The 1974 Desk

Eight-thirty is fifteen minutes away and I already know what I'm going to write in the log, because I wrote a version of it in 1974, except I didn't, because I wasn't alive, and neither was almost anyone currently employed on a trading floor. That's the part that should worry you more than the number itself.

Yesterday's producer price index came in at 5.4% year-over-year — the hottest twelve-month print of 2026, up from 4.8% in July, with diesel alone up 24.1% on the month and doing more than a third of the work in the goods basket. Core PPI, stripped of food and energy, still ran 4.6%. Treasury yields jumped on the release. Stock futures dropped. And this morning, with Brent trading above $105 and WTI not far behind on the seventh month of an active U.S.-Iran war, the August Consumer Price Index lands at 8:30, five days before Kevin Warsh's Fed decides whether "not broadly restrictive" gets converted from a Jackson Hole talking point into an actual quarter-point.

I keep reaching for Arthur Burns because Arthur Burns is the only precedent that fits, and precedent-fitting is a dangerous game when the fit isn't clean. Burns inherited a Fed in 1970 committed to full employment as a policy input, watched OPEC's embargo quadruple oil prices in 1973-74, and hiked into a supply shock anyway, believing — correctly, as far as it went — that a central bank cannot let energy-driven inflation de-anchor expectations, and incorrectly, as it turned out, that hiking into a shock the Fed didn't cause would fix a shock the Fed couldn't reach. He got both stagnation and inflation for his trouble, and PPI didn't just touch 5.4% that decade, it touched 19.6% in November 1974, a number so absurd it makes today's print look like a rounding error. It wasn't a rounding error. It was the same mechanism running at lower amplitude: an energy shock working its way backward through the production chain into everything downstream of a diesel truck.

What Burns didn't have, and what should make you slightly more comfortable and slightly more nervous in equal measure, was a Volcker waiting in the wings with the political cover to break the back of the thing properly in 1979 — cover that came only after years of Burns-era credibility erosion made the pain politically survivable. Warsh doesn't need to invent that credibility from scratch. He's spent his four months in the chair doing almost nothing but building it: no August meeting, a deliberately thinned press calendar, a Jackson Hole speech that told markets, more or less, to stop looking to him for their next trade. Whatever else you think of the man, he is running the opposite communications strategy from a Fed chair drowning in a decade of forward guidance addiction. He wants distance from the tape. Burns wanted to be loved by it.

But distance from the tape doesn't buy distance from the Persian Gulf. That's the actual structural problem sitting underneath this morning's print, and it's not one Warsh's reticence solves. The Fed's reaction function has, in practical terms, been subcontracted to a shipping lane it doesn't control and a war it didn't start. CME FedWatch has September hike odds sitting around 73% this morning, up from the high 30s a month ago, and every basis point of that repricing traces back through diesel and jet fuel and gasoline futures to a conflict whose resolution timeline nobody at the Eccles Building can forecast with a straight face. You can build all the institutional credibility you want. You cannot hike your way into closing the Strait of Hormuz, and you cannot cut your way into reopening it either.

Which is what makes this CPI print genuinely load-bearing rather than merely scheduled. Consensus sits around 3.4% headline, 2.4% core, both marginally better than the PPI omen would suggest — but PPI leads CPI by design, not by accident, and a 5.4% wholesale number sitting five days before a rate decision is not the kind of thing a soft consumer print quietly absorbs. Bitcoin already knows this. It fell through $78,000 into the high $76,000s overnight on the PPI release alone, trading like a rate-sensitive risk asset rather than the inflation hedge its own mythology insists it is — proof, again, that the "digital gold" narrative evaporates precisely when the macro backdrop would seem to demand it most.

And then there's Oracle, reporting last night into all of this as if none of it were happening. Cloud infrastructure revenue up 121% to $7.4 billion against a $7.19 billion estimate, the kind of beat that in June would have bought the stock another leg of the 2025 melt-up it's spent all of 2026 giving back. Instead it walks into a tape where the 10-year sits near 4.85%, its highest since November 2023, discounting every future dollar of that AI infrastructure buildout at a rate the market didn't have to use twelve months ago. Oracle's backlog — $638 billion, 9.5 times projected revenue — is a bet on a decade. The Fed decision in five days is a bet on a quarter. Those two clocks are not running at the same speed, and nobody has reconciled them, they've just been trading past each other since June.

Burns eventually got his lesson delivered by Volcker, at the cost of a recession that took the funds rate to 20%. I don't think Warsh needs that lesson delivered to him — he's arguably already internalized it, which is the entire point of the credibility-building exercise. What I don't know, and what nobody staring at an 8:30 release this morning actually knows, is whether internalizing the lesson from 1979 helps you at all when the shock forcing your hand isn't domestic monetary excess but an active war measured in tanker strikes. Burns fought the last war. Warsh may be fighting the right one and still lose, for reasons that have nothing to do with monetary policy at all.

Eight-thirty. Here it comes.