Happiness, Economics and Politics: Towards a Multi-Disciplinary Approach

Happiness, Economics and Politics Happiness, Economics and Politics Towards a Multi-Disciplinary Approach Edited by ...

Author: Amitava Krishna Dutt | Benjamin Radcliff

93 downloads 1556 Views 1MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Happiness, Economics and Politics

Happiness, Economics and Politics Towards a Multi-Disciplinary Approach

Edited by

Amitava Krishna Dutt and Benjamin Radcliff University of Notre Dame, USA

Edward Elgar Cheltenham, UK • Northampton, MA, USA

© Amitava Krishna Dutt and Benjamin Radcliff 2009 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA

A catalogue record for this book is available from the British Library Library of Congress Control Number: 2009936225

ISBN 978 1 84844 093 7 Printed and bound by MPG Books Group, UK

Contents List of contributors Preface

vii ix

Introduction: happiness, economics and politics Amitava Krishna Dutt and Benjamin Radcliff PART I

HAPPINESS

1 The history of happiness and contemporary happiness studies Darrin M. McMahon 2 On the measurement and mismeasurement of happiness: contemporary theories and methodological directions Anthony D. Ong 3 How do we assess how happy we are? Tenets, implications and tenability of three theories Ruut Veenhoven 4 Happiness and domain satisfaction: new directions for the economics of happiness Richard A. Easterlin and Onnicha Sawangfa PART II

1

25

33

45

70

HAPPINESS AND ECONOMICS

5 Happiness when temptation overwhelms willpower Alois Stutzer 6 Happiness and the relative consumption hypothesis Amitava Krishna Dutt 7 The Easterlin Paradox revisited Robert H. Frank 8 Does inequality matter to individual welfare? An initial exploration based on happiness surveys from Latin America Carol Graham and Andrew Felton 9 Perceptions of discrimination, effort to obtain psychological balance and relative wages: can we infer a happiness gradient? Arthur Goldsmith

v

97 127 151

158

202

vi

PART III

Happiness, economics and politics

HAPPINESS AND POLITICS

10 Politics and happiness: an empirical ledger Alexander C. Pacek 11 Democracy and happiness: what causes what? Ronald Inglehart 12 The causal link between happiness and democratic welfare regimes Charlotte Ridge, Tom Rice and Matthew Cherry 13 Labor organization and the quality of life in the American states Suzanne M. Coshow and Benjamin Radcliff PART IV

231 256

271

285

WHAT IS TO BE DONE?

14 Should national happiness be maximized? Bruno S. Frey and Alois Stutzer 15 Change your actions, not your circumstances: an experimental test of the Sustainable Happiness Model Kennon M. Sheldon and Sonja Lyubomirsky 16 What is to be done? Toward a ‘happier’ world Amitava Krishna Dutt and Benjamin Radcliff

301

Index

351

324 343

Contributors Matthew Cherry, Department of Political Science, University of Iowa, Iowa City, Iowa, USA. Suzanne M. Coshow, Department of Sociology, Indiana University-South Bend, South Bend, Indiana, USA. Amitava Krishna Dutt, Department of Economics and Policy Studies, University of Notre Dame, Indiana, USA. Richard A. Easterlin, Department of Economics, University of Southern California, Los Angeles, California, USA. Andrew Felton, Federal Deposit Insurance Corporation, Washington, DC, USA. Robert H. Frank, Johnson Graduate School of Management, Cornell University, Ithaca, New York, USA. Bruno S. Frey, Department of Economics, University of Zurich, Switzerland. Arthur Goldsmith, Department of Economics, Washington, DC and Lee University, Lexington, Virginia, USA. Carol Graham, The Brookings Institution, Washington, DC and School of Public Policy, University of Maryland, College Park, Maryland, USA. Ronald Inglehart, Department of Political Science, University of Michigan, Ann Arbor, Michigan, USA. Sonja Lyubomirsky, Department of Psychology, University of CaliforniaRiverside, California, USA. Darrin M. McMahon, Department of History, Florida State University, Tallahassee, Florida, USA. Anthony D. Ong, Department of Human Development, Cornell University, Ithaca, New York, USA. Alexander C. Pacek, Department of Political Science, Texas A&M University, College Station, Texas, USA. vii

viii


Benjamin Radcliff, Department of Political Science, University of Notre Dame, Indiana, USA. Tom Rice, Department of Political Science, University of Iowa, Iowa City, Iowa, USA. Charlotte Ridge, Department of Political Science, University of Iowa, Iowa City, Iowa, USA. Onnicha Sawangfa, Department of Economics, University of Southern California, Los Angeles, California, USA. Kennon M. Sheldon, Department of Psychology, University of MissouriColumbia, Columbia, Missouri, USA. Alois Stutzer, Department of Business and Economics, University of Basel, Basel, Switzerland. Ruut Veenhoven, Department of Sociology, Erasmus University, Rotterdam, the Netherlands.

Preface This volume grew out of a conference entitled ‘New Directions in the Study of Happiness: US and International Perspectives’, held at the University of Notre Dame on 22–24 October 2006. Earlier drafts of most of the chapters in this volume were presented at this conference. We are very grateful to the participants of the conference, especially those whose work is included in this volume, for their contribution, patience and encouragement. We are also deeply indebted to the Department of Economics and Policy Studies and the College of Arts and Letters, and to the Program in American Democracy, at the University of Notre Dame, for their generous financial and logistic support. We would especially like to thank Harriet Baldwin for her help with the organization of the conference. We also gratefully acknowledge the help we received from the staff at Edward Elgar, especially from Joanne Betteridge and Kate Pearce. The conference, especially because of its multidisciplinary nature, was very stimulating and enjoyable to us (and, as we could gather from their comments, the other participants), and we are very pleased to be able to prepare this volume for wider dissemination. This is doubly so since we like to believe, as our introduction makes clear, that the individual chapters add up to rather more than the sum of their parts, providing as they do a good summary of the state of the emerging field of the scientific study of happiness. We hope that this volume will also be stimulating and enjoyable – and, dare we say, increase the happiness – of a wider audience. Amitava Krishna Dutt and Benjamin Radcliff

ix

Introduction: happiness, economics and politics Amitava Krishna Dutt and Benjamin Radcliff The nature and meaning of happiness has been discussed over the centuries by religious figures, philosophers and, more recently, by social scientists. Moreover, it seems indisputable that happiness has always been the (or at least, a) goal of, or driving force for, most human beings. In recent years a new field of inquiry, the so-called ‘science of happiness’, has emerged. This field tries to examine the idea of happiness using quantifiable and measurable concepts and to analyse its determinants using the empirical and theoretical tools of the social and biological sciences. The ‘scientific’ study of happiness, however, has been conducted by scholars from different disciplines in different and largely separate ways. Psychologists have turned from an emphasis on problems such as depression to a consideration of positive affect, and attempted to measure a number of notions of happiness using survey and experimental data, as well as brain signals, and have focused on different kinds of mental processes, personality traits and environmental factors to understand the determinants of happiness. Economists have focused on the effects of income and consumption on happiness and well-being, debating, essentially, whether money buys happiness. They have also approached the study of economic development using the concepts of economic growth and improvements in well-being. Sociologists have emphasized interactions within communities, families and groups as well as the broader effects of culture on happiness. Political scientists have focused on how different systems of government – such as democracies – and different types of governmental policies, such as the size of the welfare state, affect happiness. Philosophers have speculated on the meaning of happiness – for instance, whether it involves the maximization of pleasure, the minimization of pain, or maximization of the difference between pleasure and pain, or whether it refers to human flourishing. There have, of course, been scholars who have crossed the normal boundaries of their own disciplines to incorporate factors from outside. For instance, some economists – such as Veblen (1899) and, more recently, Schor (1998) and Frank (1999) – have incorporated ideas about 1

2


status, more traditionally studied by sociologists, in analysing the relation between consumption and happiness, while some political scientists – such as Lane (2000) and Putnam (2000) – have addressed issues related to friendship and community in understanding the determinants of happiness. But the general tendency, given the disciplinary structure of the academy, has been to remain largely within a particular discipline. As students of happiness and well-being, we believe that the study of happiness is best done from a multidisciplinary perspective, and that to make progress in the subject we need greater communication and interaction among different disciplines. We therefore decided to organize a conference with scholars from different disciplines, including history, psychology, philosophy, sociology, economics and political science. Coming ourselves from economics and from political science, we are especially interested in how concepts and issues in economics and political science can be combined and cross-fertilized to deepen our understanding of happiness. Therefore, although we sought representation from a variety of disciplines, we invited more scholars from our two disciplines than from the others. This book contains versions of most of the papers presented at the conference, as well as our own reflections based on our own work and the proceedings of the conference. The book is divided into four parts, the first on happiness in general, the second on economics and happiness, the third on politics and happiness, and a final one on what is to be done. The following sections of this introduction discuss some of the key questions related to each part and how the chapters in this volume relate to these questions. Thus, the next section provides a general introduction to the history, conceptualization, theories and empirics of happiness. The following two sections then examine, in turn, the relationships between happiness, and economics, and politics. The subsequent section argues that an interdisciplinary perspective is necessary for obtaining a better understanding of happiness and its determinants, and focuses on the interaction between economic and political factors in explaining happiness. The final section discusses some implications of the analysis for what can and should be done, individually and collectively, for improving human well-being and making people happier.

I.1

HAPPINESS

Like most genuinely great ideas, happiness has successfully resisted attempts to be reduced to any single meaning or definition. We begin, thus, with an appraisal of what happiness is, and how best we might conceive of

Introduction

3

it not merely as philosophers, but also as scholars of the empirical, observable world. The volume thus begins with Darrin McMahon’s insightful discussion of Western attitudes to the meaning of happiness and the possibility of its attainment, a discussion which he has carried out at much greater length in his book on the history of happiness (McMahon, 2006). Early on, happiness was considered a matter of luck, or something that was not achievable in the here and now, at least by more than a handful of people. He argues that the notion of happiness being our being’s ends and goal is an Enlightenment creed which has crowded out other ways of looking at the world and human purpose in it. McMahon sees the triumph of this Enlightenment creed as a larger dynamic which explains the wide contemporary appeal of happiness studies. But, he argues, happiness studies would do well to be cognizant of this history of happiness. Thus, McMahon ‘enter[s] a humble plea as a historian and scholar of the humanities for a certain humility as we approach our subject, pointing out the un-nerving tendency of happiness to frustrate and circumvent those who would try to grasp it in pursuit’. The need for that humility is readily seen in any survey of the meaning of happiness. We see a concern with happiness in the earliest philosophers and religions in both the East and the West. For instance, the ancient belief systems of India, which are now collectively referred to as Hinduism, took the view that ‘true’ happiness requires the realization of a person’s oneness with Brahman, the universal soul, and the liberation from worldly desires. Buddhism also stressed the importance of overcoming desire and of avoiding extremes of pleasure and austerity, of following ‘the middle path’. The major Greek philosophers did not especially value everyday sensual pleasures, with Plato favoring the controlling of appetites rather than being enslaved by them, and Aristotle valuing fulfilling activities which led to eudaimonia or flourishing, exemplified at the highest level by intellectual contemplation. Christianity, Judaism and Islam regard happiness as the goal of religious life and beliefs to be attempted through knowing and serving God. However, true happiness, they imply, can only be achieved in afterlife, in the form of union with God. Later Christian philosophers, like St Thomas Aquinas, tried to synthesize the Aristotelian view about happiness in this world and the early Christian view that true happiness can only be achieved in afterlife by suggesting that happiness in this life, while imperfect, can approximate the true happiness of afterlife through virtuous living. With the weakening of the hold of religion in everyday life and general secularization that accompanied the Enlightenment and the Industrial Revolution, it came to be increasingly accepted that happiness is something that all people could at least aspire to in their lives on earth. However, a

4


variety of meanings was attached to happiness by philosophers. Some, such as the Utilitarians, emphasized that increasing happiness entailed that we take seriously the seemingly mundane observation that we should increase pleasure and decrease pain. Others, such as Schopenhauer, were closer to Eastern traditions by focusing mostly on minimizing pain, while still others, such as Nietzsche – and creative writers such as Joyce – emphasized the maximization of pleasure, the pursuit of ‘passionate joy’, whatever the potential price to be paid. Many social scientists, instead of trying to define happiness in these ways, rely on asking people how happy they actually feel. There is a large amount of survey data now available which asks people to report on, say, on a 10-point scale, how happy or satisfied they are with their lives, all things considered. This approach in effect allows subjects to decide what they mean by happiness, and lets them decide how happy they are given their own definition. This approach to happiness is often referred to as subjective well-being, in an effort to apply a more precise and less emotionally laden term than happiness. While some have argued that such measures are valid and reliable, and that levels of subjective well-being and how they change over time and place can tell us about the objective conditions that make people better off, others remain skeptical, because of cultural differences across people and because people are known to adapt to their environments. Thus, different cultures can interpret happiness and its desirability differently, and the poor can get used to their destitution and feel reasonably happy. Despite the growing popularity of measures of subjective well-being, social scientists have continued using less subjective measures of happiness and well-being. Thus, psychologists distinguish between different levels of happiness, a first referring to (transient) emotions and feelings of joy and pleasure, a second to people’s judgments about their feelings over a long period of time (that is, subjective well-being), and a third to whether one fulfills one’s true potential, or whether one flourishes, and is closer to the Greek notion of the good life or eudaimonia (see, for instance, Nettles, 2005). In his contribution to this volume, Anthony Ong argues that developments in the psychology of happiness and well-being suggest that wellbeing cannot be well represented by a single dimension, so that researchers have to examine both eudaimonic and hedonic aspects of happiness and well-being to obtain a more complete understanding of positive human health. Ong points out that current psychological research measuring wellbeing takes two major alternative paths. One measures subjective wellbeing, which involves a person’s cognitive judgment of life satisfaction, and an emotional aspect consisting of independent positive and negative affect components. The other derives from eudaimonic expressions of

Introduction

5

virtue, and the striving for the realization of one’s true potential, and is measured by the concept of psychological well-being, involving different aspects of human actualization, such as autonomy, personal growth, life purpose and social connectedness. Ong then discusses methods with which it is possible to examine the relationship between these two aspects of well-being, and how this varies within persons (with most of the current research examining variations across persons for a single aspect). He examines methods involving longitudinal panel designs and intensive bursts designs, as well as less widely used growth curve modeling and dynamic systems analysis using differential equations, and explores how these methods can deal with rounded and irregular structures (rather than rectilinear systems) necessary for understanding the complexity of the human brain. While this work, like McMahon’s, tells a cautionary tale in how readily we are willing to accept the conventional, mainstream use of survey-based indicators of happiness or life satisfaction, much of the rest of the chapters in this volume, reflecting the norms of the literature, do rely upon fairly straightforward measures of well-being. Economists, while strongly wedded to the notion of utility and its maximization by economic agents as the method of explanation behavior and evaluating states in terms of individual utility (which may be measured with subjective well-being surveys), often rely more heavily on real income and production as measured by the national accounts because of their relative ease of measurement, assuming that more (in terms of products and income) is better (in terms of utility or happiness). Moreover, recognizing that people adapt to their states in terms of subjective feelings (like the poor person getting used to poverty), some economists (see especially Sen, 1999) favor less subjective concepts such as functionings (whether people are able to achieve particular good results in terms of things – such as good health, adequate nutrition and even adequate self respect – that are considered valuable) or capabilities (whether people have the opportunity to achieve these functions, rather than whether they actually achieve them). Turning from the meaning of happiness to theories of happiness, in the social scientific literature on subjective well-being, three broad kinds of theories have emerged for predicting how happiness varies across countries or individuals. Using the terminology suggested by Veenhoven and Ehrhardt (1995), and further elaborated by Ruut Veenhoven in his contribution to this volume, one can think of theories focusing on (1) traits, (2) social comparisons or (3) satisfaction of needs. Trait theories contend that happiness is a relatively fixed personality trait, similar to other aspects of personality, resulting from a combination of genetic inheritance, cultural socialization and other early life experiences. Individuals, then, tend toward a basic level of happiness,

6


in the same way that they tend toward a certain level of extroversion or optimism. Psychologists often think thus in terms of ‘set-points’ to which individuals are prone. Life events may move them to either more or less happiness, but such changes tend to be temporary, with individuals eventually returning to something like their original set points over time. The same logic applied to the national level suggests that people in some countries just naturally tend to be happier than those in others, as a part of the national character – that is, a national set-point, reflecting a kind of national personality. Comparison theories, often favored by economists such as Easterlin (1974), are predicated upon the notion that individuals appraise the quality of their lives in relative rather than absolute terms. In it simplest incarnation, people look to society’s ‘consumption norm’, and evaluate their lives relative to that norm, such that those with a higher level of consumption view themselves as more satisfied, and those below, less so. The obvious conclusion is that, presuming the consumption norm is society’s median, the mean level of happiness will everywhere be the same, and will not change significantly over time (since, of course, raising the social median level of consumption cannot change the fact that there are always as many people above it as below it). There are more complicated and sophisticated notions of relativity (as reviewed by Diener et al., 1999), but all depend upon the fundamental notion of comparing one’s position (typically in terms of income, social status or other measures of consumption) against some external standard. Both trait and comparison theory, then, stand in contrast to what Veenhoven has rightly characterized as the ‘common sense’ view that satisfaction with life is determined by the amount of needs that one’s life circumstances allow to be meet. Thus, following the conventional idea of a hierarchy of needs (Maslow, 1954), the more needs that are met, the more rewarding and fulfilling one is likely to find life. Veenhoven’s contribution to this volume continues his work in evaluating trait and comparison theories, and in articulating and defending the needs-based theory. As his chapter notes, the stakes involved in which theory we find most compelling are enormous, in that trait and comparison-based theories tend toward the view that national levels of happiness simply cannot be lastingly improved, since if happiness is a trait, levels will return to their set-points whatever interventions we make in the world to make life more secure or otherwise enjoyable, and if happiness is relative, these same kind of changes will merely raise the consumption norm, and thus leave the overall level of happiness unchanged. In sum, how we theorize about happiness determines whether it is sensible to even try to improve living conditions across the world, given that both trait and comparison theory suggest that doing

Introduction

7

so will not increase the amount of actual happiness in the world. We return to this issue in the conclusion of this book. The set-point notion of happiness, of course, need not lend itself to the conclusions Veenhoven sees when applied at the purely individual level (and hence even over aggregates). Thus, one can posit that people do have set-points, perhaps even as a hardwired, genetic aspect of a personality, while admitting that individuals might be able to reprogram themselves for greater happiness by using their power of agency to play against a reversion to set-point (that is, ‘adaptation’). This may be of scant comfort to those of us, like Veenhoven, concerned with the possibility of using public policy to improve objective living conditions, and thus, subjective appraisal of life (given that a country’s ‘national character’ as expressed by its cultural norms can hardly be consciously reprogrammed), but it does open the door to specific strategies individuals may use in their day-to-day lives to make life more satisfying. It is this very possibility that has given rise to the ‘positive psychology’ movement. In their contribution to this volume, Kennon Sheldon and Sonja Lyubomirsky provide an overview of one of the most promising developments in that field in the form of the ‘Sustainable Happiness Model’. They grant that one’s happiness is largely determined by genetics and other circumstantial factors beyond the individual’s control, but also argue that one’s activities also play a role, such that the set-point should be construed ‘as a range’ so that it is possible ‘to construct one’s life in such a way that one stays in the upper half of one’s set range, finding ways to remain at a level of happiness that is higher than one’s genetics alone would dictate’. Proponents of the comparison theory, also, do not argue that happiness levels cannot change over time and space. Even if for some things, like income and consumption, people obtain happiness mainly from what they have in comparison to others, rather than from their absolute levels, it does not follow that all things on which individual happiness depends involve such comparisons. For instance, individual happiness may depend strongly on absolute levels of leisure, time spent with family and friends, good health and a clean environment. Indeed, as we shall discuss later, they argue that individuals actually devote too much effort in obtaining things involving comparisons (to get ahead of others) and too little on things that do not involve comparisons. A final theoretical distinction in the study of happiness relevant to our concerns is the distinction between global evaluations of life in general, which is the principal focus of most research, and the evaluations of happiness with particular aspects, or ‘domains’, of life. In this latter interpretation happiness may be defined as the feeling one gets by being with one’s family, having better health, having a rewarding job or having more

8


money. This observation also suggests empirical analysis of what actual conditions determine overall happiness. There is considerable literature examining the determinants of happiness, more specifically, exploring how happiness depends on personal factors such as income, health conditions, whether one is employed, conditions of work, time spent with family and friends, status, religiosity, personality traits, age, marital status, gender and race, overall economic conditions such as inflation, unemployment and income distribution, overall political conditions such as nature of the government, trust in government and political instability, and other general factors like the weather (see, for instance, Argyle, 1987, 1999; Frey and Stutzer, 2002; Layard, 2005). In their contribution to this volume, Richard Easterlin and Onnichi Sawangfa bring these two strands of the literature together, and examine how trends in overall indicators of happiness relate to satisfaction people report from the widely discussed domains of finance, family life, work and health. They find that the importance of any given domain depends on actual life circumstances regarding socioeconomic status (as measured by years of schooling), time (year dummies), age or birth cohort, and that no individual domain is invariably the key determinant of overall happiness. This analysis also enables them to explain a number of empirical trends that have been found in the happiness data for the USA, that is, the positive cross-sectional relation between happiness and socio-economic status, the horizontal time series trend in happiness, the hill pattern of life cycle happiness and decline in happiness across generations.

I.2

HAPPINESS AND ECONOMICS

Of obvious relevance to economists is the question whether happiness increases when people consume more goods and services. It is widely believed by people that more is better and that money buys happiness. This belief is reflected in the theory used by mainstream economists, which assumes that individuals maximize utility, and that utility increases with the individual’s consumption or income without bound (although there may be diminishing marginal utility). It is also reflected in the way mainstream economists usually evaluate states of the economy using the notion of efficiency. Although efficiency is formally described as a situation in which no one can be made better off without making someone else worse off, in actual applications the concept is usually taken to require that there is production efficiency, that is, the economy cannot produce any more of any good or service without reducing the production of something else: everything has an opportunity cost. Economists also typically equate economic development

Introduction

9

with the growth of per capita income or product, and even when they take equity or distributional issues into account, they usually measure inequality and poverty in terms of real income or consumption. But is more really better as judged by the people who have more? Starting with Easterlin’s (1973) pioneering contribution, some economists have found that happiness in the sense of subjective well-being doesn’t seem to rise systematically with income (see Easterlin, 2001; Frey and Stutzer, 2002; Layard, 2005). Across countries most studies find that, at least beyond a certain level of income, happiness does not rise significantly with income. Time series data for rich countries, such as the USA and Japan, suggest that happiness does not increase over time despite significant increases in income. Individuals in rich countries who experience significant increases in time do not report significant increases in happiness. The only kind of empirical analysis that unequivocally supports the notion that more is better is cross-sectional analysis of individual countries: people with higher levels of income tend to be happier than those who are poorer. What explains these empirical findings, which contradict some of the basic assumptions of many economists? Suitable answers have to explain both why people increase their income and consumption and why they are not happier as a result, a phenomenon that has been called the Easterlin Paradox. Economists have come up with a number of plausible explanations which can be classified into two groups. One group focuses on the individual decision maker and argues that for a variety of reasons individuals make poor decisions which make them consume more without becoming happier as a result. The other group examines the individual decision maker in a social context where, although in isolation they could make decisions which make them happier, such decisions are influenced by the behavior of others, and the actions of all individuals taken together result in conditions which make individuals not happier. In the first group are explanations which focus on psychological processes involved in decision making and the feelings individuals experience after they make their decisions. Individuals make decisions to do certain things – like consuming more – but they make decisions which make them no happier because they either do not know what makes them happier, or because they do not have the willpower to do what makes them happier. In his contribution to this volume, Alois Stutzer uses the tools of happiness studies to examine whether individuals act ‘rationally’ in their self-interest, or whether they sometimes yield to temptation which overwhelms their willpower, resulting in outcomes that are not optimal for them. He examines this question by reviewing analysis and evidence on smoking and television viewing, and discusses his own recent work on obesity. He finds that

10


individuals frequently report to being unhappy with their decisions about smoking, television viewing and eating, and their consequences, and this loss in happiness is positively related to their own judgments about their lack of self-control. Thus, when doing something leads to immediate (even minor) gratification, and the costs – even if well understood and significant – come later, individuals perform these activities. These findings go against one of the standard assumptions of neoclassical economics – that what people choose is the best for them because they choose ‘rationally’. In the second group are explanations which state that people, when interacting with others, make decisions which do not make them happier because they are influenced by the behavior of others, and because the feelings they experience after making their decisions are affected by what others do. For instance, following the work of Veblen (1899), Schor (1998) and Frank (1999) many others have argued that individuals try to consume more to increase their status in society, but if all of them do so, no one is better off. In fact, they may be worse off if their attempts to increase consumption result in their working more hours at the expense of leisure and time spent with friends and family, and their incurring more debt which makes them financially insecure. Increases in consumption over time can also result in higher levels of aspirations and social conventions about needs, so that there may be no increases in happiness (see Easterlin, 2001). Moreover, by increasing sales promoting expenditures, firms may influence consumers into buying goods which ultimately raise their expectations and do not make them happier. Such explanations are examples of what, following Veenhoven, was earlier referred to as comparison theory. However, this approach does not imply that happiness or well-being cannot depend on specific circumstances. Several of these arguments are closely related to the relative consumption hypothesis which states efforts by all to increase their consumption will lead to no significant changes in happiness since their relative consumption does not change. The chapter by Amitava Dutt reviews different explanations about why people increase their consumption without experiencing increases in their happiness, and argues that the relative consumption effect, especially those operating through status-seeking and norm-based consumption, seem to have the widest applicability. This chapter also discusses a simple theoretical formulation of the relative consumption hypothesis and examines some implications of it, such as the role of religiosity in promoting happiness, to show that they are consistent with the empirical evidence. The analysis implies that, at least after a point, more need not be better. It may be noted that, in addition to explaining why consumption may increase without increasing happiness, both of these groups of explanations have important and wide-ranging implications for the methodological

Introduction

11

underpinnings of mainstream or neoclassical economics. The first group questions a basic axiom of neoclassical economics, that is, economic actors are ‘rational’ optimizers. The second group questions the entire notion of efficiency, that is, the idea that more (goods and services) is better. Beyond their interest in whether money buys happiness for individuals, economists are also concerned with how happiness and well-being depend on other macroeconomic factors, such as unemployment, inflation, income distribution and fairness, economic growth and the state of the environment. These relations are not only important in themselves for understanding what conditions make people better off, but also affect the relation between consumption and income, on the one hand, and happiness and well-being on the other. For instance, higher levels of individual consumption can collectively damage the environment which in turn may have a negative effect on happiness. Also, efforts by people to increase their consumption can lead them to support policies – such as less support for worker rights – that may worsen the distribution of income which may in turn reduce happiness. It is also possible that higher levels of consumption boost aggregate demand, increase output and reduce unemployment, thereby making people happier. Two chapters in this book examine the relation between happiness and well-being and such aggregate conditions. Robert Frank examines whether, in light of the Easterlin Paradox which states that increases in income (at least beyond a certain point) do not increase happiness, economic growth is a desirable goal. He first points out that the incomehappiness relation remains a controversial one, as the recent debate between Stevenson and Wolfers (2008) and Easterlin suggests. However, the debate about the income-happiness relationship does not resolve the question about the desirability of growth. While self-reported happiness is good, so are other things, such as autonomy, good health and safe neighborhoods. Frank argues that economic growth does, in fact, improve the human lot, because it reduces child mortality and hunger, and because it produces preconditions for political and social progress, including environmental improvements, reduction of discrimination against minorities and policies which help the poor. Evolutionary selection rewards organisms which have a high probability of survival and therefore are better able to adapt to their environments, but not necessarily happier people. People therefore adapt to the good and the bad, but that is not to say that growth is not good. Carol Graham and Andrew Felton provide a brief review of the empirical literature on the effect of inequality on subjective well-being and then analyse in detail data from a large survey for Latin America. While some earlier studies – including Alesina et al. (2004) – suggest that inequality

12


has a negative effect on happiness in the USA and Europe, not all studies confirm this. Graham and Felton find that the effect of inequality on happiness in Latin America depends on the concept of inequality used (that is, overall inequality in the country, the relative income or wealth to different reference groups and perceived status of a person, or perceptions about overall inequality or social mobility). Their results show that relative income and wealth have a positive, significant effect on happiness, supporting the relative consumption hypothesis and implying that the poor are made less happy by increasing inequality and the rich made happier. They also find that perceptions about inequality, rank and opportunity are at least as important as differences in relative income and wealth, and that inequality in Latin America is perceived as a sign of persistent advantage for the rich and disadvantage for the poor, rather than future opportunities. An important mechanism by which income inequality can arise is discrimination. The effect of discrimination on subjective well-being and happiness, and its consequences on behavior, are examined by Arthur Goldsmith in his examination of the issue of wage discrimination by race. Their analysis starts from the fact that black workers in the USA have a lower wage than white workers of comparable background and characteristics, and from the idea that relative income is an important determinant of happiness and well-being. Using data on perceptions of discrimination they examine how workers who feel that they are discriminated against (in terms of hiring, advancement and general workplace discrimination) react to this – either by waiting to allow employers to learn about their productivity and end statistical discrimination, or by reducing their effort, or by increasing their effort. They do so by examining the impact of these reactions to wage differences between black workers who perceive discrimination, on the one hand, and black workers who do not and white workers, on the other. They find that the data are consistent with the hypothesis that black workers who are discriminated against at hiring attempt to overcome the problem by increasing effort, while those who are discriminated against in the workplace or for advancement reduce their effort.

I.3

HAPPINESS AND POLITICS

Political scientists are latecomers to the study of happiness and subjective well-being. It is one of the goals of this book to bring specifically political concerns, and more specifically still, the concerns of political science as a scholarly discipline, into the academic study of happiness.

Introduction

13

Thus, the contribution by Alexander Pacek provides a general summary of the extant work on the role of the political, broadly construed, as a factor determining the quality of life that citizens experience. Pacek provides a nearly encyclopedic review, cataloguing and critically evaluating the research in a way never previously attempted. As he discusses in more detail, the central areas of political inquiry relate to the role of democracy, social capital (and the governmental mechanisms thought to inhibit or promote it), the role of organized labor as an interest group capable of articulating the interests of the broad class of wage-earners in capitalist society, the size and qualitative characteristics of the welfare state (in conjunction with the social democratic, labor and other progressive parties conventionally understood as agents supportive of state efforts at income maintenance and redistribution), and, finally, the overall taxing and spending policies of governments. Three of these topics are further addressed by contributions to this volume. Ronald Inglehart addresses what is surely the single most basic and compelling question that a political scientist might ask about wellbeing: does the institutionalization of the democratic process, and with it the concomitant civil and political liberties democracy implies, contribute to greater happiness? Perhaps surprisingly to those outside the field, this question may not have the easy answer we might expect. First, there are good reasons to wonder if democracy, so often thought of as a panacea for social problems, really lives up to the hopes we reflexively tend to invest in it. To be sure, democracy might well be valued intrinsically, as the protection against tyranny and exploitation it is synonymous with, but it does not automatically follow that democracy per se makes people’s lives more rewarding. Thus Robert Lane (2000) argues that democracy increases the costs that citizens face, in both the literal and the psychological senses. Simplifying somewhat, Lane suggests that democracy may induce the same kind of anxiety and frustration that existentialists see in modern life: the individual appears, via the existence of representative institutions, to share in the responsibility for events, but at the same time feels powerless as an individual in the face of the collective outcomes over which one has no control. There is, too, the possibility that citizens may collectively make poor choices at the ballot box, politicizing decisions that are better left to the market or technocrats, with deleterious consequences for social life. Another line of argumentation, which is the immediate focus of Inglehart’s work, is the direction of causality between the evidence that we do have linking democracy to greater well-being. In essence, Inglehart asks, does democracy promote satisfaction, or are satisfied citizens a necessary condition for the successful operation of democratic processes?

14


Strikingly, the other political factors that appear in the list of the ‘usual suspects’ in happiness studies all relate to questions of political economy. That receiving the most attention has been the controversial work by Benjamin Radcliff (for example, Radcliff, 2001), suggesting that an expansive welfare state is strongly associated with higher levels of subjective appreciation of life. The fundamental argument at hand is the perennial one between, in the argot that political scientists have come to accept, that of ‘markets vs politics’. We face here the two basic schools of thought in political economy: should human welfare be left primarily to a self-regulating market, with a minimalist state providing the least possible ‘safety net’ and the lowest possible level of government regulation of the economy in general and the labor market in particular, or should the state intervene in the market, through a program of what Radcliff (2001) calls ‘emancipation’ from the market through the joint mechanisms of greater regulation and a greater commitment to income maintenance? Generalizing, the debate becomes one of the overall public policy regimes that governments pursue, with the basic options remaining those available since the advent of the modern industrial economy, viz. the conventional programs of left and right. In his chapter in this volume Pacek reviews at some length the considerable evidence in favor of the contention of Radcliff’s (2001) and others that it is the program of the left that, empirically, appears to offer the surest path to the greatest good for the greatest number. The chapter by Charlotte Ridge, Tom Rice and Matthew Cherry offers a contrary view, suggesting that the apparent statistical relationship between social democratic policy regimes and happiness may neither be as theoretically coherent or as empirically robust as it may appear. As they note, arguments of the sort offered by Radcliff demand a certain amount of skepticism, given the stakes involved. As they write: ‘Not only does [Radcliff] claim that life is better under one type of government than others, it makes the claim with respect to human happiness, perhaps the most meaningful measure of a good society.’ The same tension between market and politics also animates the research on the role of labor organization as an agent of human wellbeing. Political scientists rightly focus on labor unions as political institutions in a variety of respects. Unions are, first, political actors in the electoral and policy making processes, being important elements in the coalition of left and center-left parties throughout the Western world. Capitalist societies are by their very nature class societies, such that organized labor emerges as the only interest group with either the resources or the inclination to represent the interests of working-class citizens. Thus, how countries (or, in the USA, states) differ in the extent to which workers

Introduction

15

are organized is frequently argued to be one of the most salient features in determining their political make-ups. The strength of the labor movement is also commonly viewed as the result of the political process, in that the level of organization is largely (though by no means exclusively) determined by the legal structures that determine how readily workers can form unions. Finally, the workplace itself is an inherently political institution, and the one in which workers spend more waking hours of their lives than any other single activity. Given that the work experience affects not only one’s economic well-being, but one’s sense of dignity and self-respect, the institution of the labor union, as the advocate and agent of the worker in obtaining both income and dignity, has an obvious potential for affecting an individual’s sense of satisfaction with life. Pacek’s chapter reviews the cross-national evidence on the role of unions in fostering well-being, whereas the chapter by Suzanne Coshow and Benjamin Radcliff explores this issue further by focusing on the role of organized labor in the USA per se. They attempt to determine if the relatively modest differences in the power of organized labor across the US states appear to have equally strong effects on well-being as the larger differences across countries do. This question is especially compelling given that the national political context in the contemporary USA is one in which labor is universally regarded as a weak (and ever more progressively weak) political actor.

I.4

HAPPINESS, ECONOMICS AND POLITICS

As noted earlier, the central theme of this book is that the study of happiness is best conducted from a consciously multidisciplinary perspective. To illustrate this, the book has focused on the interaction between economics and politics in understanding happiness. In general, there are several reasons why a multidisciplinary approach is useful. First, different disciplines have evolved in specific ways because of accidents in the development of these disciplines. Some issues have been given the label ‘economic’ and others the label ‘political’ not because they are intrinsically unrelated, but because dominant traditions within them happened to focus on them while excluding others. Second, the way that dominant traditions have separated different disciplines can also be argued to be flawed. For instance, in some popular interpretations ‘economics’ has focused on the operation of ‘markets’ and ‘politics’ on the operation of ‘the state’. The problem with such dichotomies is that they overlook the fact that what are called ‘markets’ and ‘the state’ are intrinsically interrelated. Thus, markets do not exist in a vacuum, but are regulated by social norms and political processes, without which they

16


simply cannot exist, let alone operate smoothly. Moreover, the state is deeply influenced by market processes through, for instance, their effects on income distribution, to say nothing of the dependency of the government on the ‘economy’ to maintain employment, its own tax revenues and, thus ultimately, the maintenance of public order. Of course, as the stock of knowledge continues expanding, specialization in areas of knowledge becomes necessary. However, it is not clear that such specialization is best done in terms of disciplines and sub-disciplines, rather than in terms of issues. Thus, we argue that it is preferable for scholars to focus on the issue of happiness, rather than simply applying the ideas of particular disciplines to the study of happiness. But given our disciplinary backgrounds, this is easier said than done. Indeed, the book is divided into separate parts which deal with economics and happiness, and with politics and happiness. However, while we are at least to some extent prisoners of our disciplines, we realized from our conference that we have much to gain in our study by struggling to break free of our disciplinary moorings. We can illustrate this through four examples of what the chapters in the book suggest can be gained by combining the insights of economics and politics to the study of happiness. The first example concerns the concept and measurement of happiness. Economics has a long history of conceiving of happiness in terms of subjective well-being. The Utilitarians mostly thought of well-being in terms of how people evaluated their circumstances, and Easterlin’s early work used evidence from happiness surveys to quantify the ‘human lot’. As noted earlier, some economists, such as Sen (1999) have discussed serious problems with overemphasis on utility or subjective well-being, and advocated the use of other indicators, such as people’s ability to achieve valued ‘goods’. Political scientists, who seem to be overly wedded to measuring happiness using the notion of subjective well-being by using happiness surveys can do worse than take into account the warnings posed by these economists. The second example relates to whether happiness depends on some outcomes or of some processes. Most economists entertain a consequentialist bias, evaluating well-being in terms of actual outcomes. Political scientists, however, have frequently been more interested in processes, such as whether governments are democratic, and to what extent individuals participate. To be sure, there are some economists who seem to be interested in processes – such as, for instance, those promoting the freedom of choice and the equality of opportunity and reducing economic insecurity – while, as discussed in some of the chapters in this book, some political scientists are interested in the outcomes – such as the election of left parties – on happiness. Happiness and well-being, of course, can be said to depend on both outcomes and processes, and an approach that is informed by the

Introduction

17

work of both economists and political scientists is more likely to recognize this. While these two examples related to the concept and broad approach to the study of happiness, our remaining examples relate to how multidisciplinary study involving both economics and politics can improve our understanding of the determinants of happiness. The third example concerns efforts by economists to understand the determinants of happiness. If increases in income and consumption do not make people significantly happier, at least beyond a certain level, what economic choices and conditions do? Economists and other social scientists have examined the evidence to find that higher socio-economic status, lower unemployment rates, lower inflation, more economic equality (although this is somewhat controversial), the kind (that is, their intrinsic nature) of work people do, more time not spent at work and lower levels of consumer debt contribute to happiness. These findings also suggest that greater economic insecurity – the possibility of losing jobs, for instance – affect subjective well-being. Economic well-being – in terms of functioning and capabilities – is also positively affected by greater access to good health and education. Some of these determinants of happiness can be chosen by individuals, but many of these can be affected by economic conditions. Other determinants, like income distribution and unemployment, are beyond individual control. If there is some validity to these empirical findings, the question then arises: what kinds of overall economic conditions and policies are more likely to increase happiness and well-being? This, in turn, raises the question: what political factors make it more likely that these policies will actually be adopted? For instance, given that the public policies that protect workers from insecurity (through, say, unemployment insurance), or the structure of laws that make union organizing easier or more difficult, are simultaneously economic in their impact on individuals, but also the result of a political process. It is thus necessary to understand economic and political factors and their interaction to answer these questions. As the labor union example illustrates most strikingly, seemingly economic factors are often inherently political, and can only be adequately understood as political phenomena. Fourth, and this time starting from the work done by political scientists (as Pacek’s chapter reviews) suggests that the election of parties on the political left (as compared to those on the political right) increases subjective well-being. These contributions argue that this finding can be explained in terms of the economic security and social safety nets provided by such left-leaning governments However, left-leaning governments are not all the same and there is a great deal of heterogeneity about what policies they pursue. It is worthwhile for political scientists to draw on the

18


work of economists to examine what specific types of policies promote economic security in any given context. We end this section with two comments. First, although we argue that there needs to be a greater recognition of the interaction between the economic and political, and more collaborative work between economists and political scientists, we do not endorse any particular type of what has been called political economy. In particular, we do not endorse that all of this interdisciplinary work needs to follow what is called ‘rational choice’ approach of self-interested, optimizing agents. This approach can illuminate some issues, but we find – for instance, because of what was argued earlier about ‘rationality’ – that for many questions this approach can be extremely misleading. Second, although we have concentrated here on economics and politics, this argument applies to other disciplines – such as psychology, sociology, philosophy and theology – as well, as implied by several chapters in the book.

I.5

TOWARD A HAPPIER WORLD?

A final concern that must animate any study of human happiness is the obvious one of what should be done by people and societies to improve their lives, individually and collectively. The first question is whether happiness itself is a goal that should be pursued. Despite the fact that for so many people happiness is the goal of life and the motivation for living, and the widespread acceptance of the legitimacy of this goal (as ensconced, for instance, in the US Declaration of Independence’s statement that ‘the life, liberty and the pursuit of happiness’ are the ‘inalienable rights’ of people), some have questioned whether the goal is worth pursuing. Wilson (2008), for example, argues that ‘happiness as immediate gratification, happiness as superficial comfort, happiness as static contentment’ could well be a ‘dystopia of flaccid grins’ (pp. 8–9). He argues that the search for happiness is delusional and inauthentic in a world full of insecurity, and that it can lead to the destruction of a thriving culture, by removing the muse of great literature, painting, music and innovation. While Wilson’s admonitions need to be taken seriously, and are closely related to Alasdair MacIntyre’s insightful remarks at the conference, they do not negate the fact that it may be desirable to reduce insecurity of certain kinds for many people in an insecure world in ways that do not provide only immediate gratification through the mere accumulation of ‘stuff’. Moreover, as Wilson is careful to point out, he is ‘thinking only of . . . [the] specific American type of happiness. I am not questioning joy in general . . . [For instance,] I am not criticizing that slow-burning

Introduction

19

bliss that issues from a life spent helping those that hurt. . . . Likewise, . . . I don’t want to romanticize clinical depression’ (Wilson, 2008, p. 7). Even if happiness, appropriately defined, is accepted as a legitimate goal, it can be asked whether we should try to maximize happiness as a matter of public policy. In their chapter in this volume Bruno Frey and Alois Stutzer discuss whether recent advances in the measurement of happiness, especially using data from surveys which ask people to evaluate the quality of their own lives, should make societies try to maximize aggregate national happiness. Although they find that these measures improve on standard measures of evaluation like GDP, they argue that governments should not try to maximize total happiness because doing so fails to take into account the complexities of the political process, fails to take into account the fact that people adapt to their situations (perhaps being on a kind of aspirations treadmill), results in governments trying to manipulate happiness measures to suit their purposes, and induces people to misrepresent their happiness levels strategically to make policies favor them. But even if pursing the formal maximization of happiness is problematic, in Frey and Stutzer’s view happiness research is valuable because it can help people make informed choices on how to best pursue their happiness (broadened to include other goods such as loyalty, self-respect, freedom and personal development) privately and collectively, and to identify institutions which help to achieve these goals. The brief concluding chapter by Dutt and Radcliff addresses the issue of what can be done to help people lead happier and more fulfilling lives? They argue that it is necessary to take not only a multidisciplinary approach to what can be done to increase happiness and well-being, but that it is appropriate to approach this question at different levels: at the level of the individual, of groups, of nations and of the world as a whole. Individuals can do many things to make appropriate decisions. As we have previously noted, the chapter by Sheldon and Lyubomirsky examines directly how, and to what extent, individuals can act to improve the degree to which they find life rewarding. The chapter by Dutt suggests people can be happier by consuming those things which provide gains that endure rather than those things which increase their status or those things to which they quickly get habituated. However, in some cases individuals may lack the incentive to do things on their own, because the happiness they will get from their actions depends not only on what they do, but what others do. For instance, it is possible that if individuals try to work fewer hours to spend more time with family or friends, they may create the impression that they are not dependable and committed employees and possibly lose their jobs. Thus groups of workers can collectively try to reduce their working hours, and governments can impose laws to limit

20


working hours which could increase happiness. But here, too, individuals with a better understanding of the determinants of happiness can make more informed choices about which policies to support. Finally, especially in an increasingly globalized world, the consequences of individual decisions for happiness and well-being may well depend on what people around the world do – the issue of global warming is an obvious example – such that achieving happiness can be seen as a truly human undertaking, not merely a personal or national one.

