Seven Holds: The Bank of Canada Shifts Its Risk to the Upside

On September 2, 2026, the Bank of Canada held its policy rate at 2.25%. That is the seventh consecutive hold. Through the first few, the live question was one of timing: when does the next cut land. That question has just changed shape, and the change is the real news in the announcement. For the political backdrop to this decision, see our piece on the last Bank of Canada decision before the vote.

What seven holds eventually tell you

A single hold says almost nothing. It signals that the central bank would rather watch. Seven in a row say something else. The last actual change to the rate was the cut of October 29, 2025. Since that date the rate has not moved in either direction, which turns 2.25% into a level the Bank is sitting at rather than a step on the way down.

The distinction matters for anyone carrying a mortgage. As long as the market read the run as a pause inside a cutting cycle, waiting had a logic to it: the next move would be favourable, patience would be rewarded. That logic rests on a premise that is no longer demonstrated.

The risk moved, and what that is not

The new element in this announcement is not the level, which is unchanged, but the way inflation risk is now described: tilted to the upside, with new tariffs named as a factor. It is worth being precise about that phrase, because it invites a misreading.

It is not an announcement of a rate increase. It is not even a forecast. It identifies the side on which a surprise is judged most likely. A central bank describing its risks as tilted upward is saying that between erring on the restrictive side and erring on the accommodative side, it now sees the second as the costlier mistake. The practical consequence is straightforward: the next move no longer has a presumed direction.

GDP up 3.3% in the second quarter

Gross domestic product rose 3.3% in the second quarter. That is not background colour, it bears directly on the reading. A central bank cuts when activity weakens and needs support. An economy growing at that pace does not present that profile.

Be careful not to make the number say more than it holds. It removes an argument from the case for a near-term cut. It does not supply an argument for a hike: solid growth alongside contained inflation would justify doing nothing at all for a long stretch. That is precisely the position the Bank occupies, and it is why what comes next cannot be deduced.

Two readings, neither settled

Both belong here, because both are circulating and neither has been confirmed by events.

First reading. Some forecasters now put a first increase on the table for 2027. The argument rests on the combination described above: an economy that is growing, inflation risk tilted upward, and a policy rate that is already low against that backdrop. On this view the question is no longer when the rate comes down, but whether it has to go up and starting when.

Second reading. A BMO economist takes the opposite position, judging that the Bank will not be inclined to raise anytime soon, with the trade war weighing on the housing recovery. The argument runs on a different axis: a trade conflict that slows activity and drags on residential construction and resale makes an increase hard to justify, whatever direction the stated inflation risk points in.

These two readings will not be settled by a blog post, and nothing in the data published so far separates them. We are not ranking them. What is useful here is to note that they exclude one another on the conclusion while each rests on real evidence.

Deciding while the debate stays open

The common practical error is to adopt whichever reading suits your own plan, then decide as though it were certain. A buyer in a hurry will happily take the second, an anxious seller the first, and each will have built a decision on half the available information.

The workable approach runs the other way. It asks whether a plan survives both cases. A payment that is affordable if the rate does not move for two years, and still affordable if it climbs at renewal, is a robust plan. One that only works under a single scenario is not cautious, it is a bet. That test runs on a real budget with real amounts, not on a borrowed forecast.

The next marker has a date

The Bank of Canada's next announcement is October 28, 2026. That is the next dated marker and there is nothing scheduled between the two. Waiting for it before deciding means accepting several weeks of stillness on a file that keeps moving anyway: a rate hold expires, a seller accepts another offer, a pre-approval runs out.

There is a useful asymmetry to notice. The October decision may confirm one of the two readings, or confirm neither and extend the run. In that last case, whoever was waiting ends up in the same place with a few weeks less runway. The central bank's calendar and a real estate file's calendar are not synchronised, and there is no reason to align the second to the first.