Beyond Greed and Fear Read online

  Real world events that occurred in the wake of the bursting bubble reinforce and confirm many points I make in the book. Initial public offerings or IPOs, constitute the subject matter of chapter 17. In the first paragraph of that chapter I state: “In the case of Internet stocks, the editor of one IPO newsletter has described investor activity as ‘insanity.com’ trading.” The insanity continued through 2000, as VA Linux Systems set a new record for a first-day gain, rising 698 percent on December 9, 1999, to displace former IPO record holder theglobe.com (described on pages 245, 246).

  A most remarkable example is the IPO of Palm, the firm that makes the Palm Pilot. In March 2000, Palm was spun out of 3Com. The first trading day for Palm’s shares was March 2, 2000, a few days before the bubble peaked. Chapter 17 is entitled “IPOs: Initial Underpricing, Long-term Underperformance, and ‘Hot-Issue’ Markets.” Long-term underpricing means that the initial offer price is too low, relative to the price set in the market on the first day of trading. Were Palm’s shares underpriced? The offer price was $38. At that price, Palm held the record for the highest market capitalization of any high-technology IPO in United States history. Its associated $22 billion market capitalization made it the fourth-largest technology firm, behind Cisco Systems, Microsoft, and Intel, and the 49th most valuable firm in the U.S. On March 2, Palm opened at $165 a share. It closed the day at $95.

  On March 2, 2000, 3Com retained 94 percent of Palm’s shares. What is especially interesting is that at the end of the first day of trading, the market value of Palm’s shares exceeded the market value of 3Com’s shares by about $25 billion. In a working paper entitled “Can the Market Add and Subtract? Mispricing in Tech Stock Carve-Outs,” Owen Lamont and Richard Thaler conclude that the market judged the non-Palm portion of 3Com to have negative value.

  3Com structured its spinoff of Palm so that each 3Com stockholder would receive 1.5 shares of Palm for every share he or she held of 3Com. In describing this situation, Richard Thaler asked: Can the market multiply by 1.5? He noted that one can ask the same question about the relative valuations of Royal Dutch and Shell, that I discuss on page 7 in chapter 1. By the terms of their charter, the joint cash flows of Royal Dutch/Shell are split between Royal Dutch and Shell Transport so that Royal Dutch receives 1.5 times the cash flows that Shell receives. Yet, the relative market valuations of Royal Dutch and Shell frequently deviate from 1.5 by a significant amount.

  Long-term underperformance means that an investor who buys and holds stock in the first few days that a firm’s shares are traded publicly will earn inferior returns on average. In this respect, I note that at the close of trading on March 2, Palm, with fewer than 700 employees, had a market cap of $53.4 billion. At $165 per share, Palm had one of the highest market capitalizations in the United States. In respect to long-term underperformance, the price of Palm’s stock went from its all-time high of $165 on March 2, 2000 to $1.11 on June 7, 2002. As to whether Palm’s IPO took place in a “hot-issue” market, let me simply say that the IPO took place at the height of the bubble.

  In chapter 5, I describe a key difference between the way individual investors form market predictions and the way institutional investors form market predictions. Individual investors suffer from extrapolation bias, and naively extrapolate recent trends. Institutional investors suffer from gambler’s fallacy, and are overly prone to predicting reversals. For example, on page 46, I state: “In the wake of above average performance in 1995 and 1996, did the strategists predict that 1997 would feature below-average performance? Indeed they did. They predicted that the Dow would actually decline by 0.2 percent, well below the 8.6 percent annual rate that the Dow had grown between 1972 and 1996.”

  To predict a market decline in 1997, after two years like 1995 and 1996 when the Dow returned over 33 percent and 26 percent respectively, is to succumb to gambler’s fallacy. The error involves going overboard. More appropriate is to predict that the Dow would rise by an amount closer to 8.6 percent—that is, to predict a lower increase than had occurred in the prior two years. In fact, as I point out in chapter 5, the Dow returned more than 22 percent in 1997.

