uneven stabilization. Fewer competing listings can help prices stop falling even before sales return to a strong level. Some local markets have fewer new listings than a year ago, which can support prices; nationally, seasonally adjusted new listings rose 3.3% in August. A broader recovery also needs improving incomes and confident buyers. The evidence that would strengthen the outlook is repeated sales improvement with stable employment and no renewed inventory build. One better price average is not enough.

Ontario’s average price was $788,835 in the August aggregate snapshot, down 1.7% year over year. Sales fell 6.0%, but active listings also fell 3.9%. British Columbia’s average was effectively flat at $926,375; sales fell 4.7% and inventory declined 5.2%. Less resale supply helps explain why a weak demand story does not translate automatically into accelerating price declines. It still leaves a large affordability gap for many households.

Within BC, the Greater Vancouver and Fraser Valley board areas carried roughly 7.9 and 8.9 months of actual inventory. Victoria sales increased 12.9% year over year. Those markets should not receive a single script. The GVR release also publishes its own Metro Vancouver sales and benchmark measures, which differ from the dashboard’s aggregate-vintage figures. The table below keeps one consistent CREA snapshot; it does not blend an official benchmark with an aggregate average.

Alberta’s average price rose 3.8% to $519,845 even as sales fell 11.5% and listings accumulated. That is a cooling activity story alongside a firmer average price. The Calgary board region and City of Calgary are different geographies. CREB’s city release showed a $569,800 total residential benchmark, about 1% below a year earlier, with apartment benchmarks down roughly 8%. A rising board-area average cannot establish that every Calgary property type appreciated.

Saskatchewan combined a 2.6% average-price gain with lower inventory. Saskatoon and Regina each had about 2.4 months of actual inventory, a different negotiating environment from the high-inventory BC boards. Manitoba moved in another direction: average prices fell 2.0%, while active listings rose 16.9%. Winnipeg’s inventory increase was 19.0%. The Prairies share some affordability advantages, but their current changes are not identical.

Quebec’s average price rose 3.6% to $572,652 while sales fell 7.3% and active listings rose 18.8%. A price gain can coexist with a loosening market. This data package supports Quebec’s provincial resale aggregate; the local-board catalogue does not provide comparable Montréal and Québec City resale series. Separate QPAREB releases provide current local evidence: Montréal CMA sales were 2,853, down 13%, with active listings up 18%; Québec City sales rose 6% to 677 while inventory rose 27%. Those are official CMA releases, separate from the CREA aggregate table. Construction and rental charts add further metropolitan detail.

Atlantic Canada is also diverging. Nova Scotia’s average price was broadly flat while inventory rose 13.6%. New Brunswick’s average fell 1.0%; Prince Edward Island’s rose 5.0%. Newfoundland and Labrador recorded a 4.0% average-price gain and a 17.2% inventory decline. Halifax-Dartmouth, Moncton, Saint John and Fredericton show different combinations of sales and listings. Saint John, New Brunswick, must not be confused with St. John’s, Newfoundland and Labrador.

The table also covers smaller Ontario and BC boards, rural markets, Yukon and Northwest Territories. Small transaction counts can produce large percentage moves or unstable averages. There is no comparable Nunavut resale series in this catalogue. For a local client report, use the nearest defensible geography and explain the boundary, rather than promising unsupported detail. The most useful comparison is often a property segment’s inventory, realistic financing and competing homes, not a national league table.

TRREB recorded 5,057 August sales and 12,075 new listings. Its current release reports sales down 2.1% and new listings down 14.1% year over year. The average price was $993,410, the median $850,000 and the composite HPI benchmark $925,900. These are three different descriptions of the market, not three competing estimates of what a particular property is worth.

A seller may face fewer competing listings and still need patience. Average property days on market reached 51, compared with 35 listing days. The property measure follows a broader listing history, making it useful when relisting has reset the visible clock. The distinction matters in a client conversation about how long a sale really took.

Detached, semi-detached, townhouse and condominium markets have different price points and buyer pools. Even within a type, an average can change because more expensive or cheaper properties happen to sell. The municipality and home-type charts provide a second screen; an actual pricing recommendation still needs comparable properties, condition, carrying costs and competing inventory.

Condominium purchasers and investors should separate purchase price from the ongoing economics: condominium fees, property taxes, insurance, maintenance, financing, vacancy and a realistic achievable rent. For pre-construction, add closing timing and the risk that an appraisal differs from the contract price. A market can become cheaper without becoming cash-flow positive, and a cheaper resale unit can compete directly with a newly completed unit seeking a tenant.

