Home > Subjective Well-Being and Utility in Psychology and Economics
Equation
Chapter 1 Section 1
Utility
and Happiness
by Miles
Kimball and Robert Willis1
University of
Michigan
March
3September 11, 20065
Abstract: Psychologists have developed effective survey methods of measuring how happy people feel at a given time. The relationship between how happy a person feels and utility is an unresolved question. Existing work in Economics either ignores happiness data or assumes that felt happiness is more or less the same thing as flow utility. The approach we propose in this paper steers a middle course between the two polar views that ��happiness is irrelevant to Economics�� and the view that ��happiness is a sufficient statistic for utility.��
We argue that felt happiness
is not the same thing as flow utility, but that it does have a systematic
relationship to utility. In particular, we propose that happiness
is the sum of two components: (1) elation--or short-run
happiness--which depends on recent news about lifetime utility and (2)
baseline mood--or long-run happiness--which is a subutility function
much like health, entertainment, or nutrition. In principle,
all of the usual techniques of price theory apply to baseline mood,
but the application of those techniques is complicated by the fact that
many people may not know the true household production function for
baseline mood.
If this theory is on target,
there are two reasons data on felt happiness is important for Economics.
First, short-run happiness in response to news can give important information
about preferences. Second, long-run happiness is important for
economic welfare in the same way as other higher-order goods such as
health, entertainment, or nutrition.
1. Introduction
On first impression, ��utility��
and ��happiness�� seem to refer to the same concept.
However, over the last century, economists and psychologists respectively
have developed technical meanings for the words ��utility��
and ��happiness�� that refer to logically distinct concepts.
The success of the Ordinalist
Revolution of Lionel Robbins (1932) and of John Hicks and R. G. D. Allen
(1934)—codified as ��Revealed Preference�� by Paul Samuelson (1938,
1947)2—has fixed the meaning of ��utility��
for more than a half-century of economists as a representation of an
individual��s preferences over alternatives. The practice
of Economics has made this concept of utility immensely valuable in
thousands of applications.
In the aftermath of the Cognitive
Revolution, the success of Hedonic Psychology—exemplified in the volume
edited by Daniel Kahneman, Ed Diener and Norbert Schwarz (1999)—has
fixed the scientific meaning of ��happiness�� within Psychology
as the overall goodness or badness of an individual��s felt experience
at any point in time. To be more explicit, operationally, psychologists
define current happiness as how people answer questions such as ��On
a scale from one to seven, where one is extremely
unhappy and seven is extremely happy, how do you feel right now?��
This concept of happiness has attracted increasing interest among economists
in recent years.
Throughout this paper, we follow
the convention that the technical meaning of ��utility�� is
determined by the tradition in Economics, while the technical meaning
of ��happiness�� is determined by the tradition in Hedonic
Psychology. Thus, utility is a reflection of people��s
choices; happiness is a reflection of people��s feelings.
Once one recognizes these two concepts as distinct, discovering the
nature of the empirical relationship between utility and happiness
stands out in sharp relief as one of the central questions at the frontier
between Economics and Psychology.
In the existing
literature attempting to link utility and happiness, the dominant explicit
or implicit hypothesis is that current felt happiness is equal to flow
utility.3 We argue that the hypothesis
that felt happiness equals flow utility is empirically untenable.
Instead, to oversimplify our discussion below, we argue, in effect,
that a large component of happiness is much more like the recent
change or innovation in lifetime utility than it is like
flow utility.
Of course,
even unchanging, predictable circumstances can have an effect on happiness,
so it is important to allow for another, longer-lasting component of
happiness. We argue that this long-run component of happiness is
not always aligned with utilityalso distinct from utility
itself, since people often knowingly and without regret make decisions
that sacrifice a pleasant mental state day after day for the sake of
some other goal.
Thus, in our view, happiness
is the sum of a transitory response to good and bad news and a long-run
response of mood to circumstances that is distinct from utility.
To be specific, we propose that happiness is the sum of two components:
(1) elation--or short-run happiness--which depends on recent
news about lifetime utility and (2) baseline mood--or long-run
happiness. Baseline mood
--which is a subutility function—
(or output of a household production function—)
much like health, entertainment, or nutrition. In other words,
long-run happiness is a ��valuable commodity,�� that cannot be purchased
directly, though inputs to it can be.
Such a theory of happiness
not only makes sense of existing happiness data, but provides a road
map for future research. According to this theory, data on felt happiness
can make two contributions to Economics.
First, short-run happiness in response to news can give important information
about preferences. Second, long-run happiness is important for
economic welfare in the same way as other higher-order goods such as
health, entertainment, or nutrition.
Policy issues surrounding long-run happiness arise
because of the value of producing and
disseminating knowledge about the household production function for
happiness and from any externalities
in the causes or effects of long-run happiness.
Desmond Morris, at the
outset of his wonderful little book
The Nature of Happiness, writes:
��The true nature
of happiness is frequently misunderstood. It is often confused
with contentment, satisfaction or peace of mind. The best way
to explain the difference is to describe contentment as
the mood when life is good, while happiness is the sensation we experience
when life suddenly gets better. At the very moment when something
wonderful happens to us, there is a surge of emotion, a sensation of
intense pleasure, an explosion of sheer delight—and this is the moment
when we are truly happy. Sadly, it does not last very long.
Intense happiness is a transient, fleeting sensation. We may continue
to feel good for quite a while, but the joyful elation is quickly lost.��
Morris��s
description of ��happiness�� emphasizes what we call
elation—the word Morris also uses to describe this type of
happiness. The ��contentment�� he refers to is close to our
concept of baseline mood, which unlike Morris, we also consider
a fully legitimate component of happiness, since both the contentment
when life is good and the joy when life gets better are likely to affect
measured subjective well-being.
A word
is in order about the length of this paper. We have learned from
experience in talking to colleagues and others
that because of the widely varying preconceptions
almost everyone has on the subject of happiness,
the perspective we propose on happiness is easy to misunderstand.
Therefore, we make an effort to lay
out the issues very carefully.
Moreover, as we discuss below, there is an existing consensus among
most psychologists and economists who are involved in studying happiness
with which we disagree. It is incumbent upon us to make clear exactly
why we disagree with the existing consensus, which requires a reexamination
of all of the key types of evidence that are used to back up that consensus.
The remainder of the paper
can be divided into two halves. The
first half, Sections 2-5, is conceptual.
In it we make the case for utility and happiness as logically and empirically
distinct concepts. The second half, Sections 6-10,
is mathematical. In it we lay out a specific model of the relationship
between utility and happiness, along with
interpretations, extensions and applications.
Such a theory of happiness not only makes sense of existing
happiness data, but provides a road map for future research. According
to this theory, data on felt happiness can make two contributions to
Economics. First, short-run happiness in response to news can
give important information about preferences. Second, long-run
happiness is important for economic welfare in the same way as other
higher-order goods such as health, entertainment, or nutrition.
2. Distinguishing Between Utility and Happiness
A. The Need to Establish
Clear Terminology. One of the difficulties we face in explaining
our this viewpoint is that the tradition
of equating ��happiness�� to flow utility runs deep in the history
of economic thought. Indeed, Jeremy Bentham��s (1781) first definition
of ��utility�� made the equation of utility and happiness explicit:
��By the principle of
utility is meant that principle which approves or disapproves of every
action whatsoever according to the tendency it appears to have to augment
or diminish the happiness of the party whose interest is in question
��.��
The ��Revealed Preference�� definition of utility—to which we resolutely adhere—is closer to Bentham��s second, more inclusive, definition of utility, in the immediately following paragraph:
��By utility is meant
that property in any object, whereby it tends to produce benefit, advantage,
pleasure, good, or happiness, (all this in the present case comes to
the same thing) or (what comes again to the same thing) to prevent the
happening of mischief, pain, evil, or unhappiness to the party whose
interest is considered: if that party be the community in general, then
the happiness of the community: if a particular individual, then the
happiness of that individual.��
Another difficulty we face
in distinguishing utility and happiness is that, while ��Revealed Preference��
guides economic research, a more naïve Marginalism has remained very
common in economic teaching. For example, ��Principles of Economics��
courses often teach about diminishing marginal utility by engaging students��
intuitions about how happy they would feel in consuming different consumption
bundles.
Let us state clearly that,
throughout this paper, when we discuss utility, we do so from the perspective
of Paretian Welfare Economics. Whether explicitly or implicitly, welfare
questions motivate a large share of economic research; an orientation
toward welfare questions is particularly important in informing our
assessment of utility in cases where people are liable to mistakes.
As for the focus on Pareto optimality, in our view, the use of happiness
data is not a Philosopher��s Stone that magically solves the
difficulties in comparing utility interpersonally, but happiness data—used
judiciously—can give useful information about individual preferences.4
Any adequate theory of utility
and happiness must explain why the meanings of happiness and utility
seem so similar. The right nuances for explaining the semantic
relationship between ��happiness�� and ��utility�� can be found
in the first two definitions for ��happy�� in the American Heritage
Dictionary (1976, Houghton Mifflin):
happy �� 1. Characterized by luck or good fortune; prosperous. 2. Having or demonstrating pleasure or satisfaction; gratified.����
The second definition is the
meaning of ��happy�� in Psychology. The first definition talks
about prosperity, which seems closely linked to utility, but there is
a hint of a stochastic element in the nature of happiness: ��luck or
good fortune.�� Our view of happiness emphasizes recent
good luck by positing that an important component of happiness has to
do with an individual��s reaction to recent news about lifetime utility.
Although the differences are important, news about lifetime utility
and lifetime utility itself are linked tightly enough that it is not
surprising to find a certain confusion between the two in the structure
of the lay lexicon. In other words, if people feel happy whenever
they receive good news about lifetime utility, it is not hard to see
why they would sometimes use the word ��happiness�� to describe lifetime
utility itself. Yet scientifically, we consider it crucial
to have two distinct, clearly delineated concepts for revealed preference
��utility�� and ��happiness��
in the psychological sense of current feelings. Maintaining two distinct
concepts—on an equal footing—in a situation where each has a certain
tendency to subordinate or engulf the other, is one of the main contributions
of this paper.
One way to think about the distinction between utility and happiness is that one��s commitment to an Ordinalist, ��revealed preference�� definition of utility is confronted with an acid test when confronted with happiness data. There is a sense in which the most radical implications of the Ordinalist Revolution are apparent only in the light of data on experienced happiness.
Both felt happiness and choice-based
utility are well-defined, observable concepts. Our aim is to determine
the dynamic relationship between the standard psychological concept
of current affect—felt happiness—and the standard economic concept
of lifetime utility. Establishing any systematic relationship
between happiness and utility would provide an important bridge between
Psychology and Economics, allow psychological data and theory to be used
in Economics in a way that is complementary to standard economic data
and theory, and enable economists to bring to bear all the tools of
economic theory toward understanding happiness.
B.
Distinguishing Between Utility and Happiness
as a Matter of Logic. In Psychology, the term ��subjective
well-being�� refers to a multidimensional concept that includes
evaluations of one��s life-as-a-whole and of specific life-domains
as well as the pleasantness of one��s average experienced affect. Though
the terminology has not been entirely standardized in the literature,
affect is a useful term to refer to how happy a respondent currently
feels, as opposed to judgments about his or her whole life. An
attractive feature of affect measures is that the cognitive burden they
place on respondents is modest in contrast to the extremely difficult
cognitive task of forming a judgment about the quality of one��s entire
life. Throughout this paper, we use ��current affect�� and ��happiness��
interchangeably.
Economists have been slower
than psychologists to focus on subjective well-being data. But
a growing economic literature has made use of subjective well-being
data. Richard Layard��s (2005) book gives a good introduction
to this literature and Bruno Frey and Alois Stutzer (2002) give a partial
survey review.5 This literature lays out many
provocative findings, but with a few exceptions, the focus of this literature
has been on the cross-sectional and trend properties of subjective well-being
rather than on its detailed dynamic properties. Two key motivations
for the use of subjective well-being data in Economics (shared in large
measure by Hedonic Psychology itself) have been (i) the desire to study
the welfare implications of non-traded goods6 (something that is especially important
for older people for whom market work is a less dominant part of their
lives) and (ii) the desire to study welfare implications in contexts
where preferences are potentially inconsistent and to diagnose optimization
mistakes.7
Despite this growing literature,
many economists are still very skeptical of the use of subjective well-being
data,8 in large part because the theoretical
status of affect--��happiness��—within economic theory is unclear.
A simple multiple-choice question illustrates this lack of clarity:
What is Happiness?
To begin to answer this kind
of question, it is important first to distinguish utility and happiness
as a matter of logic. Then the relationship between utility and
happiness will ultimately be an empirical matter. Using the shorthand
��lifetime utility��
to refer to an individual��s overall objective function—including
things the individual cares about that occur after his or her
death—we can distinguish lifetime utility and current affect (��happiness��)
as follows:
In thinking about lifetime
utility, it is important to remember that people��s choices clearly
show that they value a wide range of goods that are not traded in markets
or only partially traded in markets. Thus, the economic concept
of lifetime utility is not limited to what are sometimes called ��economic
goods�� but includes the value an individual places on non-traded
goods such as respect, freedom, clean air, a vibrant community,
being married to a particular person, and such partially-traded goods
as time allocations--which are partially traded because people are paid
for work time---and health and longevity--which are partially traded
because people pay for health care.
C. Utility and Happiness
as Empirically Distinct Candidates for a Welfare Measure.
Lifetime utility is the standard welfare measure in economics at the
individual level. It is often thought of as a discounted sum over
time of ��flow utility.�� As a counterpoint to this, Kahneman
(1999), in a chapter that has been influential among psychologists who
study well-being, has urged a discounted sum over time of affect (momentary
experienced happiness) as the appropriate measure of overall individual
welfare.9 A prima facie
case can be made for each of these views. Both subjective well-being
and utility are based on trusting an individual��s own judgment, but
different judgments are trusted in each case: as a welfare measure,
lifetime utility puts trust in an individual��s (conscious and subconscious)
judgments as reflected in choices, while the discounted sum of affect
puts trust in an individual��s (largely subconscious) judgments as
expressed in feelings.
It would be very convenient
if flow utility and affect were essentially equivalent; in that case
the standard economic measure of individual welfare would match Kahneman��s
(1999) proposed measure of individual welfare. Unfortunately,
things are not so easy. In brief, as we discuss below, affect
seems to behave very differently from at least our traditional notions
of the behavior of flow utility:
Clearly, one could attempt
to modify either one��s ideas about utility or the mode of measuring
happiness to try to bring flow utility and affect closer together.
We argue that it is better to accepttake as given
utility as determined by standard, best-practice economic methods
of measurement, and affect as determined by standard, best-practice
psychological methods of measurement, then pose as an open-ended
question the nature of the relationship between these two
concepts. We make a specific proposal for what this relationship
might be, but we consider the question—thus posed— more important
than our attempt at an answer.
3.
Measuring Happiness
The logical distinction between
happiness and utility becomes clearer the more
closelyr one pays
the attention paid
to the way each concept is measured. In this section we argue that
psychologists can reliably measure happiness, in the carefully defined
sense of how people feel at a given time. Of course, that leaves
the question of what happiness is.
To say the same thing in a
different way, some economists think happiness can��t be measured well.
This is just not true. Happiness (current affect) is one of
the easiest of all subjective concepts to measure. What is true
(that these economists are intuiting) is that once happiness is measured,
we don��t know what it means in terms of economic theory.
Psychologists have taken
measurement issues in assessing emotions
in general, and happiness in particular,
very seriously. Randy Larsen and Barbara Fredrickson (1999) give a
survey of research touching on this issue.
Self-report measures of happiness and sadness
(the most common type of measure) have been related to
impressionistic observer ratings of happiness, highly-structured coding
of facial expressions by trained observers,
assessment of voice tone, electromyographic measurement of face muscle
activation, measurement of skin conductance, heart-rate, blood pressure
and respiration, electro-enchephalograms, positron emission tomography
and functional magnetic resonance imaging of brain activity (where
����approach related positive emotions are associated with left anterior
activation whereas withdrawal related negative emotions are associated
with right anterior activation�� Larsen and Fredrickson (1999), p.
53.) Self-report measures of happiness and sadness
have also been shown to predict many types of
cognition and behavior in the laboratory,
including writing speed and performance speed on other tasks, judgments
of probabilities, the output of free word association
and word completion under time pressure,
speed of judging positive and negative
words versus nonwords, and the speed of the startle reflex after a loud
sound. All of these experiments
add up to consistent evidence that
happiness is a measurable psychological state.10
Among
self-report measures of happiness, tThe gold standard
for measuring happiness is experience sampling, in which people
are signaled at random intervals to report their current happiness.
Kahneman, Alan Krueger, David Schkade, Schwarz, and Arthur Stone (2004)
argue that the day reconstruction method is a close second. Measuring
happiness as part of a large-scale survey presents an extra issue in
that the survey itself may represent a significant slice of a day. To
avoid too much emphasis on the feeling states engendered by the interview
process itself one can ask about happiness over a longer, but still
relatively short span of time.11
The Health and Retirement Study measures affect by the following series
of questions:
��Now think about the past week and the feelings you have experienced. Please tell me if each of the following was true for you much of the time this past week: 12
a. Much of the time during the past week, you felt you were happy. (Would you say yes or no?)
b. (Much of the time during the past week,) you felt sad. (Would you say yes or no?)
c. (Much of the time during the past week,) you enjoyed life. (Would you say yes or no?)
d. (Much of the time during the past week,) you felt depressed. ( Would you say yes or no?)��
Operationally, one can treat
happiness as the latent variable behind these four yes/no questions.
This series of questions on the Health and Retirement Study is a subset
of a series of questions that
represent one version of the Center for Epidemiologic Studies
Depression (CES-D) measure of depressive symptoms.13
These questions illustrate what we mean when we say that the concept
of happiness we are referring to is about current feelings. These
questions ask about easily accessible feelings and memories of feelings
in the past week. One indication of how readily respondents
answer these questions is that the average amount of survey time required
for all four questions put together is less than 35 seconds.
In terms of external validity, self-reported current happiness measures
of this general type are strongly correlated with
frequency of smiling, others�� ratings of how happy the respondent
is, social rank and high activity in the left pre-frontal cortex and
low activity in the right pre-frontal cortex. (Brain activation
patterns can be seen as a validation of a happiness measure since high
activity in the left pre-frontal cortex and low in the right can also
be induced by seeing pictures of a smiling baby, while the opposite
can be induced by seeing pictures of a deformed baby.14)
It is important to contrast
current affect measures like those on the HRS
, with life satisfaction measures, such as those on the German
Socioeconomic Panel—��On a scale from 1 to 10, how
satisfied are you with your life?��—and ��global happiness��
questions, such as those on the World Values Survey:
��Taking all things together, would you say you are
An extensive body of psychological
research explores the cognitive processes underlying global judgments
of happiness and life-satisfaction (for a review and process model see
Schwarz and Strack, 1999). It converges on the following conclusions15:
This context-dependence of
evaluative measures of well-being attenuates any meaningful relationship
with objective circumstances of life and motivates approaches to the
measurement of well-being based on people��s momentary affective experience.
(Kahneman, 1999).
In comparison to global evaluations
of one��s life-as-a-whole, assessments of current affect pose more
reasonable cognitive demands. As noted in point 3 above, experimental
evidence suggests that survey responses to questions about overall life
satisfaction or about global happiness with life rely heavily on the
readily accessible internal information a respondent has about current
affect (Schwarz and Clore, 1983).17 Thus, how a respondent feels
right now has a strong effect on answers to overall life-satisfaction
and global happiness questions, whether we like it or not. We
maintain that it is clearer to focus on current happiness directly,
so that we know what we are getting, in a transparent way. Finally,
to the extent that respondents are not using current affect as a shortcut
to make an overall evaluation of life satisfaction or global happiness,
there is a serious danger that they will report how happy or satisfied
they think they should feel about
with their lives according to whatever folk theories they
have about happiness and satisfaction.18
4. Measuring
Utility
For economists, a discussion
of measuring utility is only a reminder. Utility is defined by
revealed preference—the information gleaned from the choices people
make. Some of the accumulated wisdom from economic research
is encoded in standard functional forms that are repeatedly applied
and tested. The techniques of revealed preference can be applied
to tradeoffs and preferences
over seemingly incommensurable values, and apply even to situations
involving choices over time.
