US Business Activity Just Hit a 62-Month High — and the Bond Market Is Now Pricing the Second Warsh Hike

S&P Global's flash release, dated September 23 and covering data collected September 10–22 from roughly 650 manufacturers and 500 service providers, showed the headline Composite Output Index rising from 56.0 in August to 58.4 in September — a 62-month high and the fastest expansion since July 2021. The acceleration was the fourth in a row and, critically, was led by both pillars at once rather than by a one-sided swing in services.

For the Fed, the input-price gauges are the more important signal — and they broke out. Overall input cost inflation hit its highest reading since October 2022, and the service-sector input-cost gauge jumped to its fastest since November 2022. The trigger, per S&P Global's narrative, was higher fuel and transport costs during the survey window, with supplier delivery times lengthening to their worst since July 2022.

Selling-price inflation picked up from August but stayed below the March–July run-rate, because service-sector competition kept a lid on passthrough. That combination — accelerating costs, only partially passed through — is precisely the pattern that traps the Fed: corporate margins compress in the near term, but the impulse to push prices harder comes back if demand holds.

Chair Warsh's September 16 hike — unanimous at 12-0, the first move of his chairmanship — landed with a marked-up inflation path in the Summary of Economic Projections: PCE for 2026 was revised to 3.7% (from 3.6% in June), and the Fed does not see 2% again until 2029, a year later than it did in June. The median dot landed at 4.1% by year-end 2026, implying exactly one more 25bp move before the December meeting.

The bond market moved in lockstep. The 10-year Treasury yield — which first touched 5% on September 14 — pushed to 5.08% on September 23, up 11bp on the session, with the 10-year real yield at 2.6% and the 10-year breakeven at 2.36%. The 30-year sits at 5.33%, the 2-year at 4.75%. The iShares 20+ Year Treasury Bond ETF — pure duration — has lost3.34% YTD on a total-return basis, with a 30-day SEC yield of 5.23%.

For the money-center banks, every additional 25bp at the front end is worth real money. JPMorgan Chase has guided 2026 net interest income to roughly $103 billion — up 7% year over year — and that guidance was set with two cuts in mind, not zero cuts. Q2 2026 NII alone printed $25.5 billion, with average loans up 10% year over year and deposits up 7%. Wells Fargo, Bank of America, and Citigroup all run asset-sensitive balance sheets that benefit when the front end rises further.

The catch — and why a flat 10-year would have been even better for the sector — is curve shape. The2-year at 4.75% versus the 10-year at 5.08% leaves a curve that is steep but not extreme; the bigger tailwind is the level of short rates, not the slope. As long as the Fed's path is 'higher for longer' rather than 'higher then crashing growth,' NIM keeps expanding.

Higher rates are not happening in a vacuum — they are happening because oil and fuel are forcing the input-price gauge higher. The Iran war has kept Brent crude in the $90s and pushed US diesel to a record $5.85 a gallon in September. That backdrop is what powered ExxonMobil's $14.7 billion of adjusted Q2 earnings — its strongest quarter since Q3 2022 — and a record $12.1 billion quarterly profit at Chevron. Average oil in Q3 2026 ran $84.39, roughly 30% above Q3 2025.

The PMI's fuel-and-transport line is the bridge from the macro print to the upstream operators: every basis point of input-cost acceleration tied to diesel and freight accrues to refiners and integrated producers first, then to oilfield-services as capex budgets re-open. The Energy Select Sector SPDR ETF has compounded roughly 48% over the trailing year — almost entirely a crude-price story.

The flip side of 'banks and energy' is duration assets. Realty Income — the cleanest retail-net-lease duration proxy in the S&P 500 — trades at 41× trailing earnings and 33× forward, a multiple compressed almost entirely by the discount-rate channel. With the 10-year at 5.08%, the S&P 500's forward P/E has already compressed from roughly 23× to 19× over the past year as real yields moved materially higher. Multiples that were paid on the assumption of a 4-handle on the 10-year are being repriced for a 5-handle.

The mechanism is mechanical, not thematic. A 100bp move in the 10-year cuts roughly 8–10% off the present value of a 15-year cap-rate stream. For REITs, that is direct equity-value compression on top of higher refinancing costs when secured debt matures. With the Fed projected to sit at 4.1% through year-end, the duration trade has nothing to lean on until the inflation picture cools — and today's PMI makes that less likely, not more.

The market narrative entering September was 'when does the Fed start cutting.' The narrative exiting the September 16 meeting was 'one more hike and then a long pause.' Today's PMI, with input costs at a 4-year peak and services employment accelerating, makes the pause case harder to defend. The October 27–28 meeting is now a live decision: the dot plot says one more move; the data supports it; the bond market has done half the work already by pushing the 10-year to 5.08%.

The investable translation is asymmetric. Bank NIM keeps expanding as the front end stays high; energy earnings keep surprising on refining margins and oil realizations; duration assets keep bleeding. The risk is not that the Fed cuts — the dot plot says it doesn't — but that growth finally cracks under the higher-rate regime. The PMI says we are not there yet.