MACKINTOSH AT HIS LEVEL BEST!
Why Bubbles Can Keep Inflating in Plain Sight
ET
Everyone’s talking about an AI bubble. For the bulls, that is proof that there is no such thing. If, the argument goes, people really believed prices were wildly overinflated, they would sell and it would pop. How can there be a bubble if everyone knows it?
As a natural contrarian, I like the idea. It bothers me when I’m part of the crowd. But it is simply wrong to think a bubble can’t continue to inflate amid widespread warnings. History proves that—and psychology and financial theory explain why.
The dot-com era offers the most obvious case of a bubble that kept inflating amid cries of excess. I suppose some die-hards might actually have thought that adding “.com” to a company name justified an average 74% rise in the stock over the next week and a half or that it really made sense to value companies by measuring their price per online click.
But there was no shortage of very loud warnings that a bubble was under way. This newspaper (and every other respectable journal) was full of articles in 1999 comparing the boom in internet stocks, many with zero revenue, to tulips and the South Sea Bubble. Big fund managers’ warnings about speculative froth were frequently quoted. Yet stocks just carried on up—until they didn’t.
The reason is that lots of people were buying shares not because they thought the prospects for the company were good or the clicks would one day turn into revenue, but rather because they saw their friends getting rich and wanted some of it.
Charles Kindleberger, the great historian of financial crises, summed it up neatly: “There is nothing so disturbing to one’s well-being and judgment as to see a friend get rich.” In 1999, anyone who secured an allocation of stock in a dot-com IPO instantly made big money. The average first-day return on IPOs that year was above 70%, by far the highest in data back to 1980 compiled by Jay Ritter, professor emeritus at the University of Florida.
SHARE YOUR THOUGHTS
Do you think the market is experiencing an AI bubble? Join the conversation below.
Yet, most people bought after the first-day pop because they didn’t get an allocation. Why did they buy? Not because these were great companies. Many looked as though their business plans were drawn up on a napkin and their stocks rose faster the more rapidly they burned through their cash.
Instead, it was a mix of companies selling few shares and the greater fool theory: Someone else will be willing to pay more tomorrow, so even though the stock is obviously overpriced, it is still worth buying now. In finance, this goes by the staid name of a “rational bubble.”
Bubbles are typically made worse by the breakdown of standard finance. Usually, stocks are kept within reasonable limits not only by owners who choose to sell when the price gets too high, but also by hedge funds and others shorting the stock, borrowing it to sell in hopes of buying it back more cheaply after it falls.
However, dot-com stocks were hard to short because there was little available to borrow. Even if it was possible to borrow the stock, it could be suicidal given the size of moves—Amazon.com, for example, had eight days of 10% or more gains in the final six months. Anyone caught short on a day like that risked being crushed.
Artificial intelligence is a little different, but a lot the same. The biggest name in AI today is OpenAI, developer of ChatGPT, but it is impossible to bet against it because it is private. That surely made it easier for OpenAI to reach an extraordinary valuation of more than 38 times revenue.
Finally, professional fund managers find it hard to sit on the sidelines. Once a bubble develops, it can go on for years. Betting against it, or even just avoiding it, means lagging behind other investors and the benchmark, and risking being fired as clients abandon the underperforming fund.
This career risk led to one of the biggest ironies of the dot-com bubble when Tony Dye, head of Phillips & Drew Fund Management, part of UBS, was ejected in March 2000 after avoiding tech stocks and thus falling behind rivals’ returns. PDFM went on to be a standout performer that year, as the bubble burst the very month he left. The lesson for fund managers was clear: Your job is more secure if you stick with the crowd, even if you end up losing money when everyone else does.
Of course, there is still the question of whether there is an AI bubble. Big Tech stocks are extremely expensive but have been for years.
If OpenAI quickly comes up with a vital service everyone proves willing to pay big bucks to use, maybe even its price can be justified. After all, the only absolute proof of a bubble comes when it bursts.
Write to James Mackintosh at james.mackintosh@wsj.com
Copyright ©2025 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8
Appeared in the October 25, 2025, print edition as 'Why Bubbles Can Keep Inflating in Plain Sight'.
Comments
Post a Comment