For non-economists, one can
say that the concept of utility relies on an individual��s judgment
of his or her priorities, as reflected especially in his or her actual
choices when faced with a tradeoff, or, at a minimum, his or her choices
in a hypothetical situation. Higher utility out of any two choices
is defined by what the individual chooses (or would choose) when presented
with those two choices. Thus, utility is a measure of the extent
to which people get what they want, and differences in utility are predictors
of behavior. This allows a deep connection between positive (descriptive)
and normative (prescriptive) aspects of utility theory in Economics.
A.
The Upward Trend in Utility.
In view of the Easterlin (1974, 1995, 2003)
Paradox (Easterlin, 1974, 1995, 2003)
of secularly nontrending happiness, an important application of the
principle of revealed preference is to a hypothetical choice between
the comprehensive extended
consumption bundle (including all externalities, public goods and time
use patterns) now and the comprehensiveextended
consumption bundle fifty years ago in the U.S. Real
per capita income has grown dramatically over that period of time, which
means that the total set of marketed consumption bundles that people
can choose from has expanded. Higher real per capita income
allows people more choices, out of which they typically elect
choose to spend in ways they could not previously afford--rejecting
the available option of continuing to spend in the same way they did
at the lower income level. Average work hours have trended slightly
downward. Moreover, as Easterbrook (2003) points out, a large
set of goods not traded or only partially traded in the market have
either stayed about the same or improved over the last fifty years.
Among partially-traded goods, medical care and longevity have been improving
dramatically, while household conveniences have reduced the time necessary
for housework and increased the time available for genuine leisure even
for many who do spend longer hours in market work. The internet
has provided a vast fund of free entertainment and advancing scientific
knowledge available in any public library;
Eequality between the sexes and races, while far
from complete, is much better than two generations ago; the number of
democratic nations is on the rise; and even the War on Terror, which
at worst could involve the nuclear destruction of a large portion of
Manhattan, is an improvement over the Cold War, which at
worst could have destroyed human life from the face of the earth.
Finally, many of the non-traded goods that worsened for a while after
1955, such as rates of crime, teenage pregnancy and drug abuse, have
turned the corner and begun trending in a favorable direction for the
last two decades. In short, although many problems remain,
and are the focus of nightly news reports, we argue that it would be
a bad deal to trade the problems we face today for the problems of yesteryear,
implying that utility is higher than fifty years ago.
Of course there are those who
look back at the past with nostalgia. In part, the increasing
individual freedom that comes with higher per capita income may have
some undesirable side effects such asin a diminished
sense of community, of the sort Robert Putnam (2000) describes in his
book Bowling Alone: The Collapse and Revival of American Community.
(Trends in the divorce rate and other aspects of family structure can
be seen as part of the same phenomenon.) Similarly, the rise in
per capita income may have increased the availability of illegal drugs
and access to a wide variety of delicious foods and drinks that create
important intra-psychic conflicts. Still, how many would really want
to go back to the way it used to be if they saw clearly the way it
really used to be? It is easy to forget the legitimate and
the irrational fears engendered by the Cold War, the toll of racial
and other injustices on those mistreated, the enforced conformity that
went along with the greater sense of community in the past, and how
effective the long-available drug of alcohol is at messing up the lives
of those who are prone to intra-psychic conflict. And it
is easy to take for granted boons such as word processing, the ability
to watch any of a huge range of movies at home, the existence of J.
K. Rowling��s Harry Potter series, and inexpensive access by
means of free internet access at the public library to a huge range
of fascinating scientific findings that were not known fifty years ago,
let alone available at the click of a mouse.
Besides ordinary selective
memory that often leads us to
forgets former difficulties once they have been surmounted,
there is another kind of bias that helps to fuel such nostalgia: despite
recent trends in historiography toward telling the stories of those
at the bottom of the social ladder19 as well as the stories of common men
and women, our image of the past is still often dominated by the biographies
of those near the top of the social ladder who were much better off
than the average person in their time. Even when we assess the
past by thinking of the experience of our own grandparents, they are
far from a random sample of people in their time. The ancestors
of a randomly chosen individual in the present are likely to be people
who were more successful than average in number of descendants--and
likely to be above average in the degree of success in their life experience
more generally.20 The experience of those who
died young in the past or who never found mates is not remembered as
well as the experience of those who did. The travails of those
in the present who die young or who never find mates is more apparent.21
One reason it is important
to important to hold imagined social rank constant when assessing
the past versus the present is that most people care a lot about social
rank.22 From the standpoint of revealed
preference, it is not difficult to observe people making choices that
sacrifice other valuable things in order to attain higher social rank.
As a consequence, it is not incredible that someone might choose to
be a king or queen in a bygone era rather than a middle-class person
today, even if the real value of the market consumption bundle of the
middle-class person today is worth much more. Individuals�� positional
concerns are not irrational. In many domains of life, relative standing
is more crucial to obtaining desired outcomes than absolute standing
(for discussions see Frank, 1985, 1999; Hirsch, 1976; Sen, 1983). Social
rank yields hard-to-measure but real benefits in terms of respect and
favorable treatment by others. Because, in practice, social rank
is so highly correlated with income, at least in the United States,
secular comparisons are useful for distinguishing concern with income
from concern with social rank.23
The point of this extended
discussion is to argue that average lifetime utility for people of a
given age is higher than it was fifty years ago, even after accounting
for a wide range of tradeoffs going far beyond those that are in obvious
monetary terms. Of course, the choice between the comprehensive
consumption bundle of fifty years ago and the comprehensive consumption
bundle now is a hypothetical choice. But every year millions of
people make a choice that is similar in important respects by migrating
from a poor home country to the U.S. and other rich countries. Many
leave behind tight-knit communities in which they have high local social
status for a foreign land where they will be at the bottom of the social
status ladder and where they cannot even speak the language. Clearly,
in making the enormous effort of migrating, with all the psychic costs
of being uprooted from one��s familiar cultural surroundings, they
are choosing something that they value highly—the modern consumption
bundle in the U.S. and other rich countries.24
B.
Mistaken Choices. The greatest difficulties in measuring utility
arise when people make mistaken choices or have inconsistent preferences.
Garden-variety mistakes based
on a lack of knowledge of objective facts are the easiest to deal with.
Consider the case of someone who chooses a particular car, thinking
that it will get good gas mileage, but then discovers that it gets bad
gas mileage. Once she learns this, the purchaser regrets the earlier
decision and wishes that she had chosen a different car. In this
case, in judging utility one needs to either use the choice the purchaser
would have made if fully informed, or
the choice the purchaser did make between ideas of cars with assumed
characteristics.
Sometimes a mistake arises
not from lack of basic facts, but from a failure of computation.
For example, one of the authors has only very recently begun to adjust
his book-buying habits for the high shadow-cost of available book-reading
time that is generated by the large number of books already in his personal
library, by thinking ��What are the chances that on any future date
this book will win out over all of the competition?�� This is
the kind of reasoning that does affect one��s choices when the calculation
has finally been made. Taking Paretian Welfare Economics as our
touchstone, we consider it relatively uncontroversial to suggest that
utility (strictly speaking, preferences) be measured according to what
people would choose when not only well-informed in terms of raw information,
but also when they are aware of relevant calculations and lines of reasoning.25
A third type of mistake is
making mistakes about what one��s subjective experience will be after
a given choice. There is nothing disruptive of standard economic theory
about the existence of experience goods, such as a new flavor of ice
cream, for which preferences are known only after trying some of the
good. Marketing strategies by firms selling experience goods
vary from actively providing a free taste to forcing people to buy a
substantial package based on guesswork. When free samples are
not provided, it is easy to make mistakes due to not fully knowing one��s
own tastes, even if the physical properties of the product itself are
well-known.
For an expensive durable good,
an optimal decision of whether to purchase the good should involve considering
the time-path of one��s subjective experiences with the good.
It is not hard to imagine someone changing her mind about buying an
expensive car upon being shown evidence that after a year people report
roughly the same experience when driving an expensive car as when driving
a much cheaper car—say because driving is one of those activities
that becomes reasonably automatic and so fades into the background of
awareness—pushed out by thoughts of where one is going to and where
one has just been. Indeed, people might not even need formal
evidence for things like this; their own past experience of being less
excited by a new good after the first few months could inform their
later decisions. However, a considerable body of evidence reviewed
in George Loewenstein and David Schkade (1999) indicates that people
make serious mistakes in predicting future affect. Though it is
larger in size and scope, one can view a mistake in predicting future
affect as akin to a mistake in predicting whether one will like a particular
type of ice cream. It is not clear whether people are making the
right decisions or not until they are well-informed about the modification
of the time-path of affect that will actually result from a purchase.
Some psychologists have gone
further, to maintain that the fact that reported happiness with a new
car is often high in the first few months after purchase and much lower
thereafter—in a way that people are bad at predicting—necessarily
means that someone has made a mistake in purchasing it. To our
way of thinking, this is going too far. The key question is whether
a correct knowledge of the modification of the time-path of affect that
will actually be induced by a good would make a material difference
to a decision. The issue becomes clearest if we consider the choice
of a purchaser who is fully conversant with the Hedonic Psychology literature
and carefully observant of the pattern of his or her own affective reactions.
As long as the purchaser is aware of and thoughtfully considers the
fact that happiness with a new durable is likely to fade after a time
(and absent the kind of inconsistency discussed below), it seems appropriate
to defer to that well-informed, thoughtful decision in judging utility,
regardless of the time-path of subjective experience with the new car.
Indeed, there is every reason to think that people care about many attributes
of a car other than its price and the subjective experience they will
have with it—such as its ability to get reliably to work and back.
Even the set of all indirect effects of a purchase on affect (for example,
including the reduced likelihood of sorrow from being scolded for getting
to work late) should not necessarily be dominant in the decision of
whether to make a purchase. The concept of utility (or equally
in this context, preferences) involves deferring to each individual��s
own view of how much to factor in the modification of the time-path
of affect that would result from a purchase when making the decision
of whether to buy or not to buy. Since the consequences for affect
are only one aspect of a good, it would not necessarily be irrational
to give those consequences only a small weight even after understanding
them fully.
A fourth type of mistake is
described by Barry Schwartz (2004) in his well-publicized recent book
The Paradox of Choice. He emphasizes the mistake of trying
��too hard�� to optimize. The ��maximizers�� he identifies
by an abbreviated personality test seem to optimize without regard to
the costs of the time and effort devoted to deliberation about a choice.
Of course, this is not true optimization in a larger sense; a fully
optimal choice must take into account deliberation costs. However,
this raises an important issue for the measurement of utility.
We argue that the utility function for everything other than deliberation
costs should be measured by the choices that would be made by exactly
such an agent who disregards deliberation costs.26
Where people��s preferences are similar to one another, this concept
of utility approximates the utility that can be achieved when a small
jury selected from a large group of people with similar preferences
pays the deliberation costs for everyone in the whole group. For
example, Consumer Reports is a practical effort to help people
approximate this level of utility.
The general point that arises
from thinking about mistakes is that, even when underlying preferences
are fully consistent, choices arise from the interaction of preferences,
ordinary constraints and information structures.
The relevant information structures can include both external information
constraints and internal cognitive constraints. In principle,
internal cognitive constraints are no harder for economic theory to
deal with than external informational constraints. For example,
Woodford��s (2002) model of monetary nonneutrality
(based on the model of rational inattention in Christopher
Sims, 2002) a lack of common knowledge
has been criticized for relying on extremely low bit transmissions rates.
One way to defend Woodford��s model is to locate the low bit transmission
rates inside the small portion of the typical decision-maker��s psyche
devoted to thinking about macroeconomics.
Rayo and Becker (2005) provide
an interesting example of explicitly modeling decision-making as the
outcome of an underlying utility function (��evolutionary efficiency��)
filtered through internal informational constraints. Assuming
a limit on the total number of gradations into which values of the underlying
utility function can be distinguished, they show that an optimal deployment
of that limited available total precision is to make fine distinctions
in the neighborhood of values where an agent will actually be operating
but only gross distinctions at outlying values.27
Could consideration of mistakes
overturn the conclusion of subsection A that utility is rising?
It is instructive to consider again the choice hundreds of thousands
of people make every year to migrate from a poor home country to a foreign
rich country. Certainly some regret their decision ex post
and wish they had never migrated, and some even pay the fixed cost of
returning to their home country to a situation no richer than before.
But among those who are able in the end to bring their families as well,
there is no evidence of widespread regret at migrating to a rich country.
There is even less regret among the grandchildren of those who migrated,
who escape most of the large fixed costs of migrating.
C.
Inconsistent Preferences.
Inconsistent preferences are more difficult to deal with than mistakes.
Just as mistakes involving one��s own preferences can be modeled as
an underlying utility function together with internal information acquisition,
transmission and processing constraints, inconsistent preferences are
now routinely modeled as an intra-psychic game between multiple agents
within the same person, each having a distinct set of preferences.
However, if there is more than one set of preferences operating within
a single individual, normative analysis faces a version of the Social
Choice problem even for evaluating individual welfare. Unlike
the standard Social Choice problem, there is no reason for a presumption
of equal ethical value for all the different intra-psychic agents.
In order to do normative welfare
analysis even at the individual level, one must take some stand on this
intra-psychic social choice problem. Our proposal in this regard
would be to rely, in all ordinary cases, on the psyche��s own dispute
resolution system. As Camerer, Loewenstein and Prelec (2005) discuss,
the brain��s controlled cognitive system is often brought into action
to use deliberation to resolve disputes between other systems in the
brain. Thus, we consider well-informed, thoughtful, revealed preference
to be the best practical gold standard for an individual��s preferences
for the purposes of welfare analysis at the individual level.28
When multiple preferences coexist
within the same individual, the whole-person utility function that is
the solution to the intra-psychic social choice problem may not be the
utility function that has the tightest relationship with happiness.
Here, a key issue is which ��selves�� give affect reports to an interviewer.
A relatively straightforward case is when the problem is short-sightedness
in the sense of hyperbolic discounting of the sort described by David
Laibson (1997), where there is one ��self�� in command at each point
in time, in a known sequence. Other cases could be more complex.
Although we consider it a high priority for the future, modeling the
relationship between reported happiness and either whole-person utility
or the set of utility functions within the psyche is beyond the scope
of this paper. From here on, in discussing the relationship between
happiness and utility, we will assume that the individual has only one
set of preferences, which are internally consistent.
D.
Is Happiness in the Utility Function?
The principle of revealed preference indicates that happiness is in
the utility function. Hundreds of thousands of people spend
thousands of dollars each on therapy that is not reimbursed by insurance
in hopes of becoming happier or at least less unhappy. Millions
of people endure the significant negative side effects of chemical antidepressants
in order to feel happier. Self-help books and magazines featuring cover
articles on happiness sell briskly. Moreover, many products
that may not actually make one��s psychological state significantly
more positive are advertised as if they will, as described in great
detail by Melinda Davis (2002). Advertising aimed at suggesting
that a product will improve one��s brain state would not be so prevalent
if a desire for positive affect were not an element of preferences.
In sum, many people want to feel happier and are willing to sacrifice
other things in order to attain that psychological state.
On the other hand, it seems
clear that happiness is neither the only thing in the utility function
nor a sufficient statistic for all of the goods that are in the utility
function. To make this clear it is best to use the technical term
��affect�� for happiness as a reminder that we are talking about feelings.
People care about things other than how they feel. Most obviously,
they sometimes sacrifice current affect for a later benefit; for example
(to take one of people��s lowest affect activities according to Kahneman,
Krueger, Schkade, Schwarz and Stone, 2004) it makes sense to spend time
on household chores despite the low momentary affect associated with
that activity because of the later benefits of having a clean house.
There might be a later increase in affect as a result, but it is not
clear that the later benefits show up in affect and utility in the same
way.29 Second, people often sacrifice
their own affect to benefit their children, as when one spends long
hours at a grueling job to finance college educations for one��s children.
In standard economic models, the benefits to one��s children show up
in one��s own utility function, but it is not clear that the benefits
to others show up in one��s own affect in the same proportion as in
utility. Third, some people genuinely care about things that contribute
to their lives but do not on average contribute to affect. For
example, it would not necessarily be a bad decision to pursue excellence
even if one knew that the effort would lower one��s expected level
of affect over a lifetime. Even striving for social rank—a dimension
in which affect and utility track each other especially well—provides
good examples of the divergence between affect and utility. Think
of how many people have knowingly and deliberately sacrificed happiness
(affect) for the sake of ambition. Some of these people would
do it all over again if they had their lives to live over.
E.
Persuasion about Preferences.
Since most people do care about happiness at least somewhat to begin
with, extolling the wonders of happiness and
exhortations to value happiness more highly
can often be effective tools for those who
desire to persuade others to change their priorities—that is, to change
their preferences and the utility function that would be needed to represent
those preferences. To the extent that people
are genuinely persuaded by such arguments, their utility function
will shift to be somewhat more tightly related to happiness.
The adverb ��genuinely�� in ��genuinely persuaded�� is needed to
subtract out the effects of social pressure in which people are brought
to outwardly assent to something they do not really
agree with. From the standpoint of Pareto optimality, there is
no reason to question the new, post-persuasion utility function if the
persuasion is, indeed genuine. However, it seems only fair that people
be made aware that it is not illogical
to put a low valuation on happiness in one��s preferences, if one so
chooses in a top-down process of concretizing one��s own preferences.30
5.
Evidence that Happiness is Not the Same Thing as
Flow Utility
Having laid out the definitions
of felt happiness and choice-based utility both conceptually and operationally
let us look more closely at their relationship. It would be convenient
for many reasons if happiness in the sense of current affect were proportional
to flow utility. Not only would this make a welfare measure based
on a discounted sum of current affect equal to lifetime utility, but
such a simple measure of flow utility would make utility empirics more
like production empirics, where one actually gets to see output directly.
However, there are two reasons why it is very difficult to maintain
that happiness and flow utility are even close to the same thing: the
Easterlin Paradox and Hedonic Adaptation.
A.
The Easterlin Paradox.
As Easterlin (1974, 1995, 2003) observes, real per capita GDP and real
consumption expenditure in the United States has risen dramatically
in the last fifty years, but the percentage of people saying they are
��very happy�� has been falling slightly. The story is even more dramatic
in Japan, where the percentage rise in per capita GDP is even more rapid,
but the graph of subjective well-being is essentially flat. In
other words, in developed societies, profound increases in average real
income and in the objective standard of living over the last 50 years
have not been associated with increases in the average happiness of
their citizens. As argued above, it is not just a matter of money
not buying happiness, since there are many other positive trends.
In short, there is strong evidence that utility has gone up, but happiness
has not.
B. Hedonic Adaptation.
In addition to the difference in the trend behavior of utility
and affect, the shorter-run dynamic behavior of affect is also quite
different from the dynamic behavior of flow utility as normally modeled.
As an empirical matter, affect in response to a number of important
categories of changes in circumstances is subject to hedonic adaptation--regression
of affect toward its previous level. Some of this evidence is
surveyed in Frederick and Loewenstein (1999). In response
to discrete negative events with lasting practical consequences, significant
hedonic adaptation over time is observed for incarceration (Zamble and
Proporino, 1990; Zamble, 1992), the loss of the use of limbs, (Wortman
and Silver, 1987) and for serious burns (Patterson, et al., 1993). The
death of a spouse seems to have a particularly long-lasting effect on
affect. But Kaprio, Koskenvuo, and Rita��s (1987) finding that
suicide rates the week after a spouse��s death are elevated almost
tenfold for women and almost seventyfold for men suggests especially
low affect immediately after the loss, which then moderates to some
extent.
Some of the
most striking data is that on lottery winners. Less than
a year after winning the lottery, Brickman, Coates and Janoff-Bulman
(1978) find that winners of large lotteries displayed only slightly
higher life satisfaction. Frederick and Loewenstein (1999) interpret
this as evidence suggestive of substantial hedonic adaptation since
it is likely that many winners of large lotteries are ecstatic immediately
after winning. More recently, Gardner and Oswald (2001) look at
people receiving a windfall--primarily lottery winners--in the British
Household Panel Survey. They find that winning £10,000 raises
affect by six times as much in the first year as £10,000 per
year in additional income. This comparison is suggestive of income having
been subject to greater hedonic adaptation than the hedonic adaptation
to the relatively recent windfall.
Brickman and Campbell (1971)
refer to the implications of hedonic adaptation for the trend in affect
the hedonic treadmill. Because of the close apparent connection
between the Easterlin Paradox and the phenomenon of hedonic adaptation,
it seems appropriate to search for a joint explanation.
