The strip enters the second weekend of August with the September cut priced firmly above 90 percent on CME FedWatch, and the desk carries the calendar between now and the September 16 to 17 FOMC decision as six data prints and one non-data event, each with an independent read on where the cut probability sits when the Committee walks into the Eccles building. The sequence is not additive. Two of the seven inputs carry front-end repricing weight that dwarfs the other five combined, and one of the seven is a confirming line rather than a moving one. The desk reads the calendar in that order.

The three high-weight windows

Three inputs carry the largest independent repricing weights inside the August window.

The first is the July CPI release in the second week of August. The higher-signal line inside the release is core services ex-shelter (supercore), not the headline. Supercore is the read the Fed prices against because it strips out the categories that are either mechanical (shelter, which lags observable market rents by roughly twelve to eighteen months) or supply-side driven (energy, food). A supercore three-month annualized print that walks the trailing series from the June reading through the mid-3 percent band walks the September cut probability further inside 99 percent and keeps a second cut at the December meeting inside the 75 percent band. A supercore print that walks the three-month annualized rate back above 4 percent walks the September cut probability out of 95 percent and into a materially two-sided distribution for the first time since May. The core goods line inside the release is the second read: any acceleration off the tariff pass-through carries directly into the goods-heavy PCE weights and pulls the year-end 2026 core PCE track higher.

The second is the FOMC minutes from the July 29 to 30 meeting, released three weeks after the decision in the third week of August. The strip reads the minutes for two specific things: the language around the balance of risks (whether the Committee framed the labor-market softening as consistent with the dual mandate or as running ahead of the inflation return-to-target), and the number and language of any dissents. The July decision landed as a hold with the SEP dot median walking to two 2026 cuts from three in the June SEP, and the minutes will carry the internal argument that produced that revision. A minutes readout that surfaces two or more dissenters against the hold (either dovish or hawkish) walks the September cut probability materially in the direction of the dissent count. A minutes readout that surfaces broad support for a September cut inside the internal deliberation walks the December second-cut probability from 68 percent Thursday close through 80 percent without the intervening data needing to move.

The third is Powell’s Friday morning keynote at the Kansas City Fed’s Jackson Hole Economic Symposium in the fourth week of August. The strip prices the Jackson Hole slot heavier than any single FOMC statement between press-conference meetings, because Powell has used the slot since 2022 to reset the framework read that governs the following twelve months of policy. A Jackson Hole speech that leans into the labor-market softening as the near-term dominant risk walks the September cut into a certainty and opens the door to a 50 basis point move. A speech that reasserts inflation-return-to-target as the binding constraint walks the September cut probability back inside the 85 to 90 percent band and closes the door on 50. A speech that stays inside the balanced-risks framing walks nothing on its own, and the strip reprices off the Q and A parsing at the end of the day rather than the prepared remarks.

The three lower-weight prints

Three inputs carry independent but smaller repricing weights.

The PPI release the day after CPI carries most of its signal through its pass-through to the following Friday’s core PCE nowcast. The categories that feed PCE directly (healthcare services PPI, portfolio management fee PPI, transportation services PPI) do more work inside the strip’s month-end read than the headline PPI print does on release day. A PPI healthcare services print that walks the healthcare PCE weight (roughly 17 percent of core PCE) higher than the CPI medical services line implied walks the July core PCE nowcast up by two to four basis points against consensus. The tape reads PPI as a two-day event: the release itself, and the following morning’s PCE nowcast revision.

Retail sales in the middle of the third week reads through the control group, not the headline. The control group excludes autos (which land at manufacturer-reported unit pace on the first business day of the month), gasoline (which tracks the retail energy price directly), building materials (which is home-improvement lumpy), and food services (which lands separately). The control group is the input to the personal consumption expenditure services line that feeds the Q3 GDP nowcast. A control group print that walks the trailing three-month annualized pace back above 4 percent walks the Q3 GDP nowcast from its current 2.4 percent band toward 3.0 percent and pulls the September cut probability inside 95 percent without the CPI print needing to run cool.

Core PCE on the last Friday of August is the confirming line. By the time core PCE prints, the strip has already priced the read off the CPI plus PPI plus retail sales sequence three weeks earlier, and the PCE nowcast has walked to within one or two basis points of the actual print in eleven of the last twelve months. Core PCE moves the strip only when it prints materially outside the walked nowcast band, and the modal outcome inside the current cycle is that it prints inside the band. The desk reads the release as a confirmation gate: a print inside the nowcasted band closes the September decision as effectively decided by the second week of August, while a print materially outside the band reopens the front-end distribution three weeks before the FOMC decision.

The one non-data event outside Jackson Hole

The seventh input inside the window is the Kansas City Fed manufacturing survey release in the third week of August. It is not a needle-moving print on its own. The strip reads it as one of the last regional Fed manufacturing surveys before the Chicago PMI at month-end (which is the last national manufacturing read that lands before the FOMC blackout). The Kansas City survey matters at the margin because it captures the energy-belt manufacturing footprint that carries higher-than-average sensitivity to the summer WTI trajectory and the natural-gas-turbine buildout demand that the reliability RTOs have been running against.

What the sequence carries

The desk reads the August calendar as a sequence, not a set. The July CPI print in the second week collapses the near-term distribution. The FOMC minutes and the Jackson Hole speech in the third and fourth weeks re-form the distribution against whatever the CPI print did. Core PCE at month-end confirms or reopens. The seven inputs together carry enough signal that by August 30, the strip’s September cut probability will sit within one to two percentage points of where it lands on the FOMC decision day, absent a September 3 to 5 payrolls print that breaks the two-standard-deviation band around the six-month average.

The single highest-weight day inside the calendar is not any of the seven inputs individually. It is the Friday morning of the Jackson Hole symposium, because it is the day that walks the framework read against everything the preceding two weeks of prints established. The strip has priced 90-plus percent September cut probability into the tape. Powell decides whether that probability is a floor or a ceiling.

Sources