REFERENCES Alesina, Alberto, Rafael Di Tella and Robert MacCulloch (2004), ‘Inequality and happiness: are Europeans and Americans different?’, Journal of Public Economics, 88, 2009–42. Argyle, Michael (1987), The Psychology of Happiness, 2nd edn, 2001, London: Routledge. Argyle, Michael (1999), ‘Causes and correlates of happiness’, in D. Kahneman, E. Diener and N. Schwarz (eds), Well-Being: The Foundations of Hedonic Psychology, New York: Russel Sage Foundation, pp. 353–73. Diener. E., E.M. Suh, R.E. Lucas and H.L. Smith (1999), ‘Subjective well-being three decades of progress’, Psychological Bulletin, 125, 276–302. Easterlin, Richard (1974), ‘Does economic growth improve the human lot? Some empirical evidence’, in Paul David and Melvin Reder (eds), Nations and Households in Economic Growth: Essays in Honor of Moses Abramovitz, Palo Alto, CA: Stanford University Press, pp. 98–125, reprinted in Easterlin (2002). Easterlin, Richard (2001), ‘Income and happiness: towards a unified theory’, Economic Journal, 111, July, 465–84, reprinted in Richard Easterlin (ed.) (2002), Happiness and Economics, Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing. Easterlin, Richard (ed.) (2002), Happiness in Economics, Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing. Frank, Robert (1999), Luxury Fever. Why Money Fails to Satisfy in an Era of Excess, New York: The Free Press. Frey, Bruno S. and Alois Stutzer (2002), Happiness and Economics, Princeton, NJ: Princeton University Press. Lane, Robert E. (2000), The Loss of Happiness in Market Democracies, New Haven, CN: Yale University Press. Layard, Richard (2005), Happiness. Lessons from a New Science, London: Penguin Press. Maslow, Abraham (1954), Motivation and Personality, New York: Harper. McMahon, Darrin M. (2006), Happiness. A History, New York: Atlantic Monthly Press. Nettles, Daniel (2005), Happiness. The Science Behind Your Smile, Oxford: Oxford University Press. Putnam, Robert (2000), Bowling Alone. The Collapse and Revival of American Community, New York: Touchstone Books.

Introduction

21

Radcliff, Benjamin. (2001), ‘Politics, markets, and life satisfaction: on the political economy of human happiness’, American Political Science Review, 95 (4), 939–52. Schor, Juliet (1998), The Overspent American. Upscaling, Downshifting and the New Consumer, New York: Basic Books. Sen, Amartya (1999), Development as Freedom, New York: Anchor Books. Stevenson, Betsey and Justin Wolfers (2008), ‘Economic growth and subjective well-being: reassessing the Easterlin Paradox’, Brookings Papers on Economic Activity, 1, 1–87. Veblen, Thorstein (1899), The Theory of the Leisure Class. An Economic Study of Institutions, New York and London: Macmillan. Veenhoven, R. and J. Ehrhardt (1995), ‘The cross-national pattern of happiness: test of predictions implied in three theories of happiness’, Social Indicators Research, 43, 33–86. Wilson, Eric (2008), Against Happiness, New York: Farrar, Straus and Giroux.

PART I

Happiness

1.

The history of happiness and contemporary happiness studies Darrin M. McMahon

I feel that I am singularly unfit to answer the question: ‘what is happiness?’ in large part because of my training as a historian, which makes me, I fear, unduly attentive to the way in which words and concepts change their meanings over time. To be perfectly frank, I’m partial to Immanuel Kant’s observation that ‘the concept of happiness is such an indeterminate one that even though everyone wishes to attain happiness, yet he can never say definitely and consistently what it is what he really wishes and wills.’ But clearly that is not really going to be good enough for this chapter. So how to answer the question ‘what is happiness?’ I might point out, as I do in my book, the strong and stubborn etymological link between happiness and luck in every Indo-European language (see McMahon, 2006). The old Norse and Old English root ‘hap’, like the old French heur or the Mittelhockdeutsch ‘Glück’ simply means luck or fortune. We have mishaps when bad things happen to us. And when good things happen to us – when we are lucky – we are happy, Glücklich, filled with bon-heur. I might, to take another tack, note the equally long and stubborn connection relating happiness and good fortune to fortune itself – to wealth, prosperity, fertility and abundance. It is not coincidental that the early Greeks spoke of the gods as olbios or makarios – as blessed or happy – not least because of their material prosperity. Thus the Homeric ‘Hymn to Hermes’ uses a form of makarios to describe the cave dwelling of the god Hermes and his mother, which is full of ‘nectar and lovely ambrosia’, with much ‘silver and gold’, fine clothing, and other things ‘such as are kept in the sacred houses of the happy’. Nor is it coincidental that the Romans placed the goddess Felicitas on the back of coins, with a horn of plenty in one hand, symbolizing abundance, fecundity and bounty. Nor is it coincidental that they referred to the destitute as ‘miser’ – wretched, unfortunate, poor – the root, of course, of our modern English term miserable. Still another tack – less historical and more analytical would be to borrow from the positive psychologists and social scientists – to note the various dimensions of happiness on a synchronic as opposed to a 25

26


diachronic plane. And here we might, following say Marty Seligman, or the English psychologist Daniel Nettle, among others, point to a first level set of associations linking happiness to positive emotion and good feeling – feelings of pleasure and joy. A second level order of happiness having to do with the longer term, encompassing a sense of satisfaction and well being, a set of judgments about one’s condition or state. I am happy with my marriage, with my work and so forth. Finally, we might distinguish a third level of happiness – the broadest of all – having do with quality of life as a whole, with human flourishing, with fulfilling one’s potential, with human excellence and the good life. In short, use a definition that accounts for some of the distinctions between hedonic and eudemonic conceptions of happiness discussed in Ong’s chapter in this volume. Now there is much to recommend this latter approach – and for practical purposes it may well be among the most satisfactory. But historians have the luxury of not being practical, as my wife – a practical woman – constantly reminds me. And so let me be myself – impractical, perhaps even something of an annoyance – by taking a page from Wittgenstein to note, as a historian, that no matter how hard we try to pin down happiness, its meaning, or meanings, will always be pregnant with the past and with its past uses. Happiness, like all of us, cannot entirely escape the past. And so it strikes me that it is worthwhile thinking seriously about those past uses before we, in the present, make use of happiness ourselves. What I’d like to do, in effect, is to repay a debt. I have benefited a great deal from the work that is done in contemporary happiness studies, and so it is clear to me at least what it can do for history. But what, it may be asked, can history do for happiness studies? And here I would hope not just to reiterate some of the past uses of happiness that inevitably bear on those in the present, but also to try to place contemporary happiness studies in a somewhat broader context, to help situate the present moment vis-à-vis the past. For though it may be true, as William James observed in the Varieties of Religious Experience that ‘How to gain, how to keep, how to recover happiness, is in fact for most men at all times the secret motive of all they do, and of all they are willing to endure’, human beings have never been as preoccupied, never been as obsessed, I would argue, with happiness as they are right now. Indeed, it is really only in the eighteenth century that considerable numbers of people began to think of happiness as a this-worldly possibility. Hitherto, happiness, at least in Western societies, had been considered by and large either as a condition of the future (of the millennium, say, or the second coming, or when the children of Israel are fully redeemed in the promised land – next year in Jerusalem), or in the past (in the Garden of Eden, in a primordial golden age, a Prelapsarian time of innocence). Or,

The history of happiness and contemporary happiness studies

27

alternatively, in another dimension of space and time altogether (heaven, or those ‘blessed or happy isles’ of the Greeks). Happiness in the here and now –in the normal conditions of life – wasn’t really considered an earthly prospect, or at least wasn’t considered by most as such. Now of course it is certainly true that one has the tradition of classical philosophy initiated by Socrates towards the end of the fifth century bce – a tradition that presented happiness or human flourishing (eudaimonia) as a function of human virtue. This is tradition that is developed by a great many Greek and Roman moralists, though none so centrally as Aristotle, for whom eudaimonia, as you know, was the goal or end, the telos of human activity. But the point I want to stress here is that for Aristotle – and in this respect he is perfectly in keeping with virtually every prominent Classical moralist after Socrates – happiness, though yes an earthly prospect, was not a habitual reward. On the contrary, happiness was a prize to be won over the course of a lifetime only by the virtuous – the happy few – those whose excellence of conduct and character allowed them to rise above normal human conditions, to live what Aristotle describes in the Nichomachean Ethics, as a ‘god-like’ life. To be happy might be within human power, but it was a power that would only ever be realized by a very small percentage of the human population. How radically different this is from that goal, first stated in the eighteenth century, to pursue the greatest happiness for the greatest number, to seek to attain, what the French Constitution of Year 3 during the French Revolution described, in its very first article, as ‘common happiness’, the happiness of all. ‘Does not every man have a right to happiness’, asks the author of the article on happiness in Diderot and D’Alembert’s great Encyclopédie, the Bible as it were of the European Enlightenment. A right to happiness! Think about it: this is revolutionary talk! And in that respect, the French revolutionary St Just was perfectly right to announce, as he did in the National Convention in 1794, that ‘happiness is a new idea in Europe’. In some real ways it was. But what was a revolutionary pronouncement in the age of Enlightenment – a right à la Jefferson to the pursuit of happiness, or a right à la St Just and the French revolutionaries to its attainment – has steadily become less and less revolutionary and more and more a part of our received assumptions about the way human life should be. Far from thinking about happiness as a miracle of the universe – or as the attainment of a god-like few – people in the developed world tend to think of happiness, today, I would argue (however they define it) as the natural human state, the way men and women ought to be if they are not abused, or prone to depressive illness, or unjustly deprived of their natural human endowments.

28


It is worth stressing that this conviction involves an assumption about human nature and human experience – about the purpose of human lives – about the way we are intended to be here on earth. It is something of a teleological assumption – one that is nice to believe, one that is comforting, one that is humane and which may in fact be true. It is an assumption, nevertheless, that can’t really be proved. The assumption rests, to some degree, on an article of faith. I talk quite a bit about this in my book – about the way in which the belief in happiness as what Alexander Pope called in the eighteenth century human ‘being’s end and aim’ not only constituted an article of faith in its own right, but an article of faith that steadily challenged, and gradually replaced, however fitfully and imperfectly, that of the reigning Judeo-Christian belief in individual salvation by God. To be sure, not all bowed either immediately or easily to the new faith in earthly happiness. I’m fond of quoting Thomas Carlyle as an example of one such apostate. That cranky, irascible Scot could write in the mid nineteenth century: Every pitifulest whipster that walks within a skin has had his head filled with the notion that he is, shall be, or by all human and divine laws ought to be, ‘happy.’ His wishes, the pitifulest whipster’s, are to be fulfilled for him; his days, the pitifulest whipster’s, are to flow on in ever-gentle current of enjoyment, impossible even for the gods. The prophets preach to us, Thou shalt be happy; thou shalt love pleasant things, and find them. The people clamour, Why have we not found pleasant things?

Carlyle was perhaps something of an exception in his blunt refusal to countenance the new faith. But others who shared it were at least prepared to acknowledge that it was in part just that, an article of faith. Darwin is interesting on this score, as is John Stuart Mill, as is Freud, who of course was ultimately a skeptic. In his Das Unbehagen in der Kultur, Civilization and its Discontents, a work, revealingly, which was originally titled Das Unglück in der Kultur, Unhappiness in Civilization, Freud, after parsing the pleasure principle which in his view ‘decides the purpose of life’, concluded that the pleasure principle was ‘at loggerheads with the whole world, with the macrocosm as much as with the microcosm. There is no possibility at all of its being carried through,’ he declared, ‘all the regulations of the universe run counter to it’. Threatened always by the suffering of our own bodies which are doomed to decay and dissolution and which cannot even do without pain and anxiety as warning signals; from the external world, which may rage against us with overwhelming and merciless forces of destruction; and finally from our relations to other men, Freud concluded that the barriers to sustained happiness were insuperable. Human beings


29

might take happiness to be the ‘purpose and intention of their lives’. ‘They strive after happiness; they want to become happy and to remain so.’ But they were, according to Freud, mistaken or deceived. ‘One feels inclined to say,’ he concluded, ‘that the intention that man should be “happy” is not in the plan of “Creation.”’ Freud was writing in the aftermath of the First World War and on the brink of the Second, not long before Hitler came to power, and so we might forgive him his pessimism. But it is, I think, an index of how far we have moved since that time that Freud’s tragic view of the inevitable frustrations, conflicts and irresolvable tensions of the human condition has long been out of favor. The Enlightenment belief in happiness – the Enlightenment faith in happiness – has totally triumphed in the developed world in the second half of the twentieth century, commensurate with – it is surely not irrelevant to note – the greatest cumulative economic expansion in human history. If, as we know from our sociologist colleagues’ work on reported subjective well-being, men and women in the USA and Europe have not, it seems, gotten appreciably happier since the 1950s, despite the massive gains in cumulative GNP. It is the case, I would argue, that men and women’s sense that they should be happy has in fact increased a great deal. Paradoxically this increase in expectations may actually decrease happiness by increasing disappointment. What I call the unhappiness of not being happy is a phenomenon one can detect in Western culture since the eighteenth century, but it has probably never been as acute as it is today. So happiness as our being’s end and aim – this Enlightenment creed has triumphed, and in the process it has tended to crowd out, discredit or co-opt other ways of looking at the world and the human purpose in it. Religion provides an interesting case in point. You would be hard-pressed to find in, say, the early part of the seventeenth century, a Christian religious apologist arguing that religion was a means to happiness (at least a means to happiness in this life). Religion – and the salvation that it offered – were considered, rather, ends in themselves. And yet increasingly, beginning in the latter part of the seventeenth century, religious apologists themselves have tended to genuflect before the new god of human happiness on earth. True Pleasure, Chearfulness, and Happiness, The Immediate Consequence of Religion was the way one American author in the 1760s titled his book on the merits of Christianity, while reminding his readers that Christ’s first miracle was to create more wine to keep the party going – and there are many Catholic analogues for this too. By the early nineteenth century, this tendency was considerably developed, prompting Alexis de Tocqueville to observe in Democracy in America that whereas Old World priests had once spoken ‘of nothing but the other life,’ and ‘hardly took any trouble

30


to prove that a sincere Christian might be happy here below,’ preachers in America were ‘continually coming down to earth’ Indeed they find it difficult to take their eyes off it. The better to touch their hearers, they are forever pointing out how religious beliefs favor freedom and public order, and it is often difficult to be sure when listening to them whether the main object of religion is to procure eternal felicity in the next world or prosperity in this.

Living, as I now do in the South, a transplant from New York City, I can affirm that this form of exhortation is very much alive. Get religion, get happy. When the latest Pew Foundation findings linking evangelical Protestantism to subjective well-being were released recently, people in my neck of the woods were doubly elated. So even among those who might be expected to be preaching fire and brimstone, earthly happiness today in effect is the highest good. The French philosopher Pascal Bruckner goes so far as to observe that it (happiness) has become ‘the sole horizon of our democracies’. Taking into account that general Enlightenment triumphs, I think, may help us to situate contemporary happiness studies in a somewhat broader context than you might get simply from studying its place in the history of psychology, say, or economics or sociology. And that is a reflection that I hope will give you a slightly better understanding of the extraordinary contemporary appeal of work on happiness in a number of fields, for that work is in many respects the culmination and perfect expression of precisely the dynamic I have traced briefly here. Happiness, we might say, is all that many have left, and so it is only natural to conclude that we should do everything in our power to figure out how to secure it. Notwithstanding the creativity and insight of many of those who study happiness today, this broader dynamic, I think, helps to account even more than their own labour for the tremendous popularity of their work. Let me make one more reflection – or series of reflections – based on my observation about the uses to which happiness has been put in the past and how that bears on the present. I think one needs to recognize that when happiness began to occupy the space formerly occupied by religion – when it became, to quote a letter from Voltaire in 1726, ‘the great and only concern’, there was born a concept of extraordinary power and allure. For what had for so long resided on the horizon of human experience, outside our temporal bounds, the source and repository of all our hopes and longings and dreams, had now been pulled down from heaven to earth, and dangled before us, every one of us, as a legitimate prospect in the here and now. With the result that those who could marshal those hopes, who could claim to lead us towards the coveted promised land of happiness on earth


31

were necessarily vested with extraordinary power. At the very time that St Just was proclaiming happiness as a new idea in Europe, his Jacobin colleagues were preaching secular sermons in former Catholic churches about happiness on earth – ‘real’ happiness, as opposed, they claimed, to the false kind that came surrounded by angels – real happiness, while the guillotine did its work. ‘The overcoming of religion as the illusory happiness of the people is the demand for their real happiness’, Marx would later write in his Contribution to the Critique of Hegel’s Philosophy of the Right. It is one of those terrible ironies of history that in taking up this injunction, Stalin liked to describe himself as the Constructor of Happiness. So what am I implying with these ominous allusions? That authorities on happiness are somehow dangerous, despite their best intentions? That the search for earthly happiness must end in blood? Of course not. And yet I would enter a humble plea as a historian and scholar of the humanities for a certain humility as we approach the study of happiness, pointing out the un-nerving tendency of happiness to frustrate and circumvent those who would try to grasp it in pursuit. . . . Fortune’s wheel turns treacherously And out of happiness brings men to sorrow

the monk observes, in Chaucer’s The Canterbury Tales. We risk missing something in today’s post-Enlightenment world, I would argue, when we fail to acknowledge that un-nerving tendency, and when we fail pay heed to those traditions of knowledge – be they Classical (think of Greek Tragedy), or Jewish (think of the tale of Job), or Christian (think of the account of original sin) – which emphasized the elusive nature of happiness and its quest, the difficulty of ever fully securing it in our grasp, the little piece of us that, however happy we might seem, always seems to cry out for more. These are insights that one finds again and again in Western history. From Horace’s lapidary reflection: ‘Nihil est ab omni parte beatum.’ Nothing is completely happy. Or Rousseau’s frank avowal: ‘I doubt whether any of us knows the meaning of lasting happiness. Happiness leaves us, or we leave it.’ Or John Stuart Mill’s insight that if you ‘Ask yourself whether you are happy . . . you cease to be so.’ There are many other such poignant reflections. I like to point out that something of this same elusive quality is creeping about in the phrase the ‘pursuit of happiness’ itself. We focus, rightly, on the word happiness, but pursuit is interesting too. In the eighteenth century it had a somewhat harder meaning than it does today, closely related, in fact, to its cognates, ‘prosecute’ and ‘persecute’. If you look, for example in Samuel Johnson’s great eighteenth-century Dictionary of the English Language, you’ll find:

32


‘To Pursue . . . 1. To chase; to follow in hostility.’ ‘Pursuit . . . 1. The act of following with hostile intention’. Johnson being the one who observed that a man may be happy in the future or in the past, but in the present, never, except when drunk. But this harder sense of ‘pursuit’ is interesting. The French talk about la chasse au bonheur, the hunt for happiness, as if one were in the process of stalking a wild, and potentially dangerous, beast. A beast, presumably, that one has to kill when it is finally cornered and can’t flee any more. So all this by way of registering a reminder that the pursuit of happiness and the uses to which happiness has been put have not always been happy. A reminder that no matter how hard we try to fix its meaning, the word and concept will always come to us charged with its religious and metaphysical past as the ultimate human end, the final place of rest, the solution and salvation to human dissatisfaction, the answer to the riddle of existence. In the early middle ages Boethius could observe that ‘God is happiness itself.’ I don’t think it is entirely an exaggeration to say that for many, today, happiness has become a sort of god. Which means that we, as its interpreters and perhaps prescribers, share in something of a priestly craft – at the very least share a moral responsibility that is greater than we might always appreciate at first glance. In this respect, the Oxford Don and Anglican Archbishop of Dublin, Richard Whately, was certainly right when he said in the nineteenth century that happiness is no laughing matter.

REFERENCES Carlyle, Thomas (1965), Past and Present, New York: New York University Press, edited by Richard D. Altick, 157. de Tocqueville, Alexis (1988), Democracy in America, New York: Harper Perennial, translated by George Lawrence and edited by J.P. Mayer, 2: 530. Freud, Sigmund (1989), Civilization and its Discontents, New York: W.W. Norton, translated and edited by James Strachey, Introduction by Peter Gay, 24–5. James, William (1994), The Varieties of Religious Experience, New York: Modern Library Edition, 78. McMahon, Darrin M. (2006), Happiness. A History, New York: Atlantic Monthly Press. Marx, Karl (1844 [1983]), ‘Contribution to the Critique of Hegel’s Philosophy of the Right: Introduction’, in The Portable Karl Marx, edited by Eugene Kamenka, New York and London: Penguin, 115.

2.

On the measurement and mismeasurement of happiness: contemporary theories and methodological directions Anthony D. Ong*

It has become clear that the phenomena referred to as human well-being is a mosaic of many component parts. This mosaic can be partitioned into a parsimonious set of dimensions, indicating measurements, that fairly completely account for individual differences among a large number of these components. The components that social scientists have been able to measure probably do not represent the entire range of experiences that constitute human well-being, but they are a goodly sample. Hundreds of different instruments have been designed to assess various features of human health and well-being. Analyses of these different instruments indicate that what is measured in common is fewer than a dozen broad dimensions indicating major kinds of positive human experience. Scientific understanding, thus, has moved away from the idea that human well-being can be well represented by a single dimension (often referred to as happiness). Well-being, to be sure, is many faceted. But to recognize that well-being has many facets is merely to start to understand it. Just what are the facets; how do they emerge in culture and in individuals? What are their functions? The volume makes it clear that well-being involves deep insight into the meaning and purpose of life. Yet well-being is not merely knowledge and comprehension of the moral imperatives of the good life, even as such insight and comprehension must be part of it. The moral authority of human well-being must extend beyond that needed to ensure the continuance of any particular group; it must deal adequately with the continuance of all humankind. Yet it must derive from acculturation within a family, within a kin group, within a particular society. Well-being must also derive from experience, but experience is not sufficient to produce it: great and diverse experience need not result in well-being. Finally, human well-being is distinct from any of what is referred to as human happiness, as Carol 33

34


Ryff has pointed out (Ryff, 1989), but it must involve and stem from much of that happiness. So we begin with the understanding that human well-being is too complex to fully understand. Nevertheless, here an attempt is made to describe an important feature of well-being that involves some aspects of mature human happiness, even as it is more than such happiness. To clarify the distinction between well-being and happiness, I will first describe what the term ‘human happiness’ has come to mean within the context of the science of hedonic psychology (Kahneman et al., 1999). I will then describe how this concept must be broadened to include the development of eudaimonia (Waterman, 1993), one form of which is an important feature of human well-being. While an in-depth review is beyond the scope of this chapter, I do strive to critically evaluate and address conceptual and methodological issues surrounding the need for (1) reliable and theory-driven measures of positive health and well-being, (2) study designs that link information at different levels of analysis, and (3) innovative methodological approaches that are sensitive to complex dynamic relationships.

2.1

EMPIRICAL INVESTIGATIONS OF HUMAN WELL-BEING

Two lines of research bring us to our current theoretical understanding of the nature of human well-being. One results in what is referred to as the theory of subjective well-being (SWB; Diener, 1984). The other results from the theory of psychological well-being (PWB; Ryff, 1989). This chapter is directed at merging these two theories. To show why the unification is necessary, I will first outline the basis for the two theories and then describe how the two can and should be united. 2.1.1

Hedonic Well-being

Equating happiness with hedonic pleasure has a lengthy history that dates back to antiquity. Beginning with Epicurus, philosophers such as Erasmus, More, Hume, Hartley and Bentham believed that the goal of life is to experience the maximum amount of pleasure, and that happiness is the totality of one’s hedonic moments. Psychologists who have adopted the hedonic view have tended to focus on a broad conception of hedonism that includes preferences and judgments about the good/bad elements of life (Kahneman et al., 1999). Hedonic well-being is thus a scientific description of human happiness. There is a problem in describing human well-being in this way, however. Happiness is a singular word. But the

On the measurement and mismeasurement of happiness

35

accumulated evidence indicates that there is more than one kind of quality that can be said to be characteristic of human happiness. Diener (Diener, 1984; Diener et al., 1999) recognized this in his tripartite account of subjective well-being (positive affect negative affect and life satisfaction). Thus, the problem is that use of the singular word ‘happiness’ fosters belief that different positive human states are all forms of one thing, happiness, but this does not seem to be the case. It is thus better to use the plural of happiness and describe SWB as a theory of several forms of happiness. Better yet, the theory can be described simply as hedonic well-being. SWB theory is largely descriptive – an account of what are the adaptive characteristics that distinguish the human capacity for generating and coping with emotional reactions to life events. But the theory is also a description of variables with which these adaptive characteristics correlate, an account of how and why such relationships come about. It is thus also explanatory. Researchers have identified two facets of SWB: a cognitive judgment of life satisfaction and an emotional aspect consisting of independent positive and negative affect components. A person’s evaluation of their life may thus be in the form of cognitions and affect. Individuals are said to have high SWB if they experience high life satisfaction and frequent pleasant emotions such as joy and affection, and only infrequently experience unpleasant emotions such as sadness and anger. In contrast, individuals are said to have low SWB if they are dissatisfied with life, experience little joy and affection, and frequently feel negative emotions such as anger or anxiety (see ibid. for a review). 2.1.2

Eudaimonic Well-being

Despite the prevalence of the hedonic view, many philosophers have denigrated happiness per se as a principal standard of well-being. Aristotle, for example, posited that true happiness is found in the expression of virtue – that is, in doing what is worth doing. Eudaimonism is an ethical theory that refers to well-being as distinct from happiness. Eudaimonic theories maintain that not all desires would yield well-being when achieved. That is, even though they are pleasure producing, some outcomes are not ‘good’ for people and would not promote wellness. Waterman (1993) argued that whereas happiness is hedonically defined, the eudaimonic conception of well-being calls upon people to live in accordance with their ‘daimon’ or true self. He suggested that eudaimonic well-being occurs when people’s life activities are most congruent with deeply held values. Thus, from the eudaimonic perspective, happiness cannot be equated with human well-being. Embracing the concept of eudaimonia or self-realization as a central

36


definitional aspect of well-being, Ryff and her colleagues (Keyes et al., 2002; Ryff, 1995; Ryff and Singer, 1998) have explored the question of eudaimonic well-being in the context of developing a lifespan theory of optimal experience. Echoing Waterman’s concerns, Ryff argues for wellbeing not simply as the attaining of pleasure, but as ‘the striving for perfection that represents the realization of one’s true potential’ (Ryff, 1995, p. 100). Ryff and Keyes (1995) presented a multidimensional approach to the measurement of psychological well-being (PWB) – as distinct from SWB – that taps six distinct aspects of human actualization: autonomy, personal growth, self-acceptance, life purpose, mastery and positive social connectedness. The extant evidence indicates that although measurement of SWB and PWB is not error-free, considerable progress has been made in identifying and measuring the separate elements of SWB and PWB. Reliable measures of these elements have been developed – the Positive and Negative Affect Schedule (PANAS; Watson et al., 1988), the Satisfaction With Life Scale (SWLS; Diener et al., 1985) and the Psychological Well-Being Scales (Ryff, 1989). Different forms of evidence have been put forth to indicate the validity of these elements. Evidence of discriminant validity of SWB elements has been supported with multi-trait, multi-method analyses (Lucas et al., 1996). Evidence of convergent validity of PWB elements has been indicated with common factor analyses (Ryff and Keyes, 1995). And evidence for the convergent and discriminant validity of all nine SWB and PWB elements has been supported with confirmatory factor analyses (Keyes et al., 2002). Thus, it is clear that the phenomenon of human wellbeing is multidimensional.

2.2

METHODOLOGICAL INNOVATIONS IN WELLBEING RESEARCH

Although the extant evidence has provided a basis for understanding the phenomena of well-being, other basic information is necessary to establish the nature of the phenomena. That is, individuals are believed to exhibit coherent patterns of experience that cannot be fully described or explained merely by locating individuals within a fixed system of trait dimensions (Allport, 1961). Thus, although nomothetic (between-person) analyses have yielded converging evidence for the construct validity of measures of SWB and PWB, very little attention has been given to investigating idiographic (within-person) relations among these elements. Perhaps nowhere more than in well-being research is the importance of repeated measurement and analysis so essential. Studies that include only


37

one occasion of measurement provide a good example of ambiguities that arise when an assumption of stability is made. These ambiguities have been described in detail by Nesselroade (Nesselroade, 1984). When participants are measured on only one occasion, the inter-individual variability in the measurements can reflect three different sources: (1) stable differences among people (traits), (2) intra-individual variability (states), and (3) temporal measurement error. These three possible sources of variation are inextricably confounded when data are obtained on only one occasion, and it is impossible to separate them. Because phenomena may also vary reliably and lawfully within individuals, conclusions based on nomothetic research are premature without idiographic information. With few exceptions (for example, Zevon and Tellegen, 1982), however, construct validation of well-being measures has been based largely on nomothetic, rather than idiographic research. Moreover, virtually all of the within-person well-being evidence to date has centered on distinguishing the occurrence of hedonic emotional states. Little is known about whether the separate elements of SWB and PWB can be reliably and independently observed within individuals studied across time. To my knowledge, no study has provided evidence indicating that the reliability and independence of measurements that have been indicated in between-person analyses of SWB and PWB (Keyes et al., 2002) also obtains for within-person observations of these phenomena. Evidence of this possibility is needed. The fact that well-being has been observed to be a relatively stable and trait-like personality characteristic in interindividual differences research (Headey et al., 1993), raises the question of how SWB and PWB is maintained over time (for a discussion, see Ong et al., 2006). Because the process of change represents a main, central issue for the scientific study of well-being, research designs are needed that can capture ongoing processes of growth and adaptation. In this section I highlight the utility of longitudinal panel and intensive bursts designs. Arguments are presented that bear on the value of these designs as underutilized approaches that appear particularly appropriate to the investigation of intra-individual change and variability in SWB and PWB. Throughout, I argue that the strength of the process approach is an essential shift away from cross-sectional, single variable explanations toward person-centered accounts of positive health. 2.2.1

Longitudinal Panel Designs

Many of the most interesting research questions addressed in well-being research relate to how individuals change over time and what factors

38


influence the development of adaptive change. Longitudinal panel designs are particularly well suited for evaluating models of long-term change or development. In the typical longitudinal panel design (1) data are collected at two or more points in time, (2) the same sample of people is interviewed at distinct points in time, and (3) data from the respondents are compared across these time points in order to monitor patterns of change. Although longitudinal panel designs vary with respect to the composition of the sample, the number of follow-up assessments and the intervals between assessments, such designs have two defining characteristics. First, the same research participants, who constitute the panel, are measured for two or more points in time (the measurement periods or waves). Second, at least one variable is measured for two or more waves. This is the longitudinal aspect of the data, which allows the measurement of qualitative or quantitative change within individuals from one wave to the next. In contrast to the longitudinal panel design, cross-sectional designs involve the assessment of research participants at only one measurement point (for a review, see Raudenbush, 2000). 2.2.2

Intensive Bursts Designs

There are times when the investigator is interested in closely observing change while it is occurring. In comparison with longitudinal panel designs, intensive bursts designs allow researchers to observe processes of change within a short but rapidly changing window of time. The use of electronic diaries (for example, palm pilots) allows for the study of the determinants and consequences of changes in well-being within people’s everyday lives. The short time intervals between events and self-reports improves accuracy and reduces bias. In addition to these improvements in measurement precision, repeated assessments of the same person over time solve a serious problem in inference that plagues research in this area. Variables that predict differences between people on an outcome like happiness may have no effect or even the opposite effect on the same outcome when measured as a change within the person observed over time (Tennen and Affleck, 1996). Only careful studies that evaluate changes over time in both the independent and dependent variable can safely make such assertions. Finally, electronic diaries have methodological advantages that are connected to the use of intensive bursts designs. First, electronic diaries allow individuals to report their behavior and experiences over the range of situational circumstances experienced in everyday life. Second, they allow for statistical modeling of behavior over time. Third, and most important, such procedures can test, rather than assume, the validity of the nomothetic approach.


39

In addition to designing studies of change, one critical aspect of testing theories of change is fitting models of change to empirical data. Below I describe analytic possibilities that are available for longitudinal panel designs and for intensive bursts designs. I focus my comments on two general data analysis strategies, namely those associated with growth curve modeling and dynamic systems analysis, respectively. For a more thorough discussion of other statistical approaches for modeling change, the interested reader is referred to Collins and Horn (1991), Collins and Sayer (2000), Kenny and Zautra (2001), McArdle and Hamagami (2001) and Raudenbush (2001). 2.2.3

Growth Curve Modeling

One of the major goals of positive psychology is to determine factors that influence normal and optimal development. These factors may be fixed at a particular level (for example, gender, ethnicity) or variable (for example, physical health, emotions). Traditional statistical methods such as repeated measures analysis of variance cannot take into account the time-varying nature of covariates. The most commonly used approach to modeling change in continuous variables that allow for time-varying covariates is growth curve models. Growth curve models, such as hierarchical linear models (Raudenbush, 2000), fit growth trajectories for individuals and relate characteristics of these individual growth trajectories (for example, slope) to covariates. Because these models typically involve relatively few occasions of measurements, longitudinal panel designs are generally the temporal design of choice when fitting growth curve models. Growth curve modeling is an appropriate technique for studying individual change because repeated measures can be considered as nested within individuals and can be represented as a two-level hierarchical model. At the within-person level, each individual’s development is modeled as a unique growth trajectory. At the between-person level, the growth parameters of these trajectories become the outcome variables, which are then modeled as a function of person-level characteristics. For excellent overviews of growth curve and hierarchical linear models for longitudinal panel studies, the reader is referred to Raudenbush (2000) and McArdle and Nesselroade (2003). 2.2.4

Dynamic Systems Analysis

A recent implementation of intensive bursts designs is dynamic systems analysis. Fundamentally, a dynamical systems approach offers a way to formalize concepts of self-regulation. The focus is on modeling or

40


representing the relationships between the current state of a variable or an ensemble of variables and the subsequent state of such variables (Boker and Nesselroade, 2002). One key advantage of the dynamic systems approach over other approaches to modeling dynamic processes is the capacity to represent ‘shocks’ or other inputs from outside the individual. For example, consider a model of self-regulation that reflects a ‘pendulum with friction’, which is hypothesized to best exemplify the intra-individual disregulation that may result from exposure to daily stress. This model is referred to as a damped linear oscillator. The equation for the damped linear oscillator can be expressed as a linear regression formula in which the acceleration of the pendulum is the outcome variable and the position and velocity of the pendulum are the predictor variables (Boker, 2001). From a developmental perspective, velocity may refer to the linear change in the system (for example, change in mood), and acceleration may pertain to the curvature (for example, the speed with which the mood change occurs). Differential equation models express effects within a system in terms of their derivatives (that is, the instantaneous rates of change of the variables), as well as in terms of the values of the variables themselves. For example, a differential equation model of emotion regulation following stress might relate daily affect to its slope, or first derivative (that is, how rapidly an individual’s mood was changing). A more complete model might include effects related to its curvature, or second derivative (that is, how rapidly mood was accelerating and decelerating in its change). These three parameters, initial position (emotion/affect), velocity (change) and acceleration (speed of change) represent a dynamical system in which the relationships between them define a central tendency of a family of trajectories that any one individual might have (Boker and Nesselroade, 2002). The regression coefficients from this structural equation model, in turn, define order parameters (for example, frequency and decay rate) of the system that best represents the interrelations between variability in affect and stress over time. The dynamic systems approach is both efficient and powerful, since it can identify intra-individual fluctuations in dynamics using relatively sparse data.

2.3

SUMMARY AND CONCLUSIONS

I have strived to demonstrate in this chapter that scientific understanding has moved away from the idea that human well-being can be well represented by a single dimension. Evidence accumulated over the course of this century has made it clear that the phenomenon of human well-being is multidimensional. Therein lies a problem in identifying particularly happy individuals; therein lies a problem of determining where to look


41

for particularly happy individuals; therein lies a difficulty of examining a hypothesis stipulating that, on average, happy people will display more wisdom and character than unhappy people. Jahoda (1958) brought attention to this problem 50 years ago and it still does not have a ready solution. The use of eudaimonic indicators solves one problem but introduces another: in what sense is one better off with a higher ‘purpose in life’, to take one example, if unhappiness accompanies it? Researchers, thus, should strive to assess both hedonic and eudaimonic indicators of wellbeing to obtain a more complete understanding of positive human health. I have also suggested that one major limitation of current theorizing in positive psychology is inherent in the very properties of extant measurement tools. That is, most theories of well-being (SWB and PWB) are described in terms of Cartesian coordinates or factors. These factors may be rotated into an infinity of different positions, each equally adequate for describing the relationships among dimensions of well-being, but each calling for different concepts and different language for describing human well-being. A metatheory of simple structure has guided the rotation that has been accepted as the basis structure of SWB and PWB theory. This metatheory requires that manifest dimensions of well-being relate to a finite number of factors. This is a reasonable requirement for studies designed to indicate it – and many studies have been so designed – but it is not an indication of how well-being must be organized to account for relationships that are observed within and across individuals. A fundamental limitation of any theory built on a rectilinear system of factors is that it is not of a form that well describes natural phenomena: it is thus unlikely to be fully adequate. Rather, it is a system of a kind that can accurately describe rectangular structures built by humans – the angles of city streets or rooms of buildings – but not the rounded and irregular structures of nature. The phenomena of nature are not usually well described by the linear equations of a Cartesian coordinate system. A system of factors is not a system for representing rounded structures such as we see in the configurations of plants and animals or of the human brain. Nor is it a set of structural formulas such as those for the hydrocarbons, the chief constituents of living things. The equations that describe the outer structure and convolutions of brains must be parabolas, exponentials, hyperboles and the like. It is likely that the equations that best describe the inner organizations and workings of human well-being are of the same forms, not those that describe city blocks and buildings. Finally, I have underscored the importance of taking a process approach to understanding the complexity of positive human health and wellbeing. Extant theories of SWB and PWB provide few details about how well-being develops or about how positive psychological states interact

42


and work together to produce optimal human functioning. These theories, thus, do little to indicate the dynamics of human adaptation. The kind of system that ultimately will best describe such adaptation and its development, I submit, will be functional and will map on to the human brain. To represent such adaptation and development mathematically, it might be more nearly of the form of a spiral of Archimedes, out of which evolves a repetitive building on what is known (induction), which leads to deductions that generate empirical studies and more induction, which leads to further deductions, which spawn further induction and so on. In the long run, knowing that science is a never-ending search for better explanations and that no theory of reality is final, we can be confident that SWB and PWB theory will be replaced by a better theory.

NOTE * Preparation of this chapter was supported, in part, by a grant from the National Institute on Aging (R01 AG00156).

REFERENCES Allport, G.W. (1961), Pattern and Growth in Personality, Oxford: Holt, Reinhart and Winston. Boker, S.M. (2001), ‘Differential models and differential structural equation modeling of intraindividual variability’ in L.M. Collins and A.G. Sayer (eds), New Methods for the Analysis of Change, Washington, DC: American Psychological Association, pp. 5–27. Boker, S.M. and J.R. Nesselroade (2002), ‘A method for modeling the intrinsic dynamics of intraindividual variability: Recovering the parameters of simulated oscillators in multi-wave panel data’, Multivariate Behavioral Research, 37, 127–60. Collins, L.M. and J.L. Horn (eds) (1991), Best Methods for the Analysis of Change: Recent Advances, Unanswered Questions, Future Directions, Washington, DC: American Psychological Association. Collins, L.M. and A.G. Sayer (2000), ‘Modeling growth and change processes: design, measurement, and analysis for research in social psychology’, in Charles M. Judd and Harry T. Reis (eds), Handbook of Research Methods in Social and Personality Psychology, New York: Cambridge University Press, pp. 478–95. Diener, E. (1984), ‘Subjective well-being’, Psychological Bulletin, 95, 542–75. Diener, E., R.A. Emmons, R.J. Larsen and S. Griffin (1985), ‘The Satisfaction With Life Scale’, Journal of Personality Assessment, 49, 71–5. Diener, E., E.M. Suh, R.E. Lucas and H.L. Smith (1999), ‘Subjective well-being: three decades of progress’, Psychological Bulletin, 125, 276–302. Headey, B.W., J. Kelley and A.J. Wearing (1993). ‘Dimensions of mental health: life satisfaction, positive affect, anxiety and depression’, Social Indicators Research, 29, 63–82.


43

Jahoda, M. (1958), Current Concepts in Positive Mental Health, New York: Basic Books. Kahneman, D., E. Diener and N. Schwarz (eds.) (1999), Well-being: The Foundations of Hedonic Psychology, New York: Russell Sage Foundation. Kenny, D.A. and A. Zautra (2001), ‘Trait-state models for longitudinal data’, in L.M. Collins and A.G. Sayer (eds.), New Methods for the Analysis of Change, Washington, DC: American Psychological Association, pp. 243–63. Keyes, C.L., D. Shmotkin and C.D. Ryff (2002), ‘Optimizing well-being: the empirical encounter of two traditions’, Journal of Personality and Social Psychology, 82, 1007–22. Lucas, R.E., E. Diener and E. Suh (1996), ‘Discriminant validity of well-being measures’, Journal of Personality and Social Psychology, 71, 616–28. McArdle, J.J., and F. Hamagami (2001), ‘Latent difference score structural models for linear dynamic analyses with incomplete longitudinal data’, in L.M. Collins and A.G. Sayer (eds.), New Methods for the Analysis of Change, Washington, DC: American Psychological Association, pp. 139–75. McArdle, J.J. and J.R. Nesselroade (2003), ‘Growth curve analysis in contemporary psychological research’, in Wayne F. Velicer and John A. Schinka (eds), Handbook of Psychology: Research Methods in Psychology, Vol. 2, New York: John Wiley & Sons, Inc., pp.447–80. Nesselroade, J.R. (1984), ‘Concepts of intraindividual variability and change: impressions of Cattell’s influence on lifespan developmental psychology’, Multivariate Behavioral Research, 19, 269–86. Ong, A.D., J.L. Horn and D.A. Walsh (2006), ‘Stepping into the light: modeling the dynamics of hedonic and eudaimonic well-being’, in A.D. Ong and M. van Dulmen (eds), Oxford Handbook of Methods in Positive Psychology, New York: Oxford University Press, pp. 12–28. Raudenbush, S.W. (2000), ‘Comparing personal trajectories and drawing causal inferences from longitudinal data’, Annual Review of Psychology, 52, 501–25. Raudenbush, S.W. (2001), ‘Toward a coherent framework for comparing trajectories of individual change’, in A.G. Sayer and L.M. Collins (eds), New Methods for the Analysis of Change, Washington, DC: American Psychological Association, pp. 33–64. Ryff, C.D. (1989), ‘Happiness is everything, or is it? Explorations on the meaning of psychological well-being’, Journal of Personality and Social Psychology, 57, 1069–81. Ryff, C.D. (1995), ‘Psychological well-being in adult life’, Current Directions in Psychological Science, 4, 99–104. Ryff, C.D. and C.L.M. Keyes (1995), ‘The structure of psychological well-being revisited’, Journal of Personality and Social Psychology, 69, 719–27. Ryff, C.D. and B. Singer (1998), ‘Human health: new directions for the next millennium’, Psychological Inquiry, 9, 69–85. Tennen, H. and G. Affleck (1996), ‘Daily processes in coping with chronic pain: methods and analytic strategies’, in Norman S. Endler and Moshe Zeidner (eds), Handbook of Coping: Theory, Research, Applications, Oxford: John Wiley & Sons, pp.151–77. Waterman, A.S. (1993), ‘Two conceptions of happiness: contrasts of personal expressiveness (eudaimonia) and hedonic enjoyment’, Journal of Personality and Social Psychology, 64, 678–91.