C. The Implications
of Affect Data for Hedonic Adaptation and the Easterlin Paradox.
In unpublished work, Kahneman and
Schwarz seem to have discovered another important fact about hedonic
adaptation: measures of current affect such as data from experience
sampling show even stronger hedonic adaptation (mean reversion) than
life-satisfaction or global happiness measures.
(There is some discussion of this in
Kahneman, Krueger, Schkade, Schwarz
and Stone, 2004.) As discussed above in footnote 18,
because life satisfaction and global happiness evaluations incorporate
an element of autobiography and
people��s ideas about how they ��should�� feel, they will tend to
show more permanent effects of events such as unemployment, as
Lucas, Clark, Georgellis and Diener (2004) find.
What is even more serious,
the likely influence of people��s folk theories of how they
��should�� feel on life satisfaction and global happiness evaluations
may account for some of the modest relationships with income that these
measures show. Also, some of the variance in income comes from
recent enough income innovations that the dependence of happiness on
news we argue for below could account for some of the remainder of the
correlation observed between income and life satisfaction. (Recall
that life satisfaction and global happiness measures are significantly
influenced by current affect.)
Given these hints, we
predict that future research focusing on affect data as opposed to life
satisfaction and global happiness evaluations will deepen the Easterlin
Paradox and raise estimates of the extent of hedonic adaptation.
DC.
Hedonic Adaptation vs. Habit Formation.
Note that hedonic adaptation is not the same thing as habit formation.
Hedonic adaptation is a statement about happiness, as measured by psychologists.
Habit formation is a statement about utility, as measured by revealed
preference. For example, habit formation often refers to a tendency
to do something more if you have done it in the past—an effect of
past consumption on marginal utility. Of course, if
happiness were proportional to flow utility then
hedonic adaptation and habit formation would be tightly linked.
This could be empirically problematic, because data on hedonic adaptation
might then imply extremely strong habit formation. For example,
suppose utility was of the form made popular by George Constantinides
(1990), which can be represented as
where vt
is lifetime utility, �� is the discount factor for flow utility
U, current consumption is denoted C, the ��habit�� H
is a weighted average of past consumption levels, and �� is a
parameter between zero and one. Given this form of the utility
function, if happiness were proportional to flow utility, then
evidence of complete hedonic adaptation would only be consistent with
For comparison, Joseph Lupton (2002) estimates ����.75 when
estimating based on data for life-cycle portfolio choices and a value
of �� close to zero when looking at consumption choices.
The reason consumption data does not support a high value of ��
is that, unless the lags in the habit H
are quite long, a high value of �� implies there should be a
strong autocorrelation for consumption growth rates that is absent in
the data. If happiness were proportional to flow utility, matching
the speed of hedonic adaptation would require a fast-moving habit, for
which consumption data point to a value of ��
near zero. Moreover, even the higher value of ����.75
would not match the observed extent of hedonic adaptation.
A more subtle discussion is
required if—to match happiness data—someone suggests a type of habit
formation that would not show up in empirical data other
than happiness data. Suppose that everyone agreed, based on empirical
results, that current affect At
was given by To make things even better, suppose that lifetime
utility vt
could be represented as
One could then claim that affect
was equal to flow utility, where flow utility was given by
with lifetime utility vt
being given by
But a bit of algebra shows
that
Since Ct-1
is already fixed at time t, and the multiplicative factor (1-��)
does not affect preferences, this utility function represents the same
preferences over choices from time t
on as the lifetime utility function
There are
is enough degrees of
freedom in this example to force flow utility to be equal to affect.
We argue, however, that in this instructive case it is clearer and more
evocative of the existing economic literature to represent the lifetime
utility function in the equivalent, but simpler, more convenient, and
more familiar, form , where f(Ct)
is thought of as the flow utility function U(Ct).
The complexity in affect can then be represented in the relationship
between flow utility and affect. In particular, the stipulated
equationcan then be described by saying that ��affect is equal to the
first difference of flow utility.��31
While the two flow utility
functions Ut=f(Ct)
and Ut=f(Ct)-f(Ct-1)
are equivalent in the preferences they represent over choices at time
t and beyond, could the difference between them bear on the hypothetical
choice between the consumption bundle now and the consumption bundle
fifty years ago discussed in Section 4 A? One answer is to point
out that the individual and social choices we really face are those
of the next fifty years, not of the past fifty years. Looking
toward the future, we have the habits that we have from the past, and
must take those as given. From this point of view, the two utility functions
are fully equivalent.
Another answer is to carefully
pose the hypothetical choice between different comprehensive consumption
bundles in a way that takes into account all relevant habit formation.
For example, imagine that one werewas forced to
put one��s newborn child up for adoption in one of two worlds, where
one world has the comprehensive consumption bundle of fifty years ago,
while the other world has the comprehensive consumption bundle we have
now. Alternatively, assuming that per capita GDP and other objective
circumstances improve as much in the next fifty years as they have in
the last fifty years, would you rather put your newborn child up for
adoption in the world that has the comprehensive consumption bundle
we have now or the world that has the comprehensive consumption bundle
of fifty years from now? Because it is hard to imagine the future
in detail (even after conditioning on the values of some key statistics,
as here) this is a more difficult question, but an important one.32
The closely related choice
of which society one would wish to be born into is
a crucial tool in John Rawls��s (1971)
extremely influential book of political philosophy
A Theory of Justice (anticipated by Rawls 1951, 1958).
In the Economics literature, choices between societies are also a crucial
tool in John Harsanyi��s (1953, 1955)
theory of social welfare (discussed ably by
Pattanaik, 1968.) Of course, these are very difficult choices
to make. Nevertheless, revealed preference gives some guidance
here, while a simple model of 100% hedonic adaptation would guarantee
that happiness data could give no
guidance for such choices.
ED.
Local and Global Marginal Thinking vs. Focusing Illusion.
Just as the distinction between utility and happiness breaks any tight
link between hedonic adaptation and habit formation, the distinction
between utility and happiness should make one cautious in using happiness
data to assert that people are making systematic optimization mistakes.33
David A. Schkade and Daniel
Kahneman (1998) consider a thought experiment familiar to readers of
David Lodge��s (1978) comic novel, Changing
Places,
(Lodge, 1978), in which a professor from the gray English industrial
city of Birmingham has the opportunity to spend a sabbatical year at
Euphoric State on the shores of San Francisco Bay while a California
professor takes his place as a visitor at Birmingham. Schkade
and Kahneman (1998) study two groups of students, one residing in a
gray Midwestern climate and the other in the brilliant sunshine of
Southern CCalifornia. When surveyed, students
in both locations have the same distribution of subjective well being.
Both Midwestern and California students also predict that either they
themselves or a student like them would be more satisfied with specific
aspects of California including climate, outdoor activities, social
life and cultural opportunities. Schkade and Kahneman explain
their results in terms of a focusing illusion:
When a judgment about an
entire object or category is made with attention focused on a subset
of that category, a focusing illusion is likely to occur, whereby the
attended subset is overweighted relative to the unattended subset.
In particular, when attention is drawn to the possibility of change
in any significant aspect of life, the perceived effect of this change
on well-being is likely to be exaggerated. (p. 340.)
While they do not conduct such
an experiment, it appears that Schkade and Kahneman believe that a person
who actually moved to California would not experience a permanent increase
in measured happiness or satisfaction. This would seem logical
in light of the equality of overall life satisfaction they observe among
Midwestern and California students. Moreover, they argue
that the focusing effect can help explain examples in the literature
in which paraplegics and lottery winners are found to have similar levels
of life satisfaction as individuals who experienced neither of these
extreme events. Moreover, although
not tested in the Midwest-California example,
they cite other instances involving paraplegics, lottery winners and
widowed spouses in which the positive or negative effect of these events
on measured happiness is transient.
What do these results imply?
Tacitly assuming that happiness can be set equal to flow utility, Schkade
and Kahneman suggest that people mispredict utility for two reasons.
First, because of the focusing illusion they overemphasize the importance
of a particular aspect of life in California—say, climate—among
the determinants of overall satisfaction. Second, and perhaps
for the same reason, people fail to predict that their mood will adapt
to local circumstances within a relatively short period of time.
This appears to be consistent with Schkade and Kahneman��s interpretation
when they write, ��At the individual level, the focusing illusion may
lead to unnecessary initiatives. For example, it is not unlikely
that some people might actually move to California in the mistaken belief
that this will make them happier. (p. 345)
The theory we advance in this
paper would predict the same pattern of survey results about happiness,
both cross-sectionally and longitudinally, but the interpretation of
the results would be different. When the prospect of relocating,
say, from Ann Arbor to Berkeley arises, conventional economic theory
suggests that an individual needs to consider global utility maximization
by comparing the (ordinal) heights of two utility mountains, one corresponding
to attainable levels of utility in Ann Arbor and the other to attainable
levels of utility in Berkeley. The heights of these mountains depend
on location-specific nontraded goods such as climate, topography and
culture but also on variations in location-specific traded and partially
traded goods that would be available to the person given wages, prices,
employment opportunities, family and friends, leisure possibilities
and so on. In conventional economic theory, an individual would
make a migration decision by comparing the heights of the two utility
mountains. Once in Berkeley, the utility mountain in Ann Arbor
becomes irrelevant and an individual��s decisions are concentrated
on finding allocations of income and time to alternative bundles of
traded, partially traded and non-traded goods that place her as close
as possible to the summit of the local utility mountain.
We suggest that focusing is
best understood not as an illusion, but rather as a mental act that
plays a familiar role in economic theory. Conventional economic
theory suggests that a consumer chooses an allocation that maximizes
his utility subject to a budget constraint, a time constraint and other
relevant constraints such as distance from family and friends.
In finding this optimum, the consumer compares the marginal utility
gained from a good with the utility value of its marginal cost in dollars,
time, or social interaction. A mental calculation of marginal
utility—a partial derivative—requires focusing because it asks how
much utility would change holding everything else constant.
The empirical evidence of focusing described by Schkade and Kahneman
suggests that people are readily able to think about the positive or
negative impact of a particular event or state, holding other aspects
of life constant.
As is often noted both by economists
and non-economists, the optimization task assumed to take place
in standard economic theory is daunting in the complexity of its cognitive
demands on both information and calculation. To find the local
optimum associated with a given utility mountain corresponding to a
given location and a given time, an individual may need to consider
only variations in a small number of aspects of life because many others
are already settled through past decisions, trials and errors.
In most day-to-day decisions, focusing on the few dimensions at issue
yields a large savings in deliberation costs. The person already
has a job, a spouse and children, a home and, perhaps, the only significant
decision at the moment is to choose
whether to go to a Chinese or Italian restaurant tonight. By contrast,
a large decision involving changing a location, choosing a spouse or
changing jobs will cause many aspects of life to change simultaneously.
To find the optimum in such cases, the person needs some way to discover
the highest utility mountain in a vast range of (high dimensional) mountains,
each associated with a particular discrete choice. Just knowing
that the next step has higher altitude is not enough.
Even in deciding about an actual
move, a fully rational homo economicus might conduct a series
of thought experiments, similar to those on Schkade and KahnemanSK��s
questionnaire concerning satisfaction with aspects of life in the Midwest
or California, for each relevant aspect of life. If we assume
that utility is additively separable in different aspects of the locations,
then the total difference in utility from a move is
Some of these aspects will
be essentially the same in both locations, so the individual can focus
on just those that are different, together with whatever combinations
of aspects interact in a nonseparable way.
Daydreaming in a focused way
may be a very helpful way of sorting through particular aspects of a
location choice before getting on to the difficult task of making an
actual location decision—which entails a summary valuation.
After all, Hawaii, New Zealand or the South of France may have even
better climates than California. But, among these, perhaps France
and California are the best of these in culture and cuisine, on which
a given person places a higher marginal utility value. But, after
considering the value of these particular aspects of other mountains,
it may be that the advantages of the current mountain dominate because
it is close to family and friends, its properties are more certain and
staying avoids the costs of moving. In long run equilibrium, migration
takes place until the expected utility of individuals in the place they
reside is at least as high as it is in other places. Moreover,
in equilibrium, location-specific advantages such as climate
will tend to generate offsetting compensating disadvantages such as
high housing prices or low wages (Sherwin
Rosen, 1986). It would not be surprising to find that utility
is nearly equated in those locations that seem like relevant alternatives.
In this subsection, we have
argued that our theory would yield the pattern of survey results reported
by Schkade and Kahneman (1998), but that our interpretation of these
results would be quite different from their theory of focusing illusion.
This raises the question of whether there are any testable differences
between the two theories. The most obvious concerns regret.
If focusing creates an illusion that leads to the misprediction of utility,
we would expect that, on average, people who actually moved to California
would experience regret. In our theory, focusing is
just an intermediate mental step in forming a summary judgment involving
weighting a broad range of relevant issues and aspects of life.
While the summary judgment might be erroneous from an ex post point
of view, there is no reason to think that the errors are in one direction
or the other—California might turn out to be even better than one
imagined in a Midwestern college classroom. Similar testable differences
between focusing illusion and our theory could be sought from data on
regret from other sources such as new car purchases, dating behavior
and many other areas of life.
F. Is Choosing Lower
Long-Run Happiness Evidence of a Mistake or Evidence that Happiness
and Utility are Not the Same Thing?
Equating happiness with
utility is a key assumption in what has become an established theoretical
consensus among happiness researchers in Economics as well as Psychology.
This consensus challenges the validity of the foundations of conventional
Welfare Economics which lie in revealed preference theory. In this section,
we briefly describe the established consensus in the context of a specific
empirical application by Frey and Stutzer (2004b) which examines the
relationship between happiness and time spent commuting. Their
assumption that happiness and utility
are the same thing, in conjunction with the empirical relationship between
happiness and commuting, leads them to conclude that individuals systematically
mispredict utility. This conclusion, in turn, calls into question
the key assumption of revealed preference theory: namely, that the chosen
alternative yields higher utility to the consumer than those which are
not chosen. Instead, in the spirit of
��Subjective Well-Being is Desirable, But Not the Summum Bonum,��
(Diener and Scollon, 2003), we argue that Frey and Stutzer��s (2004b)
findings provide evidence that utility and happiness are empirically
distinct, but do not bear on the validity of welfare theory based on
revealed preference.
In the consensus theory,
as summarized by Frey and Stutzer (2003, 2004b),
reported subjective well-being is taken as a proxy measure for utility.
Maintaining this very strong assumption
opens up a wide range of empirical applications and allows for direct
tests of conventional theory, as Frey and Stutzer (2004b) illustrate
with their analysis of commuting time. Most people find commuting
time unpleasant, but endure it as a necessary evil in order to work
at a more interesting or better paying job while living in a nicer or
cheaper location. In equilibrium,
along the lines of Rosen (1986), they argue that individuals should
sort themselves among locations such that the disutility of additional
commuting is offset by compensating monetary or nonpecuniary benefits
associated with a better job or residential location. In such
an equilibrium, they argue, total utility should not be related to total
commuting time.
Frey and Stutzer (2004b)
test this hypothesis in a regression of happiness on commuting time
using data from the German Socioeconomic Panel (GSOEP), holding a number
of socioeconomic characteristics constant but leaving labor income free
to vary. They find a significant negative coefficient for commuting
time, contrary to the zero coefficient expected under
a Rosen-esque theory. Of course, persons with higher (non-labor)
wealth might have higher utility and choose both a better job or house
and a shorter commute, thus creating a spurious negative correlation
between happiness and commuting time. However, Frey and Stutzer
find a significant (although somewhat smaller) coefficient on commuting
time in an alternative specification in which permanent
differences in wealth or other differences are controlled with the use
of individual fixed effects. While their econometric model might
be subject to other criticisms, for our purposes we
provisionally accept their empirical finding of a negative relationship
between happiness and commuting time. We also note that other
investigators have suggested similar results for other kinds of decisions.
For example, Gruber and Mullainathan (2002) find that cigarette tax
increases raised the happiness of potential smokers; Schorr (1991) argues
that people mismanage the balance between work and leisure, tending
to overwork; and Loewenstein, Ted O��Donoghue and Matthew Rabin (2002)
suggest that misguided purchases of consumer durables such as fancy
cars occur because people overestimate the future satisfaction the purchase
will bring.
The theoretical explanation
advanced by Frey and Stutzer (2004b) and other happiness researchers
for such findings is that people systematically mispredict the future
utility or, equivalently in this view, the future happiness they will
obtain by taking a given action. In particular,
Frey and Stutzer (2004b) hypothesize that misprediction is most severe
for goods or activities with extrinsic attributes that can be purchased
in the market relative to those with intrinsic attributes involving
nonmarket social interactions. Thus, while a commuter may choose
his home and job with the expectation that the extra money he gets from
lower rent or a higher wage will offset the utility loss resulting from
spending less time with his family and friends, the negative relationship
between happiness and commuting time is interpreted to imply that people
systematically overestimate the future relative utility of the things
they obtain with the extra income.
Standard economic theory
can easily accommodate unsystematic mistakes by consumers, but has a
much more difficult time making sense of
systematic mistakes. If people get lower net utility from long
commutes, why don��t they learn this and change the
location of their home or job accordingly? Frey and Stutzer (2004b,
p. 9) explicitly address this issue. They argue that the formation
of expectations about future utility depends on reconstructions of feelings
in the past. Failure to learn from mistakes results
because ����remembered utility and predicted utility become similar
and relatively independent of actually experienced utility.��
For example, they cite studies in which participants on vacation or
holiday trips enjoyed the actual trip less than they had predicted,
but report enjoyment levels similar to the ones predicted
when they recall the experience afterward.
This is a remarkable argument.
Revealed preference suggests that the people who took a vacation gained
expected utility. Moreover, in recalling the trip they believe
they actually received as much satisfaction as they had expected to
get. Presumably, they felt no regret. Their decision appears
to involve no mistake according to standard revealed preference arguments
and certainly nothing in their recalled experience would cause them
to be less likely to take a similar trip in the future. Despite
all that, it is alleged that these people actually experienced less
utility during the trip than they had expected to receive. This
discrepancy is interpreted as a mistake and the failure to notice it
after the fact is regarded as the reason that people do not learn from
experience and correct their mistakes. Hence, misprediction of
utility is common, causing people to make wrong decisions
repeatedly which, against their own interest, result in lower levels
of experienced utility than could be achieved by alternative decisions.
The hypothesis that utility misprediction is relatively greater for
actions with extrinsic aspects suggests that materialistic people will
be most harmed by these mistakes.
In our view, to make a
convincing argument that the individuals were making a mistake, one
would need either to find evidence of regret or
indications that being presented with the purported evidence of misprediction
of utility caused people to want to change their decisions. In
the case of commuting, we do not think people would be surprised to
be told that commuting is quite unpleasant.
Learning evidence that it is difficult to buy
much happiness with money or that the effects of additional money on
happiness are transient could have more impact on people��s decisions.
In the case of trips, forgetting some of the annoyances of travel may,
in fact, distort people��s decisions; being reminded of these annoyances
might affect their decisions to some extent. However,
some of the most important benefits of a trip are precisely the memories
one brings back. To the extent those memories are positive, the
traveler has achieved one of the main objectives of a trip—with the
forgetting of annoyances serving as a
helpful aspect of the household production function for vacation memories.
The incidence of regret and second-thoughts after being presented with
relevant data is ultimately an empirical matter for which the quantitative
size of effects is just as important as the qualitative direction of
effects.
In the absence of evidence
of regret or second-thoughts upon being presented with relevant data,
the other possibility (which we highlight)
is that utility and happiness are not the same thing. Under this
alternative, the interpretation of much of the evidence cited by happiness
researchers about utility misprediction and systematic mistakes in decisionmaking
is simply misleading. While we present our argument using a formal
model in the second half of this paper, it is useful to provide some
informal intuition now for our contention that evidence from the happiness
literature is not inconsistent with a conventional economic model of
rational (albeit not omniscient) utility maximizing consumers.
It seems quite reasonable, as Diener and Scollon (2003) argue,
to assume that maximizing subjective happiness is not the only goal
of many consumers because happiness competes with other values or objectives,
some of which do not have positive effects on affect. Concretely,
much like Becker (1965) or Lancaster (1966), we think of happiness as
the outcome of one of a number of household production processes each
of which combines inputs of goods, time,
and social and physical environment to generate outputs of final commodities
according to a household technology.
