Takeover discussions took place against an interesting financial backdrop for NCR. NCR had been forecasting robust results for 1991 and beyond. It had told AT&T that it was expecting revenue to hit $6.6 billion in 1991, $7.3 billion in 1992, $10.3 billion in 1995, and $16.84 billion by the year 2000. NCR’s income forecast was equally upbeat: $386 million in 1991, $545 million in 1992, $897 million in 1995 and $1.56 billion in 2000. However, NCR’s 1990 performance had actually been poor. In 1990 NCR posted net income of $369 million, a 10 percent decline from the previous year, on revenue of $6.29 billion, a 6 percent increase. The fourth quarter of 1990 had not ended well. In January 1991, NCR announced earnings of $1.71 per share compared to $2.02 per share a year earlier.
Were NCR’s financial projections well founded? Was 1991 going to turn out much better than 1990? On July 17, 1991, NCR stated that it was unlikely its 1991 results would meet 1990 levels. In this regard, NCR cited several factors, including the U.S. recession. About a month later, the company filed a statement with the SEC, saying that its 1991 revenue, income before taxes, and net income would be “materially below” projections made to AT&T during the merger negotiations.
NCR blamed the change in outlook on the slumping market in Europe and on weakening orders for its products in the Pacific region. It indicated that revenues and income from its new line of high-power computers, the 3000 series, would not begin to flow until 1992. However, NCR put its disappointing results into context by pointing out that other computer makers, such as IBM and Digital Equipment Corp., had experienced similar problems.
AT&T’s reaction to these disclosures was interesting. It stated that NCR’s disclosure had not altered either AT&T’s view of the merger or the price it was paying for NCR. An AT&T spokeswoman was quoted in the September 9, 1991 issue of the Wall Street Journal as saying: “We’re merging with NCR for the long-term strategic fit. Both sides still believe this is a perfect match. We did our homework on this.”10
Think about NCR’s rosy projections through the year 2000. Had AT&T really done its homework? Or did it underreact? Did AT&T ignore disconfirming information about NCR’s financial performance during 1991, thereby succumbing to the illusion of validity? Perhaps. Having committed themselves to the takeover, did AT&T also succumb to loss aversion?
Remember Richard Shaffer’s asking AT&T’s Robert Kavner to name a successful technology merger? Did the AT&TNCR deal turn out to be a major success story, did it turn out like IBM-Rolm, or was it a “calamity” as Charles Exley had warned? In 1991, many speculated that Robert Allen’s legacy at AT&T would be determined by how well the NCR deal worked out. Were they right?
In September 1991, NCR became a unit of AT&T, with a new name—Global Information Services (GIS). The results of 1992 did not match NCR’s forecast. AT&T had not done its homework very well. The plan was for GIS to move away from NCR’s highly focused strategy and become a computer company with broad appeal. This did not work out favorably. Key managers from the original NCR left GIS, and longtime customers felt abandoned. The competitive environment also changed, shifting toward low-margin personal computers and away from the larger systems in which NCR specialized. In July 1995, AT&T announced a layoff plan for GIS and stated its intention to get out of the consumer goods, transportation, and government markets and concentrate instead on the financial, retail, communications, services, and media markets.
Robert Allen’s Legacy
The NCR fiasco did not turn out to be Robert Allen’s legacy. His legacy is the second breakup of AT&T, the largest voluntary corporate breakup in history. In September 1995, Allen announced that AT&T was planning to split itself into three separately quoted companies: telecommunications services, telephone equipment manufacturing, and computing.11 The telecommunications firm retained the name AT&T Corporation, the equipment manufacturing firm took the name Lucent Technologies, and the computing firm was “spun off” under its old name, NCR. (I place “spun off” in quotation marks because it arises as an issue in later discussion about equity carve-outs.)