44


Watson, D., L.A. Clark and A. Tellegen (1988), ‘Development and validation of brief measures of positive and negative affect: the PANAS scales’, Journal of Personality and Social Psychology, 54, 1063–70. Zevon, M.A. and A. Tellegen (1982), ‘The structure of mood change: an idiographic/nomothetic analysis’, Journal of Personality and Social Psychology, 43, 111–22.

3.

How do we assess how happy we are? Tenets, implications and tenability of three theories Ruut Veenhoven1

3.1

THE PROBLEM

Happiness is highly valued in present day society. Not only do people aim at happiness in their own life but there is also growing support for the idea that we care for the happiness of other people and that governments should aim at creating greater happiness for a greater number of citizens (Bentham, 1789). This classic philosophy is not only more accepted these days, but also more practicable, now that scientific research provides more view on the conditions for happiness (Veenhoven, 2004). In that context, happiness is commonly understood as how much one likes the life one lives, or more formally, the degree to which one evaluates one’s life as a whole positively. A central element in this definition is subjective ‘evaluation’ or ‘liking’ of life, also referred to as ‘satisfaction’ with life. These words refer to a mental state but leave some ambiguity about the precise nature of that state. That question is differently answered in three theories linked to different theories about how we evaluate life. Set-point theory sees the evaluation as a stable attitude towards life and focuses more on the mental processes that maintain this attitude than on the processes that have brought it about. Comparison theory sees evaluation rather as a continuous judgment process involving the comparison of perceptions of life-as-it-is with notions of how-life-should-be. Affect theory sees happiness also as a continuous mental process, but now as an appraisal of how well one feels usually. These different descriptive theories of how we assess how happy we are have great implications for prescriptive theories of happiness. Set-point theory, and to a lesser extend also comparison theory, implies that there is little value in happiness and that there is also little chance of furthering happiness enduringly and this goes against the utilitarian tenet that we should aim at greater happiness for a greater number. 45

46


This begs the question whether these theories adequately reflect reality or not. Do they apply at all and, if so, do they apply equally well or do some apply more than others? Over the last 15 years I have addressed these questions in several publications (Veenhoven, 1991, 1994, 1995, 1997). In this chapter I develop the argumentation further, linking up with an evolutional perspective and taking new empirical findings into consideration. I will also reflect on Cummins’s (Cummins et al. 2002) recent ‘homeostatic’ theory of happiness. Below I will start with a closer look at the concept of happiness and next review each of the above-mentioned theories about how we assess how happy we are. Each of these theories will be discussed in the following way. First, I describe the main tenets and variations. I then discuss in more detail what moral implications these theories have. Next, I evaluate each of these views by considering their theoretical plausibility and the empirical support. I start with a precise definition of happiness.

3.2

CONCEPT OF HAPPINESS

The word happiness is used in different meanings that are often mixed up. To avoid such confusion, I will review the main connotations and select one of these, which I analyse in more detail. 3.2.1

Meanings of the Word

When used in a broad sense, the word happiness is synonymous with ‘quality of life’ or ‘well-being’. In this meaning it denotes that life is good, but does not specify what is good about life. The word is also used in more specific ways, and these can be clarified with the help of the classification of qualities of life presented in Table 3.1. Four qualities of life This classification of meanings depends on two distinctions. Vertically there is a difference between chances for a good life and actual outcomes of life. Horizontally there is a distinction between ‘external’ and ‘internal’ qualities. Together, these distinctions mark four qualities of life, all of which have been denoted by the word ‘happiness’. Livability of the environment The left top quadrant denotes the meaning of good living conditions. Often the terms ‘quality of life’ and ‘well-being’ are used in this particular meaning, especially in the writings of ecologists and sociologists. Economists sometimes use the term ‘welfare’ for this

How do we assess how happy we are?

Table 3.1

Life-chances Life-results Source:

47

Four qualities of life Outer qualities

Inner qualities

Livability of environment Utility of life

Life-ability of the person Satisfaction

Veenhoven (2000a).

meaning. ‘Livability’ is a better word, because it refers explicitly to a characteristic of the environment. Politicians and social reformers typically stress this quality of life and sometimes refer to it as happiness. I rather see it as a condition for happiness and not happiness as such. One can live in excellent circumstances but still be unhappy, because of an inability to reap the chances. Life-ability of the person The right top quadrant denotes inner lifechances. That is: how well we are equipped to cope with the problems of life. This aspect of the good life is also known by different names. Doctors and psychologists especially also use the terms ‘quality of life’ and ‘wellbeing’ to denote this specific meaning. There are more names however. In biology the phenomenon is referred to as ‘adaptive potential’. On other occasions it is denoted by the medical term ‘health’, in the medium variant of the word.2 Sen (1992) calls this quality of life variant ‘capability’. I prefer the simple term ‘life-ability’, which contrasts elegantly with ‘livability’. This quality of life is central in the thinking of therapists and educators. I also see this as a prerequisite for happiness and not as happiness itself. Even a perfect person will be unhappy when living in Hell. Utility of life The left bottom quadrant represents the notion that a good life must be good for something more than itself. This presumes some higher value, such as ecological preservation or cultural development. Moral advisors emphasize this quality of life. This usefulness of life has also been denoted with the word happiness, but again I do not follow that use of words. In my language one can lead a useful life but still be unhappy. Satisfaction with life Finally, the bottom right quadrant represents the inner outcomes of life. That is the quality in the eye of the beholder. As we deal with conscious humans, this quality boils down to subjective appreciation of life. This is commonly referred to by terms such as ‘subjective well-being’, ‘life-satisfaction’ and also ‘happiness’. I follow this latter use of the word.

48

Table 3.2


Four kinds of satisfaction

Part of life Life as a whole

Passing

Enduring

Pleasure Top-experience

Part-satisfaction Life-satisfaction

Four kinds of satisfaction This brings us to the question of what ‘satisfaction’ is precisely. This is also a word with multiple meanings and again we can elucidate these meaning using a simple scheme. The scheme in Table 3.2 is based on two distinctions; vertically between satisfaction with ‘parts’ of life versus satisfaction with life ‘as a whole’, and horizontally between ‘passing’ satisfaction and ‘enduring’ satisfaction. These two bi-partitions yield again a four-fold taxonomy. Pleasures Passing satisfaction with a part of life is called ‘pleasure’. Pleasures can be sensoric, such as a glass of good wine, or mental, such as the reading of this text. The idea that we should maximize such satisfactions is called ‘hedonism’. The term happiness is sometimes used in this sense and then denotes a particular pleasant experience. I do not use the term happiness for this matter. Part-satisfactions Enduring satisfaction with a part of life is referred to as ‘part-satisfaction’. Such satisfactions can concern a domain of life, such as working life, and aspects of life, such as its variety. Sometimes the word happiness is used for such part-satisfactions, in particular for satisfaction with one’s career. I do not use the term happiness in this meaning. Peak-experience Passing satisfaction can be about life as a whole, in particular when the experience is intense, pervasive and ‘oceanic’. This ecstatic kind of satisfaction is usually referred to as ‘peak-experience’ or ‘bliss’. When poets write about happiness they usually describe an experience of this kind. Likewise religious writings use the word happiness often in the sense of a mystical ecstasis. Another word for this type of satisfaction is ‘Enlightenment’. I do not use the term happiness in this sense. Life-satisfaction Enduring satisfaction with one’s life as a whole is called ‘life-satisfaction’ and also commonly referred to as ‘happiness’ and as ‘subjective well-being’. I do use the word happiness in this meaning, and will use it interchangeably with ‘life-satisfaction’.


3.2.2

49

Definitions of Happiness as Life-satisfaction

This brings us to the question of what ‘life-satisfaction’ is precisely. A review of the various definitions reveals that this concept is often linked to mental processes supposed to be involved, definitions of happiness reflecting theories of happiness. Affective definitions Several definitions depict happiness as an affective phenomenon. For instance, Wessman and Ricks (1966, pp. 240–1) wrote: ‘Happiness appears as an overall evaluation of the quality of the individual’s own experience in the conduct of his vital affairs. As such, happiness represents a conception abstracted from the flux of affective life, indicating a decided balance or positive affectivity over long periods of time.’ In a similar vein Fordyce (1972, p. 227) states ‘Happiness is a particular emotion. It is an overall evaluation made by the individual in accounting all his pleasant and unpleasant experiences in the recent past.’ These definitions are close to Jeremy Bentham’s (1789) famous definition of happiness as ‘the sum of pleasures and pains’, which also involves the notion of an ‘affect balance’. A contemporary variation on this theme is proposed by Daniel Kahneman (2000) in the notion of ‘objective happiness’, which is the ‘raw’ affective experience that underlies the overall evaluation of life.3 Cognitive definitions Happiness is also defined as a cognitive phenomenon, that is, as the result of a deliberate evaluation process. In that vein McDowel and Newell (1987, p. 204) describe life-satisfaction as a ‘Personal assessment of one’s condition compared to an external reference standard or to one’s aspirations’. Likewise, Shin and Johnson (1978, p. 478) defined life-satisfaction as a ‘global assessment of a person’s quality of life according to his chosen criteria’. Some of the definitions in this line stress the active achievement of life goals (for example, Annas, 2004), while others rather stress the absence of unfulfilled aspirations, for example, Schmitz (1930, p. 234) who depicted happiness as a ‘state of being without desires’. In all conceptualizations happiness is deemed to be higher, the smaller the distance between standard and reality. Attitudinal definitions Happiness has also been depicted as a happy disposition and as a positive attitude towards life. In this line Lieberman (1970, p. 40) wrote ‘at some point in life. Before even the age of 18, an individual becomes geared to a certain stable level of satisfaction, which – within a rather broad range of

50


environmental circumstances – he maintains throughout life’. Some of the definitions of this kind stress the consistency in affective response, while others rather see it as a belief system. Mixed definitions Several definitions combine one or more of the above elements. For instance, Diener defines Subjective Well-Being (SWB) as being satisfied with life (attitude), while feeling good (affect). In his own words: ‘Thus a person is said to have high SWB if she or he experiences life satisfaction and frequent joy, and only infrequently experiences unpleasant emotions such as sadness or anger. Contrariwise, a person is said to have low SWB if she or he is dissatisfied with life, experiences little joy and affection and frequently feels negative emotions such as anger or anxiety’ (Diener et al., 1997, p. 25). All three elements are involved in Chekola’s (1974, p. 2002) definition of happiness as ‘realization of a life-plan and the absence of seriously felt dissatisfaction and an attitude of being displeased with or disliking one’s life’. Likewise Sumner (1997, pp. 145–6) describes ‘being happy’ as ‘having a certain kind of positive attitude toward your life, which in the fullest form has both a cognitive and an affective component. The cognitive aspect of happiness consists in a positive evaluation of your life, a judgment that at least on balance; it measures up favorably against your standard or expectations . . . The affective side of happiness consists in what we commonly call a sense of well-being, finding your life enriching or rewarding or feeling satisfied or fulfilled by it’. 3.2.3

My Conceptualization of ‘Overall’ Happiness and ‘Components’

In my own conceptualization of happiness similar distinctions are used, but in a more systematic way. I distinguish between ‘overall’ happiness and ‘components’ of happiness and assume that the latter function as ‘sub-totals’ in the overall evaluation of life. Overall happiness Overall happiness is defined as ‘the degree to which an individual judges the overall quality of his life as a whole favorably’ (Veenhoven, 1984, pp. 22–4). Thus defined, happiness appears as an attitude towards one’s own life that has some stability of its own and that involves related feelings and beliefs. These feelings and beliefs are seen as ‘components’ of happiness. Components of happiness When evaluating their lives, people can use two more or less distinct sources of information: their affects and their thoughts. We can ‘observe’

How do we assess how happy we are? Global assessment

51

OVERALL HAPPINESS Satisfaction with one’s life as a whole

Sub-totals

Hedonic level of affect Balance of pleasant and unpleasant affect

Contentment Perceived realization of wants

Information basis

Affective experience

Cognitive comparison

Figure 3.1

Happiness and its components

that we feel fine most of the time, and we can also ‘judge’ that life seems to meet our (conscious) demands. These appraisals do not necessarily coincide. We may feel fine generally, but nevertheless be aware that we failed to realize our aspirations. Or we may have surpassed our aspirations, but nevertheless feel miserable. The relative weight in the overall evaluation is variable in principle; it is an empirical question to what extent one component dominates the other (Figure 3.1). Hedonic level of affect We experience different kinds of affects: feelings, emotions and moods and these experiences have different dimensions, such as active – inactive and pleasant – unpleasant. This latter dimension is called ‘hedonic tone’. When we assess how well we feel we typically estimate the pleasantness in feelings, in emotions, as well as in moods. I call this ‘hedonic level of affect’ and this concept fits the above-mentioned ‘affective’ definitions of happiness. A person’s average hedonic level of affect can be assessed over different periods of time: an hour, a week, a year as well as over a lifetime. The focus here is on ‘current’ hedonic level. This concept does not presume subjective awareness of that average level. One can feel good most of the time, without being fully aware of that. Therefore this concept can be applied to beings who cannot reflect on their own life, such as animals and little children. Contentment Unlike animals and little children most adults can also evaluate their life with the use of reason and compare life as it is with notions of how one wants life to be. The degree to which individuals perceive their wants to be met is called ‘contentment’ and this concept equals the above-mentioned ‘cognitive’ definitions of happiness.

52


This concept presupposes that the individual has developed some conscious wants and has formed an idea about their realization. The factual correctness of this idea is not at stake. This conception of happiness as a ‘trinity’ helps to place different theories about how we assess how happy we are.

3.3

SET-POINT THEORIES OF HAPPINESS

Set-point theories of happiness hold that we are programmed to experience a certain degree of happiness, largely irrespective of how well we are doing. In this view happiness just happens to us. 3.3.1

Variants

A classic religious version of this theory is Devine predestination, God having decided that some people will be happy and others not, just as he foresees who will enter Heaven and who will be dammed to Hell. Secular variants assume that happiness is geared by mental inclinations that are also beyond a person’s control. Genetic disposition This variant holds that happiness is largely determined by an innate disposition to enjoy life or not. A spokesman of this view is Lykken (1999), who claims to have shown that about 50 percent is heritable. There is uncertainty about the nature of this disposition, some see it in the reward system of the brain and link it to positive or negative ‘affectivity’ while others hold secondary effects responsible, such as inborn physical health. In the latter case happiness is essentially a variable state, though it tends to remain at the same level because of constancy in its determinants. Below I will not discuss this variant of set-point theory. Personality trait Another current view is that happiness depends very much on personality traits, that is, predispositions to react in a certain way. One of these ways is liking things or not and personality traits such as ‘extraversion’ and ‘neuroticism’ are seen to determine our affective reactions to and perceptions of things that happen to us. It is generally assumed that these traits have a genetic component. In this view personality molds the evaluation of life. Personality can also affect happiness through its impact on the course of life-events, and this is


53

central in the dynamic-equilibrium theory of Heady and Wearing (1992). Yet again, I do not consider that a set-point theory, because happiness itself is essentially a variable state in this idea. Cultural view A macro-level variant of this theory is that the view on life is embodied in the national character. In this line Inglehart (1990, p. 30) wrote that cross-national differences in happiness ‘reflect cognitive cultural norms, rather than individual grief and joy’. In an earlier paper I have depicted that view as the ‘Folklore theory of happiness’ (Veenhoven, 1995, p. 35). Homeostatic maintenance While the above set-point theories aim at explaining differences in happiness, there are also theories of this kind that focus at the general level of happiness. These are motivational theories that assume that we tend to maintain a comfortable level of happiness, even in adverse conditions. In that line Cummins et al. (2002) hold that we unconsciously keep happiness between 7 and 8 on a 10-step scale, just as we maintain a body temperature of 32°C. 3.3.2

Implications

These theories imply that there is little chance of creating greater happiness for a greater number, since happiness is a stable trait rather than a variable state and as such not responsive to external conditions. In this view one can at best try to raise that fixed level a bit, be it with genetic engineering or training. The theory also implies that there is little sense in raising happiness, since happiness is unrelated to the wider thriving of the individual. In this view being happy or not is comparable to liking chocolate or not; fine if you do but no real problem if you don’t. 3.3.3

Theoretical Plausibility

It is plausible that differences in stable conditions for happiness create stable differences in level of happiness and conditions for happiness can be external or internal (see Table 3.1). It is also plausible that happiness tends to remain at a similar high level in the favorable and stable conditions of modern society. Yet set-point theory holds that the stability is not in the preconditions, but in the evaluation itself and that is not so plausible.

54


Why, then, do we evaluate life at all if we always end up with the same conclusion? In this light it is difficult to see why happiness is so prominent in people’s minds, for example, that they think of it almost every day (Freedman, 1978). One also wonders why evolution has developed the ability to evaluate if the evaluation of life boils down to a fixed response. Set-point theory implies that happiness does not serve any function in human life and that being happy or unhappy is as trivial as having brown or blond hair. It can hardly be reconciled with the fact that happiness is universally pursued and neither with the fact that being happy or not appears to be closely linked to how well we thrive. Taken to the extreme, set-point theory would predict that we are equally happy in Heaven as in Hell and this is hard to believe. 3.3.4

Empirical Support

At first sight there is strong empirical support for the set-point theory, happiness tending to be stable over time. Follow-up of individuals shows little change in happiness from one year to another and if changes occur, these are typically short-lived. Trend analysis of average happiness in nations also shows much stability. Yet, at a closer look, we also see change. Long-term follow-up of individuals reveals considerable mobility along the life-satisfaction ladder in modern society, Ehrhardt et al. (2000) estimated that over a lifetime only 30 percent of the original rank order in happiness will be left. Follow-up studies have also shown that at least some life-events bring about a permanent change in life-satisfaction, for instance, getting married or losing one’s job. Though people tend to adjust to external shocks in their lives, that adjustment is not always complete (Diener et al., 2006). Likewise, average happiness in nations appears not to be immutable, average happiness has risen gradually in most nations over the last 30 years (Veenhoven and Hagerty, 2006) while in some countries an abrupt fall in happiness occurred, for example, in Russia after the ‘Rubel crisis’ in the late 1990s (Veenhoven, 2000b). At first sight there is also empirical support for Cummins’s theory that we tend to maintain a level of satisfaction between 7 and 8. Studies in modern Western nations showed indeed a concentration of responses in these categories, but surveys in other parts of the world show another picture, for example, an average of 3.2 in Tanzania and in the abovementioned case of Russia a dip from 5.1 to 4.1. Another finding that contradicts this theory is the high number score of 9 and 10 in some Western nations, for example, a 20 percent score of 10 in Switzerland.


3.4

55

COGNITIVE THEORIES OF HAPPINESS

Cognitive theories hold that happiness is a product of human thinking and reflect discrepancies between perceptions of life as it is and notions of how life should be. Notions of how life should be are assumed to root in collective beliefs and to vary across cultures. This view on happiness is dominant in philosophy and also pervades the thinking of many social scientists. 3.4.1

Tenets

The basic assumption of this theory is that happiness is based on the comparison with standards, though there is a difference on the nature of these standards and ways of comparison. Another basic assumption is that collective beliefs are involved. Comparison The theory assumes that we have ‘standards’ of a good life and that we constantly weigh the reality of our life against these standards. Standards are presumed to be variable rather than fixed and to follow perceptions of possibilities. In other words, we would tend to judge life by what we think it can realistically be. Different theories stress different standards. In the variant of life-time comparison, the focus is on whether we are doing better or worse than before. In this view a happy youth will not add to happiness in adulthood. The social comparison variant stresses how well we are doing relative to other people, and in particular people like us. In this view happiness is surpassing the Jones’s. Several of these theories are combined in Michalos’s (1985) ‘Multiple Discrepancies Theory’ of happiness, which assumes that we not only compare with what we want and with what others have, but also with what we need and with what we deem fair. Social construction The idea that we compare to standards begs the question of where these standards come from. This is typically seen as an outcome of socialization, involving the adoption of collective notions of the good life, sometimes with minor modifications. These collective notions of the good life are seen as ‘social constructions’ that draw heavily on the wider culture and shared history. In this line some sociologists argue that happiness as such is also a social construction. In this view happiness is a culturally variable concept, comparable to the notion of ‘beauty’.

56


Reflected appraisal A sociological variant holds that we not only compare life our self with our own standards, but that we also appraise our life through the eyes of others, in other words, that in assessing how happy we are we estimate how happy other people think we are. If so, this enhances the salience of shared standards of the good life. This theory is summarized in Figure 3.2. 3.4.2

Implications for Happiness Promotion

This theory holds that happiness does not depend on objective conditions of life, but on the standards by which these conditions are judged. As such, it also implies that there is little value in happiness. One reason is that happiness may be bought by a lowering of standards, as advocated in some variants of Buddhism. A second reason is the relativistic argument that all standards of the good life are mere collective illusions, with limited appeal in a particular time and place. Most cognitive theories imply also that there is little chance of creating greater happiness for a greater number, in particular the theories that assume that standards adjust to reality. Some variants of this theory predict that happiness will vary around the neutral level (for example, Unger, 1970), while some variants even predict that most people will be unhappy, for example, theories that stress the social salience of success in advertisements and the news. 3.4.3


It is reasonable to assume that we use our thinking in appraising the quality of our life. Yet it is not so reasonable to assume that thinking is the only way to assess how happy we are. If so, little children cannot be happy, because they lack the ability to define standards of the good life and compare with reality. If thinking were the only way of assessing how we are doing one also wonders what our affect system is good for and why affective experience is so pervasive. Is affect then a mere remnant of the past? Still another qualm is that standards of the good may be less clear than assumed. We mostly have some notions in mind, but typically not a clear hierarchy of wants and having a ‘rational life plan’ seems to be more exception than a rule. There is also a problem with the implication that happiness does not depend on real conditions of life but on the intellectual yardsticks by which these are valued. This would mean that one can be perfectly happy


Global assessment

Sub-assessment

Information basis

Underlying process

Substrate

Figure 3.2

57




Standard setting

Culture

Cognitive theory of how happiness is assessed

in Hell, provided that one does not know better or that one is socialized to believe that this is the best place to be. In this view there is no adaptive value in happiness and, in fact, not in thinking either. This problem is mainly in the assumptions of how collective notions of the good life come about. If one assumes that these are unique constructs, following the internal logic of particular belief systems one ends up concluding that happiness is of no consequence, which I deem implausible from an evolutionary point of view. If, on the other hand, one assumes that these notions reflect accumulated experience with the realities of life, the conclusion is rather that living up to these standards is mostly wise and that happiness is therefore an indication of proper living. As we will see below, this view is compatible with the ‘need’ theory of happiness.

58

3.4.4


Empirical Support

Correlational studies typically show a strong relationship between overall happiness and contentment. The smaller the gap between standard and reality, the higher the level of happiness (for example, Michalos, 1985). This correlation is generally interpreted as proof for the theory that happiness depends on the outcome of a comparison process, but causality could also work otherwise, happiness determining comparison, particularly the estimation of the size of gaps. This so-called ‘top-down’ effect was demonstrated in a follow-up study in Australia by Headey et al. (1991) for satisfaction with one’s standard of living and with one’s job. Correlational studies further show relationships between happiness and perceived achievement of specific goals, such as completing a study or raising a family. Yet this research also shows that success in some goals counts more than success in other goals, and in particular that success in material goals is relatively weakly related to happiness. It seems that achievement of intrinsic goals adds more to happiness than success in extrinsic goals (Kasser and Ryan, 1993) and this contradicts the idea that happiness is geared by socially constructed standards in the first place. There is also empirical support for the assumption that standards adjust over time and that effects of life-events on happiness are therefore shortlived. For instance, follow-up of people who had had a financial windfall showed an uplift of happiness that lasted only one year (Gardner and Oswald, 2001). Yet entering marriage appears to have more lasting effects on happiness, in particular for people who were not too happy when single, and severe physical handicaps, such as spinal cord injury, appear to reduce happiness permanently. For reviews of the data see Veenhoven (1994) and Diener et al. (2006). Another disconfirming finding is that most people tend to be happy most of the time, while life-time comparison theory would predict that the average is about neutral and some variants of social comparison theory imply that the average must be below neutral. A further fact that does not fit the theory is the close relationship between average happiness in nations and objective quality of life. Average happiness differs widely across nations (between 8.2 and 3.2 on scale 0–10) and about 75 percent of these differences can be explained by variation in ‘hard’ societal characteristics, such as economic affluence, freedom and democracy (Veenhoven, 2004). These findings contradict the idea of culturally unique standards and adjustment to the possible. Interestingly, there appears to be neither relationship between average happiness and income-inequality in nations nor a relationship with state welfare effort,


59

while these matters are widely seen as desirable. So, if notions of the good life affect happiness at all, not all affect happiness equally much.

3.5

AFFECTIVE THEORIES OF HAPPINESS

Affect theory holds that happiness is a reflection of how well we feel generally. In this view we do not ‘calculate’ happiness, but rather ‘infer’ it, the typical heuristic being ‘I feel good most of the time, hence I must be happy’ (Schwartz and Strack, 1991). 3.5.1

Tenets

In this line of thought one question is how we take stock of our affective experience. Another question is what makes us feel good or bad and this links up to the wider question about the functions of affect. Frequency of affect It would seem that the overall evaluation of life is geared by the most salient affective experiences and that these are typically intense affects. This view is common in fiction and is more or less implied in life-reviews. Yet research using the Experience Sampling Method shows that it is rather the relative frequency of positive to negative affect that matters (Diener et al., 1991). Mood as informant How do we assess that relative frequency? The cognitive view on affect procession suggests that we compute an affect balance in some way, using estimates of frequency and duration. A competing view is that this occurs automatically and that the balance reflects in mood. In this view mood is an affective meta-signal that, contrary to feelings and emotions, is not linked to specific objects. Emotions denote an affective reaction to something and prepare the organism to a response, while negative mood signals that there may be something wrong and urge to find out what that is. Gratification of needs Why do we feel good or bad at all? Probably because this informs us in how well we are doing. Affects are an integral part of our adaptive repertoire and seem to be linked to the gratification of human needs. ‘Needs’ are vital requirements for survival, such as eating, bonding and exercise. Nature seems to have safeguarded the gratification of these needs with affective signals such as hunger, love and zest. In this view positive mood signals that all needs are sufficiently met at the moment. ‘Needs’ in this

60


theory should not be equated with ‘wants’ in the above discussion of cognitive theories. Needs are inborn and universal, while ‘wants’ are acquired and can be variable across cultures. Wants can concur more or less with needs. Motivation to act In this view negative and positive moods function as red and green lights on the human machine, indicating either that there is something wrong or that all systems are functioning properly. If so, this is likely to have behavioral consequences, negative mood urging to cautions and positive mood encouraging going on. This is what Fredrickson’s (2004) ‘broaden and build’ theory is about. This theory is summarized in Figure 3.3. 3.5.2

Implications for Happiness Promotion

In this view happiness is a desirable state, both because it signals good adaptation and because it enhances behavior that apparently works out well. This is at least so if one accepts that it is good that we live up to our nature. In this view it is also possible to create greater happiness for a greater number. If happiness depends in the end on the gratification of human needs, we can advance happiness both by improving the livability of the environment (left top quadrant in Table 3.1 and by enhancing individual life-abilities (right top quadrant in Table 3.1). There are limits to this, but even if the average happiness of 8.2 in present-day Denmark might be the highest possible level, there is still much room for improvement in the rest of the world. 3.5.3


It is hard to imagine someone saying to enjoy life when feeling depressed most of the time. Such a person may say that their life is nevertheless ‘meaningful’ but that is not the same as ‘satisfying’; remember the distinction in Table 3.1 between the ‘usefulness’ of a life and satisfaction with life. This theory also makes sense in an evolutionary perspective. It is likely that evolution has developed ways of monitoring needs gratification, in particular in organisms that can choose. It is unlikely that rational thinking is the main way, since this developed late in evolution. It is quite likely that adaptation is guided by affective signals in the first place and that all higher animals can feel more or less well. It is unlikely


Global assessment

Sub-assessment



Information basis


Underlying process

Need gratification

Substrate

Figure 3.3

61

Human nature

Affect theory of how happiness is assessed

that humans are an exception to this rule. The ability to think was added to an existing affect system and did not replace that. This can be seen in the structure of the human brain, where the affect system is located in the older parts that we have in common with other animals and where the capability to think is situated in the neo-cortex that is typical for humankind. 3.5.4

Empirical Support

Unlike ‘wants’, ‘needs’ cannot be measured and neither can ‘need gratification’. A direct test of this theory is therefore not possible. Still we can test the implications of this theory. One implication is that people will be unhappy in conditions where basic human needs remain unmet, such in chronic hunger, danger and loneliness. This prediction is supported by the finding that average happiness

62


Global assessment

OVERALL HAPPINESS Satisfaction with one’s life as a whole 1

Sub-assessment


2


Information basis


3


4

Underlying process

Need gratification

Standard setting 5

Substrate

Figure 3.4

Human nature

6

Culture

Causal effects in the evaluation of life

is low in poor countries with failed states. Support can also be seen in the rising happiness in modern nations (Veenhoven, 2005). At first sight the prediction is contradicted by absence of a correlation between individual happiness and income in rich nations, but this may mean that the material needs of even the poor are gratified. Gratification of social needs is less well secured in rich nations and consequently we do see a substantial impact of marriage and friendship on happiness. Another testable implication is that happy people must thrive better biologically. This appears indeed in greater longevity of the happy. Well controlled long-term follow-up studies show sizable effects, comparable to smoking or not (Veenhoven, 2008).

3.6

HOW APPRAISALS RELATE

So far I have depicted these ways of evaluating life as separate appraisals, which each influence the overall evaluation of life in their own way. Yet these mental processes are linked in several ways. Figure 3.4 summarizes some probable interactions.


3.6.1

63

Set-points Root in Earlier Appraisals

If understood as a stabilized attitude, set-points must have developed in the past on the basis of experience. This is not necessarily only one’s own experience, since attitudes can also be copied. Still, in the case of this attitude towards one’s own life it is likely that one’s own experiences plays a role and as such it is likely that set-points root in earlier affective and cognitive appraisals. In this view set-points are an echo of the past that are likely to wane in the course of time and then be revised, in particular when major life-change urges to a reappraisal. In this case affective and cognitive appraisals appear on the scene again. 3.6.2

Hedonic Affect Influences Contentment

In this line it seems probably that hedonic level of affect plays a role in the comparison process, in particular in the assessment of the gap between want and reality. When feeling good we will tend to see small gaps and when feeling bad we may attribute that feeling to wide gaps. This affective ‘bias’ is probably stronger at the higher level of aggregation, it may not affect appraisals of success to specific standards too much, such as the appraisal of whether your dissertation met your scientific aspirations, but is likely to influence estimates of success in meeting all standards of the good life. In relation with overall life satisfaction this is known as the ‘top-down’ effect and this is depicted with arrow 1 in Figure 3.4. In this reasoning we could call it the ‘mood’ effect (arrow 2). Affective experience may also gear cognitive appraisal of life at a deeper level. Shared standards of the good life are likely to build on earlier experience of what leads to a satisfying life and in this way connect to human needs (arrow 5). In this view wants will typically be vessels for needs and will ‘false wants’ be an exception rather than the rule. A reversed effect is unlikely; cultural standards of the good life have no influence on innate human needs. Likewise, wider human nature influences wider human culture, or sets at least limits to cultural variation (arrow 6), while human culture does not shape human nature. 3.6.3

Comparison Impinges on Affect

An extreme version of cognitive theory holds that hedonic affect is entirely due to goal attainment; whatever that goal is (for example, Oatley, 1992). This is clearly not true, not only because we can feel good or bad for no apparent reason, but also because not all achievements are equally satisfying (cf. Kasser and Ryan, 1993).

64


Still we do react affectively on meeting some standards and this is particularly true for meeting standards of performance. Possibly this is mainly due to the gratification of related needs, such as the needs for self-esteem and self-actualization, but we cannot rule out that the meeting of the standard in itself also generates positive affect. Affective reactions to comparison are particularly likely in the case of meeting standards in the eyes of others. Like other social animals we seem to have an innate need for acceptance by the congeners around us (cf. Maslow’s need for social respect) and one can well imagine why such a need has developed in evolution. If so, we are likely to enjoy meeting the shared standards of performance, even if the performance itself is not gratifying. This is another instance where needs and wants overlap and this effect is denoted with arrows 3 and in Figure 3.4.

3.7

WHY AFFECTIVE INFERENCE DOMINATES

The theories discussed above are not mutually exclusive. In assessing how happy we are we may draw on both affective experience and cognitive evaluation and it is also possible that we tend to stick to an idea of how happy we are once we have made up our mind. Still it seems to me that the reading of affects dominates the evaluation of life. There are four reasons to think so. (1) Affect theory does best as a complete explanation, while the other two theories rather depict an aspect of the appraisal process. (2) Affect theory fits in better to the other theories than reversely. (3) Affect theory is the most plausible in an evolutionary perspective. (4) Affect theory fits better to the available data. 3.7.1

Affect Theory Provides the Most Complete Explanation

There is probably some truth in all three of these theories; they have all intuitive appeal and supportive evidence. That is not to say that they qualify as a major explanation of how we appraise how happy we are, the theories may merely highlight an aspect of the mental process. As we have seen above, set-point theory highlights the tendency to stick to a particular view, unless circumstances urge to a re-evaluation. This is a common heuristic that operates also in other attitudes. I see this as a minor process and not as a main way of appraising satisfaction with life. If taken as the main mechanism, this theory debouches in absurdities, such as that happiness is insensitive to actual weal and woe. In the same vein, cognitive theory can be seen to highlight a part of the appraisal process and in particular the ‘checking’ of intuitive affective


65

overall appraisal by conscious judgments of success on specific criteria. Cognitive theory cannot explain very well how one could calculate an overall evaluation, since clear priorities are mostly lacking. Cognitive theory can neither explain very well how standards come about and why there is so much resemblance in standards across cultures. If taken as the only way in which we assess how happy we are, cognitive theory leads into absurd conclusions, such as that we can be happy in Hell. Affective inference is more likely to function as the main manner of assessing satisfaction with life, in particular in combination with the assumption that hedonic level of affect reflects need gratification. It is hard to imagine how one could assess ‘satisfaction’ with life without considering how well one feels most of the time and the assumption that that heuristic dominates does not lead into bizarre consequences. 3.7.2

Affects Influence Set-point and Comparison More Than Reversely

In Section 3.6 I discussed the interrelations between the three theories and noted that affective experience is likely to play a major role in the development of attitudes to one’s life that is in the crystallization of set-points. The reverse effect is less likely, though an established attitude can influence the reading of our affects, it is unlikely to mold affective experience as such. In Section 3.6 I also argued that affective experience is likely to influence the cognitive appraisals of life. I distinguished four levels at which affects influence cognitive appraisal and claimed that only on one of these levels there is a comparable influence of cognition (cf. Figure 3.4). If so, affective experience is the most dominant force. 3.7.3

Affect Theory is Most Plausible in Evolutionary Perspective

Another way of appraising the plausibility of theories is considering how well they fit the wider perspective that humans are a product of evolution and that many human behaviors are typically functional in some way. Above I have considered all three theories in that light, below is a summary. Applies not only to human adults As discussed above, affective theory applies to all human beings, and possibly higher animals, while cognitive theory and set-point theory apply only to thinking beings. That would mean that the assessment of happiness changes profoundly when we grow up and that it changes again when we get demented. This does not seem probable to me. I can imagine that the development of abstract thinking adds something to the process

66


of evaluation, but not that the affective information stream is turned off. Neither do I buy the implication that the happiness of children does reflect need gratification and the happiness of adults not. Functional This brings us to the wider point of adaptive significance. As noted above, set-point theory implies that there is no information value in happiness, since happiness is a fixed mind-set and not responsive to objective thriving. To a lesser extend this is also implied in cognitive theory, in which effects of improvement or deterioration are short-lived and where standards of comparison root in collective beliefs that vary across cultures. This boils down to the conclusion that happiness does not matter and that conclusion is absurd. Affective theory sees happiness as a reflection of need gratification and this makes more sense, especially in the context of a functional view on human consciousness and motivation. 3.7.4

Affect Theory Fits the Available Data Best

There is no direct evidence for the dominance of affective inference in the evaluation of life. Though this can be tested to some extent, nobody has done so as yet, at least not to my knowledge. The wider theory that hedonic balance of affect reflects need gratification can hardly be tested at all, since we cannot measure needs very well and particularly not psychological needs. Still there are several pieces of indirect evidence, most of which are already mentioned above. Primacy of affect A point, not yet mentioned above, is that evaluation appears to be an affective process in the first place. In a classic paper Zajonc (1984) has shown that affective appraisal precedes cognitive evaluation. Likewise, Damasio (1994) has shown that injuries in the parts of the brain where affects are processed leave patients unable to make choices, even when their thinking is still intact. This is another indication that cognition has not replaced affect in human evolution and that cognitive appraisals play at best an additional role in the assessment of happiness. Happiness linked to actual thriving Set-point theory and cognitive theory imply that we can evaluate life positively while doing badly from a biological-adaptive point of view. Affect theory rather holds that happiness reflects how well life fits the demands implied in human nature. This latter view is confirmed in two pieces of evidence. People tend to be happier when living in favorable


67

conditions rather than in misery and happiness goes together with mental and physical health. Universal conditions for happiness Cognitive theory implies that conditions for happiness can differ wildly across cultures; while affect theory rather predicts that there will be much similarity in conditions for happiness. This latter point is confirmed in two lines of research. First, comparison of average happiness across nations has shown that 75 percent of the differences can be explained with the same societal characteristics. Second, analysis within nations shows striking similarities all over the world, for instance, being married appears to go with greater happiness all over the world. This point is discussed in more detail in Veenhoven (2010).

3.8

CONCLUSION

There are different theories of how we assess how happy we are: (1) the theory that we echo an earlier evaluation and try to maintain it, (2) the theory that we calculate happiness constantly by comparing life as it is with standards of how life should be, and (3) that we infer happiness from ongoing affective experience and that this affective experience reflects need gratification. These three theories are not mutually exclusive but may differ in import. Affective inference seems to dominate the appraisal of life.

NOTES 1. I thank Mark Chekola for his valuable comments. 2. In the ‘core-affect’ variant happiness set-point is an affective phenomenon, be it a matter of non-responsive affect. Likewise, in the ‘outlook’ variant set-point is a cognitive phenomenon and a tendency to see the glass as either half full or half empty. 3. In this view ‘subjective’ happiness results from the cognitive processing of this affective information.

REFERENCES Annas, J. (2004), ‘Happiness as achievement’, Deadalus: Journal of the American Academy of Arts and Science, 133, 44–51 Bentham, J. (1789), Introduction to the Principles of Morals and Legislation, London: Payne. Chekola, M.G. (1974), ‘The concept of happiness’, PhD Dissertation, University of Michigan, Ann Arbor, USA.

68


Cummins, B., E Gullone and L.A. Lau (2002), ‘A model of subjective well-being homeostasis: the role of personality’, in E. Gullone and R.A. Cummins (eds) The Universality of Subjective Well-being Indicators, Netherlands: Kluwer, pp. 7–46. Damasio, A. (1994), Descartes Error, New York: Putman. Diener, E., W. Pavot and E Sandwick (1991), ‘Happiness is the frequency, not intensity, of positive versus negative affect’, in F. Strack, M. Argyle and N. Schwarz (eds), Subjective Well-Being, Oxford: Pergamon. Diener, E., E. Suh, and S. Oishi (1997), ‘Recent findings on subjective wellbeing’, Indian Journal of Clinical Psychology, 24, 25–41. Diener, E., R.E. Lucas and C.N. Scollon (2006), ‘Beyond the hedonic treadmill: revising the adaptation theory of wellbeing’, American Psychologist, 61, 305–14. Ehrhardt, J.J., W.E. Saris and R. Veenhoven (2000), ‘Stability of life-satisfaction over time: analysis of change in ranks in a national population’, Journal of Happiness Studies, 1, 177–205 Fordyce, M.W. (1972), ‘Happiness, its daily variation and its relation to values’, PhD Dissertation, US International University, San Diego, CA. Fredrickson, B.L. (2004), ‘The broaden-and-build theory of positive emotions’, Philosophical Transactions, Biological Sciences, 359, 1367–77. Freedman, J.L. (1978), Happy People, New York: Harcourt Brace Jovanovich. Gardner, J. and A. Oswald (2001), ‘Does money buy happiness? A longitudinal study using data on windfalls’, Working Paper, University of Warwick. Headey, B., R. Veenhoven and A. Wearing (1991), ‘Top-down versus bottom-up: theories of subjective well-being’, Social Indicators Research, 24, 81–100. Headey, B. and A. Wearing (1992), Understanding Happiness; A theory of Subjective Well-being, Melbourne, Australia: Longman Cheshire. Inglehart, R. (1990), Culture Shift in Advanced Industrial Society, Princeton: Princeton University Press. Kahneman, D. (1999), ‘Objective happiness’ in D. Kahneman, E. Diener and N. Schwarz (eds), Well-Being: The Foundations of Hedonic Psychology, Russell Sage Foundation, New York, pp. 3–25. Kahneman, D. (2000), ‘Experienced utility and objective happiness: a moment based approach’, in D. Kahneman and A. Tverski (eds), Choices, Values and Frames, New York: Cambridge University Press. Kasser, T. and R.M. Ryan (1993), ‘The dark side of the American dream, correlates of financial success as a central life aspiration’, Journal of Personality and Social Psychology, 65, 410–22 Lieberman, L.R. (1970), ‘Life satisfaction in the young and the old’, Psychological Reports, 27, 75–9. Lykken, D.T. (1999), Happiness: What Studies on Twins Show Us About Nature, Nurture and the Happiness Set-point, New York: Golden Books. McDowell, I. and C. Newell (1987), Measuring Health: A Guide to Rating Scales and Questionnaires, New York: Oxford University Press. Michalos, A.C. (1985), ‘Multiple Discrepancies Theory (MDT)’, Social Indicators Research, 16, 347–413. Oatley, K. (1992), Best Laid Schemes: The Psychology of Emotions, Cambridge: Cambridge University Press. Schmitz, O.A. (1930), ‘Glück und lebenskunst (Happiness and the art of living)’, Psychologische Rundschau, 2, 233–8.


69

Schwarz, N. and F. Strack (1991), ‘Evaluating one’s life: a judgment model of subjective well-being’, in F. Strack, M. Argyle and N. Schwarz (eds), Subjective Well-being, Oxford: Pergamon, pp. 27–47. Shin, D. and D.M. Johnson (1978), ‘Avowed happiness as the overall assessment of the quality of life’, Social Indicators Research, 5, 475–92. Sumner, L.W. (1996), Welfare, Happiness and Ethics, New York: Oxford University Press. Unger, H.E. (1970), ‘The feeling of happiness’, Psychology, 7, 27–33. Veenhoven, R. (1984), Conditions of Happiness, Dordrecht: Reidel. Veenhoven, R. (1991), ‘Is happiness relative?’, Social Indicators Research, 24, 1–34. Veenhoven, R (1994), ‘Is happiness a trait? Tests of the theory that a better society does not make people any happier’, Social Indicators Research, 32, 101–60. Veenhoven, R. (1995), ‘The cross-national pattern of happiness. Test of predictions implied in three theories of happiness’, Social Indicators Research, 34, 33–68. Veenhoven, R. (1997), ‘Advances in the understanding of happiness’, (in French), Revue québécoise de psychologie, 18(2), 29–74. Veenhoven, R. (2000a), ‘The four qualities of life: ordering concepts and measures of the good life’, Journal of Happiness Studies, 1, 1–39. Veenhoven, R. (2000b), ‘Are the Russians as unhappy as they say they are? Comparability of self-reports across nations’, Journal of Happiness Studies, 2, 111–36. Veenhoven, R. (2004), ‘Happiness as an aim in public policy: the greatest happiness principle’, in P.A. Linley and S. Joseph (eds), Positive Psychology in Practice, New York: Wiley, pp. 658–78. Veenhoven, R. (2005), ‘Is life getting better? How long and happily people live in modern society’, European Psychologist, 10, 330–43. Veenhoven, R. (2008), ‘Healthy happiness: effects of happiness on physical health and the consequences for preventive health care’, Journal of Happiness Studies, 9, 449–64. Veenhoven, R. (2010), ‘How universal is happiness?’, in E. Diener, J.F. Helliwell and D. Kahneman (eds), International Differences in Well-being, Oxford: Oxford University Press. Veenhoven, R. and M. Hagerty (2006), ‘Rising happiness in nations, 1946–2004. A reply to Easterlin’, Social Indicators Research, 77, 1–16. Wessman, A.E. and D.F. Ricks (1966), Mood and Personality, New York: Holt, Rinehart and Wilson. Zajonc, R.B. (1984), ‘On the primacy of affect’, American Psychologist, 39, 117–23.

