Warsh's First Test: Is the AI Boom 1996 or 1999?
Alina Collins
The new Fed chair's defining challenge is not whether to raise or cut rates — it is whether the current AI boom resembles 1996's benign expansion or 1999's pre-bubble peak. That single judgment will set his entire policy path.
One decade, two opposite endings — why does it matter?
Warsh keeps citing the 1990s as his historical anchor, but that decade contains two completely different scripts.
1996: Greenspan held steady through rapid growth, betting expansion would not ignite inflation. He was right — growth continued for years and earned him the title "the Maestro."
1999: stocks surged, the labor market kept tightening, and Greenspan shifted to consecutive rate hikes — ending with the dot-com crash. This means → the same person, in the same decade, reached opposite outcomes because he read the nature of the boom differently.
What exactly is Warsh betting on?
Warsh's logic boils down to one claim: AI-driven productivity gains have not yet shown up in official data and may take years to appear.
In plain terms = if the Fed insists on waiting for hard data before standing pat, it will misread healthy growth as overheating, hike rates, and kill the very expansion that could have held inflation down.
The bet, at its core, is to replace lagging data with a forward-looking narrative as the basis for policy. Asked whether AI is currently boosting demand or expanding supply, Warsh said only that "demand is easier to measure" — deliberately avoiding a direct answer.
The pushback: could anticipated gains actually trigger overheating?
Chicago Fed President Austan Goolsbee offered the most systematic challenge: whether a productivity boom lets the central bank stand pat depends on whether the boom is a surprise.
This means → a boom everyone can see coming is actually more dangerous — people spend against future wealth before the productivity gains arrive, and the economy overheats prematurely.
Goolsbee pointed to real-world signals: AI data-center construction is pushing up land, power, and chip prices, squeezing resources from other sectors. Apple's price hikes this week, driven by rising costs, are evidence this mechanism is already operating.
Does the "they can't borrow to front-run" rebuttal hold?
Fed Governor Christopher Waller countered Goolsbee at the same Stanford conference: the front-running mechanism only works if people can borrow to spend early.
In plain terms = many households are constrained by current income — if they cannot borrow, they cannot front-run, and the entire "anticipation triggers overheating" logic chain breaks.
This reflects how deep the internal split runs: the latest dot plot shows nearly half of officials expect a rate hike this year, while the rest expect the opposite — there is no consensus even on direction.
The tool Warsh most wants to scrap may be the one he needs most?
Warsh has been explicit: he wants a Fed that "doesn't show its hand early," scaling back forward guidance — the practice of signaling rate moves in advance.
The irony is that forward guidance was built in 1999 by Greenspan himself — precisely to keep markets from being blindsided.
This means → if the economy follows the optimistic script, Warsh never needs the tool. But if it follows the other script, he faces a dilemma: either revive the very convention he wants to abolish, or stay silent and risk a violent market reaction. Put simply = the answer still comes back to the same question: is this 1996, or 1999?
Content is for reference only, not financial advice.