  Did strategists succumb to gambler’s fallacy after the bubble burst? In December 2000, the Wall Street Journal elicited the market predictions of eight highly respected strategists. When these strategists were asked to state their predictions for the closing value of the S&P 500 at year-end 2001, the index had declined by 7 percent during 2000. The strategists’ average forecast called for an increase in excess of 17 percent! In other words, gambler’s fallacy continued. Just for the record: the S&P 500 returned -11.9 percent during 2001.

  The message of chapter 16, “Corporate Takeovers and the Winner’s Curse,” is that because of hubris and overconfidence, the managers of acquiring firms often overpay for their targets. On page 328, I note that as I began work on the chapter in 1998, the computer firm Compaq was in the process of acquiring Digital Equipment Corp. The history of prior technology deals did not offer much in the way of promise for success. How did the merger turn out? Consistent with the hubris hypothesis, Compaq executives were surprised by the difficulty of integrating the two firms. This outcome appears to have been a primary factor in the resignation of Compaq’s chief executive officer Eckhard Pfeiffer. It took Compaq more than a year to absorb Digital Equipment, during which time it lost valuable market share to Dell and Sun Microsystems.

  At the time of my writing this new preface, Compaq has not recovered. However, this is not the end of the story. In 2002, Hewlett-Packard acquired Compaq in a shareholder battle that was front-page news. The battle pitted Carleton (Carly) Fiorina, CEO of Hewlett-Packard, against board member Walter Hewlett, son of William Hewlett, one of Hewlett-Packard’s founders. I imagine that the outcome of that merger will be the subject of a future edition.

  Acquisition activity peaked at $1.8 trillion in 2000, more than triple the level in the mid-1990s. Between 1995 and 2000, the average acquisition price in the United States rose 70 percent, to $470 million. Since that time, the extent of the winner’s curse has become apparent. In April, 2002 AOL Time Warner Inc. wrote off $54 billion of “goodwill” to recognize AOL’s overpayment for Time Warner. This was one of one of the largest writedowns in corporate history. Joining AOL Time Warner in the writedown category were Vivendi Universal SA of France, Cordiant Communications Group, Boeing Co., Vodafone Group PLC., Tyco International Ltd., WorldCom Inc., and AT&T Corp. The appearance of AT&T on this list is especially interesting, in view of its long history experiencing the winner’s curse: see pages 228–233 of chapter 16. Behaviorists stress that although people do learn, they learn slowly.

  In chapter 18, “Optimism in Analysts’ Earnings Predictions and Stock Recommendations,” I pointed out that the games analysts play with investors have a “wink, wink, nod, nod character.” Specifically, investors do not appear to appreciate the role that analysts’ recommendations play in attracting investment banking business to their firms, so that what “investors hear is not always what analysts mean.” The issues described in the first part of chapter 18 came to the fore as the prices of technology stocks plummeted after March 2000. As stocks peaked in March 2000, nearly 73 percent of all analyst recommendations were “buy” and “strong buy.” Notably, at year-end 2000, the percentage of recommendations that were “buy” was still above 70 percent. The contrast between recommendations and performance was stark enough to attract the attention of Congress as well as the attorney general for the State of New York, Eliott Spitzer.

  In the course of investigating Merrill Lynch, Spitzer’s office uncovered some interesting facts. At the same time that Merrill analysts were issuing buy recommendations for particular stocks to the public, within their firm they were describing these same stocks as “crap” in email messages to each other. In 2002, Merrill Lynch agreed to pay $100 million in order to settle a case brought against them by the attorney general. Spitzer’s office was particularly interested in the analysts covering technology
stocks, such as Merrill’s Henry Blodget, and Morgan Stanley’s Mary Meeker, who had been dubbed “the queen of the dot-coms.”