Between 1993 and 1996, GIS lost over $3.85 billion. The financial press reported that by the time NCR was “spun off” in December 1996, AT&T had lost approximately $7 billion on its investment during the five years AT&T ran the company. From 1993 through 1996, it threw more than $3 billion in good money after bad.
AT&T’s experience with NCR is a dramatic example of winner’s curse, driven by hubris, optimism, the illusion of control, loss aversion, and the illusion of validity. But is the AT&T takeover of NCR case just another interesting story, or does it hold lessons that are more general? Just how typical is it?
Corporate Takeovers and the Winner’s Curse: The General Case
The article best known for arguing that hubris leads acquiring firms to suffer from winner’s curse is by Richard Roll (1993). He argues that one of the first things an acquiring firm does when examining a potential takeover target is to consider the target’s market value. The acquirer then makes its own independent assessment of value, an assessment that would factor in any perceived synergy. If the acquiring firm’s own evaluation lies below current market value, then it regards the target as overvalued and would not choose to enter a bid. But if its evaluation lies above market value, then the acquiring firm may choose to enter a bid. Roll points out that this leaves the acquiring firm open to winner’s curse because it acquires the target only when its estimate is too high.
A central aspect of this argument involves the accuracy of market value. Market value reflects the collective judgment of many investors, not just the judgment of the acquiring firm’s managers. Imagine that this collective judgment is correct and properly reflects all the information known about the target firm, even the information that is held by insiders. In other words, the market is strong-form efficient.12 Roll suggests that we imagine a world where there are no inefficiencies with respect to operations or management’s choice of strategy. He calls this the “frictionless ideal benchmark” from which to judge the evidence. In this type of world, the managers of the acquiring firm incorrectly place greater confidence in their independent estimate of value than in the collective wisdom of all other market participants. They do so because of hubris.
Suppose that the hubris hypothesis is valid, in its strong-form efficient characterization. Think about what would happen to market prices if an acquiring firm made its intentions known through an announcement that had not been anticipated by the market. In this case, the acquiring firm’s actions would cause the price of the target firm’s stock to rise. But because there is no value to the acquisition in this frictionless ideal world, and there are direct takeover costs, rational investors would bid down the price of the acquiring firm’s stock. Why? Because the combined entity is worth less than the sum of its parts. Of course, if the acquiring firm were to abandon its bid, then the price would revert. However, if the bid were to prove successful, then the price of the target firm would rise still further, whereas that of the acquiring firm would decline even more.13
Recall how the market reacted to AT&T’s announcement that it intended to acquire NCR. The price of NCR stock soared, while that of AT&T declined. The combined market values fell slightly. This pattern is consistent with the hubris hypothesis described above. Roll (1993) surveys the evidence about how takeovers affect the market values of acquiring firms, target firms, and the combined entity. Begin with target firms. Michael Bradley, Anand Desai, and E. Han Kim (1983), and Paul Asquith (1983), find that the target’s market value goes up by an average of 7 percent on the announcement of a bid. However, if the bid is withdrawn within sixty days, and no further offer is subsequently made for a year, then market value falls back to a little below its starting point.
Evidence about total gains is difficult to measure from returns, especially when the acquiring firm is a lot larger than the target. The conclusion is ambiguous. We do not really know what happens on average. Paul Asquith, Robert Bruner, and David Mullins (1983) find that bidde
r returns are higher when bidders acquire larger targets, suggesting that this finding is consistent with positive overall gains. Yet Roll is not persuaded by their argument, because their finding could also mean simply that losses tend to be lower for larger targets.
Two other studies suggesting that corporate takeovers generate positive value added are those by Michael Jensen and Richard Ruback (1983) and Steven Kaplan (1989). Kaplan analyzed the largest management buyouts between 1980 and 1986. He finds that three years after a buyout, target firms had big boosts in both income and net cash flow, as well as lower capital expenditures. Shareholders earned an average 42 percent premium gain from the buyout, as measured by the shareprice boost between the price two months before the bid and the price upon completion of the buyout. Kaplan found that the efficiencies came from better incentives for management and workers rather than from layoffs.
