Reading a CPI Print: The Six Numbers That Move Rates and FX

Once a month, at 8:30 AM Eastern, the US Bureau of Labor Statistics releases the Consumer Price Index report. It is the single most-watched inflation number on the planet, and in a Fed-focused market it moves dollars, yields, and gold faster than almost any other scheduled release. Yet the same problem that plagues NFP reads plagues CPI reads: retail traders anchor on the one headline number, miss the internals, and misread the direction the tape takes.

This article walks the CPI release the same way we walked the payrolls release: the six numbers that actually matter, how the Fed reads each of them, and how to interpret the tape reaction against what was priced.

What CPI actually is

The Consumer Price Index measures the change in prices paid by urban consumers for a representative basket of goods and services. The Bureau of Labor Statistics surveys prices monthly at retail outlets, service establishments, and rental units across 75 urban areas, then combines them into indexes.

The release comes out on a fixed schedule the BLS publishes each December for the following year. It usually lands in the second or third week of the month, covering the prior month’s data (September’s CPI print reports on August prices). Related reading: central bank meetings explained for how CPI feeds directly into the Fed’s rate decision.

There are two headline series and a family of underlying components. Understanding what each measures is the first step in reading the release intelligently.

The six numbers that actually matter

1. Headline CPI (all-items)

The all-items index. It includes food and energy and captures the full basket of consumer prices. Consensus estimates come in two flavors:

Month-over-month percent change (e.g. 0.2%, 0.3%, 0.4%). This is the short-run signal. A 0.1% MoM surprise, positive or negative, is meaningful in the current regime.
Year-over-year percent change (e.g. 2.9%, 3.1%). This is the level the Fed and the market anchor on for the medium-term inflation story.

Headline moves markets because it is what shows up on cable news and what non-economists watch. But it can be noisy month to month because food and energy are volatile, and the Fed explicitly targets a broader definition of inflation. So headline moves the tape in the first 30 seconds, then core takes over.

2. Core CPI (all-items less food and energy)

The Fed’s preferred CPI measure. It strips out food and energy, which are volatile and often driven by supply shocks rather than underlying demand. Core is the number every rates trader and every FOMC participant is watching most closely.

Consensus is again expressed in MoM and YoY:

Core MoM: 0.2% to 0.4% has been the recent range. A 0.5% MoM print is hot and typically dollar-positive, yield-positive, gold-negative.
Core YoY: the medium-term anchor. Above the Fed’s 2% target for years now. The market cares about the trend more than the level: is it decelerating month over month, or is it sticky?

Core surprises usually move rates more than headline surprises. If headline is hot but core is in line, the market often fades the initial move as it reads through to the Fed’s reaction function.

3. Super-core services (services ex-shelter)

This is the number the Fed added to its lexicon in 2022 and has been talking about ever since. It measures services inflation excluding rent and owners’ equivalent rent. Think haircuts, restaurant meals, insurance premiums, medical services, transportation services.

Why the Fed cares: services ex-shelter tracks wage-driven inflation. It moves slowly, and it is the last piece of inflation to come down in a disinflation cycle. If super-core is running hot while headline is cooling, the Fed will lean hawkish because the disinflation is not yet reaching the sticky, wage-sensitive parts of the basket.

Super-core is not always reported in every consensus preview, but it is in the release, and it is what the Chair will emphasize in the next press conference. If you are trading around a CPI print into an FOMC meeting, this is the number to weight highest.

See reading a central bank statement for how the Chair typically frames super-core in the accompanying press conference.

4. Shelter (rent and owners’ equivalent rent)

Shelter is roughly one-third of the total CPI basket and roughly 40 percent of core CPI. It is enormous, and it moves slowly.

The two components:

Rent of primary residence: what tenants actually pay in rent.
Owners’ equivalent rent (OER): a survey-based estimate of what homeowners would pay to rent their own home. This is imputed, not observed.

Shelter is the slowest-moving component of CPI. It lags the actual rental market by roughly 12 to 18 months because the BLS methodology samples lease renewals rather than new leases. So when real-time rent indexes (Zillow, Apartment List) are showing rents falling or flat, the CPI shelter component can still be printing hot for another year.

Traders who understand this lag can read future CPI prints in advance: if real-time rent data is decelerating today, shelter CPI will decelerate 12 to 18 months from now. This is one of the highest-signal setups in the release.

5. Energy

Gasoline, fuel oil, natural gas, and electricity. This is the most volatile component. It swings with crude oil prices, refinery margins, and seasonal demand.

Energy is often the single largest driver of the difference between headline and core. A hot headline with a cool core almost always traces back to an energy spike. The Fed looks through most energy moves because they are largely supply-driven and mean-reverting.

