Canada’s Mortgage Renewal Cliff: Five-Year Terms, Variable Rates and the Big Banks

Canada’s “mortgage renewal cliff” is the stress test of a system that deliberately shares interest-rate risk between borrowers and banks, and by October 2026 it has been absorbed better than almost anyone predicted in 2023. Where a US homeowner can lock in a rate for 30 years, a Canadian typically fixes for five, and a large minority float. That design kept Canadian banks safe through the 2008 crisis but left households exposed when the Bank of Canada lifted its policy rate from 0.25% to 5% between March 2022 and July 2023. The years 2024 to 2026 were when the loans written at the bottom of the cycle came due. This article explains how Canadian mortgage structure works, who was exposed to the renewal wave, how the Bank of Canada’s cuts, OSFI’s stress test and the new loan-to-income limits shaped the outcome, what the arrears data show, and why the final leg of the cliff in 2027 still matters for the Big Six banks. It is part of the Canadian Company Stories hub.

How is a Canadian mortgage different from an American one?

A Canadian mortgage separates the amortisation period, usually 25 years, from the term, usually five years or less, so the interest rate is renegotiated several times over the life of the loan. In the United States the 30-year fixed-rate mortgage, made possible by Fannie Mae and Freddie Mac securitisation, lets a borrower hold one rate for three decades and refinance only when rates fall.

The Canadian structure exists because banks fund mortgages largely from deposits and short-dated bonds, and because the Bank Act historically limited prepayment protection to five years, so lenders are unwilling to lock in long rates. The consequence is that interest-rate risk sits with the household, not the bank. When rates rise, Canadian monthly payments rise at the next renewal; when they fall, borrowers benefit quickly. American households were insulated from the 2022–23 rate shock; Canadian households were not.

The mix of products matters. CMHC’s spring 2026 Residential Mortgage Industry Report found that in February 2026, 42% of new mortgages at chartered banks were variable-rate, 35% were fixed for three to five years, 11% were fixed for five years or more and the rest were shorter. During the pandemic, five-year fixed loans at under 2% dominated; by 2023, as borrowers bet on cuts, shorter terms and variables took over. Variable-rate loans themselves come in two forms: adjustable-payment loans where the monthly amount moves with the prime rate, and fixed-payment variables where the payment stays constant but the amortisation stretches, sometimes to the point where the payment no longer covers interest.

Scale is the other difference. Outstanding residential mortgage debt passed $2.4 trillion in December 2025, up 4.8% year over year, against a household debt-to-disposable-income ratio of 173.3 in the fourth quarter of 2025. The Big Six banks hold 75.1% of outstanding mortgages but originated only 54.8% of new loans in the third quarter of 2025, a share that has fallen about seven points in a year as credit unions, monoline lenders and mortgage investment entities took more of the flow.

Who took out the pandemic mortgages that are now renewing?

The renewal wave consists of borrowers who bought or refinanced between mid-2020 and early 2022 at five-year fixed rates below 2%, plus variable-rate borrowers from 2021–22 whose payments reset as prime rose. The Bank of Canada’s May 2025 Financial Stability Report estimated that about 60% of all outstanding mortgages would renew during 2025 and 2026, and that around 60% of those renewing would face higher payments, most of them five-year fixed loans taken out or renewed during the pandemic.

The typical case is a household that borrowed $500,000 in early 2021 at 1.8%, paying roughly $2,070 a month, and renewed in early 2026 at around 4.2%, pushing the payment above $2,600. The Bank’s 2026 report put the average payment increase for the final cohort of fixed-payment pandemic mortgages, about 12% of all mortgages, at roughly 15%. Rates fell from about 4.8% in January 2025 to about 4.2% in January 2026, which took the edge off, and the number of borrowers renewing in 2026 is 13% lower than in 2025, so the peak dollar value of renewals has already passed.

The second cohort is variable-rate and short-term fixed borrowers who took loans after the 2022–23 increases, about 14% of outstanding mortgages. For them, renewal in 2026–27 is expected to bring no change or a decrease in payments on average, because they are rolling off rates of 5% to 6% into rates of 4% to 4.5%. This is the group that effectively pre-paid the cliff.

