Beyond Greed and Fear

by Hersh Shefrin

  On page 26 I introduce the notion of mental accounting, and in chapter 10, on page 125, I discuss how the concept of mental accounting applies to portfolio choice. Two recent articles extend the ideas developed in this chapter. The first is “Behavioral Portfolio Theory” (Journal of Financial and Quantitative Analysis, 2000), by Meir Statman and me. The second is “Mental Accounting, Loss Aversion, and Individual Stock Returns” (Journal of Finance, 2001) by Nicholas Barberis and Ming Huang. Barberis has several recent pieces that are relevant to this discussion. In “Investing for the Long Run When Returns Are Predictable” (Journal of Finance, 2000), he analyzes what investors need to consider about the proportions of their portfolios to hold in stocks, given the limits of our understanding about the drives of returns. In “Style Investing” (Journal of Financial Economics, 2002), he and co-author Andrei Shleifer analyze the ramications stemming from investors’ tendency to categorize investments into groups, such as growth and value.

  Chapter 9 is devoted to the disposition effect, especially “get-evenitis,” the tendency to gamble by holding onto losers too long. On page 24, I mention the case of Nicholas Leeson, who lost Barings Bank over $1.4 billion because he could not come to terms with a loss and engaged in highly speculative currency trading in an attempt to get even. In February 2002, the Wall Street Journal reported the largest currency trading scandal since the case of Nicholas Leeson. One John Rusnak, a trader at Allied Irish Banks PLC, lost his firm $691 million, in an attempt to get even. The way to deal with get-even-it is to employ stop-loss orders, either explicitly or through a self-imposed rule. Apparently, neither Leeson nor Rusnak learned the lesson.

  The behavioral elements that influence investment decisions are hardly unique to Americans. In an article entitled “What Makes Investors Trade?” (Journal of Finance, 2001), Mark Grinblatt and Matti Keloharju use data from the Finnish Central Securities Depository to analyze the behavior of Finnish traders. Grinblatt and Keloharju find strong evidence of get-evenitis. They differentiate between an extreme loss (in excess of 30 percent) and a moderate loss. Grinblatt and Keloharju report that an extreme capital loss makes it 32 percent less likely that an investor will sell a stock, whereas a moderate capital loss makes such a sale 21 percent less likely. The discussion in chapter 9 explains that this phenomenon reverses itself in December. Grinblatt and Keloharju report that in December, investors are 36 percent more likely to sell extreme losers than they are in the rest of the year. Interestingly, investors wait for the last eight trading days to sell extreme losers, almost the last minute.

  On page 34 readers will find a section entitled “The Failure to Diversify.” On pages 132–133, I describe the results in a working paper by Brad Barber and Terrance Odean, a paper that has now been published as “Trading Is Hazardous to Your Health: The common Stock Investment Performance of Individual Investors” (Journal of Finance, 2000). Barber and Odean report that in their sample of 78,000 households with accounts at a major discount broker, the median number of stocks held was between 2 and 3.

  One of the most striking findings in the Barber-Odean study involves the percentage of investors who managed to beat the market. When I ask people to guess what fraction of individual investors beat the market, the typical response lies between 5 percent and 20 percent. Barber and Odean find that 49.3 percent of investors beat the market before trading costs, and 43.4 percent beat the market after trading costs. This fact astonishes people, because they connect performance to skill. However, the market return serves to average the returns to all stocks; hence, half of stocks beat the market. And the average individual investor only holds 2 or 3 stocks. Lack of diversification leads about half of investors to beat the market before trading costs. At the same time, this lack of diversification produces a very wide range in performance, from -95 percent to over 11,000 percent, measured on annual basis.

  The Barber-Odean study pertained to January 1991 through December 1996. Recent research by William Goetzmann and Alok Kumar documents that although investors have made some progress on this dimension, they continue to hold portfolios that fall far short of being well diversified. Their working paper is entitled “Equity Portfolio Diversification.”