For instance, in the commuting example of Frey and Stutzer (2004b),
an individual may endure unpleasant commuting in part because it affords
additional money or a more desirable residential location that enables
him to buy nice things for his wife and children, to have his children
attend a better school or to be able to contribute to a charity to relieve
the suffering of others. An empirical question, mostly not addressed
in the happiness literature, is whether each of these ways to use money
has the same effect on subjective mood or happiness. It seems
possible that they do not but, nonetheless, that the individual would
be willing to sacrifice his own happiness to benefit others. If
so, the negative correlation between commuting distance and happiness
observed by Frey and Stutzer (2004b) is quite consistent rational
with utility maximizing behavior by persons whose preferences include
goals beyond narcissistic fixation on their own pleasure.
Though an altruistic motivation makes the example especially clear,
the same logic applies if the objective
the individual is pursuing in preference to happiness is a non-altruistic
goal.
E.
The Implications of Affect Data
for Hedonic Adaptation and the Easterlin Paradox.
In unpublished work, Daniel Kahneman and Norbert Schwartz
seem to have discovered another important fact about hedonic adaptation:
measures of current affect such as data from experience sampling show
even stronger hedonic adaptation (mean reversion)
than life-satisfaction or global happiness measures.
As discussed above, because life satisfaction
and global happiness evaluations incorporate
an element of autobiography and
people��s ideas about how they ��should�� feel, they will tend to
show more permanent effects of events such as unemployment, as
Lucas, Clark, Georgellis and Diener (2004) find.
What is even more serious, the likely influence of people��s folk theories
of how they ��should�� feel on life satisfaction and global happiness
evaluations may account for some of the modest relationships with income
that these measures show. Also, some of the variance in income
comes from recent enough income innovations that the dependence of happiness
on news we argue for below could account for some of the remainder of
the correlation observed between income and life satisfaction.
(Recall that life satisfaction and global happiness measures are significantly
influenced by current affect.)
Given these hints, we predict that future research focusing on affect
data as opposed to life satisfaction and global happiness evaluations
will deepen the Easterlin Paradox and
raise estimates of the extent of hedonic adaptation.
G. Summary of the Argument that Utility and Happiness are Empirically
Distinct. Here is the underlying structure of the argument
that utility and happiness are empirically distinct. First,
using
standard utility representations, utility has a strong upward trend,
while happiness has very little trend.
Moreover, happiness is strongly mean-reverting even after permanent
changes in circumstances, while utility is not.
Second, if one is willing to use nonstandard utility representations
(including the flexibility one has in choosing flow utility functions
that add up to equivalent lifetime utility functions), one can
say the following:
(a)
On one hand, if changes or innovations in lifetime utility were the
only component of happiness, then maximizing happiness and maximizing
lifetime utility would be essentially equivalent; indeed, happiness
could even be viewed as an exotic way of representing lifetime utility
except that since happiness is
focused on changes, it still could not represent
preferences over initial levels or initial paths.
To put the issue dramatically, though
happiness is quite tightly linked to utility in this case,
because it is focused on changes, happiness provides no representation
of people��s views over which society it is best to be born into.
(b)
The frequent use of the concepts of utility and happiness to make social
welfare statements makes it ill-advised to dismiss the representation
of preferences over which society to be born into as unimportant or
meaningless. Indeed, this kind of preference is closely related
to important conceptions of social justice. These preferences over different
comprehensive social situations do not
necessarily line up with measured happiness.
(c)
Any evidence for persistent, predictable
effects of choice variables on happiness imples that
changes or innovations in lifetime utility are
not the only component of
happiness.
(d)
The fact that at times people knowingly, thoughtfully and without regret
make choices that predictably lower their
mood, day after day, implies that utility and happiness are empirically
distinct.
All of the statements (a—d) remain true regardless of what utility representation one uses for a given set of preferences. Further discussion of these arguments must wait until we have laid out our model.
Given the empirical evidence
that utility and happiness behave very differently, what is the relationship
between happiness and utility? In this section we will explain
the main elements of our answer in a model with more structure than
is really needed. Within that structure we propose that
affect =
baseline mood + elation,
where elation is short-run
happiness—which depends on recent news about lifetime utility and
baseline mood is long-run happiness—which is a subutility function
much like health, entertainment, or nutrition. There is a two-way
linkage between affect and utility in this theory. First, baseline
mood is an argument of the flow utility function. Second, elation
is a function of news about lifetime utility.
One weakness of the approach
in this section—using
with relatively well-defined functional forms—is that,
because of the illusion of cardinality for von Neumann-Morgenstern utility,
it does not sufficiently emphasize the ordinal nature of utility.
The Appendix takes a more general axiomatic approach, which allows us
to more clearly demonstrate the consistency of our theory with Ordinalism.
A.
News and Short-Run Happiness. To motivate the mathematical
model below, let us begin with the observation that—although the relationship
between circumstances and happiness is weak in the long run—all the
evidence suggests that subjective well-being responds in an intuitive
and important way to news about objective circumstances.
For example, subjective well-being rises significantly after experimental
subjects find a dime and falls significantly after experimental subjects
are given negative test results (e.g., Schwarz, 1987). The theoretical
outline we propose builds on these observations by positing that a major
component of affect depends directly on news about objective
life circumstances that has arrived over the last few months rather
than on the level of circumstances. This assumption is consistent
with the general observation that people evaluate changes rather than
states, an assumption that is also central to Prospect Theory (Kahneman
and Tversky, 1979).
We call the component of happiness
due to recent news about lifetime utility elation. Dismay
is the algebraic opposite of elation: dismay = -elation.
If expectations are rational, standard results about rational expectations
imply that news—dynamic revisions to rational expectations—will
be zero-mean and unpredictable. As a result, elation—which
is a function of recent news—will be strongly mean reverting.
Intuitively, news doesn��t stay news for very long. At
the psychological level, the initial burst of elation dissipates once
the full import of news is emotionally and cognitively processed.
B. Happiness in the Utility
Function. Since Gary Becker��s (1965) pioneering work, much
of the activity of a household outside of paid work has been reconceived
as household production of goods. Becker (1965) emphasized the concept
of household production as a way to study the structure of the household��s
utility function. For example, a household may undertake many
activities and purchases all focused on preserving health, such as buying
and consuming vitamins, exercising, and going to the doctor. It
often aids intuition to think of the health subutility function as giving
the output of a household production function for health. We think
of the part of happiness not
due to recent news about lifetime utility as this kind of subutility
function—or equivalently as the output of a household production function.
We call the part of happiness
not due to recent news about lifetime utility baseline mood.
In particular,
1. Any predictable aspect
of happiness is part of baseline mood. This includes any
persistent aspect of happiness.
2. Any aspect of happiness
that would be predictable if the relevant arguments of the subutility
function were predictable is a part of baseline mood. The pleasantness
of one��s current activity falls into this category. 34
Physical health provides a
good analogy for baseline mood. Like health, baseline
mood
Ultimately, it is an empirical matter what baseline mood depends on, but provisionally, we view baseline mood as depending on factors such as:
a. genes35
b. psychologically active drugs, such as Prozac
c. sleep
d. exercise36
e. eating habits
f. time spent with friends37
g. social rank38
h. the pleasantness of one��s current activity.39
Viewing baseline mood as one
of the arguments of flow utility allows the powerful language of price
theory to be applied to baseline mood, just as to health. For
example, Hall and Jones (2004) argue that health is a luxury good in
the sense that continuing increases in per capita income will increase
the budget share devoted to health-related expenditures. Similarly,
one might argue that continuing increases in per capita income are likely
to increase the budget share devoted to baseline-mood-
related expenditures.40
A key limitation on our ability
to apply price theory to baseline mood is the possibility that people
may not have accurate knowledge of the production function for baseline
mood. People��s expenditures of time and money will depend on their
beliefs about the production function for baseline mood rather than
the true function. Pursuing the analogy to health again, it seems
reasonable that, just as people don��t know the true production function
for health, they may not know the true production function for baseline
mood. In principle, the discovery and dissemination of facts
about the determinants of baseline mood could have large positive welfare
effects.41
One factor that could make
it especially difficult for people to figure out the determinants of
baseline mood is the salience of the component of happiness due to elation.
Although the elation mechanism has its own functions, from the standpoint
of figuring out the determinants of baseline mood, elation acts as noise.
In principle, the discovery and dissemination of facts about the determinants
of baseline mood could have large positive welfare effects.42
To the extent that people
do understand the determinants of baseline mood,
price theory can contribute in important ways to an understanding of
long-run happiness. Consider, for example, the negative correlation
that has sometimes been found between
��materialism�� and happiness. Robert Lane (2000) gives a discussion
of the mixed empirical evidence for such a negative correlation.
In assessing the evidence, it is also important to be aware of the partial
tautology in relating measures of unhappiness to
materialism indices that contain many
survey items measuring dissatisfaction and griping. Nevertheless,
in order to make the logical point as clearly as possible, suppose it
could be documented conclusively that materialism, in the narrow sense
of valuing material goods highly, lowers happiness. Price theory suggests
that as long as there is any tradeoff between happiness and material
goods, those who value material goods more compared to happiness will
choose a bundle with more material goods (as often found for those who
are more materialistic) and less happiness. The mechanics of the
tradeoff could, for example, be due to decisions such as the decision
of whether to commute further to a higher paying job discussed in Section
5F. Materialism lowering happiness would be similar to the
effect preferences have on any choice between two distinct goods—such
as when those who place an extremely high
value on career success have worse
physical health because they do not
make time to exercise or see the doctor.
Another
One important
key application of price theory is
to baseline mood probably can survive close scrutiny of
how much people really know.
the Easterlin Paradox itself.
Even after accounting for the elation mechanism,
sSince baseline mood is likely to
be a normal good, there is still a version of the Easterlin Paradox
that we must confront. With people much richer now, why don��t
they purchase more baseline mood? Trends in
Some of the externalities related to community and social
rivalry43 and any exacerbations of internal
conflicts can certainly contribute toward an explanation,
help, since most of these externalities and internal conflicts
are likely to figure into happiness at least as strongly as they figure
into utility. Lack of knowledge of the true production function
for baseline mood could also contribute in an important way toward explaining
this version of the Easterlin Paradox.
But , but there may also be
is also a price-theoretic element to the
explanation. Although income has gone up, the price of
baseline mood may have risen. The most likely reason for this is
if many of the inputs into baseline mood are time-intensive, such as
exercise or time spent with friends. With the price of baseline
mood higher, people may choose to expand their consumption of other
goods rather than baseline mood. The greater people��s willingness
to substitute between baseline mood and other goods, the smaller the
price rise necessary to explain the Easterlin Paradox.
C.
A Formal Model of Utility and Affect.
The formal model in this subsection assumes a fully rational optimizing
agent, with an internally consistent utility function, who is well informed
about the nature of his or her own preferences. Indeed, we posit
a lifetime utility function that is totally standard in how it is built
up from flow utility U. The one difference from the standard
case is that the flow utility function U
is a comprehensive function that includes baseline mood M
as an argument:
Kt
is a potentially large vector of state variables encoding every aspect
of the past that carries over to affect the present in a way that matters
for utility, such as wealth, weight, level of fatigue, one��s spouse
being alive, oneself being alive, genes, etc., concatenated with a vector
of exogenous variables (variables over which the individual has no control)
such as the weather, the state of the entire economy and the level of
consumption of the average person in society (to allow for direct social
rivalry in consumption) and other external determinants of social rank.
Xt can be a large vector of control variables
representing aspects of the current actions that can be chosen, such
as time allocations (including exercise, time with friends, and sleep),
consumption (including psychologically active drugs and the services
of psychotherapists), and portfolio choices. Baseline mood
M(Kt, Xt)
is written as a general function of Kt and
Xt—which also appear as direct arguments in
the flow utility function. Thus,
in one sense, the flow utility function is no different from a function
of the vectors Kt and Xt,
directly:
U(Kt,
Xt, M(Kt, Xt))=U(Kt,Xt)flow
utility..
Therefore,
at one level, the flow utility function is no different from a function
of the vectors Kt and Xt,
directly. However, in applications, the dependence on the
baseline mood subutility function M
can provide additional structure to the flow utility function.
Moreover, the specification of M
makes predictions about what will be observed in affect data.
The lifetime utility function
can also be written recursively as
pinned down also by the terminal
condition that lifetime utility is uniformly zero after the end of all
things the agent cares about: vT+1��0.
(This recursive form takes a step toward the Bellman equation without
yet assuming optimization.)
We define the lifetime utility
innovation ��t (��iota��) as
The lifetime utility innovation
��t is a precise way of formalizing the concept
idea of ��news about lifetime utility.�� Since the
lifetime utility innovation is proportional to
the surprise in lifetime utility at time t, rational expectations
implies that the lifetime utility innovation
��t is mean-zero and unpredictable by all
information available at time t-1
or earlier.
The recursive expression for
lifetime utility can be lagged and rearranged to yield this equation
for the lifetime utility innovation:
Thus, the lifetime utility
innovation is almost, but not quite, equal to a simple change in lifetime
utility. It differs by removing the predictable part of the movement
in lifetime utility due to the passage of time, whether from discounting
or from flow utility becoming ��water under the bridge.��
Elation, in turn, is an increasing
function of current and past lifetime utility innovations:
Finally, affect At
itself (��happiness��) is the sum of baseline mood and elation:
Notice that in this framework,
the utility function is defined first, in a way that is a straightforward
extension of a standard form. Then elation (the news component
of happiness) is modeled as dependent on lifetime utility innovations.
Baseline mood (the non-news component of happiness) is modeled in a
fairly agnostic way as a function of current variables and implicitly
of lagged variables through the state vector Kt.
To match empirical evidence about baseline mood, it is important to
include non-marketed goods such as social rank in the arguments of baseline
mood.
D.
Evidence that Expectations Matter for Affect.
One of the central predictions of this model is that expectations will
matter for affect, since the lifetime utility innovations are given
by , and elation is a function of current and past lifetime utility
innovations. The importance of expectations for affect is
indicated by the evidence surveyed in Frederick and Loewenstein (1999)
that advance notice of the death of a spouse reduces the size and duration
of the drop in affect after the actual death of the spouse. The
following passage from Frederick and Loewenstein (1999, p. 315) is especially
close to the spirit of the model here: ��Even if advance notice does
improve post-outcome well-being, its overall
effect on well-being is ambiguous, since receipt of the bad news may
diminish the well-being of the person between the time the notice is
received and the time the event actually occurs.�� In the model
here, it is the processing of bad news that generates a period of lower
affect, whether the primary bad news occurs before the actual death
of the spouse or only at the time of the actual death.
Camerer, Loewenstein and Prelec
(2005, p. 28) give a good summary of some remarkable neurobiological
research relevant to the role of expectations in determining affect:
An important feature of
many homeostatic systems is that they are highly attuned to changes
in stimuli rather than their levels. A dramatic demonstration
of such sensitivity to change came from single-neuron studies of monkey
responding to juice rewards (see Wolfram Schultz and Anthony Dickinson
2000). These studies measured the firing of dopamine neurons in
the animal��s ventral striatum, which is known to play a powerful role
in motivation and action. In their paradigm, a tone was sounded,
and two seconds later a juice reward was squirted into the monkey��s
mouth. Initially, the neurons did not fire until the juice was
delivered. Once the animal learned that the tone forecasted the
arrival of juice two seconds later, however, the same neurons fired
at the sound of the tone, but did not
fire when the juice reward arrived. These neurons were not responding
to reward, or its absence �� [ellipses and all italics in original]
they were responding to deviations from expectations. (They are
sometimes called ��prediction neurons.��) When the juice was
expected from the tone, but was not delivered, the neurons fired at
a very low rate, as if expressing disappointment.
These results are just
the tip of the iceberg in the neurobiology literature. A great
deal of evidence points to machinery in the human brain that generates
sophisticated short-run expectations—expectations that people are
not always consciously aware of.
See for example John O��Doherty et al. (2003),
Jay Gottfried, O��Doherty and Raymond Dolan (2003),
Ben Seymour et al. (2004), Seymour et al. (forthcoming) and O��Doherty
(2005).
E.
The Evolutionary Significance of Elation.
Though any such claim is highly speculative at this point, we are inclined
toward Randolph Nesse��s (2000, 2001, 2004, forthcoming) functional
interpretation of affect as part of the motivational system
for processing of utility-relevant information. If
something good happens, elation motivates the individual to think about
what went right (in case there is a way to make it happen again) and
how to take advantage of any new opportunities that may have arisen.
If something bad happens, dismay (negative elation) motivates the individual
to think about what went wrong (in case there is a way to avoid it in
the future), and how to mitigate the harm of the new situation.
On this view, elation and dismay are in the same genus as curiosity,
which is part of the motivational system for processing information
that is neither obviously good nor bad, but for which there may be value
to finding out more. Indeed, experimental inductions of elated
and depressed moods have been found to change individuals�� strategy
of information processing across a variety of tasks (for reviews see
Schwarz, 1990, 2002 and William Morris, 1999). Elated people
are especially good at seeing opportunities, while dismayed people are
especially good at seeing dangers.
F. The Evolutionary Significance
of Hedonic Adaptation. Thinking of a temporary jump in affect
occurring after utility-relevant news as functionally related to information-processing
makes the functional significance of hedonic adaptation similar to the
functional significance of adaptation in other aspects of perception.
Frederick and Loewenstein (1999, p. 303) make this comparison explicitly:
��Adaptive processes serve
two important functions. First, they protect organisms
by reducing the internal impact of external stimuli��. Second, they
enhance perception by heightening the signal value of changes from
the baseline level��.��
��Hedonic adaptation may
serve similar protective and perception-enhancing functions��. persistent
strong hedonic states (for example, fear or stress) can have destructive
physiological concomitants �� Thus, hedonic adaptation may help to
protect us from these effects.��
��Hedonic adaptation may
also increase our sensitivity to, and motivation to make, local changes
in our objective circumstances��.��
Rayo and Becker (2005) construct
a formal model that spells out the logic of Frederick and Loewenstein��s
(1999) claim.
G.
Speculations on the Evolutionary Significance of Baseline Mood.
Certain kinds of persistent situations could call for heightened sensitivity
toward opportunities or toward dangers. For example, moderately
high social rank or good physical health may make it safe to look more
for opportunities than for dangers. Thus, it could make sense for these
situations to stimulate the same machinery that is turned on by the
receipt of good news. The high variance of persistent individual
differences in baseline mood suggests a frequency dependence in which
there is an advantage to being a pessimist looking for dangers when
most of the surrounding people are optimists who might miss dangers,
while there is an advantage to being an optimist who sees opportunities
if there are plenty of pessimists around to alert one to possible dangers,
and few other optimists around to boldy
seize opportunities.
One of the most interesting
possibilities is that important aspects of the determination of baseline
mood are just quirks in the affective system that have no functional
significance. The mixed-strategy evolutionary
equilibrium in which the fitness of moderately happy and moderately
unhappy people is equal would reduce the strength of any evolutionary
pressure against such quirks.
Regardless of how the ��production
function�� for baseline mood arose, now that it is presentthere,
it makes sense to exploit it, just as Stephen Pinker (1997) argues that
we exploit our sense of taste (designed, say, to motivate the search
for honey and
nuts and ripe fruits) with cheesecake and our musical sense
(designed, say, to help us distinguish the sounds of different kinds
of objects) with symphonies and Rock and Roll.
H.
Implications of the Integrated Framework for Utility and Happiness.
There are three key implications of this benchmark model for the relationship
between affect and utility. First, there is a clear distinction
between the psychological concept of affect and the economic concept
of flow utility. Affect is not
equal to either flow utility or to the overall objective function.
Second, the elation component
of affect depends primarily on unexpected changes
in lifetime utility. For applications, the most important point
about elation is that the theory here contradicts the notion that a
temporary movement in affect is unimportant because of its short duration.
To the contrary, a temporary movement in affect may be
extremely very important as a
signal of important utility-relevant news related to the long-term welfare
of the individual.
Third, baseline mood, while
not a summary measure of flow utility, is something that people care
about. As with health, the relative concern with raising baseline
mood compared to raising consumption of other goods may increase along
with per capita income, implying that the average share of effort and
expenditures devoted to raising baseline mood may increase in the future.