  On pages 266 and 267 of chapter 18, I provide an example that took place in 1997, involving the stock of chip manufacturer Intel to describe biases in analysts’ earnings estimates. The Spring 2001 issue of Financial Management contains an article titled “Is the Response of Analysts to Information Consistent with Fundamental Valuation? The Case of Intel,” by Brad Cornell. He analyzes the market reaction to a press release by Intel on Thursday, September 21, 2000. The press release indicated that Intel’s expected revenue for its third quarter would grow between 3 percent and 5 percent, not the 8–12 percent that analysts had been projecting. In response to news that was less than earth shattering, Intel’s stock price dropped by 30 percent over the next five days! Intel’s chairman, Craig Barrett, commented on the reaction, stating: “I don’t know what you call it but an overreaction and the market feeding on itself.”

  One of the most interesting findings in Brad Cornell’s analysis is that of the 28 analyst reports on Intel that he examined, not a single one contained a discounted cash flow model from which to infer fundamental value. Yet, in the week following Intel’s press release, many analysts lowered their recommendations on the stock.

  If analysts do not rely on discounted cash flow analysis to gauge whether a stock is fairly priced, what do they use? On pages 81–83, I discuss the heuristic “good stocks are stocks of good companies.” In line with this heuristic, Cornell suggests that analysts rate the company instead of the investment; that is, they react to bad news in the same way that a bond-rating agency reacts to bad news by downgrading the firm’s debt. In other words, analysts base their judgments on a heuristic that leaves them vulnerable to bias.

  Chapter 14 provides an analysis of the behavioral elements that led to the Orange County, California, bankruptcy in 1994, the largest municipal bankruptcy in history. In December 2001 Enron, the seventh largest firm in the United States, filed for bankruptcy protection. Here too, behavioral elements were paramount. Enron management, apparently overconfident from its success in the natural gas business in the early 1990s, sought to repeat that success in markets where they lacked expertise. Enron invested more than $10 billion in ventures that produced a near-zero return. As a result, Enron returned less than its cost of capital to investors, thereby destroying shareholder value. Moreover, Enron executives employed an opaque framing strategy, obscuring the financial implications of their investments through the use of limited partnerships that were off-balance sheet items.

  Enron’s bankruptcy triggered widespread concern among investors about whether they could trust the information conveyed in corporate financial statements. Indeed in 2002, the entire accounting profession found itself in a crisis about the extent to which auditors permit firms to engage in opaque framing. Even worse, employees at Arthur Andersen, the accounting firm that audited Enron, had engaged in major modification and shredding of documents related to Enron. The U.S. Department of Justice filed suit, and a jury found the accounting firm guilty of obstructing justice. Arthur Andersen had also audited the financial statements for WorldCom, whose chief financial officer improperly booked $3.8 billion in expenses as capital expenditures, a maneuver that enabled the firm to report positive earnings in 2001 rather than a loss.

  The issue of opaque accounting became a dominant theme in 2002. Although required by law to file financial statements with the Securities and Exchange Commission (SEC) that conform to Generally Accepted Accounting Principles (GAAP), many firms use different definitions of earnings for their news releases and forecasts than the GAAP-definition. The alternative definitions are known as “pro forma” earnings, to mark their “as if” nature. Typically, pro forma earnings only pertain to a subset of line items, mostly relating to operations, and exclude items such as restructuring charges. As a result, pro forma earnings often appear to be more favorable than GAAP earnings. More important there is no uniform definition of what pro forma earnings entail. The definition of pro forma earnings varies from firm to firm, thereby allowing firms to engage in opaque framing, if they wish. Interestingly, analysts forecast and track pro forma earnings. Notably, the focus of First Call’s reports is pro forma earnings, not GAAP earnings. In an effort to make earnings more transparent, the SEC has been engaged in a major effort to prevent firms from using the pro forma definition in their news releases and forecasts.

  New Research in Behavioral Finance

  Having to send Beyond Greed and Fear off to press in the midst of a stock market bubble felt like turning in a mystery novel without the last chapter. Of course, psychological factors are always at work, but the last few years do seem to have been an extraordinary period in underscoring just how important psychological factors can be. In the remainder of the preface, I continue to discuss new developments in the context of existing chapters. However, instead of focusing on recent events, the updates will focus on recent contributions by academic scholars.