  Chapter 11 describes the behavioral factors that influence retirement saving and spending. As I say on page 139, “investors need to overcome myopia and exercise self-control in order to save for retirement.” On page 141, I indicate that most Americans have not been able to save adequately. A 2000 working paper by David Wise and Steven Venti, entitled “Choice, Chance, and Wealth Dispersion at Retirement,” provides evidence that the dominant factor that determines the wealth households possess at retirement is their past saving rate. In the aggregate, the rate at which households save over their lifetimes is more important than the size of their medical bills, the impact of inheritances, and whether they have invested conservatively or aggressively over their lifetimes.

  One of the most important recent developments in regard to retirement saving is a program developed by Richard Thaler and Shlomo Benartzi, called “Save for Tomorrow” (SMT). In January 2002, the Wall Street Journal and New York Times ran stories describing the program. In their paper, “Save More Tomorrow: Using Behavioral Economics to Increase Employee Saving,” Thaler and Benartzi describe the characteristics of their plan, along with preliminary results.

  The SMT plan has four ingredients, all designed to deal with issues identified in the behavioral literature. First, because people tend to accord future unpleasantness much less weight than immediate unpleasantness, employees are approached about increasing their contribution rates a considerable time before they begin to participate. Second, people hate situations where they perceive themselves as incurring losses. Consider the challenge of how to frame matters so that employees do not perceive a loss in their take-home pay. Thaler and Benartzi suggest that if employees join SMT, their contribution to the plan should begin with the first paycheck after a raise. Third, the contribution rate continues to increase on each scheduled pay raise until it reaches a preset maximum. In this way, inertia and status quo bias work toward keeping people in the plan. Fourth, the employee can opt out of the plan at any time.

  Thaler and Benartzi report that the first implementation of the SMT plan took place in 1998 at a midsized manufacturing company. Prior to adoption, the company suffered from low participation rates as well as low saving rates. Thaler and Benartzi report three findings about the post-adoption period. First, 78 percent of those who were offered the SMT plan elected to use it. Second, 98 percent of those who joined remained in the plan through two pay raises, and 80 percent remained in the plan through the third pay raise. Third, the average saving rates for SMT plan participants increased from 3.5 percent to 11.6 percent over the course of 28 months. Thaler and Benartzi are in the process of applying the lessons of their successful pilot study to other firms.

  Thaler and Benartizi constructed their SMT plan with great care. To be sure, automatic enrollment in 401(k) plans does not guarantee that employees will save more. This finding is described in the working paper entitled “For Better or For Worse: Default Effects and 401(k) Savings Behavior” by James Choi, David Laibson, Brigitte Madrian, and Andrew Metrick. These authors find that employees participants tend to become anchored at low default savings rates and in conservative default investment vehicles.

  Chapter 12 describes how investment companies strategically use opaque framing in their interactions with investors. On page 171, I describe the game of opaque fees. In this regard, Brad Barber, Terry Odean and Lu Zheng have written a paper entitled “Out of Sight, Out of Mind: The Effects of Expenses on Mutual Fund Flows.” The main point of the paper is that that mutual fund investors are more sensitive to salient inyour-face fees, such as loads and commissions, than they are to operating expenses. The authors document a negative relationship between fund flows and load fees, between fund flows and commissions charged by brokerage firms, but not between operating expenses and fund flows.


  One of the most interesting applications of the ideas in chapter 12 involves the Masters 100 Fund run by well-known fund manager Ken Kam at Marketocracy. Ken Kam’s strategy is based on the Olympic coins framework (gold, silver, and bronze), which I describe on pages 161–165. In that framework, the coins are weighted and represent intrinsic ability. Those who toss gold coins have more intrinsic skill than those who toss silver coins or bronze coins. Heads represents success. However, luck also plays a role—through luck alone, someone tossing a bronze coin may still toss a long sequence of heads. In chapter 12, I describe the odds of being able to filter out skill on the basis of past performance.