The third and most stressed cohort is small but concentrated: buyers in 2022–23 in Toronto who stretched at the peak and now hold loans larger than their homes are worth. The Bank of Canada estimated these borrowers at about 2% of outstanding balances, showing the most acute stress, and the condo investors who had to close on pre-construction units into higher rates and lower appraisals are the extreme version.

What did the Bank of Canada do to soften the landing?

The Bank of Canada cut its policy rate from 5% in May 2024 to 2.25% by October 2025, with seven consecutive cuts in 2024 and early 2025 before pausing at 2.75% from March to September 2025 and then cutting twice more. It has held at 2.25% at every decision since, including 2 September 2026, with the next announcements due on 28 October and 9 December 2026.

The cuts transmitted quickly because of the mortgage structure. Variable-rate borrowers saw payments fall within weeks, and the five-year fixed rate, which tracks the five-year Government of Canada bond, fell from above 5% at the 2023 peak to the low 4% range by 2025. In October 2026 the best five-year fixed rates available through brokers were around 4.3% to 4.7%, with the lowest Big Six rate at about 4.7% and variable rates near 3.3% to 3.4%. The gap between the 2021 rate and the 2026 rate therefore closed by roughly a third before most renewals arrived.

The Bank’s September 2026 statement was cautious. GDP grew 3.3% annualised in the second quarter after a weak first quarter, unemployment edged down to 6.4%, headline inflation hovered near 3% on gasoline prices while core measures stayed close to 2%. But the collapse of Canada–US trade talks and new tariffs on both sides, together with high energy prices from the Middle East conflict, pushed up inflation risks, and the Bank said it was prepared to adjust policy in either direction. Its October 2026 business and consumer surveys showed sentiment weakening and inflation expectations rising.

The policy rate matters less to the cliff than the bond market does. Five-year Government of Canada yields around 3.1% in mid-2026 set the floor under fixed mortgage rates regardless of what the Bank does with the overnight rate, and the Bank’s own May 2026 report noted that nearly all borrowers facing large payment increases would have renewed by the second half of 2027. The last leg of the cliff is therefore priced off a bond yield that has barely moved in a year.

How does the stress test work and did it do its job?

The stress test requires uninsured borrowers at federally regulated lenders to qualify at the greater of their contract rate plus 2 percentage points or 5.25%, under OSFI’s Guideline B-20, with the Department of Finance applying the same minimum qualifying rate to insured mortgages. Introduced for insured loans in 2016 and extended to uninsured loans in January 2018, the test was designed precisely for the scenario that arrived in 2022–23: a sharp rise in rates between origination and renewal.

The evidence suggests it worked. The Bank of Canada’s 2025 report found that more than 90% of five-year fixed borrowers would face payment increases smaller than the stress test had required them to absorb, and its 2026 report found that more than 90% of borrowers who renewed in the previous 12 months did so at rates below their qualifying rate. About 85% of renewing households could cover their expected payment increase for 12 months or more from existing financial assets. Average loan-to-value across the mortgage stock fell from 74% at the end of 2019 to 71% at the end of 2024 as prices rose, and HELOC utilisation fell from 35% to 28%.

The test has also been loosened at the margin. Since November 2024 borrowers switching lenders at renewal with no change to the loan amount or amortisation no longer have to re-qualify, which restored competition among lenders for renewal business and is one reason the Big Six’s share of originations has slipped. Borrowers who stay with their existing lender have never had to re-qualify, which removes the risk that a household could be refused a renewal outright.

Its limits are the borrowers who were never subject to it. Mortgage investment entities, which hold about 1.3% of outstanding mortgages and 4.5% of originations, lend at an average rate of 9.6% to borrowers with high debt-service ratios; their stage-3 impairment rate was 5.5% and foreclosure rate 2.9% in the third quarter of 2025, with exposure concentrated in Ontario (48.8%) and British Columbia (37.3%). That is where the stress has gone, not to the banks.

What are OSFI’s loan-to-income limits and why were they introduced?