Since elation depends on (mean-zero)
news about lifetime utility, rather than on the level
of lifetime utility, elation has no trend. Thus,
utility can rise with per capita income while
elation has no trend and
happiness has only the trend imparted by the growth rate of baseline
mood. This guarantees that the economic concept of lifetime utility
and the psychological concept of the temporal sum of affect over time
put forward by Kahneman (1999) will be numerically distinct approaches
to assessing overall welfare. Distinguishing clearly between utility
and happiness allows scientific questions about utility and happiness
to proceed in a way that respects the insights of both Psychology and
Economics without prejudging the deep
ethical question of the proper contribution of each concept to the assessment
of overall welfare--–an ethical question that
revolves fundamentally around the extent to which one should trust people��s
immediate feelings and the extent to which one should trust people��s
choices as an indications of what most enhances
their welfare. In this ethical debate, traditional Welfare Economics
has implicitly staked out a position in favor of utility as the better
measure of overall welfare, but the case for Kahneman��s (1999) proposal
deserves to be thoughtfully considered as well.44
Maintaining a clear distinction
between affect and flow utility also makes it possible to see where
the psychological approach toward welfare assessment and the economic
approach toward welfare assessment are pulling in the same direction.
For example, social rank—,
whether appearing as an effect of other people��s consumption or time
use on baseline mood or on flow utility directly—
will matter for both the psychological and economic measures of overall
welfare. As another example, as long as baseline mood is an argument
of the flow utility function, any advance in scientific understanding
of determinants of baseline mood, and the dissemination of
scientific knowledge about baseline mood to individuals in society will
be important for both measures of overall welfare.
7. Elation Theory and the Confusion Between Utility and Happiness
Any adequate account of the
relationship between utility and happiness must explain why these two
concepts are often confused. Why is it that they often seem to
mean the same thing? To answer this question, it is useful to
compare maximizing lifetime utility with Kahneman��s (1999) proposal
of maximizing the true mathematical expectation of the present discounted
value of happiness45 in the context of the theory presented
above.
A.
A. Maximizing
the Present Discounted Value of Happiness versus Maximizing
Lifetime Utility. To the extent that baseline mood is different
from flow utility and to some extent controllable, maximizing the expected
present discounted value of happiness as Kahneman (1999) recommends
will be different on that account from maximizing lifetime utility.
But what about maximizing the expected present discounted value of happiness
when baseline mood is beyond the individual��s control? In that
case only elation will matter in maximizing the presented discounted
value of happiness. Proposition 1 addresses this case:
Proposition 1:
Given (i) rational expectations, (ii) perfect memory, (iii) happiness
that is the sum of baseline mood and elation, (iv) baseline mood that
is exogenous to the individual, and (v) elation that is a positive linear
combination of lifetime utility innovations, as of time t, maximizing
the expected present discounted value of affect is equivalent to maximizing
lifetime utility.
Proof:
Let elation et be given by
Tthen
the expected present discounted value of happiness is
where
as long as time t
is at least n periods away from death, and somewhat less if
t is less than n periods from death. Using the definition
of lifetime utility innovations, perfect memory and the fact that the
expectation of lifetime utility innovations conditional on previous
information is zero, one can simplify the expected present discounted
value of happiness further, to
Given the exogeneity of baseline
mood M and the perspective of time t, everything in this
expression is fixed except for b0,t vt.
Thus, maximizing the expected present discounted value of happiness
is equivalent to maximizing b0,t vt,
which in turn is equivalent to maximizing vt.46
B.
Maximizing Current Happiness.
Note that under the assumptions of Proposition 1, maximizing current
happiness alone is also equivalent to maximizing lifetime utility, since
Given the assumed exogeneity
of baseline mood Mt, the only thing that is
not fixed in this expression as of time t
is the term a0vt, so one
does the same thing to maximize current happiness as to maximize lifetime
utility. The reason a present discounted value of happiness is
not required is that elation is already forward-looking.47
C. Why Utility and Happiness
are Often Confused. Psychological evidence is accumulating
that baseline mood can in fact, be modified deliberately—and in ways
that go beyond the zero-sum game of acquiring social rank. But
a lack of understanding of the determinants of baseline mood can make
baseline mood seem exogenous. As noted above, one reason
for this lack of understanding may be that a large fraction of the time-series
variance of happiness may be accounted for by elation and dismay.
To the extent that elation and dismay dominate people��s perception
of happiness, Proposition 1 indicates why people might think that utility
and happiness are essentially the same thing.
It is when people do begin
to recognize that baseline mood might be controllable that the distinction
between utility and happiness becomes crucial. Understanding the
ways in which baseline mood is controllable clearly matters for optimization.
Understanding the distinction between utility and happiness is becoming
important precisely because we are beginning to see a wider variety
of ways to raise utility by raising happiness rather than
being limited to raising happiness (temporarily) by
raising utility.
8. Utility
of Elation.
To the extent that people value
transient happiness as well as lasting happiness, elation may enter
the utility function as well as baseline mood. Because
elation depends in turn on news about lifetime utility, putting elation
in the utility function requires one to solve simultaneously for elation
and lifetime utility. For that reason, we have delayed the discussion
of elation in the utility function until this point.
A.
A. Adding Elation when Elation is a Linear Function
of Lifetime Utility Innovations.
One key result, showing the robustness of the model of Section 6 to
the addition of elation, is the following:
Proposition 2:
Given rational expectations, adding to the flow utility function
a linear function of lifetime utility innovations (with positive coefficients
summing to less than one) to the flow utility function
has no effect on the preferences represented by the utility function.
Proof:
Using an asterisk to represent the modified flow utility and lifetime
utility functions, let
where . Note that the
relevant lifetime utility innovations will be those for the modified
lifetime utility function. Modified lifetime utility is then
where,
as above,
as as
long as time t is at least n periods away from death, and somewhat less
if t is less than n periods from death. The essential structure
here is that modified lifetime utility v*t is equal to the
original lifetime utility vt plus the expected value of a
linear combination of the modified lifetime utility innovations with
positive coefficients running from n periods back, up to the lifetime
utility innovation in the agent��s last period. Because
lifetime utility innovations have mean zero conditional on previous
information, one can simplify this further to
The condition that guarantees
that ��b0,t<1. Therefore, one can
solve for v*t:
Because is fixed as of time
t, as a representation of preferences over choices at time t,
is equivalent to vt/(1-��b0,t),
which in turn is equivalent to vt itself.
To recap the proof, when a
linear combination of lifetime utility innovations is added to the lifetime
utility function, (1) the future lifetime utility innovations do not
affect decisions because their expectation is zero,
; (2) the past lifetime utility innovations do not affect
decisions because they are predetermined and (3) the current lifetime
utility innovation does not affect decisions because, to the extent
it is not predetermined, it is perfectly correlated with the original
lifetime utility function.
B.
Manipulating the Timing of News and Manipulating Expectations.
In the proof of Proposition 2, have we tacitly assumed a fixed information
structure? Does adding elation that is a positive linear combination
of lifetime utility innovations to the utility function affect preferences
over information structures, even when both information structures would
lead to the same decisions over other variables? For example,
could it make people want to delay when they hear news in order to manipulate
their own feelings? The answer is no. Because rational
expectations take into account the information structure, there is no
way to game the system with any rule set up in advance. Suppose
for example, that you told your friend to tell you good news right away,
but to withhold bad news. The Bayesian inference in rational expectations
would cancel out any effect on the expected lifetime utility innovation,
though it would certainly affect the ex post
distribution of lifetime utility innovations. Formally, v*t
can be expressed as a linear function of vt
and past expectations about lifetime utility. Choosing
a different information structure now
can only affect current and future expectations about lifetime utility.
That includes choosing an information structure when your friend says
��I know what happened, do you want me to tell you or not?�� since
any revelation is still in the future, if in the near future.
Of course, even a mean-zero
effect on the distribution of lifetime utility innovations will affect
lifetime utility when added elation is a nonlinear function of lifetime
utility innovations or elation enters the utility function nonlinearly,
as we discuss below. Also, imperfect memory of past expectations
may provide an opening for gaming the system by trying to reduce one��s
remembered past expectations.
This may be particularly relevant for the memories
of past expectations parents transmit to a child about the child��s
prospects: the gap between parent and child can be one source
of imperfect memory in a dynasty. More generally, an attitude
of gratitude (whose value is not diminished by the triteness of the
phrase) can serve the same purpose as manipulable memory. It often
involves substituting comparisons with others in a worse situation for
comparisons with one��s own remembered past expectations or one��s
own deductions of what one ought to have expected in the past.
Given perfect memory, but irrational
expectations, it may be harder to beneficially manipulate expectations
than one might at first think, since then lowering one��s expectations
adds to flow utility in the future, but subtracts from flow utility
now.48
(As long as one is more than n
periods away from death, this will have no effect on lifetime utility.
Closer to death, the direct effect of
pushing one��s expectations down is harmful and pushing one��s expectations
up is beneficial.) It is when one can manage high expectations
now, but remember them in the future as if they were low expectations
that there is a real opening for beneficial manipulation of beliefs.
C. Do Mistakes about the
Rate of Hedonic Adaptation Matter?
We argued above that because utility and happiness are distinct, the
psychological phenomenon of hedonic adaptation does not have any necessary
implications for the shape of the utility function. In particular,
if flow utility depends only on baseline mood and not on elation, as
in the model above, the determination of elation, including the rate
of hedonic adaptation, has no effect on the lifetime utility function.
Thus, when elation is not an argument of the utility function, misprediction
of hedonic adaptation causes no material harm to utility maximization,
contrary to the claims of Schkade and Kahneman (1998).
To pursue the question further,
consider how much harm there is to mistakes about the rate of hedonic
adaptation in the context of the model of subsection A, with a positive
linear combination of lifetime utility innovations added to the flow
utility function. Mistakes about the rate of hedonic adaptation are
mistakes about the true values of the coefficients . Since the
modified lifetime utility function is equivalent to the original utility
function regardless of the values of the coefficients (as long as they
are positive and add to less than one), mistakes about the rate of hedonic
adaptation will not distort decisions at all and so will be costless!
In the light of the lack of harm to optimization from misperception of the rate of hedonic adaptation in this benchmark case, any serious claim of quantitatively significant harm to optimization from misperception of the rate of hedonic adaptation would require careful modeling. For example, when elation is a nonlinear function of lifetime utility innovations, or flow utility is a nonlinear function of elation, there is likely to be at least some harm from misperception of the rate of hedonic adaptation, but it is not clear how large this harm would be.
In the case of imperfect memory,
misperception of the rate of hedonic adaptation might cause one to exert
too much or too little effort toward manipulating one��s memories,
but whether this results in a serious reduction in lifetime utility
depends on how great the scope is for manipulation of memory.
One of the most important effects
of underestimating the rate of hedonic adaptation is that it will cause
an overestimation of the unconditional variance of elation, since the
effects of unforeseen increases or decreases in lifetime utility seem
like they will be long-lasting. An overestimation of the unconditional
variance of elation should, in turn, cause an individual to overestimate
the fraction of the variance of happiness due to elation and underestimate
the fraction of the variance of happiness due to baseline mood.
As shown above, this overestimation of the persistence of elation does
not necessarily interfere with maximizing lifetime utility, but it
would tend to push Kahneman��s suggested alternative of maximizing
the expected present discounted value of happiness in the direction
of maximizing lifetime utility. Since elation embodies movements
in lifetime utility, anything that exaggerates the importance of elation
in happiness is likely to make maximizing happiness more like maximizing
utility, as indicated by the extreme case of Section 7, where elation
is the only controllable component of happiness.
D.
Elation Nonseparable in the Utility Function.
Given Proposition 2, the key issues arising from elation in the utility
function are (1) imperfect memory and departures from rational expectations,
discussed briefly above, (2) nonseparability of elation in the utility
function, and (3) nonlinearity of the utility function in lifetime
utility innovations once elation is substituted out.49
Here we will barely mention the possible consequences of nonseparability.
Nonseparability of lifetime utility innovations could make manipulating
the timing of news optimal. For example, an altruistic person
might want to throw a surprise party to take advantage of a complementarity
in the recipient��s utility function between elation and the presence
of friends. Nonseparability of elation in the utility function
can also generate wealth effects that modify the size of the effects
of imperfect memory, nonrational expectations and nonlinearity.
E. Elation Nonlinear in
Lifetime Utility Innovations as a Foundation for Prospect Theory.
It is worth discussing in some depth the effects of nonlinearity of
the flow utility function in lifetime utility innovations, because this
can lead very naturally to a version of Prospect Theory.
Rather than discuss the effects of nonlinearity in general, we focus
on a particular type of nonlinearity, motivated by the evolutionary
interpretation of elation in Section 6-E. Consider
the following set of features one might wish a model to exhibit:
a. Elation is proportional to the rate of cognitive processing of news.
b. Within bad or good news, the total amount of processing needed is proportional to the magnitude of the news.
c. Bad news requires more processing per unit of lifetime utility innovation than good news. (This implies a partial separation of the cerebral system for working through good news from the cerebral system for working through bad news.)
d. It takes longer to process
a big chunk of news than a small chunk of news.
Let us present a toy model
that illustrates how these features could arise. It is in continuous
time to provide detail of how news is being processed and for convenience.
News arrives once, at time 0. G is a positive state variable
representing the cumulative amount of unprocessed good news.
B is a negative state variable for the cumulative amount of unprocessed
bad news magnified by the parameter ��>1 to represent the additional
difficulty of processing bad news.50 Formally, at the instant of news,
either G or B jumps according to
The capacity for cognitive
processing is ��, so that in the absence of additional news, and as
long as G>0,
Similarly, in the absence of
additional news, and as long as B<0,
Elation is �� for some
time after good news, while elation is -–��
for some time after bad news.51 The time it takes to get
over a chunk of news is proportional to the size of the news.
It takes �� times as long to get over bad news.
The von Neumann-Morgenstern
lifetime utility function at the moment the news is received
at time zero is given by
with ��<��-1.
If the lifetime utility innovation ��0 is greater
than zero, the present discounted value of elation g(��0)
is given by
If ��0<0,
Graphically, g(��0)
is concave for positive lifetime utility innovations and convex for
negative lifetime utility innovations, with a kink at zero.
(The slope is 1 for small positive values, but ��>1 for small
negative values.) g(��0) has an
asymptote at ��/�� at +�� and an asymptote at -–��/��
at -��. This is a shape familiar from Prospect Theory (Kahneman
and Tversky, 1979).52
Let v-
be the level of lifetime utility immediately before the news at time
zero, while v+
is the lifetime utility immediately after the news. The news tells the
constant level of consumption C
that will prevail from time zero on. Therefore,
which can be solved uniquely
for v+, since the maximum slope of ��g
is �Ȧ�<1:
Also,
which has a unique solution
for v- since h
is decreasing in v-. To get a little
more intuition, think of what happens for small ��. If
�� is small, then
Thus, the lifetime utility
function takes some of its curvature from f(C) and some
of its curvature from g. If f(C) has the
functional form of decreasing absolute risk aversion, then as the agent
becomes richer, more and more of the curvature of the lifetime utility
function will come from g. Also, since g
is kinked at zero, the function g—which comes from the agent��s
awareness of the affective consequences of good and bad news—will
dominate the agent��s choices between small risks. For choices
among large enough risks, the fact that g
has a limited range, bounded between two asymptotes, could make the
curvature of f the dominant factor.
One aspect of these equations
that may seem esoteric is the self-referential aspect of lifetime utility
depending on its own innovation. In order to see more clearly
how that self-dependence is resolved, it is helpful to look at a different
approximation: the approximation for small lifetime utility innovations.
For small positive lifetime utility innovations,
so that
and
For small negative lifetime
utility innovations,
so that
and
There are two interesting results
apparent from this approximation. First, the realization of consumption
at which there is no surprise is where , just as it would be if elation
were not in the utility function. Second, the kink at this level
is made sharper by the way �� appears in the factor (1-�צ�)-1.
This factor looks formally like a Keynesian multiplier—reflecting
the self-referential aspect of lifetime utility depending on its own
innovation.
The model above is only a toy
model, but we think it accurately indicates the likely flavor of a more
general treatment of nonlinear von Neumann-Morgenstern preferences over
elation: given the structure of elation, Prospect Theory can easily
arise from rational preferences over one��s own emotions.53
Such an affective foundation for Prospect Theory puts Prospect Theory
in context. For example, this kind of model predicts that
Prospect-Theory-like behavior will arise where the affective consequences
of a choice are much larger than the non-affective consequences of a
choice. Also, to the extent that Prospect Theory arises from the
affective consequences of choices, affect data will be helpful in understanding
people��s choices, even though it will not be the whole story.
9.
Implications of the Theory for Happiness Empirics
Even without the extensions
discussed in Section 8, the integrated framework for utility
and happiness laid out in Section 6 has many important implications
for empirical work using happiness data. One of the most basic
tests of the value of our framework is whether pursuing these implications
for empirical work turns out to be fruitful.
A.
A. The Time-Series Properties of Happiness Matter.
The most obvious implication of our framework is the need for more research
on the time-series properties of happiness. For example,
We have work in progress along
both of these lines, but the details must to left to other papers.
B. Price Theory Can be Used
to Study Baseline Mood. Second, the theory of baseline mood
implies that standard price theoretic tools can be applied to the low-frequency
movements of happiness. For example, the dollar value people place
on feeling permanently happier can be gauged by how much they are willing
to pay for psychotherapy in time and money (beyond what insurance
pays for), divided by the effect of the psychotherapy on happiness.
The less effective psychotherapy is at actually raising happiness, the
higher the implied valuation on happiness. In the case of antidepressants,
in addition to the monetary and time cost, one would have to determine
how much people would be willing to pay to have an antidepressant free
of side effects and add that value before dividing by the effect on
happiness. Such ratios can begin to identify the marginal value
of happiness.
Several other price-theoretic
issues have been discussed above. Even in the context of our framework,
normality of baseline mood still leads to a version of the Easterlin
Paradox. It is important to construct measures of the price of
happiness over time to see if an upward trend in that price can explain
why people are not choosing higher baseline mood in their ever-expanding
consumption bundle.
As mentioned above, one limitation
inon the use of these price theoretic tools is that
they depend on knowing people��s beliefs about the household production
function for baseline mood. Would people do more things
that add to happiness if they just knew what they were?
For example, there are some hints that, in addition to its other
benefits, getting more
additional sleepp might add significantly
to happiness.54 If this is true, and people
knew it, this could place a strong upper bound on the value people place
on happiness (the hourly wage divided by the effect of an extra hour��s
sleep on happiness), but such a conclusion would only be warranted if
people really knew exactly how much benefit an extra hour��s sleep
would have for happiness. On might obtain a more reasonable estimate
of the value of happiness by conditioning on people��s reports of how
much they believe an extra hour of sleep each night would add
to happiness.
Given measures of the marginal
value of happiness, any evidence about the determinants of happiness
should be included in cost-benefit analyses. If the marginal dollar
value of happiness is high, it could motivate ever more careful empirical
work to measure the strength of the effects of variables on long-run
happiness. In particular, it could motivate many clever minds
to look for good instruments for the possible determinants of long-run
happiness.55
In principle, the application
of price- theoretic tools to baseline mood should
yield tests of the theory as well as applications. This kind of
test of the theory is likely to emerge over time as the measures of
the relevant concepts are refined.
C. The Elation Theory is
Readily Testable. Because the theory of elation is the most
highly structured aspect of our theory, it is also the most readily
testable. In particular, we hope to test whether or not people��s
hypothetical choices between alternatives A and B always match their
predictions of how happy they would be immediately after receiving the
news that A had happened or that B had happened with no action on their
part. Our theory predicts that people will choose the alternative
that would seem like the best news to them (as indicated by their happiness
immediately after hearing). To the extent that this does not seem
like a very daring prediction, it indicates that the reader has a prior
belief in favor of one of the key linchpins of our theory. Nevertheless,
it is a testable prediction. It is not
true by definition.
D. Elation Provides Information
about Preferences and Expectations.
TThird, the theory of elation implies that, if
it is possible to control for variation in baseline mood, the response
of happiness to news will give direct information about preferences.
Indeed, the elation component of happiness is essentially an ��excess
returns�� measure for lifetime utility. Therefore, in principle,
happiness data can serve as the basis for exactly the same kinds of
��event studies�� as those carried out using data on excess financial
returns.