4.

Happiness and domain satisfaction: new directions for the economics of happiness Richard A. Easterlin and Onnicha Sawangfa1

The purpose of this chapter is to see to what extent the domain satisfaction model of psychology explains four different patterns of happiness in the USA: (1) the positive cross-sectional relation of happiness to socioeconomic status, (2) the nearly horizontal time series trend, (3) the hill pattern of life cycle happiness, and (4) the decline across generations. The domain model sees each of these happiness patterns as the net result of the corresponding patterns of satisfaction that people have in each of several realms of life – in the present analysis, finances, family life, work and health. These domain satisfaction patterns do not simply replicate the happiness pattern – with regard to age, for example, happiness may go up, but satisfaction with finances, down. Thus, given that the domain satisfaction patterns may differ from that for happiness, and also among themselves, the questions of interest here are specifically the following. Do the patterns by socio-economic status of satisfaction with each of the following – finances, family life, work and health – come together in a way that predicts the positive cross-sectional relation of happiness to socioeconomic status? Do the life cycle patterns of satisfaction in each of these four domains account for the hill pattern of life cycle happiness? Do the time series trends in satisfaction with finances, family life, work and health explain the nearly horizontal time series trend in happiness? Finally, is the decline in happiness across cohorts the net outcome of the cohort patterns of satisfaction in each of the four domains?

4.1

CONCEPTUAL FRAMEWORK

Economists typically adopt the view that well-being depends on actual life circumstances, and that one can safely infer well-being simply from observing these circumstances. The influence of this view is apparent even 70

Happiness and domain satisfaction

71

in the burgeoning literature on the economics of happiness where, despite frequent acknowledgment of subjective factors, many studies consist mainly of regressing happiness on an array of objective variables – income, work status, health, marital status and the like. (See the surveys in Clark et al., 2008; DiTella and MacCulloch, 2006; Frey and Stutzer, 2002a and 2002b; Graham, 2005 and 2009; Layard, 2005.) Psychologists, in contrast, view the effect on well-being of objective conditions as mediated by psychological processes through which people adjust somewhat to ups and downs in their life circumstances. Their skepticism of the economists’ view is well represented by psychologist Angus Campbell’s complaint over three decades ago: ‘I cannot feel satisfied that the correspondence between such objective measures as amount of money earned, number of rooms occupied, or type of job held, and the subjective satisfaction with these conditions of life, is close enough to warrant accepting the one as replacement for the other’ (Campbell, 1972, p. 442; cf. also Lyubomirsky, 2001). This statement appeared in a volume significantly titled The ‘Human’ Meaning of Social Change (emphasis added). In contrast to economists’ focus on objective conditions, Campbell proposed a framework in which objective conditions were replaced by reports on the satisfaction people expressed with those conditions (Campbell et al., 1976; Campbell, 1981). This approach is sometimes termed multiple discrepancy theory (Michalos, 1986, 1991; cf. also Diener et al., 1999b; Solberg et al., 2002). In this framework global happiness or overall satisfaction with life is seen as the net outcome of reported satisfaction with major domains of life, such as financial situation, family life and so on. Satisfaction in each domain is, in turn, viewed as reflecting the extent to which objective outcomes in that domain match the respondent’s goals or needs in that area, and satisfaction may vary with changes in goals, objective conditions or both. In economics similar models comparing attainments to aspirations date back to March and Simon (1968); for a recent example, see de la Croix (1998). An advantage of this approach is that judgments on domain satisfaction reflect both subjective factors of the type emphasized in psychology and objective circumstances stressed by economics. In the domain of family life, for example, one’s goals, simply put, might be a happy marriage with two children and warm family relationships. Satisfaction with family life would reflect the extent to which objective circumstances match these goals – the greater the shortfall, the less the satisfaction with family life. Over time, subjective goals, objective circumstances or both may change, and thereby alter judgments on domain satisfaction. Given objective conditions, goals may be adjusted to accord more closely with actual circumstances, in line with the process of hedonic adaptation emphasized by

72


psychologists. Given goals, objective circumstances may shift closer to or farther from goals, altering satisfaction along the lines stressed by economists. Thus, in contrast to the objective measures used in economic models – in the case of family life, such things as marital status and number of children – reports on satisfaction with family life reflect the influence of subjective norms as well as objective circumstances. Another advantage of Campbell’s domain approach is that it classifies into a tractable set of life domains the everyday specific circumstances to which people refer when asked about the factors affecting their happiness (Cantril, 1965; Kahneman et al., 2004; Robinson and Godbey, 1997, ch.17). Of course, there is not complete agreement on what domains of life are conceptually preferable, and the classification of life domains remains a subject of continuing research. Virtually all life domain studies agree, however, that four domains are of major importance – finances, family circumstances, health and work. These four, for example, with slightly different labels, are at the head of Cummins’s (1996) meta-analysis of the domains of life satisfaction. It is these four that are studied here as predictors of the patterns of happiness by socio-economic status, time, age and cohort.

4.2

PRIOR WORK

Economic research on domain satisfaction has heretofore been quite limited, and much of what has been done focuses on explaining, not overall happiness, but satisfaction with specific economic circumstances, for example, job satisfaction, housing satisfaction, financial satisfaction, satisfaction with income, satisfaction with standard of living and so on (Diaz-Serrano, 2006; Hayo and Seifert, 2003; Hsieh, 2003; Solberg et al., 2002; Vera-Toscano et al., 2006; Warr, 1999). Few studies explore the relation of global happiness to the different domains. An important exception is the work of van Praag and Ferrer-i-Carbonell (2004), which examines the extent to which differences among individuals in overall satisfaction are related to satisfaction with a variety of life domains, several of which correspond to those studied here (see Chapters 3 and 4; also van Praag et al., 2003). Their results, based on data for the UK and Germany, support the importance of the domains studied here, and suggest that domain satisfaction variables provide a better statistical explanation of happiness than objective conditions (for a similar result with regard to wages and hours of work, see Clark, 2005). In another interesting study Rojas (2007) uses the domain satisfaction approach to study individual happiness in Mexico, focusing on domains deriving from the philosophical rather than social science literature. In a recent article Easterlin (2006) uses


73

the domain approach to study happiness over the life cycle. The present chapter extends the Easterlin analysis to the variation in happiness by socio-economic status, time and birth cohort. Outside of economics work relating happiness to domain satisfaction is more extensive (see, for example, the bibliography in Veenhoven, 2005, section 12-a). One of the most ambitious projects brings together studies of individual data for 12 European countries of both domain satisfaction and satisfaction with life in general (Saris et al., 1996). The domains vary somewhat among countries, but one result common to all countries is that two domains are consistently positively related to overall life satisfaction – material living conditions (captured in satisfaction with housing and satisfaction with finances) and ‘social contacts’, reflecting the importance to well-being of personal relationships (ibid., p. 227; on personal relationships and well-being, see Ryff, 1995; Ryan and Deci, 2000). The counterparts of these two in the present study are financial satisfaction and satisfaction with family life. All of these earlier studies, both within and outside of economics, focus on explaining how happiness varies in relation to one particular variable, usually among individuals at a point in time. In contrast, the aim here is to test how well the domain satisfaction approach explains mean happiness within the US population in relation to each of four different variables – by socio-economic status (education), over time (year), over the life cycle (age) and across generations (birth cohort). For each variable the test is the same, to see how well the actual relation of happiness to that variable can be predicted from the corresponding patterns for the four domain satisfaction variables – financial situation, family life, work and health.

4.3

DATA AND METHODS

The data are from the US General Social Survey (GSS) conducted by the National Opinion Research Center (Davis and Smith, 2002). This is a nationally representative survey conducted annually from 1972 to 1993 (with a few exceptions) and biannually from 1994 to 2006. The present analysis is based on data for 1973–94, because two of the variables of interest, family and health satisfaction, are included in the GSS only during this time span. The GSS is a survey of households, and weighted responses are used here to represent more accurately the population of persons (ibid., pp. 1392–3 of codebook). For happiness there are three response options; for financial satisfaction, also three options; for job satisfaction (including housework), four options; family satisfaction, seven options; and health satisfaction, seven

74


options. The specific question for each variable is given in Appendix 4.A1. In the present analysis the response of an individual to each question is assigned an integer value, with a range from least satisfied (or happy) equal to 1, up to the total number of response options (for example, 3 for happiness, 7 for health satisfaction). Socio-economic status is measured by years of schooling, ranging from zero to 20. The age range is from 18 to 89; birth cohort, from 1884 to 1976. Year is in terms of time dummies with 1973 being the reference year. The use of time dummies enables us to separate period from age and cohort effects (cf. Blanchflower and Oswald, 2007). Descriptive statistics are given in Appendix 4.A2. The basic procedure consists of the following steps: 1.

A regression of happiness on age, cohort, education, gender, race and year (in dummy form) is estimated from the individual data for 1973–94 (see Appendix 4.A3, column 1). Both linear and quadratic forms are tried for the age, cohort and education variables, and the form yielding the best fit in terms of significant t-statistics is selected. This single regression is then used to estimate how happiness varies in relation to each of the four variables (age, cohort, education and year) controlling for the other three. We call these estimated values ‘actual happiness’. Actual happiness differs from the raw mean of happiness in relation to any given variable, such as education, in that it controls for other, essentially fixed, characteristics of individuals. Thus, in the case of education, the estimated happiness-education pattern controls for differences from one level of education to another in the composition of the population by age, cohort, gender and race, and also in period effects. Formally, we have: Happy 5 C(1)(age, cohort, education, gender, race, year dummy), where C(1) denotes Equation (4.1) in Appendix 4.A3. The typical pattern of variation of happiness in regard to a given variable, say, years of schooling, is then estimated by entering in this regression the mean values of the other variables (age, cohort, gender, race and year), while allowing that variable to range from its minimum to maximum value as given in Appendix 4.A2 (for education, from zero to 20 years of schooling). Thus, Happy for the ith year of schooling 5 C(1)(age, cohort, male, black, educationi, t1973, t1974, . . ., t1994)


75

where age is the mean age, cohort is the mean birth cohort and so on. Since years of schooling range from 0 to 20, following the above procedure for each level of education results in a series: Happy_Ed(0), Happy_Ed(1), . . ., Happy_Ed(20), where Happy_Ed(j) is actual happiness of a person with j years of schooling. This series is plotted in Figures 4.1(a) and 4.2(a) as actual happiness, the happiness pattern that is to be predicted. 2. A similar procedure is followed to derive the typical pattern of variation of satisfaction in each of the four domains. First, a regression is estimated of satisfaction in a given domain in relation to age, cohort, education, gender, race and year (in dummy form) as presented in Appendix 4.A3, columns 2–5. Then, the typical pattern of variation of satisfaction in that domain in regard to a given variable, say, education, is estimated by entering in the regression the mean values of all other variables, while allowing that variable to range from its lowest to highest value as given in Appendix 4.A2. Thus, similarly to the computation of actual happiness by level of education, the series of actual domain satisfaction by education can be obtained from C(2), C(3), C(4) and C(5), respectively: (a) (b) (c) (d)

3.

Satfin_Ed(0), Satfin_Ed(1), . . ., Satfin_Ed(20); Satfam_Ed(0), Satfam_Ed(1), . . ., Satfam_Ed(20); Satjob_Ed(0), Satjob_Ed(1), . . ., Satjob_Ed(20); and Sathealth_Ed(0), Sathealth_Ed(1), . . ., Sathealth_Ed(20).

These series are plotted in Figure 4.3(a). A regression is estimated from the individual data of the relation of happiness to the four domain satisfaction variables – financial satisfaction (Satfin), family satisfaction (Satfam), work satisfaction (Satjob) and health satisfaction (Sathealth) – to establish the relative impact of each domain on happiness (Appendix 4.A4). Formally, Happy 5 D(Satfin, Satfam, Satjob, Sathealth). All domains turn out to have a significant positive effect on happiness, as one might expect, with family and financial satisfaction having the greatest weight. Although there is some variation in the domain weight by demographic characteristics, such as sex and age, they are not sizeable enough to alter the basic results obtained here.

76

4.


A prediction of the variation of happiness with regard to each variable (education, time, age and cohort) is obtained by substituting in the step 3 regression equation the domain satisfaction values estimated in step 2. For the cross-section analysis, for example, predicted happiness for a given education level is estimated by entering in the step 3 regression equation the four domain satisfaction values for that level of education derived in step 2. Thus, mean predicted happiness for zero years of schooling, controlling for age, cohort, gender, race and year, is computed as: PredHappy_Ed(0) 5 D(Satfin_Ed(0), Satfam_Ed(0), Satjob_Ed(0), Sathealth_Ed(0)). Similarly, PredHappy_Ed(1) 5 D(Satfin_Ed(1), Satfam_Ed(1), Satjob_Ed(1), Sathealth_Ed(1)). This procedure is repeated for all other levels of education to obtain the predicted pattern of happiness in relation to education. The series PredHappy_Ed(0), PredHappy_Ed(1), . . ., PredHappy_Ed(20) is then plotted as predicted happiness in Figure 4.2(a).

The regression technique used is ordered logit, because responses to the several variables are categorical and number three or more. Ordinary least squares regressions yield virtually identical results, suggesting that the findings are robust with regard to methodology. In step 3, in estimating the relation of happiness to domain satisfaction from individual data, a question arises about possible bias in reports on satisfaction (cf. Diener and Lucas, 1999a, p. 215; van Praag and Ferreri-Carbonell, 2004, ch. 4). Responses on satisfaction – whether with life in general or an individual domain – are known to be influenced by personality traits. Consider two persons with identical objective conditions and subjective goals. If one of them is neurotic, then it is likely that this person’s responses on satisfaction with both life in general and the various domains of life will be lower than the other person’s, because neurotics tend to assess their circumstances more negatively than others (Diener and Lucas, 1999a). However, a purpose of the step 3 regression is to establish


77

the relative weights in determining happiness of the four domain satisfaction variables. Because the happiness and domain satisfaction responses for any given individual would be similarly biased by personality, the estimate of relative weights for that individual, and correspondingly for the population as a whole, should be free of personality bias. Another purpose of the step 3 regression is to predict actual happiness from domain satisfaction. If personality bias exists in an individual’s report on happiness, then actual happiness, which is based on this report, is biased by personality. Similarly, personality bias in an individual’s report on domain satisfaction leads to personality bias in actual domain satisfaction. Predicted happiness, derived from actual domain satisfaction, is then also biased by personality. But since the personality bias in actual happiness is the same as that in actual domain satisfaction and, thus, in predicted happiness, the predictive power of the domain model judged by comparing actual happiness with predicted happiness should not be influenced by personality bias.

4.4

RESULTS

Actual Happiness The happiness patterns to be explained are both familiar and unfamiliar. Most familiar, perhaps, is the positive cross-sectional association of happiness to socio-economic status (Figure 4.1a). Also well-known is the fairly flat relation of happiness to time (Figure 4.1b). (The fluctuations in the figure are due to the use of time dummies.) Less familiar are the patterns in relation to age and cohort. Over the life cycle happiness rises slightly to mid-life and declines slowly thereafter (Figure 4.1c; cf. also Easterlin, 2006; Mroczek and Spiro, 2005). Although the swing in happiness is mild, it is statistically significant. The pattern differs from the usual U-shaped relation to age reported in the economics literature, because the U-shaped happiness-age relation is the result of a multivariate regression in which controls are included, not only for the variables used here (education, time, cohort, gender, race), but also for life circumstances (income, work status, marital status, health) (Blanchflower and Oswald, 2004, 2007). Controls for life circumstances would be inappropriate here and also in regard to the happiness patterns for education, time and cohort, because the specific purpose of the analysis is to test whether satisfaction with life circumstances, which reflects both objective life circumstances and subjective norms, explains the happiness patterns observed.

2.5

18

0

24

2

30

4

36

6

10

12

42

48 54 Age

60

(c) By Age

14

66

Years of Schooling

8

(a) By Years of Schooling

72

16

78

84

18

90

20

1884 1894 1904 1914 1924 1934 1944 1954 1964 1974 Cohort

(d) By Birth Cohort

Year

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993

(b) By Year

Mean actual happiness by years of schooling, year, age and birth cohort, 1973–94

Values in each panel are after controlling for the three variables heading the other panels, and also gender and race. See Appendix 4.A3.

Figure 4.1

Note:

Mean Happiness

2.4

2.3

2.2

2.1

2

2.5

2.4

2.3

2.2

2.1

2

Mean Happiness

2.5 2.4 2.3 2.2 2.1 2 2.5 2.4 2.3 2.2 2.1 2

78

2.5

2.4

2.3

2.2

2.1

2

4

6

Predicted

Actual

(c) By Age

8 10 12 14 Years of Schooling

16

18

20

Predicted

Actual

18 24 30 36 42 48 54 60 66 72 78 84 90 Age

0

(a) By Years of Schooling

Predicted

Actual

1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 Year

(b) By Year

(d) By Cohort

1884 1894 1904 1914 1924 1934 1944 1954 1964 1974 Cohort

Predicted

Actual

Mean predicted and actual happiness by years of schooling, year, age and birth cohort, 1973–94

Values in each panel are after controlling for the three variables heading the other panels, and also gender and race. See Appendix 4.A3.

Figure 4.2

Note:

Mean Happiness

Mean Happiness

2

2.5

2.4

2.3

2.1

2

2.2

2.5 2.4 2.3 2.1 2 2.1 2

2.2

2.3

2.4

2.5

2.2

2.5 2.4 2.3 2.1 2 2.1 2

2.2

2.3

2.4

2.5

2.2

2.5 2.4 2.3 2.2 2.1 2 2.5 2.4 2.3 2.2 2.1 2

79

Figure 4.3

4.7

5

80 Satfam

Satfam

5.3 5.6 5.9 6.2 6.5 6.8

4.7

5

5.3 5.6 5.9 6.2 6.5 6.8 Satfin

2.1 2.2 2.3 2.4

1.7 1.8 1.9

2

2.1 2.2 2.3 2.4

0 2

SATFIN

4

1977

SATFAM

HAPPY

1973

2

1981 1985

Year

(b.1) By Year

1989

20

1993

SATFIN

18

SATFAM

16

HAPPY

6 8 10 12 14 Years of Schooling

(a.1) By Years of Schooling

(continues on p. 81)

Satfin 1.7 1.8 1.9

1.7 1.8 1.9

2

1.7 1.8 1.9

2.1 2.2 2.3 2.4

Happy

2 2.1 2.2 2.3 2.4 Happy Satjob

Satjob 2.5

2.71

2.92

3.13

3.34

2.5

3.55

2.71

2.92

3.13

3.34

3.55

1973

0 2

SATJOB

1977

4 6

SATHEALTH

1981 1985

Year

SATHEALTH

16

1989

SATJOB

(b.2) By Year

8 10 12 14 Years of Schooling

(a.2) By Years of Schooling

20

1993

18

4

4

4.3 4.6 4.9 5.2 5.5 5.8 6.1

4.3

4.6

4.9

5.2

5.5

5.8

6.1

Sathealth

Sathealth

2.1 2.2 2.3 2.4 2.5

2

Satfam

7

SATFIN

SATFAM

HAPPY

(c.1) By Age

SATFIN

SATFAM

HAPPY

SATJOB

SATHEALTH

1884 1894 1904 1914 1924 1934 1944 1954 1964 1974 Cohort

(d.2) By Birth Cohort

18 24 30 36 42 48 54 60 66 72 78 84 90 Age

SATJOB

SATHEALTH

(c.2) By Age 5.8 6.1

Mean domain satisfaction and actual happiness by years of schooling, year, age and birth cohort, 1973–94

Cohort

1884 1894 1904 1914 1924 1934 1944 1954 1964 1974

(d.1) By Birth Cohort

18 24 30 36 42 48 54 60 66 72 78 84 90 Age

See note to Figure 1.

Figure 4.3

Note:

Satfin

Satfin

1.8 1.9

2.1 2.2 2.3 2.4 2.5

4.9 5.2 5.5 5.8 6.1 6.4 6.7

7

6.7

6.4

Satfam 5.8 6.1

5.5

5.2

4.9

1.8 1.9 1.9

2

2.1

2.2

2.3

2.4

2.5

2

2.1 2.2 2.3 2.4 2.5 Happy Happy

3.8 3.59 3.38 3.17 2.96 2.75

1.8

2

1.8 1.9

Satjob Satjob

3.8 3.59 3.38 3.17 2.96 2.75

Sathealth Sathealth

5.2 5.5 4.3 4.6 4.9 4 5.8 6.1 5.5 5.2 4.9 4.6 4.3 4

81

82


Least studied is how happiness varies by cohort.2 For cohorts born between the late nineteenth century and the 1970s, the relation of happiness to cohort is negative and curvilinear, with the lowest happiness levels found in the cohorts born in the mid-1950s (Figure 4.1d). Thus, the happiness of younger cohorts is, on average, significantly less than older, except that among the most recent cohorts there is a slight upturn in happiness. The magnitude of the happiness differences among cohorts is not very great, but it is somewhat larger than the changes found in the time series and life cycle patterns. The difference among cohorts found here is after controlling for cohort differences in age, education, period influences, gender and race. If data were available for a single year only, then it would be impossible to distinguish the cohort pattern from that for age. If, for example, in the survey year 1980 mean happiness were to increase from age 20 (that is, persons born in 1960) to age 80 (persons born in 1900), then the cohort pattern would be negative, the reverse of that for age, with happiness declining from the cohort of 1900 to that of 1960. (If the age pattern were hill-shaped moving from left to right on the x-axis, the cohort pattern would be hillshaped too – in effect, the cohort pattern traverses the same hill in reverse fashion.) With data for only one year, there would be no way of deciding whether one is observing the relation of happiness to age or to cohort. Our data, however, span 21 years, and thus in deriving cohort effects we compare the happiness of 21 different cohorts at a given age, and, correspondingly, in deriving age effects the happiness at 21 different ages of a given cohort. The fact that our age and cohort patterns of happiness are not simply the reverse of each other (as is true also of the age and cohort patterns for the individual domains) indicates that we are successfully differentiating between age and cohort influences. 4.4.1

Predicted Happiness

There are four fairly disparate patterns of happiness to be explained – a positive cross-sectional relation to education, a fairly flat relation to time, the ‘hill’ pattern of the life cycle and a negative curvilinear relation across cohorts. How well do the domain satisfaction patterns predict these patterns of happiness? The answer, based on the procedures outlined in steps 2–4 above, is reasonably well. The cross-sectional relation of happiness to education is closely predicted by the cross-sectional patterns of happiness to education derived from the domain model (Figure 4.2a). The predicted time series pattern of happiness based on the time series patterns of satisfaction in each domain corresponds closely to the actual horizontal time series


Table 4.1

83

Mean squared error of the prediction of happiness

Variable Years of Schooling Year Age Birth Cohort

Mean Squared Error 0.00 033 0.00 054 0.00 028 0.00 152

pattern (Figure 4.2b). The life cycle happiness predicted by the life cycle patterns of domain satisfaction follow the ‘hill’ pattern of actual happiness, although the predicted movement peaks slightly earlier, at age 43 compared with 52, and the amplitude is slightly less than the actual (Figure 4.2c). Least satisfactory is the prediction of the cohort pattern. Although happiness of younger cohorts is correctly predicted to be less than older, the predicted curve is virtually linear rather than concave upward, so that the upturn among the youngest cohorts is missed (Figure 4.2d). Table 4.1 compares the mean squared error of the prediction of happiness by education, year, age and cohort. The most satisfactory prediction is that for life cycle happiness. It is closely followed by the predictions for years of schooling and the time series pattern of happiness. Confirming the visual observation of Figure 4.2, the least satisfactory is the prediction of the cohort pattern, with a mean squared error more than five times that of the life cycle prediction. 4.4.2

Domain Satisfaction

As a general matter, the four domain patterns for any one variable typically differ from each other, and the domains dominating the prediction of happiness are not the same for all four variables. This is brought out in Figure 4.3, which presents for each variable the actual domain patterns and, for comparison, that for actual happiness. The left hand panel presents the patterns for the domains of family life and financial satisfaction; the right hand panel, the patterns for satisfaction with work and health. By comparing the actual domain patterns with that for actual happiness, one is able to form a tentative impression of which domains are chiefly responsible for the happiness pattern for any given variable. Perhaps most striking is that more educated persons are happier because they enjoy greater satisfaction in all four realms of life. For family life, finances, work and health, satisfaction trends upward with the level of education (Figure 4.3, panels a.1 and a.2). The rate of change, however, varies among the domains. Satisfaction with family life and health

84


increases at a decreasing rate, while satisfaction with finances grows at an increasing rate. Only for satisfaction with work is the trend linear like that for actual happiness. The fairly flat relation of happiness to time appears from the figure to reflect similar patterns in the four domains (panels b.1 and b.2). However, if one fits ordinary least squares trend lines to the fluctuating lines in the figures, some subtle differences emerge. All of the patterns have very slight, but significant trends. Actual happiness has a small uptrend, amounting to a total increase for the period of 0.013 on the happiness scale of one to three. This is the equivalent of a net upward shift by one response category – say, from ‘pretty happy’ to ‘very happy’ – of 1.3 percent of respondents over the entire 21 year period. This is not very much of a shift, although it is statistically significant. Based on the fitted trends, the corresponding shift for each domain (all of them significant) are for financial satisfaction 13.2 percent, work satisfaction 12.3 percent, family life satisfaction −0.5 percent and health satisfaction 11.4 percent. Thus, the very slight uptrend for actual happiness is the net outcome of the slight positive trends in satisfaction with finances, work and health outweighing the slight negative trend in satisfaction with family life. Turning to the age patterns, one finds that the increase to mid-life of life cycle happiness is due to increasing satisfaction with family life and work outweighing negative changes in satisfaction with finances and health (Figure 4.3, panels c.1 and c.2). The decline of happiness beyond mid-life occurs because declines in satisfaction with family life and work join the downtrend in satisfaction with health. The adverse impact on happiness of these negative trends is moderated, however, by increasing satisfaction that people express with their financial situation as they move into older age. Finally, the lower happiness of younger compared with older cohorts is due to downtrends in satisfaction in three domains – finances, work and health (Figure 4.3, panels d.1 and d.2). Satisfaction with family life does not differ between older and younger cohorts, despite the striking differences in family life between today’s cohorts and those of their parents and grandparents. The slight upturn in actual happiness among the youngest cohorts cannot be explained by the domains studied here, because none of the domains shows an improvement of younger relative to older cohorts. One important conclusion that emerges from surveying the domain patterns is that no single domain is the key to happiness. Rather, happiness is the net outcome of satisfaction with all of the major life domains, and the domain patterns frequently differ from each other. Moreover, the importance of any given domain varies depending on the happiness relationship being studied – cross-sectionally by education, over time, through the life cycle or across generations.


4.5

85

CONCLUSIONS

How well does the domain satisfaction model predict the way in which mean happiness varies by socio-economic status, year, age and birth cohort? The answer is quite well for the first three – education, year and age – and not too badly for the fourth, cohort. Some skeptics might say the success of the predictions is no great surprise – due to the common influence of personality, reports of happiness and satisfaction with finances, family life, work and health all are highly correlated in data for individuals. But we are analysing here the relations among group means, not individual data, and as can readily be seen from Figure 4.3 the patterns of the domain means often differ from that for happiness and among themselves. The observation that the happiness and domain satisfaction variables are highly positively correlated among individuals does not imply, for example, that the mean values by age of happiness and the four domain satisfaction variables will also be positively correlated. Personality tends to be stable from one age to another while the means for happiness and domain satisfaction follow different and sometimes contrary paths with regard to age. Similarly, unless one believes that personality varies systematically by socio-economic status, birth cohort and over time, there is no reason to suppose that the individual-based correlation among the happiness and domain satisfaction variables would result in corresponding correlations among the group means for these categories. Put differently, the point is not whether personality varies among individuals in a way that would explain individual differences in happiness, but whether, on average, it varies by age, level of education, year and cohort in ways that would generate the prediction of happiness by these characteristics. To our knowledge there is no research demonstrating that personality varies by age, level of education, year and cohort in the same way as happiness. Some might contend that because actual happiness is derived from the same explanatory variables as are the domain means (Appendix 4.A3), it is inevitable that actual happiness and happiness as predicted by these domain means would come together well. An argument similar to that just given applies here. There is no reason to suppose that a weighted average of the four domain means at, say, a certain age should equal actual happiness at that age. Different life cycle patterns among the four domains need not together produce a life cycle pattern of happiness that fits nicely with the life cycle pattern of actual happiness. It would be interesting to see if happiness regressions of the type found in the economics literature, based only on objective variables, do as well in predicting the happiness patterns as the domain satisfaction variables used

86


here. If, for example, one estimates the life cycle pattern of such ‘objective’ variables as income, marital status, employment status and health, would one be able to predict from these patterns the actual life cycle pattern of happiness? We venture that the answer is no – that Angus Campbell is right when he says that subjective well-being depends not on objective conditions alone, but on the psychological processing of objective circumstances, as captured in reports on satisfaction with these conditions. The fact that the domain patterns studied here come together reasonably well to predict actual happiness provides new support for the meaningfulness of subjective data on well-being and its components. Thus, while a skeptic of the present analysis might point to the startling contrast between the life cycle pattern for happiness and that for satisfaction with finances – almost diametrical opposites – it turns out that when the movements in the other domains are accounted for, along with that for financial satisfaction, the hill pattern observed for actual happiness is predicted fairly closely by the domain patterns. This close prediction would be unlikely to occur if no credence could be given to what people say about their feelings. In addition, the similarity between the present patterns of predicted and actual happiness supports the conclusion that the four domains studied here are probably the most important in determining happiness, a result consistent with the literature on domain satisfaction. But these four domains do not tell the whole story of happiness movements, as is made especially clear here by the disparity between the predicted and actual happiness patterns by birth cohort. Finally, depending on the happiness relationship being studied – by socio-economic status, time, age or cohort – the role played by different domains in determining happiness tends to vary. Happiness is the net outcome of satisfaction with all of the major domains of life, and no single domain is sufficient to explain the various patterns of overall happiness. This is in many ways a first pass at testing fairly comprehensively Campbell’s domain satisfaction model, and while the model performs reasonably well, the results raise a number of questions for further research. For example, would increasing the number of life domains improve the predictions? What chiefly determines the domain satisfaction patterns – objective conditions like those emphasized in economics or subjective factors stressed in psychology? To what extent are there interrelations among the various domains themselves? The domain satisfaction model provides a new and reasonable start on unraveling the mysteries of happiness – a new direction, perhaps, for research on the economics of happiness. But it is only a start.


87

NOTES 1. For valuable comments we are grateful to Andrew Clark, Andrew Oswald, John Strauss, Anke Zimmermann, and participants in the Conference on New Directions in the Study of Happiness: US and International Perspectives, University of Notre Dame, 22–24 October 2006. Laura Angelescu provided excellent research assistance; financial help was provided by the University of Southern California. 2. An exception is the article by Blanchflower and Oswald (2000), which focuses on the trend of happiness among younger persons since 1972. However, their analysis controls for differences among cohorts in life circumstances, whereas the present analysis does not.

REFERENCES Blanchflower, D.G. and A. Oswald (2000), ‘The rising well-being of the young’, in D.G. Blanchflower and R.B. Freeman (eds), Youth Employment and Joblessness in Advanced Countries, Cambridge, MA: NBER, pp. 289–328. Blanchflower, D.G. and A. Oswald (2004), ‘Well-being over time in Britain and the USA’, Journal of Public Economics, 88, 1359–86. Blanchflower, D.G. and A. Oswald (2007), ‘Is well-being U-shaped over the life cycle?’, IZA Discussion Paper No. 3075, Bonn: Institute for the Study of Labor (IZA). Campbell, A. (1972), ‘Aspiration, satisfaction, and fulfillment’, in A. Campbell and P.E. Converse (eds), The Human Meaning of Social Change, New York: Russell Sage, pp. 441–66. Campbell, A. (1981), The Sense of Well-Being in America, New York: McGrawHill. Campbell, A., P.E. Converse, and W.L. Rodgers (1976), The Quality of American Life, New York: Russell Sage Foundation. Cantril, H. (1965), The Pattern of Human Concerns, New Brunswick, NJ: Rutgers University Press. Clark, A.E. (2005), ‘What makes a good job? Evidence from OECD countries in S. Bazen’, in C. Lucifora and W. Salverda (eds), Job Quality and Employer Behaviour, Basingstoke: Palgrave MacMillan, pp. 11–30. Clark, A.E., P. Frijters and M.A. Shields (2008), ‘Relative income happiness and utility: an explanation for the Eastern Paradox and other puzzles’, Journal of Economic Literature, 46 (1), 95–144. Cummins, R.A. (1996), ‘The domains of life satisfaction: an attempt to order chaos’, Social Indicators Research, 38, 303–28. Davis, J.A. and T.W. Smith (2002), General Social Surveys, 1972–2002. University of Connecticut, Storrs, CT: The Roper Center for Public Opinion Research. Machine-readable data file: Principal Investigator, James A. Davis; Director and Co-principal Investigator, Tom W. Smith; Co-principal Investigator, Peter Marsden; National Opinion Research Center, producer; The Roper Center for Public Opinion Research, University of Connecticut, distributor. One data file (43 p 698 logical records and one codebook (1769 pp.). de la Croix, D. (1998), ‘Growth and the relativity of satisfaction’, Mathematical Social Sciences, 36, 105–25.

88


Diaz-Serrano, L. (2006), ‘Housing satisfaction, homeownership and housing mobility: a panel data analysis for twelve EU countries’, IZA Discussion Papers No. 2318, Bonn: Institute for the Study of Labor (IZA). Diener, E. and R.E Lucas (1999a), ‘Personality and subjective well-being’, in D. Kahneman, E. Diener and N. Schwarz (eds), Well-being: The Foundations of Hedonic Psychology, New York: Russell Sage, pp. 213–29. Diener, E., E.M. Suh, R.E. Lucas and H.L. Smith (1999b), ‘Subjective well-being: three decades of progress’, Psychological Bulletin, 125, 276–302. DiTella, R. and R. MacCulloch (2006), ‘Some uses of happiness data in economics’, Journal of Economic Perspectives, 20, 25–46. Easterlin, R.A. (2006), ‘Life cycle happiness and its sources: intersections of psychology, economics and demography’, Journal of Economic Psychology, 27, 463–82. Frey, B.S. and A. Stutzer (2002a), Happiness and Economics, Princeton, NJ: Princeton University Press. Frey, B.S. and A. Stutzer (2002b), ‘What can economists learn from happiness research?’, Journal of Economic Literature, 40, 402–35. Graham, C. (2005), ‘Insights on development from the economics of happiness’, World Bank Research Observer. 1–31. Graham, C. (2009), ‘Happiness, economics of’, in S. Durlauf and L. Blume (eds), The New Palgrave Dictionary of Economics Online, 2nd edn, Basingstoke: Palgrave-Macmillan, last accessed 2 May 2009. Hayo, B. and W. Seifert (2003), ‘Subjective economic well-being in Eastern Europe’ Journal of Economic Psychology, 24, 329–48 Hsieh, C.M. (2003), ‘Income, age and financial satisfaction’, International Journal of Aging & Human Development, 56, 89–112. Kahneman, D., A.B. Krueger, D.A. Schkade, N. Schwarz and A.A. Stone (2004), ‘A survey method for characterizing daily life experience: the Day Reconstruction Method (DRM)’, Science, 306, 1776–80. Layard, R. (2005), Happiness: Lessons from a New Science, New York: Penguin Press. Lyubomirsky, S. (2001), ‘Why are some people happier than others? The role of cognitive and motivational processes on well-being’, American Psychologist, 56 (3), 239–49. March, J.G. and H.A. Simon (1968), Organizations, New York: John Wiley. Michalos, A.C. (1986), ‘Job satisfaction, marital satisfaction and the quality of life’, in F.M. Andrews (ed.), Research on the Quality of Life, University of Michigan, Ann Arbor: Survey Research Center, Institute for Social Research, pp. 57–83. Michalos, A.C. (1991), Global Report on Student Well-being, Vol. I. Life Satisfactions, New York: Springer-Verlag. Mroczek, D.K. and A. Spiro, III (2005), ‘Changes in life satisfaction during adulthood: findings from the veterans affairs normative aging study’, Journal of Personality and Social Psychology, 88 (1), 189–202. Robinson, J.P. and G. Godbey (1997), Time for Life: The Surprising Ways Americans Use Their Time, 2nd edn, University Park, PA: Pennsylvania State University Press. Rojas, M. (2007), ‘The complexity of well-being: a life satisfaction conception and domains-of-life approach’, in I. Gough and J.A. McGregor (eds) Well-being in Developing Countries: From Theory to Research, New York: Cambridge University Press, pp. 259–80. Ryan, R.M. and E.L. Deci (2000), ‘Self-determination theory and the facilitation of


89

intrinsic motivation, social development, and well-being’, American Psychologist, 55 (1), 68–78. Ryff, C.D. (1995), ‘Psychological well-being in adult life’, Current Directions in Psychological Science, 4, 99–104. Saris, W.E., R. Veenhoven, A.C. Scherpenzeel, and B. Bunting (eds) (1996), A Comparative Study of Satisfaction with Life in Europe, Budapest: Eötvös University Press. Solberg, E.C., E. Diener, D. Wirtz, R.E. Lucas and S. Oishi (2002), ‘Wanting, having, and satisfaction: examining the role of desire discrepancies in satisfaction with income’, Journal of Personality and Social Psychology, 83, 725–34. van Praag, B.M.S. and A. Ferrer-i-Carbonell (2004), Happiness Quantified: A Satisfaction Calculus Approach, Oxford: Oxford University Press, ch. 3. van Praag, B.M.S., P. Frijters and A. Ferrer-i-Carbonell (2003), ‘The anatomy of subjective well-being’, Journal of Economic Behavior and Organization, 51, 29–49. Veenhoven, R. (2005), ‘World Database of Happiness’, http://worlddatabaseofhappiness.eur.nl, last accessed 28 April 2009. Vera-Toscano E., V. Ateca-Amestoy and R. Serrano-del-Rosal (2006), ‘Building financial satisfaction’, Social Indicators Research, 77, 211–43. Warr, P. (1999), ‘Well-being and the workplace’, in D. Kahneman, E. Diener and N. Schwarz (eds), Well-being: The Foundations of Hedonic Psychology, New York: Russel Sage Foundation, pp. 392–412.

90


APPENDIX 4.A1

QUESTIONS AND RESPONSE CATEGORIES FOR HAPPINESS AND SATISFACTION VARIABLES

HAPPY: Taken all together, how would you say things are these days – would you say that you are very happy, pretty happy or not too happy? (Coded 3, 2, 1 respectively.) SATFIN: We are interested in how people are getting along financially these days. So far as you and your family are concerned, would you say that you are pretty well satisfied with your present financial situation, more or less satisfied, or not satisfied at all? (Coded 3, 2, 1 respectively.) SATJOB: (Asked of persons currently working, temporarily not at work or keeping house.) On the whole, how satisfied are you with the work you do – would you say you are very satisfied, moderately satisfied, a little dissatisfied or very dissatisfied? (Coded from 4 down to 1.) SATFAM: For each area of life I am going to name, tell me the number that shows how much satisfaction you get from that area. Your family life 1. 2. 3. 4. 5. 6. 7.

A very great deal A great deal Quite a bit A fair amount Some A little None

(Reverse coded here.) SATHEALTH: Same as SATFAM, except ‘Your family life’ is replaced by ‘Your health and physical condition.’


APPENDIX 4.A2

91

DESCRIPTIVE STATISTICS

Table 4.A1 Variable

Number of Observations

Mean

Standard Deviation

Minimum Maximum

Happy Satfin Satjob Satfam Sathealth

29 651 29 728 23 808 23 207 23 252

2.22 2.04 2.66 4.66 4.24

0.63 0.74 0.92 1.62 1.68

1 1 1 1 1

3 3 4 7 7

Age Birth Cohort (1880 = 0) Years of Schooling Male Black

29 853 29 853

43.89 60.10

17.18 18.34

18 4

89 96

29 853 29 853 29 853

12.35 0.45 0.11

3.12 0.50 0.31

0 0 0

20 1 1

t1973 t1974 t1975 t1976 t1977 t1978 t1980 t1982 t1983 t1984 t1985 t1986 t1987 t1988 t1989 t1990 t1991 t1993 t1994

29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853 29 853

0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.10

0.22 0.22 0.22 0.21 0.22 0.22 0.22 0.22 0.22 0.22 0.22 0.22 0.22 0.22 0.22 0.21 0.22 0.23 0.30

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

92


APPENDIX 4.A3 Table 4.A2

Independent Variable Age Agesq Coh Cohsq Educ

STEPS 1 AND 2 EQUATIONS

Regression of happiness and each domain satisfaction variable on specified independent variables: ordered logit statistics Dependent Variable Happy (1) 0.022 772 (0.001)** −0.00 022 (0.001)** −0.02 851 (0.000)** 0.00 019 (0.000)** 0.055 533 (0.000)**

Educsq Male Black t1973 t1974 t1975 t1976 t1977 t1978 t1980 t1982 t1983 t1984 t1985 t1986

−0.09 836 (0.000)** −0.69 836 (0.000)** ------------0.035 135 (0.648) −0.09 603 (0.199) −0.05 226 (0.471) 0.043 856 (0.538) 0.10 802 (0.110) 0.006 386 (0.927) 0.017 871 (0.793) −0.12 173 (0.062)+ 0.07 414 (0.281) −0.14 544 (0.024)* 0.022 754 (0.732)

Satfin (2)

Satfam (3)

−0.04 344 (0.000)** 0.000 514 (0.000)** −0.01 713 (0.000)**

0.044 079 (0.000)** −0.00 043 (0.000)**

0.035 612 (0.050)+ 0.002 054 (0.005)** 0.012 652 (0.596) −0.61 869 (0.000)**

0.075 085 (0.000)** −0.00 192 (0.023)* −0.17 479 (0.000)** −0.45 077 (0.000)**

------------−0.00 994 (0.886) −0.05 083 (0.466) −0.01 693 (0.800) 0.224 268 (0.001)** 0.133 804 (0.051)+ −0.08 415 (0.197) −0.1 206 (0.054)+ −0.1 065 (0.099)+ 0.021 233 (0.737) 0.024 798 (0.697) 0.090 113 (0.178)

Reference Year 0.077 578 (0.281) 0.171 288 (0.019)* −0.05 673 (0.419) 0.039 493 (0.584) 0.007 053 (0.922) 0.199 211 (0.007)** 0.303 385 (0.000)** −0.03 378 (0.637) 0.208 851 (0.005)**

−0.12 431 (0.085)+

Satjob (4) 0.043 668 (0.000)** −0.0 004 (0.000)** −0.03 556 (0.000)** 0.000 164 (0.018)* 0.047 991 (0.000)**

0.021 695 (0.423) −0.43 641 (0.000)** ------------−0.07021 (0.373) 0.187315 (0.021)* 0.060301 (0.441) 0.00468 (0.949) 0.159489 (0.034)* −0.04797 (0.520) 0.02714 (0.707) 0.141 839 (0.045)* −0.01 138 (0.881) 0.073 894 (0.302) 0.231 755 (0.002)**

Sathealth (5) −0.01 198 (0.060)+ −0.0 001 (0.018)* −0.0 089 (0.064)+

0.210 914 (0.000)** −0.00 579 (0.000)** 0.138 811 (0.000)** −0.16 603 (0.000)** ------------0.038 031 (0.583) 0.014 605 (0.826) −0.00 767 (0.906) 0.079 327 (0.254) −0.01 189 (0.856) 0.177 236 (0.009)** 0.352 281 (0.000)** -0.0 688 (0.319) 0.212 522 (0.003)**

−0.11 249 (0.141)


Table 4.A2 Independent Variable t1987

(continued) Dependent Variable Happy (1)

Satfin (2)

−0.02 126 (0.759) 0.162 877 (0.016)* 0.068 958 (0.307) 0.15 064 (0.032)* 0.013 087 (0.852)

0.172 639 (0.006)** 0.158 673 (0.015)* 0.119 368 (0.076)+ 0.051 672 (0.473) 0.077 566 (0.252)

t1994

−0.12 405 (0.059)+

0.120 657 (0.060)+

cut1:Constant

−2.03 755 (0.000)** 0.79 881 (0.052)+

−2.19 873 (0.000)** −0.14 451 (0.713)

t1988 t1989 t1990 t1991 t1993

cut2:Constant

93

cut3:Constant cut4:Constant cut5:Constant cut6:Constant Observations 29 651 Pseudo 0.014 R-squared Chi2 607.512 Log Likelihood −27 328.6

Satfam (3)

Satjob (4)

0.06 154 (0.410) 0.107 415 (0.196) 0.066 921 (0.407) 0.052 268 (0.526) 0.065 499 (0.420) 0.0 394 (0.621) 0.028 085 (0.780)

−0.04 379 (0.554) 0.121 667 (0.107) 0.09 812 (0.194) 0.115 772 (0.145) 0.100 156 (0.200)

−2.92 248 (0.000)** −2.0 171 (0.000)** −1.36 141 (0.000)** −0.47 882 (0.006)** 0.285 209 (0.102) 1.808 822 (0.000)**

−3.07 068 (0.000)** −1.72 014 (0.000)** 0.234 316 (0.608)

Sathealth (5) 0.163 291 (0.039)* 0.087 871 (0.309) 0.001 555 (0.986) 0.141 139 (0.136) −0.01 768 (0.855) 0.019 645 (0.845)

0.12 954 (0.074)+ −3.54 667 (0.000)** −2.42 659 (0.000)** −1.77 893 (0.000)** −0.65 057 (0.189) 0.075 143 (0.879) 1.529 096 (0.002)**

29 728 0.031

23 207 0.007

23 808 0.02

23 252 0.016

1 734.832 −30 710.4

371.977 −31 389.3

879.21 −25 354

1 052.814 −37 052.7

Note: + significant at 10%; * significant at 5%; ** significant at 1%; robust p-value in parentheses.

