Wharton Impact

Researchers have shown that many initial public offerings start out well, but then later stock prices slump once investors realize that the companies are weaker than they expected. A new study reveals that if stock offerings are overpriced, a major reason may be that sell-side analysts are too optimistic in their predictions about these companies’ earnings growth.

When Yahoo!, an Internet search-engine company, went public this year, its stock at first seemed to live up to its name. The initial public offering (IPO), managed by Wall Street’s well-known Goldman Sachs, did phenomenally well, and Yahoo!’s stock, priced at $13, soared to $43 on the first day of trading. Soon, though, Yahoo! lost its sheen. By August the stock had slumped to $18; in early December, it was trading in the mid-20s. And Yahoo! is scarcely the only Internet company to have taken this roller-coaster ride.

Other Internet companies that went public during the past year have also seen their stocks shoot to dizzying heights before gravity yanked them back to earth.

If all this has a familiar ring, it is because Wall Street has been there, seen that, so to speak. Researchers who follow stock movements have shown that many IPOs start out well, but then later slump once investors realize that the companies are not as strong as they expected.

And the phenomenon seems to occur in waves. Today Internet companies are all the rage. Yesterday it was biotechnology firms. Tomorrow … who knows? But you can bet there will be some company in some industry that will be on analysts’ lists of hot tips.

New-Issue Puzzle

Observers have tried to explain why this happens. In academia, they call it the so-called “new issue puzzle.” If markets are efficient, as theory suggests they are—which means investors take all available information into account before they agree to pay a certain price for a stock — why should IPOs perform badly over time? Why do new offerings decline in value over the years and companies that once seemed promising lose millions of dollars in market capitalization?

Some people attribute the trend to normal market risk. Others suggest that perhaps the initial expectations were based on flawed research. At a gut level, though, many have suspected that IPOs get hyped because the analysts who present seemingly objective research about the offerings in fact work for the very investment banks that are trying to market the stock to investors.

That gut feeling now has substance. In a recent paper, Wharton’s Patricia M. Dechow, Anheuser-Busch Term Assistant Professor of Accounting, and Richard G. Sloan, assistant professor of accounting, and Harvard’s Amy P. Sweeney, say that IPOs are overpriced mainly because sell-side analysts are too optimistic in their predictions about companies’ earnings growth potential. And investors, rather than recognizing that sell-side analysts are part of a company’s marketing effort for its stock offering, naively accept these projections, which pushes up the price of the stock.

Dechow, Sloan and Sweeney found that the earnings forecasts of so-called affiliated analysts — analysts who work for the investment bank underwriting the offering — are more biased than those of analysts who work for other investment banks. The researchers say these hyped-up expectations largely explain the long-run underperformance of IPOs.

The researchers based their study on data relating to 7,636 stock offerings made between 1981 and 1990. They obtained information about analysts long-term forecasts as well as the names of the analysts’ employers. After eliminating companies about which they had incomplete information, they worked with a sample of 1,179 offerings.

Next, Dechow, Sloan and Sweeney separated the analysts into two groups: affiliated and non-affiliated. Finally, they compared the actual earnings growth in these companies over five years with the forecasts of the affiliated and non-affiliated analysts. They also adjusted their calculations to account for major market swings. For example, the impact of the 1987 crash that rocked the whole market was factored out of the reckoning.

Wages of Bias

The results were illuminating. On average, the companies’ earnings grew by 5.7 percent a year over five years. Contrast that with the fact that analysts had predicted growth of 16.2 percent a year. Dechow, Sloan and Sweeney show that analysts over-estimated earnings growth by roughly 10 percent a year for the five years following the stock offerings.

The predictions of affiliated analysts were much further off the mark than those of unaffiliated analysts. For example, the forecast error for the entire sample was 10.6 percent—but while unaffiliated analysts erred by 10.3 percent in their predictions, the affiliated ones were off by 14.4 percent.

Disturbingly, these biases worked their way into the prices that investors paid for the stocks. The fact that initial price spurts were followed by slumps suggests that investors, at least at first, bought the analysts’ pitch.

Dechow, Sloan and Sweeney say bluntly: “An important role of the underwriter is to manipulate investors’ expectations of the future earnings prospects of their clients in order to secure a higher offer price.”

What accounts for this bias among analysts? Dechow believes that both affiliated and unaffiliated analysts are reluctant to make negative forecasts because neither group benefits from asking investors to sell the stock. “If they did that, they might lose contact with management,” she notes.

Affiliated analysts face much greater pressure than unaffiliated ones because the investment bank’s underwriting side wants people to buy the stock. “From my understanding of the way these organizations work, the analysts often receive bonuses for bringing a client to the investment bank so that the corporate finance section can underwrite the stock offering,” she says. As a result, affiliated analysts have more incentives not to reveal bad news.

Dechow says while she and her co-authors have not worked on data for the 1990s, she sees no reason why the trend should be any different today. “It’s not as if the Securities and Exchange Commission has come in and said, ‘This has to stop.’ And I haven’t heard any investment banks say they are cracking down on this practice. They have always said that there isn’t a problem here. So I think the trend will continue, because this is where the investment banks make money.”

Market Implications

While the paper does not make any specific recommendations about solving these problems, Dechow suggests a few steps. For one thing, since the forecasts of unaffiliated analysts are less optimistic than those of affiliated ones, it would help if more unaffiliated analysts gave out their forecasts.

Secondly, she recommends more public scrutiny of analysts, so that investors are left in little doubt about their affiliation. “I’d like investors to be aware of this problem, so that if an affiliated analyst is making very optimistic forecasts, they should be wary about believing it,” she says. “I’d like more disclosure. You’d almost want to know if the analyst is getting a bonus or not, or if the analyst’s compensation is somehow tied to how the issue goes off.”

Another key step, Dechow says, would be to change the incentive structure for analysts, “which would mean their pay would be less dependent on investment banking underwriting deals.” Before the 1970s, for example, analysts got most of their pay through the brokerage department’s commissions. But in the 1970s, after the market was deregulated, commissions became a small part of investment bankers’ incomes. Now the money that pays for analysts’ research often comes from the underwriting side of investment banks.

“The problem has become worse than it might have been in the 1970s because of this institutional change and deregulation,” she says.

Dechow and her co-authors plan in the future to look at related issues. For example, they might study what happens if the company making the offering has big institutions among its stockholders. “A lot of companies like to have big institutional holders because it stabilizes their stock,” Dechow says. “Big institutions are also good clients for investment banks, so the investment banks may have less incentive to make these big buyers bear the cost.”

Wharton Magazine - Background

Type to Search

See all results