2026-08-17 → 2026-08-20

July inflation is the last consumer-price read before the Bank of Canada's 2 September decision, and the headline is expected to swing from -0.4% to +0.4% on the month. Both legs of that swing are mostly gasoline: June's decline was the same oil-price unwind that dragged the US headline negative, and July's rebound is the base washing out.

The backdrop is a central bank that has earned its stillness. The Bank has held the overnight rate at 2.25% since its cuts ended last autumn, June headline inflation decelerated to 2.8% from 3.2%, and the preferred core measures sit at the bottom of the 1-3% control range's midpoint. A sixth consecutive hold on 2 September is close to fully priced.

The monthly headline will make the noise, but it is the least informative number in the release. A +0.4% print that is all gasoline changes nothing for a committee that looks through energy; the risk scenario is a firm month with energy flat, which would say underlying momentum is rebuilding just as the Bank has stopped watching for it.

For CAD, the asymmetry is quiet. An in-line report leaves the currency trading US rate pricing rather than its own, as it has most of the summer. Only a genuine core surprise, in either direction, gives the loonie a domestic story this week.

Median CPI is one half of the pair the Bank of Canada actually steers by, and consensus has it ticking up to 2.0% from 1.9%, dead on the midpoint of the control range.

A central bank with core inflation at target and a 2.25% policy rate is the closest thing the G10 has to a finished cycle. That is precisely why the print matters: the hold-forever pricing only survives while these two series behave.

At 2.2% or above, the September hold stays safe but the market starts asking whether the next move discussion has the wrong sign; Canada would be re-joining the sticky-inflation world its peers never left. At 1.9% or lower, the Bank's position gets more comfortable still, and the front of the Canadian curve has nothing to reprice.

The trimmed mean is expected unchanged at 1.8%, which would put it below the 2% midpoint for a second month and make it the softer half of the core pair.

Read the two measures together, not separately. Median at 2.0% with trim at 1.8% is a distribution whose centre sits at target with the tails quiet; that combination is what lets the Bank of Canada hold at 2.25% while the Fed, the Bank of England and the RBA all still contain members arguing for hikes.

The divergence is the tradeable fact. Canada is the control group of this cycle: if its core measures stay pinned near 2% while everyone else's stick above target, the case that restrictive-for-longer elsewhere is the right policy, not an overreaction, gets stronger every month.

Consensus expects claims to rise 11.2K after 6.7K, which would extend a run of gently rising claimant numbers: a labour market loosening at the edges rather than breaking.

The timing gives it more weight than usual. The Monetary Policy Committee held Bank Rate at 3.75% in July on a 6-3 vote with the three dissenters wanting a hike, and last week's GDP surprise, +0.3% in June against a consensus contraction, handed the hawks fresh cover. Rising claims are the counterweight: tightening into a softening labour market is a harder argument than tightening into a growing economy.

On its own this print moves little; it lands twenty-four hours before the number that does. But a claims figure north of 20K would colour how Wednesday's CPI gets traded, because a hawkish surprise in inflation is a different proposition when the jobs data underneath is deteriorating.

For sterling, the asymmetry mirrors the gilt market's: bad labour news gets amplified this week, good labour news gets shrugged, because the market's live question is whether the BoE hikes into the autumn, not whether it cuts.

The week's most consequential European release. Consensus expects headline inflation to jump to 2.9% from 2.6%, a three-tenths acceleration that would be the Bank of England's own forecast arriving on schedule: the July Monetary Policy Report projected inflation rising through the second half as energy costs feed through.

That forecast is exactly why three MPC members voted to raise Bank Rate to 4% in July. Their argument was pre-emption: hike before the energy re-acceleration hardens expectations. A print at 2.9% validates the worry without settling the argument; a print at 3.0% or above starts moving votes, and the September meeting becomes genuinely live.

The economy is no longer the hawks' obstacle. Q2 GDP came in at 0.4% with June alone at +0.3% against a consensus contraction, so the activity excuse for patience has thinned. What is left of the doves' case is the labour market, which is why Tuesday's claimant count reads as a prelude to this print.

For gilts the stakes are mechanical. A 10-year near 4.96% already charges more than 120bp over Bank Rate for inflation and fiscal risk; an upside CPI surprise presses directly on that premium, and the long end has shown all summer that it does not rally properly on good news. Sterling takes the same trade with more velocity.

The composition matters as much as the headline. An energy-driven 2.9% with services inflation cooling is the tolerable version the majority can look through. Services re-accelerating alongside energy is the version that makes the 6-3 vote arithmetic unstable.

Minutes of the 28-29 July meeting: the fifth consecutive hold at 3.50-3.75%, and the one that drew three dissents in favour of a hike. The vote count is known; what the minutes reveal is how broad the sympathy for that dissent ran among the non-voters, and that breadth is the single most important thing in the document.

Read them with their sell-by date in mind. This meeting happened before the -23K payroll print and before July's tame CPI, so the committee described in these pages was looking at a hotter economy than the one that exists now. A hawkish tone is partly stale by construction; the market knows it, and the reaction function should discount accordingly.

Two passages deserve the close read. First, the discussion of the inflation framing: the July statement leaned on supply shocks, energy included, to explain elevated inflation, and the minutes will show whether that framing was consensus or contested. Second, any conditions members attached to a September move, because those conditions can now be checked against data that has since arrived and mostly come in soft.

The market walks in pricing a September hike at roughly 42%, down sharply after CPI week. Minutes that show the hike camp extending beyond the three dissenters would push that pricing back up and reverse some of last week's front-end rally; minutes that show a comfortable majority behind the hold confirm the drift toward no-move-in-September.

The larger frame is the following week. Jackson Hole runs 27-29 August and Kevin Warsh delivers his first keynote there as chair, three weeks before the 16 September decision. These minutes are the last official record of the committee's thinking before that speech; whatever tension they reveal, Warsh gets the stage to resolve it.

Consensus expects +11.4K jobs after the prior month's +76.3K blowout, a deceleration that would read as normalisation rather than deterioration: the Australian survey is volatile, and giant months are routinely followed by flat ones.

The context makes this the first real test of last week's RBA hold. The board kept the cash rate at 4.35% on 11 August but disclosed that it discussed a hike, and Governor Bullock declined to rule out further tightening, saying inflation remains too high and that a period of subdued growth will be needed to bring it down sustainably. A central bank talking that way is one strong data run from moving again.

A print above roughly +40K, on top of last month's 76.3K, is exactly that data run: it would say the labour market is re-tightening while inflation sits above target, and the market would start pricing the September-quarter meetings as live. AUD firms and the front of the curve sells off.

A soft or negative print delivers the subdued growth the board says it wants, validates the pause, and lets the market keep treating 4.35% as the peak. The kiwi-style risk, tightening bias colliding with a visibly weakening labour market, is not Australia's problem yet; this release decides whether it starts to be.

Expected unchanged at 4.4%. The rate has been the stable anchor of the Australian labour data while the monthly employment counts swing, and stability at 4.4% is consistent with a market that is tight but no longer tightening.

The RBA's framing gives the number its edge. A board that wants subdued growth to bring inflation down needs the unemployment rate to drift up, gently, over time. A fall to 4.3% or below, alongside a strong employment count, argues the economy is absorbing 4.35% without slowing, and strengthens the case the board already debated: that the cash rate is not restrictive enough.

As ever with this release, check participation before trading the headline. A rate that falls because the labour force shrank is a different economy from one where hiring outran a growing workforce, and the two argue for opposite policy conclusions.