The week the curve reversed, and not for the reason anyone wanted. Wednesday delivered a split inflation print: July headline PCE rose 0.2% on the month and 3.7% on the year against a 3.6% consensus, running hotter than expected, while core came in at 0.2% and 3.3%, in line. The second estimate of second-quarter GDP was revised down to 1.5% from 2.1%. Nvidia followed after the close with a blockbuster outlook that carried the Nasdaq into Friday. Then Kevin Warsh took the Jackson Hole podium for his first keynote as chair and said it directly: inflation is running too high, price stability is the predominant focus, and the central bank still has work to do. He described himself as impressed with the economy's overall strength while flagging that underlying inflation trends have not improved, and doubled down on giving markets less explicit forward guidance. The rates response was immediate. The two-year yield jumped more than six basis points to 4.298% while the thirty-year eased two basis points to 5.168% and the ten-year held near 4.676%. Equities moderated from early gains, with the Nasdaq giving back a 0.5% advance to trade down 0.4% into the afternoon while the S&P 500 still tracked a weekly gain. PayPal fell 16% after Advent and Stripe abandoned their pursuit of the company.
The long end finally backed down, which is exactly what edition #032 said this universe needed, and it happened for the opposite reason from the one that would have helped. A hawkish chair pulled the thirty-year lower by convincing the market he will not tolerate inflation, which compresses the term premium that broke loose last week. The price of that relief was six basis points at the front end and a September meeting that is live again. This is a credibility trade rather than an easing trade: the curve flattened because the market now believes the Federal Reserve will do the work, not because the work is finished. For a library of companies that finance against the long end and get repriced by the front end, that is a two-sided outcome rather than a win.
The data underneath deserves plain language. Headline PCE at 3.7% ran above consensus, core held at 3.3%, and both sit well north of a 2% target while second-quarter growth was revised down to 1.5%. Slowing growth with sticky prices is the combination that leaves a central bank no comfortable option, and it is why Warsh chose candor over guidance. The counterweight is real and should not be dismissed: Nvidia's outlook says the artificial intelligence capital cycle is intact, and that single fact is holding up a tape that macro alone would have sold. Two prints now stand between this speech and the September 16 decision, the August jobs report and August CPI, and the standing line from #022 still governs the second of them, because the Strait keeps crude inside every inflation forecast.
Coverage ledger: front-end risk returning is the variable to respect, because a live September meeting reprices risk appetite at the speculative end of the market first, and that is where this library lives. Long-end relief helps any name that must raise capital, so NEOV gets a modest tailwind it did not have last week. BTBT sits directly downstream of the Nvidia outlook and the capital expenditure cycle it confirms. GLMD and ACON keep the cash floors that make a live hike meeting a question rather than a threat, and CLPS remains a function of a risk appetite that a hawkish chair just made more expensive. Full dossiers in the research library.