Thursday 24 September 2026

Canadian fixed mortgage rates rise again before the 28 October Bank of Canada decision, with the 5-year Canada yield moving back toward the 3.65% it reached in mid-September. The Bank itself holds at 2.25%.

Why we think so. Wednesday was the worst day of the year for government bonds. The US 30-year Treasury bond yieldBond yield The annual return an investor gets for lending to a government or company through a bond. Yields rise when bond prices fall. Fixed mortgage rates are priced from the 5-year Government of Canada yield. rose to 5.39%, a level last seen in July 2004, and the 10-year closed at 5.11%, up from 4.96%. Every Treasury maturity beyond two years now yields more than 5%. The trigger was strength, not weakness: S&P Global’s flash PMI put US composite activity at 58.4, its highest in 62 months, with manufacturing at 57.0 and services at 58.7, and input costs rising at the fastest rate since October 2022. A selloff driven by growth and inflation data does not reverse in a week. Canada was pulled along, with the 10-year rising to about 3.86% and the 5-year, which sets fixed mortgage pricing, closing Tuesday at 3.54%, up from 3.35% a month earlier. Lenders fund fixed mortgages from the bond market: when the 5-year rose about a quarter point in the second week of September, several banks raised fixed rates by 10 to 20 basis points within days.

What the market is pricing. Forecasters expected US manufacturing at 53.6 and services at 55.8, so both beat by a wide margin. Markets now price better than even odds of further quarter-point Fed increases in October and December, on top of the move to 3.75% to 4.00% on 16 September. In Canada, the consensus is a hold at 2.25%, and we agree. The Bank cannot cut into a global bond rout and 3% inflation, and has no reason to hike when tariffs are cutting growth. The overnight rate has not moved since March.

What it means. Anyone renewing or closing a mortgage in the next 120 days can ask for a rate hold today rather than next week; it is free and it caps the damage if the next move is up. The gap between variable and fixed has narrowed, and a variable rate only wins if the Bank cuts, which is not on the table this autumn. A GIC saver gets the other side of the trade: higher yields mean better rates on new one to five-year terms, so a maturing GIC is worth comparing with the old ladder. The gap between a 2.25% Bank of Canada and a Fed near 4% keeps the dollar weak at 70.94 US cents. A household can check how much of its debt reprices in the next two years, and whether the budget still works at a renewal rate half a point higher.

What would change our mind. The US 10-year back below 4.8% or the Canada 5-year below 3.4% within two weeks, which would mean the market judged Wednesday’s PMI a one-off; or a September CPI print in mid-October with CPI ex-gasolineCPI ex-gasoline The Consumer Price Index with gasoline removed. When it sits well below the headline rate, the inflation is coming from fuel rather than from the broad economy. clearly falling, which would reopen the case for a cut. Either and we would say so.

What to watch. US initial jobless claims today, expected at 201,000; US new home sales for August, expected at 615,000 annualised; today’s Fed speakers, Barkin, Hammack and Paulson; Canada’s September CPI in mid-October; and the Bank of Canada decision and Monetary Policy Report on 28 October.

Wednesday’s selloff is a repricing, not the start of a bear market, because it is happening on strong growth data and profits are still expected to rise sharply this quarter. But the portfolio mix that made sense at 1% yields is not the mix that makes sense at 4% to 5%.

Why we think so. The S&P/TSX Composite fell 584.18 points, or about 1.6%, to 35,751.43, its biggest one-day drop in weeks. In New York the S&P 500 fell 0.8% to 7,706.03, the Nasdaq 1.1% and the Russell 2000 1.8%, a third straight decline for the Dow. The reason was the bond market. A 10-year Treasury at 5.11% and a 30-year at 5.39% give investors a competitor for equities that did not exist a year ago. When a safe bond pays more, the price investors will pay for future company profits falls, and the effect is largest for assets that pay little or no income. Gold, which pays no yield, fell US$58.00 to US$4,318.40 an ounce and was the first casualty; the Toronto index, heavy in miners, was the second. Energy was the one sector to rise, with November crude up US$1.64 to US$92.16. Bitcoin fell 2.6% over 24 hours to US$84,076 and Ethereum 2.8% to US$2,674, moving with risk assets rather than against them.

