Utility and Happiness Miles Kimball and Robert Willis

Utility and Happiness Miles Kimball and Robert Willis University of Michigan, May 2005

A Growing Economic Literature Uses Happiness Data • Provocative findings—see Layard’s Happiness • Mostly focuses on the cross-section and the long -run trend. • Motivations of the researchers: – to study the welfare implications of non-traded goods – to study welfare implications in contexts where choice behavior is potentially inconsistent. Many economists are still skeptical of the use of subjective well-being data in economics. • Many other economists remain skeptical. • Theoretical status of happiness is unclear.

What is Happiness? a. Flow utility? b. The individual’s overall objective function? c. The part of the individual’s objective function that abstracts from the desire to do one’s duty? d. The individual’s objective function plus pleasure from memory? e. None of the above?

Introduction 1. Distinguishing utility and happiness as a matter of logic. 2. Why we care about utility and happiness. 3. Why the relationship between them can’t be simple (short version). 4. Our take on the relationship between utility and happiness.

1. “Affect” and “Utility” • Lifetime Utility = The extent to which people get what they want, where what they want is indicated by their choices. • Current Affect = How positive people’s feelings are at a given time.

2. Overall Welfare • Our theory starts from the presumption that both feelings and choices are likely to be useful indicators of overall welfare—that is, of what makes people better off in the sense relevant for policy.

3. The Easterlin Paradox and Hedonic Adaptation Taking both feelings and choices seriously runs into the difficulty that affect and utility seem to behave quite differently. • Easterlin Paradox: Measured utility trends strongly upwards, while measured affect has little trend. • Hedonic Adaptation: Utility is affected permanently by permanent changes in external circumstances, but the effects on affect seem shorter-lived.

4. The Relationship Between Happiness and Utility is Unresolved Existing work in Economics either A. B. ignores happiness data, e. g. : “Happiness is irrelevant to Economics” OR assumes happiness=flow utility: “Happiness is a sufficient statistic for utility. ”

The Middle Way In this paper, we steer a middle course between these two extremes: • Happiness ≠ Flow Utility, BUT • Happiness has a systematic relationship to utility.

Central Aim of this Research Program • Our central aim is to determine in detail the dynamic relationship between standard psychological concept of current affect and the standard economic concept of lifetime utility. • This is an open empirical question. Both affect and choice-based utility are welldefined, observable concepts. Previous work shows their relationship is not as simple as one might expect, but does not show exactly what that relationship is.

Significance Establishing any systematic relationship between affect and utility would 1. provide an important bridge between psychology and economics. 2. allow psychological data and theory to be used in economics in a way that is complementary to standard economic data and theory. 3. allow all the tools of economics to be brought to bear toward understanding happiness.

Sketch of our Integrated Theory of Utility and Happiness Experienced happiness is the sum of two components: • elation: short-run happiness that depends on recent news about lifetime utility • baseline mood: long-run happiness that is a subutility function (like health, entertainment, or nutrition. )

Why Happiness Matters for Economics (Our View) 1. short-run happiness in response to news can give important information about preferences. 2. long-run happiness is important for economic welfare in the same way as other composite goods such as health, entertainment, or nutrition.

Plan of Talk 1. Utility ≠ Happiness A. Measuring Happiness B. The Easterlin Paradox C. Measuring Utility D. Hedonic Adaptation 2. Utility and Happiness are Related A. Explaining the Easterlin Paradox and Hedonic Adaptation: Elation and Baseline Mood B. A Formal Model of Utility and Happiness 3. What the Relationship Means for Economics A. Extensions B. Axiomatics C. Implications for Empirical Research

Psychologists Reliably Measure Happiness, But What Is It? • Some economists think happiness can’t be measured well. This is just not true. Current happiness (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.