OSFI’s loan-to-income (LTI) limits, applied from the first quarter of 2025, cap the share of each federally regulated bank’s new uninsured mortgages with a loan-to-income ratio above 450% at 15% of quarterly originations. The measure targets the tail of highly leveraged borrowers rather than the average, and it applies at the institution level, so a bank can still lend to a stretched borrower but must ration how many such loans it writes.

The regulator’s concern was that the stress test alone did not stop borrowers from taking loans equal to five or six times their income when rates were near zero. In 2021 roughly a quarter of new uninsured mortgages at some lenders exceeded 450% LTI, and those borrowers were disproportionately in Toronto and Vancouver. The Bank of Canada’s 2026 report found that borrowers with large balances relative to income account for about 17% of outstanding mortgage balances and that their arrears are rising, which is exactly the population the LTI limit is designed to shrink over time.

The limit has had little visible effect so far because the market did it first. With prices about 20% below the 2022 peak and incomes up 16% between 2021 and 2025, far fewer new loans exceed 450% LTI, and most banks report being well inside the 15% ceiling. The binding constraint will come in the next upswing, when the cap prevents the kind of leverage build-up seen in 2020–22.

OSFI has paired the LTI limit with other changes. The 2024 B-20 revisions, the removal of the stress test for straight switches, and the federal government’s December 2024 package raising the insured price cap to $1.5 million and allowing 30-year amortisations for first-time buyers and new builds together form a system that loosens access for first-time buyers while tightening the tail. Peter Routledge, OSFI’s superintendent, has framed this as a trade: more flexibility on terms in exchange for hard limits on leverage.

What do the arrears and provisioning data actually show?

Arrears have risen from record lows but remain far below historical norms: CMHC’s 90-day delinquency rate was 0.24% in the fourth quarter of 2025, up from 0.21% a year earlier, with chartered banks at 0.25%. The Bank of Canada’s broader measure, the share of borrowers more than 60 days late on any account, was about 1.3% for mortgage holders and 2.5% for households without a mortgage, both broadly stable through the year to May 2026.

The banks’ own numbers show the pressure in consumer credit rather than mortgages. Royal Bank of Canada reported third-quarter 2026 net income of $6.0 billion, up 11%, with total provisions for credit losses of $1.0 billion (36 basis points of loans), of which $979 million was on impaired loans. TD Bank Group reported net income of $4.6 billion and total provisions of $917 million, with its Canadian personal and commercial segment’s impaired provisions up 19% on consumer credit migration. TD’s chief risk officer Ajai Bambawale described delinquencies as stable overall but slightly higher in real estate secured lending, and the bank’s RESL book grew 4% year over year.

Regional concentration is where the risk lives. The Bank of Canada estimated that about 4% of mortgages renewing in 2027 nationally would be unable to refinance on standard criteria (loan-to-value above 80%, gross debt service above 39% or total debt service above 44% after exhausting a 25-year amortisation), rising to about 9% in the Toronto area; a further 10% fall in prices would lift those figures to 7% and 12%. Condo loan-loss allowances at lenders rose 39% year over year in 2025, and CMHC’s condo risk work singles out Toronto and Vancouver.

Borrowers have adapted through amortisation. About 10% of 2022 borrowers had refinanced by mid-2026, and roughly 70% of those extended their amortisation by an average of six years. Fixed-payment variable mortgages that went into negative amortisation in 2023 have largely been restored to schedule as rates fell. The system is absorbing the shock by lengthening loans, not by foreclosing, which keeps arrears low at the cost of slower household deleveraging.

How are the Big Six banks managing the renewal book?

The banks have treated renewals as a retention battle rather than a credit problem, offering existing borrowers rates at or near the best available to stop them switching now that straight switches no longer require re-qualification. TD’s Canadian personal banking head Sona Mehta told analysts in August 2026 that the bank had recorded record proprietary originations while holding pricing discipline, and RBC’s personal banking volumes grew 2% year over year.

The economics are favourable to the banks even if volumes are flat. A loan written at 1.8% in 2021 renewing at 4.2% in 2026 roughly doubles the interest income on the same balance, and because deposit costs have fallen with the policy rate, net interest margins on mortgages have widened. That is why the Big Six reported record or near-record profits through 2025 and 2026 despite rising provisions, and why mortgage lending remains the core of Canadian retail banking rather than a loss leader.