To implement this insight about
the use of happiness data for ��event studies,�� it would be very
helpful to have a regular monthly, or even daily, time series on average
aggregate affect.56
This would allow a test of average preferences over aggregate
events. In particular, after accounting for the lagged effects
of the previous month��s�� news, the theory
of elation implies that whether affect goes above baseline or below
baseline indicates whether the month��s news has been on average good
news overall or bad news overall. Since many things happen in
a month, each month��s data would give information about whether a
different innovation vector for the expected consumption bundle represented
an increase or a decrease in lifetime utility. Over time,
this would tell a great deal about average preferences for aggregate
events. Information about preferences for aggregate events is
particularly valuable because many of these things do not have regular
markets. For example, one might want to know about the relative
importance people put on geopolitics compared to economics. Any
month in which there is good economic news but bad geopolitical news,
or the reverse, would provide relevant information. Election returns
are often read as giving this kind of information, but affect data immediately
on the heels of news may give more detail. At a minimum,
high-frequency happiness data serves
as a kind of general-purpose poll question that can give
useful insight into how people feel about any big event that comes along.
Kimball, Helen Levy, Fumio
Ohtake and Yoshiro Tsutsui (2006) report a pilot study using a few months��
worth of happiness data on the University of Michigan Surveys of Consumers.
They find a significant dip in measured happiness both
in the first week of September 2005, right after
Hurricane Katrina, and in the week
after the earthquake in Pakistan that occurred in October 2005.
Adaptation to the hedonic effects of these national and international
news events was close to complete after two weeks. The dip in
happiness after Hurricane Katrina was significantly greater in the South
Central region of the United States, closest to the hurricane��s landfall.
The size of the average U.S. reaction to the earthquake in Pakistan
is almost as great as the size of the
average U.S. reaction to Katrina.
Although the human toll from the earthquake in Pakistan was much greater
than from Katrina, this still indicates a surprisingly high degree of
concern for people on the other side of the world if we are correct
in our hypothesis that the size of short-run spikes in happiness indicates
the magnitude of the implications of news for lifetime utility.
Tsutsui, Kimball and Ohtake
(2005) apply a similar event-study methodology to the hedonic reactions
in a Japanese sample to the overwhelming electoral victory of Japanese
Prime Minister Koizumi in October 2005.
In situations where preferences
are clear, the theory of elation draws a strong link between happiness
and expectations. This provides another avenue for testing the
theory. Section 6
D discusses some of the supporting evidence that has already been established
on this score. More can be done in this area. One of the
most interesting tests would be in areas where people are known to violate
rational expectations or where the memory of past expectations is likely
to fade. Here the test would be to see if the pattern of people��s
reported happiness matched the quirks in their expectation formation
and memory.
Assuming that the elation theory
is valid, it may have relevance for the survey measurement of preferences
and expectations quite broadly. On the preference side, since
the elation mechanism seems to be fairly automatic as a psychological
process, it may be that it is easier and more reliable for respondents
to predict their happiness after option A and after option B than it
is for them to make a direct choice. On the expectations side,
given the unfamiliarity of precise probabilities compared to the familiarity
of happiness, it may be easier and more reliable for respondents to
report happiness than for them to directly report probabilities.
For example, after setting the stage by asking how happy a sample of
people on one side of the political divide would feel (A) if their preferred
presidential candidate won or (B) if their less preferred presidential
candidate won, their average happiness in the days before the election
might be an efficient way to assess their subjective probabilities of
victory for their preferred candidate.
We have an example from personal
experience of using elation to gather information about the strength
of preferences. One of us was present when a daughter opened letters
from the admissions departments of the colleges she had applied to.
The evident strength of the daughter��s positive affective reaction
was persuasive in establishing the extra value she placed on going to
her much more expensive first-choice college, as opposed to her much
less expensive second-choice college. Of course, this did not
indicate what the ultimate wisdom of each choice would be, but it did
indicate her preferences given her beliefs about what it would be like
to go to each college.
E.
Sufficient Statistics. As we have argued at length, elation
and dismay measure the effect of news on expected lifetime utility.
Somewhat more formally, elation provides a
sufficient statistic that captures the effect on
the (expected) lifetime utility of current or future events that had
not previously been anticipated. This interpretation of elation
is similar to the more familiar idea that in analyzing lifecycle maximization
problems the marginal utility of income, a scalar quantity denoted by,
��serves as the sufficient statistic which captures all information
from other periods that is necessary to solve the current-period maximization
problem.�� (Blundell and MaCurdy, 1999, p.1594). In particular,
measures the utility value of a dollar saved for expenditure in the
future relative to a dollar spent on consumption today and also the
marginal utility value of additional
time spent on leisure or household production per unit of foregone earnings.
Unanticipated changes in the marginal utility of income provide a signal
about changes in the optimal allocation of consumption and leisure over
the lifecycle.
It is interesting to consider
the roles of these two quantities for consumer behavior. The role
of the marginal utility of income is well known. Any news about
changes in expected future income or price which causes the marginal
utility of income to fall serve as a signal to a utility-maximizing
consumer to increase current consumption and leisure by reducing saving;
news that leads to an increase in the marginal utility of income is
a signal to reduce consumption and leisure and increase saving.
Both elation and are derived from the
(expected) lifetime utility function. It follows that news that
affects probability beliefs about future incomes, prices, health or
any other variables that affect preferences or constraints that causes
changes inwill also tend to cause elation or dismay.
Mathematically, there
are two differences between elation
in our theory and as sufficient statistics. First, elation is
about the total lifetime utility,
while is about the derivative
of lifetime utility with respect to wealth. Second, it is
temporary spikes in the level of
elation that indicate a change lifetime utility, while it is
permanent changes in the level of
that indicate a change in the first derivative of lifetime utility.57
In principle, economists
could gain insight into the effects of news by studying the dynamics
of consumption and labor supply jointly with longitudinal data on subjective
well-being. Observed changes in savings, consumption or labor
supply respond to news that influences the marginal value of a dollar
while spikes in subjective well-being reflect changes in total lifetime
utility. It is easy to show that elation and changes in the marginal
utility of income are not necessarily correlated in a simple way.
For example, a person who receives news that he has been promoted and
will be receiving a higher salary next year will feel elation—a signal
that his lifetime utility has gone up—and his marginal utility of
income will fall—a signal that he should allocate more of his wealth
to current consumption. Today, he might choose to celebrate his
promotion with dinner at a fancy restaurant. Consider a less fortunate
person who has just heard a jury convict him of a long prison sentence.
Obviously, this person suffers dismay. However, given the difference
in the availability of fine food in prison and in town, his marginal
utility of income also falls and he may also choose a
fancy restaurant meal today, assuming that his appeal allows him to
stay out of prison for a while. As another example, a person (with
full medical insurance) who has had a successful operation on a malignant
tumor will experience elation and an increase
in his marginal utility of income, signaling that he should increase
his saving to accommodate his longer life expectancy.
Although elation and changes
in the marginal utility of income could
be correlated in either direction, it is likely that most news in the
economic domain reflects good or bad news about future income and wealth.
That is, macroeconomic news about expansions or recessions or microeconomic
news about one��s own promotion or layoff tend to produce both unexpected
gains or losses in lifetime utility and signals to increase or decrease
current consumption. This might be consistent with psychological
evidence that has been interpreted as
suggesting that an up mood is a signal to move forward with bold plans,
putting aside worries, while a down mood is a signal to focus on things
that might go wrong and to proceed with caution. It would
be interesting to examine whether the correlation between mood and consumption
is usually positive and also to test whether this correlation is reversed
in the less usual situations such as those described above when mood
and the marginal utility of income move together rather than in opposite
directions.
A good way to discuss the policy
implications of our framework is to contrast the views that we believe
follow from our framework with those of Layard (2005). Layard
(2005) is very bold in making policy recommendations based on happiness
theory and empirics. Although he is especially bold, we consider
the general tenor of his recommendations to be reasonably representative
of views expressed in much of the existing happiness literature.
Layard explicitly accepts Kahneman��s
(1999) proposal to use the expected present discounted value of measured
happiness averaged across people as the social welfare function.
Besides the issues we discuss in this paper, Layard is assuming a solution
to interpersonal comparability issues that we think have not been solved,
but let us leave that aside, since all of our social welfare measures
share that difficulty in all but the easiest applications.58
Many of Layard��s recommendations
depend only on happiness being more valuable than current public policy
recognizes. The general discounting of intangibles in policy
discussions makes this likely. Generating and popularizing
happiness accounts in parallel to GDP accounts is a reasonable step
to rectify insufficient attention to these intangibles. Taking
happiness more seriously also suggests that
many other concrete steps, such as fighting the stigma to antidepressants
and psychological treatment, and devoting more resources to mental health
care, mental health research, happiness research, and public education
about the determinants of long-run happiness.
Other recommendations depend
on the externalities inherent in people caring about social rank.
Since both revealed preference and happiness data indicate that social
rank is important, these recommendations remain on the mark. Quantitatively,
a revealed preference measure of the importance of social rank may be
different from an affective measure, but qualitatively, the implications
of social rank mattering are the same either way.
Affective data also provide
a good reminder of the importance of many other externalities and public
goods—an importance that can be verified by revealed preference.
The sense of community matters, the strength and quality of marriages
and families matter:, and the responsiveness of
government matters. Again, a revealed preference measure may differ
quantitatively in the importance it suggests for these externalities,
but it is likely to agree qualitatively.
There are two areas where we
differ with Layard. First, Layard makes many recommendations based
on Kahneman��s (1999) social welfare measure, where we would turn to
utility based on revealed preference as the appropriate welfare measure.
This leads to stark differences in recommendations about tradeoffs between
economic growth and other values. For example, Layard argues that
since economic growth does very little to raise happiness, while being
forced to move from one city to another lowers happiness significantly,
it is worth sacrificing a great deal of economic growth in order to
slightly reduce the need for mobility. To us, this either tacitly
assumes that feeling happy is the only thing people care about (which
we dispute), or it requires forcing upon them the objective of maximizing
happiness when, given the choice, they reject this objective for themselves.
There are many indications that economic growth is, in fact, important
to people, even if it does not raise happiness. In principle,
the dollar value of happiness could be high enough to make such a sacrifice
worthwhile even if happiness is not the only thing in the utility function,
but it would require take
an extremely high value. Even given existing lay knowledge
about the determinants of baseline moodbeliefs, if
the value people place on happiness were high enough to make this kind
of sacrifice worthwhile, we should see many more people seeking psychological
treatment, sleeping more, exercising more, eating better, pursuing meditative
practices, and so on, than we observe. Of course, if stability
of residence enters the utility function beyond its effect on happiness,
its valuation could be higher, but this
is not Layard��s argument.59
It is worth being very
explicit about why Kahneman��s (1999) social welfare measure differs
so much in its implications from standard social welfare measures in
Economics. Our theory implies that the present discounted
value of affect Kahneman points to is the sum of two very different
components: the present discounted value of baseline mood and the present
discounted value of elation.
Conceptually, we view the present discounted value of baseline mood
as something like the present discounted value of any other sub-utility
function, such as the present discounted value of a health measure.
While likely to be correlated with lifetime utility,
this present discounted value of baseline mood
represents only one of the things people care about.
(Less importantly, this present discounted value also has
in it no representation of any decreasing
returns to baseline mood in the utility function.) By contrast,
the present discounted value of elation is a very interesting quantity
that (at least approximately) represents the
cumulative innovation in lifetime utility over the interval of time
covered in the discounted sum. In other words, the present discounted
value of elation answers the question of how well one��s life has turned
out compared to what one expected at an earlier time, perhaps long ago,
when the discounted sum began. If for example, one could separate
out elation from baseline mood in measured affect and constructed present
discounted sums of elation from a long panel of the adult population
over 18, the average discounted sum of elation would represent how much
better or worse people��s lives turned out than they expected at age
18. As a social welfare measure, this intriguing quantity
has one serious problem: it does not credit as social improvement any
improvement in how people��s lives in a society look as of age 18.
Even if the panel were extended back to
five-year olds, the present discounted sum of elation from that age
on would not give due weight to improvement in life prospects as life
prospects appear as of age 5.
In our view, many of the most valuable aspects of progress
over the past few centuries, or even the past few decades,
are ones that would be highly valued by
five-year olds, not just in the moment, but as they think about what
their lives will be like when they grow up.
Second, we consider Layard too quick to believe that people are making
systematic mistakes in optimization.60 People no doubt do make mistakes,
but because happiness is not the only thing people care about, happiness
data alone is seldom enough to identify optimization mistakes.
The key types of evidence we would point to for identifying mistakes
are regret and people changing their minds on a decision after thinking
more carefully or getting better information. Also, not
all factual mistakes lead to optimization mistakes. In Section 8 C,
we argue at length that mistakes in predicting the dynamics of affect
do not necessarily lead to optimization mistakes.
11.
Conclusion
Happiness research matters
because—In this Conclusion, we begin with
some speculative views about the importance of research on happiness
and end with some thoughts about our approach as a research strategy.
A. Why Happiness Research Matters.
eEven if economic progress continues unabated
over the next 50 years in the U.S. and other advanced countries—,
whether the citizens of these countries end up rich and happy or rich
and unhappy depends on whether money can buy happiness and
on whether the additional economic resources will, in fact, be used
to obtain additional happiness. To the extent there is a tradeoff
between happiness and other values, the increases in income and wealth
that accompany economic progress are likely to make improvements in
subjective well-being increasingly important for welfare compared to
further improvements in other areas.61
One area where trends
in happiness could have important macroeconomic effects is in the area
of happiness on the job. For example, it is possible that, in
the coming decades, advances in subjective well-being at work
could alter people��s relationship to work in a way that significantly
raises the average retirement age.
Happiness on the job is likely to be an increasingly important element
of competitive advantage—particularly for firms that need to attract
skilled workers who may place a higher dollar value on happiness.
B. Research Strategy.
The ��Hedonic Treadmill,�� ��Easterlin Paradox�� and
��Progress Paradox�� all refer to the lack of secular improvement
in subjective well-being in the face of major increases in per capita
income, improvements in health, and improvements in many other social
indicators. On its face, this paradox seems to present a serious
challenge for Economics. A thoroughgoing resolution of this
paradox is essential for effective integration of subjective well-being
data into Economics.
In order for happiness research to fully tap into the vast accumulated
human capital of the Economics profession, we consider it important
to develop a theory that respects the canons of Economics as well as
the findings of Psychology. One of the most important canons
of Economics is Ordinalism, or the principle of Revealed Preference.
Revealed Preference, applied
to the Easterlin Paradox of nontrending happiness in the face of
dramatic improvements in per capita income and many other areas of life,
, clearly distinguishes utility from happiness. Utility
is the extent to which people achieve what they care about, as indicated
by their choices; happiness is how they feel. This distinction
is important. In particular, the distinction between utility and
happiness leads to many insights and productive questions that would
be difficult to see as long as utility and happiness are confused with
one another.
Appendix: Axiomatics
This paper is not the
place to deal thoroughly with axiomatic issues
about the link between happiness and utility, but it is useful to briefly
address some issues about the preference for baseline mood and the link
between innovations in lifetime utility and elation.
Preference for Baseline
Mood: One element of our discussion above is the
postulate that people value long-term happiness positively.
This postulate cannot be made meaningful without auxiliary hypotheses
distinguishing things that matter for utility only through their effect
on happiness—which is a statement outside of strict revealed-preference
analysis over the goods in the utility function more elementary than
happiness. To illustrate this logical issue, suppose the
flow utility function can be described by the function
U(Kt, Xt,
Mt), where baseline mood
Mt is given by the function
Mt = ��(Kt,
Xt) and the partial derivative
UM
> 0. This flow utility function is obviously
equivalent to the alternative flow utility function
�� defined by ��(Kt,
Xt, Mt)
�� U(Kt,
Xt, 2��(Kt,
Xt)-Mt), but the partial
derivative ��M < 0.
This is not a new issue.
It arises for any Becker-esque treatment of goods produced by a household
production function. For example,
it would not be unreasonable to say
a priori (at least as an approximation)
that, on the benefit side, a washing machine is only valued for
its laundering services, which in turn are only valued for their contribution
to clothing services. This stipulation would then be important
to an analysis of the demand for washing machines.
In an applied context,
we think auxiliary hypotheses of the type needed
to study the place of baseline mood in the utility function
can, in fact, be reasonable. For example,
it might be reasonable to assume that
other than its time and money costs, talk therapy enters the utility
function only through its effect on
baseline mood. This auxiliary assumption, together with evidence
on the size and duration of the effect of
talk therapy on baseline mood would
allow what is otherwise a revealed preference evaluation to be placed
on happiness. To take a more complex example,
one might assume that besides its time and money cost, an antidepressant
medication enters the utility function only through
baseline mood and its medical side-effects.
Preferences over another medication that generates
similar side effects but has few benefits for the individual
might make it possible to evaluate the cost of the side effects.
Then, other than time and money costs,
the vector good of switching from the other ineffective medication to
the effective antidepressant would enter the utility function only through
its effect on baseline mood.
Happiness and News
about Lifetime Utility. In connection with elation, the key
axiomatic issue is whether it is possible to express our main claims
about happiness and news in a way that is independent of any particular
representation of the lifetime utility function.
Let the stochastic process of the control variable vector
X and the state variable vector
k be the fundamentals that lifetime utility depends on.
Define K as a vector giving the history of
k through time t and the history of
X through time t-1.
That is,
Call lifetime utility
vt, as before.
Note that vt
depends on the information at time
t about the future, so news is reflected in changes in
v. We propose the following three axioms relevant to happiness
and news:
Remarks:
These axioms are all ordinal.
The would not be changed in meaning by monotonically increasing transformations
of v and h.
These axioms can be applied
readily to any lifetime utility function that can be expressed
by the terminal condition (perhaps with
in the end) and the recursive relationship
, where is the probability distribution function for
vt+1 .
(Although there would be excess baggage in using the expanded state
vector Kt, this recursive equation can also
be written .) Note that expected utility maximization is
not required for the axioms to be meaningful.
Assuming
the simultaneous equations ultimately
yield a well-defined and well-behaved lifetime
utility function, these axioms can also be applied to lifetime utility
functions like those in section 8 for which, in light of
the expression for happiness given by
News and Happiness Axiom 1, one can represent the inclusion of
overall happiness in the utility function by
an intertemporal equation of the form
together with the terminal
condition .
In conjunction with the
additively time-separable intertemporal expected utility function of
Section 6, for which using the reduced form flow utility function U, and
k for the state variable vector that directly matters for flow
utility, the equation means that we can define a function
Ht
so that
In words,
the entire history of k and
X included in K
allows one to calculate the history of flow utility, which allows one
to back out the history of lifetime utility from
the history of lifetime utility innovations and
. This equation makes it easy to apply Happiness and News
Axioms 2 and 3. Axiom 2 implies that
In addition to , Axiom 3 implies that
for any integer j from 0 to
t-1. This says that recent news about future events will have
a bigger effect on happiness than older news about future events.
(The factor merely puts the comparison between
the effects of lifetime utility innovations
at different lags on a present-value rather than a current-value basis.)
This inequality allows the possibility that distant
enough lags of lifetime utility innovations could have a negative effect
on happiness. Though we do not think
this possibility is empirically relevant, we
also do not think it should be ruled out
a priori.
Finally, we argue that,
other than for the application of Happiness and News Axiom 3, it is
reasonable to include the initial value of lifetime
utility in the comprehensive history state variable vector
K. We interpret as the view of lifetime utility in the
instant before birth begins, when the individual has no information
about her or his life prospects other than the information that is embodied
in genes and body structure at that point.
Because the individual��s information set is biologically limited
up until birth, it is appropriate to view
as an element of the state variable vector
that is not subject to subtle expectational effects.
After this inclusion, we can write happiness as
Happiness At depends on the current (expanded) state variable vector Kt, the current control variable vector Xt, and the history of lifetime utility innovations. This is our essential claim about the nature of happiness given an additively time-separable intertemporal expected utility function. The additivity in the main text equation between the function M(Kt , Xt) of K and X and the function of lifetime utility innovations is only a mathematical and expositional convenience.
References
Argyle, Michael, 1999. ��Causes and Correlates of Happiness,�� chapter 18 in Daniel Kahneman, Ed Diener and Norbert Schwarz eds., Well-Being: The Foundations of Hedonic Psychology, Russell Sage Foundation, New York.
Aristotle, fourth century
B.C.E. Nicomachean Ethics, available in many modern editions.
Becker, Gary, 1965. ��A
Theory of the Allocation of Time,�� Economic Journal
75, pp. 492-517.
J. Bentham (1789) Introduction
to the Principles of Morals and Legislation. 1948 edition, New York:
Hafner.
Biddle, S., and N. Murtrie,
1991. Psychology of Physical Activity and Exercise,
London: Springer/Tavistock/Routledge.