Measuring Current Happiness (‘Affect’). “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: 1. Much of the time during the past week, you felt you were happy. (Would you say yes or no)? 2. (Much of the time during the past week, ) you felt sad. (Would you say yes or no? ) 3. (Much of the time during the past week, ) you enjoyed life. (Would you say yes or no? ) 4. (Much of the time during the past week, ) you felt depressed. (Would you say yes or no? )”

The Validity of Self-Reported Happiness Correlated with • frequency of smiling. • others’ ratings of how happy someone is. • social rank. • high activity in the left pre-frontal cortex and low activity in the right pre-frontal cortex--which can also be induced by seeing pictures of a smiling baby and reduced by seeing pictures of a deformed baby.

Other Measures of Subjective Well. Being: Life Satisfaction On a scale from 1 to 10, how satisfied are you with your life?

World Values Survey Global Happiness Question "Taking all things together, would you say you are 1. Very happy 2. Quite happy 3. Not very happy 4. Not at all happy 9. Don’t Know [Do NOT READ OUT]”

Problems with these Alternative Measures of Subjective Well-Being • • Judging overall life-satisfaction or overall happiness in life is a complex cognitive task. Evidence on the sensitivity of of subjective well -being data to context indicates that respondents use shortcuts involving readily accessible information, such as – – How happy the respondent feels right now How happy the respondent thinks he or she should feel, given objective circumstances.

Advantages of Affect Measures (Current Happiness Measures) • By contrast, affect measures depend on much more accessible information: – How R feels right now. – How R felt the past week. • • Very little judgment is required. How R feels right now affects the overall lifesatisfaction or global happiness questions anyway. It is clearer to focus on that current happiness component directly. Then we know what we are getting.

Modified World Values Survey Question "Taking all things together, in the past week, would you say you have been: 1. Very happy 2. Quite happy 3. Not very happy 4. Not at all happy 9. Don’t Know [Do NOT READ OUT]”

Why Not ‘Happiness = Flow Utility’? • • • Assuming that current happiness is equal to flow utility immediately has many strong implications. In particular, a large amount of data on happiness exists, with many characteristics that do not match usual ideas about utility. Measured happiness 1. has no strong trend. 2. is strongly mean-reverting.

The Easterlin Paradox

Happiness is Not Improving Despite Other Positive Trends • The electronics revolution and the Internet have vastly expanded access to a rapidly growing quantity of culture and science. • Crime, teenage pregnancy and drug abuse trending downward. • Greater equality between races and sexes. • War on Terror better than Cold War. • Better medical care and greater longevity.

Life Expectancy

Utility has gone up, even though happiness hasn’t. • Utility is Defined by Revealed Preference • Would you want to go back to the way things used to be? – No ice cream – No computers or electronics – No modern music – Jim Crow, strong male dominance – Cold War – No modern medicine, dying young

The Ordinalist Revolution 1. The Ordinalist, or “revealed preference” revolution in Economics developed techniques for measuring individual welfare based on choice data alone. 2. This clearly defines utility as a distinct concept from happiness. • • Utility is the extent to which people get what they want. Happiness is how people feel.

Hedonic Adaptation (Mean Reversion of Affect) Cross-sectional evidence of hedonic adaptation for • incarceration • loss of the use of limbs • serious burns • death of a spouse • winning the lottery – winning £ 10, 000 raises affect by six times as much in the first year as £ 10, 000 per year in additional income.

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 economists. – Often refers to tending to do something more if you have done it in the past (effect on marginal utility) – To the extent it refers to past actions lowering utility from similar actions now, shifting toward the action when near death might be evidence

Hedonic Adaptation and the Easterlin Paradox • Brickman and Campbell (1971) call the implications of hedonic adaptation for the trend in affect the hedonic treadmill. • Mechanically, if the impact of good events on happiness is transitory, that will help account for the Easterlin Paradox, so the two phenomenon are related.

Experience Data Show Even Stronger Hedonic Adaptation • • Data on felt happiness from experience sampling reverts to its previous level even more completely than life satisfaction and global happiness assessments (Kahneman and Schwartz, unpublished work). Why? Life satisfaction and global happiness assessments incorporate – – an element of autobiography people’s ideas about how they “should” feel

Hypotheses to Explain the Easterlin Paradox and Hedonic Adaptation a. Objectives other than affect in the objective function. b. Persistent mistakes or inconsistent preferences. c. Non-marketed goods, such as positive social interaction and social rivalry. d. Elation and dismay: a component of affect that has to do with the cognitive processing of recent news.