The competitive threat is in origination. The Big Six’s share of new mortgages fell nearly seven points in a year to 54.8% as brokers steered first-time buyers to monolines and credit unions, and as insured 30-year amortisations and the $1.5 million cap widened the pool of CMHC-insured borrowers whose loans any lender can fund cheaply through Canada Mortgage Bonds. The banks keep 75% of the stock because renewals default to the incumbent, but the flow is more contested than at any time since the financial crisis.

Credit quality is being managed through amortisation and forbearance. Banks have allowed borrowers to extend amortisations at renewal, restructured negatively amortising variable loans and, in Toronto condo cases, worked with developers and buyers on closing shortfalls rather than enforcing. Because almost all high-ratio loans are CMHC-insured and the banks’ uninsured books have average loan-to-values around 70%, the capital at risk in a genuine downturn sits with the federal insurer and the household, not the lender, which is the design of the Canadian system.

What would turn the slope back into a cliff?

The scenario that converts a manageable renewal wave into a credit event is a rise in unemployment coinciding with the final 2027 cohort and a further fall in Toronto and Vancouver prices. The Bank of Canada’s May 2026 report identified job losses from trade uncertainty as the key risk to households, and the breakdown of Canada–US trade talks in the autumn of 2026 made that risk concrete.

The arithmetic is unforgiving in Toronto. With home prices about 20% below peak and condo prices 24% below, borrowers from 2022–23 have little or no equity; the Bank’s estimate that 9% of Toronto-area renewals in 2027 could not refinance on standard criteria would become 12% on a further 10% price fall. Those borrowers cannot sell without a loss and cannot refinance to lower payments, so a job loss leaves no option but arrears. Mortgage investment entities, with 5.5% of loans impaired already, would be the first to show it.

Inflation is the other threat. The Bank of Canada’s September 2026 statement flagged rising upside risks from tariffs and energy prices, and its October surveys showed inflation expectations climbing. If the Bank were forced to hold or raise rather than cut, five-year fixed rates would stay above 4% through the 2027 renewals and the expected 15% payment rise for the last pandemic cohort would not shrink. A rate increase, which markets did not price in October 2026, would reverse the relief that variable-rate borrowers have enjoyed.

Against that, the structural cushions are substantial: $1.73 of debt per dollar of disposable income is below the 2022 peak, average loan-to-value is around 70%, 85% of renewers have a year of payment cover, and the stress test means nearly all borrowers were underwritten at rates above what they now face. Canada’s renewal cliff was the most anticipated household credit event of the decade, and anticipation, in the form of regulation, provisioning and borrower behaviour, is why it has so far been a slope.

What can founders and investors learn from Canada’s mortgage renewal cliff?

The first lesson is that risk you can see coming is risk you can manage. The renewal wave was identified in 2022, quantified by the Bank of Canada every year since, and regulated in advance through the stress test and loan-to-income limits; borrowers saved, banks provisioned and lenders adjusted terms. The lesson for any business facing a known refinancing or repricing date is to treat it as a project with a timeline, not an event to survive.

The second is that where interest-rate risk sits determines who adapts. Because Canadian households bear the rate risk, they changed behaviour fast, shifting to variable and short terms in 2023 and extending amortisations in 2025–26, while the banks’ profits rose. Founders negotiating debt should know which side of that allocation they are on and price it accordingly.

The third is that regulation shapes market share. The removal of the stress test for straight switches, the higher insured cap and the 30-year amortisation changed the competitive balance between the Big Six and smaller lenders within a year, cutting the big banks’ origination share by seven points. The fintech challengers building around open banking are counting on the same dynamic in deposits and payments.

The fourth is that averages conceal the tail. National arrears of 0.24% describe a healthy system; 9% of Toronto renewals that cannot refinance and 5.5% impairment at private lenders describe a stressed one. Investors analysing Canadian consumer credit, real estate or the banks themselves should build their view from the segments, especially the rental and condo markets of Ontario and British Columbia, rather than from the national figure. The same discipline applies in other markets with short-term mortgage structures such as Australia and the United Kingdom, which readers can compare in the Australian Company Stories hub.

Frequently Asked Questions