The Narrative Life of a Bubble
How do narratives about a bubble shape the bubble itself?
In the winter of 1996, Alan Greenspan coined a term: irrational exuberance. It “came to me in the bathtub,” he later recalled of the moment he scribbled notes for a speech he would deliver to the American Enterprise Institute, a conservative think tank. Making his way from the bathtub to the ballroom, Greenspan, then chair of the Fed, went on to deliver a speech that suggested that “irrational exuberance has unduly escalated asset values.”
From Tokyo to New York, markets tumbled in response. But it wasn’t the kind of market correction that Greenspan had hoped for with his carefully worded “Fedspeak.” Over the following months, markets didn’t just rebound—they exploded. By July of 1997, the Dow Jones Industrial Average had reached 8000 while the Nasdaq broke 1,500, both record highs. Even with this widespread sense of optimism, though, few believed Lazlo Birinyi Jr., a global strategist for Deutsche Bank Securities and the owner of an investment research firm, when he predicted the Nasdaq would reach 2,800 by the end of 1999. When it climbed to 4,000, Bloomberg journalist Anthony Bianco described him as “The Prophet of Wall Street.”
When the market finally crashed in March 2000, destroying over $5 trillion in market value, the only narrative that seemed to hold sway was Greenspan’s “irrational exuberance,” a term that economists such as Robert Shiller later elaborated in his book of the same title. For Schiller, who had also testified before the Federal Reserve Board two months prior to Greenspan’s speech, explaining that “market levels were irrational,” the term “irrational exuberance” meant “extravagant expectations.” It was “wishful thinking on the part of investors that blinds us to the truth of our situation.”1
The Dot-Com Bubble, the historical episode that materialized in this new term and reshaped our thinking about financial crises, has gained relevance. It has become a reference point in an ongoing conversation about the AI Bubble, with some seeing an uncanny relationship with this moment of irrational exuberance to the “extravagant expectations” in the potential of AI. Even Sundar Pichai, the head of Alphabet, said there were “elements of irrationality through a moment like this.” Since his statement back in November 2025, financial observers have continued to weigh in on the issue. William Janeway, professor of economics at the University of Cambridge, thinks that “financial speculation is outpacing productivity gains.” Financial observers such as Bill Conerly, meanwhile, believes the “underlying fundamentals are extremely positive for artificial intelligence companies.”
While there seems to be a growing consensus that there is, in fact, a bubble, these kinds of observations create more questions than answers. What we do know for sure, though, is that there is a growing narrative about an AI Bubble. And if narratives, as a growing number of economists suggest, shape expectations about the future—which, in turn, shape behavior—then how is all this bubble talk shaping the economy? To a degree, maybe all this bubble talk doesn’t just describe the current economic conditions. Maybe it plays an active role in shaping perceptions of risk, expectations, and economic behavior.
It might also be interesting to consider how much bubble talk impacts expectations and behavior at different stages of the narrative. For instance, when bubble talk goes mainstream, with many different journalists and other financial observers pointing to the possibility at once, it may have a more significant impact on expectations than later stages of the narrative. In contrast to the early stage, when there may be a greater sensitivity to risk, investors become desensitized to the narrative in the later stages. It becomes background noise, making the prospect of a bubble seem less urgent. Where is the inflection of bubble talk, when investors become the most sensitive to the narrative? When do warnings about excess begin to alter behavior—and when do they simply fade into background noise? If economic expectations are shaped by stories, then the story about an AI bubble may matter as much as the bubble itself.
Robert J. Schiller, Irrational Exuberance (Princeton, NJ, 2000), xii, xvi, 4.