Consensus Wants a 1% CPI Month That Never Touches Core
The March inflation estimates embed the widest headline-versus-core wedge in the recent sequence — and the same consensus expects consumer sentiment to barely move. Both cannot be right.
The March inflation estimates in this week's calendar are not a forecast of higher inflation. They are a forecast about composition: consensus puts headline CPI at 0.9-1.0% month-over-month against a 0.3% previous, while core is expected at just 0.3% against 0.2%. That is a wedge of roughly 0.6-0.7 percentage points opening in a single month, and it embeds an assumption most commentary will skip past — that a shock large enough to lift the annual headline rate from 2.4% to 3.3-3.4% leaves the underlying series essentially where it was.
This matters today because the assumption is testable inside five days. Core CPI month-over-month decides whether the shock is classified as a relative-price event or the start of a broadening; Michigan consumer sentiment, released the same afternoon, shows whether households are willing to treat it as consensus does; and the producer-price print on April 14 — estimated at 1.1% against a 0.5% previous — is the corroborating evidence on the cost side.
| Headline CPI MoM (%) | Core CPI MoM (%) | |
|---|---|---|
| 2026-02-01 | 0.3 | 0.2 |
| 2026-03-01 | 1.0 | 0.3 |
The load-bearing number is core, not headline
A headline print near 1.0% would annualize above 11% if sustained, and on the year-over-year rows consensus asks for a jump of roughly 0.9-1.0 points in a single month against a core year-over-year move of only 0.2 points, from 2.5% to 2.7%. That asymmetry is precisely why the headline number carries the least revision risk on the sheet: an outsized print merely confirms what estimates already contain. The variable that would force investors to revise is core month-over-month, where the distance between 0.3% and 0.4% looks trivial on paper and is not — 0.3% compounds to roughly 3.7% annualized, 0.4% to roughly 4.9%. Core is the series a central bank reaction function leans on, because energy and food price levels are not policy-controllable and are conventionally treated as noise around the trend; a 0.4% handle removes the ability to describe the month as noise at all.
Consensus is not forecasting inflation this week; it is forecasting that inflation stays in the categories investors have been told will be looked through. That is a far more specific — and far more fragile — claim than a directional call on prices.
The second dataset that does not agree with the first
The same calendar carries Michigan Consumer Sentiment for April, estimated at 52.0 against a 53.3 previous — a decline of 1.3 points, released roughly ninety minutes after the CPI data. Set the two expectations side by side and the internal tension becomes visible. Household inflation perception is formed disproportionately by the prices purchased most frequently and most visibly, which are the food and energy components sitting outside core. Consensus is therefore expecting the largest headline monthly acceleration in this sequence to coincide with a sentiment reading that barely moves.
One of those two estimates is likely to be wrong, and the direction of the error is informative rather than symmetric. If sentiment falls materially past 52.0, it suggests the shock is registering in expectations, which is the channel through which a goods-and-energy event migrates into the wage and services categories that define core. The absolute level compounds the fragility: a reading in the low 50s already sits near the depressed end of the index's historical range, leaving little buffer before the consumer datapoint itself becomes the story rather than the inflation print.
The cost side offers the cleanest corroboration, and it arrives April 14. Producer prices are estimated at 1.1% month-over-month against a 0.5% previous — a prior print that was already running well ahead of core consumer inflation. Two independent series pointing the same way changes the reading: a single elevated headline CPI month can be dismissed as a commodity print, but producer prices accelerating from an already-elevated base alongside it is more consistent with a genuine input-cost shock working through the chain. That distinction is what separates a one-month statistical artifact from a margin problem.
Which is the second-order consequence worth sizing. If producer prices rise at 1.1% while the consumer core component companies can actually pass through rises at 0.3%, the gap lands in gross margin rather than in the CPI. The first read on whether that gap is showing up in corporate language comes immediately: JPMorgan and Johnson & Johnson report April 14 against consensus of $5.47 and $2.68 respectively, followed by TSMC ($3.31) and ASML ($7.72) on April 15. The relevant content in those releases is not the earnings comparison but the input-cost and pricing commentary — semiconductor capital equipment and foundry guidance are unusually early indicators of whether cost pressure is being absorbed or passed on.
What else could explain the estimates
The strongest competing interpretation is that forecasters have identified a specific, known, one-off step change — an administered price, tariff, or energy adjustment with a defined effective date — and have correctly modelled it as mechanically confined to non-core categories. On that reading the wedge is not an assumption at all but arithmetic, and core at 0.3% is simply the underlying trend continuing undisturbed. The bundle does not identify the driver, so this cannot be dismissed. This publication still favors the primary reading for two reasons: the producer-price estimate shows the acceleration is not confined to the consumer basket, and the sentiment estimate implies household expectations remain almost unaffected by a shock consensus itself sizes at roughly a full percentage point of annual headline inflation. One-off shocks can be arithmetically contained in a price index; they are harder to contain in expectations.
The demand-side cross-check runs through housing. Existing home sales for March, due April 13, are estimated at 4.06 million against 4.13 million previously — a modest decline consistent with rate-sensitive demand already operating under pressure. If the inflation data forces a repricing of the rate path, housing becomes the fastest-transmitting evidence of whether the shock is being met by an offsetting demand slowdown or is layering onto one. Disclosure channels added nothing directional this window: among the 662 congressional transactions reported, the largest sampled entries were a Microsoft purchase, a Chevron sale, and a series of Treasury-bill buys with trade dates back in December, and the bundle recorded no insider buy clusters and no institutional position changes.
Portfolio impact map
The wedge holds: core prints 0.3%, sentiment lands near 52.0, PPI comes in below 1.1%
The shock is classifiable as a relative-price event, the look-through framing survives, and the load-bearing variable resolves in favour of the existing rate path.
Typically supported
- Duration-sensitive fixed income exposure
- Rate-sensitive equity sectors including housing-linked demand
- Long-duration growth equity valuations
Typically pressured
- Commodity-linked inflation hedges
- Positioning sized for a near-term policy repricing
The wedge breaks: core rounds to 0.4%, or sentiment falls well past 52.0, or PPI confirms 1.1%
The shock stops being separable from trend inflation, and both the policy path and the margin assumption embedded in April guidance come under revision.
Typically supported
- Shorter-duration and floating-rate exposure
- Businesses with demonstrated pricing power
- Real-asset and commodity-linked sensitivities
Typically pressured
- Long-duration assets exposed to further rate repricing
- Margin-thin manufacturers absorbing input costs
- Consumer-facing demand tied to sentiment
- Housing-linked cyclical earnings estimates
What would change this read
- Core CPI MoM (Mar), estimate +0.3% vs +0.2% previous — the single load-bearing number. A 0.4% handle annualizes near 4.9% and removes the ability to describe the month as a contained relative-price shock.
- Michigan Consumer Sentiment (Apr), estimate 52.0 vs 53.3 — released the same afternoon as CPI. A decline materially larger than 1.3 points would be consistent with the headline shock migrating into household expectations, the channel by which non-core prices reach core.
- Producer Price Index MoM (Mar), April 14, estimate +1.1% vs +0.5% — the corroboration test. Confirmation on top of an already-elevated prior print favours the input-cost reading over the one-off explanation, and puts the PPI-versus-core gap into margins.
- April 14-15 reporting cluster: JPM, JNJ, then TSM and ASML — input-cost and pricing language, not the EPS comparison, is the evidence on whether the cost gap is being absorbed or passed through.