This paper was prepared for John B. Taylor and Michael Woodford, Editors

4

Siegel (1994, p. 20). However, Siegel notes that the U.S. equity premium was only 1.9% per year

1816–70 and 2.8% per year 1871–1925.

5

Siegel (1994, p. 31).  It should be noted that one must push the investor horizon up to a fairly

high number, around 30 years, before one finds that historically stocks have always outperformed

bonds since 1871; for ten year periods of time one finds that bonds often outperform stocks.  There

are not many thirty-year periods in stock market history, so this information might be judged as

insubstantial.  Moreover, Siegel notes that even with a thirty-year period stocks did not always

outperform bonds in the U.S. before 1871.

7

wonders why people invest at all in debt if it is so outperformed by stocks.

4

  Those who have

tried to reconcile the equity premium with rational investor behavior commonly point out

the higher risk that short-run stock market returns show:  investors presumably are not fully

enticed by the higher average returns of stocks since stocks carry higher risk.  But, such

riskiness of stocks is not a justification of the equity premium, at least assuming that

investors are mostly long term.  Most investors ought to be investing over decades, since

most of us expect to live for many decades, and to spend the twilight of their lives living off

savings.  Over long periods of times, it has actually been long-term bonds (whose payout is

fixed in nominal terms), not the stocks, that have been more risky in real terms, since the

consumer price index has been, despite its low variability from month to month, very

variable over long intervals of time, see Siegel (1994).  Moreover, stocks appear strictly to

dominate bonds: there is no thirty-year period since 1871 in which a broad portfolio of

stocks was outperformed either by bonds or treasury bills.

5

Benartzi and Thaler show (1995) that if people use a one-year horizon to evaluate

investments in the stock market, then the high equity premium is explained by myopic loss

aversion.  Moreover, prospect theory does not suggest that in this case riskless real interest

rates need be particularly high.  Thus, if we accept prospect theory and that people frame

stock market returns as short-term, the equity premium puzzle is solved.

Benartzi and Thaler (1996) demonstrated experimentally that when subjects are asked

to allocate their defined contribution pension plans between stocks and fixed incomes, their

responses differed sharply depending on how historical returns were presented to them.  If

they were shown 30 one-year returns, their median allocation to stocks was 40%, but if they

were shown 30-year returns their median allocation to stocks was 90%.  Thaler, Tversky,

Kahneman and Schwartz (1997) shows further experiments confirming this response.

Loss aversion has also been used to explain other macroeconomic phenomena, savings

behavior (Bowman, Minehart and Rabin, 1993) and job search behavior (Bryant, 1990).

Regret and Cognitive Dissonance

There is a human tendency to feel the pain of regret at having made errors, even small

errors, not putting such errors into a larger perspective.  One “kicks oneself” at having done

something foolish.  If one wishes to avoid the pain of regret, one may alter one’s behavior

in ways that would in some cases be irrational unless account is taken of the pain of regret.

The pain of regret at having made errors is in some senses embodied in the Kahneman–

---

8

Tversky notion of a kink in the value function at the reference point.  There are also other

ways of representing how people behave who feel pain of regret.  Loomes and Sugden

(1982) have suggested that people maximize the expected value of a “modified utility

function” which is a function of the utility they achieve from a choice as well as the utility

they would have achieved from another choice that was considered.  Bell (1982) proposed

a similar analysis.

Regret theory may apparently help explain the fact that investors defer selling stocks

that have gone down in value and accelerate the selling of stocks that have gone up in value,

Shefrin and Statman (1985).  Regret theory may be interpreted as implying that investors

avoid selling stocks that have gone down in order not to finalize the error they make and not

to feel the regret.  They sell stocks that have gone up in order that they cannot regret failing

to do so before the stock later fell, should it do so.  That such behavior exists has been

documented using volume of trade data by Ferris, Haugen and Makhija (1988) and Odean

(1996b).

Cognitive dissonance is the mental conflict that people experience when they are

presented with evidence that their beliefs or assumptions are wrong; as such, cognitive

dissonance might be classified as a sort of pain of regret, regret over mistaken beliefs.  As

with regret theory, the theory of cognitive dissonance (Festinger, 1957) asserts that there is

a tendency for people to take actions to reduce cognitive dissonance that would not normally

be considered fully rational:   the person may avoid the new information or develop

contorted arguments to maintain the beliefs or assumptions.  There is empirical support that

people often make the errors represented by the theory of cognitive dissonance.  For

example, in a classic study, Erlich, Guttman, Schopenback and Mills (1957) showed that

new car purchasers selectively avoid reading, after the purchase is completed, adver-

tisements for car models that they did not choose, and are attracted to advertisements for the

car they chose.

McFadden (1974) modelled the effect of cognitive dissonance in terms of a probability

of forgetting contrary evidence and showed how this probability will ultimately distort

subjective probabilities.  Goetzmann and Peles (1993) have argued that the same theory of

cognitive dissonance could explain the observed phenomenon that money flows in more

rapidly to mutual funds that have performed extremely well than flows out from mutual

funds that have performed extremely poorly:  investors in losing funds are unwilling to

confront the evidence that they made a bad investment by selling their investments.  

Anchoring

It is well-known that when people are asked to make quantitative assessments their

assessments are influenced by suggestions.  An example of this is found in the results survey

researchers obtain.  These researchers often ask people about their incomes using

questionnaires in which respondents are instructed to indicate which of a number of income

brackets, shown as choices on the questionnaire, their incomes fall into.  It has been shown

that the answers people give are influenced by the brackets shown on the questionnaire.  The

tendency to be influenced by such suggestions is called “anchoring” by psychologists.

---