As absolutely everyone who has been following the markets knows, the stock of VA Linux, a builder of powerful Intel-based servers tailored to run the Linux operating system, skyrocketed on Thursday to close $239 a share, an astounding 698 percent gain on its first day of trading. Early in the day, it went as high as $320.
Investors, including a few lucky E-Trade account holders, who got shares at the offering price of $30 a share whooped with joy. Meanwhile, other observers gasped in horror. Stock message boards on discussion sites like Raging Bull and Silicon Investor were filled with posts expressing astonishment that anyone would pay this price for a hardware company with $17 million in sales last year.
And, indeed, there is good reason to think that buying into a super-hot IPO, however promising the company, is not a great idea. Studies have shown that explosive IPOs often under perform the market. Such stocks tends to rise sky-high, and then drop. Often it keeps falling for a very long time. Of 15 IPOs that doubled on their first day in 1995 or 1996, nine are now below their first-day closing prices, according to data compiled by Professor Jay Ritter of the University of Florida. The Globe.com, the previous record holder for price run-up in the first day of trading, went public at $9, opened at $87 and closed at $63.50 (all pre-stock-split prices); it now languishes just above $10 a share.
By Friday afternoon, anybody who had bought VA Linux at $320 had already lost $102 a share, as the new issue closed the week at $218.
But there's something missing from this picture. Everybody knows that IPOs that run up to astounding heights lose a big part of their value within days. So who are the investors naive enough to buy into them?
Conventional wisdom holds that these stocks are bought by small investors who are willing to buy into a good story at any price. According to this theory, the buyers for VA Linux are true believers who think that Linux-based computers will eventually displace the Microsoft monopoly.
In this case, the conventional wisdom seems wrong. In scouring hundreds of messages on investor bulletin boards, I did not find one person who admitted to buying VA Linux at $320. This is significant, because investors who like a stock tend to be vocal about it. Many will happily go on Internet message boards to defend their purchases - in part because the buzz helps keep up the price of their holdings. If individual investors are not out in force defending the companies prospects, you have to figure they're not holding the stock.
There is another possible theory, and that is that the investors who bought VA Linux the moment it opened did so largely by accident. Something like that happened in November 1998 with TheGlobe.com. Naive individual investors put in orders to buy the stock before it opened without specifying a "limit," or top price. The next day some of them were shocked that instead of buying the stock at a little bit over the IPO price, their trades had been executed at $90 a share.
In this case, I don't think that happened, either. After TheGlobe.com went public, many brokerages instituted tighter control to make sure this didn't happen again. Most will now not let investors ask to buy a new issue without putting in a limiting price.
So who was it who bought VA Linux at these astonishing pricing?
There's one theory that has gotten a lot less notice than it should. Earlier this year, TheStreet.com's Cory Johnson reported that investment banks were asking institutional investors who got shares in hot IPOs to buy additional shares after trading opened. In other words, a large mutual fund might get 30,000 shares of a hot stock at the low offering price - but only if fund managers indicated they would buy 60,000 more shares after the stock opened for trading.
"It might not be stated explicitly," says Johnson, "But investment banks will give shares to funds that will buy more shares in the after-market and support the stock."
Why would investment bankers want institutional investors to do this?
Simple. It assures that when trading opens there will be significant demand for the shares. That keeps the price up, gives company insiders a healthy profit, and makes the investment bank that underwrote the offering look good. (It's the investment bank's job to generate investor interest, and there's no better proof of investor interest than a blockbuster first-day opening price.)
Now look at it from the point of view of the mutual fund or other large institutional investor that might have purchased these shares. Let's say Bigshot Internet Fund got 10,000 shares of VA Linux at $30 a share. Then let's say it bought 20,000 shares more at $280 a share. The stock ended the day $239 - let's say $240 to keep the math simple.
Here's the final result. On paper, Bigshot Internet Fund made $210 on each of the shares it got at the offering price. It lost $40 on each share it bought at $280. That's a $2.1 million paper gain, and an $800,000 paper loss - a total profit of $1.3 million. So it works out well for everyone involved. The company has a great opening day, the investment bankers look good, and Bigshot Internet Fund still makes plenty of money.
The funny thing is that there's a very good chance that Bigshot Internet Fund sold its low priced initial allocation. In other words, Bigshot Internet Fund could be both buying and selling stock in VA Linux.
I can't prove that's what happened. But right now it sounds like a better explanation than the idea that there are tens of thousands of small investors putting lots of money into a stock that has jumped above $200 - if only because I can't find those naive investors.
I like this theory, also, on epistemological grounds. The market is a lot more sophisticated than it is given credit for being. It is generally a bad idea to assume that investors who buy a stock do so because they are simply silly or inexperienced. Sometimes, of course, that is the case, but more often it turns out that the market behaves quite rationally, and it is only the observers who are naive.