Week one: the data said the opposite
We spent August warning that a 50% tariff wall would break Canada. The wall went up on Aug 22. Then Canada printed 3.3% annualized GDP growth, the TSX finished August at record highs, and the country's third straight quarter of population decline turned into the first real per-person income gain in years. This is the worst scorecard we have published, and we are going to walk through every line of it. But the more important finding is where the risk actually went: to a Federal Reserve that, as of Friday, is more likely to hike than cut in September. For the first time since we started this map, American recession odds now exceed Canadian ones.
The scorecard
On Aug 22 we published three paths and a pre-committed odds rule. We said the market had spent Friday pricing a deal and would have to reprice on Monday. We said Canada's export economy was walking into a wall. Here is what actually happened, graded without spin.
| What we said | What happened | Grade |
|---|---|---|
| Monday gaps down — Friday closed pricing a deal (TSX +250) that no longer exists | The TSX rose on Monday and kept rising. It closed Aug 25 at 36,958 (+243), hit record highs during the month, and finished August +3.5% at 36,554. The S&P 500 ended the week up 0.5%. | Miss |
| Tariffs stick and grind the economy down (45% path) | Tariffs did stick, and Trump escalated on Aug 24 with a threat of 50% on all Canadian autos and steel from Jan 1, 2027. But Q2 growth was already banked, and the damage window is Q3–Q4, not this print. | Too early |
| The loonie is the tell — watch it, not the equity index | Correct, and it was the only market that behaved as predicted. CAD slipped to 71.9¢ (1.3900) while the TSX rallied. The currency priced the trade war; the index priced bank earnings and gold. | Hit |
| Don't trust the +75K LFS jobs surge — SEPH will be the tiebreaker | SEPH June came in at +4,800 — flat. Manufacturing −7,200, food service −6,100, retail −3,900, professional services −1,600. The single largest gain was public administration +10,600. Employer records show a private sector that has stopped hiring. | Hit |
| Canada's recession odds belong above America's | Reversed this week. Canada grew 3.3% with no contraction quarter on the books; the US has flat real consumer spending, 3.3% core PCE, and a Fed leaning toward a hike. | Miss |
Why 3.3% is weaker than it looks
The number is real. The composition is a warning.
Exports rose 3.6% — the biggest jump since Q1 2023 — led by passenger cars and light trucks at +27.0%. That is not a healthy export boom. That is every automaker in Ontario shipping everything they could across the border before the Aug 19 deadline. Front-running is borrowed growth: the vehicles that crossed in June were going to cross in September, and now they won't.
Inventories subtracted 1.3 percentage points, with businesses drawing down $17.0 billion of stock. The buffer that absorbs a tariff shock has already been spent. Q3 has to be produced, not un-warehoused.
Per-capita GDP rose 1.0% — but read the reason. Statistics Canada notes the population declined for a third consecutive quarter. Output per person improves when the denominator shrinks. That is arithmetic, not prosperity, and it is the same denominator effect that has been flattering the unemployment rate all year.
The household saving rate is 3.7%, and housing investment rebounded 2.5% after two down quarters. Households are spending into a tariff shock with a thin cushion.
So: a strong quarter, built substantially on a one-time pull-forward, with the inventory buffer gone and the payback due in the quarter that is running right now. It is genuinely good news that Canada entered the trade war from a position of growth rather than contraction. It does not tell us much about how it exits.
The jobs verdict
This one deserves a direct answer, because it was the question that started this whole series: should we have believed the +75K jobs print?
Partly no. SEPH is the employer-payroll record — actual remittances, not a household survey — and for June it shows payroll employment up 4,800 on 20 million jobs. That is zero. Underneath it, the goods and consumer-facing economy is shedding: manufacturing −7,200, accommodation and food −6,100, retail −3,900. The one sector doing real hiring was public administration.
Job vacancies sit at 509,100 for a vacancy rate of 2.8% — unchanged from May and from a year ago. Employers are not competing for workers.
The honest caveat: SEPH June and LFS July are different months, so this is not a clean grading of that specific print. But the pattern across both surveys is consistent and it is the pattern that matters — headline employment holds up while the composition rots, propped by self-employment in the household survey and by government payrolls in the employer records. Skepticism was warranted. It was warranted for compositional reasons, not because the statisticians were wrong.
The bigger story: the Fed turned
On Friday at Jackson Hole, Fed Chair Kevin Warsh said: "Inflation is running above our 2% target. So the Fed's predominant focus right now should be on prices."
Market-implied odds of a September rate hike went from roughly one-in-three to better than a coin flip. Bond markets began, in the phrasing of the wire reports, building bets that the Fed may raise rates to get inflation under control. The Russell 2000 — the index most exposed to domestic borrowing costs — fell 1.4% on the day and 1.5% on the week while the S&P rose.
He has the data for it. July PCE ran 3.7% headline, 3.3% core, both up 0.2% on the month. And the part that should worry you: real consumer spending was flat — up less than 0.1% after inflation, down from 0.4% in June — while the saving rate climbed to 3.0%. Income rose 0.4%; people banked it instead of spending it.
That combination is the classic pre-recession setup: prices high enough to force a central bank to tighten, demand already soft enough that tightening lands on a weakening consumer. A Fed that cuts into a slowdown is a cushion. A Fed that hikes into one is an accelerant. This is why the US number moves up.
Odds, moved
Canada — recession within 12 months
Cut. You cannot hold a 45–50% call through a quarter that grew 3.3% with no contraction on the board, a housing rebound, and record equity prices. The tariff damage is still coming and the pull-forward still has to be paid back — that is why this is 35–40% and not 25%. But the starting position is stronger than we assumed.
United States — recession within 12 months
Raised. Core PCE at 3.3%, real spending flat, a savings rate climbing on caution, and a Fed chair who just told the market that prices come first. The path where the Fed hikes in September into an already-stalling consumer is now the single largest recession risk in North America.
What we are watching
Our rule for next time
We got the mechanisms right and the timing wrong, twice in a row. The correction we are making, published in advance so you can hold us to it:
1. A structural risk is not a market call. A tariff that will damage an economy over four quarters says nothing about what an equity index does on the following Monday. We will stop implying it does.
2. Sector mix beats narrative. The TSX is roughly half financials and materials. Canadian bank earnings and record gold prices were always going to matter more to that index than an auto tariff. We knew the composition and did not use it.
3. Odds move on realised data, not on headlines. We moved Canada to 45–50% on the day the tariffs went live, before a single data point measured the damage. From now on, event-driven odds changes are capped at half the move until hard data confirms — full move on the print, not the press conference.