Core CPI Is 2.4% and Core PCE Is 3.0%. Wednesday's Report Tests Which Inflation the Fed Is Reading

The September consumer price report is due at 08:30 US Eastern time on Wednesday 14 October 2026 (18:00 in India). The forecasts cluster around a headline rate of 3.6% to 3.7% and a core rate, which excludes food and energy, of about 2.4% to 2.5%. If core comes in at 2.4%, it would match August as the slowest since March 2021. Yet the inflation measure the Federal Reserve targets, core PCE, was 3.0% in August. That is a 0.6 percentage point gap with the usual order reversed: over the long run, consumer price index inflation has run above the PCE price index, not below it. This piece explains why the two diverge, how much of the current gap is a measurement artefact, and what Wednesday's number can and cannot change. It is written before the data release, so nothing here is a result.

What is expected on Wednesday, and by whom

The forecasts come from different sources and should not be averaged. Admiral Markets' preview lists a Street consensus of 3.7% headline year over year (as of 6 October), the Cleveland Fed's inflation nowcast of 3.60% headline and 2.39% core (7 October), and Kalshi market pricing of 3.63% headline. Continuum Economics, in a 29 September preview, expects headline up 0.6% on the month to 3.7% and core up 0.2% to 2.4%. A Reuters poll cited by Kitco on 9 October is slightly different: 3.6% headline and 2.5% core. For reference, August was 3.4% headline (0.4% on the month) and 2.4% core (0.3% on the month).

The headline is driven by energy. Continuum forecasts energy up 5.5% on the month, which would be its largest contribution since a 10.9% surge in March, with gasoline leading. Admiral's table shows why: in August, gasoline was up 27.4% year over year and energy 16.3%. Those are the numbers behind a headline that is rising even while core cools.

Why core may cool, and why that is a narrow signal

Continuum argues August's core upside was narrowly based. Core was led by telephone services (up 5.4%, adding about 0.096 percentage points to core), lodging away from home (up 2.4%) and used cars (up 0.4%). The telephone jump followed a methodology change in how that component is measured and is treated as a one-time effect. The Cleveland Fed's median CPI, which strips out the extremes, rose only 0.17% in August, down from 0.26% in July. Continuum's forecast for core excluding food, energy and shelter is +0.17% against +0.31% in August. If that is right, September core cools because three August spikes do not repeat, not because underlying pressure has broadly eased.

The forecast spread is itself informative. The Cleveland Fed's 2.39% rounds to 2.4% while the Reuters poll's 2.5% rounds one tenth higher, so a core reading of 2.4% versus 2.5% falls inside the range of the forecasters and should not be treated as a surprise in either direction.

The wedge: core CPI 2.4%, core PCE 3.0%

CPI and PCE measure different things. The Cleveland Fed notes that CPI covers out-of-pocket spending by urban households, while PCE also covers rural households and includes spending made on consumers' behalf, such as employer-provided health insurance, Medicare and Medicaid. PCE weights update monthly and CPI weights annually, which lets PCE reflect substitution faster and is one reason PCE inflation tends to run lower. Housing has a smaller weight in PCE. Since 2000, the Cleveland Fed says, CPI inflation has averaged 0.39 percentage points above PCE. A core CPI that is 0.6 points below core PCE is therefore not a small deviation from normal: measured against that average, it is a swing of about one full point, though the average is for all items and the current gap is for core, so treat the comparison as indicative.

Most of the swing is attributed to a few PCE categories, not the broad economy. PIMCO, in an analysis published 5 August, said the gap between core PCE and core CPI can largely be explained by two categories. Portfolio management fees rise with equity prices because they are measured in dollars of assets under management, even though PIMCO says fees as a percentage of assets have been declining. And software price measures, in PIMCO's view, do not fully capture quality gains from cloud services and AI features, which likely overstates inflation. PIMCO put underlying US inflation at 2.2% to 2.8% through June, against a core PCE then reported at 3.3%. Employ America's July analysis separately attributed 27 basis points of PCE overshoot to equity effects through portfolio fees, 37 basis points to tech equipment, and further amounts to airfares and health care, with the caveat that its contributions overlap and should not be summed. These are analysts' views, not official findings.

The wedge has already narrowed. On 30 September the Bureau of Economic Analysis released August data along with an annual update that changed how it measures portfolio management fees, legal services and computer software. Core PCE for July was cut from 3.3% to 3.0%, according to FXStreet, and August came in at 3.0% (0.2% on the month, against a 0.3% forecast; headline PCE was 3.4%, matching headline CPI). Babypips reported that the BEA's changes may have pulled core PCE lower. In other words, the 0.3 point reduction came from the annual update's revisions and method changes rather than from new price moves. Before the revision the gap to core CPI (2.5% in July) was about 0.8 points; after it, about 0.6.

What Wednesday can and cannot change

The Fed raised its target range to 3.75%–4.00% on 16 September, its first increase since 2023, by a unanimous 12–0 vote, and the median projection implies one further quarter-point rise by year-end, to 4.1%, according to Admiral Markets. Its next meeting is 27–28 October. After the August PCE report, FXStreet reported CME FedWatch pricing of about 35% for an October hike, 100% for a hike by 9 December, and 16% for two hikes by then; New York Fed president John Williams had said on 29 September there was no need for urgency. Kitco reported on 9 October that traders have reduced October hike bets further. Those odds are a few days old and move with data.

A soft core CPI would cut against the argument that inflation is broadening, but it would not by itself settle the Fed's decision, for two reasons. The Fed's target is PCE, which carries the very categories driving the wedge, and September PCE is not due until 29 October, the day after the Fed meets, so the central bank will decide with CPI and producer price data in hand but not the PCE print. And a core CPI of 2.4% sits 0.4 points above the Fed's 2% target while core PCE sits a full point above it. A hot core CPI, say 2.6% or more, would be a larger surprise, since forecasts are clustered at 2.4% to 2.5%; it would revive hiking bets, as Kitco's market commentary suggests.

The 10-year Treasury yield, last around 5.23% per Kitco on 9 October after reaching its highest level in 24 years, adds a second channel. A soft core could ease long yields; a firm headline driven by energy could keep them high. Kitco relays strategists' warnings that a 5% yield has been seen as a threshold for equity pressure.

What the evidence does not establish

The wedge explanation rests on analysts' attribution of PCE categories, not on a decomposition published by the statistical agencies. The sources differ on how much of the gap is software, how much is portfolio fees, and how much is airfares and health care. The nowcasts disagree by 0.1 points. And the August core upside was narrowly based, so one soft month would not show disinflation. Treat the 14 October print as one data point about CPI and not as a verdict on which gauge the Fed should trust.