Blanchflower, David G., and
Andrew Oswald, 2004. ��Well-Being
Over Time in Britain and the USA,��
Journal of Public Economics, 88, 1359-1386.
Brickman, P., and D. Campbell,
1971. ��Hedonic Relativism and Planning the Good Society,��
in M. H. Appley ed., Adaptation Level Theory: A Symposium,
New York, Academic Press, pp. 287-302.
Brickman, P., D. Coates and
R. Janoff-Bulman, 1978. ��Lottery Winners and Accident Victims:
Is Happiness Relative?�� Journal of Personality and Social
Psychology, 36 (8), pp. 917-927.
Camerer, Colin, George Loewenstein
and Drazen Prelec, ��Neuroeconomics: How Neuroscience Can Inform Economics,��
Journal of Economic Literature
43 (March 2005), pp. 9—64.
Charles, Kerwin K., 2002.
��Is Retirement Depressing?: Labor Force Inactivity and Psychological
Well-Being in Later Life�� NBER working paper #9033.
Clark, A., 1999. ��Are wages
habit-forming? Evidence from micro data,�� Journal of Economic Behaviour
and Organisation, 39, 179-200.
Conlisk, John, 1996.
��Why Bounded Rationality,�� Journal of Economic Literature
XXIV:2 (June), 669-700.
Constantinides, George, 1990.
��Habit Formation: A Resolution of the Equity Premium Puzzle,��
Journal of Political Economy 98 (June), 519-43.
Davis, Melinda, 2002.
The New Culture of Desire: The Pleasure Imperative Transforming Your
Business and Your Life,
Free Press.
Di Tella, Rafael, Alberto Alesina,
and Robert MacCulloch. "Happiness and Inequality: Are Europeans
and Americans Different?" Journal of Public Economics 88,
nos. 9-10 (August 2004): 2009-42.
Di Tella, Rafael, and Robert
MacCulloch, 1999. ��Partisan
Social Happiness,��
Review of Economic Studies (forthcoming).
Di Tella, Rafael, Robert MacCulloch,
and Andrew J. Oswald, 2003. ��The Macroeconomics of Happiness.��
Review of Economics and Statistics 85 (4), 793-809.
Di Tella, Rafael, Robert MacCulloch,
and Andrew Oswald, 2001. ��Preferences
over Inflation and Unemployment: Evidence from Surveys of Happiness,�� American Economic Review
91 (March).
Diener, E., and Lucas, R. E.,
1999. Personality and subjective well-being. In D. Kahneman, E. Diener
and N. Schwartz (Eds.), Well-Being: The Foundations of
Hedonic Psychology (pp. 213-229). New York: Russell Sage Foundation.
Diener, Ed, and Christie Scollon,
2003. ��Subjective Well-Being is Desirable But Not the
Summum Bonum,�� mimeo, University of Illinois at Urbana-Champaign.
Easterbrook, Gregg, 2003.
The Progress Paradox: How Life Gets Better While People Feel Worse,
Random House, New York.
Easterlin, Richard A., 1974
. ��Does Economic Growth Improve the Human Lot?: Some Empirical
Evidence,�� in P. A. David and W. R. Levin eds., Nations and Household
in Economic Growth, Stanford University Press, pp. 98-125.
Easterlin, Richard A., 1995.
��Will Raising the Incomes of All Increase the Happiness of All?��
Journal of Economic Behavior and Organization
27, pp. 35-47.
Easterlin, Richard A., 2003.
��Explaining Happiness,�� Proceedings of the National Academy of
Science, 100 (19), pp. 11176-11183.
Frank, R. H., 1985. Choosing
the right pond. New York: Oxford University Press
Frank,
R., 1999. Luxury Fever: Why Money Fails to Satisfy in an Era of Excess,
New York: N.Y. Free Press.
Frederick, Shane, and George
Loewenstein, 1999. ��Hedonic Adaptation,�� chapter 16 in Daniel
Kahneman, Ed Diener and Norbert Schwarz eds., Well-Being: The Foundations
of Hedonic Psychology, Russell Sage Foundation, New York.
Frey, Bruno S., Simon Luechinger
and Alois Stutzer, 2004. ��Valuing Public Goods: The Life Satisfaction
Approach,�� Center for Economic Studies and Ifo Institute for Economic
Research Working Paper #1158, University of Zurich.
Frey, Bruno S. and Stutzer,
Alois, 2003, ��Testing Theories of Happiness�� In
Paradoxes of Happiness in Economics, Conference held at the University
of Milano-Bicocca, March 21-23, 2003.
Frey, Bruno S., and Alois Stutzer,
2000. ��Happiness, Economy and Institutions,��
The Economic Journal, 110 (October), 918-938.
Frey, Bruno S., and Alois Stutzer,
2002. ��What Can Economists Learn from Happiness Research?��
Journal of Economic Literature, 40 (June), pp. 402-435.
Frey, Bruno S., and Alois Stutzer,
2004a. ��Happiness Research: State and Prospects,�� Institute
for Empirical Research in Economics Working Paper #190, University of
Zurich.
Frey, Bruno S., and Alois Stutzer,
2004b. ��Economic Consequences of Mispredicting Utility,�� Institute
for Empirical Research in Economics Working Paper #218, University of
Zurich.
Gardner, Jonathan, and Andrew Oswald, 2001. ��Does Money Buy Happiness? A Longitudinal Study Using Data on Windfalls,�� mimeo, available at
http://www2.warwick.ac.uk/fac/soc/economics/staff/faculty/oswald/
Geneakoplos, John, David Pearce
and Ennio Stacchetti, 1989. ��Psychological Games and Sequential
Rationality,�� Games and Economic Behavior 1, 60-79.
Gottfried, Jay A., John
O��Doherty and Raymond J. Dolan, 2003.
��Encoding Predictive Reward Value in Human Amygdala and Orbitofrontal
Cortex,�� Science 301 (22 August), 1104-1107.
Gruber, Jonathan, and Sendhil
Mullainathan, 2002. ��Do Cigarette Taxes Make Smokers Happier?��
NBER Working Paper #8872.
Hamermesh, Daniel S., forthcoming.
��Subjective Outcomes in Economics,�� Southern Economic Journal.
Hall, Robert E., and Charles
I. Jones, 2004. ��The Value of Life and the Rise in Health Spending,��
National Bureau of Economic Research Working Paper #10737.
Harsanyi, John C., 1953.
��Cardinal Utility in Welfare Economics and in the Theory of Risk-Taking,��
Journal of Political Economy, 61 (October), pp. 434-435.
Harsanyi, John C., 1955.
��Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons
of Utility,�� Journal of Political Economy, 63, pp. 309-321.
Heckman, James J. and
MaCurdy, Thomas E. (1980) ��A Life Cycle Model of Female Labor Supply.��
Review of Economic Studies 47: 47-74.
Helliwell, John F., 2002.
��How��s Life? Combining Individual and National Variables to
Explain Subjective Well-Being,�� NBER Working Paper #9065.
Hicks, J. R. and Allen, R. G. D., ��A Reconsideration of the Theory of Value��, Economica,
February/May, 1934, pp.52-60
and 196-219.
Hirsch, F. (1976). Social
limits to growth. Cambridge, MA: Harvard University Press.
Hirshleifer, David, and Tyler
Shumway, 2003. ��Good Day Sunshine: Stock Returns and the Weather,��
Journal of Finance, 1009-1032.
Kahneman, Daniel, Ed Diener
and Norbert Schwarz, 1999. Well-Being: The Foundations of Hedonic
Psychology, edited volume, Russell Sage Foundation, New York.
Kahneman, Daniel, 1999.
��Objective Happiness,�� chapter 1 in Daniel Kahneman, Ed Diener and
Norbert Schwarz eds., Well-Being: The Foundations of Hedonic Psychology,
Russell Sage Foundation, New York.
Kahneman, Daniel, Alan B. Krueger,
David Schkade, Norbert Schwarz, and Arthur A. Stone, 2004.
��Toward National Well-Being Accounts,�� American Economic Review,
94, pp. 429-434.
Kahneman, Daniel, Alan B. Krueger,
David A. Schkade, Norbert Schwarz and Arthur A. Stone, 2004. ��A
Survey Method for Characterizing Daily Life Experience: The Day Reconstruction
Method (DRM),�� Science
(December), 1776-1780.
Kahneman, Daniel, and Amos
Tversky, 1979. ��Prospect Theory: An Analysis of Decision Under
Risk,�� Econometrica, 47 (2), 263-291.
Kaprio, J., M. Koskenvuo, and
H. Rita, 1987. ��Mortality after Bereavement: A Prospective Study
of 95,647 Widowed Persons,�� American Journal of Public Health
77, pp. 283--287.
Kapteyn, A., and Wansbeek,
T., 1985. ��The individual welfare function: A review,�� Journal
of Economic Psychology, 6, 333-363.
Kezdi, Gabor, and Robert J.
Willis, 2003. ��Who Becomes a Stockholder? Expectations,
Subjective Uncertainty, and Asset Allocation,�� mimeo, University of
Michigan.
Kimball, Miles, Helen
Levy, Fumio Ohtake and Yoshiro Tsutsui, 2006.
��Unhappiness after Hurricane Katrina,�� NBER Working Paper #12062
(March).
Kreps, David, and Evan Porteus,
1978. ��Temporal Resolution of Uncertainty and Dynamic Choice
Theory,�� Econometrica, 46, 185-—200.
Laibson, David, 1997.
��Golden Eggs and Hyperbolic Discounting,�� Quarterly Journal of
Economics, 62 (2), pp. 443--478.
Layard, Richard, 2003.
Lane, Robert E., 2000.
The Loss of Happiness in Market Democracies, Yale University
Press, New Haven, pp. 141-158.
Larsen, Randy J., and
Barbara L. Fredrickson, 1999.
��Measurement Issutes in Emotion Research,��
chapter 3 in Daniel Kahneman, Ed Diener and Norbert Schwarz eds.,
Well-Being: The Foundations of Hedonic Psychology, Russell Sage
Foundation, New York.
Layard, Richard, 2005. Happiness.
London, UK: Penguin Press. This book is based to a large extent
on Layard��s 2003 Lionell Robbins Memorial Lectures, London School
of Economics.
Lewinsohn, P. M., J. M. Sullivan
and S. J. Grosscup (1982). ��Behavioral Therapy: Clinical Applications,��
in A. J. Rush (ed.), Short-Term Therapies for Depression
(pp. 50-—87), New York: Guilford.
Lodge, David. (1978) Changing
Places. Penguin Books.
Loewenstein, George, Ted
O��Donoghue and Matthew Rabin (2002),
��Projection Bias in Predicting Future Utility,�� mimeo. Carnegie-Mellon
University.
Loewenstein, George, and David
Schkade, 1999. ��Wouldn��t It Be Nice? Predicting Future
Feelings,�� chapter 5 in Daniel Kahneman, Ed Diener and Norbert Schwarz
eds., Well-Being: The Foundations of Hedonic Psychology, Russell
Sage Foundation, New York.
Lucas, Richard E., Andrew Clark,
Yannis Georgellis, and Ed Diener, 2004. ��Unemployment Alters
the Set Point for Life Satisfaction,�� Psychological Science
15 (1), pp. 8--13.
Lupton, Joseph, 2002.
��The Habit Liability in Life-Cycle Consumption and Portfolio Choice,��
mimeo, Federal Reserve Board.
Luttmer, E.F.P., 2004. ��Neighbors
as negatives: Relative earnings and well-being,�� NBER Working Paper,
No. w10667.
Lykken, D., and Tellegen, A.,
1996. Happiness is a stochastic phenomenon. Psychological Science,
7, 186-189.
Marmot, Michael, 2004.
How Our Position on the Social Gradient Affects Longevity and Health,
London: Bloomsbury.
MaCurdy, Thomas E. (1980)
��An Empirical Model of Labor Supply in a Life-Cycle Setting.��
Journal of Political Economy 89: 1059-1085.
MaCurdy, Thomas (1985)
��Interpreting Empirical Models of Labor Supply in An Intertemporal
Framework with Uncertainty.�� In J.J. Heckman and B. Singer, eds. Longitudinal
Analysis of Labor Market Data (Cambridge University Press, Cambridge
UK).
MaCurdy, Thomas E. and
Blundell, Richard (1999) ��Labor Supply: A Review of Alternative Approaches.��
In O. Ashenfelter and D. Card., eds. Handbook of Labor Economics, Vol.
3A. Elsevier: 1559-1695.
Maslow, Abraham, 1943.
"A Theory of Human Motivation"
Psychological Review, 50, pp.
370-396.
Morris, Desmond, 2004.
The Nature of Happiness, Little Books Ltd., London.
Morris, W. N., 1999. The mood
system. In D. Kahneman, E. Diener and N. Schwartz (Eds.),
Well-Being (pp. 169-189). New York: Russell Sage Foundation.
Nesse, Randolph M., 2000. Is
depression an adaptation? Archives of General Psychiatry, 57,
14-20.
Nesse, Randolph M., 2001. Motivation
and melancholy: a darwinian perspective. Nebr Symp Motiv, 47,
179-203.
Nesse, Randolph M., 2004. Natural
selection and the elusiveness of happiness. Philos Trans R Soc Lond
B Biol Sci, 359, 1333-1347.
Nesse, Randolph M., forthcoming.
��Natural Selection and the Regulation of Defenses,�� Evolution
and Human Behavior.
Nickerson, Carol, Norbert Schwarz,
Ed Diener, and Daniel Kahneman, 2003. ��Zeroing in on the Dark
Side of the American Dream: A Closer Look at the Negative Consequences
of the Goal for Financial Success,�� Psychological Science,
14 (November), 531-536.
Nisan, M.,
1972. ��Dimension of time in relation to choice behavior and achievement
orientation,�� Journal of Personality and Social Psychology,
21, 175-182.
O��Doherty, John P.,
2005. ��Reward Representations and Reward-Related Learning in
the Human Brain: Insights from Neuroimaging,�� Wellcome Department
of Imaging Neuroscience.
O��Doherty, John P.,
Peter Dayan, Karl Friston, Hugo Critchley and Raymond Dolan, 2003.
��Temporal Difference Models and Reward-Related Learning in the Human
Brain,�� Neuron 28 (April 23), 329-337.
O��Guinn, T.C. and Shrum,
L. J., 1997. ��The role of television in the construction of consumer
reality��. Journal of Consumer Research,
23, 278-294
Pareto, Vilfredo, 1906.
Manuel of Political Economy trans. by Ann S Schwier, edited by Ann S.
Schwier and Alfred N. Page (New York: Augustus M. Kelley, 1971).
Pattanaik, P. K., 1968.
��Risk, Impersonality and the Social Welfare Function,��
Journal of Political Economy, 76 (December), 1152-1169.
Patterson, D. R., J. J. Everett,
C. H. Bombardier, K. A. Questad, V. K. Lee, and J. A. Marvin, 1993.
��Psychological Effects of Severe Burn Injuries,�� Psychological
Bulletin 113, pp. 362--378.
Pinker, Stephen, 1997.
How the Mind Works, Norton.
Putnam, Robert D., 2000.
The Collapse and Revival of American Community,
Simon & Schuster.
Rawls, John, 1951.
��Outline of a Decision Procedure for Ethics,��
Philisophical Review, 60, pp. 177-197.
Rawls, John, 1958. ��Justice as Fairness,�� Philosophical Review, 67, pp. 164-194.
Rawls, John, 1971.
A Theory of Justice, revised edition 1999, Belknap Press of Harvard
University Press, Cambridge, MA.
Rayo, Luis and Gary S. Becker,
2005. ��Evolutionary Efficiency and Happiness,�� mimeo, University
of Chicago.
Reich, J. W., and
J. Zautra, 1981. ��Life Events and Personal Causation: Some Relationships
with Satisfaction and Distress,�� Journal of Personality and Social
Psychology, 41, 1002-—1012.
Robbins, Lionel C., 1932.
An Essay on the Nature and Significance of Economic Science. 1984
edition, New York: New York University Press.
Robinson, Michael D. and
Gerald L. Clorem 2002. ��Belief
and Feeling: Evidence for an Accessibility Model of Emotional Self Report,��
Psychological Bulletin, 128, 934-960.
Rosen, Sherwin (1986) "The
Theory of Equalizing Differences," in O. Ashenfelter and R. Layard
(eds.), Handbook of Labor Economics
Samuelson, Paul A., 1938.
��A Note on the Pure Theory of Consumer's Behaviour,�� Economica
February), 61—71.
Samuelson, Paul A., 1947.
Foundations of Economic Analysis.
Cambridge: Harvard University Press, (enlarged edition, 1983), esp.
chapters V, VII, and VIII.
Schkade, D. A. and Kahneman,
D. (1998). Does living in California make people happy? A focusing illusion
in judgments of life satisfaction. Psychological Science, 9 (5),
340-529.
Schor, Juliet B. (1991)
The Overworked American: The Unexpected Decline of Leisure.
New York. Basic Books.
Schultz, Wolfram, and Anthony
Dickinson, 2000. ��Neuronal Coding of Prediction Errors,��
Annual Review of Neuroscience, 23: 473—500.
Schwartz, Barry, 2004.
The Paradox of Choice: Why More is Less, New York: HarperCollins.
Schwarz, N. (1987).
Stimmung als Information: Untersuchungen zum Einfluß von Stimmungen
auf die Bewertung des eigenen Lebens. [Mood as information.] Heidelberg,
Germany: Springer Verlag.
Schwarz, N. (1990). Feelings
as information: Informational and motivational functions of affective
states. In E. T. Higgins and R. M. Sorrentino (Eds.), Handbook
of motivation and cognition: Foundations of social behavior (Vol.
2; pp. 527-561). New York, NY: Guilford Press.
Schwarz, N., 1996. Cognition
and communication: Judgmental biases, research methods and the logic
of conversation. Hillsdale, NJ: Erlbaum.
Schwarz, N., 1999. ��Self-reports:
How the questions shape the answers,�� American Psychologist, 54,
93-105.
Schwarz, N. (2002). Situated
cognition and the wisdom of feelings: Cognitive tuning. In L.
Feldman Barrett and P. Salovey (Eds.), The wisdom in feelings
(pp.144-166). New York: Guilford.
Schwarz, N., and Bohner, G.,
2001. ��The construction of attitudes,�� In A. Tesser and N. Schwarz
(Eds.), Blackwell handbook of social psychology: Intraindividual
processes (Vol.1, pp. 436-457). Oxford, UK: Blackwell.
Schwarz, N., and Clore, G.L.,
1983. Mood, misattribution, and judgments of well-being: Informative
and directive functions of affective states. Journal of
Personality and Social Psychology, 45, 513 - 523.
Schwarz, Norbert and Fritz
Strack, 1999. ��Reports of Subjective Well-Being: Judgmental
Processes and Their Methodological Implications,�� chapter 4 in Daniel
Kahneman, Ed Diener and Norbert Schwarz eds., Well-Being: The Foundations
of Hedonic Psychology, Russell Sage Foundation, New York.
Seligman, Martin E. P., 2002.
Authentic Happiness: Using the New Positive Psychology to Realize
Your Potential for Lasting Fulfillment, Free Press, New York.
Sen, Amartya, 1983. ��Poor,
relatively speaking,�� Oxford Economic Papers, 35, 153-169.
Seymour, Ben, John P.
O��Doherty, Martin Koltzenburg, Katja Wiech, Richard Rackowiak, Karl
Friston and Raymond Dolan, forthcoming.
��Opponent appetitive-aversive neural processes underlie predictive
learning about pain relief,�� Nature Neuroscience.
Seymour, Ben, John P.
O��Doherty, Peter Dayan, Martin Koltzenburg, Anthony K. Jones, Raymond
J. Dolan, Karl J. Friston, 2004.
��Temporal Difference Models Describe Higher-Order Learning in Humans,��
Letters to Nature 429 (June), 664-667.
Shepperd, J. A., Findley-Klein,
C., Kwavnick, K. D., Walker, D., and Perez, S.,
2000. ��Bracing for loss,�� Journal of Personality and Social Psychology,
78, 620-634.
Shepperd, J. A., Ouellette,
J. A., and Fernandez, J. K., 1996.
��Abandoning unrealistic optimism: Performance estimates and the temporal
proximity of self-relevant feedback,��
Journal of Personality and Social Psychology, 70, 844-855.
Sims, Christopher, 2002.
��Implications of Rational Inattention,��
unpublished paper, Princeton University.