94


APPENDIX 4.A4 Table 4.A3

STEP 3 EQUATION

Regression of happiness on domain satisfaction variables:ordered logit statistics

Independent Variable

Happy

Satfin

0.573 019 (0.000)** 0.498 200 (0.000)** 0.460 422 (0.000)** 0.242 419 (0.000)** 4.299 545 (0.000)** 7.743 151 (0.000)**

Satjob Satfam Sathealth cut1:Constant cut2:Constant Observations Pseudo R-squared Chi2 Log Likelihood

18 440 0.133 3 200.648 −14 855.8

Note: + significant at 10%; * significant at 5%; ** significant at 1%; robust p-value in parentheses.

PART II

Happiness and economics

5.

Happiness when temptation overwhelms willpower Alois Stutzer1

5.1

INTRODUCTION

Saint Anthony of Egypt was tempted by the devil, who appeared in the guise of a monk offering Anthony bread while he was fasting. Anthony overcame the temptation and pursued his long-term plans. As mere mortals we sometimes lack the willpower to withstand the seductions of window displays, to curb our hunger for salty and fatty titbits, to control anti-social emotions and so on. Accordingly, some of us end up obese, addicted to drugs, indebted, with poor job market outcomes or with unsuccessful relationships. Independently of whether the involved behavior is perceived morally condemnable, it is often understood to reduce individuals’ welfare. But how do we know that some choices are suboptimal according to people’s own evaluation? On what foundation do we judge whether people make mistakes? Addressing these questions is highly relevant for public policy if the goal is to understand environments that make people best off. If some choices are suboptimal, conditions can be searched for that make these mistakes less likely. The scientific analysis of suboptimal choices is not an easy task though. The rational choice perspective in traditional economic theory is purely equipped to offer guidance in studying systematic errors in consumption choice. According to its basic view, individuals know what they choose. They are able to predict costs and benefits of pursuing some activity or consuming some good now and in the future. After people have chosen some options, these options are implemented without problems. The preferred course of action can be pursued and people’s behavior in expectations maximizes their welfare. This implies that behavior reveals consistent preferences. Systematic errors in consumption choice are thus ruled out by assumption. This approach makes it almost impossible to detect and understand suboptimal consumption decisions due to problems of, for example, limited willpower.2 97

98


We propose the economics of happiness3 as an alternative approach to study phenomena where temptation might overwhelm willpower. This chapter contributes to the cross-disciplinary field of economics and psychology (see, for example, Camerer et al., 2004; De Cremer et al., 2006; Frey and Stutzer, 2007; Rabin, 1998) and establishes a closer link between the study of suboptimal choices and the research on subjective well-being (see also Hsee et al., 2008). Based on individuals’ judgments of the quality of their lives, it is, in principle, possible to derive whether some observed behavior is suboptimal and is therefore reducing a person’s welfare. This approach draws on empirical research in which the dependent variable is reported subjective well-being or life satisfaction and consumption behavior serves as the main explanatory variable. This approach is promising as it puts forward a proxy for individuals’ welfare to evaluate choice behavior. However, the approach is subject to the same econometric difficulties faced by studies that examine the determinants of behavior, that is, the possibility of omitted variable bias and endogeneity bias. Here, the economics of happiness is applied to further the understanding of important consumption decisions. We mention recent evidence on TV viewing and smoking. In more detail, we discuss eating habits in Western countries where obesity has become a major health issue. Research in economics has provided important insights into how technological progress reduced the relative price of food and contributed to the increase in obesity. However, the increased availability of food might well have overstrained willpower and led to suboptimal consumption decisions relative to people’s own standards. It is indeed found that obesity decreases the well-being of individuals who report limited self-control while it does not do so for others. In the following section theoretical and empirical challenges of using the happiness approach to study suboptimal choices are discussed. First, the key characteristics and the normative basis of the approach based on individuals’ judgments about the quality of their lives are disclosed. Second, issues in the empirical identification are raised. Sections 5.3–5 present evidence on willpower and subjective well-being for three important domains of consumption: TV viewing, smoking and eating. Section 5.6 offers concluding remarks.

Happiness when temptation overwhelms willpower

5.2 5.2.1

99

THE HAPPINESS APPROACH TO SUBOPTIMAL DECISIONS AND INDIVIDUAL WELFARE Normative Basis and Evaluation Metric

Limits of the revealed preference approach The normative basis of the study of consumer choice in economics is the idea of consumer sovereignty. Individuals’ choices are considered to be the result of rational utility maximization. This view is, however, challenged by research in economics and psychology that reports a large number of different anomalies in a real-life decision-making context. Anomalies are understood in the sense of individual behavior violating certain axioms underlying the rational consumer hypothesis (Kahneman et al., 1991). One of the most challenging deviations from utility maximizing consumption choice is due to people having limited willpower. Standard economics is mute about willpower and assumes that people are able to make and implement decisions according to their long-term preferences. Viewed this way, consuming goods and pursuing activities that some people consider addictive, or at least forming bad habits, such as smoking cigarettes or taking cocaine, watching TV or driving expensive cars are considered a rational act. Contrary to this view, many people judge their own and other people’s consumption behavior as irrational in the sense that they think that they would be better off if they would consume less of these goods and care more for their future well-being. Such self-control problems involve two aspects: myopia and procrastination. In both cases the present is emphasized at the expense of the long term. When affected by myopia, people focus on consuming in the present and lack discernment or long-range perspective in their thinking and planning, thus undermining their well-being over time. In this respect, generally goods offering immediate benefits at negligible immediate marginal costs are tempting. Procrastination focuses on putting off or delaying an onerous activity more than a person would have liked when having evaluated it beforehand. In other words, willpower is necessary to stand temptations to fulfill immediate desires that conflict with long-term goals.4 Based on revealed preference, it is difficult, if not impossible, to discriminate between the view of consumers as rational sovereign actors and consumers facing limited willpower. Promises of a complementary happiness approach Two extensions of the traditional emphasis on ex ante evaluation and observed decision are insightful in the study of individual welfare. First, the standard economic concept of revealed preference is complemented

100


with the concept of individual judgment on the quality of one’s life, what might be called the happiness approach. This separation of concepts makes it possible that judgments about experiences systematically diverge from orderings of options derived from observed behavior. The second extension is closely related to the first, and emphasizes ex post evaluations as a valuable source of information about the possibility of bounded rationality in people’s decision making. How do people fare and judge their situation after they have made decisions? The key idea is thus that judgments on people’s life are captured as a proxy for their individual welfare. Thereby the standards underlying people’s judgment are assumed to be those that the individual would like to pursue in order to maximize welfare. Thus the identification of mistakes hinges on the presumption that individuals pursue individual welfare based on some stable evaluation standards. Moreover, whether mistakes are properly identified depends on whether the evaluation metric fits people’s judgment of their life. The normative basis of the approach thus goes beyond assuming the pursuit of happiness but also involves the choice of the concrete evaluation metric to elicit people’s judgments.5 Thus ambiguities remain when choosing the empirical concept in order to measure individual welfare.6 Some people might favor a distant look reflecting on one’s life after the fact, while others favor the reasoned ex ante evaluations as their standards. Still others might give priority to how they felt when experiencing the course of life. Imagine those people who see happiness or high individual welfare as something like the ‘positive, persistent attitude towards both particular experiences and life experience more generally that a person feels upon repeated reflection’ (Kelman, 2005, pp. 408ff.). For them, general evaluations of their satisfaction with life as a whole might be an appropriate metric to capture judgments about individual welfare. For those people who equate individual welfare with moment-tomoment affect, individual welfare might be measured relatively best by approaches like the experience sampling method (Csikszentmihalyi and Hunter, 2003; Scollon et al., 2003) or the day reconstruction method (Kahneman et al., 2004). The relevance of the (normative) choice of an elicitation mechanism is underscored for the specific application to willpower. While the benefits of immediate gratification seem easily accessible in moment-to-moment measurement of individual affect, this same measure is likely to miss the value that people attribute to resisting a temptation and to exercising willpower. When looking for an empirical tool to collect information about people’s judgment, it is thus important to reveal the concrete metric.


5.2.2

101

Empirical Identification of Suboptimal Choices Due to Limited Willpower

Before we discuss the testing strategies based on individual welfare judgments, we briefly mention two previous approaches. Both of them look for patterns of behavior that cannot easily be reconciled with standard utility maximization. Prediction of behavior with indicators of limited self-control This approach starts out with a standard model of individual behavior. It is studied whether the explanatory power of the (empirical) model is increased when the variation in people’s level of self-control is taken into account. Empirically, limited self-control is captured (1) by using behavioral markers, like not having a bank account or having had many hangovers from alcohol consumption in the recent past (see, for example, DellaVigna and Paserman, 2005), (2) by letting people in experiments choose between immediate payoffs and higher delayed payoffs (Thaler, 1981), or (3) by measures of self-report. There is a rich literature in psychology on developing and applying survey measures of self-control (for example, Tangney et al., 2004) or related psychological measures like conscientiousness7 (for an application in economics, see Ameriks et al., 2007) and mastery8. In addition, there are more specific survey measures relying on scenarios (ibid.) or direct reporting of self-control problems in some specific aspect of consumption. Self-infliction of costs A problem of self-control is diagnosed if people are observed spending a lot of time or money on changing their behavior. For instance, in the context of obesity, this ‘smoking gun’ approach looks for evidence like, for example, spending a lot of money on staying at a clinic where mainly sugarless tea is served. More generally, self-binding mechanisms are voluntarily chosen to reduce the utility of some activity or to make short-term revisions of consumption plans less attractive. Ex post evaluation based on individual judgments on the quality of one’s life This approach puts forward a proxy for individual welfare to evaluate choice behavior. Expressed in simple terms, it is studied whether some specific behavioral patterns are related to higher or lower reported subjective well-being. However, the approach is subject to the same econometric difficulties faced by studies that examine the determinants of behavior, that is, the possibility of omitted variable bias and endogeneity bias. Two specific strategies to deal with the issue of endogeneity will be discussed in the following sections on TV viewing and smoking.

102


In this section a general strategy is proposed that combines various sources of information to what might be called a thick description of limited willpower and subjective well-being. It is studied whether the ex post evaluation of some behavior systematically varies between groups of people who have differing amounts of willpower. As we will apply this approach to the phenomenon of obesity in Section 5.5, it is also introduced in this context in order to clarify the ideas and arguments. The central hypothesis states obesity makes people worse off in terms of reported subjective well-being if the increased body mass is due to a self-control problem. However, if people are not lacking willpower, a body mass index (BMI) above 30 does not enter negatively into the evaluation of people’s well-being. Any correlation between the level of willpower and subjective well-being as such is statistically captured in the constant term. Three comments serve to clarify the underlying assumptions, strengths and weaknesses of the approach: 1.

It is no problem for the approach if – in terms of an application to obesity – fat people are jollier. The approach does not rely on a specific benchmark correlation between the phenomenon under study and subjective well-being. A strong preference for food (and thus a higher BMI) can be positively or negatively correlated with reported well-being. It is predicted that obese people judge their overall wellbeing less favorably than people of normal weight when they indicate limited willpower rather than when they do not. 2. The approach explicitly and directly tests for effects on individual welfare. Importantly, individuals are free to judge and evaluate a certain life-path. This involves, for example, the weighting of shortterm pleasures versus the pursuit of long-term goals. 3. The testing strategy relies on two qualities of the self-control problem under study. First, individuals are aware of their limited willpower (as self-reports are used). Second, self-control behavior is generalized across different consumption decisions. This means that limited willpower affects behavior across-the-board. Depending on the application, this latter assumption might be questionable. In the case of obesity, it means that a high BMI need not be positively correlated with the exertion of willpower in other areas. There is, however, the possible scenario that people prefer to be weak-willed and fat rather than to be weak-willed and a chain smoker. Weak-willed people who are obese would then not necessarily judge their well-being less favorably than those who are not obese, and the testing strategy fails to identify limited willpower reducing individual welfare.


103

The assumption that self-control behavior generalizes across activities, however, can be relaxed in the concrete application. Consumption activities that are close substitutes for people with limited willpower can be simultaneously taken into account in the empirical analysis. In sum, this general strategy combines the idea of exploiting supplementary information on the variation in a person’s level of willpower and the idea of an ex post evaluation based on reported subjective well-being.

5.3 5.3.1

TV VIEWING A Controversial Mass Phenomenon

There are countries where people over their lifetime spend as many hours watching TV as they devote to paid work (Corneo, 2005). TV viewing is thus by far the most important leisuretime activity in modern societies. The largest number of heavy TV viewers9 in Europe is found in Greece. As much as 36.8 percent of the population (age 15 and older) reports that they spend three hours a day or more watching TV. At the other end of the ranking, there are 8.4 percent heavy TV viewers in Switzerland (Frey et al., 2007 based on the European Social Survey). Revealed preference therefore suggests that, for many people, TV consumption is an important source of well-being. This assessment is in contrast to the mixed appraisal of TV viewing in society. Television has been called a ‘plug-in-drug’, keeping people glued to the screen and impeding the enjoyment of more valuable experiences. Accordingly, the expansion of TV consumption has a negative connotation, being associated with a decline in social capital, an increase in violence and crime, and a weakening of democracy.10 In sum, there is a strong popular notion that people watch too much TV. They watch more than they would like to watch, both ex ante and ex post.11 The reason why TV may lend itself to overconsumption is mainly due to the immediate benefits and the negligible immediate marginal cost of engaging in this activity. One just has to push a button. In contrast to going to the cinema, the theater or any outdoor activity, there is no need to be appropriately dressed before leaving the house, and there is no need to buy a ticket or to reserve a seat in advance. Watching TV does not require any special physical or cognitive abilities (Kubey and Csikszentmihalyi, 1990, p. 173). Unlike other leisure activities, TV viewing does not need to be coordinated with other persons. It is quite possible to sit alone in front of the TV, while other leisure activities, such as tennis or golf, require a partner with similar time availability and similar preferences. As

104


a consequence, watching TV has, compared to other leisure activities, an exceedingly low or nonexistent entry barrier. At the same time, it offers entertainment value and is considered to be one of the best ways of reducing stress. Moreover, while watching TV immediate marginal costs are even lower and having a remote control is an invitation to ultra short-term optimization (zapping). Many of the costs resulting from such consumption behavior are not experienced immediately, or not predicted at all. These characteristics of the consumption good induce many individuals to fall prey to excessive TV viewing. Here, the role of limited willpower in TV viewing is addressed with regard to consumers’ welfare. It is hypothesized that, for people facing similar restrictions, heavy TV viewing indicates overstrained willpower rather than a love of TV. Accordingly, heavy TV consumption is expected to result in lower reported subjective well-being. In such a situation an increase in the price of TV viewing would be expected to increase the well-being of TV viewers with a self-control problem as the higher price were to serve as a selfbinding mechanism (analogous to the argument in Gruber and Mullainathan, 2005). For most consumers, the price of viewing an additional hour of TV, however, is zero. It is thus not easy to pursue this empirical approach to test the rational consumer hypothesis (at least as long as pay per view is not more common). An interesting alternative might be the extreme case of no TV. While it is definitely not optimal, it might be attractive compared to unrestrained consumption. The introduction of TV would represent a situation for a possible comparison. However, in most countries this technological innovation gained ground too early in the twentieth century in order to be able to match it with data on reported subjective well-being. There are, however, some natural experiments about access to TV that provide some insights as to the consequences of TV for factors that are closely related to individual well-being. A certain Canadian city was unable to receive any TV signals up until 1973, due to its location in a steep valley. Otherwise, it was similar to two cities in the vicinity used as control cases. A study by Williams (1986) suggests that the introduction of TV crowded out other activities, in particular those outside the home, such as taking part in sports’ activities or attending clubs. It also reduced the reading abilities and creative thinking of children and fostered more aggressive behavior and stereotyped ideas about gender roles. TV also reduced the problem solving capacities of adults. Another study by Hennigan et al. (1982), based on a natural experiment, takes a look at the advent of TV in the USA which, due to technical reasons, took place at different times in different places. Petty crime, but not violent crime, increased. Observing the same time period in the USA, Gentzkow (2006) finds that the advent of TV reduced voter turnout.


5.3.2

105

The Subjective Well-being of Heavy TV Viewers

Direct evidence on the relationship between TV viewing and reported subjective well-being is scarce. So far, mainly the subjective well-being of heavy TV viewers is analysed, controlling for many individual characteristics. Such an approach is followed in a large study on TV viewing and life satisfaction for 22 European countries in 2002–03 (Frey et al., 2007). It was found that the more people spend time watching TV, the lower their reported satisfaction with life, ceteris paribus. The result of the econometric analysis is consistent with the hypothesis that heavy TV viewers suffer significant reductions in their utility because they are unable to fully control their TV consumption: They watch too much according to their own evaluation.12 Where do the costs of the misallocation of time come from? There are lost alternatives in the present, like engaging in more stimulating activities or socializing. It is, for example, found that people watching a lot of TV spend less time with family and friends and invest less in relational goods in general (Bruni and Stanca, 2008). But there are also costs in the future. One might be tired the next morning because of not getting enough sleep. Seen long term, people might change their beliefs about the world and about the sources of well-being. In particular, the exposure to the healthy, wealthy and good-looking people on TV is expected to increase people’s aspirations with regard to their own body, but also with regard to their consumption standard. There is substantial research on the relationship between TV viewing and materialism (for example, Kasser, 2002) and financial satisfaction (Bruni and Stanca, 2006; Layard, 2005). Most studies find a positive correlation between extensive TV consumption and those outcomes that are related to lower subjective well-being. In the study for 22 European countries mentioned above (Frey et al., 2007), half of the correlation between TV consumption and life satisfaction can be attributed to heavy TV viewers having lower financial satisfaction, attributing more importance to being rich, feeling less safe, trusting other people less and thinking that they are involved less in social activities than their peers.13 Because these costs are not experienced immediately, individuals with time inconsistent preferences are unable to adhere to the amount of TV viewing they planned or which, in retrospect, they would consider optimal for themselves. 5.3.3

Willpower and the Role of the Choice Set

An alternative way of testing whether heavy TV viewers experience reduced individual well-being because of their consumption choice refers

106


to people’s opportunity set. Benesch et al. (2006) study whether the effect of having a larger number of TV channels available, that is, a larger choice set, raises people’s subjective well-being, as would be expected by standard economic theory. The expansion and diversification of media supply, due to VCR, cable or satellite has, in many countries, gone hand in hand with increased television viewing time (see the collected studies in Becker and Schoenbach, 1989). Again, the longer time spent in front of the TV set is consistent with rational consumers, as well as with TV viewers who are subject to a self-control problem. A study on the introduction of cable TV in Israel (comparing neighborhoods with a difference-in-difference approach) reports people’s evaluation of their consumption choice (Weimann, 1996). It is found that with cable TV and thus more channels there is a significant increase in the percentage of viewers agreeing to the statements ‘I often watch television more than I intend to’ (28 percent before cable introduction and 41 percent one year after) and ‘watching television is often a waste of time’ (24 percent before cable introduction and 36 percent after). The expanded choice set due to the technological change seems to have led to an increase in the number of people watching more TV than they planned to, or more than they think is good for them.14 Benesch et al. (2006) test the hypothesis based on recent data from the European Social Survey, World Values Survey and Television Key Facts from IP Germany. In a first step consumers who possibly have a self-control problem are identified as those people with a large positive residual in a regression explaining the amount of TV viewing according to individual socio-demographic characteristics (referred to as ‘heavy viewers’). In a second step the effect of a higher number of TV channels on subjective well-being is estimated for heavy TV viewers in comparison with moderate TV viewers. Based on more than 70 000 individual observations from 45 country samples, they find a statistically significantly negative interaction term between (residual) TV viewing and the number of TV channels, and calculate a negative marginal effect of additional TV channels on the wellbeing of heavy viewers. This is consistent with the hypothesis of limited self-control being involved in the phenomenon of mass TV consumption.

5.4

SMOKING

Recently, many Western countries have introduced extended smoking restrictions. Based on the standard economic model, we would predict that smokers would heavily oppose them. Smokers could always voluntarily refrain from smoking and these restrictions are thus only perceived as an additional restriction on the choice set. However, if it is taken into account


107

that some smokers might suffer limited willpower to quit smoking, they might welcome smoking restrictions as a kind of self-control mechanism. Hersch (2005), in fact, finds for the USA using data from the Current Population Survey that smokers who plan to quit smoking are more supportive of smoking regulations than are other smokers. A similar reasoning holds for tobacco taxes. Different behavioral models can make systematically different predictions for the effect of excise taxes on individual welfare, while they may all predict reduced consumption of the good that is taxed. People suffer a loss when a normal good is taxed, but experience increased utility when the tax helps to overcome a bad habit like smoking. Accordingly, people might oppose sin taxes as being discriminatory against particular pleasures in life or advocate them to encourage individuals to improve their lot. In a nutshell, the standard economic model predicts that recent increases in cigarette taxes and restrictions on smoking reduce smoking and make individuals worse off. A model that incorporates limited willpower, however, predicts that smoking is reduced while individual utility might be increased. Research on happiness can contribute to this debate and directly study the effect of tobacco taxes on people’s subjective well-being. In two longitudinal analyses across the USA and Canadian states Gruber and Mullainathan (2005) perform such a test with data from the General Social Survey. They analyse the effect of changes in state tobacco taxes on the reported happiness of people who are predicted to smoke at the prevailing tobacco tax. They arrive at the result that a real cigarette tax of 50 cents15 significantly reduces the likelihood of being unhappy among those with a propensity to be smokers. In fact, they would, with a 50 cents tax, be just as likely to report being unhappy as those not predicted to be smokers (that is, the proportion of smokers in the lowest happiness category would fall by 7.5 percentage points). This result favors models of time-inconsistent smoking behavior, in which people have problems with self-control.16 Moreover, the result shows that price increases can serve as a self-commitment device. Problems of self-control with smoking also arise due to temptation (Bernheim and Rangel, 2004). Alternative tests would relate the happiness of potential smokers to clean air laws. These tests would capture exogenous changes in cues or moments of temptation. A comparison of results would allow assessing the boundaries of prices as a means of affecting self-regulation. Research findings on subjective well-being with regard to self-control problems with smoking complement other evidence suggesting self-control problems in a systematic way. There is a large market offering all kinds of drugs and therapies to people who want to stop smoking. In fact, eight out

108


of ten smokers would like to quit smoking and try it every eight and a half months on average (Gruber and Koszegi, 2001).

5.5 5.5.1

OBESITY Can Obesity be Explained by Genes and Relative Prices?

Obesity is on the rise in many Western countries and, with it, illnesses such as diabetes and heart disease. Observers call it an ‘obesity pandemic’, comparable to big threats such as global warming and bird flu, or talk of it as the epidemiological landslide of the last two decades. Overweight and obesity is defined relative to people’s weight to height ratio in metric units, as captured in the body mass index (BMI): BMI 5 kg/m2. Adults with a BMI ≥ 30 kg/m2 are classified as obese and those with a BMI ≥ 25 kg/m2 as overweight. In many European countries the prevalence of obesity has risen three-fold or more since the 1980s (Sanz-de-Galdeano, 2005; WHO Europe, 2005). In Europe adults today have an average BMI of almost 26.5. Worldwide, the percentage of obese adults varies greatly: around 3 percent in South Korea and Japan, 8 percent in Switzerland, over 22 percent in the UK and more than 30 percent in the USA (Figure 5.1). In the USA adult obesity rates have more than doubled since the 1980s. In the year 2000, three in ten adults were classified as obese (Flegal et al., 2002). Overweight accounts for 10–13 percent of deaths and 8–15 percent of healthy days lost due to disability and premature mortality (DALY) in the European Region (World Health Organization, 2002). A debate has started about the economic causes of this phenomenon, as well as its consequences (see, for example, Cutler et al., 2003; Finkelstein et al., 2005; Rashad, 2006). Increased obesity has been explained as the relationship between energy expenditure and energy intake. Energy expenditure is lower nowadays because manual labor has been replaced by more sedentary work due to technological changes (Lakdawalla and Philipson, 2002). However, this trend started long before the obesity endemic took off. The increase in calories consumed fit the obesity pattern better and is of sufficient magnitude to account for its increased prevalence (Putnum and Allshouse, 1999). In particular, higher snack calories are responsible for higher energy intake for men, and for even higher energy intake for women (Cutler et al., 2003). What is the economic rationale behind the shifting energy household? Looking at relative prices suggests that, since the early 1980s, there has been a decrease in price for calorie-dense foods and drinks compared to


109

35 30

Percentage

25 20 15 10 5

Ko Sw J rea itz apa er n N lan or d w ay Ita Au ly s F tri D ran a N en ce et m he a rl rk Sw and e s Po den Be lan lg d Ic ium el a S nd Fi pain n Po lan G rtug d er a C m l ze ch Ire any R lan ep d N C ub ew a lic Ze nad a a H Lu u lan xe ng d m ar Sl o A bo y U vak us urg ni R tra te e li d pu a Ki b ng lic d U ni M om te ex d ic St o at es

0

Note: Percentage of population aged 15 and over, with a BMI greater than 30 (2003 or latest available year). Source:

OECD (2005).

Figure 5.1

Obesity across countries

fruit and vegetables, which are less energy-dense (Finkelstein et al., 2005). These price reductions were made possible by new technologies in food production, in particular for prepackaged and/or prepared food. People have reacted by eating more frequently (snacking), eating bigger portions and spending less time on preparing meals. The question arises how these increases in body weight, causing considerable harm to people’s health, are to be evaluated. Do people eat too much? What is the standard for ‘too much’ if people can choose when and how much they want to eat? Traditional economics advises us to resort to consumer sovereignty under such conditions. Even with full information about the benefits of physical activity, the nutrient content of food, and the health consequences of obesity, some fraction of the population will optimally choose to engage in a lifestyle that leads to weight gain because the costs (in terms of time, money, and opportunity costs) of not doing so are just too high. (Ibid., 2005, p. 252)

This might apply even more because health insurances and taxpayers finance a large amount of the monetary costs of obesity. Moreover,

110


obesity can be rationalized, assuming a high discount rate for future outcomes. The argument for variation in individual discount rates as an explanation for increased obesity is put forward, for example, by Komlos et al. (2004).17 However, the possibility of individuals consuming ‘too much’ food is excluded by assumption in the revealed preference approach. Yet, in order to justify this view, one would have to reconcile the prevalence of obesity with other behavioral regularities, like people spending large sums of money on diets and health clubs, or people’s weight yo-yoing as they go from one diet to the next.18 An alternative approach accepts that people might face self-control problems when exposed to the temptation of immediate gratification from food when they are hungry or have a craving for something sweet, fatty or salty (see, for example, Offer 2001). There is a rich literature on the control of eating, emphasizing physiological mechanisms (Blundell and Gillett, 2001; Smith, 2006). In particular, humans are endowed with a system of weight regulation that favors weight gain over weight loss to reduce any future risk of starvation. While this ability was evolutionary advantageous, it is a challenge to conscious control of food intake today. People consume more food and calories and eat more frequently than that they consider good for themselves when they think about and plan their diet. People are aware of this phenomenon, but more so in others than themselves (Taylor et al., 2006). They judge their own and other people’s consumption behavior as irrational in the sense that they think that they would be better off if they would consume less and care more about their future well-being.19 5.5.2

Previous Evidence on Obesity and Subjective Well-being

Reported subjective well-being provides information about people’s evaluation of their situation after they have decided about their food and beverage consumption. Two predictions summarize the conflicting views on the role of limited willpower in obesity. If technical ‘progress’ in producing fatty food is indeed a major driving force behind obesity, the standard economic model predicts that individuals will become heavier and happier. However, if individuals have self-control problems, we would expect them to become heavier and less happy. There is a growing literature on empirical research, studying whether obese people are less satisfied. According to an empirical investigation for roughly 8000 young women, obesity is related to lower satisfaction with work, family relationships, partner relationship and social activities (but not satisfaction with friendships) (Ball et al., 2004). Other studies


111

report correlations between obesity and symptoms of depression, whereby the risk of depression is higher for obese women than obese men (for example, McElroy et al., 2004; Needham and Crosnoe, 2005). These findings, however, provide only limited insights, as the correlations can be due to third variables affecting both eating behavior and subjective wellbeing, or because low life satisfaction and stress can lead to obesity. The latter has been studied in a longitudinal analysis for 5867 pairs of twins (Korkeila et al., 1998). It is found that a high level of stress, as well as a low level of life satisfaction, are both predictors of weight gain over six years and for certain groups of people over 15 years of age. Another panel study addresses the reverse relationship. Taking baseline mental health into account, it analyses the long-term consequences of obesity, finding an increased risk for depression (Roberts et al., 2002). These results are valuable to assess the relevance of the phenomenon, but they have to be supplemented with further evidence to identify the contribution of selfcontrol problems to the link between obesity and subjective well-being. Alternatively, it is possible to characterize conditions where attempts to recapture self-control are encouraged. It is to be expected that those people who stand to lose a lot from being obese, or who have access to resources, are more successful in controlling their behavior. For example, obese women seem to suffer a salary and promotion penalty even more than obese men (see, for example, Baum and Ford, 2004; Finkelstein et al., 2005). They have strong incentives to control their body weight and might suffer more when their lack of willpower leads to failure. Consistent with this point of view, people in the top income quintile, or in professions with a low prevalence of obesity, report the largest well-being costs of obesity (Felton and Graham, 2005). People with a higher education or income level are more likely to view themselves as overweight, keeping the level of the BMI constant (Oswald and Powdthavee, 2007). 5.5.3

Willpower, Obesity and Subjective Well-Being

Hypothesis and data Based on the preceding subsection, it is hypothesized that the well-being of people with limited self-control is reduced when they are obese, while the well-being of people with strong willpower is not affected. This hypothesis follows the general testing strategy proposed in section 5.2 in order to better understand suboptimal consumption decisions. Here an empirical analysis is presented based on a unique data set combining information on people’s weight, height, perceived control over their life and eating behavior, as well as a multi-item measure of subjective well-being. The original analysis is developed and described in detail in Stutzer (2007).

112


The data are from the Swiss Health Survey 2002 compiled by the Swiss Federal Statistical Office. They combine responses from a telephone survey and a questionnaire mailing going to the same people. The sampling population was the resident population aged 15 and over. Interviewed were 19 706 individuals and 16 141 of them responded to a supplementary written questionnaire. For 19 471 respondents there is complete information about their body mass. Based on a weighted distribution of the BMI index, the percentage of obese people in the adult population amounted to 7.7 percent in 2002, an increase of 2.3 percentage points since 1992. Overweight are 29.4 percent and normal weight are 49.9 percent. There is also a substantial fraction of 13.0 percent having a BMI below 20 and thus being underweight. People’s subjective well-being is assessed using eight questions from the Bern Questionnaire of Subjective Well-Being (Grob et al., 1991). The questions are reported in the Appendix 5A.1. The main analysis is for the compound measure based on the eight items. Variation in self-control between people is assessed using a general measure of reported mastery and a specific measure of reported willpower in pursuing a healthy diet. The mastery scale is from Pearlin et al. (1981), whereby four out of seven questions were included in the survey (see Appendix 5A.1). In the analysed sample 24.1 percent of the subjects report that they feel in control of their life, that is, all questions of limited mastery are completely denied. Domain-specific willpower is derived from the following survey item: ‘Many people – maybe you too – attach importance to a healthy diet. Do you see obstacles for somebody who pursues a healthy diet? Please tick all reasons that apply! . . . “lack of willpower, lack of belief in success” ’. Limited willpower with regard to a healthy diet is reported by 25.5 percent of people in the sample. Results People’s body mass is compared to their subjective well-being.20 According to the basic hypothesis, obesity is expected to negatively affect the subjective well-being of those with limited willpower. For them, obesity is not meant to be the outcome of rational food consumption but rather of time-inconsistent behavior. The dependent variable ‘reported subjective well-being’ is now an ordinal measure. Ordered probit regressions are estimated and marginal effects are calculated for the top category of subjective well-being. For dummy variables, the marginal effects indicate a change in the probability of reporting high subjective well-being. In Table 5.1 results for two different specifications are presented. Specification I includes people’s body mass, dividing it into four categories, as well as categorizations of their age and sex. This specification assures that no other choice variables pick up any potential negative consequence of obesity


Table 5.1

113

BMI and reported subjective well-being Full Sample

General Indicator of Self-control: Mastery

Specific Indicator of Self-control: Lack of Willpower is an Obstacle to a Healthy Diet

Limited Full Limited Full Self-control Self-control Self-control Self-control Marginal Effects for the Top Category of Subjective Well-being Specification I Underweight Normal weight Overweight Obese Control variables Baseline prob. No. of obs. Pseudo R2 Specification II Underweight Normal weight Overweight Obese

−0.012** (4.43e-3) −0.003 (3.46e-3) −0.014** (5.16e-3) 0.099 15108 0.002 −0.008 (*) (4.32e-3) −0.004 (3.32e-3) −0.008 (5.17e-3)

Control variables Baseline prob. No. of obs. Pseudo R2

0.092 15108 0.021

−0.010** 0.029 −0.012 (3.13e-3) (2.05e-2) (8.18e-3) Reference group −0.001 −0.018 −0.010 (2.66e-3) (1.26e-2) (6.46e-3) −0.012** −0.003 −0.034** (3.76e-3) (2.09e-2) (7.63e-3) Age categories and sex included 0.055 0.230 0.086 10681 3392 3458 0.002 0.004 0.003

−0.011* (5.39e-3)

−0.007* −0.028 −0.010 (3.01e-3) (2.04e-2) (7.71e-3) Reference group −0.002 −0.012 −0.010 (*) (2.49e-3) (1.26e-2) (6.07e-3) −0.008* 0.013 −0.027** (3.69e-3) (2.15e-2) (7.49e-3) All factors included (see appendix A.1) 0.050 0.225 0.078 10681 3392 3458 0.021 0.019 0.027

−0.006 (5.31e-3)

−0.001 (4.18e-3) −0.002 (6.83e-3) 0.097 10117 0.002

−0.002 (4.01e-3) 0.005 (6.88e-3)

0.090 10117 0.021

Note: Marginal effects based on ordered probit regression. Standard errors in parentheses. Significance levels: (*) 0.05 < p < 0.1, * 0.01 < p < 0.05, ** p < 0.01. Source:

Stutzer (2007) based on Swiss Health Survey 2002.

on well-being. Specification II includes a large set of covariates of subjective well-being. The Swiss Health Survey provides sufficient information about individual characteristics to specify a microeconometric well-being function that is similar to the ones usually applied when testing economic issues. Table 5.A1 in the Appendix 5A.1 presents the results for such a specification, including all the control variables and covering the full

114


sample. They confirm previous findings in the literature on the correlates of happiness. The findings for the BMI are also shown in the lower half of the first column in Table 5.1. In both equations with the full sample, obesity is negatively correlated with subjective well-being. However, the partial correlation is not statistically significant in specification II. Moreover, the partial correlations are not yet a test of the theoretical prediction. A very high BMI is hypothesized to negatively affect well-being if it is the result of limited self-control, but not otherwise. Therefore, the partial correlation between obesity and subjective well-being are estimated separately for people with full and limited self-control. Both indicators of self-control are applied: mastery and domain-specific willpower. With two specifications each, results from eight estimations are summarized in Table 5.1. Consistent with the basic hypothesis, obesity is related with lower subjective well-being when people have limited self-control but no statistically significant effect is found for the sample of people classified as having full self-control. The marginal effect is largest with specification I for the sample of people who report a lack of willpower as being an obstacle to a healthy diet. The probability of reporting high subjective well-being is 3.4 percentage points lower for people who are obese rather than normal weight, whereby the baseline probability for people in the reference group is 8.6 percent. Together, the pieces of evidence from the study of people’s subjective well-being indicate that the phenomenon of obesity can only be understood when going beyond revealed preference and the assumption of unlimited consumer sovereignty, but taking limited willpower into account. 5.5.4

Three Open Issues

The general approach proposed in this chapter to study the effect of limited willpower on individual welfare raises three related issues that we have not yet taken up. These are the generalization of self-control behavior, reverse causality, and the distinction between outcome and process as possible reasons for reduced well-being when somebody is a heavy TV viewer, a smoker or obese. The three issues are again discussed for the case of obesity. The first issue has to do with the nature of limited willpower. People are exposed to many opportunities, with low immediate marginal costs, but high marginal benefits. The question arises whether people with a selfcontrol problem make myopic decisions when faced with all, or most, of these opportunities, or whether they are able to control some challenges to willpower, but find it too difficult to control all of them. The latter view fits in with the idea that there is a limited capacity for self-regulation.


115

Resisting one temptation may result in poorer regulation of a concurrent desire for immediate gratification, or vice versa (Muraven et al., 1998). For the identification of limited willpower in the proposed approach, it has to be assumed that either self-control behavior generalizes across-theboard or tempting activities that are close substitutes have to be taken into account statistically.21 The idea that self-control resembles a muscle might be particularly relevant in understanding the interplay between obesity and smoking (Gruber and Frakes, 2006). People who work at controlling their eating habits might give up on resisting smoking and vice versa. In the presented study by Stutzer (2007) on the negative effect of obesity, information about whether somebody is a smoker or non-smoker is also included. For the main specification II, with the specific indicator of selfcontrol, almost identical marginal effects as before are found. For people with limited self-control, the marginal effect of obesity is −0.026 (−0.027 before). For people with full self-control, the marginal effect of obesity is 0.005 (0.005 before). The main result is thus robust to the inclusion of the closest substitute to yielding to the temptation to overconsume. The second issue concerns causality. To what extent do the consequences of obesity due to limited willpower reduce subjective well-being and to what extent does the experience of reduced well-being lead to stress/ frustration eating and obesity? This is a valid concern, even though based on the approach proposed in this chapter the correlation between obesity and subjective well-being is not interpreted as such, but rather the differential effect for people with full and limited self-control. The data set used in Stutzer (2007) captures whether somebody turns to eating when stressed. Stress eating is found to negatively correlate with subjective well-being in the full sample, keeping body mass constant. Moreover, the marginal effects are sizeable. While the baseline probability of people reporting subjective well-being in the top category is 8.4 percent, this probability is reduced by between 1.7 and 3.7 percentage points if stress eating is not ‘very untypical’ but is either ‘rather untypical’ or ‘very typical’. However, the differential effect of obesity on subjective well-being between people with limited and full self-control is not explained by stress eating. For the general indicator of self-control, the difference in the marginal effects of obesity on individual well-being is slightly larger when comparing people with limited and full selfcontrol. For the specifications applying the specific indicator of self-control, the difference in marginal effects is slightly reduced. However, people who lack the willpower to stick to a healthy diet still report a significantly lower subjective well-being when they are obese (marginal effect is −2.0 percentage points), while there is no such negative effect for people who report full selfcontrol. A direct test for reverse causation thus cannot explain the reduced well-being of obese people in the case of limited willpower.

116


The third question is whether reduced willpower as such, rather than its consequences, is responsible for lower well-being. Limited willpower might well repeatedly lead to experiencing frustration because plans regarding one’s diet are not realized. People who experience self-control problems then suffer reduced self-esteem, and thus lower subjective well-being. Related empirical evidence is found in a community sample of 2000 adults (Greeno et al., 1998). In addition to a higher BMI, the lack of perceived eating control was associated with lower satisfaction with life. For men, it was only the lack of eating control that was correlated with reported subjective well-being. This line of reasoning is important in order to understand the relationship between appearance norms, body image and eating disorders (Derenne and Beresin, 2006). In the analysis presented above, this aspect is not studied directly. However, the effect of limited willpower on the level of people’s well-being is statistically taken into account when estimating separate equations as it enters into the constant term.

5.6

CONCLUDING REMARKS

This chapter has started out with the provocative question on how to judge whether people make suboptimal consumption choices according to their own evaluation. Within standard neoclassical economics this issue would not have been raised. Standard economic theory assumes that individuals do not commit any systematic errors in their consumption decisions, because they know their own preferences best and are able to make and implement the consequent choices. Limited willpower is no concern. No doubt, it is very likely that individuals are quite capable of making satisfactory consumption decisions for most of the goods most of the time. The main message of this chapter is that it is necessary to go beyond this narrow approach. One should take into account the methodological advances made possible by happiness research. They allow us to empirically test whether individuals do or do not make suboptimal decisions due to limited willpower, rather than simply assuming that they do not, as is the case in revealed preference theory. The possibility to proxy individual welfare in a satisfactory way using data on reported subjective well-being enables economists to empirically study the difference between decisions made and the individual welfare produced. We see a large potential in this approach to study many areas of consumption choice for which popular discourse and most other disciplines acknowledge people’s difficulties in the pursuit of happiness. The proposed approach, however, also requires particular transparency as every empirical analysis gives priority to a specific measure of subjective


117

well-being as a criterion of individual welfare. This involves a strong normative judgment. We are aware that different indicators of individual welfare potentially offer different evaluations of what are suboptimal decisions. They thus capture different aspects of people’s pursuit of happiness. For example, people might gain utility from exerting willpower because they have resisted a temptation. Or they judge their life favorably because their life course contributes to positive self-signaling, identity, goal completion, mastery or meaning (for economic analyses of these human motives see, for example, Akerlof and Kranton, 2000; Loewenstein, 1999). When thinking about policy proposals, it is important to know whether and to what extent people face a self-control problem when tempted by the abundance of consumption opportunities. For instance, take the abundance of food and the observed obesity. Is obesity rational and reflecting an appetite for food so that it can be reduced to an issue of externalities in a publicly regulated and funded health system? Or does it reflect ignorance and a lack of imagination of its consequences, requiring some sort of information policy? Or must obesity be treated like smoking, where some people lack willpower to control their behavior? The challenge for research is to disentangle the various behavioral reasons for obesity. But even if these questions can be answered for the three specific areas discussed, the suboptimal consumption patterns due to limited willpower are no cause for immediate government intervention. Moreover, it is doubtful whether ‘the government’ is able to make better decisions in the interests of the persons concerned (see Frey and Stutzer, 2006). It is presumably more effective to support individuals subject to self-control problems by providing ways of overcoming their weakness, for example by proposing self-binding mechanisms (for a broader discussion see, for example, O’Donoghue and Rabin, 2005). It should thus be clear that this analysis is not a normative evaluation from the point of view of a benevolent social planner. Rather, the focus is on the suboptimal choices in consumption that individuals commit according to their own perception, placing people in a less favorable position in terms of their own welfare evaluation. Thus the policy goal remains a search for institutions under which suboptimal choices are less likely and people are better off accepting their limited willpower.