But if energy is stubbornly high for many months in a row, it starts to feed into inflation expectations and wage negotiations, and the Fed can no longer look through it. This is the “second-round effects” story the ECB and Fed both talk about.

6. Food

Food-at-home (grocery) and food-away-from-home (restaurants). Food-at-home is more commodity-driven and volatile; food-away-from-home is stickier and more wage-driven.

Food is smaller than energy but politically potent, because everyone buys groceries and everyone notices when they get more expensive. The Fed looks through short-run food spikes for the same reasons it looks through energy spikes. But sustained food inflation feeds into inflation expectations the same way sustained energy does.

The tape reaction and how it usually unfolds

The 30-second, 5-minute, one-hour framework from the reading an NFP print piece applies to CPI too, with one important difference: CPI is a slower release to fully digest because there are more components and the components interact.

First 30 seconds: algorithmic reaction to headline and core MoM. The initial move is roughly proportional to the surprise vs consensus.
First 5 minutes: super-core, shelter, and energy get integrated. Sometimes the initial move partially reverses if these components tell a different story than the headline.
First hour: the sector breakdown, the diffusion index of price changes, and economist commentary crystallize the narrative. Fed reaction function updates start priced.

A hot headline with a cool core and a cool super-core is often faded within 15 minutes because the Fed will not react to it. A cool headline with a hot super-core often extends its initial move only briefly, because the Fed will still see sticky services inflation as the reason not to cut.

The Fed reaction function overlay

Every CPI print is read against the Fed’s current stance, exactly as with NFP:

Fed hawkish (fighting inflation): super-core and core MoM are the two numbers to weight most. A hot super-core kills a rate cut narrative even if headline is cooperating.
Fed cutting (labor market focus): headline is what gives them the political room to cut. A cool headline and cool core are what they want to see to justify continued easing.
Fed on hold (data-dependent): everything matters, and CPI is often the single most market-moving release in this regime.

The real yields and FX piece walks through the transmission from an inflation surprise to Treasury yields to the dollar. In short: hot CPI raises nominal yields more than it raises inflation expectations, which raises real yields, which is dollar-positive. Cool CPI does the reverse.

Common misreadings

“The headline number is the print.” No. Same mistake as with NFP. The release is six numbers plus a hundred sub-components. The headline is one input.

“Core is always more important than headline.” Usually, but not always. When energy is doing something extreme (a geopolitical shock, an OPEC decision), headline moves markets on its own because it feeds into inflation expectations even if the Fed initially looks through it.

“CPI is the Fed’s target.” Actually no. The Fed targets PCE (Personal Consumption Expenditures) inflation, which is calculated by the Bureau of Economic Analysis and uses a different basket weighting. CPI is what the market reacts to in real time because it publishes earlier and more frequently, but the Fed will point to PCE in speeches. The two usually move together but can diverge in specific months.

“A cool CPI print means the Fed will cut.” Not by itself. The Fed needs a trend, not a single print. Two or three consecutive cool prints matter; one cool print in the middle of a hot trend does not shift the reaction function much.

“CPI is deterministic.” The BLS methodology has real uncertainty. Weights are updated annually, seasonal adjustments are revised, and the shelter series lags real-time rents. Read the release with the same skepticism you would bring to any survey-based statistic.

What to do the week of a CPI print

If you are trading around a CPI print:

Read the consensus preview 24 hours ahead. Reuters and Bloomberg publish economist ranges; the range width tells you how much uncertainty is priced.
Know where positioning sits. Is the market long dollars into the print, or short? A hot CPI print into short-dollar positioning produces a bigger move than the same print into flat positioning.
If you are going to trade the release, do it on the 5-minute reaction, not the 30-second reaction. Retail traders who click in the first 30 seconds get filled at very poor prices. See news trading the honest take for the transaction-cost math.
If you are already in a position, understand what a hot vs cool print does to your thesis. A macro dollar-long position through a hot CPI print is on the right side of the market for the first hour; through a cool print, it is on the wrong side. Size accordingly ahead of time.

The one-sentence summary

CPI is six numbers plus a hundred sub-components, not one headline number, and the market’s job on release day is to update its estimate of the Fed reaction function based on how headline, core, super-core, shelter, energy, and food all land together; retail traders who watch only the headline number miss the actual signal and get filled at poor prices in the first 30 seconds.

Related reading: reading an NFP print for the companion piece on the other most-watched US data release, central bank meetings explained for the Fed context every CPI is read against, and real yields and FX for how inflation surprises transmit to dollar direction.