The Economy — Digest
TL;DR
The Fed's October hike is now effectively dead: a BEA methodology revision cut core PCE more than expected (3.0% vs 3.3% consensus), October odds collapsed from over 80% at their peak to ~37%, and Goldman and JPMorgan both moved their next-hike call to December. But the long end refused the relief — the 10-year closed near 5.30% and the 30-year at its highest since 2002 — while the pump squeeze eased only fractionally, with gasoline setting a record for the time of year.
The October hike died by revision — and the revision itself is now the story
The biggest single repricing of the cycle came from a data release, not a speech. On September 30 the BEA's stacked report showed core PCE at 3.0% y/y against a 3.3% consensus, with methodology changes (software, portfolio management fees, legal services) lowering core inflation by about 36 basis points — more than the 20-30bp economists had expected — and July rewritten from 3.3% to 3.0% (Finding: PCE revision+1). October hike odds fell from above 80% earlier in the month to ~37%, with the next expected increase pushed to December, and Goldman formally moved its call — seeing "a strong chance the Fed ultimately decides no additional hikes are necessary if disinflation persists" (Roic News).
The skeptics are lined up and worth taking seriously. Santander's Stephen Stanley: "Two months' worth of data is not enough to point to a new trend, especially when the monthly reading for August reaccelerated" — August's monthly gain was driven by a 4.4% gasoline rebound (Finding: PCE revision+1). And the revision lands in a charged context: the BEA is now covered by a Commerce scientific-integrity policy that deleted anti-interference language, Sen. Warren has set an Oct 7 deadline for answers, and a 36bp downward cut to the Fed's target gauge four weeks before the midterms guarantees scrutiny from both directions (Finding: PCE revision
+1). The thread on the methodology implementation resolved this week — with the revision doing even more work than projected.
What to watch: whether the Fed responds to Warren by the Oct 7 deadline — the data-integrity overlay now attaches to every future print.
The long end is pricing a different economy than the front end
The softest inflation print in months did nothing for bonds. On September 30 the 10-year rose more than 4bp to 5.298% — near 2007 highs — and the 30-year gained to 5.632%, "around its highest level since 2002," after yields initially fell on the PCE release and then reversed as traders looked ahead to Friday's jobs report (Finding: long end). This extends last digest's contradiction: front-end relief is not translating into long-end relief, and the curve is steepening into a Fed that may be nearly done hiking.
The transmission to households is direct and Fed-proof. With the 10-year pinned near 5.3%, mortgage rates stay in the 7%+ zone "regardless of what the Fed does in October" (Finding: long end). The market is pricing term premium, supply, and energy risk — not the policy path.
What to watch: whether the 30-year holds above 5.6% through Friday's payrolls — a hot print sends yields higher from here, not lower.
The labor story flipped from "cooling fast" to "rebounding" in 24 hours
Within a day of JOLTS showing openings at a five-month low, ADP printed a clear beat: private employment +90,000 in September against a ~68-70K consensus, "hiring accelerated for the first time since May" (Finding: ADP). ADP chief economist Dr. Nela Richardson: "It's a strong report… After a three-month slowdown, job creation rebounded and pay growth remained solid." The composition carries the real signal: goods-producing added 31K even with diesel at records, while financial activities (-16K) and professional/business services (-11K) show the white-collar soft patch deepening.
This matters for the October 27-28 meeting because payrolls are the last substantive gate: "If Friday's numbers come in hotter than expected — as did Wednesday's ADP private payrolls report — it could send yields higher" (Finding: ADP). The caveat stands — ADP has historically been a poor predictor of the BLS series — so treat direction, not level.
What to watch: Friday's September payrolls (~+84K consensus) — the one print that could revive October pricing from its deathbed.
The pump squeeze eased a hair, but Washington is now drafting pre-midterm interventions
Retail diesel fell 14.7 cents from its all-time record $6.529 to $6.382/gal — still up $2.628 year over year — while regular gasoline at $4.48 was "setting a record high for this time of year," a season when prices normally fall (Finding: pump prices). Brent closed September up ~14%, its largest monthly gain since July, as US-Iran talks stayed deadlocked over the sequencing of any Hormuz reopening (Finding: oil
).
The policy response has gone from watching to drafting: the White House has urged the EU to release emergency diesel reserves, is considering easing red-dyed diesel restrictions — "aiming to provide some price relief for consumers ahead of the November midterm elections" — and has floated diesel export restrictions (Finding: pump prices). Meanwhile the household cushion is thinning from the other side: the saving rate dropped to 4.1%, the lowest since November 2022, as spending outpaced flat real disposable income (Reuters). Last digest's sub-2.0% tripwire is further away than the old data implied, but the direction is unchanged.
What to watch: Sunday's OPEC+ meeting (Oct 4) — a hold keeps the diesel squeeze intact into heating season with distillate inventories at 105.2m barrels and falling.
What surprised us
- The methodology revision did more than anyone modeled. Economists expected 20-30bp; they got 36bp, enough to single-handedly kill the October hike. When a statistical overhaul moves markets more than the underlying data, the revision process itself becomes a market factor — and with the integrity-policy cloud over Commerce, both parties will read the same number in opposite ways.
- Goldman went beyond December. Not "later hike" but possibly "no additional hikes necessary" — sell-side is now pricing an end to the tightening cycle entirely, well ahead of the Fed's own median, which still embeds one more move to a 4.1% year-end midpoint.
- The long end ignored everything good. Soft PCE, a dovish Fed repricing, and yields still closed at 2002/2007 highs. The term-premium story has now survived two consecutive weeks of ostensibly bullish news.
- White-collar recession, blue-collar resilience. ADP's bifurcation — construction and manufacturing hiring while finance and professional services shed — is the mirror image of every prior downturn pattern in this cycle.
Open threads worth a vote
- [watch] OPEC+ sets November output targets (Oct 4 meeting) — a surprise increase would pressure pump prices before the Nov 3 midterms; the expected hold keeps the diesel squeeze intact.