News and Happiness • The relationship between circumstances and happiness is weak in the long run, BUT • No one disputes that in the short run happiness responds in an intuitive way to news about lifetime utility. • Thus, we argue that an important component of happiness is due to recent news about lifetime utility.

‘Elation’ and ‘Dismay’ • ‘elation’ = the component of happiness due to recent news about lifetime utility. • ‘dismay’ = -elation

Elation and Hedonic Adaptation • • If expectations are rational, standard results about rational expectations imply that elation will be strongly mean reverting. Intuitively, – News doesn’t stay news for very long. – The initial burst of elation dissipates once the full import of news is emotionally and cognitively processed.

Integrated Theory of Utility and Happiness affect = baseline mood + elation.

Household Production Since Gary Becker’s pioneering work, much of the activity of a household outside of paid work has been reconceived as household production of goods.

‘Baseline Mood’ • baseline mood = M(Kt, Xt) • Xt = vector of control variables: time use, spending pattern, portfolio choice, etc. • Kt = vector of – state variables encoding every aspect of the past that matters for utility: wealth, weight, habits, level of fatigue, one’s spouse being alive, etc. – variables exogenous to the individual: weather, state of macroeconomy, consumption patterns of others in society, etc. CONTINUED ON NEXT SLIDE →

Baseline Mood and Flow Utility • flow utility = U(Kt, Xt, M(Kt, Xt)) • We think of baseline mood M(Kt, Xt) as the component of happiness produced by a household production function. • A good analogy is to health. Like health, baseline mood – can be measured independently of Kt and Xt – is only one argument of the flow utility function – depends on different things than flow utility does on the same things with different weights) – has a complex household production function (or

What does Baseline Mood Depend on? 1. Any persistent aspect of happiness is part of baseline mood. Genes are the biggest factor. Also, there is some evidence that each of the following has a persistent effect on happiness: a. Prozac b. sleep c. exercise d. good eating habits e. social rank 1. + pleasantness of one’s current activity

Do People Know the Production Function for Baseline Mood? 1. Just as people don’t know the true production function for health, they may not know the true production function for baseline mood. 2. Lack of understanding of the dynamics of the elation mechanism could make it difficult for individuals to parcel out the determinants of baseline mood. 3. The discovery and dissemination of facts about the determinants of baseline mood could have large positive welfare effects 4. A big deal if the share of the money and time budget devoted to baseline mood trends up.

Linking Elation to the Bellman Equation (Stoch. Dyn. Prog. ) • Vt(Kt): maximum value of lifetime utility attainable vector Kt exogenous variables and state variables inherited from the past. • β: utility discount factor, • Et: rational expectation conditional on time t information • εt+1: vector of random variables • Γ: law of motion of Kt

The Bellman Equation, the Bellman Innovation and Elation

Extensions to the Theory of Elation • • If expectations are not rational, it may be possible to manipulate elation in ways that raise its mean. Elation may respond more to news about whether one’s choices worked out than to news about things beyond one’s control. – This would make it possible to manipulate elation by labeling good events as due to one’s efforts, while bad events were beyond one’s control.

The Evolutionary Psychology of Elation and Dismay • Functionally, elation and dismay motivate cognitive processing—much like curiosity. – Elation: after good news, it pays to • think what you did right, so you can do it again • think how to take advantage of the new opportunities – Dismay: after bad news, it pays to • think what you did wrong, so you can avoid doing it again • think how to mitigate the harm of the bad news – Curiosity: after news that is neither clearly good nor bad, it pays to learn more for the sake of option value • Economic implications of this functional role of elation: such directed information acquisition could affect probability assessments in systematic ways.