Bradley, Desai, and Kim (1982) attempt to measure dollar gains directly, rather than by using returns, and conclude that there is an increase in combined market value (though not by a statistically significant amount). They find that the average amount in their study, $17 million, is split into a +$34 million piece for the target and a –$17 million piece for the bidder!
Bradley, Desai, and Kim (1982) exclude offers that are not “control oriented” and find that both bidder and target benefit from increased market value. Paul Malatesta (1983) finds a positive effect for targets but a negative effect for bidders. He studied the period covering the sixty months that preceded approval by the target board. During this period, the increase in combined value turns out to be positive but very small. In addition, targets tend to perform poorly during this period, but bidders appear to do well. Yet during the merger months, the acquiring firms do not increase significantly in value, unlike the target firms.
Using British data, Michael Firth (1980) finds strong evidence for the –/+/0 effect in which the value of the acquiring firm declines (–), that of the target firm increases (+), but the combined value of the two stays the same (0). He finds that the change in combined market value is negative, with over half of the mergers experiencing a combined loss. Specifically, Firth finds that the premium paid by the bidder is proportionally related to the decline in the bidder’s market value, suggesting that the market actually understands winner’s curse! Nikhil Varaiya (1985) finds a positive but insignificant increase in combined market value, with the bidder losing and the target gaining. Moreover the larger the gain for the target, the worse the loss for the bidder.
What happens on the day of the announcement? Here the evidence is mixed. Asquith finds positive, but insignificant, returns to the bidder on the day of the announcement. However, Peter Dodd (1980) finds the opposite. Carol Eger (1983), studying noncash offers, finds negative returns. Interestingly, Varaiya finds negative returns on announcement day, with the effect stronger when there are rival bidders.
Equity Carve-Outs
When AT&T announced that it would split itself into three companies, the price of AT&T’s shares rose $2.625 to close at $67.375, a 4.1 percent increase. The average increase is 2.3 percent. John Hand and Terrance Skantz (1997) report that this upsurge is a common occurrence in equity carve-outs. But they also suggest that it is indicative of mispricing, because when the carved-out subsidiary goes public, there is a corresponding mean excess stock return of –2.4 percent. On average, the time between an announcement and the subsequent IPO is forty-five trading days. The situation with AT&T was considerably more complex than most, and took about one year.
Hand and Skantz suggest several possible reasons for this mispricing. One is the tendency of the financial press to describe equity carve-outs as “spinoffs.” They state that “[s]pinoffs are more common events that are well known to increase parent stock prices by an average of 3%.” A second possible reason stems from accounting changes associated with carve-outs that may mislead investors.
General Biases
The case study involving AT&T documented several behavioral biases that afflict corporate executives, most notably excessive optimism, the illusion of control, and loss aversion. These biases have been addressed in a series of articles.
Meir Statman and David Caldwell (1987) discuss a variety of reasons why executives are reluctant to terminate losing projects. In a follow-up article, Statman and James Sepe (1989) study how the stock price reacts to the announcement of a project termination. They find that stock prices typically rise when a project termination is announced, suggesting that investors greet the announcement as good news. Why good news? Investors are aware that being human, executives suffer from loss aversion and throw good money after bad. This stops with project termination.
The recognition of biases on the part of others is also a theme in the work of Statman and Tyzoon Tyebjee (1985). They find that in an experimental setting, managers seek to offset the optimistic capital budgeting projections made by their corporate colleagues.
J. B. Heaton (1998) has developed a theory of corporate financial decision making, based on optimism. His work builds on the studies of psychologist Neil Weinstein (1980), who finds that excessive optimism is especially pronounced when individuals believe they exert a measure of control (again the illusion of control) and their commitment levels are high. Heaton notes that optimism appears to be most severe among individuals of high intelligence and those who are dealing with career events.14
Summary
Corporate executives are vulnerable to both heuristic-driven bias and frame dependence. In particular hubris, a form of overconfidence, predisposes executives engaged in takeovers to overpay for targets. The act of overpaying is called winner’s curse.
