A Saturated Credit Channel: What a Rate Cut Can No Longer Restart

As September 2, 2026 approaches, one expectation keeps circulating: that the Bank of Canada eases and the market restarts. That expectation carries an assumption nobody states out loud, namely that the cost of credit is what is holding buyers back. The last five months let you test that assumption directly, and the result is inconvenient. On the nature of the shock the central bank is facing, see our piece on the tariff-driven supply shock; this one is about something else, whether the instrument can still produce transactions.

The channel, link by link

A rate decision never acts on a housing market directly. It travels a five-link chain. The central bank moves the cost of very short-term funding. Lenders pass that cost into their products. A borrower's monthly payment falls. Their borrowing capacity, computed off that payment, rises. And in the final link, that borrower, now solvent at a higher price, actually acts.

A chain transmits only as well as its weakest link, and this one has a property that is easy to forget: the last link is conditional. It assumes there exists a population of buyers sitting just under the qualifying threshold, for whom a few dollars of monthly payment decide between an offer made and an offer not made. If that population is thin, the first four links can work perfectly while nothing happens at the end.

The test five months just handed us

The right way to test an instrument is to watch the target variable while the instrument stays put. The policy rate was held at 2.25% at the July 15, 2026 announcement. Over the same stretch, the QPAREB release of August 6, 2026 reports 3,338 sales in July in the Montreal census metropolitan area, a fifth consecutive monthly decline and the steepest drop since February 2026.

Five months of continuous decline with the instrument frozen is a clean observation. It does not say the rate is unimportant. It says the move observed did not have the rate as its cause. A cause that did not vary cannot explain a variable that varied five times in a row. Whatever is pushing sales down sits elsewhere, and a decision that moves the rate will reach it only by accident.

What marginal effect actually means

The useful question is not whether a rate cut has an effect, but how many extra transactions a quarter point genuinely produces. That is its marginal effect, and the quantity is not constant: it depends entirely on how buyers are distributed around the qualifying threshold.

In a market where many households are declined by a hair, a quarter point flips a large number from the wrong side to the right side, and volume responds immediately. In a market where the buyers who are stuck are stuck on their down payment, or on a gap between price and income too wide to bridge, that same quarter point crosses no threshold at all. The instrument is identical, its yield is not. That is what saturation of a channel means: it still works mechanically, it simply no longer produces much.

Two constraints the rate does not move

The down payment is the first. It is built by saving and computed as a percentage of the purchase price. A rate cut adds not one dollar to it. It would even raise the amount required if it ended up supporting prices, since the percentage applies to a higher price. That is a case where the instrument works against the goal people assign to it.

The second is the gap between the target price and household income. Once that gap runs well beyond what a lighter monthly payment can close, easing arrives too late and too small. What these two constraints share is that neither appears in the central bank's release, which is why they vanish from the commentary on announcement day even though they decide the fate of most files.

The effect that does exist, and is not that one

Easing that brings in no new buyer is not without effect. It lightens the load on those already borrowing: renewals in progress, variable-rate loans, home equity lines of credit. That is an income effect for existing borrowers, real and measurable in a monthly budget.

But an income effect and a volume effect do not show up in the same data and do not call for the same decisions. Conflating them leads a seller to wait for a demand rebound that will not come, and a buyer to postpone on a bet that a wave of competitors is about to arrive, which will not happen either. Both effects flow from the same announcement, which is exactly what makes the confusion easy.

The test to run on your own file

It takes one operation, and it beats any forecast. Recompute your borrowing capacity with a quarter point less, then look at what that extra amount concretely changes about the list of properties within reach. If the answer is that no new property enters the range, the September 2 decision is moot for that file, whichever way it goes.

The test has the merit of replacing a collective expectation with a personal answer. It also moves the question to where it is actionable: not what the central bank will decide, but what is really blocking this particular file, and whether that blockage is of the kind credit can lift.