What if Elation is in the Felicity Function? So far, a. flow utility = U(Kt, Xt, M(Kt, Xt)) What if b. flow utility = U(Kt, Xt, M(Kt, Xt)) +b 0ιt+b 1ιt-1+b 2ιt-2+… c. flow utility = U(Kt, Xt, M(Kt, Xt), f(ιt, ιt-1, ιt-2, …)) )

Elation and Prospect Theory • Elation theory yields prospect theory very naturally if a. Elation is roughly proportional to the rate of cognitive processing of news. b. Bad news requires more processing than good news. c. Within bad or good news, the total amount of processing needed is roughly proportional to the magnitude of the news. d. It takes longer to process a big chunk of news than a small chunk of news. e. People use expected elation (perhaps with distorted probability weights) as a heuristic for decisions in the face of risk. (e. is where the irrationality comes in. )

Reprise: Integrated Theory of Utility and Affect Happiness is the sum of two components: • • elation: short-run happiness that depends on recent news about lifetime utility baseline mood: long-run happiness that is a subutility function (like health, entertainment, or nutrition. ) – In principle, all of the usual techniques of price theory apply to baseline mood, but – The dynamics of elation make it harder for people to learn the true household production function for baseline mood.

Integrating Happiness into Mainstream Economics • Happiness needs to be integrated in a way that respects the canons of Economics. • Two key dimensions for integrating happiness into economics: – First, the short-run responses of happiness to news provide important information about preferences. – Second, long-run happiness is important in its own right.

The Integrated Theory of Utility and Affect: Axioms 1. Happiness and news: other things being equal, a. the agent will feel happier if current expectations are of a preferred future. b. the agent will feel less happy if past expectations were of a preferred future.

The Integrated Theory of Utility and Affect: Axioms 2. Preference for happiness. Between two fully anticipated life courses with the same outward circumstances, the agent prefers the one in which he or she feels happier. 3. Social Rivalry. The empirical possibility of social rivalry is represented by the absence of an axiom that preferences depend only on the agent’s own situation.

Preference for Happiness: Discussion The preference for happiness shows up in both household and firm behavior: • Purchases of therapy, Prozac, self-help books, magazines featuring “happiness. ” • Advertising that tries to suggest that a product will make one feel happy.

Preference for Happiness: Discussion The preference for happiness axiom is a relatively cautious assumption. • It allows for the possibility that people have other pragmatic concerns beyond the extent to which those concerns show up directly in felt happiness. For example, they may care about satisfying basic drives, the happiness of family members, health, ethical concerns and other goals.

Preference for Happiness: Discussion • One could make stronger assumptions but the axioms above are enough for our approach. • Cautious assumptions have the advantage of being more likely to gain broad acceptance within Economics.

An Alternative, Stronger, Assumption • Sufficient Statistic Hypothesis: some combination of current and expected future happiness is a sufficient statistic for the agent’s preferences. • The Preference for Happiness Axiom allows for the Sufficient Statistic Hypothesis as a special case that, in principle, is empirically testable. -Happiness can be measured by feelings. -Preferences can be measured by choices whenever people are well informed and have time to make a careful choice.

Implications of the Theory for Happiness Empirics • We need to focus on the dynamics of happiness. • It is important to have happiness data with many observations in the time dimension. • It is the high frequency movements of happiness that tell us most about the utility function. • Low frequency movements in happiness tell us only about a sub-utility function.

Proposal • I am proposing the collection and analysis of monthly data on subjective well-being in the Survey of Consumer Attitudes and on a monthly Japanese survey. • This will allow us to look at the dynamic response of happiness to national and international events and the interaction between happiness and consumer confidence. • The monthly frequency (with date and time stamps within the month) is important because hedonic adaptation may be complete within a few months for a typical items of news.

Significance 1. Even if economic progress continues unabated over the next 50 years in the U. S. 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.

Significance 2. To the extent there is a tradeoff between subjective well-being and other values, the increases in income and wealth that accompany economic progress are likely to make improvements in subjective wellbeing increasingly important for welfare compared to further improvements in other areas.

