Canada ran the experiment everyone is arguing about

Every argument in markets right now reduces to one question: does an oil shock become inflation, or does it just become an expensive month? On Monday Canada published a data set that answers it about as cleanly as real economies ever do.

Canadian headline inflation accelerated to 3.0% year over year in July, up from 2.8% in June. That puts it at the ceiling of the Bank of Canada's 1% to 3% control range. On the month, prices rose 0.5%, or 0.3% seasonally adjusted (Statistics Canada).

Gasoline rose 25.7% year over year. That single line is doing essentially all of the work.

Strip it out and inflation was 2.2%, for the third consecutive month. Not falling, not rising. Flat, three months running, while the headline number climbed (Statistics Canada).

The core measures the central bank actually watches barely moved. CPI-median printed 2.0% and CPI-trim 1.9%, averaging 2.0%, which is the target (RBC Economics). Elsewhere in the basket, groceries slowed to 3.1% and shelter ran at 1.3%.

Economists read it as a non-event for policy. The consensus response was that the rebound is not enough to sway the Bank of Canada, whose next decision comes on 2 September (CTV News).

Why this matters well beyond Canada

The whole hawkish case in the developed world rests on a specific claim: that a supply shock does not stay in the category it hits. The argument is that expensive fuel raises freight, raises input costs, raises wage demands, and eventually shows up in the price of things that have nothing to do with oil. If that is happening, headline inflation is a leading indicator and central banks are behind. If it is not happening, headline inflation is a tax on households that will drop out of the annual comparison in twelve months, and reacting to it is a policy error.

In Canada in July, it did not happen. Gasoline up 25.7%, everything else steady at 2.2% for three straight months, core at target.

That is one country and one month, and it is not proof. Canada is smaller, more energy-producing, and has a different housing and wage structure than the US. But it is measured evidence on a question that is currently being answered mostly with assertion, and it arrived on the same day the US bond market took the opposite view: the 30-year Treasury yield rose to 5.31%, the highest since June 2007, with persistent inflation and government borrowing named as the drivers (CNBC), while Brent hit $90 a barrel and September Fed hike odds fell to roughly 30% to 31% (Kitco).

So on Monday the long end of the US curve priced more inflation risk from oil at the same moment the nearest available real-world test suggested oil was not passing through.

So what for an allocator

This is the difference between a headline number and a decision-useful number. Under ABC, most portfolio decisions that get made on inflation are really decisions about the Cash sleeve and about duration. A 3.0% headline print and a 2.0% core print imply different things about how long rates stay where they are, and the gap between them is entirely one line item. Anyone sizing a portfolio off the headline is sizing off gasoline.

The habit worth building is to always ask what the number looks like without its biggest mover. Headline inflation, index returns, revenue growth, a portfolio's own performance: in each case the aggregate hides whether the result is broad or whether one component is carrying it. Canada published both versions. Most data does not, and you have to go looking.

Three dated tests are coming and they are all falsifiable. UK CPI on Wednesday, Japanese CPI on Friday, and US core PCE on 26 August. If two or more show the same shape, energy in the headline but not in the core, then the long end's inflation worry is being priced at exactly the moment the data starts arguing against it. If core measures start climbing alongside the headline instead, the hawks were early rather than wrong, and it is the front end that is mispriced. There is a real chance of either, which is why the useful thing this week is knowing in advance which evidence would change your mind.

One takeaway: an inflation number is an average, and averages hide the only thing you need to know, which is whether the price rise is spreading or standing still.

Analytics & education, not advice. DYOR.