For about two years, a public company could add a word to its name and watch its stock triple. Most of what moved prices in the dot-com bubble was not a crime; it was a story the market wanted to buy, sold by people who were paid when it bought. Here is how the machine worked, what it did to the companies that rode it, and where it ended.
The noun did the work
In 1998 and 1999 the cheapest way to move a stock was to say "internet". A study of firms that added ".com" or a web term to their names found about a 74% abnormal return in the ten days around the announcement, whether or not the business did anything online [1].
K-tel, best known for compilation records sold on late-night TV, said it would sell music online; its stock went from about $6.63 to about $28.69 in a week [2]. The bookstore chain Books-A-Million announced an "enhanced" website and traded from $3.06 to around $47 intraday within days, then back to $16 inside a week [3]. A fish-meal company, Zapata, planned to rename itself "Zap" and ran newspaper ads offering to buy people's websites [4].
The first day was the product
New listings were priced low and opened high, and the first-day "pop" became the headline. The early social network theglobe.com rose 606% on its first day of trading in November 1998, on almost no revenue [5]. A year later VA Linux, a maker of Linux servers, priced its shares at $30 and closed its first day near $239: up 698%, the biggest first-day gain on record at the time [6].
Who got those shares at the offer price mattered. Regulators later found that "hot" IPO shares had been handed out improperly at several large banks, among other conflicts [7].
Everyone was a Strong Buy
Research analysts were supposed to be the referees. At the big banks they were also part of the sales team: a bank won a company's stock offering partly on the promise of glowing coverage afterwards. In early 2000 the large majority of covered stocks were rated Buy, and less than 1% were rated Sell [8].
In April 2003, ten of the largest Wall Street firms settled charges that investment banking had held "inappropriate influence" over their analysts from about mid-1999 onward. They paid roughly $1.4 billion in penalties, disgorgement and funding for independent research, without admitting or denying the allegations [7].
Losing money on purpose
Profit was optional; growth was not. Companies were valued on "eyeballs", visitors and sign-ups, and spent heavily to buy them. Pets.com shipped heavy bags of pet food below cost and bought a Super Bowl ad; it was liquidated 268 days after its February 2000 IPO [9]. Webvan promised grocery delivery in 26 cities, ordered a billion dollars of warehouses, was valued at more than $4.8 billion at its November 1999 IPO, and was bankrupt by July 2001 [10].
How it ended
The warning came early. In December 1996 the chairman of the Federal Reserve asked out loud how anyone would know when "irrational exuberance has unduly escalated asset values" [11]. The market went up for more than three more years.
The Nasdaq Composite, home of most of these stocks, peaked in March 2000. By October 2002 it had fallen about 78%, and it did not get back to its 2000 high until 2015 [12]. theglobe.com was delisted in 2001 [5]; VA Linux fell to about a dollar within two years [6].
The game on top
Synthetic Demand is a satire of exactly this. You are the CEO of an invented company and you have one job: make the stock go up before you sell. A melt-up like 1999 is one of the markets you can be dealt, though nobody tells you the year until it is over. You can rename the company, tease a product, hire a promoter or simply say the right noun, and the tape answers.
Every call you make, and every one you don't, is graded A to F. The good grades go to CEOs who read the room. Nobody in 1999 did.