Invest1 publisher2 min readPublished
A 0.2-point revision to core PCE would still leave it 1.1 points above the Fed's target
The BEA publishes its annual national accounts revision on September 30, fourteen days after the FOMC votes, and Goldman Sachs and JPMorgan both model it taking 0.1 to 0.2 points off core inflation back to 2021.
The Investor · Invest desk
What happened
- Markets are pricing roughly 90% odds of a rate hike at the Federal Reserve's September 15-16 meeting, after a hotter-than-expected CPI report unsettled traders and policymakers.
- The Bureau of Economic Analysis publishes its August PCE report on September 30 alongside its 2026 annual national accounts revision, which recalibrates technology and service prices back to 2021.
- Those methodological revisions are expected to lower core PCE readings retroactively by 0.1 to 0.2 percentage points.
- Goldman Sachs and JPMorgan both estimate that May 2026 core PCE falls from a reported 3.4% year-over-year to between 3.2% and 3.3% once the revisions apply.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- exposure A hike at 90% odds is already paid for, so the position that can still move is the one carried into September 30, when the restated series prints.
- constraint The adjustment is at most 40% of the half-point spread in the August monthly trackers, so it cannot settle whether core is running nearer 3.2% or 3.7%.
- precedent A lower 2023 to 2025 core series would put the calibration of past tightening decisions on the table at every future meeting where officials cite that history.
Apply the top of the range Goldman Sachs and JPMorgan have modelled, 0.2 points, to July's 3.3% core PCE reading, and it becomes 3.1% [5][2][14]. Against the 2% target that is an overshoot of 1.1 points where the published series says 1.3, so the revision takes about a seventh off the gap [14][15].
The August tracking estimates are noisier than the revision itself. Firms are penciling in monthly core gains of 0.26% to 0.30% [6]; compounded over twelve months, those two numbers annualise to 3.2% and 3.7% [16]. The month-to-month uncertainty is therefore about half a point wide, and the methodological adjustment is 0.1 to 0.2 [18]. Crypto Briefing reports that the trackers were built off recent CPI trends and do not include the new methodology [6].
What the BEA is changing is how it prices cloud computing, streaming services and healthcare delivery, areas where the agency's existing tools have lagged actual pricing [8]. The recalibration reaches back to 2021, five years of history [3]. Anna Wong at Bloomberg Economics has noted that the revisions are likely to soften what have appeared to be stubborn spikes in core inflation [7].
I would weight the dull version most heavily: the recalibration lowers the 2021 through 2025 level, the recent monthly prints hold in the 0.26% to 0.30% band, and a hike ends up looking defensible on the numbers the committee had in front of it [3][6]. Or rather, the more interesting version is that the new methodology also pulls down the August increment, and that is the case where November pricing swings toward a pause [12]. There is a third: the published revision lands outside the 0.1 to 0.2 points two banks modelled, in either direction [2][4].
Treating a hawkish repricing on the 16th as mismeasured requires the second case, and the evidence for it is two sell-side models, not a BEA release [4]. Tom di Galoma of Mischler Financial has cautioned against hiking on data that is about to be rewritten, pointing to divisions inside the Fed [9]. Those divisions are where the timing bites. Officials uneasy about tightening into a cooling labor market argue their case on the old series, while officials citing the 2% target get their vote before the restated numbers exist [10][2]. The committee decides on September 16 and sees the revision on September 30, fourteen days later [11][17].
What to watch
- Whether the September 30 release moves only the historical level or also the recent monthly increments, the case a November pause needs.
- Where November FOMC pricing sits in the days after the revised series lands, against the roughly 90% attached to September.
- Whether the BEA's published revision falls inside the 0.1 to 0.2 point range Goldman Sachs and JPMorgan modelled.