The Federal Open Market Committee next meets on 15 and 16 September. Between now and then it receives two inflation reports, the minutes of its own last meeting, and a further employment report — a compressed sequence that will do more to determine the outcome than anything any official says in the interim.
That sequence matters more than usual because the Committee is divided. July’s decision to hold the target range at 3-1/2 to 3-3/4 percent passed 9–3, with Beth Hammack, Neel Kashkari and Lorie Logan all preferring a quarter-point increase. When a committee splits that way, incoming data does not merely inform the decision. It arbitrates an argument.
Here is what arrives, and what each release can and cannot settle.
12 August — July CPI
The most consequential release in the window. Headline CPI ran at 3.5 percent year-on-year in June, above the Fed’s 2 percent objective, as it has been since 2021.
The number that will move markets is not the headline. It is the month-on-month core rate — prices excluding food and energy — annualised and compared against the recent trend. Headline inflation over twelve months is heavily determined by what happened eleven months ago, and base effects can produce a falling annual rate during a period when prices are accelerating. The monthly core figure is the closest thing to a current reading.
Two components deserve specific attention:
- Shelter. Roughly a third of the CPI basket, and measured through a methodology that lags observable market rents by a year or more. Shelter inflation is therefore partly a report on the past. Both hawks and doves use this to dismiss prints they dislike, which is a reason to look at core services excluding shelter separately.
- Core services excluding shelter. The category most closely tied to wage costs, and the one Fed officials have historically pointed to when arguing that inflation is domestically generated rather than imported. This is where the dissenters’ case is strongest or weakest.
What it cannot settle: one monthly core print is well within the range of normal statistical noise. It will move September pricing regardless.
13 August — July PPI
Producer prices measure what firms receive rather than what households pay. The report is treated as a leading indicator of consumer inflation on the theory that input costs eventually pass through.
That theory is weaker than its popularity suggests. Pass-through depends on margin structure and competitive position — firms with pricing power pass costs on, firms without absorb them into margins. PPI has repeatedly moved without CPI following.
Its more reliable use is as an input to the PCE price index, the Fed’s actual preferred measure. Several PCE components are constructed from PPI source data rather than CPI, so analysts use the two reports together to estimate PCE before it is published. That estimate, not the PPI headline, is what desks will be circulating on the 13th. We set out the differences between these indices in how to read inflation data without being misled.
19 August — minutes of the July meeting
The minutes will describe the discussion behind the 9–3 vote. They are the only account of that argument that will be available for five years, which is when full transcripts are released.
Read them for the distribution of opinion rather than the conclusions. Minutes use deliberately calibrated quantifiers — “a few,” “several,” “many,” “most” — and the useful question is how many participants were sympathetic to tightening without formally dissenting. Three recorded dissents plus four or five sympathisers is a meaningfully different Committee than three isolated hawks.
The important limitation: the minutes describe a meeting held before the July employment and inflation data existed. They tell you where the Committee stood three weeks ago, not where it stands.
4 September — August employment report
The final employment report before the meeting, and the one that will carry the most weight, because it also revises July and June.
Payroll growth has been decelerating: 129,000 in May as revised, 57,000 in June. If that pattern extends, the majority’s case for holding strengthens considerably regardless of what inflation does, because a labour market weakening that fast will suppress wage pressure with a lag. If payroll growth reaccelerates while core inflation holds near current levels, the dissenters’ position becomes substantially harder to argue against.
As always, the revisions matter as much as the headline, and for the reasons set out in why the jobs number gets revised twice.
15–16 September — the meeting
September carries a Summary of Economic Projections, which makes it more informative than a decision-only meeting. Every participant — including the regional presidents who do not hold a vote that year — submits a projection for growth, unemployment, inflation and the appropriate policy rate.
The resulting dot plot is the clearest available measure of whether the three dissenters are an isolated wing or the visible edge of a larger shift. Three dissents alongside a median dot showing no change is a contained disagreement. Three dissents alongside a median dot that has moved up is the beginning of a policy turn.
One standing caution about the dot plot: it is a projection conditional on each participant’s own forecast, not a commitment, and it has been a poor predictor of actual policy over horizons longer than a couple of quarters. It is a snapshot of opinion, and it is routinely reported as a plan.
What markets are pricing
Equities have been strong through this period — the S&P 500 closed above 7,700 for the first time on 3 August and set a record the following session — which is not obviously consistent with a market expecting tighter policy.
That tension is worth holding onto. Either the market believes the dissenters will not prevail, or it believes an economy strong enough to warrant a hike is good enough for earnings to absorb one. Both readings are defensible, and they imply very different behaviour if the September projections surprise. Federal funds futures give a live read on the first of those, and are a more reliable guide to expectations than equity levels.
Whatever the outcome, the mechanism by which any of it reaches the real economy runs on a lag of one to two years — which means the September decision will be judged on conditions that do not yet exist, using data that has not yet been collected, and which will itself be revised.
Sources: FOMC statement and meeting calendar, Board of Governors of the Federal Reserve System; Bureau of Labor Statistics release schedule; market levels as reported 3–4 August 2026.