This is not to say that overpayment is inevitable. In fact, as a general matter, the evidence on the hubris hypothesis is mixed. Nevertheless, corporate executives contemplating a takeover need to be mindful of winner’s curse. Most important, they need to think about which other firms have been successful in similar situations, and they must be prepared to justify why they are at least as capable of performing the same feat as these other firms. Executives might also benefit from being more attentive to the red flags that get raised when they announce their intention to acquire the target.15 It is not that successful takeovers are impossible. But they are more difficult and less frequent than overconfident, optimistic executives tend to believe.
Chapter 17 IPOs: Initial Underpricing, Long-Term Underperformance, and “Hot-Issue” Markets
Many investors have experienced an IPO-adrenaline rush on the first trading day as they search for the “next Microsoft.” In the case of Internet stocks, the editor of one IPO newsletter has described investor activity as “insanity. com trading.”1
There are three behavioral phenomena associated with initial public offerings (IPOs). These have been termed (1) initial underpricing, (2) long-term underperformance, and (3) “hot-issue” market. Before explaining what each of these terms means, let me note that there are three main parties to IPOs: the issuing firm, the underwriter, and investors. Although the role of all parties in the three phenomena is discussed, the emphasis is on the role of investors.
Initial underpricing occurs when the offer price is too low. That is, the issue will be underpriced and its price will soar on the first trading day. But price may overshoot fundamental value, in which case, it will fall back over time, giving rise to long-term underperformance. IPO activity also appears to move in hot and cold cycles. A hot-issue market is a period where investor demand for IPOs is especially high.
Are the three IPO phenomena consistent with market efficiency? I argue not. In a hot-issue market, excessive optimism on the part of investors leads IPO prices to rise above fundamental value on the first trading day and remain so for long periods. This optimism is a manifestation of heuristic-driven bias. Investors may also be affected by other heuristics, including instances of similarity, betting on trends, and representativeness. Also, in a hot-issue market, the possibility of regret looms large in the minds of
investors.2
This chapter discusses the following:
• two cases, one dealing with Boston Chicken and the other with Netscape Communications. Both cases feature the three IPO phenomena, as well as several behavioral elements
• the general evidence about IPOs
• some theories that explain the three IPO phenomena
Case Study 1: Boston Chicken
Boston Markets is a franchise owned by Boston Chicken, Inc.; Boston Market rotisserie chicken has been a popular dinner choice across the country. Several years ago, Boston Chicken was a favorite among diners and investors alike, appearing attractively priced to both. At most outlets, a quarter chicken, two side items, and corn bread started at $3.99. In a 1996 episode of the NBC sitcom The Single Guy, the main character’s father describes his few regrets in life: not spending more time with his son and not investing in Boston Chicken stock.
The case of Boston Chicken vividly illustrates initial underpricing and long-term underperformance. The firm went public on November 8, 1993, with a 1.6 million share offering at $20. On its first trading day, Boston Chicken shares soared 142.5 percent, opening at $45.25 and closing at $48.50 per share. This was the best first-day performance for an IPO in the preceding two years.
How did the stock fare over the next five years? Figure 17-1 shows the dramatic rise and fall of the stock, after adjusting the price for a subsequent 2-for-1 split. After a dramatic climb in 1995 and 1996, the stock price closed at $8.47 a share on November 10, 1997. In fact, in 1997 Boston Chicken was among the 10 worst-performing stocks on NASDAQ, declining by 82 percent. Then, on October 5, 1998, Boston Chicken filed for bankruptcy protection and closed 178 restaurants across the country. The price of its stock fell to 46.75 cents.
Beyond Greed and Fear Page 30