Beyond Greed and Fear

by Hersh Shefrin

  Let’s say you received a hefty bonus, so you want to add $5,000 to your mutual fund investments. If you do so all at once, you run the risk that the market will fall, and you will have invested at the wrong time. … On the other hand, if you put $500 into the fund each month for the next ten months, each time the market dips, you’ll cheer. …

  Dollar-cost averaging for lump sums may not make sense in the early stages of a bull market, or in a bear market that promises quick recovery. When prices are climbing rapidly, you will benefit far more from investing your money early, taking full advantage of the early growth in value, rather than spending your investment over several months and losing out on the early price surge. (pp. 77–82)

  Finance academicians have criticized the practice of dollar-cost averaging for years. See George Constantinides (1983) and Michael Rozeff (1994). But as Statman (1995a) points out, the standard critique misses the important framing advantages offered by dollar-cost averaging. These advantages are implicit in the description provided by Wall above.

  Consider a variant of Wall’s $5,000 bonus example. Suppose that an investor in receipt of a bonus considers investing half the amount this month and the remaining half the next month. Let the current share price be $50. In this case, the investor purchases 50 shares this month. Imagine that the share price falls dramatically during the month, to $12.50. By following the dollar-cost-averaging rule at the beginning of the next month, the investor purchases an additional 200 shares. The average price per share is $31.25, the midpoint of $50 and $12.50. But the investor does not buy the same number of shares in the two months. Therefore the average cost to the investor on a per share basis is actually lower than $31.25, because more shares are purchased when the price is less. In this example, the average cost per share is $20, the ratio of the total expenditure $5,000 to the number of shares purchased, 250.

  So $20 is less than $31.25, which is something that a dollar-cost averager can take comfort in. Remember how Wall described it: “The result is a lower cost per share than if you bought a set number of shares each month.” Had the investor bought the same number of shares in the two months, the cost per share would have been $31.25. So, is there anything wrong with that? That is, if an investor measures the $20 he paid against a reference point of $31.25, then what’s wrong with generating a guaranteed gain?

  To answer this question, consider a variation of a problem presented by Rozeff (1994). Suppose that the investor with the $5,000 bonus compares two strategies: (1) investing the $5,000 in stock immediately, the lump-sum strategy; and (2) parking $2,500 in cash and following a two-month dollar-cost-averaging strategy. For the sake of illustration, let the expected monthly return on the stock be 1 percent, and the monthly return standard deviation be 1 percent. The yield on cash is zero.

  If the investor follows the lump-sum strategy, he can expect to have $5,100.50 at the end of the two-month horizon. If he uses dollar-cost averaging, then he can expect to have $5,075.50. This illustrates Wall’s cautionary note about investing early in bull markets. Of course, only a crystal ball can say before the fact that prices are about to climb rapidly. The fact is that the expected return to dollar-cost averaging is less than the expected return to lump-sum investing. But so too is the risk. A little arithmetic shows that the return standard deviation of the former strategy is $56.57 while that of the latter is $72.13.

  Is this the academician’s complaint, lower risk? Actually, it’s not. The academician’s complaint is that an investor who puts 75 percent of his bonus into stocks right away and keeps the remaining 25 percent in cash for the entire two-month horizon can expect to do as well as with dollar-cost averaging, but with less risk. The 75/25 strategy has a return standard deviation of $54.10.

  Why is dollar-cost averaging riskier? To answer that question, consider a slightly different problem. Suppose a second investor has $2,500 to invest. She can put 50 percent into stocks and 50 percent into cash, and keep the position for two months. Or she can hold the $2,500 in cash for the first month and then invest it all in stocks for the second month. Which is the riskier strategy?

  A typical investor prefers the first strategy because it seems less risky. And the first strategy is less risky. Now suppose that we give the same investor an additional $2,500 of stock that must be held for two months, and ask her whether that would change what she does with the first $2,500. Most investors do not see why having an additional $2,500 in stock would affect their preferences for what to do with the first $2,500.

  But this is a framing issue. In this frame the typical investor says she prefers to split the first $2,500 evenly between cash and stock for the two months, rather than lurch from 100 percent cash to 100 percent stock at the end of the first month. But it turns out that the first strategy gives her the same result as the first investor’s lump-sum strategy with $5,000, and the second gives her the same result as his dollar-cost averaging. If both investors are identical in their views about risk, their different choices will stem from frame dependence.

  Of the two frames, the first is more common. In this frame, dollar-cost averaging is attractive for the same reason that people prefer multiple bets to one-shot risks: It mitigates loss aversion. Moreover as Statman (1995a) argues, financial loss is usually accompanied by regret. Lump-sum investors are likely to blame themselves more than dollar-cost averagers if the stock takes a dive shortly after they bought it. Why? Regret is especially painful if the action that was taken is perceived as departing from what is habitual, or what Kahneman and Tversky (1982) call conventional. By its nature, dollar-cost averaging is habitual, like making a mortgage payment on the first of every month. But lump-sum investing is not routine, and it appears to be less prudent than the cautious step-by-step approach of dollar-cost averaging. A lump-sum investor leaves himself more vulnerable to regret than a dollar-cost averager.

