When the next round of finance texts is written, the American dot-com bubble of the late 1990s is sure to take its place with the classics--the tulip bubble, the South Seas bubble, the run-up to the great crash of 1929. But what caused it? According to one theory, the problem was a shortage of short selling, say Wharton finance professors Christopher C. Geczy and David Musto.
But what caused it?
Bubbles can rise out of investors' unbridled enthusiasm. But the potential for irrational exuberance is always present, and countervailing forces usually bring it under control. Why did common sense, sound analysis or betting against the trend fail to curb the enthusiasm in this case?
According to one theory, the problem was a shortage of short selling, says Christopher C. Geczy, a finance professor at Wharton. The story suggests that "if you can't short, and there's no one to force the price down in that way, what's going to happen?" he says, explaining the theory. "Well, the price would stay high."
He and two colleagues--Wharton finance professor David K. Musto, and Adam V. Reed, finance professor at the Kenan-Flagler Business School at the University of North Carolina at Chapel Hill--examined the market to see if this was so. The result is their article in the Journal of Financial Economics entitled, ?Stocks are Special Too: An Analysis of the Equity Lending Market.?
In the process, they looked at several theoretical trading strategies involving pairs of long-short bets to determine whether the textbook profit predictions would work in the real world. Their finding: Strategies such as buying value stocks and shorting growth stocks could indeed generate profits, despite the costs incurred. "We found it didn't cost so much, so that you could capture that (profit)," Geczy says.
Short sellers borrow shares, sell them and hope to make a profit when the stock price falls and they can replace the borrowed shares with others bought more cheaply. Short sellers must analyze stocks just as carefully as ordinary investors who buy stocks, or go "long," in hopes that prices will rise. When short sellers flock into a stock, their belief it will fall helps dampen the market consensus and prevent runaway enthusiasm.
Geczy says some observers speculate that short sellers were unable to perform this function during the dot-com bubble because they could not find enough shares to borrow. Since many of the dot-coms were young companies that had only recently gone public, large blocks of shares remained in the hands of founders and other insiders, according to this theory. In order to keep share prices rising, insiders may have hoarded their shares, denying them to short sellers and perpetuating the bubble.
|In order to keep share prices rising, insiders may have hoarded their shares, denying them to short sellers and perpetuating the bubble.|
But was this theory true?
To test it, the researchers obtained a year's worth of short-sale data--from November 1998 through October 1999--from a large custodian bank, which they agreed not to identify. These banks hold shares for the accounts of institutional investors such as mutual funds, pension funds and hedge funds. Like an ordinary bank that uses customers' deposits to make loans, custodian banks loan shares of stock to brokers whose customers wish to borrow the shares or directly to short sellers.
The researchers found that, in fact, the bank had made extensive loans of shares in the dot-com stocks. Short sellers, it turns out, were able to obtain the shares they needed in order to bet against the continuing rise of dot-com stocks. Hence, the bubble was what it had appeared to be--a result of investor enthusiasm gone wild rather than market manipulation by insiders.
"It doesn't seem that (dot-com) prices were high because investors couldn't short them," Geczy notes. By several measures, there was plenty of available short exposure to dot-coms and, for that matter, recent initial public offerings.
The bank data also enabled the researchers to determine whether borrowing costs would be so high as to wipe out the profits that theorists have said could be made with certain paired long-short bets.
One of those strategies would involve buying a basket of value stocks and shorting a basket of growth stocks. Historically this has earned profits because value stocks tend to provide slightly higher return than growth stocks. In a rising market, the value stocks on the long side of the bet would make money, while the growth stocks on the short side would lose money. But since value stocks have slightly higher returns, they would make slightly more than the growth stocks would lose. In a falling market, the trader would lose slightly less on the value stocks he owned than he would make on the falling growth stocks he was short.
Paired bets are used by professionals, such as hedge fund managers, to filter out effects of the market as a whole. By going long and short, the trader can, in theory, make money regardless whether the overall market goes up or down--so long as the relationship between the two baskets of stocks sticks to the normal pattern.
The researchers found that this value-growth strategy could have earned the trader 30 to 50 basis points per month, or 0.3 to 0.5 of a percentage point, assuming there were no costs involved, Geczy says.
But what would happen in the real world where some profits would be given up to "friction," such as the payments to the bank for borrowing shares? Using the bank data, which included the short-selling lending fees, they found that most of the profit would be left even after these costs were paid; friction might be as low as 3 basis points a month. "We found that that the shorting costs were just not much of an issue," Musto points out. "You can get at least some of the profits of these long-short strategies even net of all these short-selling frictions."
Several other strategies also had a high probability of producing profits in the real world. Those include ways of capturing differences in the performance between large stocks and small ones, between high-momentum and low-momentum stocks, and between stocks in initial public offerings and the market as a whole. IPOs, for example, have a history of falling in price after they enter the market--and of falling further and faster than the overall market tends to rise. Hence, the trader could buy a basket of stocks representing the market and short a basket of IPOs.
|The researchers found that this value-growth strategy could have earned the trader 30 to 50 basis points per month, or 0.3 to 0.5 of a percentage point, assuming there were no costs involved, Geczy says.|
Of the strategies examined, the only one that appeared to have diminished profits strongly was merger arbitrage. Historically, target companies' prices rise during mergers, and acquirers' prices fall. So this strategy calls for buying shares of target companies in mergers and shorting shares of the acquirers. The researchers found, however, that there were difficulties in finding shares of target firms to borrow, making this strategy much less profitable than would seem possible on the face of things.
In each, case, the research looked not only at how institutional investors would fare but at the outcomes for individual investors who would not get the volume discounts enjoyed by the pros. Even individuals often could make these bets profitably, they found.
The research looked at market-wide results and did not determine how many stocks a trader would have to own or short in order to match the results.
But Musto points out that the group took a conservative approach that assumed the trader was simply buying or shorting all stocks available at the bank. A real trader might well do better by buying just the stocks that seemed to have the best prospects for gains and shorting only the ones most likely to fall.
Also, he says, the period studied predated the rise of low-fee exchange-traded funds, or ETFs, that make it easy and inexpensive for ordinary investors to bet on baskets of stocks. ETFs are similar to mutual funds but trade throughout the day like stocks. And ETFs can be shorted, while ordinary mutual funds cannot.
Of course, there are no guarantees that the historical relationships in these strategies will hold. In fact, the historical outperformance of small stocks versus large ones has not held in recent years, Musto says.
And there is no guarantee that overpriced stocks will fall quickly enough for short-sellers who ultimately have to pay off their loans, regardless of which way the stocks move. Indeed, short sellers lost fortunes during the late '90s, when irrational exuberance continued to drive stocks upward.
To read more articles like this one, visit Knowledge@Wharton.
All materials copyright © 2002 of the Wharton School of the University of Pennsylvania.