NOTES 1.

I am grateful to Thomas Brändle, Bruno Frey and Michael Zehnder and the participants of the Conference on New Directions in the Study of Happiness at the University of Notre Dame for helpful comments.

118 2. 3. 4.

5. 6. 7.

8. 9. 10. 11. 12. 13. 14.

15. 16.

17. 18.

Happiness, economics and politics In a recent debate on behavioral welfare economics the possibilities and limits of a choice-based approach to identify mistakes have been controversially discussed (see Bernheim and Rangel, 2007; Gul and Pesendorfer, 2007; Koszegi and Rabin, 2007). For an introduction to the economics of happiness, see, for example, Frey and Stutzer (2002a, b), Layard (2005) and Frey (2008). In economics time inconsistent preferences are most prominently formulated in models of hyperbolic discounting (see, for example, Laibson, 1997). A low discount factor (that is, a discount factor decreased by b, b [ (0,1)) is applied between the present and some point in time in the near future and a constant discount factor d thereafter. An excellent account of the recent extensive empirical and theoretical literature on time inconsistent preferences is provided in Frederick et al. (2002). An excellent account of the ambiguities of welfare in the context of economics and hedonic psychology is provided in Kelman (2005). There is in fact a debate in happiness research about how to measure subjective wellbeing that might be better understood from the angle of proponents’ preferred welfare concept (see Helliwell, 2006; Kahneman and Riis, 2005; Kahneman et al., 2004). ‘Conscientiousness describes socially prescribed impulse control that facilitates taskand goal-directed behaviors, such as thinking before acting, delaying gratification, following norms and rules, and planning, organizing, and prioritizing tasks’ (John and Srivastava, 1999, p. 121). ‘Mastery refers to the extent to which people see themselves as being in control of the forces that importantly affect their lives’ (Pearlin et al., 1981, p. 340). By ‘heavy TV viewers’, people are meant who spend a great deal of time watching TV, and not TV viewers who are overweight (although watching a lot of TV is sedentary and invites people to snack, which can in turn lead to obesity). The case for negative impacts of TV consumption on society is, for example, made by Kubey (1996), Putnam (2000), Sparks and Sparks (2002) and Gentzkow (2006). Regarding television consumption, there is some (anecdotal) evidence that individuals may have self-control problems. Forty percent of US adults and 70 percent of US teenagers admit that they watch too much TV (Kubey and Czikszentmihalyi, 2002). However, the potential correlation is also consistent with a hypothesis according to which unhappy people resort to TV viewing. This might lead to a mutually reinforcing relationship. While these correlations are suggestive, it has to be kept in mind that third factors could be driving differences in the different attitudes as well as in TV viewing. Asking people directly whether they think that they watch too much TV could lead to answers that are motivated by social desirability. It should be noted that surveys on general life satisfaction are less likely to be affected by such a bias, at least not one that is systematically correlated with some specific consumption behavior. The average real (in 1999 US$) cigarette tax in the USA is 31.6 cents in the sample (Gruber and Mullainathan, 2005, p. 5). In another study the negative internality from suffering a self-control problem and being a smoker is assessed (Jürges, 2004). The monthly compensation required to make a smoker as well off as a nonsmoker is estimated to be approximately 500 euros. However, the effects of smoking on life satisfaction were not identified, based on changes in exogenous conditions restricting the possibilities to smoke. In an empirical analysis for the Netherlands Borghans and Golsteyn (2006) conclude, however, that it is unlikely that BMI increased because of an increase in the timediscount rate. This argumentation on the revealed preference approach does not exclude that observed behavior can give clear indications of a problem with the control of body mass, for example, when people inflict costs on themselves in order to make eating chocolate less attractive. However, the revealed preference approach gives no reason to search for such contradictory patterns in consumption behavior. On the contrary, it urges the researcher to look for rationalizations.

Happiness when temptation overwhelms willpower 19.

20.

21.

119

With regard to obesity, the self-control issue is explicitly addressed in Cutler et al. (2003), whereby its relevance in the assessment of consumers’ welfare is discounted because it would require only some exercise on the part of overweight people to balance their energy household. Observed inactivity thus seems to indicate that overweight people do not suffer from their body mass. However, the trade-off is calculated, assuming that people have self-control problems with eating, but not with taking physical exercise. This asymmetry does not fit our casual observations. Stutzer (2007) also studies the covariates of body mass in a multinomial regression analysis. In addition to standard demographic and socio-economic factors, indicators of ignorance and limited willpower are included. It is found that people who report that health is not a relevant issue for them are statistically significantly more likely to be obese than people who report that health is either relevant or very important for them. The relative risk ratio indicates a 1.51 greater probability. Consistent with this, people who care about their diet are less likely to be obese, with a risk ratio of 0.81 relative to people who do not care. Moreover, people who report a lack of willpower when it comes to a healthy diet have a higher probability of being obese, the relative risk ratio being 1.41. This condition can also explain why a domain-specific indicator of willpower is a better predictor of obesity and reduced well-being of obese people than a general indicator of perceived control.

REFERENCES Akerlof, George A. and Rachel E. Kranton (2000), ‘Economics and identity’, Quarterly Journal of Economics, 115 (3), 715–53. Ameriks, John, Andrew Caplin, John Leahy and Tom Tyler (2007), ‘Measuring self control problems’, American Economic Review, 97 (3), 966–72. Ball, Kylie, David Crawford and Justin Kenardy (2004), ‘Longitudinal relationships among overweight, life satisfaction, and aspirations in young women’, Obesity Research, 12 (6), 1019–30. Baum, Charles L., II and William F. Ford (2004), ‘The wage effects of obesity: a longitudinal study’, Health Economics, 13 (9), 885–99. Becker, Lee B. and Klaus Schoenbach (eds) (1989), Audience Responses to Media Diversification: Coping with Plenty, Hillsdale, NJ: Lawrence Erlbaum Associates. Benesch, Christine, Bruno S. Frey and Alois Stutzer (2006), ‘TV channels, self control and happiness’, IEW Working Paper No. 301, University of Zurich. Bernheim, Douglas and Antonio Rangel (2004), ‘Addiction and cue-triggered decision processes’, American Economic Review, 94 (5), 1558–90. Bernheim, B. Douglas and Antonio Rangel (2007), ‘Toward choice-theoretic foundations for behavioral welfare economics’, American Economic Review, 97 (2), 464–70. Blundell, John E. and Angela Gillett (2001), ‘Control of food intake in the obese’, Obesity Research, 9 (Suppl. 4), 263S–70S. Borghans, Lex and Bart H.H. Golsteyn (2006), ‘Time discounting and the body mass index, evidence from the Netherlands’, Economics and Human Biology, 4 (1), 39–61. Bruni, Luigino and Luca Stanca (2006), ‘Income aspirations, television and happiness: evidence from the World Values Survey’, Kyklos, 59 (2), 209–25. Bruni, Luigino and Luca Stanca (2007), ‘Watching alone: relational goods, television and happiness’, Journal of Economic Behavior & Organization, 65 (3–4), 506–28.

120


Camerer, Colin, George Loewenstein and Matthew Rabin (eds) (2004), Advances in Behavioral Economics, New York: Russell Sage Foundation. Corneo, Giacomo (2005), ‘Work and television’, European Journal of Political Economy, 21 (1), 99–113. Csikszentmihalyi, Mihaly and Jeremy Hunter (2003), ‘Happiness in everyday life: the uses of experience sampling’, Journal of Happiness Studies, 4 (2), 185–99. Cutler, David M., Edward L. Glaeser and Jesse M. Shapiro (2003), ‘Why have Americans become more obese?’, Journal of Economic Perspectives, 17 (3), 93–118. De Cremer, David, Marcel Zeelenberg and J. Keith Murnighan (eds) (2006), Social Psychology and Economics, Mahwah, NJ: Lawrence Erlbaum. DellaVigna, Stefano and M. Daniele Paserman (2005), ‘Job search and impatience’, Journal of Labor Economics, 23 (3), 527–88. Derenne, Jennifer L. and Eugene V. Beresin (2006), ‘Body image, media, and eating disorders’, Academic Psychiatry, 30 (3), 257–61. Felton, Andrew and Carol Graham (2005), ‘Variance in obesity across cohorts and countries: a norms-based explanation using happiness surveys’, Mimeo, The Brookings Institution. Finkelstein, Eric A., Christopher J. Ruhm and Katherine M. Kosa (2005), ‘Economic causes and consequences of obesity’, Annual Review of Public Health, 26, 239–57. Flegal, Katherine M., Margaret D. Carroll, Cynthia L. Ogden and Clifford L. Johnson (2002), ‘Prevalence and trends in obesity among US adults, 1999–2000’, JAMA, 288 (14), 1723–7. Frederick, Shane, George Loewenstein and Ted O’Donoghue (2002), ‘Time discounting and time preference: a critical review’, Journal of Economic Literature, 40 (2), 351–401. Frey, Bruno S. (2008), Happiness: A Revolution in Economics, Cambridge, MA: MIT Press. Frey, Bruno S. and Alois Stutzer (2002a), Happiness and Economics: How the Economy and Institutions Affect Well-Being, Princeton, NJ and Oxford: Princeton University Press. Frey, Bruno S. and Alois Stutzer (2002b), ‘What can economists learn from happiness research?’, Journal of Economic Literature, 40 (2), 402–35. Frey, Bruno S. and Alois Stutzer (2006), ‘Mispredicting utility and the political process’, in Edward J. McCaffery and Joel Slemrod (eds), Behavioral Public Finance, New York: Russell Sage Foundation, pp. 113–40. Frey, Bruno S. and Alois Stutzer (eds) (2007), Economics and Psychology. A Promising New Cross-Disciplinary Field, Cambridge, MA: MIT Press. Frey, Bruno S., Christine Benesch and Alois Stutzer (2007), ‘Does watching TV make us happy?’, Journal of Economic Psychology, 28 (3), 283–313. Gentzkow, Matthew (2006), ‘Television and voter turnout’, Quarterly Journal of Economics, 121 (3), 931–72. Greeno, Catherine G., Christine Jackson, Elizabeth L. Williams and Stephen P. Fortmann (1998), ‘The effect of perceived control over eating on the life satisfaction of women and men: results from a community sample’, International Journal of Eating Disorders, 24 (4), 415–19. Grob, Alexander, Ruth Luthi, Florian G. Kaiser and August Flammer (1991), ‘Berner Fragebogen zum Wohlbefinden Jugendlicher (Bfw)’. The Bern subjective well-being questionnaire for adolescents (Bfw), Diagnostica, 37 (1), 66–75.


121

Gruber, Jonathan H. and Botond Koszegi (2001), ‘Is addiction “Rational”? Theory and evidence’, Quarterly Journal of Economics, 116 (4), 1261–303. Gruber, Jonathan H. and Michael Frakes (2006), ‘Does falling smoking lead to rising obesity?’, Journal of Health Economics, 25 (2), 183–97. Gruber, Jonathan H. and Sendhil Mullainathan (2005), ‘Do cigarette taxes make smokers happier?’, Advances in Economic Analysis and Policy, 5 (1), 1–43. Gul, Faruk and Wolfgang Pesendorfer (2007), ‘Welfare without happiness’, American Economic Review, 97 (2), 471–6. Helliwell, John F. (2006), ‘Well-being, social capital and public policy: what’s new?’, Economic Journal, 116 (510), C34–C45. Hennigan, Karen M., Linda Heath, J.D. Wharton, M.L. Delrosario, T.D. Cook and B.J. Calder (1982), ‘Impact of the Introduction of Television on Crime in the United-States: empirical-findings and theoretical implications’, Journal of Personality and Social Psychology, 42 (3), 461–77. Hersch, Joni (2005), ‘Smoking restrictions as a self-control mechanism’, Journal of Risk and Uncertainty, 31 (1), 5–21. Hsee, Christopher K., Yuval Rottenstreich and Alois Stutzer (2008), ‘Suboptimal choices and the need for experienced individual well-being in economic analysis’, Mimeo, University of Basel. John, Oliver P. and Sanjay Srivastava (1999), ‘The big five trait taxonomy: history, measurement, and theoretical perspectives’, in Lawrence A. Pervin and Oliver P. John (eds), Handbook of Personality: Theory and Research, New York: Guilford Press, pp. 102–38. Jürges, Hendrik (2004), ‘The welfare costs of addiction’, Schmollers Jahrbuch, 124 (3), 327–53. Kahneman, Daniel and Jason Riis (2005), ‘Living, and thinking about it: two perspectives on life’, in Felicia A. Huppert, Nick Baylis and Barry Keverne (eds), The Science of Well Being, Oxford: Oxford University Press, pp. 285–301. Kahneman, Daniel, Jack L. Knetsch and Richard H. Thaler (1991), ‘The endowment effect, loss aversion, and status quo bias: anomalies’, Journal of Economic Perspectives, 5 (1), 193–206. 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’, Science, 306 (5702), 1776–80. Kasser, Tim (2002), The High Price of Materialism, Cambridge, MA: MIT Press. Kelman, Mark (2005), ‘Hedonic psychology and the ambiguities of “Welfare”’, Philosophy and Public Affairs, 33 (4), 391–412. Komlos, John, Patricia K. Smith and Barry Bogin (2004), ‘Obesity and the rate of time preference: is there a connection?’, Journal of Biosocial Science, 36 (2), 209–19. Korkeila, M., J. Kaprio, A. Rissanen, M. Koskenvuo and T.I.A. Sorensen (1998), ‘Predictors of major weight gain in adult Finns: stress, life satisfaction and personality traits’, International Journal of Obesity, 22 (10), 949–57. Koszegi, Botond and Matthew Rabin (2007), ‘Mistakes in choice-based welfare analysis’, American Economic Review, 97 (2), 477–81. Kubey, Robert (1996), ‘Television dependence, diagnosis, and prevention: with commentary on video games, pornography, and media education’, in Tannis M. Williams (ed.), Tuning into Young Viewers: Social Science Perspectives on Television, Thousand Oaks, CA: Sage, pp. 221–60.

122


Kubey, Robert and Mihaly Csikszentmihalyi (1990), Television and the Quality of Life. How Viewing Shapes Everyday Experience, Hillsdale, NJ: Lawrence Erlbaum Associates. Kubey, Robert and Mihaly Czikszentmihalyi (2002), ‘Television addiction is no mere metaphor’, Scientific American, 286 (2), 74–80. Laibson, David (1997), ‘Golden eggs and hyperbolic discounting’, Quarterly Journal of Economics, 112 (2), 443–77. Lakdawalla, Darius and Tomas Philipson (2002), ‘The growth of obesity and technological change: a theoretical and empirical examination’, NBER Working Paper No. 8946, Cambridge, Massachusetts. Layard, Richard (2005), Happiness: Lessons from a New Science, New York: Penguin. Loewenstein, George (1999), ‘Because it is there: the challenge of mountaineering . . . for utility theory’, Kyklos, 52 (3), 315–44. McElroy, Susan L., Renu Kotwal, Shishuka Malhotra, Erik B. Nelson, Paul E. Keck and Charles B. Nemeroff (2004), ‘Are mood disorders and obesity related? A review for the mental health professional’, Journal of Clinical Psychiatry, 65 (5), 634–51. Muraven, Mark, Dianne M. Tice and Roy F. Baumeister (1998), ‘Self-control as limited resource: regulatory depletion patterns’, Journal of Personality and Social Psychology, 74 (3), 774–89. Needham, Belinda L. and Robert Crosnoe (2005), ‘Overweight status and depressive symptoms during adolescence’, Journal of Adolescent Health, 36 (1), 48–55. O’Donoghue, Ted and Matthew Rabin (2005), ‘Incentives and self-control’, Mimeo, Cornell University and University of California at Berkley. OECD (2005), OECD Factbook 2005: Economic, Environmental and Social Statistics, Paris: OECD. Offer, Avner (2001), ‘Body weight and self-control in the United States and Britain since the 1950s’, Social History of Medicine, 14 (1), 79–106. Oswald, Andrew and Nattavudh Powdthavee (2007), ‘Obesity, unhappiness, and the challenge of affluence: theory and evidence’, Economic Journal, 117 (523), F441–59. Pearlin, Leonard I., Elizabeth G. Menaghan, Morton A. Lieberman and Joseph T. Mullan (1981), ‘The stress process’, Journal of Health and Social Behavior, 22 (4), 337–56. Putnam, Robert D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York: Simon & Schuster. Putnum, Judith Jones and Jane E. Allshouse (1999), ‘Food consumption, prices, and expenditures, 1970–1997’, Food and Rural Economic Division, Economic Research Service, Statistical Bulletin No. 965, US Department of Agriculture, Washington, DC. Rabin, Matthew (1998), ‘Psychology and economics’, Journal of Economic Literature, 36 (1), 11–46. Rashad, Inas (2006), ‘Structural estimation of caloric intake, exercise, smoking, and obesity’; Quarterly Review of Economics and Finance, 46 (2), 268–83. Roberts, Robert E., William J. Strawbridge, Stephane Deleger and George A. Kaplan (2002), ‘Are the fat more jolly?’, Annals of Behavioral Medicine, 24 (3), 169–80. Sanz-de-Galdeano, Anna (2005), ‘The obesity epidemic in Europe’, Discussion Paper No.1814, IZA, Bonn.


123

Scollon, Christie Napa, Chu Kim Prieto and Ed Diener (2003), ‘Experience sampling: promises and pitfalls, strengths and weaknesses’, Journal of Happiness Studies, 4 (1), 5–34. Smith, Trenton G. (2006), ‘Reconciling psychology with economics: obesity, behavioral biology, and rational overeating’, Working Paper No. 2006-4, School of Economic Sciences, Washington State University. Sparks, Glenn G. and Cheri W. Sparks (2002), ‘Effects of media violence’, in Jennings Bryant and Dolf Zillmann (eds), Media Effects. Advances in Theory and Research, Mahwah, NJ: Lawrence Erlbaum Associates, pp. 269–85. Stutzer, Alois (2007), ‘Limited self-control, obesity and the loss of happiness’, WWZ Discussion Paper No. 0707, University of Basel. Tangney, June P., Roy F. Baumeister and Angie Luzio Boone (2004), ‘High selfcontrol predicts good adjustment, less pathology, better grades, and interpersonal success’, Journal of Personality, 72 (2), 271–324. Taylor, Paul, Cary Funk and Peyton Craighill (2006), ‘Americans see weight problems everywhere but in the mirror. A social trends report’, Pew Research Center, Washington, DC. Thaler, Richard H. (1981), ‘Some empirical evidence on dynamic consistency’, Economic Letters, 8 (3), 201–7. Weimann, Gabriel (1996), ‘Cable comes to the holy land: the impact on cable TV on Israeli viewers’, Journal of Broadcasting & Electronic Media, 40, 243–57. WHO Europe (2005), ‘The challenge of obesity in the WHO European Region’, Fact Sheet EURO, 13. Williams, Tannis Macbeth (ed.) (1986), The Impact of Television. A Natural Experiment in Three Communities, Orlando, FL: Academic Press. World Health Organization (2002), The World Health Report 2002, Geneva: World Health Organization.

124


APPENDIX 5A.1

SCALES APPLIED IN THE SWISS HEALTH SURVEY AND STUDIED IN STUTZER (2007)

Subjective Well-being (translated from Grob et al., 1991) To what extent do the following statements apply to you? ● ● ● ● ● ● ● ●

My future looks bright. I enjoy life more than most people. I am satisfied with how my life plans materialize. I deal well with those things in my life that cannot be changed. Whatever happens, I make the best out of it. I enjoy my life. My life is meaningful to me. My life is on the right track.

Possible answers: 1 5 completely wrong, 2 5 very wrong, 3 5 rather wrong, 4 5 rather accurate, 5 5 very accurate, 6 5 completely accurate. The responses are added together (SWB_tot) and summarized on a six point scale according to the following criteria: SWB_tot .5 44 & SWB_tot ,5 48 S SWB 5 6 SWB_tot .5 40 & SWB_tot , 44 S SWB 5 5 SWB_tot .5 36 & SWB_tot , 40 S SWB 5 4 SWB_tot .5 32 & SWB_tot , 36 S SWB 5 3 SWB_tot .5 28 & SWB_tot , 32 S SWB 5 2 SWB_tot .5 8 & SWB_tot , 28 S SWB 5 1 Mastery (based on four out of seven questions from Pearlin et al., 1981) When you think about your life, how strongly do you agree or disagree with these statements about yourself? ● ● ● ●

There is really no way I can solve some of the problems I have. Sometimes I feel that I’m being pushed around in life. I have little control over the things that happen to me. I often feel helpless in dealing with the problems of life.

Possible responses: completely agree, rather agree, rather disagree, completely disagree.


Table 5.A1

125

Covariates of subjective well-being in Switzerland, 2002 Ordered probit regression Coefficient

BMI Underweight Normal weight Overweight Obese Demographic factors Age 15–19 Age 20–24 Age 25–29 Age 30–34 Age 35–39 Age 40–44 Age 45–49 Age 50–54 Age 55–59 Age 60–64 Age 65–69 Age 70–74 Age 75–79 Age 80 and older Female Level of education Mandatory schooling Secondary general edu. Secondary prof. education Tertiary professional edu. University Marital status Married Single Widowed Divorced Separated Household composition 1 adult 2 adults

−0.053(*) −0.022 −0.047 0.125(*) 0.293** 0.195** 0.120**

z-value

Marginal effect for a score of 6

−1.91 −0.008(*) Reference group −1.08 −0.004 −1.42 −0.008

z-value

−1.96 −1.09 −1.46

−0.064(*) −0.040 −0.037 0.034 0.135** 0.353** 0.357** 0.405** 0.525** 0.086**

1.74 0.022 5.04 0.058** 4.39 0.036** 3.35 0.021** Reference group −1.78 −0.010(*) −1.01 −0.006 −0.88 −0.006 0.83 0.006 3.04 0.024** 5.86 0.071** 5.39 0.072** 5.74 0.084** 6.68 0.117** 4.11 0.014**

1.62 4.33 3.96 3.15 −1.84 −1.03 −0.90 0.82 2.83 4.96 4.53 4.73 5.27 4.12

−0.047 0.030

Reference group −1.09 −0.008 1.07 0.005

−1.12 1.07

0.104**

2.71

0.018*

2.57

0.021

0.50

0.004

0.49

−0.214** −0.077(*) −0.113** −0.238**

Reference group −6.37 −0.033** −1.76 −0.012(*) −3.02 −0.018** −3.18 −0.034**

−6.78 −1.84 −3.22 −3.78

0.098**

Reference group 3.28 0.016**

3.29

126

Table 5.A1


(continued) Ordered probit regression Coefficient

3 adults 4 adults and more No children 1 child 2 children 3 children and more Citizenship status Foreigner Main life circumstances Full-time job Part-time job Family business In education Unemployed Housework Retired Chronically ill Income Ln(equivalence income) No. of obs. Pseudo R2

0.095* 0.182** 0.049 0.063(*) 0.123*

z-value

Marginal effect for a score of 6

2.40 0.016* 3.97 0.033** Reference group 1.55 0.008 1.83 0.011(*) 2.53 0.022*

z-value

2.28 3.61 1.51 1.78 2.36

−0.119**

−4.06

−0.018**

−4.33

−0.059* 0.121 −0.004 −0.367** −0.027 −0.182** −0.578**

Reference group −2.56 −0.010** 1.04 0.022 −0.08 −0.001 −5.46 −0.048** −1.02 −0.004 −3.46 −0.028** −9.23 −0.065**

−2.62 0.97 −0.08 −7.20 −1.03 −3.75 −14.40

1.415** 15 108 0.021

8.43

0.422**

6.56

Note: Ordered probit regression. Further control variables not shown are ‘education not defined’, ‘other paid activity’, ‘other life circumstances’, ‘income not available’, ‘interview in French’, ‘interview in Italian’. Significance levels: (*) 0.05 < p 100,001 Capitol

City size 0.5

Percentage

0.4

0.3

0.2

0.1

0 1

Figure 8.2

2 City size

3

Histogram of respondents by city size

Tables 8.1 and 8.5a, in general respondents are happier in smaller cities and less happy in big ones. This also allows us to focus on the difference between rural areas, normal cities and large metropolitan areas. See Figure 8.2 for the histogram of city sizes. We then repeated the above exercise, but calculated average wealth for each city size level within each country. With this less aggregated specification for average wealth, we are able to include country dummies. In

Does inequality matter to individual welfare?

175

this instance, we again get the positive sign on individual wealth, but a negative and significant sign on average wealth. Thus, in Latin America, having wealthier neighbors or city-mates, controlling for an individual’s own wealth, lowers self reported happiness. This is similar to what Luttmar finds for earnings areas/PUMAs in the USA. Relative differences matter to respondents in Latin America, above and beyond the effects of individual income. The above regressions used the following formula, where X is a vector of individual characteristics that have been found to matter to happiness, such as marital status, education, health and so on: Y 5 X b 1 avg wealth b2 1 wealth b1. This is equivalent to the approach used in the Di Tella and MacCulloch paper described above.27 In addition, though, they decompose income into average national income and relative income, which is the difference between individual income and average income. We do the same for our wealth index, labeling the former variable avgwealth and the latter, relwealth. The sum of the two is individual income. This means that if the coefficients on the two variables are the same in a happiness regression, then happiness is increasing in wealth with no regard to relative status. For example, if average income increases by one measurement unit but a person’s income remains constant, then that person’s happiness increases by the coefficient on avgwealth but decreases by the coefficient on relwealth. If they are the same, then the person’s happiness is unchanged. If relwealth is more important than avgwealth, as one studying these variables might posit, then happiness would decrease. The equivalence between the Di Tella and MacCulloch and Luttmer techniques is demonstrated below: Y 5 X b 1 avgwealth b3 1 relwealth b4 (DiTella and MacCulloch) 5 X b 1 avgwealth b3 1 (wealth 2 avgwealth) b4 5 X b 1 avgwealth (b3 2 b4) 1 wealth b4 (Luttmar) Therefore, the Di Tella and MacCulloch approach provides the same information as the Luttmer technique, but making explicit the effects of relative as well as average wealth on happiness. Di Tella and MacCulloch use data from the US General Social Survey and the Eurobarometer and find that the effect of each of these components is the same – with a coefficient of 0.5 on each. Thus they reject the hypothesis that relative income

176


per se – above and beyond being a concern for personal income – matters. We repeat this exercise with our data for Latin America, although we must use our 0–11 wealth index rather than income, and thus do not take logs. In strong contrast to the findings for the USA and Europe, we find that the coefficient on average wealth is insignificant, while the coefficient on relative wealth is positive and significant (Table 8.3). The implication is that only relative income, above and beyond average wealth, matters positively to well-being in the region. Thus relative wealth contributes to greater than average happiness for those that are above mean income – the wealthy. It results in lesser than average happiness for those who are below mean income – the poor (as the value on relative wealth for those below mean income is negative, making them that much less happy). We repeated the same regressions with our country/city size specification of average and relative wealth, including country dummies. Each observation for relative wealth is the respondent’s distance from the mean wealth level of other respondents in similar sized cities in their country. As in the case of the country-level specification, we get an insignificant sign on average wealth, and a positive and significant sign on relative wealth, confirming the importance of relative wealth to Latin American respondents, this time using a different reference norm (Table 8.3). Unlike the results for Europeans and Americans in country-level and state-level studies, Latin Americans seem to be concerned with relative differences above and beyond their being a product of total individual income.28 The high levels of inequality in Latin America may underlie our respondents’ higher levels of concern for relative than absolute differences. We also explored the effects of relative and absolute wealth according to which quintile respondents were in. Much of the theory – and some of the empirical work on the role of relative versus absolute income – suggests that absolute income gains matter more to those below a certain minimum level of income. Relative income matters more, meanwhile, as people get wealthier and are no longer concerned about meeting basic needs. In an analogous sense, cross-country happiness comparisons find that economic growth leads to higher average happiness levels at low levels of per capita incomes but not at higher ones. Our results do not necessarily fit the theory. We grouped respondents into quintiles for our sample, based on our wealth index, to see if the coefficients on relative and absolute wealth differed by quintile. Thus, in each quintile category, the observation on average wealth is the average wealth for the respondent’s country; the respondent gets a 0 for the quintiles that they are not in, and the average wealth figure for the quintile they are in. Relative wealth works similarly; respondents get zero values for the


177

quintiles they are not in, and then the value of each respondents’ particular relative wealth is recorded in the quintile group that they correspond to. When we include our quintile variables in the regression, we find that average wealth remains insignificant, while individuals in quintiles 1, 2 and 5 retain concerns about relative wealth. (The coefficient on relative wealth for the fifth quintile is positive and significant at the 15 percent level only.) The coefficient on relative wealth for the fourth quintile is significant and negative, meanwhile, but only at the 10 percent level. This suggests that relative income differences make these respondents less happy, even though they are above mean income. This may be because their distance from the mean and/or the poor does not seem big enough, because they think their distance from the rich is too great, or both. The most significant effects seem to be those for respondents in the lowest two quintiles. As they are below mean income, the positive coefficient on relative wealth translates into lower happiness levels (Figures 8.3a and 8.3b). Inequality in Latin America seems to make the poor much less happy and the rich moderately happier. We then repeated our work at the country/city size reference group level. As above, this was a simple grouping of respondents by wealth quintile – in this instance based on the city size/country intersection. In this case respondents are grouped in quintiles which correspond to their country and also to their city size – small, medium and large. Thus respondents who live in big, wealthier cities are likely to be in a higher quintile when grouped at the country level than when compared to wealthier respondents in their city size reference group. We ran the same regression as above but for country/city size average and relative wealth, and including country dummies. In this instance, though, we again get an insignificant effect on average wealth, and strong (positive) effects of relative wealth for the wealthiest quintile. Relative wealth is positive and significant at the 10 percent level for those in the first and fourth quintiles (and a negative but insignificant effect on quintile 3). Thus the effect holds weakly for the poorest, but flips for those in the fourth quintile. Some of this may be specification driven: those respondents that are in quintile 4 at the country level are likely to be in quintile 3 when compared with other respondents in big cities, for example (Figures 8.3a and 8.3b). With the city size rather than country level reference group, the effects of relative wealth seem to be stronger for the rich rather than for the poor. It may well be that when compared to those in a smaller reference group, the poor feel less distanced from the rich, and therefore suffer less negative effects of inequality. The rich, meanwhile, may feel relatively better off with a smaller reference group than they do in a larger one. In other words, a respondent who is wealthy compared to those in their small town

178


Average wealth quintile

1 2 3 4 5

coefficient 0.0495 0.0552 –0.0114 0.1067 0.0613

Relative wealth quintile

1 2 3 4 5

0.1690 0.5994 0.5442 –0.2873 0.0450

z-score 0.61 0.80 –0.14 1.25 0.85 3.12** 3.35** 1.77 –1.82 1.49

0.6 avg wealth coefficient relwealth coefficient 0.4

0.2

0.0

–0.2

1

2

3 Quintile

4

5

Notes: Average wealth computed by country. Control variables: standard demographic variables, clustered by country.

Figure 8.3a

Relative wealth and happiness, by wealth quintiles

reference group is probably less wealthy in relative terms when compared to the larger, country-level reference group. Regardless of the nuances, relative differences seem to matter to well-being in the region, even when a different reference group is used. These differences seem to matter most to those at the top and bottom of the distribution. To explore differences across reference groups more closely, we ran the average/relative wealth regression separately for each city size. In a departure from most of the above findings, in which average wealth is insignificant, we get a positive and significant sign on average wealth for respondents in small cities. While the sign on relative wealth remains positive and significant, the value on the coefficient is smaller than that for average wealth (although the t-statistic is much higher). This suggests that both average and relative wealth levels matter to the well-being of those

Does inequality matter to individual welfare? coefficient

z-score

Average wealth quintile

1 2 3 4 5

0.018 0.028 0.041 0.019 0.002

0.49 0.78 1.02 0.49 0.04

Relative wealth quintile

1 2 3 4 5

0.065 0.085 –0.151 0.149 0.157

1.72 1.57 –1.22 1.92 3.94**

179

Avg. wealth by country and city size 2 Average wealth Relative wealth

Coefficient

1

0

–0.1

–0.2 1

2

3 Wealth quintile

4

5

Notes: Average wealth computed at country/citysml intersection. Control variables: standard demographic variables and country dummies, clustered by country/city size.

Figure 8.3b

Relative wealth and happiness, by wealth quintiles

in the small cities, our smallest size reference group, and also those with the lowest levels of average wealth. For our larger and wealthier reference groups (the larger city and country levels), in contrast, relative wealth seems to be the only wealth variable that matters.29

8.4

WHAT DO THESE RESULTS MEAN? A SIMPLE ILLUSTRATION

What does all of this mean in plain language? We illustrate in Figure 8.4 with a simple exercise comparing a typical respondent in the bottom and top quintiles from Honduras and Chile. Average wealth levels, on our 0–11 scale wealth index, are 4.78 for Honduras and 7.75 in Chile – almost twice

180

Happiness, economics and politics Rich Hondurans: wealth = 8.0 RichChileans: wealth = 10.3 Average Honduran wealth: 4.8

Honduran gap: 3.3 Chilean gap: 2.5

POOR

RICH

Honduran gap: 2.1 Chilean gap: 2.5 Average Chilean wealth: 7.8 Poor Hondurans: wealth = 2.6 Poor Chileans: wealth = 5.3 Happiness Gap = wealth gap * coefficient ÷ 4 Calculated Happiness Gap Chile wealth gap Honduras wealth gap Chile-Honduras difference difference * coefficient/4 = Honduran happiness differential

Poor –2.489 –2.142 0.347 0.43%

Rich 2.521 3.261 0.740 0.93%

Mean Happiness (1–5 scale) Wealth quintile 1 2 3 4 5 Total

Chile 2.54 2.74 2.77 2.94 3.08 2.79

Honduras 3.11 3.15 3.17 3.13 3.30 3.17

Overall 2.73 2.85 2.91 2.97 3.08 2.88

Mean Wealth (1–11 scale) Chile 5.26 7.00 8.00 9.00 10.27 7.76

Honduras 2.64 4.00 5.00 6.00 8.04 4.78

Overall 3.12 5.00 6.00 7.46 9.63 5.81

Note: Regionwide results: rich are 3.83 points higher than mean; poor are 2.68 points lower than mean. These gaps * .05/4 5 5%. happiness for the rich and 3%, happiness for the poor.

Figure 8.4 Happiness gap in Honduras and Chile as high in the latter. Average wealth in the bottom quintile in Honduras is 2.64 and in Chile is 5.26, over twice as high in the latter. Average wealth in quintile 5 in Honduras is 8.04 and in Chile it is 10.27. If rising personal wealth is sufficient to increase happiness, then the typical respondent in Chile should be happier than in Honduras, and a poor respondent in Chile should be much happier than in Honduras, while a wealthy one should be moderately happier. Yet, as the coefficient on average wealth is insignificant, it suggests this is not the case.


181

Instead, it is relative income, or the gap between each individual’s income and the average that matters. For the typical poor (quintile 1) respondent in Honduras, the gap between their income and the average is 2.14 points. In Chile the gap between the quintile 1 respondent and the average is 2.49 points. If we multiply the difference between these figures (0.35) times the coefficient from an OLS regression on relative wealth for the region (0.05) then we can assume that poor (quintile 1) respondents in Honduras are about one-half of 1 percent (0.017 divided by the 4-point happiness scale) happier than poor respondents in Chile, even though the average wealth levels of the poor in Chile are over twice as high!30 For those in the top quintiles, meanwhile, the gap between the wealth of those in the top quintile in Honduras and the average wealth is 3.26, while that for Chile is 2.52. If we multiply this difference (0.74) times the coefficient on relative wealth, we can assume that respondents in the top quintile in Honduras are about 1 percent happier than those in Chile, even though they are significantly less wealthy. Conducting a similar exercise at the regional level, meanwhile, we see that the average wealth of respondents in quintile 5 is 9.63, or 3.83 points higher than the regional mean wealth of 5.80, while the typical respondent in the first quintile, with a mean wealth of 3.12, is 2.68 points below the mean. Multiplying these gaps times the coefficient on relative wealth (0.05) and dividing by the 4-point scale, this implies that the rich are made 5 percent happier by their relative difference between themselves and the average, while the poor are made 3 percent less happy by inequality. This is a property of the skewed nature of the wealth distribution (which is even greater when using income as the measure rather than our wealth index), as the rich are further away from the mean than the poor are. It is important to note that this is an illustrative exercise which is intended to suggest the magnitude and direction of the effects that we find, rather than to attach a real value. There are a number of issues that we cannot resolve, such as the arbitrary nature of our scaling assumptions. Short of a viable alternative, these calculations assume that a move one point up or down the happiness scale has a similar effect regardless of where on that scale the respondent is. Yet it may well be that moving from somewhat unhappy to somewhat happy matters more to individuals’ lives than does moving from somewhat happy to very happy. We unfortunately cannot resolve that question here. Our findings suggest that inequality matters much more to well-being in the region – including for those in low income groups – than the standard theory implies. The latter stresses the importance of absolute income gains for those at the bottom of the distribution. Much of the literature on the effects of inequality (discussed above) posits that in contexts where it has

182


positive effects on happiness, it is because it signals future opportunities. This can only occur if beliefs in the prospects for upward mobility are high. In Latin America, in contrast, it is likely that persistently high levels of inequality signal to the poor that there are persistent disadvantages (and possibly other kinds of discrimination which our variables do not allow us to measure) and to the rich that there are persistent advantages. On reflection, these results should not come as a surprise in a region where inequality levels are higher than in the USA or Europe, and where the institutions equalizing opportunities, such as educational and labor markets, function far less efficiently and equitably.

8.5

PERCEPTIONS OF INEQUALITY AND WELLBEING

In addition to examining the direct effects of inequality on well-being, we attempted to capture the effects of inequality per se – for example, inequality defined more broadly than in income terms. In this section we attempt to capture this broader definition of inequality through a number of different variables which capture respondents’ perceptions of inequality, status, economic success and prospects for upward mobility. In previous work we find that respondents’ prospects for upward mobility (POUM), for example, are positively correlated with happiness and even with better labor market performance in future periods.31 Here we explore the relationship of several of these variables with well-being, and how that relationship varies according to reference group size. Two questions in particular allow us to separate feelings of status from other economic concerns or utility of wealth. One of these is a catch-all question asking ‘In general, how would you describe your present economic situation and that of your family?’ This variable is consistently one of the most significant to well-being, usually more so than any other except health. The other is the economic ladder question (ELQ), included in many other well-being surveys besides the Latinobarómetro, which asks respondents to place themselves on a 10-step ladder where the poorest are on step one and the richest on step ten. This question is also an important predictor of happiness, even when other questions about wealth are included. It is purely a relative ranking of wealth. When combined with the personal economy question, it allows us to decompose the utility of wealth into status and other effects. The frame of reference for the ELQ is left up to the respondent. The question does not specify whether the ladder represents their country or a smaller or larger reference group (such as the city or the world). Responses


183

Table 8.4 (a) Summary statistics for ELQ and economic satisfaction Means of variables

economic satisfaction (1-4 scale)

ELQ (1-10 scale)

wealth (1-11 scale)

2.88 2.96 3.01

3.66 3.74 4.25

4.38 5.34 6.56

Small town Medium city Big city

education happy (1-16 scale) (1-5 scale) 7.37 7.16 9.53

2.72 2.94

Table 8.4 (b) Correlation between different measures of wealth wealth

socio-economic status

personal economic satisfaction

wealth

1

socio-economic status*

0.5112

1

personal economic satisfaction ELQ

0.2521

0.2477

1

0.3956

0.327

0.3131

Note:

ELQ

1

* As judged by the interviewer.

suggest that people in fact take all of these frames into account. Wealthier countries have higher ELQ scores, suggesting international comparisons; ELQ increases (as does wealth) with city size, suggesting country-wide comparisons; but ELQ increases more slowly with city size than wealth does, indicating local comparisons. Meanwhile, personal economic satisfaction increases with city size, but given the increase in the other variables, there is actually a negative coefficient on the big city dummy variable in the regression. Summary statistics for the ELQ and personal economy question are in Table 8.4. What do these subjective variables, personal economy and ELQ, allow us to measure that the objective variables used before do not? For one thing, they may do a better job of measuring the elusive concept of relative status than looking at relative wealth alone. When regressing happiness on four measurements of wealth (wealth, ELQ, personal economy and socio-economic status, plus standard demographic variables and country dummies), the latter two subjective variables were more significant, both statistically and practically, than the objective variables. There is obviously some collinearity among these variables, but there is also a fair

184


amount of variance (the correlation is , 0.6 between any two of them, see Table 8.4) and the results hold up using both OLS and ordered logit regressions. It also holds up when measuring relative wealth at the country and country/city size level, with and without the relevant dummies. In fact, a happiness regression with our full set of 30 control variables (but not the personal economy question) gives an R-squared value of 0.069, while using the personal economy question as the only explanatory variable gives an R-squared value of 0.038. When we include both personal economic ranking and the ELQ in a happiness regression, we find that the coefficient on the personal economic ranking is much greater than that for the ELQ (Table 8.5). Even after adjusting for scale (there are twice as many possible responses on the ELQ as there are on the personal score), this suggests that people’s subjective assessment of their overall personal situation is much more important to their happiness than is their subjective assessment of their relative position. How can we reconcile this with our previous finding that relative wealth is all that matters to happiness? Indeed, it is consistent with that result. Relative wealth is presumably an important factor in the personal economy question. Since ELQ is not perfectly correlated with personal economy, the fact that the ELQ is significant at all indicates that relative status has bearing on happiness outside of a purely economic context. We looked at the determinants of ELQ scores (in other words, using the ELQ as the dependent variable). As in the case of happiness, ELQ scores display a U-shaped relationship with age, first decreasing until approximately 57 years and then increasing (a similar shape to that of happiness). Education, wealth and self-reported health are positively correlated with ELQ scores, while men and the unemployed are more likely to report lower ELQ scores. Since men are, on average, wealthier than women, this suggests that they also have higher economic standards than women do (Table 8.6). We then looked at how these scores varied according to where people live (city sizes). Wealth levels are, on average, higher in large cities than in small ones. In contrast, we found that respondents’ subjective personal economic rankings were lower in big cities and higher in small towns! (Table 8.7a). In our view this perceptions gap is in keeping with other findings in the happiness literature. It is suggestive of Luttmer’s recent work on US earnings areas and our own findings on average country-level wealth. In both cases respondents of similar income or wealth levels are less happy when their peers or compatriots have higher levels of wealth. James Duesenberry’s classic work on savings also resonates. He finds that, holding income levels constant, respondents that live in neighborhoods with higher average levels of wealth are less satisfied with their incomes than those that live in less wealthy neighborhoods.


Table 8.5(a)

185

The economic ladder and personal economic satisfaction

age age squared education wealth married male health unemployed self-employed retired student small town big city

coefficient

z-score

−0.0773 0.0007 0.0153 0.2035 0.1069 −0.0537 0.4354 −0.4945 −0.0822 0.0704 −0.1513 0.0809 −0.1110

−14.54** 11.66** 3.71** 24.96** 3.37** −1.81* 23.73** −8.48** −2.37** 0.97 −2.4 1.63 −3.26**

Notes: Ordered logit estimation of a 1-4 scale of personal economic satisfaction. Controls include standard demographic variables and country dummies.