  Marketocracy invites people to manage fictitious portfolios over the web and chooses the top 100 to manage real money—hence the name Masters 100 Fund. More than 50,000 people have signed up at the Marketocracy website to try their hand. Hence the sample is very large. Ken Kam argues that by selecting the top 100 performers from a universe of over 50,000, the odds favor Marketocracy being able to filter out investors whose success stems from superior skill or information rather than luck. The Masters 100 Fund was introduced in November 2001 and beat the S&P 500 by about 2 percent over the next two months. For the first five months of 2002 the fund was up by 7.5 percent. In contrast, the S&P 500 was down 6.5 percent in the same period, and the Wilshire 5000 was down 4.3 percent. Notably, the fund appears to be no more volatile than the S&P 500 or the Wilshire 5000. The Wall Street Journal ranked the fund as the top-performing fund among multi-cap core funds.

  Chapter 15 describes issues involving the money management industry. I present a case, based on the experience of my own university, which is advised by the consulting firm Cambridge Associates. After Beyond Greed and Fear was published, I received a letter and Cambridge research paper from Ian Kennedy, Director of Research at Cambridge Associates. He writes in the hope of persuading me that “we are not entirely unenlightened on the issue of value-added by active management! I should add, however, that this paper, distributed to all our clients, has made no perceptible dent in the manager selection practices of the investment committees we work with. Because they are generally composed of very successful, intelligent people, they just know that they can identify superior active managers. At Cambridge Associates, we have a lively, running debate as to whether anyone could pick managers that would beat ‘the market,’ net of fees, over any extended period of time (e.g., ten years or more), and how one might reasonably go about trying to do so. I’m in the camp that thinks it might be possible, but is extremely difficult—and that the usual approach of the typical investment committee is absolutely doomed to failure (unless they just happen to get lucky.)”

  The behavior of investment committees is a new area for researchers. John Payne and Arnold Wood have report the findings of a survey they conducted of investment committee members in a working paper “Optimizing Investment Committee Decision Making.” The general behavioral decision literature documents that groups are effective when dealing with intellectual tasks where there is a correct answer, but are less effective when dealing with judgmental tasks where there is no objectively correct answer. Against this backdrop, Payne and Wood find that investment committees report that they deal more frequently with judgmental tasks than tasks involving a correct answer, and that they feel being confident in their decisions. On average, they indicate that the probability of making a correct decision is 73 percent, a finding that is consistent with Ian Kennedy’s remarks quoted in the previous paragraph.

  I have intentionally left for the end updates to chapter 7, “Picking Stocks to Beat the Market” and chapter 8, “Biased Reactions to Earnings Announcements.” I feel that academics and practitioners pay far too much attention to beating the market, and in consequence discount, or even overlook, the most important behavioral lessons discussed above.

  I began chapter 7 by making clear that some people do consistently beat the market. In this respect, I mentioned the Wall Street Journal’s contest that pitted the “pros” against the “darts.” The newspaper terminated the contest in 2002, after more than a decade of operation. Over the contest period, the average annual return for the pros’ stock picks was 10.2 percent, roundly trouncing the darts’ return of 3.5 percent and the Dow Jones Industrial Average that returned 5.5 percent. Another example I mention on page 70 involves the stocks recommended on the television program “Louis Rukeyser’s Wall Street,” the continuation of “Wall Street Week With Louis Rukeyser.”

  In chapter 7, I present evidence collected by Zacks that on average, stocks recommended by brokerage firms beat the S&P 500. At the same time, on pages 78–80 and 89, I indicated that the margin of outperformance was modest, and that the odds of picking a brokerage firm whose recommendations would outperform the market were no better than tossing a fair coin. These conclusions are reinforced in an article entitled “Can Investors Profit from the Prophets? Security Analyst Recommendations and Stock Returns,” by Brad Barber, Reuven Lehavy, Maureen McNichols, and Brett Trueman (Journal of Finance, 2001).

  Chapters 7 and 8 discuss what are called predictable patterns in individual stock returns. Most financial economists agree that stock returns exhibit momentum in the short term and reversals in the long term. However, there is disagreement as to whether or not this pattern signifies mispricing. Proponents of behavioral finance assert that it does signify mispricing, and proponents of market efficiency assert that it does not. Moreover, among proponents of behavioral finance there is a debate. Some argue that momentum stems from underreaction, while others argue that momentum stems from overreaction.