Solnick, Sara J., and David
Hemenway, 1998. ��Is More Always Better?: A Survey of Positional
Concerns,�� Journal of Economic Behavior and Organization, 37,
373--383.
Steffick, Diane E., 2000.
��Documentation of Affective Functioning Measures in the Health and
Retirement Study,�� HRS Documentation Report DR-005, available
online at http://hrsonline.isr.umich.edu/docs/userg/dr-005.pdf
Steptoe, A., J. Kimball and
P. Basford, 1996. ��Exercise and the Experience and Appraisal
of Daily Stressors: A Naturalistic Study,�� Journal of Behavioral
Medicine, 21, 363—374.
Stigler, George, 1950.
"Development of Utility Theory," Journal of Political Economy
58, pts 1-2 (August-October), pp. 307-327, 373-396.
Strack, F., Martin, L.L., and
Schwarz, N. (1988). Priming and communication: The social determinants
of information use in judgments of life-satisfaction. European Journal
of Social Psychology, 18, 429 - 442.
Strack, F., Schwarz, N., Chassein,
B., Kern, D., and Wagner, D. (1990). The salience of comparison standards
and the activation of social norms: Consequences for judgments of happiness
and their communication. British Journal of Social Psychology, 29,
303-314.
Strack, F., Schwarz, N., and
Gschneidinger, E. (1985). Happiness and reminiscing: The role of time
perspective, mood, and mode of thinking. Journal of Personality and
Social Psychology, 49, 1460 -1469.
Sulls, J., and Wills, T.A.
(Eds) (1991). Social comparison: Contemporary theory and research.
Hillsdale, NJ: Erlbaum.
Thayer, R. E., 1989.
The Biopsychology of Mood and Emotion, New York: Oxford University
Press.
Tsutsui, Yoshiro, Miles
Kimball and Fumio Ohtake, 2005.
��Koizumi Carries the Day: Happiness and Unhappiness after the Japanese
Election,�� Osaka University and University of Michigan.
Wilson, Timothy D., 2002.
Strangers to Ourselves: Discovering the Adaptive Unconscious,
Harvard University Press, Cambridge, MA and London, England, especially
chapter 6, pp. 137-159.
Wolfers, Justin, 2003.
��Is Business Cycle Volatility Costly? Evidence from Surveys
of Subjective Well-Being,�� NBER Working Paper #9619.
Woodford, Michael, 2002. ��Imperfect Common Knowledge and the
Effects of Monetary Policy,"
in P. Aghion, R. Frydman, J. Stiglitz, and M. Woodford, eds., Knowledge,
Information, and Expectations in Modern Macroeconomics: In Honor of
Edmund S. Phelps, Princeton Univ. Press.
Wortman, C. and R. Silver,
1987. ��Coping with Irrevocable Loss,�� in Cataclysms, Crises
and Catastrophes: Psychology in Action, Master Lecture Series, v.
6, Washington, D. C., American Psychological Association, pp. 189-235.
Zamble, E., 1992. ��Behavior
and Adaptation in Long-Term Prison Inmates: Descriptive Longitudinal
Results,�� Criminal Justice and Behavior
19, pp. 409--425.
Zamble, E., and F. Proporino,
1990. ��Coping, Imprisonment, and Rehabilitation: Some Data and
Their Implications,�� Criminal Justice and Behavior
17, pp. 53-70.
1
We would like to acknowledge first and foremost the substantial contributions
of Norbert Schwarz to this paper. Our discussions with him from
the very first beginnings of this paper clarified many things for us,
particularly about the empirical evidence on happiness measures.
However, we need to make clear that there are important aspects of our
theoretical position he would not agree with. In addition
to Norbert Schwarz, we would like to thank George Akerlof, Toni Antonucci,
Robert Barsky, Kerwin Charles, Fred Conrad, Mick Couper, Michael
Elsby, Gwenith Fisher, Bruno Frey,
Christopher House, Michael Hurd, Helen Levy,
Charles Manski, Randolph Nesse, Fumio Ohtake, Antonio Rangel,
Luis Rayo, Matthew Shapiro, Daniel Silverman, Alois Stutzer,
Yoshiro Tsutsui, Janet Yellen and participants in a
seminars at Osaka University, the Stanford Institute
for Theoretical Economics, University of Michigan, Harvard, and Brown
for helpful discussions and comments on early versions of this material.
They must also be absolved from complicity in any errors we perpetrate,
small or large. We are
grateful for support from National Institute on Aging grant P01 AG026571-01.
2 For more of the history of these developments, see also George Stigler (1950).
3 Kahneman (1999), Gruber and Mullainathan (2002), Frey and Stutzer (2004b), and Layard (2005) are some of the most explicit in equating happiness and flow utility.
4 One can then make the leap from individual
preferences to statements about social welfare on more or less the
an same terms as one could in the absence of happiness data.
To the extent that happiness data give the illusion of providing a cardinal
utility function, it is an illusion similar to that provided by expected
utility theory—where one may sometimes need to be reminded that a
monotonic transformation f(E(U)) of the overall
objective function E(U) leaves preferences unaltered. Just
as there is no necessary reason why the curvature of U
in the expected utility representation E(U) tells us how to aggregate
preferences interpersonally, there is no necessary
reason why whatever structure is revealed in preferences as they
relate to happiness data tells how preferences must be aggregated.
At a minimum, any debate about what happiness says about social welfare
must take into account the existing literature on social welfare and
social choice theory.
5 Some of the recent empirical papers in economics using happiness data are John Helliwell (2002), David Blanchflower and Andrew Oswald (2004), Clark (1999), Rafael Di Tella, Alberto Alesino and Robert MacCulloch (2004), Di Tella and MacCulloch (1999), Di Tella, MacCulloch and Oswald (2001, 2003), and Wolfers (2003).
6 See for example Frey, Simon Luechinger and Stutzer (2004) and Frey and Stutzer (2000, 2004a).
7 See for example, Jonathan Gruber and Sendhil Mullainathan (2002) and Frey and Stutzer (2004b).
8 See, for example, Daniel Hamermesh (forthcoming).
9 Kahneman calls momentary affect ��instant utility,�� but here, to avoid confusion, it is best to reserve the term ��utility�� for the concept of overall individual welfare in Economics.
10 In general, self-report measures of emotions can be affected by social desirability and the semantic framing effects that arise cross-culturally, and lack of conscious awareness of emotions. For the most part, social desirability and semantic framing effects should be fairly constant over time within a given culture and can be dealt with empirically using fixed effects. The likelihood that people might lack conscious awareness of emotions is a subject of debate within Psychology. Some psychologists insist on conscious awareness as part of the definition of an emotion. (Larsen and Fredrickson, 1999 reports that ��some would question whether an unperceived emotion is an emotion at all.��) But even Tim Wilson (2002), in a book-length argument for the possibility of unconscious feelings, points out that ��feelings differ from the rest of the adaptive unconscious in their potential to reach awareness�� and allows that ��It might even be the case that the default is for feelings to emerge into awareness, and that it takes special circumstances to prevent them from doing so.�� (See Wilson, 2002, p. 134.) It seems likely that the overall positive or negative aspect of feelings that we are focusing on under the label of ��happiness�� makes it into consciousness more reliably than the detailed reasons behind feelings or finer categorizations of emotions. Wilson (2002) goes on to discuss repression, inattention and ��the obscuring of feelings by the smoke screen of people��s conscious theories and confabulations.�� Repression and inattention seem unlikely to cause serious problems for the survey measurement of happiness. However, ��the smoke screen of people��s conscious theories�� about happiness is a serious issue, which we address below.
11 Michael Robinson and Gerald Clore (2002, p. 950) looked at evidence on happiness reports with different time frames. Their evidence led them to conclude that a few week��s time is the longest interval for which one can get happiness reports that are not contaminated in an important way by people��s theories of how they ��should�� feel.
12 In the first wave respondents were instead asked ��Please tell me how often you have experienced the following feelings during the past week: all or almost all of the time, most of the time, some of the time or none or almost none of the time.��
13 See Steffick (2000) for a detailed description and assessment of the CES-D questions in the HRS. Besides omitting the other less relevant questions, we have reversed the order of the first two questions even after those omissions in order to give the version of the question that we would recommend for use on other surveys that do not have a more extensive CES-D battery of questions.
14 See Layard, 2005.
15 We are particularly grateful to Norbert Schwarz for this summary of the psychological research on different subjective well-being measures.
16 The relationship of such context-dependence to decision-making is an important research question. For example, Hirshleifer and Shumway (2003) indicates that sunny days have a detectable effect on stock-market trading.
17 See also Schwarz (1996, 1999) and Schwarz and Bohner (2001).
18 For example, consider the fact reported by Lucas, Clark, Georgellis and Diener (2004), that life satisfaction is permanently dragged down by an episode of unemployment. Even if the affective sting of past unemployment has long since faded away, asking for an overall evaluation of life satisfaction invites the respondent to evaluate the past as well as the present. It is not surprising that a past episode of unemployment permanently affects one��s assessment of one��s autobiography.
19
Our choice of the ��social ladder�� metaphor is influenced by the
form of a question on the HRS leave-behind survey asking people to mark
their perceived social rank on a printed ladder, pioneered by Michael
Marmot��s Whitehall II studies of British civil servants.
(See Marmot, 2004.).
20 In the future, this bias could go the other way, since the ratio of family size for high-status parents to family size for low-status parents seems to be falling over time.
21 Note that, in comparing the past to the present, it is important to abstract from people��s preference for the familiar and status-quo bias more generally, which would have worked in favor of the actual experience in the past as much as in the present. A good way to abstract from the attraction of familiar idiosyncratic details of one��s life is to imagine a choice between being (a) thrown into the life of a randomly chosen individual in the present and (b) being thrown into the life of a randomly chosen individual in the past. Unfortunately, there is no such helpful device to help in abstracting from the familiarity of one��s entire era.
22 For some direct evidence on the strength of preferences over social rank based on hypothetical choices, see Solnick and Hemenway (1998). For a discussion of social comparison by psychologists, see Sulls and Wills (1991).
23 There are some instructive instances of social rank diverging from income rank even in the present-day U.S. Clergy and teachers (including professors) often have considerably higher social rank than their income rank. This relatively high social rank is important in making many people willing to sacrifice a significant amount of income to go into these fields.
24 Note that, at $9600, the per capita GDP of Mexico—an important source of migration to the U.S.—is not far below the U.S. real per capita GDP in 1955.
25 There is a practical problem of distinguishing between the force of a calculation or line of reasoning itself and the desire to agree with the person urging that line of reasoning. In principle there are ways to deal with that problem. For example, in presenting a hypothetical choice, it is important to even-handedly present correct calculations and lines of reasoning that favor both the pro and con side of a decision. Also, to minimize social pressure, it may be possible to present calculations and lines of reasoning by a prepared text or an interactive computer setup. Making sure the agent is able to make the decision with as much anonymity as possible may also be helpful.
26 Calculation and deliberation costs should be recognized just as much as any other costs an agent faces. The difficulties in modeling ��bounded rationality�� problems due to the ��infinite regress�� problem discussed by John Conlisk (1996) among others should not blind us to the obvious fact of deliberation costs. Because we do not view the recognition of deliberation costs as a departure from ��rationality�� at all, we favor the more neutral term ��bounded cognition�� for what has traditionally been called ��bounded rationality.��
27 In their paper, Rayo and Becker (2005) call the filtered version of underlying utility ��happiness.�� However it is unclear in what way it would relate to happiness as we are using the term. In particular, like visual processing (which they use as an analogy), the filtered version of utility they discuss might operate at a very early unconscious or ��automatic�� stage in the sense of Colin Camerer, George Loewenstein and Drazen Prelec (2005) and so could be several cerebral processing steps prior to ��happiness�� in the experiential sense.
28 Insisting on transitivity is one aspect of ��thoughtfulness�� here.Thus, in principle, in assessing preferences, we would rely on an individual��s deliberative choices for an entire menu of decisions at once, with an iterative process where the individual is forced to resolve non-transitivities.
29 This is a very interesting empirical question. In testing whether intertemporal tradeoffs in utility match intertemporal tradeoffs in happiness alone, one must address the problem that people are not good at predicting their future happiness, as pointed out by Loewenstein and Schkade (1999). It may be possible to address this problem with some combination of educating people about the likely consequences of a decision for future feelings and eliciting what their expectations about future feelings are after that education to control for any remaining mispredictions of future feelings.
30 Of course, in a large fraction of cases of attempted persuasion about preferences, the desired preferences for the other person will be given the rhetorical label ��happiness,�� ��true happiness,�� ��genuine happiness,�� or ��authentic happiness,�� regardless of how important happiness in the narrow sense of positive affect is in those preferences. For logical clarity (which can be at variance with persuasive power), the phrase ��recommended preferences�� can be substituted in place of ��true happiness�� or similar phrases. Aristotle��s use of eudaimonea (the Greek word for happiness) in the Nicomachean Ethics (fourth century B.C.E.) can be seen as an example of using ��happiness�� as a label for such recommended preferences. Saying this in no way diminishes the cogency of Aristotle��s recommendations.
31 Note that with a finite horizon, the two formally similar versions of the utility function would no longer represent exactly the same preferences. The lifetime utility function would imply a greater tendency to consume in the period immediately before death than . However, the lifetime utility function would still be equivalent to the lifetime utility function . This equivalent form with ��flow utility�� depending only on current consumption might easily be more convenient, despite the odd-looking coefficient on f(CT).
32 Note that while there is good reason to hope that utility will be higher in the future, it is not clear that the Easterlin Paradox will continue into the future. It is possible that average long-run happiness will be significantly higher in the future.
33 In the absence of an adequate theory of the relationship between utility and happiness, it is best to be cautious about asserting that people are making systematic optimization mistakes even when it is clear that people are making mistakes in predicting the dynamics of happiness. We return to this issue after presenting our theory of the relationship.
34 See Kahneman, Krueger, Schkade, Schwarz and Stone (2004) on the average level of affect experienced during different activities. As one unsurprising example, people experience higher affect while eating than the affect they experience while doing housework.
35 See Diener and Lucas (1999).
36 See Thayer (1989), Biddle and Murtrie (1991), Steptoe, Kimbell and Basford (1996) and Argyle (1999).
37 See Lewinsohn, Sullivan and Grosscup (1982), Reich and Zautra (1981) and Argyle (1999).
38 See Luttmer (2004).
39 See Kahneman, Krueger, Schkade, Schwarz and Stone (2004).
40
The hypothesis that in the future of rich countries
baseline mood will be a luxury good
is inspired by Maslow (1943), who argues that once basic needs (such
as physiological and safety needs) are satisfied, higher needs (such
as needs for love, belonging, esteem and actualization) come to the
fore. Both long-run happiness
at home and long-run happiness at work might exhibit strong income effects.
However, oOne contrary
bit of evidence running contrary to this idea
against that
baseline mood is being
a luxury good is that in the Hindhu and Buddhist traditions a great
deal of time and effort were often devoted to baseline-mood-raising
meditation even thousands of years ago at much lower levels of per capita
income than today.
41 This view of the value of pinning down the determinants of baseline mood is consistent with the program of Positive Psychology, as described by Seligman (2002).
42 This view of the value of pinning down the determinants of baseline mood is consistent with the program of Positive Psychology, as described by Seligman (2002).
43 Television may have enhanced the negative effect of social rivalry on happiness by leading people to believe the distribution of income and other advantages in society is higher than it actually is, leading people to underestimate their true social rank. See O��Guinn and Shrum (1997).
44 The strength of Kahneman��s case depends in important measure on whether, as he argues, there is no way to construct a consistent underlying set of preferences from the contradictory decisions people make, even after following the approaches discussed above in Section IV, ��Measuring utility.��
45 The extension of Kahneman��s proposal to the true mathematical expectation in uncertain situations is not explicit in Kahneman (1999), but it seems a reasonable interpretation.
46 Note that only exogeneity of the conditional mean of baseline mood is needed for this result. An ability to control the variance of baseline mood, with no effect on the mean, would still leave elation totally dominant in the expected present discounted value of happiness.
47 In an analogy to exotic financial securities due to George Akerlof when he first heard about elation, elation provides a kind of tranche of current and future effects on flow utility.
48 Nevertheless, tThere
is evidence people do some of this kind of manipulation of memories
of past expectations. Nisan (1972) finds that study participants
taking an immediate test were less confident than those taking a test
in 4 weeks. Similarly, Shepperd, Ouellette, & Fernandez
(1996) find that college seniors were more muted in estimated first-job
salaries than sophomores and juniors. (See also Shepperd,
Findley-Klein, Kwavnick, Walker and Perez (2000).,)
In each case, confidence was reduced when proximity to performance outcomes
waswere more immediate. We are grateful to Norbert
Schwarz for cluing us in to this evidence.
49 There are some other possible extensions
of the model that we cannot give a serious discussion to here.
One of the more interesting is the possibility that elation responds
more to news about whether one��s choices worked out than to news about
things beyond one��s control. That would make it possible to
manipulate elation by labeling good events as due to one��s efforts,
while labeling bad events aswere beyond
one��s control.
50
We choose �� –Greek psi
in the shape of a pitchfork—to symbolize the hellishness of working
through bad news.
51 A more realistic model might make the flow of elation and dismay increasing in the magnitude of the news, and greater in absolute magnitude for bad news than good news, but this assumption will do for our example.
52 One facet of this model worth pointing
out is that, other than the kink at zero, the curvature of g(��0)
depends on the discount rate ��. It may be that this will
not provide sufficient curvature to match the observations that motivate
Prospect Theory. But in a more complex model, hyperbolic
discounting could come to the rescue by providing a high discount rate
in the first little while. In any case, it is easy to add mechanisms
that generate more curvature. Most obviously, one can posit that
the amount of cognitive processing required goes up less than proportionately
with the magnitude of news, so that and , where the function is
increasing and concave for positive values and uniformly zero for negative
values.
53 These are psychological preferences in the same sense as those appearing in the ��psychological games�� studied by Geneakoplos, Pearce and Stacchetti (1989).
54 Norbert Schwarz, personal communication, and Kahneman, Krueger, Schkade, Schwarz and Stone (2004). The results are not definitive because of the lack of a good econometric instrument for hours of sleep that is known a priori not to affect happiness directly.
55 Kerwin Charles (2002) is a good example of the kind of attention to exogeneity in happiness research that we mean.
56 This could be done as a regular monthly rider on the Survey of Consumer Attitudes—which also has a wide range of potentially relevant expectations data— for under $40,000 per year. Day and time stamps to the data would yield daily data, after some correction for the nonrandomness of the sample within a month. (Hard-to-catch people tend to be interviewed later in the month.) We have identified funding for the first few months of such data collection, but need to find long-term funding for such a data-collection effort.
57 A similarity between elation and is that econometric identification of both spikes in elation and changes in requires subtracting out an individual fixed effect. It may also be necessary to control for a few other factors that have predictable effects on changes in subjective well-being or behavior. For example, the real interest rate can have predictable effects on the evolution of consumption and labor supply, even in the absence of news, while time-varying determinants of baseline mood can have predictable effects on overall subjective well-being.
58 The Ordinalist Revolution made it clear that the key philosophical issues in judging social welfare for purposes of public policy could not be avoided even if a perfect direct measure of individual welfare existed. Most notably, there is no easy escape from the difficulties surrounding interpersonal comparison. For example, should those with more refined tastes who can distinguish more minute differences in quality therefore be accorded greater weight in social choice? See Stigler (1950).
59 Note that for the set of things that only enter utility through happiness, valuing them at their effect on happiness times the revealed preference dollar value of happiness is a very different procedure from the common valuation procedure of dividing the effect of a variable on happiness by the effect of income on happiness. This procedure is wrong because it assumes that income only affects utility by affecting happiness—something we know to be false for income to the extent that people thoughtfully sacrifice happiness for higher income. In order to use income as a numeraire, all the benefits of income on utility need to be accounted for, not just the (possibly small) fraction of the benefits of income that show up in a higher level of happiness.
60 In this, Layard follows Gruber and Mullainathan (2002) and Frey and Stutzer (2004b).
61 One area where trends in happiness could have important macroeconomic effects is in the area of happiness on the job. For example, it is possible that, in the coming decades, advances in subjective well-being at work could alter people��s relationship to work in a way that significantly raises the average retirement age. Happiness on the job is likely to be an increasingly important element of competitive advantage—particularly for firms that need to attract skilled workers who may place a higher dollar value on happiness.
All Rights Reserved Powered by Free Document Search and Download
Copyright © 2011