Table 8.5 (b)

wealth average wealth small town big city

coefficient

z-score

0.2075 −0.1800 −0.0519 0.0647

14.22** −2.71** −0.69 1.08

Notes: Ordered logit estimation of a 1-4 scale of personal economic satisfaction. * t-statistics underneath coefficients.

Table 8.5 (c)

wealth socio-economic status ELQ persecon

coefficient

z-score

0.0361 0.0457 0.0704 0.5913

3.26** 1.83 4.76** 15.34**

Notes: Ordered logit estimation of a 1-5 scale of happiness. Controls include standard demographic variables and country dummies.

186

Table 8.6


Determinants of ELQ

age age squared education wealth married male health unemployed self-employed retired student small town big city constant

coefficient

z-score

−0.0259 0.0002 0.0587 0.1883 0.0340 −0.1075 0.2278 −0.1033 −0.0231 0.0976 0.0976 0.0472 0.0802 2.3861

−6.29** 4.93** 18.23** 30** 1.37 −4.6** 16.25** −2.27** −0.85 1.7 1.96 1.2 3.01** 20.53**

Notes: Low point of age: 57.9. OLS regression of the 1-10 scale economic ladder question. Controls: standard demographic variables including wealth and country dummies (not shown). Second regression which included educational inequality clustered by country.

ELQ, on the other hand, rises with city size (as does wealth), and even after controlling for socio-demographic data, ELQ rankings tend to be higher in big cities. Once again, this appears to be a reference group effect: people in small cities are more likely to know how others around them live than are those in medium or large ones. And, for the most part, they are fairly on par with their neighbors, as there is less variance in wealth levels in smaller cities. People in big cities, meanwhile, are probably aware that objective economic conditions in the countryside and smaller towns are worse than they are in the major cities. We next explored whether the average or relative aspects of the ELQ and personal economy rankings mattered more to happiness. We repeated the technique of separating the variables into an average component and a relative component for ELQ and the personal economy question.32 Using an F test, we could not reject the hypothesis that the coefficients for average and relative personal economy are equal and positive. On the other hand, average ELQ was completely insignificant, while relative ELQ was significantly positive. Thus, although people in, for example, large cities with wealthy neighbors realize that they are wealthier than people in rural areas, this brings them no additional happiness because they are concerned

187

−5.98** 4.56** 11.05** 21.71** 1.52 −4.29** 9.59** −2.59** −0.85

1.44 1.69 0.69 2.12**

0.098 0.098 0.047 0.080

z-score

−0.026 0.000 0.059 0.188 0.034 −0.107 0.228 −0.103 −0.023

coefficient

retired student small town big city

age age squared education wealth married male health unemployed self-employed

relative ELQ

0.091 0.091 0.214 −0.291

−0.026 0.000 0.056 0.184 0.030 −0.106 0.226 −0.105 −0.016

coefficient

Components of the ELQ and relative ELQ

1.34 1.58 4.47** −8.74**

−6.14** 4.59** 10.74** 22.21** 1.32 −4.26** 9.57** −2.6** −0.6

z-score

retired student smalltown bigcity avgELQ

age age2 yedu wealth married male health unemp selfemp

relative ELQ

Notes: OLS regression of a 1-10 scale of the economic ladder question. Controls include standard demographic variables and country dummies, clustered by country/city size. Average ELQ is computed at the country/city size level.

age age squared education wealth married male health unemployed selfemployed retired student small town big city

ELQ

Table 8.7a

0.093 0.093 0.157 −0.164 −0.341

−0.026 0.000 0.057 0.186 0.031 −0.106 0.227 −0.105 −0.019

coefficient

1.380 1.620 4.080** −5.490** −6.750**

−6.040** 4.580** 10.830** 22.000** 1.390* −4.280** 9.580** −2.600** −0.680

z-score

188

Table 8.7b


Average and relative ELQ and happiness

happy average ELQ relative ELQ

coefficient

z-score

0.1297 0.1245

1.76 6.65**

Notes: OLS regression of a 1-5 scale of happiness. Controls include standard demographic variables and country dummies, clustered by country/city size. Average ELQ is computed at the country/city size level.

happy average personal economy relative personal economy

coefficient

z-score

1.006 0.623

4.12** 14.9**

Notes: OLS regression of a 1-5 scale of happiness. Controls include standard demographic variables and country dummies, clustered by country/city size. Average personal economic satisfaction is computed at the country/city size level.

about their relative position vis-à-vis their rich neighbors in the cities. Furthermore, although a person’s ELQ rises with the average ELQ around them, that person’s relative ELQ tends to decrease with higher status neighbors. These findings are very much in keeping with our findings based on objective measures of relative and average wealth (Table 8.7b). We can use similar methods to look at intergenerational mobility. One question asks: ‘Do you believe that [your children] will live better, the same, or worse off than how you live today?’ Another question asks respondents to rank their children’s future status on the ELQ. The combination of the two can be used to examine effects of status and wealth shifts, where the first variable (POUMkids) allows us to factor out the effect of an overall rise in living standards. We can then create a variable, generational POUM, by subtracting respondents’ own ELQ score from their children’s, to look at expected shifts in status as well as wealth. At the country level, the highest average generational POUM score was for Chile (77 percent), while the lowest was for Costa Rica (19 percent). One can imagine that being in a fast growing economy with a great deal of economic change, such as Chile, would suggest better prospects for one’s children’s getting ahead than would living in one such as Costa Rica, where social insurance systems are basically sound, but where economic reform has been slow and growth performance moderate at best.


Table 8.8a

age age2 yedu wealth married male health unemp selfemp retired student smalltown bigcity

189

Generational POUM coefficient

z-score

−0.0162 0.0001 −0.0097 −0.0260 0.0185 −0.0302 0.0554 0.1203 −0.0159 −0.1544 0.0655 −0.1350 0.0935

−3.33** 2.54** −2.57** −3.52** 0.63 −1.09 3.36** 2.23** −0.5 −2.29** 1.09 −2.88** 2.98**

Notes: Low point of age: 59.43. OLS regression of a -10-10 scale of the generational POUM question. Controls include standard demographic variables and country dummies.

At the individual level, the generational POUM displayed a U-shaped age relationship, with the low point at 55 years. There was also an upside down U shaped relationship with education, with the turning point being 8.75 years of education, which is greater than primary but short of completed secondary school. This is closely linked to our findings on unemployment (discussed below), with the probability of being unemployed having a similar relationship with age and education, where the turning point of the latter is about 9.2 years of school. The unemployed are disproportionately represented among those with completed or almost completed secondary education (Table 8.9). Employed respondents with this educational profile, meanwhile, had lower expectations for their children’s mobility than did those with more or less education. Individuals with this profile have fared worse compared to those with university and higher technical skills, whose earnings have increased in both relative and absolute terms; and worse in relative terms compared to those with lower levels (basic only) of education.33 Those respondents that were actually unemployed had a higher generational POUM than the average. This probably reflects hope and optimism as much as objective conditions. Our earlier work suggests that most people retain hope for their children, even when in difficult straits.34 And given that the ELQ rankings of most unemployed people

190

Table 8.8b


Time to achieve desired standard of living How long will it take you to achieve your desired standard of living?

City size

Small town Medium city Big city Total

Will never 21-30 11-20 6-10 3-5 achieve years years years years 18% 38% 44% 100%

14% 17% 12% 12% 35% 38% 34% 36% 51% 46% 54% 52% 100% 100% 100% 100% coefficient

age age2 yedu wealth married male health unemp selfemp retired student smalltown bigcity

0.0139 −0.0001 −0.0354 −0.0558 −0.0799 0.0975 −0.2300 0.0763 −0.0845 −0.3161 0.1955 0.1343 0.0216

1-2 Have Total years achieved already 11% 7% 13% 39% 39% 36% 50% 55% 51% 100% 100% 100% z-score 2.6** −1.15 −8.93** −7.41** −2.6** 3.45** −13.3** 1.4 −2.55** −4.01** 3.52** 2.78** 0.67

Notes: Dependent variable monotonically increasing with age within the sample range. Ordered logit regression of a 1-7 scale of the time to achieve desired standard of living question. Controls include standard demographic variables and country dummies.

tend to be low, they would not have to rank their children particularly high to have a positive generational POUM. Scores were lowest in small towns and highest in the big cities, which not coincidentally have the greatest and most varied employment educational and employment opportunities. A related inequality perceptions variable was the time respondents thought it would take to reach their desired standard of living. The question was phrased as: ‘How long do you think it will take you to reach your desired standard of living?’ with possible answers ranging from ‘I already have it’ to several different year categories (1 to 2 years, 5 to 10 years and so on) to ‘never’. As shown in Table 8.9b, respondents who live in small


Table 8.9a

191

Cost of unemployment

unemployed

coefficient

z-score

−0.342

−6.05 **

Notes: Ordered logit regression of a 1-5 scale of happiness for 2004 data set. Controls include standard demographic variables and country dummies.

unemployed unemployed*gini coefficient

coefficient

z-score

−1.347 0.018

−5.18 ** 3.80 **

Notes: Ordered logit regression of a 1-5 scale of happiness for pooled 1997-2004 data set. Controls include standard demographic variables and year dummies.

unemployed (incomplete primary) unemployed (completed primary) unemployed (incomplete secondary) unemployed (completed secondary) unemployed (incomplete tertiary) unemployed (completed tertiary)

coefficient

z-score

−0.485 −0.205 −0.511 −0.562 0.027 −0.246

−3.83** −1.63 −4.46** −5.17** 0.13 −1.39

Notes: Ordered logit regression of a 1-5 scale of happiness. Controls include standard demographic variables and country dummies. Costs of unemployment by education level. Base case is illiterate.

towns are more likely to report ‘never’, while there was no significant difference in the responses of those that live in big cities from those in medium ones. It is likely that those in small towns, particularly rural ones, are well aware that the greatest opportunities for both education and employment are in larger urban areas rather than in their small towns. Meanwhile, those respondents with completed secondary school were the most likely to answer ‘never’ or the next lowest score. Again, trends in returns to education are likely playing a role. To help explain our findings, we examined a variable which asked respondents to choose what affected them most among the many reasons for which there was unequal treatment of people in their countries. Possible answers ranged from skin color to poverty to age. Respondents in small towns were more likely to say that poverty and lack of education

192


Table 8.9b

Fear of unemployment

small town big city

coefficient

z-score

−0.256 0.081

−4.34** 1.87

Notes: Ordered logit regression of a 1-5 scale of fear of unemployment. Controls include standard demographic variables (except dummy variables for jobs that are not in the workforce) and country dummies.

gini coefficient

coefficient

z-score

0.017

4.45**

Notes: Ordered logit regression of a 1-5 scale of fear of unemployment. Controls include standard demographic variables (except dummy variables for jobs that are not in the workforce).

were the primary reasons, while those in big cities were more likely to report corruption or the need to pay bribes. These findings suggest that both sets of respondents perceive that there is inequality and injustice. Yet the responses suggest that those in small towns feel that they do not have access to opportunity due to their own poverty and education (explaining a higher tendency to the ‘never’ responses on the above question), while those in big cities are more likely to believe that opportunities and access are monopolized by those with greater means or connections. Those in small towns seem more concerned about their own poverty compared to the rest of society, while those in large cities are more concerned with their access to opportunities compared with more ‘connected’ folks. In both instances the concerns cited run in the opposite direction of an interpretation in which inequality signals opportunity and mobility, which is more typical for the USA and for Europe.

8.6

THE COSTS OF UNEMPLOYMENT AND INEQUALITY

Continuing with our methodology of looking at the effects of inequality on specific subgroups, we here analyse the impact on happiness of unemployment. Previous happiness research has found that unemployment is


193

one of the most traumatic events that can happen to people. One of the reasons for this is of course the loss of income; however, there is also a cultural stigma to unemployment that impacts happiness. The typical unemployed person in our study is a male who has attended some high school (on average ten years of education). The unemployed percentage of the population increases with city size. This may be an artifact of the data, however, because people in rural areas are more likely to be outside the formal labor force altogether and unemployment is a less relevant concept for them. We sought additional information about how inequality affects welfare via our knowledge of the effects of unemployment on happiness. The strength of these effects – for example, the ‘costs’ of unemployment – tend to vary across countries and regions. We build from the work of others. Di Tella, MacCulloch and Oswald find that respondents in the USA and Europe are made more unhappy by higher unemployment rates than they are by inflation. In other words, the typical respondent – including employed respondents – would accept higher levels of inflation if it would eliminate the insecurity associated with higher unemployment rates. Several studies have shown that increased unemployment in general lessens the impact on unemployed individuals. Clark and Oswald (1994) find that the unemployed in Britain are less unhappy in districts where the unemployment rate is higher. The costs to happiness that comes from the decreased probability of finding a job seems to be lower than the gains to happiness that come from being less stigmatized and accompanied by more unemployed counterparts. Similarly, Stutzer and Lalive (2004) find that unemployed respondents are less happy in cantons that have voted to reduce unemployment benefits in Switzerland (controlling for benefit levels), as the stigma from unemployment is higher. As discussed above, Eggers, Gaddy, and Graham find that both employed and unemployed respondents are happier in regions with higher unemployment rates in Russia. We, too, find positive effects of general unemployment on happiness, both using an unemployment rate calculated from our own data and the latest statistics available from the United Nations Economic Commission for Latin America and the Caribbean (ECLAC). These are country-wide unemployment rates and have statistically significant positive effects on happiness. As in the above studies, higher overall unemployment may reduce the stigma effect on individuals. The results must be tempered, though, by the limited information that open unemployment rates can provide in a region with high levels of informal employment (exceeding 50 percent in a few countries). Inequality in countries also has an effect on happiness among the

194


unemployed. Using our pooled data set from 1997–2004, we ran a standard happiness regression, including a control variable for being unemployed, and then adding interaction terms for being unemployed in a high or low Gini country. We find that the costs to happiness of being unemployed are lower in higher Gini countries (Table 8.9a). In other words, unemployed respondents in countries with higher inequality are actually happier than those in countries with low inequality. Countries with high inequality are also, on balance, poorer than other countries, so the unemployed may have less far to fall in those countries. Another reason may be the higher levels of informal employment in the poorer and more unequal countries in the region, thereby resulting in less stigma for the unemployed. Or it may be due to some other countrylevel unobservable that we are not accounting for. And, while the costs of being unemployed are lower in higher Gini countries, fear of unemployment (among the employed) is higher, in keeping with our intuition about greater levels of informality and associated insecurity. Thus in higher inequality countries the lower stigma for the unemployed is accompanied by greater insecurity for the employed. Job instability has particularly affected those with a high school level of education, and if we look at the happiness impact of unemployment among different educational groups, it turns out that, in addition to having the highest rate of unemployment, those with a high school education are also made most unhappy by unemployment. In fact, unemployment has a statistically insignificant effect on happiness on the ends of the education spectrum (Table 8.9b). College-educated people are also less likely to fear unemployment than those with less education. And unemployment is a less relevant concept for the illiterate, who are most likely to be outside the formal labor market to begin with, and those with higher education are more likely to be able to find another job than those with secondary school education. We also examined the costs to unemployment by city size. As in the case of our Gini coefficients, we find that the costs of unemployment are lower in big cities than they are in small towns, suggesting that there is a lower stigma effect in big cities. Yet also as in the case of inequality (as measured by the Gini), fear of unemployment is higher in the big cities, presumably because labor markets are more integrated into the international economy and volatility is more of a factor, while relying on farming as a safety net is not an option the way it is in smaller towns (Table 8.9b). Our findings are suggestive of how the costs of being unemployed can vary across countries and according to different measures of inequality. Inequality seems to be correlated with a lower ‘stigma’ for the unemployed, but with a higher fear of unemployment for the employed.


8.7

195

CONCLUSIONS

This chapter was an attempt to explore the effects of relative income differences, as well as of inequality more broadly defined (inequality per se), on well-being in Latin America, the region with the highest inequality in the world. We find large and consistent effects of relative income differences (and concerns for relative income differences) on well-being. At the same time, average country and city size wealth, holding individual incomes constant, had no significant effects on well-being, with the exception of the smaller, poorer cities. This suggests that inequality or relative position matters more in Latin America than it does in other places, such as Europe and the USA. Rather surprisingly, the strong effects of inequality (or relative wealth more specifically) held for both the poorest and the wealthiest groups. The effects of relative income contribute to the happiness of those who are above average income and result in lower happiness levels for those who are below it. A back of the envelope calculation suggests that inequality in the region makes those in the highest quintiles 5 percent happier than the average and those in the poorest quintile 3 percent less happy, regardless of differences in average or individual wealth levels within and across these groups. Various studies of inequality and well-being in the USA and Europe find modest effects in one direction or the other (positive or negative), or else inconclusive evidence that inequality matters at all. A common explanation for these mixed findings is that in Europe and the USA inequality can be a signal of income mobility and opportunity as much as it is a signal of injustice. In Latin America, a region where the gaps between the poor and the wealthy are much larger and more persistent, inequality seems to be a signal of persistent advantage for the very wealthy and persistent disadvantage for the poor, rather than a signal of future opportunities. We also analysed trends in respondents’ perceptions of inequality, rank and opportunity as a means to gauge the effects of broader, non-income definitions of inequality – inequality per se – on well-being. Our findings support the importance of relative differences in these realms to wellbeing, and suggest that they may be more important than income-based differences. And concerns for status or relative differences were higher among those respondents whose reference norms are higher – in places where there is higher average wealth and with greater variance in levels (and probably more information and awareness), as in big cities. Inequality and perceived inequality play a mediating role in the effects of unemployment on well-being. Higher levels of inequality seem to lower the costs of unemployment for the unemployed (perhaps by reducing stigma), but increase insecurity or fear of unemployment for the employed.

196


Our findings are, by definition, suggestive rather than conclusive. We set out to explore the effects of relative income differences on well-being, using a range of measures, including some unconventional ones, as well as to try and shed light on an as yet loosely defined concept – inequality per se – using perceptions about status and opportunity. Most of the measures suggest that inequality has perverse effects on welfare in Latin America. It is associated with lower well-being for those at the bottom of the distribution in particular and for those with below average wealth levels in general. Our findings on perceptions of status and opportunity run in the same direction. Not all of the effects of inequality are negative; the wealthy are made happier by higher relative differences. Yet this is not necessarily optimal in a normative sense (depending on one’s priors). And while the unemployed seem to suffer lower well-being costs in contexts of higher inequality, it is also linked to higher fear of unemployment. The implications of our findings for policy are less clear. The modest evidence that we have on support for redistribution in the region suggests that there is not much support for it among the poor – precisely the group that is most hurt by inequality. At the same time, the concerns that we find among respondents about poverty and lack of equal access to education and other opportunities suggest that it would be much easier – and arguably much more efficient – to generate support for policies that can help increase access to education and opportunity. That, however, is a major challenge, and the subject for another paper.

NOTES 1.

2. 3. 4.

5. 6. 7.

The authors thank William Dickens, Richard Easterlin, Branko Milanovic, Andrew Oswald, Alois Stutzer and Peyton Young for helpful comments on earlier drafts. An earlier, summary version of this chapter was published in the Journal of Economic Inequality, January 2006. For an excellent description of the role of equity norms in mediating a number of important economic outcomes that markets alone cannot determine, see Young (1995). Nancy Birdsall and I discuss these two kinds of inequality at length in the introduction to Birdsall et al. (1998). Theoretical studies include the works of Danny Quah, Sam Bowles and Herbert Gintes, Steven Durlauf, Francois Bourguignon, Robert Frank and Roland Banabou, among others. Earlier works include those of John Rawls and A.C. Pigou. For an excellent summary of many of the issues involved, see Arrow et al. (2000). For a description of the possible biases in survey research, see Bertrand and Mullainathan (2001). For a complementary approach which focuses on procedural utility, see Frey et al. (2004). See, among others, Benabou and Ok (2001), Boeri et al. (2001), Graham (2003), Pitketty (1995) and Schwarze and Harpfer (2004). Acemoglu and Robinson (2002), meanwhile,


8. 9. 10.

11.

12. 13.

14.

15.

16. 17.

18. 19. 20.

21. 22.

197

develop a political economy theory of the Kuznets Curve in which the reduction of inequality depends on a credible threat of revolution by the poor. They also find stark differences among different income groups, however (discussed below). See Alesina et al. (2004). See Clark (2003) on Britain and Tomes (1986) on men in Canada. Both of these studies are discussed in greater detail below. Easterlin (1974, 1995, 2001, 2003) used 30 surveys from 19 countries, including some developing countries. Similar results, or minor modifications of them, have been found by both economists and psychologists. See, among others, Blanchflower and Oswald (2004) and Diener et al. (1993). For an excellent review of much of this literature, see Frey and Stutzer (2002). See Diener et al. (1993), Easterlin (1974, 2003), Frank (1999) and Luttmer (2004). A recent cross-country study by Ball and Chernova (2004), based on the World Values Survey, finds that the effects of relative income on happiness are up to four times as great as those of absolute incomes, although the effects of absolute incomes are still positive and significant. For an excellent definition of these latter inequalities, often called horizontal inequalities, see Ravallion (2004). For example, for a lognormal distribution (often used to model income/wealth distribu2 tions) based on a normal distribution N (m, s2) , the mean is em1s /2 and the median is em. Since the mean is conditional on the variance but the median is not, a mean-preserving increase in the variance will increase the ratio of the mean to the median (Aitchison, 1957; Moene and Wallerstein, 2003). Opinion polls in Russia suggest that the inequality that most matters to the average citizen is that between Moscow – the reform capital – and the rest of the country, rather than the more general cross-regional differences that are captured by the Gini (VTsIOM, 2004). Boeri et al. (2001), meanwhile, find that most Europeans want to shift expenditures from pensions to unemployment insurance. This effect is stronger where labor markets are more rigid, such as in Italy and Spain (for example, it is harder to fire people so results in less labor mobility, flexibility and higher unemployment). We also squared the wealth variable in order to see if there was a quadratic effect, which would suggest a shift in attitudes (and support for lower taxes) for the very wealthy. Yet we did not find evidence of such a shift. The question on taxes and redistribution (LOWTAX) is phrased: ‘do you support lower taxes, even if welfare spending suffers’, making it very clear to respondents that there are trade-offs to lower taxes. See Graham and Sukhtankar (2004) and Graham (2003). See, for example, Diener and Biswas-Diener (2000). There is a debate among psychologists on the optimum scale for well-being questions. While there is not complete agreement on the range, most agree that a longer scale than 1 to 4 allows for more accuracy (Cummins and Gullone 2002). Blanchflower and Oswald (2004) get a correlation coefficient of 0.56 for British data for 1975–1992 where both questions are available; Graham and Pettinato (2002b) get a correlation coefficient of 0.50 for Latin American data for 2000–01, in which alternative phrasing was used in different years. The Dominican Republic was included for the first year in 2004, raising the country total to 18. Due to logistical and other constraints, the survey only has 70 percent coverage in Chile, 51 percent in Colombia and 30 percent in Paraguay. The survey is produced by Latinobarómetro, a non-profit organization based in Santiago, Chile and directed by Marta Lagos (http://www.latinobarometro.org). The first survey was carried out in 1995 and covered eight countries. Access to the data is by purchase, with a 4 year lag in public release. Graham has worked with the survey team for years and assisted with fund raising, and therefore has access to the data.

198 23.

24.

25.

26.

27. 28. 29.

30. 31. 32. 33. 34.

Happiness, economics and politics The correlation coefficient between the interviewer’s assessment of SES and our index is 0.50. We also estimated a latent wealth variable using primary component analysis of the items in the wealth index, but this alternative does not substantively change our results (Filmer and Pritchett, 2001). Another major difference is that the self-employed are happier than average in the USA and Europe but less happy in Latin America. While these respondents are selfemployed by choice in the former context, in the latter they are in the informal sector due to lack of other alternatives. For this and all other regressions involving the Gini coefficient, we replaced the number by the standard deviation from the regional mean in order to make the coefficients easier to interpret. (In other words, we now think of the differences in terms of standard deviations rather than as incremental changes between closely bunched numbers.) We used the most recent number available; the years range from 1999 to 2004. Since the Gini coefficient changes so slowly for most countries, this should not affect the results. The mean for the countries involved was 53.7, from a minimum of 44.6 to a maximum of 59. The Gini coefficient for the USA, in comparison, is 41.8 [United Nations, 2004]. An additional issue is that the phrasing and placement of the happiness question changed slightly from 1997–99 to all of the subsequent years. In order to control for any bias introduced by this, we split the sample according to happy question type, and get essentially the same results. These split sample regressions available from the authors upon request. Di Tella and MacCulloch (2003). At the PUMAs level, Luttmer does find that Americans are concerned about relative income differences. The coefficient on average wealth for small cities is 0.245 and the t-statistic is 1.920; on relative wealth, it is 0.152 and the t-statistic is 5.815; for medium cities the coefficient for relative wealth is 0.103 and the t-statistic is 3.716; for large cities these figures are, relatively, 0.110 and 4.784. In order to calculate these coefficients, we used OLS to regress happiness, although we used ordered logistic regression in the rest of the chapter. Graham and Pettinato (2002b) and Graham et al. (2004). In other words, the average ranking for the relevant reference group – country or country/city size – and the distance of the individual respondent’s ranking from that average. Behrman et al. (2001). In perceptions surveys in Peru, for example, we found that a much higher percent of respondents ranked their own past progress negatively than assessed their children’s future prospects negatively [Graham and Pettinato, 2002a].

REFERENCES Acemoglu, D. and J. Robinson (2002), ‘The political economy of the Kuznets Curve’, Review of Development Economics, 6 (2), 183–203. Aitchison, J. (1957), The Lognormal Distribution, Cambridge: Cambridge University Press. Alesina, A., R. Di Tella and R. MacCulloch (2004), ‘Inequality and happiness: are Europeans and Americans different?’, Journal of Public Economics, 88, 2009–42. Arrow, K., S. Bowles and S. Durlauf (2000), Meritocracy and Economic Inequality, Princeton, NJ: Princeton University Press. Ball, R. and K. Chernova (2004), ‘Absolute income, relative income, and happiness’, paper presented at the International Society of Quality of Life Studies, Philadelphia, November.


199

Barro, R. (2001), ‘Inequality, growth, and investment’, in K. Hassett and R.G. Hubbard (eds), Inequality and Tax Policy, Washington, DC: The AEI Press, pp. 1–38. Behrman, J., N. Birdsall and M. Szekely (2001), ‘Economic reform and wage differentials in Latin America’, Carnegie Endowment Working Papers. Benabou, R. (2000), ‘Unequal societies: income redistribution and the social contract’, American Economic Review, 90, March, 96–129. Benabou, R. and E. Ok (2001), ‘Social mobility and the demand for redistribution: the POUM hypothesis’, Quarterly Journal of Economics, 116, 447–87. Bertrand, Marianne and Sendil Mullainathan (2001), ‘Do people mean what they say? Implications for subjective survey data’, American Economic Review, 91, 67–72. Birdsall, N. (2000), ‘Unequal societies: income distribution and the social contract’, American Economic Review, 90, 96–129. Birdsall, N., D. Ross and R. Sabot (1995), ‘Inequality and growth reconsidered’, World Bank Economic Review, 9, 477–508. Birdsall, N. and J.L. Londono (1997), ‘Asset inequality matters: an assessment of the World Bank’s approach to poverty reduction’, American Economic Review, 87, 32–7. Birdsall, Nancy and Carol Graham (2000), New Markets, New Opportunities: Economic and Social Mobility in a Changing World, Washington, DC: The Brookings Institution Press and the Carnegie Endowment for International peace. Birdsall, N., C. Graham and R. Sabot (eds) (1998), Beyond Trade-offs: Market Reforms and Equitable Growth in Latin America, Washington, DC: The Brookings Institution Press and the Inter-American Development Bank. Blanchflower, D. and A. Oswald (2003), ‘Does inequality reduce happiness? Evidence from the states of the USA from the 1970’s to the 1990’s’, Journal of Public Economics, 88, 1359–87. Blanchflower, D. and A. Oswald (2004), ‘Well-being over time in Britain and the USA’, Journal of Public Economics, 88, 1359–87. Boeri, T., A. Borsh-Supan and G. Tabellini (2001), ‘Welfare state reform: a survey of what Europeans want’, Economic Policy, London. Brown, G., J. Gardner, A. Oswald and J. Qian, (2003), ‘Does wage rank affect well being?’, Paper presented to the Brookings/Warwick Conference on ‘Does Inequality Matter? Lessons from the Economics of Happiness’, Washington, DC, 5–6 June. Clark, A. (2003), ‘Inequality-aversion and income mobility: a direct test’, DELTA Working Paper 2003-11, Paris, France. Clark, A. and A. Oswald (1994), ‘Unhappiness and unemployment’, The Economic Journal, 104, 648–59. Contreras, D., R. Cooper, J. Herman and C. Neilson (2004), ‘Dinamica de la Pobreza y Movilidad Social: Chile 1996–2001’, Mimeo, Santiago, Chile. Cummins, R. and E. Gullone (2002), ‘Why we should not use 5-point Likert scales: the case for subjective quality of life measurement’, in Proceedings: Second International Conference on Quality of Life in Cities, Singapore: National University of Singapore, pp. 74–93. Diener, E. and R. Biswas-Diener (2000), ‘Income and subjective well-being: will money make us happy?’, Mimeo, University of Indiana, Champlain. Diener, E., E. Sandvik, L. Seidlitz and M. Diener (1993), ‘The relationship between

200


income and subjective well-being: relative or absolute?’, Social Indicators Research, 28, 195–223. Di Tella, R. and R. MacCulloch (2003), ‘Income, happiness, and inequality as measures of welfare’, paper presented to the Brookings-Warwick University Conference on ‘Why Inequality Matters: Lessons from the Economics of Happiness’, Washington, DC, 5–6 June. Easterlin, R. (1974), ‘Does economic growth improve the human lot? Some empirical evidence’, in P. David and M. Reder (eds), Nations and Households in Economic Growth, New York: Academic Press, pp. 804–79. Easterlin, R. (1995), ‘Will raising the incomes of all increase the happiness of all?’, Journal of Economic Behavior and Organization, 27, 35–48. Easterlin, R. (2001), ‘Life cycle welfare: trends and differences’, Journal of Happiness Studies, 2, 1–12. Easterlin, R. (2003), ‘Income and happiness: towards a unified theory’, The Economic Journal, 111, 465–84. Eggers, A., C. Gaddy and C. Graham (2006), ‘Unemployment and well being in Russia in the 1990’s: can society’s suffering be individuals’ solace?’, Journal of Socio-Economics, 35 (2), 209–42. Filmer, D. and L. Pritchett (2001), ‘Estimating wealth effects without expenditure data or tears: an application to educational enrollments in states of India’, Demography, 38 (1), 115–32. Frank, R. (1999), Luxury Fever: Money and Happiness in an Era of Excess, New York: The Free Press. Frey, B. and A. Stutzer (2002), ‘What can economists learn from happiness research?’, Journal of Economic Literature, 20, 402–35. Frey, B., M. Benz and A. Stutzer (2004), ‘Introducing procedural utility: not only what but also how matters’, Journal of Institutional and Theoretical Economics, 160 (3), 377–401. Graham, C. and S. Pettinato (2001), ‘Happiness, markets and democracy: Latin America in comparative perspective’, Journal of Happiness Studies, 2 (3), 237–68. Graham, C. and S. Pettinato (2002a), ‘Frustrated achievers: winners, losers, and subjective well being in new market economies’, Journal of Development Studies, 38 (4), 100–40. Graham, C. and S. Pettinato (2002b), Happiness and Hardship: Opportunity and Insecurity in New Market Economies, Washington, DC: The Brookings Institution Press. Graham, C. and S. Sukhtankar (2004), ‘Is economic crisis reducing support for markets and democracy in Latin America? Some evidence from surveys of public opinion and well being’, Journal of Latin American Studies, 36, 349–77. Graham, C. and P. Young (2003), ‘Ignorance fills the income gulf’, The Boston Globe, 23 June. Graham, C. and P. Young (2004), Rags to Riches? The American Dream is Less Common in the United States than Elsewhere, Washington, DC: The Century Foundation Press. Graham, C., A. Eggers and S. Sukhtankar (2004), ‘Does happiness pay? some evidence from panel data for Russia’, Journal of Economic Behavior and Organization, 55. Hagerty, M. (1999), ‘Social comparisons of income in one’s community: evidence from national surveys of income and happiness’, Mimeo, San Diego, California.


201

Kingdon, G. and J. Knight (2004), ‘Community, comparisons and subjective wellbeing in a divided society’, WPS No. 2004-21, Centre for the Study of African Economies, Department of Economics, University of Oxford. Luttmer, Erzo (2004), ‘Neighbors as negatives: relative earnings and well-being’, Quarterly Journal of Economics, 120 (3), 963–1002. McMurrer, D. and I. Sawhill (1998), Getting Ahead: Economic and Social Mobility in the United States, Washington, DC: Urban Institute Press. Moene, K. and M. Wallerstein (2003), ‘Earnings inequality and welfare spending: a disaggregated analysis’, World Politics, 55 (4), 292–303. Oswald, A. with Jonathan Gardner (2004), ‘How is mortality affected by money, marriage and stress?’, Journal of Health Economics, 23, 1181–207. Pitketty, T. (1995), ‘Social mobility and redistributive politics’, Quarterly Journal of Economics, 110, 551–84. Ravallion, M. (2004), ‘Competing concepts of inequality in the globalization debate’, in S. Collins and C. Graham (eds) (2004), Brookings Trade Forum 2004: Globalization, Poverty, and Inequality. Washington, DC: The Brookings Institution Press. Schwarze, J. and M. Harpfer (2004), ‘Are people inequality averse and do they prefer redistribution by the state? Evidence from German longitudinal data on life satisfaction’, Frankfurt: Working Paper No. 707, German Institute for Economic Research. Stutzer, A. and R. Lalive (2004), ‘The role of social work norms in job searching and subjective well being’, Journal of the European Economic Association, 2 (4), 696–71. Tomes, N. (1986), ‘Income distribution, happiness, and satisfaction: a direct test of the interdependent preferences model’, Journal of Economic Psychology, 7, 425–46. VTsIOM (2004), ‘Moskva glazami Rossiyan’ (‘Moscow as seen by Russians’), Press Release No. 113, 3 September. Young, P. (1995), Equity: In Theory and Practice, Princeton, NJ: Princeton University Press. United Nations Development Program (2004), Human Development Indicators, www.hdr.undp.org.

9.

Perceptions of discrimination, effort to obtain psychological balance, and relative wages: can we infer a happiness gradient? Arthur Goldsmith*

1

OVERVIEW

There is ample evidence that blacks receive lower wages than whites with comparable characteristics and background (Altonji and Blank, 1999; Couch and Daly, 2002; Darity and Mason, 1998; Goldsmith et al., 2006a; Mason 1997). Estimates of the racial wage gap for males typically range between 12–15 percent.1 Social psychologists report that relative income is an important determinant of happiness or well-being. Thus, to the extent that black workers face wage discrimination there is likely to be an associated gap in well-being. This chapter offers, and tests, a theory of how a person’s perception that they face workplace discrimination influences their behavior, and hence, their wages. The theory is developed by extending the neoclassical theory of wage determination to incorporate the insights of Festinger’s (1957) theory of cognitive dissonance – one of the most innovative and prominent theories of behavior in social psychology.2 Our model advances the notion that workers simultaneously derive satisfaction from both the wage they earn and from being in psychological balance which is governed by a perception of ‘fair’ treatment. We interpret Festinger as asserting that thoughts or cognitions that do not ‘fit’ together, result in dissonance and that thoughts must be largely consistent for a person to attain psychological balance. In our view, a person who believes they face wage discrimination is thrust into an unbalanced psychological state since they think they are not treated fairly – the wage they receive falls short of their perceived contribution to the revenues of the firm that is governed by their skills and effort. This person can be expected to make cognitive adjustments in an effort to reach psychological balance.

202

Perceptions of discrimination: can we infer a happiness gradient?

203

Employers may discriminate by undervaluing the talents of minorities, due to negative stereotypes they hold. If employers learn the error of their ways by observing workers, then employees who believe they are discriminated against can simply wait for their employer to adjust their perspective and the wage they pay, leading them to psychological equilibrium. The dangers of this strategy are two-fold. First, as effort declines so will productivity and wages. Second, if effort declines too much the worker may be fired. Another approach might be to reduce their effort level so their productivity falls into line with their wage. This strategy of curtailing effort to reach psychological balance will result in a negative relation between wages and perceptions of wage discrimination. Alternatively, workers who believe they are subject to wage discrimination may decide that greater effort will break down negative stereotype beliefs about themselves held by employers. The idea is that this will lead employers to upgrade their evaluation of the worker’s productivity and to pay them a higher wage. In this scenario workers who believes they are subject to wage discrimination will actually earn a higher wage. Thus, the impact on a worker’s wage of perceived exposure to wage discrimination is unclear. A number of studies have attempted to determine if worker selfreports of exposure to workplace discrimination are closely aligned with conventional statistical measures of wage discrimination (Darity et al., 2006; Hallock et al., 1998; Kuhn, 1987). However, there is a paucity of empirical work attempting to offer an explanation for how perceptions of discrimination might influence wages. Neumark and McLennan (1995) assert that people who believe they face workplace discrimination have less of an incentive to invest in human capital leading to lower wages. Our approach differs from theirs in that we investigate how perceptions of discrimination might influence wages for workers with a given level of skills as they seek psychological equilibrium. Using data drawn from the Multi City Study of Urban Inequality (MCSUI) we estimate wage equations to shed light on how individuals who believe they face discrimination respond. This chapter is organized as follows. In Section 9.2 we present Festinger’s theory of cognitive dissonance and integrate this with the conventional neoclassical theory of wage determination. In addition, we discuss how a perception of wage discrimination may lead to cognitive dissonance that destroys psychological balance. Strategies for restoring psychological balance are identified and their influence on the racial wage gap is determined. Section 9.3 documents the frequencies of perceptions of wage discrimination because of race for our subsample of white and black workers drawn from the MCSUI. The data are described in this section along with our empirical procedures. In addition, we present estimates

204


of the influence of perceived wage discrimination on wages. Concluding thoughts are offered in Section 9.5.

9.2 9.2.1

COGNITIVE DISSONANCE AND PSYCHOLOGICAL EQUILIBRIUM Cognitive Dissonance Theory

Festinger’s theory of cognitive dissonance posits a link between attitudes and cognitive processes that may lead to behavioral change.3 His theory suggests that individuals seek harmony between their cognitions or thoughts. Festinger hypothesized that disharmony makes a person uncomfortable and tense. The discomfort fostered by dissonance motivates cognitive changes designed to restore harmony. Festinger (1957, p. 9) defined anything a person perceives to ‘know’ about themselves, others and their environment as a cognitive element. The relation between any two cognitive elements may be dissonant, consonant or irrelevant. A dissonant relation exists between two cognitive elements when, in the perceiver’s mind, they do not seem to ‘fit’ together (ibid., p. 13). Festinger proposed that the amount of dissonance associated with any two inconsistent cognitive elements grows with the importance of these elements to the perceiver. He expects the importance of a cognitive element to depend on two factors, the intensity with which an attitude or belief is held and the proximity of the element to the individual’s self-perception. Finally, as the magnitude of the inconsistency rises so does the pressure to eliminate or, at least, to reduce it. Festinger believed that dissonance is typically resolved by altering an inconsistent cognition, reducing its importance, or through the availability of new information. The next section explores how Festinger’s theory of cognitive dissonance can explain the behavior of persons that believe they face wage discrimination leading to changes in their wage rate. 9.2.2

Perceptions of Wage Discrimination, Workplace Effort and Cognitive Dissonance

Consider a firm that hires white (w) and black (b) workers from a pool of labor, and believes initially that all the members of a racial group are homogenous. Over time employers are presumed to observe worker effort and thus treat workers as heterogenous. Suppose the firm has an initial expectation (E) of worker productivity and believes that the marginal


205

product of black workers is less than the expected productivity of white workers, (MP E ) b , (MP E ) w, although black and white workers are actually equally productive. Following conventional neoclassical economic theory suppose the firm pays workers a real wage (W) equivalent to their expected marginal product MPE since firms only observe, and hence learn, an employee’s actual productivity over time. { (W) b 5 (MP E ) b } , { (W) w 5 (MP E ) W } .

(9.1)

The racial wage gap identified in Equation (9.1) is the result of statistical discrimination, since the firm has formed an inaccurate perception of black worker productivity based on stereotypical beliefs (Arrow, 1973; Coate and Loury, 1993; Lundberg and Startz, 1983).4 This type of discrimination may well occur at firms since research by social psychologists (Fiske and Ruscher, 1993; Fiske et al., 2002) provides empirical evidence that persons tend to hold negative stereotypes based on race and ethnicity.5 Workers are presumed to form their own expectation of their produc|| tivity (MP E ) . Psychologists Carver and Scheier (1981, p. 186) argue that individuals establish a target or goal called a ‘standard’ to guide their behavior. In our view, the typical person who takes a job establishes a standard of being ‘treated fairly’, which entails earning a real wage at least equal to their judgment of their marginal product. | | E) (W) $ (M P .

(9.2)

We assert that psychological balance or equilibrium occurs when workers believe they are treated fairly. In addition to having a sense of their own productivity, workers are assumed to know the firms judgment of the productivity of white workers, since firms tout this as the performance level of the ideal worker. Suppose white workers’ expectations of productivity are equivalent to their firm’s expectations of their performance, so white workers are in psychological equilibrium | |E) (M P w 5 (MP E ) w 5 (W) w.

(9.3)

Suppose there are two groups of black workers. Group 1 is composed of black employees who are in psychological balance since they form an expectation of their productivity in line with the judgment made by the employer | | E) 1 (M P b 5 (MP E ) b 5 (W) b.

(9.4)

206


Group 2 contains black workers who believe, accurately, that they are just as productive as white workers | | E) 2 [ (MP E ) b 5 (W) b ] , (M P b 5 (MP E ) w 5 (W) w

(9.5)

but, these workers are not in psychological balance since their pay is determined by the employers’ judgment of their productivity which falls short of their own assessment. All black workers at this firm are subject to statistical discrimination since the firm has formed an unjustifiably low expectation of black worker productivity. However, only the black workers in Group 2 recognize that they face statistical discrimination since Group 1 workers attribute the racial wage gap to a difference in productivity. Moreover, the workers in Group 2 experience cognitive dissonance, since their desire to be treated || fairly, (W) b 5 (MP E ) 2b, and their belief that they are not (see Equation 9.5 above), due to discrimination, are cognitions that do not match. Workers in this predicament may attempt to restore psychological balance by altering their effort on the job. Of course one way to eliminate this source of cognitive uncomfort would be to simply quit their job, but this reaction is often not pragmatic so we focus on alternative strategies below. A standard assumption of neoclassical economics is that a worker’s effort on the job, e, contributes to their productivity, MP(e), and that as ( (e) ) their effort rises so does their marginal product, 0 MP . 0. Worker effort 0 (e) is expected to depend on a number of factors including competition from the jobless and the extent of employer monitoring (Shapiro and Stiglitz, 1984), and pay relative to what they might expect to earn for comparable work with other employers (Akerlof, 1982).6 Following Akerlof we assert that a worker’s perception of how they are treated by their employer with respect to wages may influence their level of effort. However, our approach differs from Akerlof’s in two ways. First, the extent to which people alter their effort is governed by a desire to be in psychological equilibrium. Second, we allow effort to act as a signal of skills or as a mechanism for breaking down stereotypes about talent. In the next section we identify and describe three different strategies – retaliatory exertion or shirking, waiting and notification – a worker can adopt to restore psychological balance if they believe they face statistical discrimination. 9.2.3

Statistical Discrimination: Worker Strategies to Restore Cognitive Equilibrium

When a worker believes they are treated unfairly, leading them to experience cognitive dissonance, they can alter their level of effort in hopes of


207

restoring cognitive balance by creating a situation where the wage falls in line with their perceived productivity. However, the direction of the change in effort is expected to depend on their assumptions about the behavior of firms. Economists Farmer and Terrell (1996) developed a theory of racial wage differences based on the idea that employers underestimate the skills of minority workers, a form of statistical discrimination, but that over time through direct exposure to minority workers they learn their assumptions are false and update accordingly. In a recent paper Goldsmith et al. (2006b) find evidence consistent with the notion that employers undervalue the skills that minority workers bring to the job, but that as employers get to know minority workers they raise the value they apply to skills acquired with other firms. We formalize this perspective by assuming that as time, () t, advances employers learn, L, more about their workers, 0L0 (tt) . 0. If they come to realize that the racial beliefs or negative stereotypes (S) they hold about black worker productivity are false they will adjust those stereotypes 0S 0L and they will begin to break down or improve, 0L 0 (t) , 0. This learning, in turn, will lead to a more favorable view of black worker productivity, 0 (MPE) b 0S 0L 0S 0L 0 (t) . 0. Thus, black workers facing statistical discrimination could adopt a strategy of waiting to reach a psychological balance. The idea would be to simply wait until employers revise their judgment of black workers’ productivity until it conforms with the productivity of white workers, (MP Er) b 5 (MP E ) w . (MP E ) b, with whom they are just as productive. Presumably, as the employer revises their estimate of black worker productivity they will pay black employees a higher wage. Suppose black workers do not change the assessment of their performance on the job. Recall that the initial situation for black workers in Group 1 and in Group 2 are respectively | | E) 1 { (W) b 5 (MP E ) b } 5 (M P b , (W) w

(9.6)

| | E) 2 { (W) b 5 (MP E ) b } , (M P b 5 (W) w.