  As Wall pointed out, dollar-cost averaging has other benefits as well. First and foremost, investors who adopt dollar-cost averaging cultivate an excellent saving habit. Why? Because dollar-cost averaging mimics the regular deposits in a defined contribution plan. Second, it serves as an anti–panic device when stocks fall.

  Dividends: Safeguarding That Retirement Nestegg

  Cash dividends have two tax disadvantages relative to capital gains. They are usually taxed at a higher rate, and those taxes cannot be deferred. Taxes on capital gains can be deferred by not selling the appreciated stock.

  Nonetheless, many investors prefer stocks that pay dividends. Why? This is an issue that Statman and I addressed (Shefrin and Statman 1984). Until now, I have been discussing retirement savings from the perspective of someone still working. But to understand the major reason why investors find dividends attractive, let’s move on to retirement. Suppose that you are retired and that you use three sources to fund your expenditures: Social Security (funds 20 percent of monthly expenditures); defined benefit pension plan (funds 45 percent of monthly expenditures); and dividends from two individual stocks, a utility stock and a bank stock. Together, the dividends from the two stocks fund 35 percent of your monthly expenditures. Moreover, these are the only stocks in your portfolio.

  Imagine that by sheer coincidence, the utility and the bank both experience financial distress at the same time; the value of both stocks has declined by 50 percent in the past three months, and both firms have decided to omit the next quarterly dividend. You have a choice: Either you can let your consumption expenditures decline by 35 percent over the next three months, or you can sell some of your stock and spend the proceeds on consumption.

  When I ask this question in surveys, over 70 percent of the respondents indicate that in this situation they would cut their consumption rather than sell stock.4 Is this the behavior of people who have difficulty delaying gratification? I don’t think so. In fact, these people appear to have the opposite difficulty. During their working years, they have cultivated habits to safeguard their savings against a lapse in self-control, a lapse that might lead
them to dip into their savings too early. One such habit is the rule “don’t dip into capital.” The way to use this rule, and fund consumption expenditures out of a portfolio, is to hold dividend-paying stocks and consume the dividends.

  This goes with the mental accounting framework described earlier. People often control their spending by financing consumption from their current-income accounts. Dividends are income, just like bond coupons and salaries. The cost of not controlling expenditures can be expensive. People might outlive their assets, a frightening thought for some. The additional tax burden associated with cash dividends appears to be worth the price for most. The additional discomfort of cutting consumption for three months, rather than dipping into capital, also appears to be worth the price.

  Using dividends to finance consumption is much more common among retirees than in the general investor population. In a survey conducted in 1986, Robert Shiller and John Pound (1989) asked respondents to indicate the extent to which they agree with the following question: “I use dividend income for day-to-day expenses of living, but avoid selling stock except to reinvest the proceeds.” On a scale of 1 to 7, where 1 connotes strongly agree, the average response was 3.29. When I administered the Shiller-Pound survey to a new sample in 1998, I obtained a similar response, 3.35, which I interpret as very mild agreement. But I also administered the previous “if retired don’t dip into capital” question to this same group. And over 88 percent indicated that if they were retired, they would cut consumption before dipping into capital.

  As I mentioned in chapter 3, dividends offer the opportunity to engage in hedonic framing. Hedonic framing involves the allocation of money across separate mental accounts for the purpose of producing the most pleasant psychological response. One instance of hedonic framing involves the “silver lining” effect, where a small positive outcome is used to partially counterbalance a loss that is much larger.

  Shiller and Pound also asked investors the extent to which they agree with the following statement about such an effect for dividends: “I am left with a larger silver lining in the event of a stock price decline. That is, even if the stock price later goes down I still have the consolation of knowing I will get my money back eventually in the form of dividends.” The median response they received was 3.41. The average score for the respondents in my sample was 3.09: They agreed with this statement somewhat more strongly than did the original Shiller-Pound sample. In fact, 25 percent registered strong agreement.5

  Retirement Spending: How Comfortable?

  People exhibit great variation in the amount of planning they do for retirement. As a result, they vary greatly in the type of retirement they experience.

  Douglas Bernheim, Jonathan Skinner, and Steven Weinberg (1997) used the Panel Study of Income Dynamics and Consumer Expenditure Survey to study how people adjust their consumption expenditures around the time they retire. They divided people into sixteen groups according to the actions they took in two respects during the six years leading up to retirement: the rate at which people accumulated wealth, and the rate at which they replaced one source of income with another.

  Figure 11-2 shows the results for three groups of people. You will see that there are three curves in the figure. Each curve shows what happened to the growth rate of consumption expenditures in the years around retirement. The left side of the figure depicts what occurred before retirement, while the right side indicates what happened after retirement.