  Part III of the book consists of three chapters about individual investors. Framing transparency is an issue of critical importance to individual investors, whose ability to process information is limited. Recent unpublished research identifies key factors that determine which stocks individual investors buy. (A copy of this study and other unpublished studies cited in this preface can be found at the authors’ websites or downloaded from the Social Science Research Network— SSRN.) The study, entitled “All That Glitters: The Effect of Attention and News on the Buying Behavior or Individual and Institutional Investors,” is by Brad Barber and Terrance Odean. Their study points to three indicators of attention: recent news, recent extreme price movements, and recent excess trading volume. In related work, Dong Hong and Alok Kumar ask a question that forms the title of their working paper: “What Induces Noise Trading Around Public Announcement Events?” Hong and Kumar suggest that at an aggregate level, investors exhibit a contrarian volume reaction that is primarily driven by price trends.

  Odean and Barber tell us that companies that are in the news grab the attention of individual investors. And as I mentioned in the previous section, when the news about these companies pertains to earnings, the news tends to be framed in terms of opaquely defined pro forma earnings, rather than GAAP earnings. This is not to say that the evidence supports the idea that individual investors have the sophistication to make use of accounting information, even when it is transparently framed. Indeed, the evidence strongly supports the opposite contention: investors rely on very crude valuation heuristics rather than on fundamental analysis. As noted above in respect to Intel, even financial analysts schooled in fundamental analysis rely on heuristics!

  Chapter 10 describes earlier work by Barber and Odean, whose research program offers some of the best evidence about the behavior of individual traders. Their new work has greatly added to our understanding. Consider the section “The Online Revolution” (pages 133–134), that discusses the impact on individual traders stemming from the Internet, overconfidence and the illusion of control. When I wrote this section, Barber and Odean had not yet written “Online Investors: Do the Slow Die First?” (Review of Financial Studies, 2002). Barber and Odean provide a fascinating account of the advertising strategies used by online trading firms in order to induce investors to trade online. These strategies appealed to a combination of overconfidence and the desire for control. Not all traders choose to trade online. Barber and Odean find that investors who chose to go online had experienced above average returns prior to going online, outperforming the market by 2 percent. Was the above average performance of online investors due to skill or luck? If it were due to skill, these investors would be expected to continue to outperform the market after going online. However, Barber and Odean find just the opposite. After going online, these investors traded more actively, took on more risk, and traded less profitably than they did prior to going online. They would up underperforming the market by 3 percent.
>
  Barber and Odean have also added to the findings about investment club performance described on page 131. In “Too Many Cooks Spoil the Profits: Investment Club Performance” (Financial Analysts Journal, 2000), they document that in the aggregate, investor clubs underperform not only the market, but also individual investors. Interestingly, investment club performance is related to gender. Brooke Harington documents that clubs composed of a mix of women and men outperform clubs composed of only women or only men. Her paper is entitled “‘Popular Finance’ and the Sociology of Investing.”

  Do investors understand what they are doing? Do they think they have chosen well for themselves? In a study entitled “How Much Is Investor Autonomy Worth?” Shlomo Benartzi and Richard Thaler presented individuals saving for retirement with information about the distribution of outcomes they could expect from the portfolios they picked and also the median protfolio selected by their peers. Benartzi and Thaler found that a majority of their survey participants actually preferred the median protfolio to the one they picked for themselves.

  In a separate article, “Excessive Extrapolation and the Allocation of 401 (k) Accounts to Company Stock,” (Journal of Finance, 2001), Benartzi analyzes the portion of 401 (k) portfolios that employees devote to company stock. He identifies some of the major reasons why, in the aggregate, employees concentrate roughly a third of their protfolios in the stocks of the company for which they work. Specifically, Benartzi finds that employees purchase company stock after it has already gone up, effectively succumbing to extrapolation bias. He also suggests that employees treat the matching funds from a company, that are automatically invested in company stock, as an endorsement, that the automatic investment leads employees to hold an even larger share of company stock than they would have otherwise.