The bigger picture. For most of the past fifteen years, bonds paid so little that holding them felt like a cost. That pushed many portfolios toward equities by default. With a safe government bond yielding 5% in the United States and about 3.9% in Canada, the comparison is honest again: a stock has to earn its place. That does not make stocks a bad investment. It makes the balance between growth and income a real decision.

What it means. For most Canadians the TSX is their pension, their RRSP and their balanced fund, and materials and energy are about a third of the index, so a bad day for gold can outweigh a good day for oil. The practical step is to look at the fixed-income share of a portfolio and at what it holds; a bond fund bought in 2021 has a very different yield today than its price history suggests. Rebalancing inside a TFSA or RRSP costs nothing in tax, and a portfolio that has drifted heavy in equities and gold after two strong years has a natural trim available. In a taxable account, the capital gains cost needs weighing against the concentration. The core question is whether the written asset mix was set for near-zero rates.

What would change our mind. The selling spreading to corporate credit, with spreads widening sharply, or the October earnings season showing revenue falling short of the capital being spent on AI. Either would make this a valuation correction rather than a rate adjustment.

What to watch. US jobless claims and new home sales today; today’s Fed speakers; the October earnings season; and the Bank of Canada decision on 28 October.

Volkswagen’s St. Thomas battery plant is delayed, not cancelled. But a 2029 start means the jobs and local property demand arrive a full business cycle later than planned, and anyone who priced them into a decision should take them out again.

Why we think so. PowerCo, Volkswagen’s battery subsidiary, said production will now begin in 2029 rather than 2027, citing market demand, battery technology and the group’s long-term plan, and framed the change as getting the pacing right. The evidence that the project is still alive is concrete: the 370-acre site is in place, EllisDon has just been named general contractor, and the subsidy structure is unchanged. Companies that intend to walk away rarely hire a general contractor first. The evidence that demand is the issue is equally clear: the delay follows slower electric vehicle sales in North America, and the plant is designed for 90 gigawatt-hours a year, enough for up to a million vehicles. A plant that large only makes sense if demand grows into it.

The bigger picture. The public money is large: Volkswagen’s own investment is $7 billion, Ontario committed $500 million in incentives and Ottawa up to $13.2 billion in production subsidies and tax credits, most of it payable only when cells are made. That design protects taxpayers in a delay, since little is paid out now, but it also pushes the promised return further out. About 3,000 jobs were promised at full operation, with roughly 1,300 construction workers at peak; about 60 are on site now. Two years is a long time for a town, an apprentice or a supplier that borrowed to expand on a 2027 start.

What it means. For a business owner in southwestern Ontario, the lesson is wider than batteries: a customer’s announced date is not a contract, and firms that expanded on the announcement rather than the purchase order are the ones carrying the cost. A household or small business whose income depends on one large project should treat announced dates as the optimistic case and keep enough liquidity, such as a cash buffer or an unused line of credit, to ride out a two-year slip. A tradesperson choosing where to train or relocate can weigh a project with shovels in the ground and a contractor named against one with only a press release. Homeowners and investors who expected local property demand from the plant should push that expectation out. The underlying question is concentration: how much of a family’s income, property value or business revenue rests on one employer or one project.

What would change our mind. A second slip of the date or a change to the plant’s scale in the next year, which would mean the market case has weakened rather than the timing; or a Volkswagen order book that fills the plant before it opens, which would make the delay purely technical.

What to watch. Construction progress under EllisDon and the size of the workforce on site against the 1,300 peak; any revision to the plant’s timeline or capacity; and electric vehicle sales in North America.