Commodity Shocks vs Demand Inflation: One Price or All Prices
A commodity shock is one price against all others — a relative-price change in inflation's clothes. Demand inflation is all prices against money. Opposite policy, opposite positioning.
Markets talk about inflation as if it were one substance that comes in different amounts. It is closer to two different diseases with one shared symptom. An inflation print generated by a commodity shock and an identical print generated by excess demand describe economies in opposite conditions, facing opposite policy needs, offering opposite reads for an equity portfolio. The sophisticated error is not missing the inflation — it is treating the diagnosis as a detail when the diagnosis is the entire content of the signal. Size tells you how much the symptom hurts; source tells you everything else.
The framing that does the work: one price versus all prices. A commodity shock is one price — energy, food, freight — moving against all others. It is a relative-price change wearing inflation's clothes: a scarcity somewhere in the world repricing a specific input, transferring income from the users of that commodity to its producers, and taxing real household income in the process. Nothing about the economy's total spending has changed; purchasing power has been redistributed and partially destroyed. Demand inflation is all prices moving against money: nominal spending outrunning the economy's capacity to supply, so that scarcity is everywhere at once and no individual price contains the information — the price level is the event. The first is a problem inside the price system. The second is a problem with the price system's unit of account, and only the institution that controls the unit of account can end it.
The two diagnostics: breadth and the pair
The diseases are distinguishable in real time by two observable signatures. The first is breadth. A relative-price shock is concentrated by definition: the affected categories and their immediate cost-chain neighbours inflate while the rest of the basket behaves. Demand inflation is broad by definition — when the problem is total spending against total capacity, trimmed-mean and median measures rise with the headline, and the share of categories inflating above target widens across the basket. The second signature is the pair: what inflation and real activity are doing together. A supply shock pushes them in opposite directions — prices up, real income and output down — the stagflationary pair, because the shock is a tax. Demand inflation pushes them the same way — prices up, activity hot, labour markets tight — because the same excess spending drives both. When you see the pair moving together, you are looking at a demand problem whatever the commodity charts say; when you see it split, the inflation is mostly a tax collector, not a spiral.
How each disease reaches an equity portfolio: three layers of why
The unglamorous complication is that pure cases are rare and the dangerous cases are hybrids. A commodity shock landing on an economy that is also running hot does not stay a relative-price event: the energy tax feeds costs economy-wide while excess demand lets firms pass those costs through, and the wage round — negotiated against a headline rate the shock inflated — converts the one-price event into all-prices momentum. This is the pass-through channel, and it is the bridge between this framework and its siblings: whether a supply shock stays a supply shock is decided by the wage machinery and the expectations anchor, not by the commodity market. The practical monitor is the same in every cycle — watch whether the shock's inflation is migrating from the affected categories into broad services pricing and wage growth. While it stays home, it is a tax. When it travels, it has changed species.
The illustration: the same shock, three different economies
Oil provides the controlled experiment, because the same commodity has hit economies in different states — stated here as established history. The oil shocks of the 1970s landed on economies with spreading wage indexation and a slipping expectations anchor. The one-price event travelled immediately: wages indexed to the inflated headline, prices repriced off the wage round, and a relative-price change became the engine of a general inflation that outlived the shock by years. The species change was not caused by the oil; it was caused by the machinery the oil landed on.
The oil price collapse of the mid-2010s ran the experiment in reverse. A large downward one-price move dragged headline inflation toward zero across much of the developed world, and by the symmetric logic of the naive reading, deflation psychology should have followed. It did not: anchored expectations looked straight through the move in both directions, core measures barely blinked, and the episode passed through the data on the base-effect timetable — a demonstration that with the anchor intact, even a violent commodity move stays a relative-price event. The post-pandemic episode then supplied the hybrid case: an energy shock stacked on top of genuinely hot demand and tight labour markets. Both diseases were present at once, which is precisely why the era's central debate — transitory or persistent — was so hard to settle: each camp was correctly describing one component and wrongly generalising it to the whole. The decomposition, not the debate, was the answer.
The strongest case against the distinction
The serious objection has two prongs. First, identification: every real inflation is a mixture, the decomposition is a modelling exercise with wide error bars, and by the time breadth and wage data have confirmed the diagnosis, the tradable moment has passed — so the framework offers clarity in hindsight and ambiguity in real time. Second, propagation: the "one price" label understates commodities like energy, which is not one price among many but an input to every price in the economy. A large energy move changes costs everywhere at once — transport, materials, utilities, food — which makes it functionally broad, and arguably a demand-relevant event for policy regardless of its supply-side origin.
Both prongs sharpen the framework rather than break it. On identification: the claim is not that the decomposition is precise, but that the pair and breadth are observable immediately — inflation arriving with weakening real income and concentrated category moves reads supply from the first print, and the error bars on that read are far narrower than the market's habit of reading every print as the same disease. The framework's edge does not require perfect attribution; it requires refusing the default of no attribution. On propagation: energy's input role is real, but it produces a level shift in the cost structure — a one-time repricing that flows through and stops — unless the wage and expectations machinery converts it into a spiral. Functionally broad is not the same as dynamically persistent, and the difference is exactly the pass-through monitor described above. The objection is right that an energy shock deserves respect; the framework specifies the precise conditions under which respect should escalate into repricing the policy path — and, by extension, when a central bank facing a fragile anchor may rationally tighten into a supply shock, buying insurance against the species change rather than fighting the shock itself.
What this changes for a portfolio
The consequence is a decision order: diagnose, then route. A supply-driven print routes to the income statement — map the portfolio's commodity-user margin exposure against its producer exposure, mark down consumer-volume sensitivity for the real-income tax, and leave the duration view substantially alone if the anchor evidence is clean; demand destruction from a severe shock can even push the growth-risk side of the ledger. A demand-driven print routes to the discount rate — the policy path is now engaged, and long-duration equity carries the incidence regardless of sector. The hybrid routes to both, weighted by the pass-through evidence as it accumulates. Sterling's Embedded Intelligence runs this triage on every inflation release: pair first, breadth second, pass-through monitors third, and only then a view — because the print's size, which is what the headlines carry, appears nowhere in the decision tree. Sterling's shorthand for the whole piece: a supply shock taxes the numerator; demand inflation attacks the denominator — and a portfolio that confuses a tax with an attack will hedge the wrong line of the valuation.
How to apply this framework
- Run the pair test before any other read. Inflation rising with hot activity and tight labour markets is a demand problem whatever the commodity backdrop; inflation rising with weakening real income is mostly a tax. The co-movement is observable from the first print.
- Let breadth arbitrate the mixture. Trimmed-mean and median measures rising with headline indicate the all-prices disease; a headline spike the trimmed measures ignore is concentrated, and concentrated means relative-price until the pass-through data says otherwise.
- Monitor migration, not the commodity. The shock's future is decided in services pricing and the wage round, not on the commodity exchange. Inflation staying in the affected categories is a tax being paid; inflation travelling into wages and broad services has changed species — and the policy path with it.
- Route supply shocks to the income statement, demand inflation to the discount rate. Margin incidence and consumer volumes for the one-price event; duration and the terminal rate for the all-prices event. The recurring market error — de-rating duration on an anchored supply shock — is the mispricing this framework is built to catch.