  Since the book was published, Narasimhan Jegadeesh and Sheridan Titman have updated their pioneering 1993 study of momentum.

  In “Profitability of Momentum Strategies: An Evaluation of Alternative Explanations” (Journal of Finance, 2001), they find that the conclusions of their earlier study remained robust in the post-sample period.

  In “Price Momentum and Trading Volume” (Journal of Finance, 1999), Charles Lee and Bhaskaran Swaminathan provide additional insights. They suggest that in order to exploit a momentum-based strategy more effectively, investors should take trading volume into account. Specifically, investors should buy high-volume winners and short sell high-volume losers. They note that investors would have lost money by focusing on the low volume losers: historically, low volume losers have rebounded.

  Mark Grinblatt and Bing Han have an intriguing paper entitled “The Disposition Effect and Momentum” (2001). They argue that momentum stems from the disposition effect, rather than underreaction or overreaction.

  Momentum arises in many contexts. Consider stock splits. In an article entitled “What Do Stock Splits Really Signal?” (Journal of Financial and Quantitative Analysis, 1996), David Ikenberry, Graeme Rankine, and Earl Stice document that there is also positive drift associated with stock splits. They find that firms that split their stocks earn an average abnormal return of 7.93 percent in the first year, and 12.15 percent in the first three years. The three-year effect seems to be concentrated in value stocks. For growth stocks, the effect does not extend beyond the first year. In subsequent work, David Ikenberry and Sundaresh Ramnath demonstrate that firms who decide to split their stocks tend to be those for whom coverage by analysts in respect to earnings forecasts has been pessimistic. Their article is entitled “Underreaction to Self-selected News Events: The Case of Stock Splits” (Review of Financial Studies, 2002). In addition, Ikenberry and Ramnath find that firms who announce stock splits are much less likely to experience a decline in future earnings, relative to firms with comparable characteristics.

  Researchers are also studying other securities to see whether the same issues occur there. In “The Long-Run Stock Returns Following Bond Ratings Changes” (Journal of Finance, 2001), Ilia Dichev and Joseph Piotrosdki find negative abnormal stock returns of between 10 and 14 percent in the first year following downgrades of corporate debt. They conclude that this effect stems from underreaction to
the announcement of the downgrades, rather than from lower systematic risk.

  In chapter 19, I argue that options markets also feature mispricing. In “Underreaction, Overreaction, and Increasing Misreaction to Information in the Options Market” (Journal of Finance, 2001), Allen Poteshman finds that options traders exhibit short-horizon underreaction to daily information, but long-horizon overreaction to extended periods of mostly similar daily information. Moreover, these misreactions increase as a function of the quantity of previous information that is similar.

  For some years now, those who trade options and those who study options markets have realized that the traditional theory, based on Black-Scholes pricing, does not apply. One of the first articles to document the phenomenon is “Riding on a Smile” (Risk, 1994), by Emanuel Derman and Iraj Kani. The “smile” refers to the shape of the graph that plots “implied volatility” against exercise price for options having the same expiration date. Were options priced in accordance with the Black-Scholes formula, that graph would appear as a horizontal line, rather than something resembling a smile or a smirk. I propose a behavioral explanation for the smile effect in “Irrational Exuberance and Option Smiles” (Financial Analysts Journal, 1999). After my article appeared in print, I received a supportive message from Emanuel Derman, who heads the quantitative strategies group at Goldman Sachs, and whose work I cite above. He had read my article, and sent me a copy of his own current presentation on smiles, aptly titled “Fear and Greed in Volatility Markets.”

  Historically, U.S. stocks have exhibited momentum at short horizons and reversals at long horizons. These phenomena are not exclusively American. In “Contrarian and Momentum Strategies in Germany” (Financial Analysts Journal, 1999), Dirk Schiereck, Werner De Bondt, and Martin Weber establish that the same phenomenon has occurred for stocks that trade in Germany on the Frankfurt exchange.