TO: Trading desk, macro book
FROM: [redacted]
RE: Today's 2pm liquidity event, and why silence is not a strategy, it's a hedge against being wrong out loud
Read this before you do anything stupid with the 2:00 p.m. print.
Nine to nine. That's the headline you need memorized before the minutes hit the tape. Half the FOMC wants a hike before year-end, half doesn't, and the man running the meeting submitted no dot at all. Kevin Warsh didn't just decline to guide markets — he declined to guide his own committee, then flew to Sintra and told the ECB crowd he's "not going to make a judgment now," as if that's a policy framework and not a hall pass. Core PCE got revised up to 3.3% for the year, from 2.7% in March. GDP got revised down to 2.2%. That combination has a name. It's called stagflation-lite, and normally a Fed chair facing it either picks a lane or gets run out of one. Warsh picked neither. He picked silence, wrapped it in a 130-word statement — half the length of what Powell used to publish — and called it discipline.
Here's what nobody on this desk seems to want to say plainly: silence isn't neutral. It's a transfer of volatility from the Fed's balance sheet to yours. Every basis point of uncertainty Warsh refuses to resolve gets priced somewhere, and right now it's getting priced in gold, which broke above $4,130 last week specifically on the theory that hike odds were fading, then in the September fed funds futures, where CME FedWatch has probability sitting in the low-to-mid 50s, down from 66% before the June jobs miss. Payrolls printed 57,000 in June. Unemployment fell to 4.2% anyway — a print that should not be possible under any labor market model taught before 2020, and isn't, except that the participation math is doing something ugly under the hood that the headline conveniently doesn't show you.
So which is it. Is the labor market soft enough to keep the Fed on hold, or is core inflation sticky enough at 3.3% to force nine committee members into a hike? Warsh's answer, delivered from Sintra with the polish of a man who has done this before because he has, literally, done this before: both, neither, ask me in September. That is not a reaction function. That is punting with better tailoring.
Action items for the book:
Stop trading the dot plot. There isn't one. Trade the minutes as if they are the only Rosetta Stone you're going to get until the July 29 meeting, because they are. Whatever hawkish language shows up describing the case for a September hike — and it will show up, because nine members believe it — that language moves real yields, and real yields are the only lever gold actually respects. Every basis point up on the 10-year, which is sitting at 4.48% this morning, is a basis point of opportunity cost against a metal that pays you nothing. Gold's H1 underperformance wasn't about geopolitics or de-dollarization stories the sell-side keeps recycling. It was mechanical. Real yields went up, gold went sideways-to-down, full stop.
Do not confuse Brent's retreat into the low $70s with the Hormuz risk actually clearing. It cleared enough to unwind a fear premium. It did not clear enough to remove the premium's existence, which means the next headline out of the Strait reprices oil in an afternoon, and Warsh's committee — the same one debating whether "elevated uncertainty" tied to Middle East conflict belongs in the statement — knows this better than any of us.
PepsiCo reports Thursday before the open, EPS consensus near $2.19, with multiple desks trimming targets into the print. Treat it as the actual data point Warsh claims to want: a read on whether elevated prices are still extractable from a consumer that just produced the weakest payroll number in recent memory. If PEP misses on volume rather than just pricing, that's the soft-landing thesis getting a crack nobody's pricing yet. Delta follows Friday, and airline capacity discipline into a soft consumer quarter tells you more about margin compression across the corporate complex than anything the Fed publishes this week.
On the tech rotation: Nvidia and Broadcom are eating outflows to start July, and before you read that as an AI capex story cracking, remember that a 10-year at 4.48% makes every long-duration growth name more expensive on a discount-rate basis, independent of anything happening in HBM supply or hyperscaler capex guidance. Don't let macro plumbing masquerade as a fundamental thesis change. Check both.
The through-line here isn't complicated, even if Warsh is working overtime to make it feel that way. A chair who won't forecast is still forecasting — he's just outsourced the job to us, in real time, with our own capital, and calling it central bank credibility. Nine members disagree with each other about the future. The chair disagrees with the premise that anyone should know it in advance. Position for the minutes like they're the only adult in the room who's going to say something concrete today, because structurally, they are.
Don't trade the vibe. Trade the transcript.