(9.7)

As a result of waiting and firms ultimately forming an accurate assessment of black workers the situation for black workers is now | | E) 1 [ (Wr) b 5 (MP Er) b ] 5 (W) w . (M P b

(9.6a)

| | E) 2 { (Wr) b 5 (MP Er) b } 5 (M P b 5 (W) w.

(9.7a)

Notice that black workers all received a wage increase resulting in them being paid a wage equivalent to that of white workers, and both groups of

208


black workers are now in psychological equilibrium. However, for black workers in Group 1, their new wage exceeds their own forecast of their productivity. Of course, waiting might be a lengthy process, in which case the workers who recognize they are facing discrimination, those in Group 2, will suffer cognitive discomfort for an extended period. If these workers believe that employers are slow to learn, or are only willing to partially adjust wages toward worker productivity, they may embrace an alternative strategy to establish psychological equilibrium. The black workers in Group 2 who understand that they face statistical discrimination may engage in a strategy of retaliatory exertion or shirking to eliminate the cognitive dissonance they experience and thus gain psychological balance. This strategy entails reducing their effort enough (er , e) to lower their assessment of their own marginal product to | | Er ( || (M P er , e)) 2b , (MP E (e)) 2b until it is in line with the employers’ perception of their performance level. However, the decline in effort must be below the threshold at which the employer discovers them, in which case they will fired. If Group 2 workers follow the shirking strategy then | | Er ( )) 2 { (W) b 5 (MP E ) b } 5 (M P er b , (W) w.

(9.7b)

In this case the workers gain psychological balance but the cost is twofold. First, their behavior ultimately verifies the employers’ initial prejudicial judgment that black workers are less productive than white workers, which reinforces their impulse to both hold false stereotypes and to practice statistical discrimination. Second, Group 2 workers will earn a wage that is unchanged and persistently lower than it would have been had they not faced discrimination or had they waited for employers to learn they were discriminating. Thus, they are only likely to adopt a strategy of reduced effort if they both find the cognitive costs of psychological disequilibrium to be large and sense that employers are slow to alter stereotypical beliefs. The shirking strategy results in black workers in both groups earning the same wage, as is the case if Group 2 workers adopt the waiting strategy to reach psychological balance. However, if they opt to shirk, then all black workers – those who perceive that they are being discriminated against and those who think they are treated fairly – will earn less than white workers. Faced with statistical discrimination, members of Group 2 may adopt behaviors to signal their differences from the stereotypical belief held by their employer. One strategy would be to exert such a high level of effort (es . e) that managers take notice or learn (00L (e) . 0) that the negative racial stereotype beliefs they hold are inaccurate (0 (0SL) , 0) which leads 0S 0L 7 them to reduce the extent of their negative perceptions 0L 0 (e) , 0. This


209

development, in turn, leads employersE to favorably reassess their judg( ) 2b 0S 0L ment of the workers’ productivity, 0 MP 0S 0L 0 (e) . 0. However, we assume that racial stereotypes only respond to an adjustment in effort that exceeds a threshold level e* considered extraordinary. We refer to this as the notification approach to restore psychological balance. Essentially, we assume that over time employers treat workers as heterogenous based on their effort level, but are only able to detect large changes in effort. The extraordinary level of effort associated with notification eliminates the employers’ negative stereotype regarding black workers in Group 2, making blacks in Group 2 equally as productive as white workers in the eyes of the firm. Moreover, the extraordinary level of effort put forth by workers in Group 2 also raises the employers’ assessment of their productivity relative to white workers, who put forth less effort. Thus, { (W s ) 2b 5 (MP Es (es . e*)) 2b } . (W) w.

(9.7c)

Given no alteration effort on the part of black workers in Group 1 the employer has no reason to alter their perception of the productivity of workers in this group or their wages so

{ (W) 1b 5 (MP E ) 1b } , { (W s ) 2b 5 (MP Es (es . e*)) 2b } .

The black workers in Group 2 are likely to increase their judgment of their own productivity given their extraordinary level of effort, | | Es ( || (M P es . e*)) 2b . (MP E ) 2b. However, the informed action strategy will place them in a psychological equilibrium so long as their new view of their level of productivity does not exceed the employers’ view. The strategy black workers adopt in their effort to overcome the cognitive dissonance caused by wage discrimination will influence wages between whites and blacks and between blacks based on their view of whether they face discrimination. Table 9.1 presents the predicted wage effects for each of the three strategies to yield psychological balance that are examined. If notification is adopted we expect a wage hierarchy resulting in (W) 1b , (W) w , (W) 2b. A shirking strategy produces a different outcome, (W) 1b 5 (W) 2b , (W) w, while waiting yields neither intra-group nor inter-group wage differences, (W) 1b 5 (W) 2b 5 (W) w. The next section describes the empirical procedures we use to estimate the link between workers’ perceptions that they face wage discrimination and the wages they receive. Our estimates are used to speculate on which, if any, of the strategies identified are adopted by those who believe they face discrimination.

210


Table 9.1

Expected wage impact when black workers in Group 2 believe they face statistical discrimination

Strategy for attaining psychological equilibrium Waiting Shirking Notification Reference Group

Intra-group Wages

Inter-group Wages

(w) 1b vs

(w) 1b vs (w) w

9.3.1

(w) 2b vs (w) w

(w) 1b 5 (w) 2b (w) 1b 5 (w) w (w) 2b 5 (w) 1b 5 (w) 2b (w) 1b , (w) w (w) 2b , 1 2 1 (w) b , (w) b (w) b , (w) w (w) 2b . Rows in Regression Tables Corresponding to Test of the Intra-group and Inter-group Wage Differences

Whites who believe they do Row 8 not face discrimination Whites who believe they do Row 8 not face discrimination

9.3

(w) 2b

Row 1

Row 6

Row 9

Row 7

(w) w (w) w (w) w

DATA AND METHODOLOGY Data

Data from the Multi City Study of Urban Inequality (MSCUI) are used in this study. The MCSUI is an interview-based survey of close to 9000 households and 2400 firms administered in the cities of Los Angeles, Boston, Atlanta and Detroit between 1992 and1994.8 MCSUI respondents included whites, blacks, Hispanics, Asians and persons coded as ‘other’. In conducting the Household Survey, from which we use data, attempts were made to ‘race match’, by assigning interviewers of a certain race or ethnicity to respondents of that same race/ethnicity. The MCSUI data are well suited for our study because participants were asked whether they believed they had faced workplace discrimination due to race in general and during hiring, or when promotion decisions were made. Exposure to promotion discrimination is captured by an affirmative response to the question: Have you ever felt at any time in the past that others at your place of employment got promotions or pay raises faster than you did because of your race? Contact with hiring discrimination is gauged by the question: Have you ever felt at any time in the past that you were refused a job because of your race? Respondents were also asked a more general question: During the past year were you discriminated


211

against at your work because of your race? We believe respondents interpreted this as meaning exposure to wage discrimination, since the question is asked in the section of the survey seeking information about wages and fringe benefits. We restrict the analysis to men ages 21–65 who were working, and who were not self-employed, when the MCSUI survey was conducted. Women and the elderly are excluded to minimize biases arising from selective labor force participation. We further restrict the MCSUI sample to blacks and whites to focus on black-white wage differences. Survey participants were asked to report their hourly wage. Persons who do not provide their wage are excluded from the sub-sample we analyse. In addition, workers who report an hourly wage below $2 or above $100 are considered outliers and are excluded. In the MCSUI data persons with reported earnings in excess of $100 000 are excluded as well. Moreover, we do not use data from Detroit since information from that metropolitan area was not collected on a number of variables contained in our study. Data on a rich array of socioeconomic and demographic factors are provided by the MCSUI including information on a person’s human capital, workplace characteristics if employed, the neighborhood where they reside at the time of the survey, and retrospective personal and family characteristics when the interviewee was a youth. Persons were excluded from our sample if they did not report information on the full set of variables used in our most fully specified wage equation. The MCSUI data we analyse (given the restrictions we impose) contains 948 observations, 513 whites and 435 blacks, when we estimate our preferred model specifications. Summary statistics Table 9.2 reports summary statistics for whites and for blacks for all of the variables used in our analysis. The data are reported in a series of panels which correspond to sets of variables that are recursively introduced to the empirical analysis. The panels provide information on wages and perceived exposure to discrimination, human capital, demographics, workplace characteristics, family and personal characteristics as a youth – referred as ‘pre-market factors’ – and current neighborhood characteristics. Variable definitions are presented in Appendix 9A.1 Table 9A.1.9 Mean hourly wages and perceptions of discrimination are reported in Panel A of Table 9.2. The typical black worker reports hourly pay of $12.61 while the average white respondent reports earning $15.94 per hour, a 21 percent difference. Blacks report much higher levels of perceived discrimination at work due to race than whites. Thirty percent of blacks, and 10 percent of whites, claim to have faced hiring discrimination. One in four

212

Table 9.2


Summary statistics for variables used in the econometric analysis: males

Variables

Hourly wage Workplace discrimination Schooling H.S. drop out High school

Community college Attend college College

Age Younger than 35 years of age Married

Union Work part-time Firm size/100 Atlanta

Mother high school graduate Father high school graduate

White (n = 513) 15.94 (7.73) 0.06 (0.24)

Black Variables (n = 435)

White (n = 513)

Black (n = 435)

Panel A Wages and Perceived Discrimination 12.61*** Promotion 0.06 (6.17) discrimination (0.24) 0.21*** Hiring 0.10 (0.41) discrimination (0.30)

0.25*** (0.44) 0.30*** (0.46)

14.64 (1.99) 0.03 (0.16) 0.36 (0.48)

13.84*** (2.16) 0.07*** (0.26) 0.53*** (0.50)

0.15 (0.35) 0.31 (0.46) 0.16 (0.37)

0.13 (0.34) 0.17*** (0.38) 0.09*** (0.29)

Panel B Human Capital Tenure Disability Did not complete H.S. by age 19 < 35 and Ave. H.S. grade ≤ C Self-esteem

6.53 (7.68) 0.12 (0.33) 0.46 (0.50)

5.57 (6.57) 0.14 (0.35) 0.34 (0.48)

0.07 (0.26) 3.34 (1.36)

0.06 (0.24) 3.32 (1.25)

37.64 (10.58) 0.57 (0.49) 0.61 (0.49)

Panel C Demographic Characteristics 35.71 Number of 0.60 (9.92) dependents (0.95) 0.49 Foreign Resident 0.05 (0.50) at age 16 (0.22) 0.52** (0.50)

0.23 (0.42) 0.09 (0.28) 0.58 (1.54) 0.14 (0.35)

Panel D Workplace Features and Location 0.30** Boston 0.38 (0.46) (0.49) 0.18 Los Angeles 0.48 (0.39) (0.50) 0.52 1994 0.46 (1.23) (0.50) 0.29*** (0.46)

0.77 (0.42) 0.72 (0.45)

Panel E Pre-market Factors 0.63*** Religious (0.48) attendance 0.50*** Jail as a youth (0.50)

0.28 (0.45) 0.14 (0.35)

0.86*** (1.25) 0.25** (0.44)

0.13** (0.34) 0.57 (0.50) 0.45 (0.50)

0.57*** (0.50) 0.12 (0.33)


Table 9.2

(continued)

Variables

White (n = 513)

Welfare as a youth Public housing as a youth Father raised Grandparent raised Good schools Good police

213

0.06 (0.24) 0.00 (0.06) 0.03 (0.17) 0.03 (0.16) 0.56 (0.50) 0.77 (0.42)

Black Variables (n = 435) 0.17*** (0.38) 0.04*** (0.20) 0.05 (0.21) 0.04*** (0.19)

Both parents raised Mother raised Other Raised

White (n = 513)

Black (n = 435)

0.82 (0.38) 0.13 (0.34) 0.01 (0.06)

0.67*** (0.47) 0.23*** (0.42) 0.03** (0.18)

Panel F Current Neighborhood Characteristics 0.41** Low crime 0.08 (0.49) (0.28) 0.49*** (0.50)

0.16*** (0.37)

Note: Weighted means are reported, with their standard errors in parentheses, for the sub-sample used to estimate Model 3 and Model 4. t-tests for differences in the means were conducted with *** Statistically significant at the 99% level, 95% level and 90% level identified by ***, ** and *, respectively. Source:

Multi City Survey of Urban Inequality (MCSUI).

blacks reports having been subject to promotion discrimination, but only 6 percent of whites feel they were treated unfairly regarding advancement. Moreover, 20 percent of the black respondents believe they faced wage discrimination while only 6 percent of whites hold this belief. Inspection of Table 9.2 reveals that on most variables there is substantial variation in mean values between whites and blacks. White workers have higher values on many variables that are known to contribute to wages, such as years of schooling and tenure. In addition, the typical white worker relative to the average black employee in the sample had more educated parents, was more likely to be raised in a two-parent family and was less likely to have been poor as a youth. Thus, casual inspection of the wage and characteristic data reported in Table 9.2 suggests that a portion of the higher wages earned by whites relative to blacks is due to having better productivity linked characteristics. Whether there is a link between perceived exposure to discrimination and wages, for black workers, and whether such a belief is associated with the racial wage gap is unclear. However, the theory set out in this chapter suggests that workers who believe they face discrimination may engage in strategies to reach psychological equilibrium, such as retaliatory exertion,

214


that may reduce the black-white pay gap. In the next section we conduct a rigorous and systematic examination of the link between perceptions of workplace discrimination and wages using regression analysis to determine whether a worker’s wage is influenced by their belief they have been subject to discrimination and if such a belief influences the relative wages of white and black workers after controlling for conventional wage determinants. 9.3.2

Methodology

We estimate reduced form wage equations using ordinary least squares to determine if within racial groups, and across racial groups, there is a difference in wages between those workers who believe they have been subject to workplace discrimination and employees who feel they have been treated fairly. The model we estimate is specified as follows ln wi 5 a 1 b (Blacki) 1 g (PerDisci) 1 y (Black*PerDisci) 1 l (Xi) 1 mi (9.8) where ln w is the log of the wage a worker receives on their job. Black and PerDisc are indicator variables that identify black employees and those workers who believe they have faced workplace discrimination. These indicators are interacted to determine if the impact of perceived exposure to discrimination varies for white and black workers. The vector X contains all of the other determinants of the wage rate. Equation (9.8) is estimated three different times, once for each of the three alternative measures of workplace discrimination, using data drawn from the MCSUI. White workers who believe they have not faced workplace discrimination are the reference category for the model. Given this wage specification we are able to estimate the effect on the wage rate of perceived exposure to discrimination (Wdis) and to being treated fairly (Wnodis) for white (w) and black (b) workers. Using these estimates we are able to construct a number of comparisons including those that shed light on how blacks that perceive they face discrimination nodis respond in an effort to attain psychological balance (W dis 5 b 2 Wb dis nodis nodis nodis g 1 y; W b 2 W w 5 b 1 g 1 y; and W b 2 W w 5 b) . We estimate a number of different versions of Equation (9.8). We begin our analysis by estimating a sparse OLS wage regression that contains only variables that indicate a person’s race, and their belief about whether they have faced wage discrimination (Model 1) and move to regressions that add controls for an individual’s skills and their socio-demographic characteristics (Model 2), and their work environment characteristics (Model 3) – yielding a garden variety wage equation.10 Then we augment


215

this conventional wage equation with family characteristics as a youth and current neighborhood descriptors (Model 4). Models 3 and 4 constitute our preferred model specifications, so our discussion of findings will focus on these models. We also estimate a model that extends Model 3 by adding controls for occupation of employment (Model 5). However, we recognize that a worker’s race may influence their assignment to a job or type of work in which case occupation is not exogenous. Thus, caution should be used when interpreting our results from Model 5. In the next section we present our estimates of equation 9.8 for model specifications 1–5.

9.4

EMPIRICAL RESULTS

Tables 9.3A, 9.3B, and 9.3C are summary tables that present our estimates of the impact of each of the three forms of perceived discrimination (workplace, hiring and promotion) on wages for reduced form log wage regressions. Our findings for Model specifications 1–5 are presented in these tables. Intra-group effects are reported in rows 2 and 8 while our intergroup findings are presented in rows 6 and 9. Our estimates of the racial wage gap when white and black workers report similar views on exposure to discrimination are set out in rows 1 and 7. Coefficient estimates for all of the variables included in the models are available from the authors upon request. Virtually all of the estimated coefficients have the expected sign and are highly significant at conventional levels. Appendix 9A.1 Table 9A.2 is a summary table that indicates which hypotheses our findings are consistent with for each of the measures of discrimination. Panel A of Appendix 9A.1 Table 9A.2 summarizes our findings when white workers who believe they have been treated fairly is the reference group. 9.4.1

Common Perceptions of Discrimination and the Racial Wage Gap

A racial wage gap suggests that there will be a racial happiness gap, ceteris paribus. We find no evidence of wage differences between black and white workers who believe they face discrimination (row 7) for each of our measures of discrimination (that is, W bdis 5 W dis w ). However, among workers who do not report exposure to discrimination (row 1), wages for black workers are 11.5 to 19 percent lower than for whites W nodis 5 W wnodis. Thus, b the racial wage gap, which is well documented, appears to be confined to those workers who think they are treated fairly. Before proceeding, it is important to note that evidence from Coleman et al. reveal that black employees grossly underestimate the degree to which they are subjected to discrimination while white employees vastly overestimate the degree to

216


Table 9.3A

The impact of race and perceived workplace discrimination on wages for males: summary table

Variables

Model 1 (n = 968)

Black (b) nodis (1) wbnodis 2 ww (2) PerDisc (g)

−0.287*** −0.154*** −0.164*** −0.109*** −0.134*** (0.036) (0.031) (0.030) (0.032) (0.031) −0.187* −0.154* −0.169** −0.142* −0.168** (0.103) (0.084) (0.082) (0.082) (0.081) 0.262** 0.158 0.167* 0.143 0.178* (0.119) (0.097) (0.095) (0.094) (0.094)

nodis wdis w 2 ww

(3) Black*PerDisc (Y) F-statistic for impact on wages (4) wb 2 ww (b 1 y [ PerDisc ]) (5) wdis 2 wnodis (g 1 y [ Black ]) nodis (6) wdis b 2 ww ( b 1 g 1 y) dis (7) wdis b 2 ww ( b 1 y) nodis (8) wdis b 2 wb ( g 1 y) nodis (9) wdis w 2 wb (b 2 g) F-statistic for the equation Adjusted R-squared Controls for: Human capital Demographics Workplace characteristics Family and neighborhood Occupation

Model 2 (n = 960)

Model 3 (n = 948)

−13.66*** −10.11*** −12.60*** [0.000] [0.002] [0.000] −0.05 0.00 0.00 [0.827] [0.968] [0.978] 1.64 0.01 −0.00 [0.201] [0.932] [0.961] −0.93 0.00 0.00 [0.335] [0.996] [0.954] −22.69*** 31.41*** 29.41*** [0.000] [0.000] [0.000] 0.06 0.39 0.41 yes yes

yes yes yes

Model 4 (n = 948)

Model 5 (n = 921)

−5.04** [0.025] 0.14 [0.706] 0.00 [0.987] 0.16 [0.690] 20.75*** [0.000] 0.43

−6.86*** [0.009] 0.24 [0.621] 0.04 [0.835] 0.17 [0.677] 25.78*** [0.000] 0.43

yes yes yes yes

yes yes yes yes

Notes: Estimating equation is ln wi 5 a 1 b (Blacki) 1 g (PerDisci) 1 y (Black*PerDisci) 1 l (Xi) 1 mi. Dependent variable: ln wage rate. Coefficient estimates using OLS are reported and standard errors are shown in parentheses. Variables for each set of controls are described in Appendix 9A.1 Table 9A.3. F-statistics and their associated p-values, shown in square brackets, are reported for tests of differences in the wage return for dark skinned and light skinned blacks. *** Statistically significant at the 99% level, 95% level and 90% level identified by ***, ** and *, respectively. Reference group is white and do not perceive discrimination. Per Disc = 1 if answer yes to ‘During the past year were you discriminated against at your work because of your race?’ Source:



Table 9.3B

217

The impact of race and perceived promotion discrimination on wages for males: summary table

Variables

Model 1 (n = 968)


−0.289*** −0.162*** −0.170*** −0.116*** −0.137*** (0.036) (0.031) (0.031) (0.033) (0.031) 0.028 −0.104 −0.112 −0.093 −0.114 (0.089) (0.073) (0.072) (0.071) (0.070) 0.087 0.143* 0.139 0.128 0.141* (0.105) (0.086) (0.085) (0.084) (0.083)

nodis wdis w 2 ww

(3) Black*PerDisc (Y) F-statistic for impact on wages (4) wb 2 ww (b 1 y [ PerDisc ]) (5) wdis 2 wnodis (g 1 y [ Black ]) nodis (6) wdis b 2 ww ( b 1 g 1 y) dis (7) wdis b 2 ww ( b 1 y) nodis (8) wdis b 2 wb ( g 1 y) (9) wwdis 2 wnodis b (b 2 g) F-statistic for the equation Adjusted R-squared Controls for: Human capital Demographics Workplace characteristics Family and neighborhood Occupation

−10.46*** [0.001] −4.13** [0.042] 4.28** [0.039] −12.16*** [0.001] −22.50*** [0.000] 0.06

Model 2 (n = 960)

Model 3 (n = 948)

−7.55*** −10.56*** [0.006] [0.001] −0.05 −0.15 [0.821] [0.703] 0.74 0.35 [0.338] [0.557] −0.60 −0.60 [0.439] [0.439] 31.36*** 29.30*** [0.000] [0.000] 0.39 0.41 yes yes

yes yes yes

Model 4 (n = 948)

−3.16* [0.076] 0.02 [0.882] 0.59 [0.443] −0.10 [0.755] 20.72*** [0.000] 0.43 yes yes yes yes

Model 5 (n = 921)

−6.24** [0.013] 0.00 [0.956] 0.35 [0.557] −0.09 [0.759] 25.70*** [0.000] 0.43 yes yes yes yes

Notes: Estimating equation is ln wi 5 a 1 b (Blacki) 1 g (PerDisci) 1 y (Black*PerDisci) 1 l (Xi) 1 mi. Dependent variable: ln wage rate. Coefficient estimates using OLS are reported and standard errors are shown in parentheses. Variables for each set of controls are described in Appendix 9A.1 Table 9A.3. F-statistics and their associated p-values, shown in square brackets, are reported for tests of differences in the wage return for dark skinned and light skinned blacks. *** Statistically significant at the 99% level, 95% level and 90% level identified by ***, ** and *, respectively. Reference group is white and do not perceive discrimination. Per Disc = 1 if answer yes to ‘Have you ever felt at any time in the past that others at your place of employment got promotions or pay raises faster than you did because of your race?’ Source:


218


Table 9.3C

The impact of race and perceived hiring discrimination on wages for males: summary table

Variables

Model 1 (n = 968)

Model 2 (n = 960)


−0.304*** (0.037) −0.088 (0.077) 0.202** (0.093)

−0.180*** −0.189*** −0.136*** (0.032) (0.032) (0.034) −0.087 −0.093 −0.090 (0.063) (0.062) (0.061) 0.171** 0.172** 0.176** (0.076) (0.075) (0.074)

−0.151*** (0.032) −0.066 (0.062) 0.134* (0.075)

−5.01** [0.025] −0.02 [0.898] 3.27* [0.071] −1.96 [0.162] 31.58*** [0.000] 0.39

−6.80*** [0.009] −0.07 [0.797] 2.90* [0.089] −2.17 [0.141] 29.50*** [0.000] 0.41

−1.32 [0.251] 0.32 [0.570] 3.54* [0.060] −0.51 [0.476] 20.91*** [0.000] 0.43

−3.81* [0.051] −0.06 [0.805] 2.20 [0.138] −1.70 [0.192] 25.73*** [0.000] 0.43

yes yes

yes yes yes

yes yes yes yes

yes yes yes

nodis wdis w 2 ww

(3) Black*PerDisc (y) F-statistic for impact on wages (4) wb 2 ww (b 1 y [ PerDisc ]) (5) wdis 2 wnodis (g 1 y [ Black ]) nodis (6) wdis b 2 ww ( b 1 g 1 y) dis (7) wdis b 2 ww ( b 1 y) nodis (8) wdis b 2 wb ( g 1 y) nodis (9) wdis w 2 wb (b 2 g) F-statistic for the equation Adjusted R-squared Controls for: Human capital Demographics Workplace characteristics Family and neighborhood Occupation

−14.30*** [0.000] −1.42 [0.234] 4.69** [0.031] −7.58*** [0.006] 23.07*** [0.000] 0.06

Model 3 (n = 948)

Model 4 (n = 948)

Model 5 (n = 921)

yes

Notes: Estimating equation is ln wi 5 a 1 b (Blacki) 1 g (PerDisci) 1 y (Black*PerDisci) 1 l (Xi) 1 mi. Dependent variable: ln wage rate. Coefficient estimates using OLS are reported and standard errors are shown in parentheses. Variables for each set of controls are described in Appendix 9A.1 Table 9A.3. F-statistics and their associated p-values, shown in square brackets, are reported for tests of differences in the wage return for dark skinned and light skinned blacks. *** Statistically significant at the 99% level, 95% level and 90% level identified by ***, ** and *, respectively. Reference group is white and do not perceive discrimination. Per Disc = 1 if answer yes to ‘Have you ever felt at any time in the past that you were refused a job because of your race?’ Source:



219

which they are subjected to discrimination. We assume that white workers who believe they are treated fairly are the reference group for black workers when answering questions about exposure to discrimination. We now compare the wages of blacks who believe they are treated fairly (Group 1), and those who feel they face discrimination (Group 2), against the reference group and each other to shed light on how blacks who believe they face discrimination respond. 9.4.2 Evidence on the Waiting, Shirking, and Notification Hypotheses; White Workers Who Do Not Face Discrimination as the Reference Group We find no significant wage difference between black workers who believe they face discrimination and those who think they are treated fairly (row 8) for workplace and promotion discrimination, which is consistent with both the shirking and waiting hypotheses. But, we also find that blacks who believe they face promotion or workplace discrimination (row 6) and those think they are treated fairly (row 1) earn significantly lower wages than whites who do not believe they face discrimination, both of which are predicted by the shirking response to perceived discrimination. Thus, the evidence points toward a downward adjustment of effort on the part of black workers who believe they are discriminated against so that their contribution to the firm’s revenue corresponds, rather than exceeds, their hourly pay. A striking finding is that black workers who believe they are discriminated against earn wages that are about 8 percent higher than black employees who think they are not exposed to discrimination, and this difference is statistically significant. This outcome is predicted by the notification response or extraordinary effort to overcome the adverse stereotypes that the workers think they face. The notification strategy also predicts that white workers will earn more than black workers who believe they are treated fairly, and we find that as well. However, white workers earn substantially more than the black workers who identify themselves as facing hiring discrimination. This later finding suggests that although employers seem to recognize the greater effort and productivity of the workers who believe they face discrimination, they still do not think they are as productive as white workers, or they are slowly adjusting the wages of black workers who adopt the notification strategy. 9.4.3

Is There a Wage Hierarchy?

If relative wages influence happiness, then a wage gradient or hierarchy is consistent with a happiness gradient. Our estimate reveal that W dis w

220


is less than W wnodis by 16–19 percent (row 1) based on Model 3 for the various forms of perceived discrimination. As noted earlier, we also find that blacks who believe they face discrimination have significantly lower wages, on the order of 16 percent, than white workers who believe they nodis do not face discrimination (row 6, W dis ). In addition, we find b 2 Ww statistically equivalent wages for white workers who think they are treated unfairly and black workers who believe they are treated fairly (row 9, nodis W dis ). Thus, we find evidence of a wage hierarchy where whites w 5 Wb who sense they are treated fairly earn more than whites who believe they are discriminated against and black workers, regardless of their view of how they are treated (unless the perceived source of discrimination is dis hiring); W wnodis . W nodis < W dis b w < W b . Thus, it is likely that the level of happiness is greater for whites who believe they are treated fairly, and happiness is roughly equivalent for other black and white workers.

9.5

CONCLUDING REMARKS

There is a substantial literature that finds a linkage between happiness and relative economic well-being as measured by earnings or wages. There is also a well-documented racial gap in wages. One explanation for this is disparate treatment or discrimination. Many black workers report perceiving that they face workplace discrimination in general and with respect to specific events such as hiring and promotion. This chapter explores how such workers respond to these feelings, under the assumption that perceived exposure to discrimination causes psychological discomfort that workers seek to eliminate. We identify three alternative strategies for attaining psychological equilibrium when facing discrimination – waiting, shirking and notification – each of which generates different predictions for intra-racial and inter-racial wage differences. Using data drawn from the MCSUI, which contains information on perceptions of discrimination, we derive estimates of the relative wage effects of perceived discrimination to shed light on which hypotheses are consistent with the data. We find evidence of a wage hierarchy with whites workers who believe they are treated fairly at the top and all other workers falling behind by an equivalent amount. Our evidence suggests that when confronted with hiring discrimination, black workers appear to give greater effort to overcome this hurdle, as predicted by the notification hypothesis. However, when black workers sense that they face promotion or workplace discrimination, the evidence is consistent with their reducing effort to bring their output level down, and hence in line with their pay level. A number of questions remain to be explored including: Are there


221

systematic differences in the characteristics of those who report believing they face discrimination and those who believe they are treated fairly? Does the skin shade of blacks who believe they are subject to discrimination, or the amount of time these individuals have spent with their current employer, influence the strategy they adopt to reach psychological balance?

NOTES *

1.

2.

3.

4.

5. 6. 7. 8. 9.

10.

This chapter is part of an on-going reserch effort involving the author, William Darity Jr. (Arts & Sciences Professor of Public Policy Studies and Professor of African American Studies and Economics at Duke University), and Darrick Hamilton (Assistant Professor at Milano – The New School for Management and Urban Policy, The New School). However, smaller wage differences between black and white workers, on the order of 7 percent, have been reported by Altonji and Blank (1999) and Neal and Johnson (1996) when only pre-market controls and the Air Force Qualifying Test (AFQT) are included as a wage regressor. Economists previously have used cognitive dissonance theory to explain economic development (Hirschman, 1965), the accumulation of debt (Maital, 1982), job choice over safe and hazardous employment (Akerlof and Dickens, 1982) and labor supply (Goldsmith et al. 2004). Akerlof and Dickens (1982) and Goldsmith et al. (2004) formally merge neoclassical theory and cognitive dissonance theory rather than present cognitive dissonance theory as an alternative explanation for behavior. See Earl (1992) for a review of the literature in which economists make use of cognitive dissonance theory, and Earl and Wicklund (1999) for a brief discussion of rational decision making and cognitive dissonance. There is a substantial body of empirical research showing that people who behave in different ways also vary predictably in their attitudes, which are thoughts or cognitions. For a review of the empirical literature on the relationship between attitudes and behavior, see Ajzen and Fishbein (1980). Another form of statistical discrimination is when perceived group characteristics, held by an employer, are applied to an individual. Thus, if an employer believes that people with poorer quality schooling make less satisfactory workers, and that blacks on average possess inferior schooling, then they may judge a black worker on the basis of this negative racial stereotype, S, rather than their own background – instead of putting forth the time and effort to accurately assess the quality of their schooling. See Fiske and Ruscher (1993) for a review of the literature in social psychology on negative stereotyping and minority group status. According to Akerlof workers have an impulse to increase effort due to the pleasure associated with being overpaid. Psychologist Brehm (1966) describes such a behavioral alteration as reactance. All Detroit and Atlanta respondents were interviewed in 1992 and 1993 respectively, while participants residing in Boston and Los Angeles were interviewed in either 1993 or 1994. A concern is whether our sub-sample of employed persons who meet our restrictions differs markedly from the sub-sample of persons capable of working. Comparison of the means, for those variables that do not describe features of work, reveal little difference between those capable of work and those actually employed (table available upon request). Beginning with Model 3 the vector X contains city indicators for Los Angeles and

222

Happiness, economics and politics Boston, with Atlanta (and hence 1992), serving as the reference category. A dummy variable is also included to identify if the respondents residing in Los Angeles and Boston were interviewed in 1993 or 1994.

REFERENCES Akerlof, George and W.T. Dickens (1982), ‘The economic consequences of cognitive dissonance’, American Economic Review, 72 (3), 307–19. Altonji, Joseph G. and Rebecca Blank (1999), ‘Race and gender in the labor market’, in Orley Ashenfelter and David Card (eds), Handbook of Labor Economics, Vol. 3A, Amsterdam: North Holland, pp. 3143–260. Ajzen, I. and M. Fishbein (1980), Understanding Attitudes and Predicting Social Behavior, New Jersey: Prentice-Hall. Arrow, Ken (1973), ‘The theory of discrimination’, in O. Ashenfelter and A. Rees (eds), Discrimination in Labor Markets, Princeton, NJ: Princeton University Press, pp. 3–33. Brehm, J.W. (1966), A Theory of Psychological Reactance, New York: Academic Press. Carver, Charles S. and Michael F. Scheier (1981), Attention and Self-Regulation: A Control-Theory Approach to Human Behavior, New York: Springer-Verlag. Coate, S. and Glen C. Loury (1993), ‘Will affirmative-action policies eliminate negative stereotypes?’, American Economic Review, 83 (5), 1220–40. Coleman, Major, William Darity Jr and Rhonda Sharpe (2008), ‘Are reports of discrimination valid? Considering the moral hazard effect’, American Journal of Economics and Sociology, 67 (2), 149–76. Couch, Kenneth and Mary C. Daly (2002), ‘Black-white wage inequality in the 1990s: a decade of progress’, Economic Inquiry, 40 (1), 31–41. Darity, William A., and Patrick L. Mason (1998), ‘Evidence on discrimination in employment: codes of color, codes of gender’, Journal of Economic Perspectives, 12 (2), 63–90. Earl, Peter E. (1992), ‘On the complementarity of economic applications of cognitive dissonance theory and personal construct psychology’, in S. Lea, P. Webley and B. Young (eds), New Directions in Economic Psychology, Brookfield, VT: Edward Elgar Publishing, pp. 49–65. Earl, Peter E. and Robert A. Wicklund (1999), ‘Cognitive dissonance’, in Peter E. Earl and Simon Kemp (eds), The Elgar Companion to Consumer Research and Economic Psychology, Cheltenham: Edward Elgar Publishing, pp. 81–8. Farmer, Amy and Dek Terrell (1996), ‘Discrimination, Bayesian updating of employer beliefs, and human capital accumulation’, Economic Inquiry, 34, 204–19. Festinger, Leon (1957), A Theory of Cognitive Dissonance, Palo Alto, CA: Stanford University Press. Fiske, Susan T. and Janet B. Ruscher (1993), ‘Negative interdependence and prejudice: whence the affect?’, in Diane M. Mackie and David Lewis Hamilton (eds), Affect, Cognition, and Stereotyping: Interactive Processes in Group Perception, San Diego, CA: Academic Press, pp. 239–68. Fiske, Susan T., Amy Cuddy, Peter Glick and Jun Xu (2002), ‘A model of (often mixed) stereotype content: competence and warmth respectively follow from


223

perceived status and competition’, Journal of Personality & Social Psychology, 82, 878–902. Goldsmith, Arthur H., Stanley Sedo, William Darity Jr and Darrick Hamilton (2004), ‘The labor supply consequences of perceptions of employer discrimination during job search and on-the-job: integrating neoclassical theory and cognitive dissonance?’, Journal of Economic Psychology, 25 (1), 15–39. Goldsmith, Arthur H., Darrick Hamilton and William Darity Jr (2006), ‘Does a foot-in-the-door matter? White-nonwhite differences in the wage return to tenure and prior workplace experience’, Southern Economics Journal. Goldsmith, Arthur H., Darrick Hamilton and William Darity, Jr (2007), ‘From dark to light: skin color and wages among African-Americans’, Journal of Human Resources, XLII (4), 701–38. Hallock, Kevin, F., Wallace Hendricks and Emer Broadbent (1998), ‘Discrimination by gender and disability status: do worker perceptions match statistical measures?’, Southern Economic Journal, 65 (2), 245–63. Hirschman, A.O. (1965), ‘Obstacles to development: a classification and a quasivanishing act’, Economic Development and Cultural Change, 13, 385–93. Kuhn, Peter J. (1987), ‘Sex discrimination in labor markets: the role of statistical evidence’, American Economic Review, 77 (4), 567–83. Lundberg, S.J. and R. Startz (1983), ‘Private discrimination and social intervention in competitive labor markets’, American Economic Review, 73 (3), 340–7. Maital, S. (1982), Minds, Markets and Money: Psychological Foundations of Economic Behavior, New York: Basic Books. Mason, Patrick L. (1997), ‘Race, culture, and skill: interracial wage differences among African Americans, Latinos, and Whites’, Review of Black Political Economy, 25 (3), 5–39. Neal, Derek A. and William R. Johnson (1996), ‘The role of premarket factors in black–white wage differences’, The Journal of Political Economy, 104 (5), 869–95. Neumark, David, and Michele McLennan (1995), ‘Sex discrimination and women’s labor market outcomes’, Journal of Human Resources, 30 (4), 713–40. Shapiro, Carl and Joseph E. Stiglitz (1984), ‘Equilibrium unemployment as a worker discipline device’, American Economic Review, 74 (3), 433–44.

224


APPENDIX: 9A.1 Table 9A.1

Definition of variables

Variables

Variable Definitions

Variables


W

Respondent’s hourly wage at survey date

Disability

White

1 if respondent is white, 0 otherwise

Foreign resident at 16 years of age

Workplace discrimination

1 if respondent believes they faced workplace discrimination, 0 otherwise 1 if respondent believes they faced promotion discrimination, 0 otherwise 1 if respondent believes they faced hiring discrimination, 0 otherwise Respondent’s age at survey date

Los Angeles

1 if respondent has a work limiting health condition, 0 otherwise 1 if respondent was primarily a foreign resident before 16 years of age, 0 otherwise 1 if respondent resides in Los Angeles, 0 otherwise

1 if respondent is younger than 35 years old, 0 otherwise Years of schooling completed at survey date Number of years employed by current employer at survey date

Work part-time

Promotion discrimination

Hiring discrimination

Age

Younger than 35

Schooling

Tenure

Atlanta

1 if respondent resides in Atlanta, 0 otherwise

Boston

1 if respondent resides in Boston, 0 otherwise

Union

1 if respondent is a union member, 0 otherwise 1 if respondent works part-time, 0 otherwise

Firm size

Mother education

Number of workers at respondent’s firm per 1000 workers 1 if respondent’s mother completed at least 12 years of formal schooling, 0 otherwise


Table 9A.1

225

(continued)

Variables


Variables


H.S. drop out

1 if respondent failed to complete high school, 0 otherwise

Father education

High school

1 if respondent’s highest level of schooling is completion of high school, 0 otherwise 1 if respondent’s highest level of schooling is completion of community college, 0 otherwise 1 if respondents highest level of schooling was attending college, 0 otherwise 1 if respondent completed college school, 0 otherwise

Both parents raised

1 if respondent’s father completed at least 12 years of formal schooling, 0 otherwise 1 if lived with mother and father to age 16, 0 otherwise

Community college

Attend college

College

< 35 & Ave. H.S. grade ≤ C

No. H.S. by 19 years of age

Self-esteem

1 if respondent is

Happiness, Economics and Politics: Towards a Multi-Disciplinary Approach

Amblyopia: A Multidisciplinary Approach

Amblyopia: A Multidisciplinary Approach

Hyperactivity Disorder: A Multidisciplinary Approach

Graves' Orbitopathy: A Multidisciplinary Approach

Head Injury: A Multidisciplinary Approach

Brain Metastasis: A Multidisciplinary Approach

Mantle plumes: a multidisciplinary approach

Brain Metastasis: A Multidisciplinary Approach

Paediatric and Adolescent Gynaecology: A Multidisciplinary Approach

Nutrition and behavior: a multidisciplinary approach

Happiness, Economics and Public Policy

Pelvic Floor Dysfunction - A Multidisciplinary Approach

Gastrointestinal Oncology: A Critical Multidisciplinary Team Approach

Climate Change: A Multidisciplinary Approach, 2nd Edition

Attention deficit hyperactivity disorder: a multidisciplinary approach

Deep Pelvic Endometriosis: A Multidisciplinary Approach

Applied Mathematical Modeling: A Multidisciplinary Approach

Climate Change: A Multidisciplinary Approach, 2nd edition

Allergic skin disease: a multidisciplinary approach

Debugging by Thinking: A Multidisciplinary Approach

Head Injury: A Multidisciplinary Approach (Cambridge Medicine)

Younger People With Dementia: A Multidisciplinary Approach

Happiness: A revolution in economics (Munich Lectures)

Construction Economics: A New Approach

Panama: Politics and Economics (Politics and Economics of Latin America)

Happiness Quantified: A Satisfaction Calculus Approach

Construction Economics: A New Approach

Business Economics: A Contemporary Approach

Knowledge Entanglements: An International and Multidisciplinary Approach

Happiness, Economics and Politics: Towards a Multi-Disciplinary Approach

Amblyopia: A Multidisciplinary Approach

Amblyopia: A Multidisciplinary Approach

Hyperactivity Disorder: A Multidisciplinary Approach

Graves' Orbitopathy: A Multidisciplinary Approach

Head Injury: A Multidisciplinary Approach

Brain Metastasis: A Multidisciplinary Approach

Mantle plumes: a multidisciplinary approach

Brain Metastasis: A Multidisciplinary Approach

Paediatric and Adolescent Gynaecology: A Multidisciplinary Approach

Nutrition and behavior: a multidisciplinary approach

Happiness, Economics and Public Policy

Pelvic Floor Dysfunction - A Multidisciplinary Approach

Gastrointestinal Oncology: A Critical Multidisciplinary Team Approach

Climate Change: A Multidisciplinary Approach, 2nd Edition

Attention deficit hyperactivity disorder: a multidisciplinary approach

Deep Pelvic Endometriosis: A Multidisciplinary Approach

Applied Mathematical Modeling: A Multidisciplinary Approach

Climate Change: A Multidisciplinary Approach, 2nd edition

Allergic skin disease: a multidisciplinary approach

Debugging by Thinking: A Multidisciplinary Approach

Head Injury: A Multidisciplinary Approach (Cambridge Medicine)

Younger People With Dementia: A Multidisciplinary Approach

Happiness: A revolution in economics (Munich Lectures)

Construction Economics: A New Approach

Panama: Politics and Economics (Politics and Economics of Latin America)

Happiness Quantified: A Satisfaction Calculus Approach

Construction Economics: A New Approach

Business Economics: A Contemporary Approach

Knowledge Entanglements: An International and Multidisciplinary Approach

Recommend Documents