CMHC reported an August all-area starts pace of 229,046 at a seasonally adjusted annual rate, broadly unchanged from July. The six-month trend fell 1.3% to 244,149. A housing start records the beginning of construction. An under-construction count is the stock of unfinished homes. A completion brings a unit closer to occupancy. Those measures can move in opposite directions for perfectly ordinary reasons: a large cohort can finish while fewer replacement projects begin. Calling all three ‘supply’ without their timing loses the central investment question.

The provincial starts chart uses seasonally adjusted annual rates. A pace of 50,000 does not mean 50,000 homes broke ground in that month, and it is not a forecast for the year. The city charts instead use trailing-year starts and completions, alongside the current construction stock, to reduce some of the noise from large multi-unit projects. Their CMA and CA boundaries are different from real-estate board areas.

Tenure is equally important. Purpose-built rental, condominium and homeowner starts do not compete in exactly the same way. A condominium can enter the rental market through an investor, but its financing and pre-sale process differs from a rental building held by one owner. The tenure series in this edition covers centres with at least 50,000 people; its dashboard key should not be mistaken for all-area national coverage.

For developers, resale prices, achievable rents, construction costs, development charges, infrastructure, approval timing and financing all meet in the project’s feasibility calculation. Public incentives may help a project clear that threshold, but an approval is not a completed home and a funded program is not proof of local absorption. For buyers, a weak start today can mean less competing supply later. For landlords, today’s completions can pressure lease-up even if the next wave has already slowed.

The rental charts deliberately keep annual paid-rent and vacancy measures separate from quarterly asking-rent evidence. The 2025 vacancy series covers apartment buildings with six or more units; average paid two-bedroom rents cover row and apartment structures with three or more units. Their survey universes differ. The advertised-rent index ends in the fourth quarter of 2025 and was published in June 2026. Neither is relabelled as a September 2026 rent quote.

This distinction matters when conditions turn. Existing tenants can still face annual increases while owners of vacant units offer incentives or accept lower asking rents. A newly completed building can add competition for another new building without immediately changing the average paid rent across the entire city. Slower temporary-resident inflows can affect some rental segments more than ownership markets, while household formation and interprovincial arrivals support others.

An investor should therefore underwrite an achievable local lease, including incentives, vacancy and turnover, then compare that income with the full cost of ownership. A published city average is context, not a promise about a particular suite. For a renter deciding whether to buy, compare the after-tax cash commitment, flexibility, transaction costs and time horizon rather than treating rent as inherently wasted or ownership as a guaranteed investment return.

Ontario’s enhanced new housing rebate is a material competitive factor for eligible new homes. Together with the existing provincial rebate, it provides up to $80,000 against the provincial HST component; the provincial affordability payment may provide up to $50,000 more, with offsets for federal housing rebates. Eligibility, contract dates, price and occupancy conditions matter. This is not an automatic $130,000 discount on every advertised new home.

The buyer-side implication is to compare the actual after-rebate contract economics with a genuinely comparable resale property, including closing costs and delivery timing. The seller-side implication is that a competing new project can change the reference price in a local market. The owner-occupier rebate described here has its own occupancy rules. A separate enhanced residential-rental-property rebate provides another pathway with its own eligibility conditions; investors need to use that framework. The linked CRA rules should be checked for the transaction’s particular structure.

Credit policy also needs accurate language. OSFI’s loan-to-income framework limits the share of new uninsured mortgages above 4.5 times income within a lender’s portfolio. It is not a universal ban on an individual borrower exceeding that multiple. OSFI has also clarified that its capital guidance did not prohibit lenders from considering rental income under the existing underwriting framework. A lender’s assessment of income, debt service, property and risk remains central.

Trade, energy and public policy are important because of these concrete transmission paths: employer income, imported building costs, financing, infrastructure and household confidence. I would avoid turning an exposure map into a prediction of layoffs or a proposed housing measure into completed supply. Watch what becomes effective, who qualifies and whether the resulting projects or purchases actually happen.

For buyers, the task is to test the payment and competing choices at a realistic financing rate. For sellers, it is to price against what a buyer can purchase now, including new-home alternatives where relevant. For investors, it is to stress-test rent, vacancy and total carrying costs. For Realtors, it is to translate verified local evidence into a clear recommendation without turning a national statistic into a property valuation.

My conclusion is conditional. Canada has enough regional variation for both opportunity and risk to be present at once. A more persuasive recovery needs durable incomes, workable payments and a supply balance that supports the transaction. The next releases will tell us which part of that combination is improving.

Copy the starter at the top and paste it into Homies, ChatGPT or Claude. Your assistant starts with your market, uses saved context where available and asks for missing details one at a time. The guide below carries the sourcing, chart and publishing instructions so the starting experience stays simple.