WASHINGTON – The man who taught Kevin Warsh what it means to run the Federal Reserve died last week, and the timing was almost too pointed to be accidental. Alan Greenspan, who steered the central bank through the 1990s technology boom without raising rates to choke it off, died on June 22 at the age of 100, leaving behind a legacy that is now a direct instruction manual, and a warning, for his self-declared successor.
Warsh invoked Greenspan by name at his own swearing-in ceremony in the East Room last month. “Chairman Greenspan was the first to tell me and show me what this role demands,” Warsh said. The tribute was personal. It was also a policy statement. Warsh has argued throughout his career and his confirmation process that artificial intelligence will be as disinflationary as the internet was in the 1990s, that its productivity gains will eventually show up in the data, and that the Federal Reserve should therefore resist the temptation to raise rates into what looks, from certain angles, like a classic speculative boom.
The op-eds from The Economist, Bloomberg Opinion, The New York Times, and Barron’s that flooded in after Greenspan’s death are not just obituaries. They are, in aggregate, an argument about whether Warsh is right. Greenspan’s bet on technology in the 1990s worked, and the productivity data eventually vindicated him. His second major bet, that financial innovation had made markets self-correcting and that the Fed could afford to look away from the housing market, did not. Millions of people lost their homes. The 2011 Financial Crisis Inquiry Commission named his failure to discourage subprime lending as one of the triggering causes of the worst economic crisis since the Great Depression.
The question Greenspan’s death forces into the open is which version of the lesson Warsh has actually absorbed. The Economist framed it as a problem of politics: the Fed chief must practise politics, the magazine argued in a piece published the day after Greenspan died, because no central banker survives by reading models alone. Bloomberg Opinion put it more directly, noting that Greenspan deflated one bubble in the end, his own authority, by staying silent on the housing market for too long. The New York Times and Barron’s both zeroed in on what Greenspan himself wrote in a joint op-ed with Warsh: that AI is not simply another technology cycle but a structural transformation that changes what the Fed’s models can measure and therefore what the Fed’s decisions should be.
That structural argument is what makes the current moment genuinely difficult to read. Greenspan’s productivity bet in the 1990s was controversial at the time, but the data, eventually, arrived. GDP growth ran above 4 percent. Unemployment fell to levels that economists said were impossible without igniting inflation. Inflation stayed low. The critics were wrong, and Greenspan’s willingness to trust the technology story before the statistics confirmed it looked, in retrospect, like genuine insight rather than recklessness.
The AI case is more uncertain. Productivity growth in the United States has been running above its post-2008 trend, but not dramatically so. The PCE price index hit a three-year high in May, driven partly by oil price pressure and partly by AI infrastructure costs that are themselves inflationary in the near term even if they prove disinflationary over the long term. Asset prices in AI-adjacent equities have run far ahead of earnings. The S&P 500 is trading at multiples that, on two prior occasions in the last thirty years, preceded significant corrections. None of this proves a bubble. It is also not the clean productivity story that Greenspan had by 1998.
Warsh’s response to this ambiguity has been to say less rather than more. His first post-meeting statement ran to 130 words, the shortest in years, a deliberate echo of Greenspan’s own sparse communications style. He has pulled back forward guidance, stopped telegraphing rate decisions in advance, and announced five working groups to study how AI is changing the data the Fed uses to understand the economy. The implicit message is that he does not yet know what the right answer is and is not going to pretend otherwise.
There is a reasonable case that this is exactly the right posture. Greenspan’s most damaging mistake was not that he missed the housing bubble but that he had become so convinced of his own framework, that financial innovation had made markets more stable, that he stopped interrogating it. The working groups Warsh has assembled represent the opposite instinct: an acknowledgment that the Fed’s existing models may not be adequate to an economy where AI is changing costs, labor markets, and investment patterns faster than the data can track.
The risk is different but symmetrical. Greenspan’s productivity bet in the 1990s worked because the technology eventually delivered. If AI delivers at the scale its proponents claim, Warsh’s patience will look prescient. If AI productivity gains prove slower to materialize than the asset markets are pricing, or if the current boom is being financed by the same kind of loose credit conditions that inflated the housing market, then Warsh will be left explaining why he watched it happen. Nobody, as Bloomberg’s obituary noted dryly, can reliably call a bubble in real time. Greenspan’s own “irrational exuberance” speech came in December 1996. The dot-com crash did not arrive until 2000. He was right, three years early, and the market mostly ignored him.
The deepest irony in Greenspan’s death is that it arrives at the moment when his approach is most fashionable and most contested simultaneously. Warsh has inherited an economy that does not look like the 1990s in the ways that matter most: above-target inflation for five consecutive years, a fiscal deficit running at wartime levels, a geopolitical environment that is actively disrupting the supply chains that kept goods prices low through the Greenspan era, and an AI boom whose productivity dividend has not yet shown up clearly in the aggregate data.
What Greenspan actually bequeathed to Warsh, beyond the policy philosophy, is the knowledge of what it costs to be wrong about a technology cycle. The 1990s bet paid off. The 2000s bet on financial self-correction did not. Warsh has said he intends to fill this role with energy and purpose, just the way Greenspan did. The question he has not answered, because no one can answer it yet, is which Greenspan he is actually prepared to be.
