Canada's Numberless CPI Story and the Hollow Alchemy of Crypto's Liquidity Narrative

Last Tuesday a headline crossed my feed from a crypto outlet: "Canada inflation in September, core pressures ease." I read it twice looking for the number. There wasn't one. No CPI print, no month-over-month momentum, no trimmed mean, no weighted median, no survey date, no Bank of Canada quote. A crypto publication had published a story about Canadian consumer prices that contained neither a price nor a Canadian official. Within an hour, two Telegram groups I lurk in had converted that absence into a thesis: global easing confirmed, rotate into duration. That conversion — an adjective becoming a position — is the only event here worth analyzing. Not the inflation. The transmutation. Aggregators have been eating editorial for years, but the 2025–26 search environment finished the job. Traffic collapsed the moment AI summaries started answering queries above the fold. What survived was keyword capture: small teams republishing macro wires under a crypto masthead, because "inflation" and "rate cut" are searched by people who also hold tokens. The template is modular. Any central bank plus any disinflation equals liquidity expansion equals risk assets. Swap the nouns, ship the post. I know that template intimately. In 2020 I ran three simultaneous Substacks on Aave, Curve and Synthetix, and the reason I could was reusable narrative shells — the same six paragraphs, refilled with different protocol names. It worked until it didn't; volume without structure is a treadmill. The lesson from that burnout year was that a template is honest only when the variable inside it carries real weight. When the variable is empty, the template becomes a machine for manufacturing conviction. I saw the same mechanic in 2017, decoding 42 ICO whitepapers for the Buenos Aires Crypto Circle. Half of them described technology nobody had built; the ones that raised fastest described technology nobody had built with better adjectives. The market did not price code. It priced cadence. That is what happened here. The word "Canada" does no analytical work in that headline. Neither does "September." The load-bearing word is "ease," and "ease" has no referent — not versus last month, not versus consensus, not versus the Bank of Canada's own forecast. An unfalsifiable claim republished at scale is not information. It is liturgy. Let me put a number on the transmission chain, because the chain is where the narrative breaks. The Bank of Canada is a small-open-economy central bank. Its policy rate moves the loonie, the front end of the Government of Canada curve, and — through import prices — Canadian CPI. That is a closed loop. Its spillover into global dollar liquidity is close to nil: Canada is a rounding error in reserve allocations, its bonds sit outside the major duration benchmarks, and a 25bp cut moves global risk appetite by roughly nothing that survives a weekend. When the Fed cuts, the dollar funding channel moves. When the ECB moves, euro-area credit moves. When the BoC moves, the loonie moves and then moves back, because a weaker CAD raises import prices and pushes CPI higher, capping how far the Bank can ease. The currency is shock absorber and feedback brake in the same mechanism. So the trade implied by that headline — Canadian disinflation to global easing to a crypto bid — is not a weak chain. It is the wrong chain, with the arrow pointing backwards. Crypto is not downstream of a Canadian CPI print. It isn't downstream of anything in it. Now the one word in the piece that does carry information: divergence. The body used it, then concluded "cautious." Divergence between what and what? At least four readings exist, and they point in opposite directions. Headline up, core down: energy or food lifted the top line while underlying momentum cooled. Dovish — the Bank can keep cutting. Goods down, services up: global prices and the currency did the work while wages and shelter held the floor. Hawkish, because services inflation mirrors domestic demand and expectations. Shelter up, everything else down: pure policy noise. Mortgage interest cost in the Canadian CPI is a lagged echo of hikes already delivered, not a reading on current policy. The correct response is to look through it while saying "cautious" aloud, because saying anything else unanchors expectations. CPI-trim up while CPI-median falls: an internal contradiction in the statistical apparatus itself, in which case "cautious" is a technical refusal of noisy data, not a policy signal at all. One more measurement problem, and it is the one the Bank actually cares about. Its preferred core gauges are CPI-trim and CPI-median, not ex-food-and-energy — the version aggregators reach for because it is easier to explain. When those two disagree, "ease" is an artifact, and the language hardens accordingly. Expectations sit on top of that: if core cools while the business outlook survey still shows firms pricing at three percent, the cut case is dead regardless of the print. Four paths, four different signs. The story published none of them, because it published no numbers. There is a structural wrinkle almost nobody outside Canada prices correctly. Canadian mortgages are dominated by five-year fixed terms, which means the shelter component of CPI is a slow-motion replay of rate decisions made years earlier — a renewal wall. Households roll from two-percent paper into five-percent paper on a schedule set long before this month's print. Two things can therefore be true simultaneously, and usually are: core pressures ease in the statistic while the actual debt-service burden on households rises. I rebuilt this decomposition in 2023 for a Toronto client, mapping the renewal calendar against the shelter sub-index, and the output was uncomfortable. The CPI was reporting relief from a policy stance that, in cash-flow terms, was still tightening. The statistic lagged reality by roughly the length of a mortgage term. Any macro read that treats Canadian shelter inflation as a contemporaneous signal is reading a photograph of a room that has already been rearranged. And then the loop closes. Cut rates and credit expands; credit expands and house prices firm; house prices firm and shelter CPI re-accelerates; shelter re-accelerates and the next cut gets harder. Canada's monetary policy has an endogenous brake installed in its largest CPI weight. "Cautious" isn't caution in that frame. It is a description of the machine. The contrarian read is not that the story was too dovish. It is that it was too dovish and too hawkish at once, and got away with it. The headline sold relief; the body sold restraint. Both shipped under one byline with no data to adjudicate between them, and every reader selected whichever half matched a position they already held. That is not journalism failing. That is journalism working exactly as designed for engagement. Here is the harder version, the one I would defend. The most plausible reason a central bank says "cautious" after a soft core print isn't measurement noise — it is a supply shock. Tariff regimes push in both directions at once: they lift import prices and suppress export demand. That is stagflationary, and the standard reaction function under a supply shock is to refuse to accelerate easing on a single month of data, because one tool cannot address both legs. If that hypothesis holds, the real stance is more hawkish than "cautious," and the story understated it by a full notch. The piece attributed a policy conclusion entirely to internal CPI composition and never named the largest exogenous variable in the country. That is not an omission. That is an attribution error dressed as analysis. Alchemy fails when the intent is hollow. Watch two numbers at the next release, not the headline: CPI-trim and CPI-median. If both print at or below 0.2% month-over-month, the easing slope is real and the divergence was benign. If they disagree with each other, then "divergence" was doing all the work in that story and none of the readers knew it. And keep a private tally this quarter: how many of your liquidity signals were actual data, and how many were republished adjectives? The ratio is usually worse than you think, and it is always worse in a bear market — when the temptation to alchemize a headline into a position runs highest.