  The top curve in the figure pertains to the “high fliers,” the top 6 percent or so. High fliers accumulated wealth and replaced income the most quickly. The bottom curve pertains to the group I call “unprepared,” the bottom 6 percent. These people accumulated wealth and replaced income at the slowest rate. The middle curve shows what happened to the median group.

  Not surprisingly, the high fliers maintain a steady growth rate in their consumption before retirement, but increase their consumption rapidly once they retire, especially with respect to travel and entertainment.

  Remember the case of Max Roth, Ira Roth’s father? He is a typical member of the unprepared, those who do not think ahead to retirement even when it looms in front of them. The day they retire they usually go into shock. As figure 11-2 makes apparent, their consumption expenditures sink like a stone the day they retire. They must rely exclusively on Social Security checks and, if they were fortunate enough to work for a company that offered a pension plan, on pension checks.

  Figure 11-2 Growth Rate of Consumption Expenditures around Date of Retirement

  People vary greatly in their preparation for retirement. Some, the high fliers, save aggressively during their working years, searching out income sources that will replace their wages and salaries when they retire. These people plan to spend more in retirement than they did while working. At the other end of the extreme are people who are completely unprepared for retirement. They do not make any adjustments, even when the onset of retirement is very near. They greet retirement by slashing their consumption spending dramatically. The median case lies in between the two extremes. The average person cuts personal spending noticeably at the onset of retirement.

  The median group falls between the two extremes. Members in this group reduce their consumption expenditures at retirement.

  Bernheim, Skinner, and Weinberg suggest that mental accounting drives the behavior of these groups. People fund consumer expenditures out of their current-income accounts. Those who have replaced income as retirement approached can afford to increase their consumption when they retire. Those who have not will cut consumption, in some cases drastically. These results are consistent with the earlier discussion concerning the use of dividend income.

  Economist Laurence Levin (1992) used the Longitudinal Retirement History Survey to study how people budget their income across different types of consumer expenditures in the years around retirement. The expenditure categories he examined in his study are groceries, eating out, charity, social club dues, entertainment, gifts, transportation, and vacations.

  Levin found very strong evidence of mental accounting. Retirees appear to control their spending by using rules that identify which portions of their wealth are off-limits when it comes to spending in particular categories. Levin divided wealth into several components: current income, liquid assets, home equity, and future income. He finds that liquid assets, home equity, and future income are off-limits for all consumption expenditures except eating out and vacations. All other categories are funded from current income.

  When retirees use assets to fund consumption, they use liquid assets. They very rarely use home equity or borrow against future income. Most retirees are reluctant even to downsize their homes in order to increase their consumer expenditures. In fact, the high fliers actually increase the value of their home equity.

  The reluctance to spend anything except current income is striking. Suppose we take a typical retiree, give him an extra $1,000 in current income per year, and then see how much of it he spends. Suppose the answer is $840. Imagine that instead of giving him the money in the form of current income, we give it to him as a liquid asset, say, a certificate of deposit. How much would we have to give him in order to induce the same $840 increase in consumer expenditure? Levin’s answer is $26,000. Such is the reluctance to dip into assets!

  The $26,000 figure refers to Levin’s entire sample. However, the unprepared group is much more willing to use liquid assets to fund consumer expenditures than other groups. Why? There are probably two reasons. First, they are already in dire straits. Second, their lack of preparedness is an indication that they have failed to cultivate the habits required to achieve a more comfortable retirement. Those with ample resources never touch their assets, except to eat out and take vacations.

  Summary

  A great many people find that they lack the foresight and self-control necessary to save adequately for retirement. But the spectrum is wide. Some people learn the keys to successful retirement planning. They unde
rstand that they need to deal with framing effects and behavioral biases.

  There are some aspects of mental accounting that are detrimental to accumulating retirement wealth, and other aspects that are helpful. Mental accounting can lead people to invest too conservatively by causing them to suffer from myopic loss aversion. At the same time, mental accounting can be constructively used to distinguish different categories of wealth, thereby helping people both to save money and thereafter to safeguard that money from being injudiciously spent.

  Dollar-cost averaging is intriguing from a behavioral perspective. It combines the cultivation of good savings habits to deal with temptation, framing effects that reduce the pain of loss, and conventionality that mitigates feelings of regret.

  Part IV Institutional Investors

  Chapter 12 Open-Ended Mutual Funds: Misframing, “Hot Hands,” and Obfuscation Games

  Think about a legendary mutual fund manager. Whose name comes to mind? Peter Lynch?1

  There are many lessons to be learned from Lynch’s experience managing Fidelity’s Magellan Fund. These lessons concern the way misframing and heuristic-driven bias combine to confuse investors about the relative contributions of skill and luck in fund performance. Investors invariably attribute excessive weight to skill, and consequently they tend to overweight the importance of an individual fund’s track record. Moreover, open-ended mutual funds companies tend to play to investors’ weaknesses by inducing opaque frames.

  This chapter discusses the following:

  • the nature of Peter Lynch’s record as a mutual fund manager, and his attitude towards investing