Significance 3. Economists are increasingly using subjective well-being data to address economic and public policy issues that involve non-marketed goods or inconsistent preferences. Identifying the implications of subjective well-being data for economic issues requires attention to the details of the mapping between subjective well being data and standard economic concepts.

Significance 4. Given an adequate understanding of the mapping between subjective well-being data and standard economic concepts, the use of subjective well-being data has the potential to be especially important in the economics of aging, since many of the most important goods for retired people are non-marketed goods. (Consider, for example, health, marital and family relationships, sense of purpose, and quality of leisure time pursuits. )

Significance 5. In the coming decades, advances in subjective well-being at work have the potential to alter people’s relationship to work in a way that significantly raise the average retirement age, with important implications for Social Security budget balance.

Significance 6. 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 thorough-going resolution of this paradox is essential for effective integration of subjective well-being data into Economics.

The Elation Theory and the Brain • Wolfram Schultz, Peter Dayan and P. Read Montague, “A Neural Substrate of Prediction and Reward, ” Science, March 14, 1997, * write: “It appears therefore, that learning is driven by deviations or ‘errors’ between the predicted time and amount of rewards and the actual experienced times and magnitudes. ” (p. 1593) *Thanks to Tilman Borgers for this reference.

Caption to Shultz, Dayan, and Montague (1997), Figure 1 “Fig. 1. Changes in dopamine neurons’ output code for an error in the prediction of appetitive events. (Top) Before learning, a drop of appetitive juice occurs in the absence of prediction—hence a positive error in the prediction of reward. The dopamine neuron is activated by this unpredicted occurrence of juice.

Caption to Shultz, Dayan, and Montague (1997), Figure 1 (cont. ) (Middle) After learning, the conditioned stimulus predicts reward, and the reward occurs according to the prediction—hence no error in the prediction of reward. The dopamine neuron is activated by the reward-predicting stimulus but fails to be activated by the predicted reward (right).

Caption to Shultz, Dayan, and Montague (1997), Figure 1 (cont. ) (Bottom) After learning, the conditioned stimulus predicts a reward, but the reward fails to occur because of a mistake in the behavioral response of the monkey. The activity of the dopamine neuron is depressed exactly at the time when the reward would have occurred. ”

The Elation Theory and the Brain (2) “The TD [Temporal Difference] algorithm is particularly well suited to understanding the functional role played by the dopamine signal in terms of the information it constructs and broadcasts. ” (p. 1594) The Temporal Difference Algorithm is quite explicit in defining the Temporal Difference error in a way almost identical to our definition of the Bellman innovation.

The Elation Theory and the Brain (3) “As the rat explores the maze, its predictions become more accurate. The predictions are considered ‘correct’ once the average prediction error δ(t) is 0. At this point, fluctuations in dopaminergic activity represent an important ‘economic evaluation’ that is broadcast to target structures: Greater than baseline dopamine activity means the action performed is ‘better than expected’ and less than baseline means ‘worse than expected. ’ (p. 1596)

The Elation Theory and the Brain (4) Hence, dopamine responses provide the information to implement a simple behavioral strategy—take [or learn to take] actions correlated with increased dopamine activity and avoid actions correlated with decreases in dopamine activity. ” (p. 1596 immediately following)

The Elation Theory and the Brain (5) Predictions: 1. Blocking: no effect of signals that have no additional marginal predictive power. 2. Secondary conditioning: Conditioned stimuli can condition stimuli that predict them. 3. The earliest consistent cue gets the dopamine reaction.

The Ordinalist Revolution 2. The Ordinalist revolution also 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?

Theory 11. How does flow utility (and therefore the overall objective function) depend on the components of affect? a. One possibility is that affect is an epiphenomenon—that is, affect depends on news about the overall objective function, but the overall objective function does not depend on affect. b. To the extent that, instead, flow utility depends on baseline mood, baseline mood simply acts like one more good generated by a household production function and can be handled in standard ways. c. A surprising theoretical result is that if elation enters flow utility in an additive linear way, choice behavior will be totally unaffected. d. The key aspects of prospect theory can be generated parsimoniously by a nonlinear dependence of flow utility on elation.